Haematologica, Volume 107, Issue 5 by Haematologica

haematologica Journal of the Ferrata Storti Foundation

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 Journal of the Ferrata Storti Foundation

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 Journal of the Ferrata Storti Foundation

Table of Contents Volume 107, Issue 5: May 2022 About the Cover 1017

Images from the Haematologica Atlas of Hematologic Cytology: acute myeloid leukemia with t(8;21)(q22;q22.1); RUNX1-RUNX1T1 Rosangela Invernizzi

https://doi.org/10.3324/haematol.2022.280829

Landmark Paper in Hematology 1018

Proteasome inhibition: the dawn of novel therapies in multiple myeloma Monika Engelhardt, Johannes Moritz Waldschmidt and Ralph Wäsch https://doi.org/10.3324/haematol.2022.280857

Editorials 1020

Immune thrombocytopenia: vaccination does not equal causation Allyson Pishko and Adam Cuker

https://doi.org/10.3324/haematol.2021.279727

1022

For older adults with hematologic malignancies, a comprehensive geriatric assessment matters Raul Cordoba

https://doi.org/10.3324/haematol.2021.279927

1024

Stem cell transplant for lymphoma - never too late? Amanda F. Cashen and Nancy L. Bartlett https://doi.org/10.3324/haematol.2021.280519

Articles Acute Lymphoblastic Leukemia 1026 Clofarabine increases the eradication of minimal residual disease of primary B-precursor acute lymphoblastic leukemia compared to high-dose cytarabine without improvement of outcome. Results from the randomized clinical trial 08-09 of the Cooperative Acute Lymphoblastic Leukemia Study Group. Gabriele Escherich et al.

https://doi.org/10.3324/haematol.2021.279357

Acute Myeloid Leukemia 1034 Clinical and molecular relevance of genetic variants in the non-coding transcriptome of patients with cytogenetically normal acute myeloid leukemia Dimitrios Papaioannou et al.

https://doi.org/10.3324/haematol.2020.266643

Bone Marrow Transplantation 1045 One and a half million hematopoietic stem cell transplants: continuous and differential improvement in worldwide access with the use of non-identical family donors Dietger Niederwieser et al.

https://doi.org/10.3324/haematol.2021.279189

Haematologica 2022; vol. 107 no. 5 - May 2022 http://www.haematologica.org/

haematologica Journal of the Ferrata Storti Foundation

Cell Therapy & Immunotherapy 1054 Improved outcome of patients with graft-versus-host disease after allogeneic hematopoietic cell transplantation for hematologic malignancies over time: an EBMT mega-file study Hildegard T. Greinix et al.

https://doi.org/10.3324/haematol.2020.265769

Coagulation & its Disorders 1064 A homozygous duplication of the FGG exon 8-intron 8 junction causes congenital afibrinogenemia. Lessons learned from the study of a large consanguineous Turkish family Michel Guipponi et al.

https://doi.org/10.3324/haematol.2021.278945

Hemostasis 1072 Germline GATA2 variant disrupting endothelial eNOS function and angiogenesis can be restored by c-Jun/AP-1 upregulation Giulio Purgatorio et al.

https://doi.org/10.3324/haematol.2021.278450

Hodgkin Lymphoma 1086 Older patients (aged ≥60 years) with previously untreated advanced-stage classical Hodgkin lymphoma: a detailed analysis from the phase III ECHELON-1 study Andrew M. Evens et al.

https://doi.org/10.3324/haematol.2021.278438

Immunological Dysregulation 1095 Complement dysregulation is associated with severe COVID-19 illness Jia Yu et al.

https://doi.org/10.3324/haematol.2021.279155

Myeloproliferative Disorders 1106 Hemorrhage in patients with polycythemia vera receiving aspirin with an anticoagulant: a prospective, observational study Jeffrey I. Zwicker et al.

https://doi.org/10.3324/haematol.2021.279032

Non-Hodgkin Lymphoma 1111 Toxicity and efficacy of chimeric antigen receptor T-cell therapy in patients with diffuse large B-cell lymphoma above the age of 70 years compared to younger patients – a matched control multicenter cohort study Ron Ram et al.

https://doi.org/10.3324/haematol.2021.278288

1119

Response and resistance to CDK12 inhibition in aggressive B-cell lymphomas Jing Gao et al.

https://doi.org/10.3324/haematol.2021.278743

1131

Characterization of GECPAR, a noncoding RNA that regulates the transcriptional program of diffuse large B-cell lymphoma Sara Napoli et al.

https://doi.org/10.3324/haematol.2020.267096

1144

Prephase rituximab/prednisone therapy and aging-related, proinflammatory cytokine milieu in older, vulnerable patients with newly diagnosed diffuse large B-cell lymphoma Richard J. Lin et al.

https://doi.org/10.3324/haematol.2021.278719

Haematologica 2022; vol. 107 no. 5 - May 2022 http://www.haematologica.org/

haematologica Journal of the Ferrata Storti Foundation

1153

Dose-adjusted EPOCH and rituximab for the treatment of double expressor and double-hit diffuse large B-cell lymphoma: impact of TP53 mutations on clinical outcome Anna Dodero et al.

https://doi.org/10.3324/haematol.2021.278638

Plasma Cell Disorders 1163 Natural history of Waldenström macroglobulinemia following acquired resistance to ibrutinib monotherapy Joshua N. Gustine et al.

https://doi.org/10.3324/haematol.2021.279112

Quality of Life 1172 Randomized controlled trial of geriatric consultation versus standard care in older adults with hematologic malignancies Clark DuMontier et al.

https://doi.org/10.3324/haematol.2021.278802

Letters to the Editor 1181

BNT162b2 COVID-19 and ChAdOx1 nCoV-19 vaccination in patients with myelodysplastic syndromes Sultan Abdul-Jawad et al.

https://doi.org/10.3324/haematol.2021.280337

1185

A report from the Leukemia Electronic Abstraction of Records Network on risk of hepatotoxicity during pediatric acute lymphoblastic leukemia treatment Joanna S. Yi et al.

https://doi.org/10.3324/haematol.2021.279805

1189

SF3B1-mutant myelodysplastic syndrome/myeloproliferative neoplasms: a unique molecular and prognostic entity Abhishek A. Mangaonkar et al.

https://doi.org/10.3324/haematol.2021.280463

1193

Immune thrombocytopenia following vaccination during the COVID-19 pandemic Philip Young-Ill Choi et al.

https://doi.org/10.3324/haematol.2021.279442

1197

Impaired in vivo activated protein C response rates indicate a thrombophilic phenotype in inherited thrombophilia Sara Reda et al.

https://doi.org/10.3324/haematol.2021.280573

1201

Erythrocytosis associated with EPAS1(HIF2A), EGLN1(PHD2), VHL, EPOR or BPGM mutations: the Mayo Clinic experience Naseema Gangat et al.

https://doi.org/10.3324/haematol.2021.280516

1205

Outstanding outcomes in infants with KMT2A-germline acute lymphoblastic leukemia treated with chemotherapy alone: results of the Children’s Oncology Group AALL0631 trial Erin M. Guest et al.

https://doi.org/10.3324/haematol.2021.280146

1209

Mitochondrial ATP generation in stimulated platelets is essential for granule secretion but dispensable for aggregation and procoagulant activity Paresh P. Kulkarni et al.

https://doi.org/10.3324/haematol.2021.279847

1214

Autologous hematopoietic cell transplantation in diffuse large B-cell lymphoma after three or more lines of prior therapy: evidence of durable benefit Matthew Mei et al.

https://doi.org/10.3324/haematol.2021.279999

Haematologica 2022; vol. 107 no. 5 - May 2022 http://www.haematologica.org/

haematologica Journal of the Ferrata Storti Foundation

Comments 1218

Serological response following anti-SARS-CoV-2 vaccination in hematopoietic stem cell transplantation patients depends upon time from transplant, type of transplant and “booster”dose Immacolata Attolico et al.

https://doi.org/10.3324/haematol.2022.280619

Case Reports 1219

Platelet-activating anti-PF4 antibodies mimic VITT antibodies in an unvaccinated patient with monoclonal gammopathy Andreas Greinacher et al.

https://doi.org/10.3324/haematol.2021.280366

1222

Kikuchi-Fujimoto disease associated with hemophagocytic lymphohistiocytosis following the BNT162b2 mRNA COVID-19 vaccination Giovanni Caocci et al.

https://doi.org/10.3324/haematol.2021.280239

1226

Zanubrutinib, rituximab and lenalidomide induces deep and durable remission in TP53-mutated B-cell prolymphocytic leukemia: a case report and literature review Lijie Xing et al.

https://doi.org/10.3324/haematol.2021.280259

Haematologica 2022; vol. 107 no. 5 - May 2022 http://www.haematologica.org/

ABOUT THE COVER Images from the Haematologica Atlas of Hematologic Cytology: acute myeloid leukemia with t(8;21)(q22;q22.1); RUNX1-RUNX1T1 Rosangela Invernizzi University of Pavia, Pavia, Italy E-mail: ROSANGELA INVERNIZZI - rosangela.invernizzi@unipv.it doi:10.3324/haematol.2022.280829

n the World Health Organization classification of acute myeloid leukemia (AML) specific categories are recognized on the basis of cytogenetic findings and cellular morphology with relevant prognostic significance. Distinctive morphological features of AML with t(8;21)(q22;q22.1); RUNX1-RUNX1T1 are illustrated in the Figure showing bone marrow smears. Granuloblastic hyperplasia with increased blast percentage is evident. Blasts are large, with eccentric, often indented or cleaved nucleus, basophilic cytoplasm with a clear area at the nuclear indentation and sometimes many azurophilic granules (A). Maturing myeloid cells show asynchronous nuclear:cytoplasmic maturation with open chromatin and prominent nucleoli also in the presence of many secondary cytoplasmic granules (B). In panel (C), in the center, note a blast with abundant cytoplasm, peripheral basophilia, a prominent Golgi zone, and an Auer rod. In panel (D), a neutrophil shows a very long slender Auer rod, demonstrating its derivation from a leukemic blast. Note also the neutrophil degranulation and abnormal nuclear segmentation. Blasts are not stained by periodic acid Schiff (E) but show positivity for Sudan black, sometimes in a restricted area at the nuclear indentation (F). This AML subtype is characterized by a relatively favorable outcome.1 Disclosures No conflicts of interest to disclose

Reference 1. Invernizzi R. Acute myeloid leukemia and related precursor neoplasms. Haematologica. 2020; 105(Suppl. 1):98-121.

hhaematologica | 2022; 107(5)

1017

LANDMARK PAPER IN HEMATOLOGY Proteasome inhibition: the dawn of novel therapies in multiple myeloma Monika Engelhardt, Johannes Moritz Waldschmidt and Ralph Wäsch Department of Internal Medicine I, Faculty of Medicine and Medical Center, University of Freiburg, Hugstetterstr. Freiburg, Germany. E-mail: monika.engelhardt@uniklinik-freiburg.de doi:10.3324/haematol.2022.280857 TITLE

Bortezomib or high-dose dexamethasone for relapsed multiple myeloma.

AUTHORS

Richardson PG, Sonneveld P, Schuster MW, et al.

JOURNAL

New England Journal of Medicine 2005;352(24):2487-2498. PMID: 15958804

owadays for multiple myeloma (MM) experts it is well established that the prognosis of this often evil disease has improved. Numerous novel anti-MM agents have shown impressive activity, inducing longevity or chronic disease courses. The currently available anti-MM drug options include proteasome inhibitors (PI), immunomodulatory agents (IMiD), immunotherapies, such as monoclonal antibodies (mAb), antibody drug conjugates, bispecific antibodies, chimeric antigen receptor T cells, chemotherapy (CTx) and others. Moreover, MM biology and genomic heterogeneity are better understood, suggesting that it is important to enumerate the extent of clonal heterogeneity and to interpret the results of subsequent therapy in light of this heterogeneity. Effective targeted therapy requires drug combinations which target distinct subclones, and the employment of targeted therapies only in patients for whom the drug target is entirely clonal,1 the former being

common and the latter scarce. This provides the rationale for doublet, triplet and quadruplet therapies. At the time of the APEX study,2 the inhibition of the proteasome was a completely novel therapeutic approach, with remarkable preclinical activity in MM.3 The drug bortezomib was the first in class for clinical application (Figure 1A).2 This large randomized open-label phase III study compared bortezomib/dexamethasone (VD) versus single agent dexamethasone (D) in 669 patients with relapsed/refractory MM (RRMM), who had one to three prior lines of therapy. It was also the first major international phase III study in RRMM that brought together the MM community in Europe and North America. Notably, after the interim analysis determined superiority of VD over D, patients in the D arm were permitted to cross over to receive bortezomib (Bz) after disease progression. This study illustrated higher responses, longer time to progression and extended overall survival in

Figure 1. Proteosome inhibition. (A) Mechanism of proteasome inhibition. (B) APEX trial randomizing 669 patients (pts) 1:1 to bortezomib (Bz)/dexamethasone (D) (VD) vs. D alone and the achievements evolving from this pivotal trial. (C) Key clinical trials that led to Food and Drug Adminstration (FDA) and European Medicines Agency (EMA) approval of various novel agents (NA). The proteosome inhibitor (PI) Bz in the APEX trial was established as one fundamental NA and was the clinical study that led others to follow and likewise achieve clinical significance in multiple myeloma (MM). Dex: dexamethasone; RRMM: relapsed/refractory MM (RRMM).

1018

haematologica | 2022; 107(5)

Landmark Paper in Hematology

patients treated with VD versus D alone, despite this crossover. The combined complete and partial response rates were 38% versus 18% (P<0.001), median progression-free survival was 6.22 months versus 3.49 months (hazard ratio [HR] 0.55; P<0.0001), the 1-year overall survival was 80% versus 66% (HR 0.57; P=0.001) and grade 3/4 adverse events occurred in 75% versus 60%, respectively.2 This study led to swift approval of VD in RRMM and established that novel agents (NA), far more active than dexamethasone, should be developed. In fact, with myriads of papers in MM, numerous publications exist for RRMM, illustrating the widespread use of PI and the important follow-up studies evolving after the New England Journal of Medicine publication.2 The enormous success of APEX and its worldwide implication are exemplified in Figure 1B and C, namely that PI were defined as key NA and induced impressive responses not only in RRMM, but also in induction, consolidation and maintenance, and stood at the beginning of the rapid development of several NA which have revolutionized MM treatment. APEX has stimulated multiple follow-up research, including papers on lines of therapy and response, earlier versus later relapse treatment, side effect management, cytogenetics/high-risk patients, PI retreatment, and avoidance and recovery of peripheral neuropathy. It allowed the exploration of multiple highly potent Bz combinations beyond D, namely employing IMiD, CTx,

mAb and others. It encouraged the development of second generation PI and novel PI schedules, i.e., with subcutaneous and weekly applications, rather than intravenous and twice a week applications. It has demonstrated that NA are a vital treatment armamentarium that has even challenged the replacement of standard autologous stem cell transplantation with NA treatment alone. Disclosures The authors participated in the APEX study at the University of Freiburg (UKF) / Comprehensive Cancer Center Freiburg (CCCF), but other than receiving study support (UKF/CCCF) for the patients being included and meticulously documented in APEX, declare no competing financial interest related to this historic report. Contributions ME wrote this report, JMW and RW provided comments and approved the paper. Acknowledgements The authors thank distinguished IMWG, EMN, DSMM and GMMG experts for their advice and recommendations. We apologize that only a fraction of important papers, related to the APEX study can be cited. The paper is dedicated to all MM experts worldwide, our MM patients and all colleagues involved in clinical studies and industry.

References 1. Lohr JG, Stojanov P, Carter SL, et al. Widespread genetic heterogeneity in multiple myeloma: implications for targeted therapy. Cancer Cell. 2014;25(1):91-101. 2. Richardson PG, Sonneveld P, Schuster MW, et al. Bortezomib or high-dose dexamethasone for relapsed multiple myeloma. N Engl J Med. 2005;352(24):2487-2498. 3. Orlowski RZ. The ubiquitin proteasome pathway from bench to bedside. Hematology Am Soc Hematol Educ Program. 2005;220-225.

haematologica | 2022; 107(5)

1019

EDITORIALS Immune thrombocytopenia: vaccination does not equal causation Allyson Pishko1 and Adam Cuker1,2 1

Department of Medicine, Perelman School of Medicine, University of Pennsylvania and 2Department of Pathology & Laboratory Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA E-mail: ADAM CUKER - Adam.Cuker@pennmedicine.upenn.edu doi:10.3324/haematol.2021.279727

accination is the most important tool available for decreasing the incidence of severe disease and death due to severe acute respiratory syndrome coronavirus2 (SARS-CoV-2).1 Early case reports and media coverage of patients diagnosed with immune thrombocytopenia (ITP) after receipt of a SARS-CoV-2 vaccine raised concern among patients and providers.2 In this issue of Haematologica, Choi et al. describe the clinical course of a series of patients with a diagnosis of ITP in the 6 weeks following SARS-CoV-2 vaccination as identified by a national survey of hematologists in Australia.3 Patients with features of the more lethal and rare vaccine-related phenomenon, vaccine-induced immune thrombotic thrombocytopenia, were excluded.3 At the time of this writing, 4.48 billion patients worldwide have received at least one dose of a SARS-CoV-2 vaccine.4 It is now increasingly apparent from large epidemiological studies that if there is an association with vaccination and ITP, the overall incidence is low (Table 1).5-7 In the study in this issue, Choi et al. explicitly do not set out to describe the incidence of ITP following vaccination nor determine causality. However, it is reassuring that only 14 patients (10 de novo cases) were identified with ITP in this survey of hematologists across Australia. ITP is not a particularly uncommon diagnosis, with an estimated incidence rate of three cases of primary ITP diagnosed per 100,000 adults per year.8 Given that 4.2 million adults in Australia had been vaccinated at the time of the study by Choi et al., we would expect about 126 of these individuals to be newly diagnosed with ITP during the ensuing 12 months by chance alone, regardless of vaccination status. Using this estimate, we would anticipate roughly 15 individuals to be newly diagnosed with ITP during any 6-week period, whether they received vaccination or not. This simplified analysis does not take into account seasonal variation in the incidence of ITP or differences in incidence based on age or gender.8,9 Nonetheless, the number of new cases of ITP (n=10) identified by Choi et al. in the 6 weeks following SARS-CoV-2 vaccination is in line with this

expected incidence, suggesting that some or all of the cases identified by the authors may have occurred in proximity to vaccination purely by chance. While we cannot conclude causality from the study, it is certainly plausible that vaccination could trigger an ITP episode. Any stimulus to the immune system may theoretically induce production of platelet auto-antibodies, as has been observed with certain pathogens and with live vaccines, including the measles vaccine.10 The study by Choi et al. provides important details regarding the clinical course of ITP following SARS-CoV-2 vaccination. Most cases responded rapidly to standard first-line therapies (corticosteroids, intravenous immunoglobulin), as would be expected in adults with garden-variety primary ITP. This experience suggests that patients who are diagnosed with ITP in proximity to SARS-CoV-2 vaccination may generally be treated with standard ITP therapy. With the rampant increase in the delta variant of SARSCoV-2, it is all the more important to instill confidence in the safety of vaccines, while also being careful to not be dismissive of safety concerns. The study by Choi et al. adds to literature reinforcing the safety of the SARS-CoV-2 vaccines by demonstrating that occurrence of ITP following vaccination is uncommon, perhaps no greater than what would be expected to occur purely by chance, and that most cases are treatable with standard first-line therapies. Patients with known ITP should be counseled to monitor for signs/symptoms of recurrence (e.g., petechiae, mucosal bleeding) after vaccination. Clinicians may consider obtaining platelet counts, particularly in patients without stable counts, before and after vaccination. Overall, the benefits of vaccination against COVID-19 continue to vastly outweigh the risks in patients with ITP and in the population at large. Disclosures AMP receives research funding from Novo Nordisk and Sanofi Genzyme. AC has served as a consultant for Synergy, has received

Table 1. Incidence of immune thrombocytopenia following SARS-CoV-2 vaccination from selected epidemiological studies.

Study

Vaccine

Definition of ITP

Population

Simpson et al.

ChAdOx1

Diagnosis code (read code)

Age ≥18, Scotland

Simpson et al.5

BNT162b2

Pottegard et al.6

ChAdOx1

Diagnosis code (read code) Diagnosis code (ICD-10) VAERS report of thrombocytopenia (did not specify ITP)

Age ≥18, Scotland Age 18-65 years, Denmark and Norway Adults, USA

Welsh et al.7

BNT162b2 or mRNA-1273

Incidence

Relative risk

1.13 (95% CI 0.62-1.63) aRR 5.77 cases per (95% CI 2.41-13.83)* 100,000 vaccinations NR aRR 0.54 (95% CI 0.1-3.02)* <5 cases out of NR 281,264 first doses 0.8 cases per NR million doses

ITP: immune thrombocytopenia; 95% CI: confidence interval; aRR: adjusted relative risk; NR: not reported; VAERS: Vaccine Adverse Event Reporting System. *Adjusted relative risk of ITP occurring 0-27 days after vaccination.

1020

haematologica | 2022; 107(5)

Editorials

royalties from UpToDate, and his institution has received research support on his behalf from Alexion, Bayer, Novartis, Novo Nordisk, Pfizer, Sanofi, and Spark. Contributions Both authors contributed equally.

References 1. Bernal JL, Andrews N, Gower C, et al. Effectiveness of Covid-19 vaccines against the B.1.617.2 (delta) variant. New Engl J Med. 2021;385(7):585-594. 2. Lee E, Cines DB, Gernsheimer T, et al. Thrombocytopenia following Pfizer and Moderna SARS-CoV-2 vaccination. Am J Hematol. 2021;96(5):534-537. 3. Choi PY, Hsu D, Tran HA, et al. Immune thrombocytopenia following vaccination during the COVID-19 pandemic. Haematologica. 2022;107(5):1193-1196. 4. Ritchie H, Ortiz-Ospina E, Beltekian D, et al. Coronavirus pandemic

haematologica | 2022; 107(5)

(COVID-19). https://ourworldindata.org/coronavirus. 5. Simpson CR, Shi T, Vasileiou E, et al. First-dose ChAdOx1 and BNT162b2 COVID-19 vaccines and thrombocytopenic, thromboembolic and hemorrhagic events in Scotland. Nat Med. 2021;27(7):12901297. 6. Pottegård A, Lund LC, Karlstad Ø, et al. Arterial events, venous thromboembolism, thrombocytopenia, and bleeding after vaccination with Oxford-AstraZeneca ChAdOx1-S in Denmark and Norway: population based cohort study. BMJ. 2021;373:n1114. 7. Welsh KJ, Baumblatt J, Chege W, Goud R, Nair N. Thrombocytopenia including immune thrombocytopenia after receipt of mRNA COVID19 vaccines reported to the Vaccine Adverse Event Reporting System (VAERS). Vaccine. 2021;39(25): 3329-3332. 8. Schoonen WM, Kucera G, Coalson J, et al. Epidemiology of immune thrombocytopenic purpura in the General Practice Research Database. Br J Haematol. 2009;145(2):235-244. 9. Moulis G, Guénin S, Limal N, et al. Seasonal variations of incident primary immune thrombocytopenia in adults: an ecological study. Eur J Intern Med. 2017;37:e26-e28. 10. Cines DB, Bussel JB, Liebman HA, Prak ETL. The ITP syndrome: pathogenic and clinical diversity. Blood. 2009;113(26):6511-6521.

1021

Editorials

For older adults with hematologic malignancies, a comprehensive geriatric assessment matters Raul Cordoba Fundacion Jimenez Diaz University Hospital, Health Research Institute IIS-FJD, Madrid, Spain E-mail: RAUL CORDOBA - raul.cordoba@fjd.es doi:10.3324/haematol.2021.279927

n this issue of Haematologica, DuMontier et al. address a key question in the management of older adults with hematologic malignancies by reporting the results of a randomized controlled trial of geriatric consultation plus standard care versus standard care alone.1 Cancer can be considered an age-related disease because the incidence of most cancers increases with age.2 With regards to hematologic malignancies, incidence rates increase for non-Hodgkin lymphoma, multiple myeloma, and acute myeloid leukemia, and remain relatively stable for acute lymphoblastic leukemia, chronic lymphocytic leukemia, and chronic myeloid leukemia among adults aged ≥75 years. In spite of improving supportive care, survival for patients aged ≥75 years with hematologic malignancies is generally poor, particularly for those with acute leukemia. Understanding the heterogeneity in the outcomes of patients with hematologic malignancies, as well as the treatment challenges and management of frailty and comorbidities among older patients may help physicians to better address the hematologic cancer burden and mortality in the aging population.3 Hematologic malignancies are a miscellaneous group of diseases with regard to biology, prognosis and treatment options. Treatment decisions in older patients should not only be influenced by disease characteristics such as stage, histology, cytogenetics, molecular markers, etc. but also by patient-related factors such as fitness, frailty, and patients’ preferences. Furthermore, fitness and frailty are not static, but dynamic factors that may improve or deteriorate over time in the course of a disease and its treatment. Geriatric assessment is considered an important task during the diagnostic work-up and prior to deciding treatment in older adults with hematologic malignancies.4,5 Geriatric assessment includes a careful assessment of various domains including instrumental and basal activities of daily living (IADL, ADL), mobility, nutrition, cognitive function, and mental status. Many instruments, including screening tools (e.g., G8) and hematology-specific approaches (e.g., the brief Geriatric Assessment in Hematology tool, the GAH scale) have been suggested to perform geriatric assessments in patients with hematologic malignancies.6 A commonly accepted concept is to categorize older patients with hematologic malignancies into ‘fit’ for standard treatment, ‘unfit’ for attenuated treatment, and ‘frail’/terminally ill, not suitable for specific hematologic therapies but best supportive care. The study by DuMontier et al. presents the first randomzied controlled trial of geriatric consultation in older adults with hematologic malignancies. While the study did not meet its primary endpoint of improvement in survival, consultation did increase the proportion of patients who participated in a goals-of-care discussion. The study is important to the field of hematology as it is the first

1022

randomized controlled trial of its kind in hematology, in contrast to four separate randomized controlled trials enrolling older patients with solid tumors presented at the American Society of Clinical Oncology (ASCO) anual meeting in 2020. The primary outcome was 1-year overall survival and secondary endpoints included unplanned care utilization within 6 months of follow-up and documented end-of-life goals-of-care discussions. Patients who were assigned to the intervention group received simultaneously geriatric consultation with a geriatrician in addition to their standard oncologic care. Patients were assessed following the ASCO’s Guideline for Geriatric Oncology for function and falls, comorbidity and polypharmacy, cognition, depression/mood, and nutrition.4 Recommended interventions included counseling, recommendations for non-pharmacological interventions, pharmacological interventions, and referrals to other specialties or allied healthcare. One hundred sixty patients with a median age of 80.4 years (standard deviation = 4.2) were randomized to either geriatric consultation plus standard care (n=60) or standard care alone (n=100). Of those randomized to geriatric consultation, 48 (80%) completed at least one visit with a geriatrician. Consultation did not improve survival at 1 year compared to standard care (difference: 2.9%, 95% confidence interval [95% CI]: -9.5% to 15.2%, P=0.65), and did not significantly reduce the incidence of emergency department visits, hospital admissions, or days in hospital. Consultation did improve the odds of having end-of-life goals-of-care discussions (odds ratio = 3.12, 95% CI: 1.03 to 9.41) and was valued by surveyed hematologic oncology clinicians, with 62.9%-88.2% rating consultation as useful in the management of several geriatric domains. Patient-reported outcomes and quality of life, as well as preserved function (mobility, cognition) and autonomy (ADL, IADL), appear important and likely are not sufficiently surrogated by established study endpoints such as response rates, toxicity and survival outcomes. Assessment of patient-reported outcomes and quality of life studies are both linked to geriatric assessment and are therefore warranted in older patients with hematologic cancer. Patient-related outcomes can help to narrow the gap between patients' and healthcare professionals' view of patients’ health and treatment success.7 Moreover, several novel drugs have been developed as oral agents, introducing an additional challenge in the management of patients, such as ensuring optimal adherence to therapy in order to maximize treatment efficacy. In addition to the work presented by DuMontier et al., a recently published review provides updates on the new therapies for common hematologic malignancies with an emphasis on older adult-specific evidence and the evolv-

haematologica | 2022; 107(5)

Editorials

ing role of a geriatric assessment in informing therapy selection and management.8 Disclosures No conflicts of interest to disclose. With regard to work outside this publication, I provide consultancy services for AbbVie, Janssen, AstraZeneca, Beigene, Roche, Kite/Gilead, Celgene/BMS, Takeda, Kyowa-Kirin and ADCTherapeutics. My group has received a research grant from Pfizer.

References 1. DuMontier C, Uno H, Hshieh T, et al. Randomized controlled trial of geriatric consultation versus standard care in older adults with hematologic malignancies. Haematologica. 2022;107(5):1172-1180. 2. White MC, Holman DM, Boehm JE, et al. Age and cancer risk: a poten-

haematologica | 2022; 107(5)

tially modifiable relationship. Am J Prev Med. 2014;46(3 Suppl 1):S7-S15. 3. Krok-Schoen JL, Fisher JL, Stephens JA, et al. Incidence and survival of hematological cancers among adults ages ≥75 years. Cancer Med. 2018;7(7):3425-3433. 4. Mohile SG, Dale W, Somerfield MR, et al. Practical assessment and management of vulnerabilities in older patients receiving chemotherapy: ASCO guideline for geriatric oncology. J Clin Oncol. 2018;36(22):2326-2347. 5. Abel GA, Klepin HD. Frailty and the management of hematologic malignancies. Blood. 2018;131(5):515-524. 6. Scheepers ERM, Vondeling AM, Thielen N, van der Griend R, Stauder R, Hamaker ME. Geriatric assessment in older patients with a hematologic malignancy: a systematic review. Haematologica. 2020;105(6):1484-1493. 7. Cannella L, Efficace F, Giesinger J. How should we assess patientreported outcomes in the onco-hematology clinic? Curr Opin Support Palliat Care. 2018;12(4):522-529. 8. Rosko AE, Cordoba R, Abel G, et al. Advances in management for older adults with hematologic malignancies. J Clin Oncol. 2021;39 (19):2102-2114.

1023

Editorials

Stem cell transplant for lymphoma - never too late? Amanda F. Cashen and Nancy L. Bartlett Washington University School of Medicine and Siteman Cancer Center, St. Louis, MO, USA E-mail: AMANDA CASHEN - acashen@wustl.edu doi:10.3324/haematol.2021.280519

n this issue, using the Center for International Blood and Marrow Transplant Research (CIBMTR) registry, Mei et al.1 provide a retrospective outcomes analysis of 285 patients with relapsed or refractory (R/R) diffuse large B-cell lymphoma (DLBCL) who underwent autologous hematopoietic cell transplantation (autoHCT) after achieving complete (CR) or partial response (PR) to their third or higher line of chemotherapy. Sixty three percent of patients had primary refractory disease or relapsed within 1 year of initial diagnosis. For perspective, during the same interval 577 patients with R/R DLBCL in the registry underwent autoHCT after second line therapy. Patients in the registry with detailed data including prior lines of therapy comprise approximately 8-10% of all patients in the CIBMTR registry, which in other analyses have been representative of the entire dataset (personal communication with corresponding author). Mei et al. report 5-year overall survival (OS) and progression-free survival (PFS) for the study population of 51% and 38%. As expected, patients in CR at the time of transplant had better OS and PFS, but a substantial minority of patients with PR also benefited from autoHCT, with a 5-year PFS of 34%. Relapses in patients with a PR occurred early, with a near plateau in the relapse/progression curve after 1 year. In contrast, patients in CR had a significantly lower 1-year risk of relapse but a continuous risk of relapse for at least 4 years after autoHCT, resulting in a non-significant difference in the 5-year risk of relapse between the two groups (CR 45% vs. PR 54%, P=0.14). Importantly, while nearly half of all patients were without relapse at 5 years, the 5-year PFS of 38% reflects a 5year non-relapse mortality of 12%. After lymphoma, the second most common cause of death was second malignancy, in 10% of the study cohort during the entire follow-up period. How does this retrospective registry analysis inform our decision whether to pursue autoHCT for a given patient with R/R DLBCL? Primarily, it confirms that patients with chemosensitive disease can achieve longterm remissions after autoHCT, including those who have early failure of frontline immunochemotherapy, those who achieve less than a CR to salvage chemotherapy, and those who have been treated with more than one line of salvage chemotherapy. Consistent with prior reports, Mei et al. found that, while patients in CR have a better PFS than those in PR, approximately a third of PR patients remain alive and progression-free after autoHCT. Their results mirror another CIBMTR analysis2 that reported 5-year PFS of 41% for patients with R/R DLBCL who had PET-positive PR prior to autoHCT. Even patients with primary refractory disease or early relapse have a 40-50% chance of durable PFS after autoHCT, if they respond to salvage therapy.3,4,5 Unfortunately, many patients with R/R DLBCL do not achieve a CR or PR to one or more lines of salvage

1024

chemotherapy. In the prospective, randomized CORAL study,4 63% of all patients, 51% of patients previously treated with rituximab, and 46% of patients with primary refractory disease or early relapse achieved CR or PR to RICE or R-DHAP. Of the patients who did not respond to the study-assigned salvage regimen, only 32% responded to subsequent therapy and underwent a stem cell transplant.6 In two randomized studies of anti-CD19 chimeric antigen receptor T-cell therapy (CAR-T) versus standard of care in patients with early failure of frontline immunochemotherapy, the response rates to standard second line therapies were 50%7 and 43%.8 Clearly, inadequate response to second- or third-line therapy is a significant barrier to use of autoHCT in many patients with R/R DLBCL. New agents with encouraging activity in R/R DLBCL, including bispecific T-cell engagers9 and antibody-drug conjugates,10 may improve the number of patients who achieve PR or CR to second- or third-line therapy making autoHCT an option for more patients. However, whether response to novel therapies with unique mechanisms of action will predict response to autoHCT is unknown and will be extremely challenging to study. While CAR-T is the obvious choice for patients who do not respond to standard second- or third-line regimens, the choice between autoHCT and CAR-T can be difficult for patients with chemosensitive R/R DLBCL. The potential approval of CAR-T in the second line will increase the challenge of weighing the relative benefits of autoHCT versus CAR-T for a given patient. Now more than 4 years after approval of the first CAR-T for R/R DLBCL, significant hurdles remain, including complicated and delayed insurance approvals, limitations on manufacturing slots, and the potential for progressive disease and worsening performance status during the “waiting period”. Rigorous cost-benefit analysis of CAR-T versus autoHCT for patients with chemosensitive disease in second and third line would be informative. Additionally, late effects of CAR-T and autoHCT need to be compared and considered in the treatment decision. Post-CAR-T complications, especially prolonged cytopenias and infections, can be more challenging than post-autoHCT in a significant minority of patients, and, as Mei’s study demonstrates, second malignancies are a concerning cause of nonrelapse mortality post-autoHCT. Although we have ample data on the tolerability and efficacy of CAR-T following relapse after autoHCT, there is no data on autoHCT following CAR-T cell failure. In the current use of CAR-T for chemorefractory disease or relapse after autoHCT, patients would not be candidates for autoHCT after CAR-T. However, if approvals are forthcoming for CAR-T as second-line therapy, understanding the feasibility and efficacy of autoHCT after CAR-T would help inform the decision of sequencing these therapies in patients responding to second- or third-

haematologica | 2022; 107(5)

Editorials

line chemotherapy. In the meantime, the retrospective studies by Mei et al. and others provide important benchmarks of outcomes following autoHCT, suggesting that autoHCT remains a reasonable option for the patients with chemosensitive R/R DLBCL, even in later lines of therapy. For those patients, CAR-T can be held in reserve in case of relapse after transplant. Disclosures NLB has received personal research funding from KITE/Gilead, has received research funding to the institution from ADC Therapeutics, Bristol-Meyers Squibb, Celgene, Forty Seven, Immune Design, Janssen, KITE Pharma, Merck, Millennium, Pharmacyclics,Pfizer, Roche/Genentech and SeaGen, and sits on the advisory boards of ADC Therapeutics, Roche/Genentech, SeaGen. AFC sits on the advisory board of Secura Bio and has received research funding Secura Bio. Contributions AC and NLB contributed equally to the writing of this editorial.

References 1. Mei M, Hamadani M, Ahn K, et al. Autologous hematopoietic cell transplantation in diffuse large B-cell lymphoma after 3 or more lines of prior therapy: evidence of durable benefit. Haematologica. 2022;105(5):1214-1217.

haematologica | 2022; 107(5)

2. Shah N, Ahn K, Litovich C, et al. Is autologous transplant in relapsed DLBCL patients achieving only a PET+ PR appropriate in the CAR T-cell era? Blood. 2021;137(10):1416-1423. 3. Bal S, Costa L, Suater C, Litovich C, Hamadani M. Outcomes of autologous hematopoietic cell transplantation in diffuse large B cell lymphoma refractory to firstline chemoimmunotherapy. Transplant Cell Ther. 2021;27(1):55.e1-55.e7. 4. Gisselbrecht C, Glass B, Mounier N, et al. Salvage regimens with autologous transplantation for relapsed large B-cell lymphoma in the rituximab era. J Clin Oncol. 2010;28(27):4184-4190. 5. Hamadani M, Parameswaran N, Zhang Y, et al. Early failure of frontline rituximab-containing chemo-immunotherapy in diffuse large B cell lymphoma does not predict futility of autologous hematopoietic cell transplantation. Biol Blood Marrow Transplant. 2014;20 (11):1729-1736. 6. Van Den Neste E, Schmitz N, Mounier N, et al. Outcome of patients with relapsed diffuse large B-cell lymphoma who fail second-line salvage regimens in the international CORAL study. Bone Marrow Transplant. 2016;51(1):51-57. 7. Locke F, Miklos D, Jacobson C, et al. Axicabtagene ciloleucel as second-line therapy for large B-cell lymphoma. N Eng J Med. 2022;386 (7):640-654. 8. Bishop M, Dickinson M, Purtill D, et al. Second-line tisagenlecleucel or standard care in aggressive B-cell lymphoma. N Eng J Med. 2022;386(7):629-639. 9. Budde L, Assouline S, Sehn L, et al. Single-agent mosunetuzumab shows durable complete responses in patients with relapsed or refractory B-cell lymphomas: phase I dose-escalation study. J Clin Oncol. 2022;40(5):481-491 10. Caimi P, Ai W, Alderuccio J, et al. Loncastuximab tesirine in relapsed or refractory diffuse large B-cell lymphoma (LOTIS-2): a multicentre, open-label, single-arm, phase 2 trial. Lancet. 2021;22 (6):790-800.

1025

ARTICLE Ferrata Storti Foundation

Haematologica 2022 Volume 107(5):1026-1033

Acute Lymphoblastic Leukemia

Clofarabine increases the eradication of minimal residual disease of primary B-precursor acute lymphoblastic leukemia compared to high-dose cytarabine without improvement of outcome. Results from the randomized clinical trial 08-09 of the Cooperative Acute Lymphoblastic Leukemia Study Group. Gabriele Escherich,1 Udo zur Stadt,1 Arndt Borkhardt,2 Dagmar Dilloo,3 Jörg Faber,4 Tobias Feuchtinger,5 Thomas Imschweiler,6 Norbert Jorch,7 Arnulf Pekrun,8 Irene Schmid,5 Franziska Schramm,1 Michael Spohn,9,10 Martin Zimmermann11 and Martin A Horstmann1,9 1 Clinic of Pediatric Hematology and Oncology, University Medical Center HamburgEppendorf, Hamburg; 2Department of Pediatric Oncology, Hematology and Clinical Immunology, Medical Faculty Duesseldorf, Duesseldorf; 3Department of Pediatric Hematology/Oncology, University Hospital Bonn, Bonn; 4Department of Pediatric Hematology/Oncology, University Hospital Mainz, Mainz; 5Dr. Von Hauner Children's Hospital, Ludwig Maximilian University, Munich; 6Department of Pediatric Hematology and Oncology, Helios Hospital, Krefeld; 7Department of Pediatric Hematology and Oncology, Protestant Hospital of Bethel Foundation, Bielefeld; 8Department of Pediatric Hematology and Oncology, Hospital Bremen-Mitte, Bremen; 9Research Institute Children’s Cancer Center Hamburg, Hamburg; 10Bioinformatics Core Unit, University Medical Center Hamburg, Hamburg and 11Department of Pediatric Hematology and Oncology, Medical School Hannover, Hannover, Germany

ABSTRACT

Correspondence: GABRIELE ESCHERICH escherich@uke.de MARTIN A. HORSTMANN horstmann@uke.de Received: June 1, 2021. Accepted: July 16, 2021. Pre-published: August 5, 2021. https://doi.org/10.3324/haematol.2021.279357

©2022 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

1026

ovel treatment strategies are needed to improve cure for all children with acute lymphoblastic leukemia (ALL). To this end, we investigated the therapeutic potential of clofarabine in primary ALL in trial CoALL 08-09 (clinicaltrials gov. identifier: NCT01228331). The primary study objective was the minimal residual disease (MRD)based comparative assessment of cytotoxic efficacies of clofarabine 5x40 mg/m2 versus high-dose cytarabine (HIDAC) 4x3g/m2, both in combination with PEG-ASP 2,500 IU/m2 as randomized intervention in early consolidation. The secondary objective was an outcome analysis focused on treatment arm dependence and MRD after randomized intervention. In B-cell precursor (BCP)-ALL, eradication of MRD was more profound after clofarabine compared to cytarabine, with 93 versus 79 of 143 randomized patients per arm reaching MRD-negativity (c2 test P=0.03, leftsided P [Fisher’s exact test]=0.04). MRD status of BCP-ALL after randomized intervention maintained its prognostic relevance, with a significant impact on event-free survival (EFS) and relapse rate. However, no difference in outcome regarding EFS and overall survival (OS) between randomized courses was observed (5-year EFS: clofarabine 85.7, SE=4.1 vs. HIDAC 84.8, SE=4.7 [P=0.96]; OS: 95.7, SE=1.9 vs. 92.2, SE=3.2 [P=0.59]), independent of covariates or overall risk strata. Severe toxicities between randomized and subsequent treatment elements were also without significant difference. In conclusion, clofarabine/PEG-ASP is effective and safe, but greater cytotoxic efficacy of clofarabine compared to HIDAC did not translate into improved outcomes indicating a lack of surrogacy of post-intervention MRD at the trial level as opposed to the patient level, which hampers a broader implementation of this regimen in the frontline treatment of ALL.

haematologica | 2022; 107(5)

Lymphoblastic leukaemia, paediatric, clofarabine

Introduction The prevention of relapse without increasing toxicity is a challenging goal of frontline treatment in acute lymphoblastic leukemia (ALL), which is unlikely to be achieved by recombination or intensification of established chemotherapeutic agents. Besides immunotherapeutical approaches, novel compounds must be probed to prevent the development of resistant clones or to efficiently overcome those that already exist. To this end, we evaluated clofarabine as one of the latest chemotherapeutic drugs to receive authoritative approval for the treatment of relapsed/refractory ALL in childhood. Clofarabine is a second-generation purine nucleoside analogue that combines the positive characteristics of first-generation purine nucleosides fludarabine and cladribine by retaining 2-halogenated adenines, resulting in improved resistance against deamination and phosphorolysis.1-3 Several studies have been launched which scrutinized clofarabine in combination with other cytostatic drugs as second- or thirdline therapy, or as a bridging regimen to hematopoietic stem cell transplantation.4-6 In the Children’s Oncology Group (COG) trial AALL1131, clofarabine was administered in combination with etoposide and cyclophosphamide, which were associated with severe infections and persistent myelotoxicity leading to premature closure of the experimental clofarabine arm.7 In order to assess the value of the frontline usage of clofarabine, the Cooperative Acute Lymphoblastic Leukemia Study Group (CoALL) conducted a sequential phase II/III trial embedded into the CoALL 08-09 regimen for newly diagnosed ALL patients for whom end-of-induction (EOI) minimal residual disease (MRD) imposed a greater risk of relapse. During the non-randomized phase II, all eligible patients with quantifiable EOI MRD received the combination of clofarabine 5x40 mg/m2 and pegylated asparaginase (PEG-ASP) 2,500 IU/m2 as early consolidation treatment. The results were compared to a high-dose cytarabine (HIDAC)/PEGASP control group in predecessor trial CoALL 03-07. Combined administration of clofarabine and PEG-ASP was feasible and exhibited acceptable toxicities without unexpected severe side effects.8 Herein, we describe the results of the subsequent phase III trial within CoALL 08-09, comparing the efficacy and tolerability of clofarabine/PEG-ASP versus HIDAC/PEG-ASP at early consolidation in a randomized fashion.

Methods Study design and patients CoALL 08-09 was a multi-center, randomized trial for patients under the age of 18 years with a confirmed diagnosis of acute B- or T-cell precursor leukemia. Accrual was open from 1 October 2010 to 31 December 2019. The study was approved by the competent ethics boards (Online Supplementary Table S1) and conducted in accordance with the Declaration of Helsinki. The efficacy of clofarabine/PEG-ASP was compared with HIDAC/PEG-ASP in a randomized fashion as a primary study objective. An additional randomization of anthracyclines in delayed intensification was conducted from 2010 to 2016 with the primary objective of comparing toxicities.9

Stratification and treatment All patients received the same three-drug induction with four weekly doses of daunorubicin (36 mg/m2) and vincristine (1.5

haematologica | 2022; 107(5)

mg/m2) along with oral methylprednisolone (60 mg/m2) over 28 days and a single dose of age-adapted intrathecal methotrexate. BCP-ALL with a discernible, but non-quantifiable, or quantifiable EOI MRD and T-ALL with ≥10-3 EOI MRD were eligible for randomization, receiving either clofarabine 5x40 mg/m2 or HIDAC 4x3 g/m2 in combination with PEG-ASP 2,500 IU/m2 as the first or second course of consolidation in the treatment of BCP-ALL or TALL, respectively (Figure 1A). Further treatment was administered according to respective strata (Figure 1B). By protocol, enrolled patients who achieved MRD-negativity at the end of induction or inversely showed an induction failure were not eligible for randomization (see the Online Supplementary Appendix for additional information).

Randomization The randomization was performed by the coordinating trial center after stratification had been finalized according to EOI MRD status. Each stratum (high risk [HR] patients were subdivided according to immunophenotype) underwent independent randomization on the basis of randomly permuted blocks to avoid imbalances within risk strata.

Analysis of minimal residual disease Real-time quantitative polymerase chain reaction (PCR) analyses were performed targeting immunoglobulin heavy chain (IGH) and T-cell receptor (TCR) gene rearrangements to assess MRD. Data were interpreted according to the guidelines developed by the European Study Group for MRD detection in ALL (EuroMRD ALL).10

Statistical analyses The probability of event-free (pEFS) and overall survival (pOS) was estimated using the Kaplan-Meier method and compared between subgroups using the log-rank test.11 Cumulative incidence functions of isolated CNS or any (isolated and combined) CNS relapse, as well as testicular relapse, treatment-related secondary malignancies and toxicity-related death were calculated using the Kalbfleisch and Prentice method and compared using Gray’s test.12 A c2 test, a Fisher’s exact test, and Spearman’s rank correlation analyses were applied to compare the distribution of parameters between subgroups and correlation between parameters.13 A c2 test was applied to determine the difference in the rate of MRD-positive patients, as provided in the study protocol. This was complemented by a one-sided Fisher’s exact test and a Cochran-Armitage trend test, the latter of which compared the trend in MRD values between randomized groups.14 The status of patients was monitored annually. The database was newly updated (1 December 2020) prior to usage for analysis. Analyses were carried out using SAS version 9.4. Further details of statistical analyses are provided in the Online Supplementary Appendix.

Results Overall, 303 study patients were eligible and randomized, allocating 151 patients toward clofarabine/PEG-ASP and 152 patients toward HIDAC/PEG-ASP (Figure 2; Table 1; Online Supplementary Appendix). Of those patients, the main endpoint (i.e., MRD after randomized intervention) was reached by 296 patients, in close approximation to the planned sample size (n=295) (Table 2). There were no differences in patient characteristics regarding known risk factors other than a more frequent occurrence of ETV6RUNX1 in the clofarabine-treated cohort (Table 1). The 1027

G. Escherich et al.

incidence of hematopoietic stem cell transplantation (HSCT) in first complete remission due to persistent MRD was comparable between arms (n=11 vs. n=12 HSCT in clofarabine and HIDAC cohorts, respectively). T-ALL patients were similarly underrepresented in both randomized arms compared to the whole study cohort (5.3% [n=8] in the clofarabine and 5.9% [n=9] in the HIDAC cohort vs. 14.2% [n=67] in the total cohort), mainly due to a greater proportion of T-ALL in the induction failure cohort (n=24/31 patients [77%]) and in the HR-reduced cohort (n=15/51 patients [29%]), both of which were

excluded from randomization according to the study protocol (Table 2; Online Supplementary Appendix).

Minimal residual disease response In the randomized treatment arms, we observed a rate of 44% MRD-positivity after high-dose cytarabine versus 33% MRD-positivity after clofarabine in BCP-ALL (Pchi2=0.03; left-sided Fisher test P=0.04). The overall reduction of MRD in BCP-ALL was significantly more profound after clofarabine compared to cytarabine, with 93 clofarabine-treated patients versus 79 HIDAC-treated patients reaching MRD

Figure 1. Treatment overview. (A) Randomized treatment block clofarabine vs. high-dose cytarabine, each combined with pegylated asparaginase (PEG-ASP). (B) Schematic overview of the CoALL 08-09 protocol. ADR: doxorubicin; BCP: B-cell precursor; BMP: bone marrow puncture; CNS: central nervous system; d: day; Dex: dexamethasone; DNR: daunorubicin; EOI: end of induction; HIDAC: high-dose cytarabine; I: induction; MRD: minimal residual disease; R: randomization; VCR: vincristine; CoALL: Cooperative Acute Lymphoblastic Leukemia study group.

1028

haematologica | 2022; 107(5)

Lymphoblastic leukaemia, paediatric, clofarabine

negativity, and a lower rate of patients with quantifiable MRD levels (6 patients after clofarabine vs. 18 patients after HIDAC) (Cochran-Armitage trend test P=0.01; Table 2; Online Supplementary Figure S1). This observation holds true in a sub-analysis of the patients with a higher burden of EOI MRD (≥10-3) who were stratified to the low risk (LR)or HR-intensified arms. Among those 73 patients, 27 patients were MRD-negative after clofarabine compared to 16 patients randomized to the HIDAC arm (CochranArmitage trend test P=0.02). In ETV6-RUNX1-rearranged ALL, which occurred more frequently in clofarabine-treated patients by chance, we observed an equivalent efficacy of the randomized nucleosides, reflecting a generally high sensitivity toward asparaginase in this prognostically favorable genetic subgroup of ALL (Table 1; Online Supplementary Table S3). In order to address a potential skewing effect of misbalanced ETV6-RUNX1 on the MRD outcome of ran-

domized groups, ETV6-RUNX1-negative ALL was analyzed separately, which confirmed greater activity of clofarabine compared to HIDAC (Pchi2=0.04210) (Online Supplementary Table S3). Importantly, after the randomized course in early consolidation (day 50 in B-cell precursor (BCP)-ALL and day 64 in T-ALL patients), MRD maintained its prognostic relevance, with a significant impact on EFS and relapse rate in comparison to day 29 EOI MRD (Figure 3A and B).15 T-ALL patients of both randomized arms achieved comparable MRD reductions by day 64, although the number of T-ALL patients was very small (Tables 1 and 2). Nevertheless, the test for trends in the overall cohort comprising both BCP-ALL and T-ALL confirmed that clofarabine was significantly more effective in MRD reduction compared to HIDAC (Cochran-Armitage trend test P=0.01) (Table 2).

Table 1. Demographics and clinical characteristics of randomized patients.

Immunophenotype B-precursor ALL T-ALL Sex male female Age at diagnosis < 10 years ≥ 10 years WBC < 25/nL ≥ 25/nL ETV6-RUNX1 rearrangement positive negative unknown KMT2A rearrangement positive negative Karyotype < 44 chromosomes 44-50 chromosomes > 50 chromosomes unknown Treatment response BM day 15 M1 M2 M3 not available Risk Stratification Low-risk standard Low-risk intensified High-risk standard High-risk intensified

High-dose cytarabine (n=152) No. (%)

Clofarabine (n=151) No. (%)

143 (94.1) 9 (5.9)

143 (94.7) 8 (5.3)

0.82

79 (52) 73 (48)

85 (56.3) 66 (43.7)

0.45

123 (80.9) 29 (19.1)

119 (78.8) 32 (21.2)

0.65

101 (66.4) 51 (33.6)

110 (72.8) 41 (27.2)

0.73

30 (19.7) 117 (77) 5 (3.3)

47 (31.1) 104 (68.9) 0 (0)

0.02

2 (1.3) 150 (98.7)

2 (1.3) 149 (98.7)

1.0

2 (1.3) 90 (59.2) 48 (31.6) 12 (7.9)

2 (1.3) 106 (70.2) 38 (25.2) 5 (3.3)

0.31

98 (64.5) 28 (18.4) 4 (2.6) 22 (14.5)

104 (68.9) 23 (15.2) 5 (3.3) 19 (12.6)

0.68

57 (37.5) 20 (13.2) 47 (30.9) 28 (18.4)

62 (41.1) 19 (12.6) 43 (28.5) 27 (17.9)

ALL: acute lymphoblastic leukemia; WBC: white blood cell; BM: bone marrow.

haematologica | 2022; 107(5)

1029

G. Escherich et al.

Outcome of randomized groups No significant differences in outcome regarding EFS and OS were observed between the randomized arms (Figure 3C and D), with a median observation time of 3.7 years. There were also no significant differences in Cox regression analyses regarding the covariates sex, age (<10 years vs. ≥10 years), WBC (< 25/nL vs. ≥25/nL), ETV6-RUNX1, and HSCT in first continuous remission as time-dependent variables. An additional stratified analysis confirmed that there were no significant differences in EFS or relapse rate between randomized courses according to the categories negative, positive nonquantifiable (n.q.), and quantifiable MRD on day 50.

Besides events that were anticipated upon quantifiable MRD after randomized intervention, several relapses occurred in MRD-negative and MRD-positive n.q. patients in both randomized treatment arms, accounting for the observed lack of surrogacy of MRD in the outcome analysis (Online Supplementary Table S4). There was no evidence of a mutual impact between the randomizations at early consolidation and delayed intensification in this study, as shown by very similar pEFS in the latter randomized arms (log-rank test P=0.88 for patients receiving doxorubicin and log-rank test P=0.50 for patients receiving daunorubicin during delayed intensification).

Figure 2. Trial profile. Flow diagram according to CONSORT guidelines. CoALL: Cooperative Acute Lymphoblastic Leukemia Study Group; LR-R: low risk-reduced; HRR: high risk-reduced; MRD: minimal residual disease; SAE: serious adverse events; HIDAC: high-dose cytarabine.

Table 2. Minimal residual disease response toward clofarabine/PEG-ASP versus high-dose cytarabine/PEG-ASP.

High-dose Cytarabine No. (%)

Clofarabine No. (%)

All

0.03 c2 0.04 Fisher 0.01 Cochran-Armitage Trend Test

B-precursor ALL

MRD d50 pos.

61 (44)

45 (33)

106

B-precursor ALL

MRD d50 neg. MRD d50 pos. nq MRD d50 ≥10-4 MRD neg.

79 (56.4) 43 (30.7) 18 (12.9) 16 (21.9)

93 (67.4) 39 (28.3) 6 (4.3) 27 (37)

172 82 24 43

MRD d64 neg. MRD d64 pos. nq MRD d64 ≥10-4

4 (44.4) 3 (33.3) 2 (22.2)

3 (37.5) 4 (50.0) 1 (12.5)

7 7 3

B-precursor ALL EOI MRD ≥10-3 T-ALL

0.02 Cochran-Armitage Trend Test 0.94 Cochran-Armitage Trend Test

ALL: acute lymphoblastic leukemia; EOI: end-of-induction; MRD: minimal residual disease, d50: day 50; d64: day 64; neg: negative; pos: positive; nq: non-quantifiable. PEG-ASP: pegylated asparaginase.

1030

haematologica | 2022; 107(5)

Lymphoblastic leukaemia, paediatric, clofarabine

Toxicity

Online Supplementary Table S2A and B). Remarkably, when comparing CTC grades 0 to 2 against grades 3 and 4 for hemoglobin and platelets, clofarabine was associated with significantly less severe toxicities (Online Supplementary Table S2B). Clofarabine caused a more frequent grade 4 depletion of white blood cells suggesting a greater lymphotoxicity given that grade 4 reduction in neutrophil counts was comparable between randomized arms (Figure 4; Online Supplementary Table S2A). Nevertheless, the incidence of severe infections after randomized treatment was comparable (Figure 4; Online Supplementary Table S2A and B). Finally, the incidence of serious adverse events (SAE) during the remaining treatment courses was very similar (18 and 19 SAE in the clofarabine vs. HIDAC arm, respectively).

No statistically significant differences in the incidence of severe or persistent toxicities between randomized treatment elements or in the subsequent treatment realization were documented (Figure 4; Online Supplementary Table S2A and B). In particular, severe grade 3 or 4 skin toxicities were not observed in either treatment arm, but clofarabine was more frequently associated with grade 2 skin toxicities. With regard to hepatotoxicity, an elevation of transaminases (aspartate and alanine transaminases [AST and ALT], respectively) was significantly more often reported after clofarabine than after HIDAC, and then spontaneously resolved without exception after each randomized treatment element before the start of subsequent chemotherapy. Accordingly, time intervals between the randomized courses and the subsequent treatment elements were similar, with a median of 22 days (range, 20–38 days) after clofarabine/PEG-ASP and 19 days (range, 18–38 days) after HIDAC/PEG-ASP. Incidence and degree of myelotoxicity differed slightly between clofarabine and HIDAC (Figure 4;

Discussion As demonstrated in trial CoALL 08-09, clofarabine combined with PEG-asparaginase is effective in the eradication

Figure 3. Outcome analyses in randomized patients. (A) Probability of event-free survival (pEFS) (5 years of follow-up) in randomized patients according to MRD on day 50/64 after completion of randomized treatment courses. For comparative outcome probability analyses according to MRD levels, MRD negativity is denoted as 1, non-quantifiable (n.q.) MRD positivity is denoted as 2, and MRD ≥ 1x10-4 is denoted as 3. (B) Cumulative relapse rate (5 years of follow-up) in randomized B-precursor and T-acute lymphoblastic leukemia (T-ALL) patients according to MRD on day 50/64. (C) and (D) legends are swoped. (C) Comparative probability of eventfree (pEFS) (5 years of follow-up) analysis in clofarabine/PEG-ASP-treated vs. HIDAC/PEG-ASP-treated ALL patients. (D) Comparative analysis of overall survival (pOS) (5 years of follow-up) in clofarabine/PEG-ASP-treated vs. HIDAC/PEG-ASP-treated ALL patients. PEG-ASP: pegylated asparaginase; HIDAC: high-dose cytarabine.

haematologica | 2022; 107(5)

1031

G. Escherich et al.

Figure 4. Treatment-related toxicities in randomized patients according to treatment arm and common terminology criteria. HIDAC: high-dose cytarabine; WBC: white blood cell.

of MRD and well tolerated in the frontline treatment of ALL. In comparison to high-dose cytarabine/PEG-ASP, clofarabine/PEG-ASP was superior in the overall reduction of an MRD burden. The frequency of MRD-positive BCPALL patients in the standard arm was lower than the predicted rate of 60%, likely due to a smaller sample size and the different distribution of risk strata in the preceding trial, CoALL 03-07. Although the prognostic impact of MRD in BCP-ALL is still clearly discernible in early consolidation after the randomized courses of clofarabine versus HIDAC, the greater cytotoxic efficacy of clofarabine did not translate into an obvious improvement of outcome at the trial level after a median follow-up period of 3.7 years. This lack of surrogacy of MRD at early consolidation in a survival endpoint analysis could be explained by a small effect size, taking into account that only a single course of clofarabine was compared with high-dose cytarabine as a part of a complex multiagent chemotherapy backbone, the entirety of which determines treatment efficacy. Our trial design allowed for the detection of a ~10% difference in outcome between randomized treatment arms at a power of 80%. Hence, the small sample size has to be considered with regard to the number of randomized patients required in order to perform a meaningful comparative analysis of survival in CoALL 08-09, which was a priori defined as a secondary objective in the study protocol. Overall, clofarabine increased the rate of MRD negativity by 25% compared to HIDAC, which is an incremental improvement with borderline significance in contrast to a statistically more robust overall reduction of MRD after clofarabine (Table 2; Online Supplementary Figure S1). The occurrence of relapsing disease in MRD-negative patients after clofarabine (and HIDAC) observed in this trial points 1032

at MRD as a time-dependent variable. In this regard, early achievement of MRD negativity at the end of induction is more predictive of outcome than achievement of MRD negativity later in treatment, most likely due to the emergence of resistant clones, i.e., MRD negativity does not necessarily imply true eradication of the disease, but simply reflects a decrease to a level below the detection limit of the PCR-based MRD assay. Inversely, MRD positivity more reliably reflects outcome when measured later in treatment.15,16 In addition, the rarity of events after treatment of ALL in childhood might generally compromise surrogacy of MRD as a prognostic marker of outcome at the trial level. A previous multi-trial approach including 4,830 patients with ALL demonstrated that EOI MRD failed as a surrogate for treatment effects on EFS at the trial level, when dexamethasone and prednisone were compared in induction treatment of AIEOP-BFM ALL and COG trials.17-19 This meta-analysis raised caution with regard to MRD as a surrogate marker for treatment decisions in randomized trials. In contrast to these trials, in which the stratifying decision was made after randomization, we can exclude that the evaluation of MRD after randomized intervention impacted a decision on the subsequent treatment in CoALL 08-09, since the ultimate stratification had been done before randomization on d29 in BCP-ALL and on d43 in T-ALL. In this trial, we applied clofarabine at a dose of 40 mg/m2 daily x 5 corresponding to the previously established single agent maximum-tolerated dose (MTD) in adult acute leukemia which is lower than the MTD of 52 mg/m2 x 5 determined in pediatric patients with acute leukemia.2,20 The administration of high-dose clofarabine in conjunction with PEG-asparaginase in early consolidation of CoALL 08-09 was feasible largely due to almost non-overlapping haematologica | 2022; 107(5)

Lymphoblastic leukaemia, paediatric, clofarabine

toxicities. By contrast, clofarabine given at a reduced dose level of 30 mg/m2 x 5 or 20 mg/m2 x 5, respectively, was associated with unacceptably severe infections and myelotoxicities in heavily pretreated pediatric patients with relapsed/refractory leukemia when combined with cyclophosphamide, etoposide, vincristine, and PEG-ASP in the COG trial AALL1131.7 Since MRD fell short as a surrogate marker in a true endpoint analysis of survival of randomized patient cohorts in CoALL 08-09, standard cytarabine treatment has not been replaced by clofarabine, despite its superior cytotoxic efficacy. Notwithstanding, given its favorable risk/benefit ratio, a further evaluation of clofarabine in combination with PEG-ASP might be warranted as a second-line replacement or add-on strategy in specific patients, to reduce treatment-related morbidities or to augment the depth of molecular remission after antibody-based immunotherapy.21,22 In particular, clofarabine/PEG-ASP could be tested in high-risk patients and compared with other established anti-leukemic agents that are burdened with severe acute and long-term toxicities, such as anthracyclines or the anti-metabolite methotrexate.23,24

References 1. Xie KC, Plunkett W. Deoxynucleotide pool depletion and sustained inhibition of ribonucleotide reductase and DNA synthesis after treatment of human lymphoblastoid cells with 2-chloro-9-(2-deoxy-2-fluorobeta-D-arabinofuranosyl) adenine. Cancer Res. 1996;56(13):3030-3037. 2. Jeha S, Gandhi V, Chan KW, et al. Clofarabine, a novel nucleoside analog, is active in pediatric patients with advanced leukaemia. Blood. 2004;103(3):784-789. 3. Huang M, Inukai T, Miyake K, et al. Clofarabine exerts antileukemic activity against cytarabine-resistant B-cell precursor acute lymphoblastic leukemia with low deoxycytidine kinase expression. Cancer Med. 2018;7(4):1297-1316. 4. Huguet F, Leguay T, Raffoux E, et al. Clofarabine for the treatment of adult acute lymphoid leukemia: the Group for Research on Adult Acute Lymphoblastic Leukemia intergroup. Leuk Lymphoma. 2015;56(4):847-857. 5. Wang H, Jones AK, Dvorak CC, et al. Population pharmacokinetics of cofarabine as part of pretransplantation conditioning in pediatric subjects before hematopoietic cell transplantation. Biol Blood Marrow Transplant. 2019; 25(8):1603-1610. 6. Hochberg J, Zahler S, Geyer MB, et al. The safety and efficacy of clofarabine in combination with high-dose cytarabine and total body irradiation myeloablative conditioning and allogeneic stem cell transplantation in children, adolescents, and young adults (CAYA) with poor-risk acute leukemia. Bone Marrow Transplant. 2019;54(2):226-235. 7. Salzer WL, Burke MJ, Devidas M, et al. Toxicity associated with intensive postinduction therapy incorporating clofarabine in the very high-risk stratum of patients with newly diagnosed high-risk B-lymphoblastic

haematologica | 2022; 107(5)

Disclosures No conflicts of interest to disclose. Contributions MZ, MAH and GE designed the study with input from FS and UzS; DD, JF, TF, TI, NJ, AP, IS and FS recruited patients; MAH, GE and FS collected, analysed, and interpreted data. Acknowledgments Genzyme/Sanofi provided the investigational drug clofarabine. We thank Kseniya Bakharevich for her assistance in collecting and interpreting the data. We gratefully acknowledge all patients, their families and care providers who participated in this study. Finally, we thank all the clinicians, as well as diagnostics and research personnel who were actively involved in this clinical trial. Data sharing Individual patient data from the trial will not be shared publicly, since a data-sharing plan had not been included when ethical approval was requested. All original data can be obtained by the corresponding authors, please contact Dr. Gabriele Escherich: escherich@uke.de

leukemia: a report from the Children's Oncology Group study AALL1131. Cancer. 2018;124(6):1150-1159. 8. Escherich G, zur Stadt U, Zimmermann M, Horstmann MA, CoALL study group. Clofarabine in combination with pegylated asparaginase in the frontline treatment of childhood acute lymphoblastic leukaemia: a feasibility report from the CoALL 08-09 trial. Br J Haematol. 2013;163(2):240-247. 9. Schramm F, Zimmermann M, Jorch N, et al. Daunorubicin during delayed intensification decreases the incidence of infectious complications - a randomized comparison in trial CoALL 08-09. Leuk Lymphoma. 2019;60(1):60-68. 10. Van der Velden VHJ, Cazzaniga G, Schrauder A, et al. European Study Group on MRD detection in ALL (ESG-MRD-ALL). Analysis of minimal residual disease by Ig/TCR gene rearrangements: guidelines for interpretation of real-time quantitative PCR data. Leukemia. 2007;21(4):604-611. 11. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53(282):457-481. 12. Aalen OO. 1. The statistical analysis of failure time data (2nd edn). Kalbfleisch JD, Prentice LR, Wiley-Interscience, Hoboken, New Jersey, 2002. Statistics in Medicine. 2004;23(21):3397-3398. 13. Gray RJ. A Class of K-Sample Tests for comparing the cumulative incidence of a competing risk. The Annals of Statistics. 1988;16(3):1141-1154. 14. Margolin BH. Test for Trend in Proportions. In: Kotz S, Johnson NL, Read CB, eds. Encyclopedia of Statistical Sciences. New York: John Wiley & Sons. 1988;vol. 9:334336. 15. Borowitz MJ, Devidas M, Hunger SP, et al. Clinical significance of minimal residual disease in childhood acute lymphoblastic leukemia and its relationship to other prog-

nostic factors: a Children's Oncology Group study. Blood. 2008;111(12):5477-5485. 16. Brüggemann M, Kotrova M. Minimal residual disease in adult ALL: technical aspects and implications for correct clinical interpretation. Blood Adv. 2017;1(25):2456-2466. 17. Galimberti S, Devidas M, Lucenti A, et al. Validation of minimal residual disease as surrogate endpoint for event-free survival in childhood acute lymphoblastic leukemia. JNCI Cancer Spectr. 2018; 2(4):pky069. 18. Möricke A, Zimmermann M, Valsecchi MG, et al. Dexamethasone vs prednisone in induction treatment of pediatric ALL: results of the randomized trial AIEOP-BFM ALL 2000. Blood. 2016;127(17):2101-2112. 19. Borowitz MJ, Wood BL, Devidas M, et al. Prognostic significance of minimal residual disease in high risk B-ALL: a report from Children's Oncology Group study AALL0232. Blood. 2015;126(8):964-971. 20. Kantarjian HM, Gandhi V, Kozuch V, et al. Phase I clinical and pharmacology study of clofarabine in patients with solid and hematologic cancers. J Clin Oncol. 2003; 21(6):1167-1173. 21. Locatelli F, Whitlock JA, Peters C, et al. Blinatumomab versus historical standard therapy in pediatric patients with relapsed/refractory Ph-negative B-cell precursor acute lymphoblastic leukemia. Leukemia. 2020;34(9):2473-2478. 22. Curren E, Stock W. Taking a “BiTE out of ALL”: blinatumomab approval for MRDpositive ALL. Blood. 2019;133(16):17151719. 23. Armenian S, Bhatia S. Predicting and preventing anthracycline-related cardiotoxicity. Am Soc Clin Oncol Educ Book. 2018; 38:3-12. 24. Bhojwani D, Sabin ND, Pei D, et al. Methotrexate-induced neurotoxicity and leukoencephalopathy in childhood acute lymphoblastic leukemia. J Clin Oncol. 2014;32(9):949-959.

1033

ARTICLE Ferrata Storti Foundation

Haematologica 2022 Volume 107(5):1034-1044

Acute Myeloid Leukemia

Clinical and molecular relevance of genetic variants in the non-coding transcriptome of patients with cytogenetically normal acute myeloid leukemia Dimitrios Papaioannou,1,2* Hatice G. Ozer,3* Deedra Nicolet,1,4,5 Amog P. Urs,1 Tobias Herold,6-9 Krzysztof Mrózek,1,4 Aarif M.N. Batcha,10,11 Klaus H. Metzeler,12 Ayse S. Yilmaz,3 Stefano Volinia,13 Marius Bill,1 Jessica Kohlschmidt,1,4,5 Maciej Pietrzak,3 Christopher J. Walker,1,4 Andrew J. Carroll,14 Jan Braess,15 Bayard L. Powell,16 Ann-Kathrin Eisfeld,1,4 Geoffrey L. Uy,17 Eunice S. Wang,18 Jonathan E. Kolitz,19 Richard M. Stone,20 Wolfgang Hiddemann,6-8 John C. Byrd,1,4 Clara D. Bloomfield1# and Ramiro Garzon1# The Ohio State University, Comprehensive Cancer Center, Columbus, OH, USA; 2Laura and Isaac Perlmutter Cancer Center, New York University School of Medicine, NYU Langone Health, New York, NY, USA; 3The Ohio State University, Department of Biomedical Informatics, Columbus, OH, USA; 4The Ohio State University Comprehensive Cancer Center, Clara D. Bloomfield Center for Leukemia Outcomes Research, Columbus, OH, USA; 5Alliance Statistics and Data Center, The Ohio State University, Comprehensive Cancer Center, Columbus, OH, USA; 6Laboratory for Leukemia Diagnostics, Department of Medicine III, University Hospital, LMU Munich, Munich, Germany; 7German Cancer Consortium (DKTK), Heidelberg, Germany; 8German Cancer Research Center (DKFZ), Heidelberg, Germany; 9Research Unit Apoptosis in Hematopoietic Stem Cells, Helmholtz Zentrum München, German Center for Environmental Health (HMGU), Munich, Germany; 10Institute for Medical Information Processing, Biometry and Epidemiology, LMU Munich, Munich, Germany; 11Medical Data Integration Center (MeDIC), University Hospital, LMU Munich, Germany; 12Department of Hematology, Cell Therapy & Hemostaseology, University Hospital Leipzig, Leipzig, Germany; 13Department of Morphology, Surgery and Experimental Medicine, University of Ferrara, Ferrara, Italy; 14Department of Genetics, University of Alabama at Birmingham, Birmingham, AL, USA; 15Department of Oncology and Hematology, Hospital Barmherzige Brüder, Regensburg, Germany; 16The Comprehensive Cancer Center of Wake Forest University, Winston-Salem, NC, USA; 17Siteman Cancer Center, Washington University School of Medicine, St. Louis, MO, USA; 18Roswell Park Comprehensive Cancer Center, Buffalo, NY, USA; 19Monter Cancer Center, Hofstra Northwell School of Medicine, Lake Success, NY, USA and 20Dana-Farber Cancer Institute, Harvard University, Boston, MA, USA 1

Correspondence:

*DP and HGO contributed equally as co-first authors.

RAMIRO GARZON ramiro.garzon@osumc.edu

CDB and RG contributed equally as co-senior authors.

ABSTRACT Received: July 21, 2020. Accepted: July 2, 2021. Pre-published: July 15, 2021. https://doi.org/10.3324/haematol.2020.266643

1034

xpression levels of long non-coding RNA (lncRNA) have been shown to associate with clinical outcome of patients with cytogenetically normal acute myeloid leukemia (CN-AML). However, the frequency and clinical significance of genetic variants in the nucleotide sequences of lncRNA in AML patients is unknown. Herein, we analyzed total RNA sequencing data of 377 younger adults (aged <60 years) with CN-AML, who were comprehensively characterized with regard to clinical outcome. We used available genomic databases and stringent filters to annotate genetic variants unequivocally located in the non-coding transcriptome of AML patients. We detected 981 variants, which are recurrently present in lncRNA that are expressed in leukemic blasts. Among these variants, we identified a cytosine-to-thymidine variant in the lncRNA RP5-1074L1.4 and a cytosine-to-thymidine variant in the lncRNA SNHG15, which independently associated with longer survival of CN-AML patients. The presence of the SNHG15 cytosine-to-thymidine variant was also found to associate with better outcome in an independent dataset of CN-AML patients, despite differences in treatment protocols and RNA sequencing techniques. In order to gain biological insights, we cloned and overexpressed both wild-type and variant ver-

haematologica | 2022; 107(5)

lncRNA variants in CN-AML

sions of the SNHG15 lncRNA. In keeping with its negative prognostic impact, overexpression of the wild-type SNHG15 associated with higher proliferation rate of leukemic blasts when compared with the cytosine-to-thymidine variant. We conclude that recurrent genetic variants of lncRNA that are expressed in the leukemic blasts of CN-AML patients have prognostic and potential biological significance.

Introduction Acute myeloid leukemia (AML) is heterogeneous with regard to the patients’ clinical course and the underlying molecular lesions that drive the disease.1-2 Research efforts of the past four decades have identified a growing list of genetic alterations associated with clinical outcome that could be used as biomarkers for the risk stratification of the patients’ treatment. These alterations include chromosomal abnormalities,3-5 gene mutations,6-11 and aberrant expression of RNA transcripts.12-16 The advent of next-generation sequencing has revealed that AML displays notable heterogeneity at the level of isolated cases; leukemic blasts of individual AML patients represent, in many instances, the sum of distinct clonal subpopulations, within which mutations in different genes co-exist and co-operate.17,18 While such sequencing efforts continue to expand our understanding of AML pathogenesis, the majority of them are focused on the protein-coding fraction of the genome, which is, comparatively, its smallest part.19 The non-protein-coding part of the genome, a large fraction of which is actively transcribed into non-coding RNA, is gaining gradual recognition for its important regulatory role.20,21 Long non-coding RNA (lncRNA), which are transcripts longer than 200 nucleotides and, per definition, lack protein-coding potential, regulate many key cellular functions in health and disease.22-24 Deregulated expression of individual lncRNA has been demonstrated to significantly affect the cancer phenotype and patients’ clinical outcome.25-29 We and others have previously shown that aberrant expression of small subsets of lncRNA independently associate with the clinical outcome of patients with cytogenetically normal AML (CN-AML).30-33 With regard to variations in the nucleotide sequences of lncRNA, it has previously been reported that disease-associated single nucleotide polymorphisms (SNP) are enriched in the genetic loci encoding these transcripts.34-36 In addition, acquired mutations of lncRNA that are recurrently detectable in the leukemic blasts have previously been identified.37 However, to our knowledge, their prognostic and biologic significance in AML have not yet been studied. Herein, we analyzed whole transcriptome sequencing data of younger adults with the goal to evaluate the clinical and biological relevance of lncRNA variants in CN-AML. We used a panel of currently available databases of human genomic polymorphisms to annotate recurrent genetic variants located within expressed lncRNA. We show that a subset of these variants independently associates with clinical outcome of CN-AML patients.

Methods Patients and treatment Exploratory analysis was conducted in pretreatment bone marrow (BM) or blood samples from 377 younger adults (aged <60 years; range, 18-59 years) with de novo CN-AML. Patients were treated with intensive, first-line chemotherapy on Cancer and

haematologica | 2022; 107(5)

Leukemia Group B (CALGB)/Alliance for Clinical Trials in Oncology (Alliance) trials. Confirmatory analyses were conducted in a set of 135 CN-AML patients (75 of whom were younger than 60 years and 60 were older) enrolled on clinical trials of the German AML Cooperative Group (AMLCG).38,39 All patients provided written informed consent regarding the research use of their specimens. All study protocols were in accordance with the Declaration of Helsinki and approved by Institutional Review Boards at each center.

Cytogenetic and molecular analyses Cytogenetic analyses of CALGB/Alliance patients were performed in CALGB/Alliance-approved institutional laboratories and results were confirmed by central karyotype review.40 The diagnosis of normal karyotype was based on analysis of ≥20 metaphases obtained from BM specimens subjected to short-term (24- or 48-hour) unstimulated cultures.40 Mutational analyses of patient samples were conducted with Sanger sequencing (for the CEBPA gene), fragment analysis (for detection of FLT3-internal tandem duplications [FLT3-ITD]) and targeted amplicon sequencing (for all other prognostic gene mutations), as reported previously.31,41-43 Molecular and cytogenetic profiling of the AMLCG cohort were obtained as described previously.38,39

Transcriptome analyses RNA samples of the patients treated on CALGB/Alliance protocols were analyzed with total RNA sequencing (RNA Seq) after depletion of ribosomal and mitochondrial RNA using the Illumina HiSeq 2500 platform. The results of the RNA Seq analysis have been deposited in the functional genomics data repository GEO and are publicly available under the accession number GSE137851. Patients in the AMLCG cohort were analyzed with RNA Seq following selection for poly-adenylated transcripts (poly-A RNA Seq) as previously described.16 For exploratory analyses, after quality control, adaptor-trimmed 50 base-pair-long paired-end reads were mapped to the human reference genome and variant calling was performed following the Genome Analysis Toolkit best practice recommendations for RNA Seq datasets.44 A two-pass variant calling approach was applied to ensure variant detection and depth of coverage (Figure 1). Unique variant positions were identified on non-coding transcripts that do not overlap with coding exons and are located in low-complexity regions of the genome (i.e., excluding repeat masked regions and segmental duplications). These variants were further evaluated for associations with clinical outcome and the expression levels of other RNA transcripts.

Statistical analyses Clinical endpoint definitions are provided in the Online Supplementary Appendix. For each examined lncRNA variant, only patients with detectable expression of the lncRNA and adequate coverage of the variant position (i.e., depth of coverage >8) were analyzed. The estimated probabilities of disease-free (DFS), overall (OS) and event-free (EFS) survival were calculated using the Kaplan–Meier method, and the log-rank test evaluated differences between survival distributions. Cox proportional hazard models were used to calculate hazard ratios for DFS, OS and EFS.45

1035

D. Papaioannou et al.

Multivariable proportional hazards models were constructed using a backward selection procedure.45 All statistical analyses using CALGB/Alliance data were performed by the Alliance Statistics and Data Center.

detection of wild-type genotype in at least 5% of the samples, and iii) detection of variant allele frequency above 0.4 in at least 5% of the samples (Figure 1). Based on these criteria, 981 variants were selected for further analyses (Online Supplementary Table S1).

In vitro experiments LncRNA wild-type and variant transcripts were amplified with Phusion high fidelity polymerase by polymerase chain reaction (PCR). Amplicons were cloned into pcDNA using the Gibson technique, according to standard protocols. For primer sequences and further experimental details please see the Online Supplementary Appendix. K-562 and THP-1 cells were transfected with vectors containing either a cytosine (C)-to-thymidine (T) variant in the lncRNA SNHG15 (SNHG15varT) or wild-type lncRNA SNHG15 (SNHG15wt); cells were also transfected with empty pcDNA3.1 and were used as controls. Cell viability and apoptosis were assessed with annexin V staining. The colorimetric MTT assay was used to assess the proliferative capacity of the transfected blasts.

Results Detection of genetic variants in the non-coding transcriptome of younger adults with cytogenetically normal acute myeloid leukemia In order to examine whether recurrent genetic variants are present in the non-coding transcriptomes of CN-AML patients, we first analyzed total RNA sequencing data of 377 younger adults with CN-AML. In order to identify unequivocally non-coding genetic variants and to avoid ambiguity in their genomic location, we excluded from further analyses all variants, which overlapped with exons of protein-coding genes and those that mapped to segmental duplications or other repeat regions of the genome. In order to evaluate the clinical and functional relevance of the lncRNA variants, additional filters were applied. Specifically, we focused on the variants that displayed: i) adequate expression and coverage in at least 100 samples (approximately 25% of a total number of samples), ii)

Detection of long non-coding RNA variants in the Cancer Genome Atlas (TCGA) dataset In order to examine the validity and reproducibility of our experimental pipeline and results, we queried the publicly available TCGA total RNA Seq dataset.8 It is noteworthy that the TCGA dataset was generated with a different RNA Seq technique (i.e., poly-A RNA Seq), which is less suitable for the interrogation of the non-coding fraction of the transcriptome.46 In addition, a relatively small number of patients in the TCGA cohort represent CN-AML cases (i.e., 44 of the 196 available cases).8 Despite these limitations, 277 out of the 981 variants that we tested were detectable in the transcriptomes of the CN-AML cases included in the TCGA study (Online Supplementary Table S2). For a subset of TCGA cases, DNA sequencing data from both leukemic blasts and germline material are available in addition to the transcriptome sequencing data. We therefore sought: i) to validate whether the presence of the detected variants in the transcriptome is also detectable at the DNA level and ii) to examine whether these variants are bona fide acquired genetic events or are present in the germline configuration of the AML patient genomes. As the sequencing technique that was used to analyze the TCGA dataset (i.e., exome sequencing) preferentially captures and interrogates the coding fraction of the genome, only 20 variant positions were available for analyses at the DNA level. Overall, there was a complete concordance between the detection of a variant in the transcriptome and its detection in the genome. Furthermore, 11 of these variants were detected in both leukemic blasts and non-leukemic tissues and thus could be considered as germline genetic variants, whereas nine variants were only detectable in leukemic samples and could therefore represent acquired mutations (Online Supplementary Table S2).

Figure 1. Outline of the two-pass experimental approach for the identification of recurrent genetic variants located within long non-coding RNA in younger adult patients with cytogenetically normal acute myeloid leukemia. In the first pass, variant calling was performed on alignment results (i.e., BAM files) following the Genome Analysis Toolkit (GATK) best practice recommendations for RNA sequencing datasets. Variants of non-coding transcripts that do not overlap with coding exons and are not located in low-complexity regions of the genome were selected. In the second pass, Samtools pile-up programs were used to identify sequencing depth, quality and alternative allele counts on selected unique variant positions. Resulting visual component framework (VCF) files were consolidated with annotation and the final variant call matrix was generated.

1036

haematologica | 2022; 107(5)

lncRNA variants in CN-AML

A Figure 2. Outcome of younger adult patients with cytogenetically normal acute myeloid leukemia who harbored the C-to-T variant of the RP5-1074L1.4 long non-coding RNA (IncRNA) (RP5-1074L1.4varT) and of those who had the wild-type lncRNA (RP5-1074L1.4wt). (A) Disease-free survival, (B) overall survival and (C) event-free survival.

haematologica | 2022; 107(5)

1037

D. Papaioannou et al.

Prognostic significance of recurrent long non-coding RNA variants in acute myeloid leukemia We proceeded to examine the associations of the lncRNA variants with the clinical outcome of younger patients with CN-AML. Of the variants tested, 41 associated with more than one outcome endpoint (DFS, OS and/or EFS; Online Supplementary Table S3) above a threshold level of significance (P<0.05). There was no association between the presence of lncRNA variants and complete remission rates in our cohort. Among the variants that showed significant association with clinical outcome in our dataset were a C-to-T variant in the lncRNA RP5-1074L1.4 (RP5-1074L1.4varT), and SNHG15varT, a C-to-T variant in the lncRNA SNHG15. In patients with detectable RP5-1074L1.4 expression (n=243), the presence of the RP5-1074L1.4varT was found in 156 (64%) of them. Patients with the RP5-1074L1.4varT had longer DFS (P<0.001) than patients with wild-type RP5-1074L1.4 (RP5-1074L1.4wt); 5 years after diagnosis 44% of the RP5-1074L1.4varT patients were alive and leukemia-free in contrast to only 26% of those with RP51074L1.4wt. RP5-1074L1.4varT was also associated with longer EFS (P<0.001; 5-year rates: 38% vs. 21%) and showed a trend for longer OS (P=0.09; 5-year rates: 43% vs. 34%; Figures 2 A to C; Online Supplementary Table S4). Among patients who expressed the SNHG15 lncRNA (n=306), SNHG15varT was detected in 78% of them (n=239). Patients who expressed the SNHG15varT had longer DFS (P=0.04; 5-year rates: 37% vs. 22%) than patients who expressed the SNHG15wt. SNHG15varT expressers also had longer EFS (P=0.04; 5-year rates: 31% vs. 19%) and a trend for longer OS (P=0.07; 5-year rates: 41% vs. 32%) compared with SNHG15wt expressers (Figures 3A to C; Online Supplementary Table S5). Finally, we examined whether genetic variants were detectable in the set of 24 lncRNA, whose expression levels were previously shown to associate with the clinical outcome of younger adults with CN-AML.30 We found seven such variants in three of the prognostic lncRNA (annotated in bold lettering in the Online Supplementary Table S1). A guanosine (G)-to-C variant in the lncRNA AL122127.25 (AL122127.25varC) was the only one that associated with patient outcome. AL122127.25varC was detected in 72 of the 257 patients who expressed the AL122127.25 lncRNA (i.e., 28% of the AL122127.25 expressers). The presence of AL122127.25varC significantly associated with shorter DFS (P=0.01; 5-year rates: 17% vs. 35%), OS (P=0.01; 5-year rates: 22% vs. 40%) and EFS (P=0.002; 5-year rates: 12% vs. 30%; Online Supplementary Figures S1A to C; Online Supplementary Table S6) in younger adult CN-AML patients. Notably, the presence of AL122127.25varC had no impact on the expression levels of the AL122127.25 transcript, when compared to the AL122127.25wt (Online Supplementary Figure S2).

Associations of long non-coding RNA variants with pretreatment clinical and molecular characteristics Next, we examined potential associations of RP51074L1.4varT, SNHG15varT and AL122127.25varC with pretreatment clinical characteristics and prognostic molecular features of younger adult CN-AML patients. Overall, there were only minor differences in these features between variant and wild-type lncRNA expressers. Patients expressing the RP5-1074L1.4varT were more likely to harbor tyrosine kinase domain mutations of the FLT3 gene 1038

Table 1. Multivariable analyses of outcome in younger adult patients with cytogenetically normal acute myeloid leukemia by expression of the C-to-T variant of the RP5-1074L1.4 long non-coding RNA (IncRNA) (RP5-1074L1.4varT) versus the wild-type lncRNA (RP5-1074L1.4wt).

DFS Variables in final models

HR (95% CI)

RP5-1074L1.4, 0.54 varT vs. wild-type (0.37-0.77) FLT3-ITD, 2.29 present vs. absent (1.59-3.30) WT1, 1.85 mutated vs. wild-type (1.04-3.27) MN1 expression, 1.55 high vs. low (1.07-2.24) CEBPA, double-mutated vs. wild-type NPM1, mutated vs. wild-type

P <0.001 <0.001 0.04 0.02 -

EFS HR P (95% CI) 0.54 <0.001 (0.40-0.75) 2.16 <0.001 (1.54-3.02) 1.90 0.006 (1.20-3.02) 0.36 0.001 (0.21-0.61) 0.41 <0.001 (0.28-0.58)

DFS: disease-free survival; EFS: event-free survival; FLT3-ITD: internal tandem duplication of the FLT3 gene; HR: hazard ratio; CI: confidence intervals; varT: C-to-T variant; vs.: versus. NOTE: Hazard ratios greater than (less than) 1.0 indicate higher (lower) risk for relapse or death (disease-free survival) or death (overall survival) or for failure to achieve complete remission, relapse or death (event-free survival) for the first category listed for the categorical variables. Variables considered for model inclusion were: RP5-1074L1.4 (varT vs. wild-type), age (as a continuous variable, in 10-year increments), sex (male vs. female), race (white vs. non-white), white blood cell count (as a continuous variable, in 50-unit increments), hemoglobin (as a continuous variable, in 1-unit increments), platelet count (as a continuous variable, in 50-unit increments), extramedullary involvement (present vs. absent), ASXL1 mutations (mutated vs. wild-type), CEBPA mutations (double-mutated vs. single-mutated or wild-type), DNMT3A mutations (mutated vs. wild-type), FLT3-ITD (present vs. absent), FLT3-TKD (present vs. absent), IDH1 mutations (mutated vs. wild-type), IDH2 mutations (mutated vs. wild-type), NPM1 mutations (mutated vs. wild-type), RUNX1 mutations (mutated vs. wild-type), TET2 mutations (mutated vs. wild-type), WT1 mutations (mutated vs. wild-type), ERG expression levels (high vs. low), BAALC expression levels (high vs. low), MN1 expression levels (high vs. low), miR-181a expression levels (high vs. low), miR-3151 expression (expressed vs. not expressed), and miR-155 expression levels (high vs. low).

(FLT3-TKD) than patients who expressed the RP51074L1.4wt (P=0.03; 12% vs. 4%; Online Supplementary Table S7). Patients expressing the SNHG15varT were older (P=0.03), more likely to be Caucasian (P=0.02) and more likely to have low expression of the MN1 gene (P=0.05; Online Supplementary Table S8) than SNHG15wt expressers. With regard to the lncRNA AL122127.25, there were no significant differences in the clinical features or frequencies of prognostic gene mutations or gene expression between AL122127.25varC and AL122127.25wt expressers (Online Supplementary Table S9).

Multivariable analyses In order to examine prognostic significance of the detected lncRNA variants in the context of other established clinical and molecular prognostic markers, we constructed multivariable models. Expression of RP5-1074L1.4varT was a significant marker for longer DFS (hazard ratio [HR]: 0.54; P<0.001) and longer EFS (HR: 0.54; P<0.001) after adjusting for other covariates (Table 1). Expression of SNHG15varT significantly associated with longer DFS (HR: 0.63; P=0.02), longer OS (HR: 0.63; P=0.008) and longer EFS (HR: 0.68; P=0.02) after adjusting for other variables for each outcome endpoint (Table 2). Finally, expression of the AL122127.25varC was an independent marker of shorter OS (HR: 1.57; P=0.009) and shorter EFS (HR: 1.59; P=0.004) after adjusting for other covariates (Online Supplementary Table S10). haematologica | 2022; 107(5)

lncRNA variants in CN-AML

Table 2. Multivariable analyses of outcome in younger adult patients with cytogenetically normal acute myeloid leukemia by expression of the Cto-T variant of the SNHG15 long non-coding RNA (lncRNA) (SNHG15varT) versus the wild-type lncRNA (SNHG15wt).

Variables in final models SNHG15, varT vs. wild-type FLT3-ITD, present vs. absent WT1, mutated vs. wild-type MN1 expression, high vs. low DNMT3A, mutated vs. wild-type miR-155 expression, high vs. low Extramedullary disease, Present vs. absent Hemoglobin, continuous, 1-unit increments Age, continuous, 10-year increments NPM1, mutated vs. wild-type CEBPA, double-mutated vs. wild-type

DFS

EFS

HR (95% CI)

0.63 (0.34-0.81) 1.68 (1.08-2.60) 1.91 (1.01-3.55) 1.51 (1.04-2.20) 2.11 (1.43-3.12) 1.85 (1.18-2.91) 0.61 (0.39-0.95) 0.87 (0.79-0.96) -

0.02

0.008

0.02 <0.001

0.004

0.68 (0.22-0.57) 2.08 (1.57-2.76) -

0.03

0.63 (0.45-0.89) 2.48 (1.83-3.35) 1.78 (1.21-2.63) -

<0.001

0.007

0.03

0.004

0.89 (0.83-0.95) -

0.001

0.45 (0.34-0.62) 0.35 (0.22-0.57)

<0.001

0.89 (0.83-0.96) 1.29 (1.12-1.48) 0.51 (0.37-0.71) 0.49 (0.30-0.80)

0.02 0.04

<0.001

<0.001 0.005

<0.01

DFS: disease-free survival; OS: overall survival; EFS: event-free survival; FLT3-ITD: internal tandem duplications of the FLT3 gene; HR: hazard ratio; CI: confidence intervals; varT: Cto-T variant; vs.: versus. NOTE: Hazard ratios greater than (less than) 1.0 indicate higher (lower) risk for relapse or death (disease-free survival) or death (overall survival) or for failure to achieve complete remission, relapse or death (event-free survival) for the first category listed for the categorical variables. Variables considered for model inclusion were SNHG15 (varT vs. wild-type), age (as a continuous variable, in 10-year increments), sex (male vs. female), race (white vs. non-white), white blood cell count [(WBC) as a continuous variable, in 50-unit increments], hemoglobin (as a continuous variable, in 1-unit increments), platelet count (as a continuous variable, in 50-unit increments), extramedullary involvement (present vs. absent), ASXL1 mutations (mutated vs. wild-type), CEBPA mutations (double-mutated vs. single-mutated or wild-type), DNMT3A mutations (mutated vs. wild-type), FLT3-ITD (present vs. absent), FLT3-TKD (present vs. absent), IDH1 mutations (mutated vs. wild-type), IDH2 mutations (mutated vs. wild-type), NPM1 mutations (mutated vs. wild-type), RUNX1 mutations (mutated vs. wild-type), TET2 mutations (mutated vs. wild-type), WT1 mutations (mutated vs. wild-type), ERG expression levels (high vs. low), BAALC expression levels (high vs. low), MN1 expression levels (high vs. low), miR-181a expression levels (high vs. low), miR-3151 expression (expressed vs. not expressed), and miR-155 expression levels (high vs. low).

RP5-1074L1.4 is an intronic lncRNA, which is embedded in intron 7 of the protein-coding SLC16A4 transcript. In order to ensure that the prognostic effect of RP51074L1.4varT was not due to perturbation of the expression levels of SLC16A4, we compared the SLC16A4 transcript abundance between 87 patients who expressed the RP5-1074L1.4wt and 156 patients who expressed the RP51074L1.4varT. We found no significant difference in expression levels of SLC16A4 between these two patient groups (Online Supplementary Figure S3).

Evaluation of the prognostic significance of long non-coding RNA variants in an independent cohort of cytogenetically normal acute myeloid leukemia patients In order to examine whether our findings were reproducible in an independent cohort of CN-AML patients, we examined patients treated on AMLCG protocols,38,39 who had available clinical outcome data, and were analyzed with poly-A selected RNA Seq (n=135). As was the case with the TCGA dataset,8 the use of an alternative RNA Seq technique limited the number of lncRNA variants that could be detected and analyzed. The lncRNA transcripts, whose variants showed the strongest association with prognosis in our initial cohort were not captured by the poly-A RNA Seq and could not, therefore, be analyzed for associations with clinical outcome. Despite this limitation, there was concordance in the findings between the two cohorts. SNHG15varT was the haematologica | 2022; 107(5)

one detectable lncRNA variant, which associated with multiple outcome endpoints in the AMLCG dataset. Specifically, the SNHG15varT was detected in 103 of the 120 AMLCG CN-AML patients who expressed the SNHG15 lncRNA. In agreement with our findings in the CALGB/Alliance dataset, the presence of SNHG15varT associated with longer DFS (P=0.04; Figure 3D) and EFS (P=0.007, Figure 3E), but not OS (P=0.17), in the AMLCG cohort.

Expression levels of SNHG15 in normal hematopoiesis In order to further examine the functional significance of the lncRNA, which harbor prognostic variants we sought to determine their expression patterns during normal hematopoiesis. To this end, we used publicly available datasets of normal hematopoietic cells analyzed with microarrays or RNA Seq and deposited in the BloodSpot portal (www.bloodspot.eu). Of the lncRNA with prognostic genetic variants, only SNHG15 was annotated in the database and could be further analyzed. We found that SNHG15 was most abundantly expressed in common myeloid progenitors, granulocyte monocyte progenitors and megakaryocyte-erythroid progenitors. It was also highly expressed in hematopoietic stem cells and lymphoid cell populations. Among mature cell populations, SNHG15 was overexpressed in monocytes, whereas its expression levels were the lowest in polymorphonuclear leucocytes of the bone marrow and peripheral blood (Online Supplementary Figure S4). 1039

D. Papaioannou et al. A

Figure 3. Figure continued on following page.

1040

haematologica | 2022; 107(5)

lncRNA variants in CN-AML

Figure 3. Prognostic significance of long non-coding RNA variants across different cytogenetically normal acute myeloid leukemia cohorts. (A) Diseasefree survival, (B) overall survival and (C) event-free survival of younger adult cytogeneti-cally normal acute myeloid leukemia (CN-AML) patients in the CALGB/Alliance cohort who had the C-to-T variant of the SNHG15 long non-coding RNA (SNHG15varT) and of patients with the wild-type SNHG15 lncRNA (SNHG15wt). (D) Disease-free survival and (E) eventfree survival of CN-AML patients in the AMLCG cohort with the C-to-T variant of the SNHG15 lncRNA (SNHG15varT) and of those the wild-type SNHG15 lncRNA (SNHG15wt).

Functional relevance of prognostic long non-coding RNA variants We hypothesized that in addition to associations with outcome, the presence of genetic variants in lncRNA could have a functional impact and affect AML blast viability and proliferation. We focused on the lncRNA SNHG15varT, whose prognostic significance was validated in an independent cohort of AML patients. We isolated total RNA from AML cell lines and amplified both the SNHG15wt and SNHG15varT transcripts. We cloned the amplicons into pcDNA3.1 expression vectors and transfected two AML cell lines (i.e., K-562 and THP-1 cells). Expression levels of the SNHG15 lncRNA were similar in cells transfected with the SNHG15wt and those transfected with SNHG15varT-containing vectors, when compared with the empty vector controls (Figures 4A and B). Regarding cell haematologica | 2022; 107(5)

viability, ectopic overexpression of the SNHG15wt and the SNHG15varT led to a discreet but consistent decrease in cell viability across cell lines, which did not reach statistical significance (Figures 4C and D). In order to further evaluate the effect of SNHG15 on the growth kinetics of the AML cells, we performed colorimetric MTT assays. Forced overexpression of the SNHG15varT had no significant effect on blast growth when compared with controls. In contrast, overexpression of the SNHG15wt led to increased proliferative capacity of the leukemic cells when compared with both controls and cells overexpressing the SNHG15varT (Figures 4E and F). These findings are in line with our prognostic observations and the inferior outcome of CN-AML patients who express SNHG15wt, compared with those who express SNHG15varT. 1041

D. Papaioannou et al.

Discussion lncRNA are gaining recognition as important molecular mediators and regulators of key cellular functions in health and disease.22-29 In AML, lncRNA have been shown to associate with the clinical outcome of both younger and older patients.30-33 However, these previous studies have focused on the expression levels of the lncRNA transcripts; the effect of genetic variants within lncRNA has not been extensively studied. Accumulating evidence suggests that such genetic variants could impact on the function of the lncRNA and be relevant for disease pathogenesis.47 In support of this view, disease-associated SNP are more frequently found in regions of the genome that encode for non-coding RNA transcripts in several types of solid tumors34,35 and in AML.34,36 These findings suggest that the presence of a variant in the non-coding transcriptome could be the functional link that explains how genetic variants that do not alter the structure of protein molecules associate with malignant phenotypes. Recently, Klco et al.37 have reported the acquisition of somatic mutations in lncRNA transcripts of AML patients, as demonstrated by analyses of the leukemic blasts in parallel with germline material. However, the prognostic value of these mutations could not be tested due to sample size limitations. Herein, we analyzed total RNA Seq data of younger adult patients with CN-AML with the goal to detect recurrent lncRNA variants and evaluate their prognostic and biologic significance. We used a stringent approach for detecting and filtering sequence variations of lncRNA. We generated a list of 981 recurrent variants, which are located within lncRNA in younger adult patients constituting the CALGB/Alliance cohort, and which had adequate sample sizes for meaningful survival analyses. In order to study potential associations of the detected variants with clinical outcome, we individualized analysis for each variant and limited comparisons to the group of patients that were expressers of the corresponding lncRNA transcript (i.e., we compared the expressers of a variant to the expressers of the wild-type lncRNA). Of the 981 candidate variants, a subset of 41 significantly associated with at least two clinical outcome endpoints of younger adults with CN-AML. LncRNA genetic variants RP51074L1.4varT and SNHG15varT were significantly associated with prognosis. RP5-1074L1.4 has not been previously associated with cancer pathogenesis or clinical outcome of cancer patients. In contrast, SNHG15, a MYC-regulated lncRNA, which harbors a small nucleolar RNA in one of its introns, has been implicated in the pathogenesis of multiple solid malignancies. SNHG15 has been shown to interact with the protein AIF and to associate with a clinically aggressive and prognostically unfavorable subset of colorectal carcinomas.48 SNHG15 has also been associated with aggressive phenotypes of hepatocellular and breast carcinomas via sponging and inhibiting the function of microRNAs 141-3p and 211-3p.49,50 Patients who expressed the RP5-1074L1.4varT had more favorable outcome (i.e., longer DFS, EFS and a trend for longer OS) than RP5-1074L1.4wt expressers. SNHG15varT also associated with better prognosis and SNHG15varT expressers had longer DFS and EFS compared with patients who expressed the SNHG15wt lncRNA. Further, we examined whether recurrent genetic variants could be detected in lncRNA previously reported to be prognostic in younger adults with CN-AML.30 A G-to-C variant in the prognostic lncRNA AL122127.25 1042

(AL122127.25varC) significantly associated with clinical outcome of CN-AML patients. Specifically, expression of the AL122127.25varC associated with shorter DFS, OS and EFS. Notably, the expression levels of the AL122127.25 lncRNA were not significantly affected by the presence of the AL122127.25varC. Thus, both the abundance and the nucleotide sequence of the AL122127.25 lncRNA associate with the clinical outcome of CN-AML patients via potentially independent mechanisms. Currently, there is limited availability of AML datasets analyzed with RNA Seq techniques (i.e., with total RNA Seq) that are suitable for in-depth analyses of the non-coding transcriptome. In order to examine whether our observations could be reproduced in other datasets, we examined the presence of our curated variant list in an independent cohort of CN-AML patients who were treated on AMLCG protocols. As expected, the use of a different RNA Seq technique significantly limited the number of lncRNA transcripts that we could interrogate. Nevertheless, SNHG15varT was detectable and associated with longer DFS and EFS of the AMLCG patients, similarly to its prognostic effect in the CALGB/Alliance dataset. Given the constantly increasing number of prognostic molecular alterations in AML, it is important to examine the prognostic value of novel markers in the context of other established clinical and molecular prognosticators. It is thus noteworthy that the lncRNA variants that we examined (RP5-1074L1.4varT, SNHG15varT, and AL122127.25varC) did not show associations with gene mutations that are currently used for the risk stratification of the treatment of AML patients.11 Mutations in the CEBPA, RUNX1, ASXL1 and NPM1 genes as well as the presence of the FLT3-ITD were similarly distributed between expressers of the lncRNA variants and expressers of the wild-type transcripts. In addition, in formal multivariable analyses that included prognostic clinical and molecular parameters, RP5-1074L1.4varT, SNHG15varT, and AL122127.25varC retained their prognostic significance after adjusting for other covariates. Consequently, detection of these variants could provide additional prognostic information and further refine risk-stratification of CN-AML patients. In addition to testing their prognostic significance, we sought to evaluate whether the presence of recurrent variants in lncRNA sequences has functional implications. We focused on SNHG15 lncRNA and performed overexpression experiments with SNHG15wtand SNHG15varT-containing vectors in the K-562 and THP-1 cell lines. Overexpression of the SNHG15varT had no significant effect on cell viability and no evident impact on cell proliferation, compared with controls. In contrast, overexpression of the SNHG15wt led to a significant increase in the proliferative capacity of the leukemic blasts in both cell lines that were tested. Despite the limitations of in vitro assays, these results indicate that expression of SNHG15wt associates with a more aggressive disease phenotype, when compared with SNHG15varT, and are in line with the inferior clinical outcome of SNHG15wt-expressers. In summary, we have performed comprehensive characterization of genetic variants within lncRNA in younger adults with CN-AML. We present analyses that support the prognostic and potential biologic significance of lncRNA variants in CN-AML. We believe that our work will serve as a useful starting point for further studies on the role of haematologica | 2022; 107(5)

lncRNA variants in CN-AML

Figure 4. Biologic significance of recurrent long non-coding RNA variants in younger adult cytogenetically normal acute myeloid leukemia patients. Fold changes of SNHG15 long non-coding RNA (lncRNA) expression levels (A and B), percent of viable cells (C and D) and proliferation (E and F) of acute myeloid leukemia cell lines transfected with empty pcDNA3, SNHG15wt-containing or SNHG15varT-containing vectors. Results for K-562 (A, C and E) and THP-1 (B, D and F) cells are depicted. Proliferation is assessed with the MTT colorimetric assay, by light absorbance. *P<0.05; **P<0.01; ***P<0.001; N.S: not significant.

genetic variants in the non-coding transcriptome of cancer patients. Disclosures No conflicts of interest to disclose. Contributions DP, HGO, DN, CDB and RG designed the study; DP, HGO, DN, TH, KM, AMNB, SV, ASY, MP, CJW, MB, WH, JCB, CDB and RG contributed to the data interpretation; DP, DN, KM, CDB and RG wrote the manuscript; HGO, DN and JK performed statistical analysis; DP and APU designed and performed experiments; TH, KM, KHM, AJC, JB, BLP, A-KE, GLU, ESW, JEK, RMS, WH, JCB, CDB and RG were involved directly or indirectly in the care of patients and/or sample procurement; KM and AJC oversaw the cytogenetic analyses. CDB and RG supervised the study. All authors read and approved the final version of the manuscript. Acknowlegments The authors would like to thank: the Alliance Hematology

References 1. Döhner H, Weisdorf DJ, Bloomfield CD. Acute myeloid leukemia. N Engl J Med. 2015;373(12):1136-1152. 2. Döhner H, Estey E, Grimwade D, et al. Diagnosis and management of AML in

haematologica | 2022; 107(5)

Malignancy Biorepository supported by Washington University subcontract WU-15-398/WU-16-51 and the Alliance NCTN Biorepository and Biospecimen Resource (U24CA196171) for sample processing and storage services, and Lisa J. Sterling (The Ohio State University, Comprehensive Cancer Center, Columbus, OH 43221) for data management. Celebrating the life and accomplishments of Dr. Clara D. Bloomfield (1942-2020). Funding Research reported in this publication was supported in part by the National Cancer Institute of the National Institutes of Health under award numbers U10CA180821 and U10CA180882 (to the Alliance for clinical trials in oncology), U10CA077658, U10CA180850, U10CA180861, CA140158, CA16058, UG1CA233338 and R35CA197734. This work was also supported in part by the Leukemia Clinical Research Foundation, D. Warren Brown Foundation and the Pelotonia Fellowship Program. TH was supported by a Physician Scientists grant (G-509200004) from the Helmholtz Zentrum München. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

adults: 2017 ELN recommendations from an international expert panel. Blood. 2017;129 (4):424-447. 3. Byrd JC, Mrózek K, Dodge RK, et al. Pretreatment cytogenetic abnormalities are predictive of induction success, cumulative incidence of relapse, and overall survival in adult patients with de novo acute myeloid

leukemia: results from Cancer and Leukemia Group B (CALGB 8461). Blood. 2002;100 (13):4325-4336. 4. Grimwade D, Hills RK, Moorman AV, et al. Refinement of cytogenetic classification in acute myeloid leukemia: determination of prognostic significance of rare recurring chromosomal abnormalities among 5876

1043

D. Papaioannou et al. younger adult patients treated in the United Kingdom Medical Research Council trials. Blood. 2010;116(3):354-365. 5. Mrózek K, Heerema NA, Bloomfield CD. Cytogenetics in acute leukemia. Blood Rev. 2004;18(2):115-136. 6. Papaemmanuil E, Gerstung M, Bullinger L, et al. Genomic classification and prognosis in acute myeloid leukemia. N Engl J Med. 2016;374(23):2209-2221. 7. Patel JP, Gönen M, Figueroa ME, et al. Prognostic relevance of integrated genetic profiling in acute myeloid leukemia. N Engl J Med. 2012;366(12):1079-1089. 8. Cancer Genome Atlas Research Network, Ley TJ, Miller C, et al. Genomic and epigenomic landscapes of adult de novo acute myeloid leukemia. N Engl J Med. 2013;368 (22):2059-2074. 9. Metzeler KH, Herold T, Rothenberg-Thurley M, et al. Spectrum and prognostic relevance of driver gene mutations in acute myeloid leukemia. Blood. 2016;128(5):686-698. 10. Eisfeld A-K, Kohlschmidt J, Mrózek K, et al. Mutation patterns identify adult patients with de novo acute myeloid leukemia aged 60 years or older who respond favorably to standard chemotherapy: an analysis of Alliance studies. Leukemia. 2018;32(6):13381348. 11. Mrózek K, Marcucci G, Paschka P, Whitman SP, Bloomfield CD. Clinical relevance of mutations and gene-expression changes in adult acute myeloid leukemia with normal cytogenetics: are we ready for a prognostically prioritized molecular classification? Blood. 2007;109(2):431-448. 12. Valk PJM, Verhaak RGW, Beijen MA, et al. Prognostically useful gene-expression profiles in acute myeloid leukemia. N Engl J Med. 2004;350(16):1617-1628. 13. Metzeler KH, Hummel M, Bloomfield CD, et al. An 86-probe-set gene-expression signature predicts survival in cytogenetically normal acute myeloid leukemia. Blood. 2008;112(10):4193-4201. 14. Li Z, Herold T, He C, et al. Identification of a 24-gene prognostic signature that improves the European LeukemiaNet risk classification of acute myeloid leukemia: an international collaborative study. J Clin Oncol. 2013;31(9):1172-1181. 15. Marcucci G, Yan P, Maharry K, et al. Epigenetics meets genetics in acute myeloid leukemia: clinical impact of a novel sevengene score. J Clin Oncol. 2014;32(6):548556. 16. Herold T, Jurinovic V, Batcha AMN, et al. A 29-gene and cytogenetic score for the prediction of resistance to induction treatment in acute myeloid leukemia. Haematologica. 2018;103(3):456-465. 17. Welch JS, Ley TJ, Link DC, et al. The origin and evolution of mutations in acute myeloid leukemia. Cell. 2012;150(2):264-278. 18. Klco JM, Spencer DH, Miller CA, et al. Functional heterogeneity of genetically defined subclones in acute myeloid leukemia. Cancer Cell. 2014;25(3):379-392. 19. Lander ES, Linton LM, Birren B, et al. Initial sequencing and analysis of the human genome. Nature. 2001;409(6822):860-921. 20. Kellis M, Wold B, Snyder MP, et al. Defining functional DNA elements in the human

1044

genome. Proc Natl Acad Sci U S A. 2014;111(17):6131-6138. 21. Taylor J. Clues to function in gene deserts. Trends Biotechnol. 2005;23(6):269-271. 22. Rinn JL, Chang HY. Genome regulation by long noncoding RNAs. Annu Rev Biochem. 2012;81:145-166. 23. Guttman M, Rinn JL. Modular regulatory principles of large non-coding RNAs. Nature. 2012;482(7385):339-346. 24. Wang KC, Chang HY. Molecular mechanisms of long noncoding RNAs. Mol Cell. 2011;43(6):904-914. 25. Gupta RA, Shah N, Wang KC, et al. Long non-coding RNA HOTAIR reprograms chromatin state to promote cancer metastasis. Nature. 2010;464(7291):1071-1076. 26. Leucci E, Vendramin R, Spinazzi M, et al. Melanoma addiction to the long non-coding RNA SAMMSON. Nature. 2016;531(7595): 518-522. 27. Trimarchi T, Bilal E, Ntziachristos P, et al. Genome-wide mapping and characterization of Notch-regulated long noncoding RNAs in acute leukemia. Cell. 2014;158(3): 593-606. 28. Papaioannou D, Petri A, Dovey OM, et al. The long non-coding RNA HOXB-AS3 regulates ribosomal RNA transcription in NPM1-mutated acute myeloid leukemia. Nat Commun. 2019;10(1):5351. 29. Ji P, Diederichs S, Wang W, et al. MALAT-1, a novel noncoding RNA, and thymosin β4 predict metastasis and survival in early-stage non-small cell lung cancer. Oncogene. 2003;22(39):8031-8041. 30. Papaioannou D, Nicolet D, Volinia S, et al. Prognostic and biologic significance of long non-coding RNA profiling in younger adults with cytogenetically normal acute myeloid leukemia. Haematologica. 2017;102(8): 1391-1400. 31. Garzon R, Volinia S, Papaioannou D, et al. Expression and prognostic impact of lncRNAs in acute myeloid leukemia. Proc Natl Acad Sci U S A. 2014;111(52):1867918684. 32. Beck D, Thoms JAI, Palu C, et al. A fourgene LincRNA expression signature predicts risk in multiple cohorts of acute myeloid leukemia patients. Leukemia. 2018;32(2): 263-272. 33. Mer AS, Lindberg J, Nilsson C, et al. Expression levels of long non-coding RNAs are prognostic for AML outcome. J Hematol Oncol. 2018;11(1):52. 34. Yan X, Hu Z, Feng Y, et al. Comprehensive genomic characterization of long non-coding RNAs across human cancers. Cancer Cell. 2015;28(4):529-540. 35. Iyer MK, Niknafs YS, Malik R, et al. The landscape of long noncoding RNAs in the human transcriptome. Nat Genet. 2015;47(3):199-208. 36. Schwarzer A, Emmrich S, Schmidt F, et al. The non-coding RNA landscape of human hematopoiesis and leukemia. Nat Commun. 2017;8(1):218. 37. Klco JM, Miller CA, Griffith M, et al. Association between mutation clearance after induction therapy and outcomes in acute myeloid leukemia. JAMA. 2015;314 (8):811-822. 38. Büchner T, Berdel WE, Schoch C, et al.

Double induction containing either two courses or one course of high-dose cytarabine plus mitoxantrone and postremission therapy by either autologous stem-cell transplantation or by prolonged maintenance for acute myeloid leukemia. J Clin Oncol. 2006;24(16):2480-2489. 39. Braes J, Amler S, Kreuzer K-A, et al. Sequential high-dose cytarabine and mitoxantrone (S-HAM) versus standard double induction in acute myeloid leukemia―a phase 3 study. Leukemia. 2018;32(12):25582571. 40. Mrózek K, Carroll AJ, Maharry K, et al. Central review of cytogenetics is necessary for cooperative group correlative and clinical studies of adult acute leukemia: the Cancer and Leukemia Group B experience. Int J Oncol. 2008;33(2):239-244. 41. Eisfeld A-K, Mrózek K, Kohlschmidt J, et al. The mutational oncoprint of recurrent cytogenetic abnormalities in adult patients with de novo acute myeloid leukemia. Leukemia. 2017;31(10):2211-2218. 42. Marcucci G, Maharry K, Radmacher MD, et al. Prognostic significance of, and gene and microRNA expression signatures associated with, CEBPA mutations in cytogenetically normal acute myeloid leukemia with highrisk molecular features: a Cancer and Leukemia Group B study. J Clin Oncol. 2008;26(31):5078-5087. 43. Whitman SP, Archer KJ, Feng L, et al. Absence of the wild-type allele predicts poor prognosis in adult de novo acute myeloid leukemia with normal cytogenetics and the internal tandem duplication of FLT3: a Cancer and Leukemia Group B study. Cancer Res. 2001;61(19):7233-7239. 44. Nekrutenko A, Taylor J. Next-generation sequencing data interpretation: enhancing reproducibility and accessibility. Nat Rev Genet. 2012;13(9):667-672. 45. Vittinghoff E, Glidden DV, Shiboski SC, McCulloch CE. Regression methods in biostatistics: linear, logistic, survival and repeated measures models. New York, NY: Springer; 2005. 46. Zhao W, He X, Hoadley KA, Parker JS, Hayes DN, Perou CM. Comparison of RNASeq by poly (A) capture, ribosomal RNA depletion, and DNA microarray for expression profiling. BMC Genomics. 2014;15 (1):419. 47. Khurana E, Fu Y, Chakravarty D, Demichelis F, Rubin MA, Gerstein M. Role of non-coding sequence variants in cancer. Nat Rev Genet. 2016;17(2):93-108. 48. Saeinasab M, Bahrami AR, González J, et al. SNHG15 is a bifunctional MYC-regulated noncoding locus encoding a lncRNA that promotes cell proliferation, invasion and drug resistance in colorectal cancer by interacting with AIF. J Exp Clin Cancer Res. 2019;38(1):172. 49. Ye J, Tan L, Fu Y, et al. LncRNA SNHG15 promotes hepatocellular carcinoma progression by sponging miR-141-3p. J Cell Biochem. 2019;120(12):19775-19783. 50. Kong Q, Qiu M. Long noncoding RNA SNHG15 promotes human breast cancer proliferation, migration and invasion by sponging miR-211-3p. Biochem Biophys Res Commun. 2018;495(2):1594-1600.

haematologica | 2022; 107(5)

ARTICLE

Bone Marrow Transplantation

One and a half million hematopoietic stem cell transplants: continuous and differential improvement in worldwide access with the use of non-identical family donors Dietger Niederwieser,1,2,3 Helen Baldomero,4 Nosa Bazuaye,5,6 Caitrin Bupp,7 Naeem Chaudhri,8,9 Selim Corbacioglu,10,11 Alaa Elhaddad,5,12 Cristóbal Frutos,13,14 Sebastian Galeano,13,15 Nada Hamad,16,17 Amir Ali Hamidieh,8,18 Shahrukh Hashmi,19,20 Aloysius Ho,21,22 Mary M. Horowitz,23 Minako Iida,21,24 Gregorio Jaimovich,13,25 Amado Karduss,13,26 Yoshihisa Kodera,21,24 Nicolaus Kröger,10,27 Regis Peffault de Latour,10,28 Jong Wook Lee,21,29 Juliana Martínez-Rolón,13,30 Marcelo C. Pasquini,23 Jakob Passweg,10,31 Kristjan Paulson,32 Adriana Seber,13,33 John A. Snowden,10,34 Alok Srivastava,21,35 Jeff Szer,16,36 Daniel Weisdorf,7,37 Nina Worel,38 Mickey B. C. Koh,39,40 Mahmoud Aljurf,8,41 Hildegard Greinix,42 Yoshiko Atsuta21,43 and Wael Saber23 for the Worldwide Network of Blood and Marrow Transplantation WBMT University of Leipzig, Germany; 2Aichi Medical University School of Medicine, Nagakute, Japan; 3Lithuanian University of Health Sciences, Kaunas, Lithuania; 4The Worldwide Network of Blood and Marrow Transplantation (WBMT) Transplant Activity Survey Office, University Hospital, Basel, Switzerland; 5African Blood and Marrow Transplantation Group – AfBMT, Bern, Switzerland; 6University of Benin Teaching Hospital, Benin, Nigeria; 7CIBMTR (Center for International Blood and Marrow Transplant Research), National Marrow Donor Program/Be The Match, Minneapolis, MN; USA; 8The Eastern Mediterranean Blood and Marrow Transplant Group (EMBMT), King Faisal Specialist Hospital & Research Center, Riyadh, Saudi Arabia; 9Oncology Center King Faisal Specialist Hospital & Research Center, Riyadh, Saudi Arabia; 10 European Society for Blood and Marrow Transplantation (EBMT), Barcelona, Spain; 11 Department of Pediatric Hematology, Oncology and Stem Cell Transplantation, University of Regensburg, Regensburg, Germany; 12Department of Pediatric Oncology and Stem Cell Transplantation Unit, Cairo University, Cairo, Egypt; 13Latin American Blood and Marrow Transplantation Group – LABMT, Bern, Switzerland; 14Cristóbal Frutos, Instituto de Previsión Social, Asunción, Paraguay; 15Sebastian Galeano, Hospital Británico, Montevideo, Uruguay; 16Australasian Bone Marrow Transplant Recipient Registry (ABMTRR), St. Vincent´s Hospital Sydney, Sydney, New South Wales, Australia; 17St. Vincent's Health Network, Kinghorn Cancer Center, Sydney, New South Wales, Australia; 18Pediatric Cell Therapy Research Center, Tehran University of Medical Sciences, Tehran, Iran; 19Sheikh Shakhbout Medical City, Abu Dhabi, UAE; 20 MAYO Clinic, Rochester, MN, USA; 21The Asia Pacific Blood and Marrow Transplant Group (APBMT), Aichi Medical University School of Medicine, Nagakute, Japan; 22 Singapore General Hospital Singapore, Singapore; 23CIBMTR, Medical College of Wisconsin, Milwaukee, WI, USA; 24Aichi Medical University School of Medicine, Deptartment of Promotion for Blood and Marrow Plantation, Nagakute, Japan; 25 Fundación Favaloro, Sanatorio Anchorena, ITAC, Buenos Aires, Argentina; 26Instituto de Cancerología-Clínica Las Américas, Medellín, Colombia; 27Department of Stem Cell Transplantation, University Medical Center Hamburg, Hamburg, Germany; 28Saint-Louis Hospital, 1 avenue Claude Vellefaux, Paris, France; 29Seoul St. Mary’s Hospital, The Catholic University of Korea, Seoul, Republic of Korea; 30FUNDALEU, Buenos Aires, Argentina; 31Klinik für Hämatologie, Universitätsspital Basel, Basel, Switzerland; 32 Cancercare Manitoba and the University of Manitoba and Cell Therapy Transplant Canada (CTTC), Winnipeg, Manitoba, Canada; 33Pediatric Department, Hospital Samaritano, Sao Paulo, Brazil; 34Department of Hematology, Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UK; 35Christian Medical College, Vellore, India; 36Peter MacCallum Cancer Center and Royal Melbourne Hospital, Parkville, Vicoria, Australia; 37University of Minnesota, Minneapolis, MN, USA; 38Medical University of Vienna, Department of Blood Group Serology and Transfusion Medicine, Vienna, Austria; 39Infection and Immunity Clinical Academic Group St George’s Hospital and Medical School, London, UK; 40Academic Cell Therapy Facility and Programme Health Sciences Authority Singapore, Singapore; 41King Faisal Specialist Hospital and Research Center, Riyadh, Saudi Arabia; 42Medical University of Graz, Division of Hematology, Graz, Austria and 43Japanese Data Center for Hematopoietic Cell Transplantation (JDCHCT), Nagoya, Japan

Ferrata Storti Foundation

Haematologica 2022 Volume 107(5):1045-1053

haematologica | 2022; 107(5)

Correspondence: DIETGER NIEDERWIESER dietger.niederwieser@medizin.uni-leipzig.de Received: May 13, 2021. Accepted: July 21, 2021. Pre-published: August 12, 2021. https://doi.org/10.3324/haematol.2021.279189

1045

D. Niederwieser et al.

ABSTRACT

he Worldwide Network of Blood and Marrow Transplantation (WBMT) pursues the mission of promoting hematopoietic cell transplantation (HCT) for instance by evaluating activities through member societies, national registries and individual centers. In 2016, 82,718 first HCT were reported by 1,662 HCT teams in 86 of the 195 World Health Organization member states representing a global increase of 6.2% in autologous HCT and 7.0% in allogeneic HCT and bringing the total to 1,298,897 procedures. Assuming a frequency of 84,000/year, 1.5 million HCT were performed by 2019 since 1957. Slightly more autologous (53.5%) than allogeneic and more related (53.6%) than unrelated HCT were reported. A remarkable increase was noted in haploidentical related HCT for leukemias and lymphoproliferative diseases, but even more in non-malignant diseases. Transplant rates (TR; HCT/10 million population) varied according to region reaching 560.8 in North America, 438.5 in Europe, 76.7 in Latin America, 53.6 in South East Asia/Western Pacific (SEA/WPR) and 27.8 in African/East Mediterranean (AFR/EMR). Interestingly, haploidentical TR amounted to 32% in SEA/WPR and 26% in Latin America, but only 14% in Europe and EMR and 4.9% in North America of all allogeneic HCT. HCT team density (teams/10 million population) was highest in Europe (7.7) followed by North America (6.0), SEA/WPR (1.9), Latin America (1.6) and AFR/EMR (0.4). HCT are increasing steadily worldwide with narrowing gaps between regions and greater increase in allogeneic compared to autologous activity. While related HCT is rising, largely due to increase in haploidentical HCT, unrelated HCT is plateauing and cord blood HCT is in decline.

Introduction Allogeneic and autologous hematopoietic stem cell transplantation (HCT) is considered a routine but complex therapy for patients with otherwise incurable chemo- and immune-sensitive malignant and non-malignant disorders.1 The treatment is also used for replacing deficient hematopoietic cells or cellular components and more recently for repairing hematopoietic stem cells by gene editing. Despite its increasing applications and international expansion of access, allogeneic HCT is still associated with significant morbidity and mortality and remains an example of highly specialized, high-cost medicine. It requires extensive experience, significant infrastructure and a network of specialists from all fields of medicine. Over the last two decades in particular, allogeneic HCT has undergone a constant technological evolution, with decreasing transplant related-morbidity and mortality and expansion of the donor pool. This has been achieved by optimizing indications, by manipulating alloreactive immune reactions ex vivo and in vivo and by using novel reduced or minimal intensity conditioning regimens.2 As a result, HCT is being offered to patients without a matched donor, to older patients and to those with comorbidities.3-5 The predominant autologous HCT transplant type, in contrast, relies exclusively on the highdose preparative regimen for tumor eradication or for reshaping the immune system in autoimmune diseases.6,7 Patients’ own hematopoietic stem cells are required to rebuild a normal hematopoietic system after the intensive preparative regimen. Missing graft-versus-host disease leads to extremely low mortality, but missing graft-versustumor effects to high relapse rates. It is not surprising that autologous HCT involves different treatment strategies and indications as compared to allogeneic HCT, but like allogeneic HCT it requires extensive experience, significant infrastructure and a network of specialists from all fields of medicine. However, the increasing specialization and complexity of health care systems required, threaten global equity of access to HCT. The World Health Organization (WHO; 1046

www.who.org) declared the transplantation of organs, cells and tissues a global priority and formed a task force to address quality, safety and equity of access. In order to achieve this, analysis of baseline global activity and evolving trends is essential.8 The Worldwide Network for Blood and Marrow Transplantation (WBMT; www.wbmt.org), is a non-governmental (NGO) umbrella organization in the field of HCT and in collaboration with the WHO, has taken up the challenge of collecting and disseminating global HCT activity data on a regular basis. Information on indications, the use of different technologies, donor types and trends over time provide a sound basis for physicians to provide appropriate patient counseling and for health care agencies to develop the necessary infrastructure. Informed by global activity survey data, the WBMT performs worldwide workshops to support the development of new HCT programs and to optimize existing programs. The ability to share accumulated experience covering a wide range of strategies, successes and pitfalls continues to be key in improving global HCT access for patients in need. The first WBMT HCT activity report was based on the global HCT activity in 2006.8 This was followed by an updated report in 2010.9 After reaching the global milestone of one million HCT in 2012,10 the WBMT focused on analyzing major trends from 2006 onwards11 and noted a narrowing of gaps in the African/East Mediterranean (AFR/EMRO) regions.12 The success of HCT depends on a number of factors including, the early and effective control of the underlying disease prior to HCT, risk of relapse of the underlying disease and donor characteristics. Over the past decade, the rapid evolution of molecular diagnostic and prognostic techniques has led to the emergence of more accurate prognostic tools and effective targeted molecular therapies for malignant hematological diseases. We report on the global activity trends between 2014 and 2016 compared to 2006 with a specific focus on global trends in equity of access, in indications of HCT and in donor type. haematologica | 2022; 107(5)

HCT worldwide access

Methods Study design In this retrospective observational survey we analyzed the worldwide HCT activity from the first published series of bone marrow transplants collected from the scientific literature and from member societies for very early transplants.11 After 2006, activities were obtained annually through the WBMT network using a unified center-based reporting system. Information from 2016 onwards is given as prediction based on currently available incomplete and non-validated data. Main outcome measures were the spread of HCT over time and transplant by donor type, country of origin, and WHO region. Secondary outcome measures were to document any trends in the number of HCT by donor type or region, to classify these trends, and quantify differences in the use of autologous or allogeneic HCT across indications and regions. No individual patient data were used and no ethics committee approval was mandated. Outcome information is not available from our center-specific registry and we opted against the report of fragmented outcome information of just two developed regions (EBMT and CIBMTR).

Data collection and validation Global transplant numbers by country of origin, year of transplant, disease and donor type (autologous vs. allogeneic) have been collected since 2006 (foundation year of the WBMT) in 194 WHO member states through the registries of the reporting member organizations, or national registries or transplant centers directly either in paper form or electronically using the standardized WBMT form. Detailed and validated information about main indication including stage of the disease, stem cell source, and allogeneic (family matched, family mismatched and unrelated) donor type were obtained for the years 2006 to 2016. Data were validated by a range of different independent systems; through confirmation by the reporting teams, following receipt of a computer printout of the entered data, by selective comparison with MEDA/TED datasets in the EBMT or CIBMTR data system or by crosschecking for double reporting with national registries. Data were validated by onsite visits to selected teams to verify reported data as part of the quality control program within the European, North American, Latin American and Asia-Pacific organizations. On-site visits to selected teams were part of the quality-control accreditation program of JACIE (www.ebmt.org/jacie-accreditation) or FACT (www.factweb.org). Based on quality controls and contacts with regulatory agencies or national offices, the response rate for allogeneic HCT was estimated to be >95% and for autologous HCT 80– 90%. The number of potential missing transplant numbers is estimated to be less than 5% for allogeneic HCT and less than 15% for autologous HCT. This number is much lower for Australia, Canada, Europe, Japan, and the USA. The survey focuses on the numbers of patients treated for the first time with HCT.

Participating hematopoietic cell transplantation (HCT) teams, groups, countries and continents In 2016, 1,662 HCT teams in 86 countries over six WHO continental regions delivered HCT services globally [(www.who.int/about/regions/en/). These regions included the Americas (AMR/PAHO; WHO regions North-, Middle and South America and Canada); Asia (SEAR/WPR; WHO regions South East Asia and Western Pacific Region, which includes Australia and New Zealand); Europe (EUR; which includes Turkey and Israel) and AFR/EMR (WHO regions Africa and Eastern Mediterranean). For specific analyses AMR/PAHO activities were divided in North America and Latin-America and AFR/EMR in Africa and EMR. A

haematologica | 2022; 107(5)

detailed list of organizations providing activity data and definitions used in the manuscript are reported in the Online Supplementary Appendix.

Statistical analysis The data analysis was comprised of ordinary least squares regressions for trends, c² tests for independent proportions of indications, and binomial tests for donor type. Calculations were done in Eviews8 and Excel 2010 (Microsoft).

Results Total hematopoietic cell transplantations and overall trends From 1957-2016, a total of 1,298,897 HCT (57.1% autologous) procedures were recorded. The cumulative numbers increased continuously from 10,000 in 1985, to 500,000 in 2005 and doubled to 1 million HCT by 2012. Projecting a frequency of at least 86,844 HCT for 2017 and 89,510 HCT for 2018 (incomplete and not validated data), a total of 1.5 million HCT worldwide was expected to be reached by 2019 (Online Supplementary Figure S1). The annual activity increased continuously from 46,563 in 2006 to 82,718 in 2016, amounting to a global increase of 77.6% since 2006, which was somewhat higher in allogeneic (89.0%) than in autologous HCT (68.9%, Table 1). The yearly increase was by a median of 5.9% for all HCT (allogeneic 6.8% and autologous 5.9%) to a total of 697,934 procedures (54% autologous) since 2006. The most frequent indications were lymphoproliferative disorders (LPD; n=370,884 HCT of which 88.4% were autologous) and leukemia (n=248,860 total of which 94.9% were allogeneic; see the Online Supplementary Figure S2). Global HCT team numbers plateaued in the last 4 years with a slight increase in 2016 (Online Supplementary Figure S3), while annual HCT numbers increased continuously. The increase was not a consequence of more reporting HCT teams, but of increased activity per HCT team. While the overall number of HCT per team was 35.1 in 2006, this reached 49.8 in 2016 (Online Supplementary Figure S3). Absolute numbers of all HCT per countries ranged from 0 to 19,505.

Hematopoietic cell transplantation teams activity in 2016 In 2016, the rates of HCT exceeded 80,000 HCT per year for the first time with 82,718 HCT (53.5% autologous) reported in the global HCT activity survey (Table 1). The majority of HCT were performed in Europe (45.2%) and in North America (24.4%), while SEAR/WPR contributed with 22.7%, Latin-America with 5.1% and AFR/EMR with 2.7%. The trends for TR were somewhat different, with rates highest in North America (561 TR), followed by Europe (439 TR), Latin America (77 TR), SEAR/WPR (54 TR), EMR (36 TR) and Africa (9 TR) (Figure 1). TR were higher for autologous than for allogeneic HCT in all regions except for SEAR/WPR, EMR and AFR. TR for allogeneic HCT ranged from 0.3 in Morocco to 414.0 in Israel (median 47.6); and for autologous HCT from 0.1 in Egypt to 705.9 in Iceland (median 99.1). The number of HCT teams varied considerably across regions, with the highest numbers being in SEAR/WPR and Europe (Figure 2). In contrast, TD ranged from 0.0529.4/country and was highest in Europe (7.7 TD) followed 1047

D. Niederwieser et al. Table 1. Global hematopoietic cell transplantation (HCT) activity in 2006, 2016 and changes according to disease indication, donor type and world region.

Family 2006 2016 Δ % (06-16) 7,491 14,992 100 3,587 7,531 110 1,679 3,965 136 826 624 -24 1,052 2,472 135 295 220 -25 52 180 246 LPD 1,863 2,023 9 Plasma cell disorders 490 312 -36 HD/NHL 1,373 1,673 21 Lymphoma other/unknown 0 38 Solid tumors 110 47 -57 Non-malignant disorders 1,458 3,423 134 Bone marrow failures 840 1,775 111 Hemoglobinopathies 338 962 185 Immune deficiencies 206 472 129 Inherited diseases 59 111 88 of metabolism Autoimmune disorders 6 19 217 Other non-malignant 9 84 833 disorders Others 80 109 36 EUR 4,906 7,074 44 North America 2,580 3,680 42.6 SEAR/WPR 1,948 7,392 279 Latin America 771 1,197 55.3 EMR/AFR 797 1,251 57 TOTAL 11,002 20,594 87.0 Leukemias

AML ALL CML MDS/MPS CLL Other leukemias

Unrelated Allogeneic Autologous 2006 2016 Δ % 2006 2016 Δ % 2006 2016 Δ % (06-16) (06-16) (06-16)

2006

Total 2016 Δ % 2016 (06-16) %

6,901 3,024 1,809 508 1,216 261 83 1,356 283 1,073 0 40 902 452 54 239 118

13,727 6,803 2,930 484 3,144 240 126 1,949 348 1,573 28 37 2,004 909 301 500 182

99 125 62 -5 159 -8 52 44 23 47

16,118 7,880 3,697 1,348 2,327 729 137 24,874 11,448 13,426 0 2,710 2,553 1,292 395 448 179

29,563 14,965 7,050 1,109 5,626 482 331 43,850 24,361 19,349 140 2,937 6,118 2,690 1,273 987 297

8.4 139.6 108.2 222.3 120.3 65.9

35.7 18.1 8.5 1.3 6.8 0.6 0.4 53.0 29.5 23.4 0.2 3.6 7.3 3.3 1.5 1.2 0.4

10 29

200 39

678 193

239.0 394.9

0.8 0.2

132 4,222 2,878 2,110 115 6 9,331

308 24,216 12,188 7,096 1,833 1,230 46,563

250 37,368 20,144 18,789 4,196 2,221 82,718

-18.8 54.3 65.3 164.8 128.9 80.6 77.6

0.3 45.2 24.4 22.7 5.1 2.7 100

28,719 14,334 6,895 1,108 5,616 460 306 3,972 660 3,246 66 84 5,427 2,684 1,263 972 293

99.5 116.8 97.7 -16.9 147.6 -17.3 126.7 23.4 -14.6 32.7

-7 122 101 457 109 54

14,392 6,611 3,488 1,334 2,268 556 135 3,219 773 2,446 0 150 2,360 1,292 392 445 177

-44,0 139.4 107.7 222.2 118.4 65.5

1,726 1,269 209 14 59 173 2 21,655 10,675 10,980 0 2,560 193 0 3 3 2

22 90

120 210

16 38

41 174

156.3 357.9

184 1

114 8,348 4,458 4,513 438 74 17,831

-14 98 54.9 114 280.9 1,133 91.0

212 9,128 5,458 4,058 886 803 20,333

223 15,422 8,138 11,905 1,635 1,325 38,425

5.2 69.0 49.1 193.4 84.5 65.0 89.0

96 15,088 6,730 3,038 947 427 26,230

844 631 155 1 10 22 25 39,878 23,701 16,103 74 2,853 691 6 10 15 4

-51.1 -50.3 -25.8 -92.9 -83.1 -87.3 1,150.0 84.2 122.0 46.7 11.4 258.0 233.3 400.0 100.0

637 246.2 19 1,800.0 27 21,946 12,006 6,884 2,561 896 44,293

-71.9 45.5 78.4 126.6 170.4 109.8 68.9

83.4 89.9 90.7 -17.7 141.8 -33.9 141.6 76.3 112.8 44.1

Δ (06/16) Difference from 2006 to 2016 in %; AML: acute myeloid leukemia; ALL: acute lymphoblastic leukemia; CML: chronic myeloid leukemia; MDS/MPS: myelodysplastic syndrome/myeloproliferative syndrome; CLL: chronic lymphocytic leukemia; LPD: lymphoproliferative disorders; EUR: Europe; SEAR/WPR: South East Asia Pacific Region/West Pacific Region; EMR/AFR: East Mediterranean Region/African Region.

by North America (6.0 TD), SEAR/WPR (1.9 TD), Latin America (1.6 TD), EMR (0.4 TD) and Africa (0.3 TD). Accordingly, the most HCT per team were performed in North America and EMR.

Trends in indications and hematopoietic cell transplantation type All regions showed increases in activity compared to 2006 (range, 54.3% in EUR - 164.8% in SEAR/WPR) for both transplant types (range, autologous 45.5% in EUR 126.6% in SEAR/WPR and allogeneic 49.1% North America - 193.4% SEAR/WPR; Table 1). Trends and increase in allogeneic and autologous HCT from 2006 to 2016 according to disease and regions are shown in Figure 3 and the Online Supplementary Figure S4A and B, respectively. The most common indications for autologous and allogeneic HCT (n=82,718) were lymphoproliferative diseases (53.0%), leukemias (35.7%), non-malignant disorders (7.3%) and solid tumors (3.6%; see Table 1).

Allogeneic hematopoietic cell transplantations Leukemia is the most common indication for allogeneic HCT (n=28,719) and the most frequent single indication 1048

acute myeloid leukaemia (AML) with a total of 14.334 HCT, followed by acute lymphoblastic leukemia (ALL; n=6895), myelodysplastic syndrome (MDS)/myeloproliferative neoplasms (MPS; n=5,616) and chronic myeloid leukemia (CML; n=1,108). Non-malignant disorders (n=5,427) with the subgroups bone marrow (BM) failure (49.5%), hemoglobinopathies (23.3%) and immunodeficiencies (17.9%) were identified as the second most common indication for allogeneic HCT. The third most common indication was lymphoproliferative disorder (LPD; n=3,972) of which 81.7% were Hodgkin disease (HD)/non-Hodgkin disease (NHL) and 16.6% plasma cell disorders (PCD). Increases from 2006 were noted especially in non-malignant disorders (139.4%), leukemias (99.5%) with the subgroups MDS/MPS (147.6%), AML (116.8%) and LPD (23.4%). Declines were noted in CML (-16.9%) and chronic lymphocytic leukemia (CLL) (-17.3%; Table 1). HCT were performed predominantly in AML in first complete remission (CR1) (Online Supplementary Figure S4A) with a decline in procedures in non-CR1. In ALL CR1 was also more frequent than non-CR1. HCT for MDS and, to a lower extent MPN, increased steadily over the observahaematologica | 2022; 107(5)

HCT worldwide access

Figure 1. Transplant rates. Hematopoietic cell transplantation (HCT)/10 million population according to transplant type (autologous, allogeneic, related mismatched and cord blood) and world regions in 2016. EUR: Europe, EMR: East Mediterranean Region; AFR: Africa; SEAR/WPR: South East Asia Pacific Region/West Pacific Region.

tion period. Increased allogeneic HCT activities were observed for almost all indications in all regions. The highest increase was noted for related donors in the SEAR/WPR region (279%; Table 1; Figure 3). The only observed decrease involved solid tumors (44% decrease) in almost all regions except for SEAR/WPR and LPD in North America.

Autologous hematopoietic cell transplantation The most frequent indication for autologous HCT in 2016 was LPD (n=39,878, 84.2% of all autologous HCT; Table 1) and the most frequent single indication PCD and HD/NHL with 59.4% and 40.4%, respectively. HCT for solid tumors (n=2,853) accounted for 6.4% of autologous HCT and 844 (1.9%) were performed for leukemias mostly for AML (74.8%) and ALL (18.4%). Autoimmune diseases (AID) (92.2%) were the predominant indication for autologous HCT within non-malignant diseases (n=691). Frequencies of autologous HCT increased in almost all indications and regions especially for LPD and non-malignant disorders. Decreases in autologous HCT were observed for leukemia (except SEAR/WPR) and for solid tumors in Europe. The highest increase in autologous HCT was observed in Latin America for non-malignant disorders. The frequency of autologous HCT increased by 68.9% predominantly in non-malignant disorders (258%), primarily AID (246.2%) in PCD (122%) and in lymphoma (46.7%; Online Supplementary Figure 4B; Table 1). Decreased activity of autologous HCT was reported for all leukemias (-51.1%; Figure 3).

Trends in donor type and stem cell source Autologous HCT (range, 0-11,655) were reported from 85 participating countries, while allogeneic HCT (range, 07,850) were reported from 76 countries, including procedures from unrelated donors and from CB in 55 and in 41 countries, respectively. Overall, related HCT has become haematologica | 2022; 107(5)

more frequent than unrelated HCT starting in 2014 (Online Supplementary Figure S5). The increase in related HCT was mainly due to the use of non-identical related donors (39.5% of related HCT), which increased significantly over the last 4 years, while unrelated CB showed a moderate decline. A detailed analysis of identical and nonidentical related HCT according to indication and in comparison to unrelated HCT is given in Table 2. Highest increases were observed for severe aplastic anemia and hemoglobinopathies (Delta 2007-2016 >1,000%) and for leukemia (Delta 2007-2016 >550%; especially AML CR1, ALL CR1, CML first CP with maximum 2,456%, data not shown), but also in LPD (especially in HD and NHL max Delta 2007-2016 =711%). In comparison, HCT from related identical donors showed Delta 2007-2016 of maximum 189% for ALL CR1 and from unrelated for AML CR1 of 430%. Amongst allogeneic HCT, the proportion of unrelated donor HCT ranged from 4.4% to 78.3% (median 20%), with 30 countries performing more unrelated than related donor HCT. Absolute unrelated HCT numbers ranged from 0 to 4,311 and those of CB ranged from 0 to 1,233 in individual countries. It is not surprising that more related HCT were performed in regions without an unrelated donor registry (Latin America, EMR, AFR and SEAR/WPR). Related haploidentical HCT (n=8,131) were evenly distributed in 62 countries; with absolute numbers ranging from 0 to 2,554 and proportions of all allogeneic HCT ranging between 1.5% to 77% (median 8%). TR for haploidentical HCT was highest in North America (n=38), 25 in Europe, eight in Latin America and ≤4 in Asia Pacific, AFR and EMR (Figure 1). However, TR as % of allogeneic HCT reached 34% in SEAR/WPR and 26% in Latin America, while it was only 14% in Europe and EMR. Peripheral blood (PB) was the predominant graft source in both autologous (99.7%) and allogeneic (72.8%) HCT, while CB as a source has declined (in 2016 13.9% of all 1049

D. Niederwieser et al.

Table 2. Trend of global hematopoietic cell transplantation activity according to disease indication and donor type.

Related Related Related total Unrelated identical non-identical 2007 2016 Δ%(07-16) 2007 2016 Δ%(07-16) 2007 2016 Δ%(07-16) 2007 2016 Δ%(07-16) Leukemias AML ALL CML MDS/MD/MPN CLL Other leukemias LPD Plasma cell disorders Multiple myelomas Others Lymphoma Hodgkin disease Non-Hodgkin lymphoma Solid tumors Non-malignant disorders Bone marrow failure Severe aplastic anemia Other Hemoglobinopathies Immune deficiencies Inherited diseases of metabolism Autoimmune disorders Other non malignant disorders Other TOTAL

6,858 3,249 1,635 601 1,044 279 50 1,677 542 341 9 1,135 186 796 47 1,354 789 644 145 338 140 44 12 31 61 9,997

8,906 4,453 2,256 388 1,559 166 84 1,238 234 211 23 979 205 774 4 2,185 1,136 964 172 711 234 54 10 40 66 12,399

29.9 37.1 38.0 -35.4 49.3 -40.5 68.0 -26.2 -56.8 -38.1 155.6 -13.7 10.2 -2.8 -91.5 61.4 44.0 49.7 18.6 110.4 67.1 22.7 -16.7 29.0 8.2 24.0

1,101 472 376 56 156 17 24 163 23 8 1 140 29 63 32 196 65 48 17 21 84 19 1 6 12 1,504

unrelated and 6.7% of all allogeneic HCT). Of the 38,425 allogeneic HCT, 7,868 (20%) were bone marrow (BM)derived, 27,963 (73%) were from PB and 2,594 (7%) from cord blood (CB). Of the 44,293 autologous HCT, 99.7% were from PB, 0.3% from BM and only three single HCT were from CB.

Discussion The analysis of global HCT activity based on data from 2006-2016 on 700.000 HCT over the last 10 years gathered by the WBMT gives important insights into the global trends in the field of HCT. As per previous reports,8–13 HCT activities continue to increase worldwide without plateau and exceeded 80,000 procedures annually for the first time in 2016 and are expected to exceed 90,000 in 2018 to reach a total of 1.5 million by 2019. Although allogeneic HCT has increased more than autologous in recent years, the later remains the predominant transplant type and LPD the leading disease indication. Leukemia is the second most frequent HCT indication and 94.9% of these use allogeneic donors. While HCT on patients with acute leukemias (predominantly in CR1, but not in non-CR1) and MDS/MPN increase, a decrease of frequencies in chronic leukemias (CLL and CML) was observed. Non-malignant disorders now account for 7.3% 1050

6,056 3,064 1,702 236 905 53 96 758 64 54 10 681 170 511 42 1,232 633 553 80 251 238 57 9 44 43 8,131

450.1 549.2 352.7 321.4 480.1 211.8 300.0 365.0 178.3 575.0 900.0 386.4 486.2 711.1 31.3 528.57 873.9 1,052.1 370.6 1,095.2 183.3 200.0 800.0 633.3 258.3 440.6

8,018 3,749 2,023 663 1,211 298 74 1,877 583 356 11 1,294 222 871 84 1,559 862 700 162 359 224 63 13 38 73 11,611

14,992 7,531 3,965 624 2,472 220 180 2,023 312 279 33 1,673 377 1,296 47 3,423 1,775 1,523 252 962 472 111 19 84 109 20,594

87.0 100.9 96.0 -5.9 104.1 -26.2 143.2 7.8 -46.5 -21.6 200.0 29.3 69.8 48.8 -44.1 119.56 105.9 117.6 55.6 168.0 110.7 76.2 46.2 121.1 49.3 77.4

7,868 3,610 2,015 479 1,358 316 90 1,338 245 192 7 1,093 175 683 35 984 448 320 128 66 255 98 8 109 68 10,293

13,727 6,803 2,930 484 3,144 240 126 1,949 348 316 32 1,573 181 1,392 37 2,004 909 743 166 301 500 182 22 90 114 17,831

74.5 88.5 45.4 1.0 131.5 -24.1 40.0 45.7 42.0 64.6 357.1 43.9 3.4 103.8 5.7 103.66 102.9 132.2 29.7 356.1 96.1 85.7 175.0 -17.4 67.7 73.2

of all HCT (89.2% of which are allogeneic) with BM failures as the most frequent group of disease indications (47.5%, mostly severe aplastic anemia). Hemoglobinopathies were the second most frequent indication (22.4%) with an increase of 222.3%. Finally, solid tumors were the indication in 3.6% of all HCT and were almost exclusively transplanted with autologous grafts (97%). Striking is the pronounced increase in haploidentical HCT in hemoglobinopathies, severe aplastic anemia, leukemia and LPD (especially MM and NHL) in developing countries. Findings are partly in agreement on a larger scale with the development in Europe and US, but also diverge especially on the use of haploidentical HCT for non-malignant diseases, in the use of autologous HCT in autoimmune disease (LABMT) and in leukemias (SEAR/WPR) and in the use of allogeneic HCT in autoimmune diseases in SEAR/WPR. There was substantial divergence in the rate of growth of HCT between different regions, with increases of more than 129% in SEAR/WPR and Latin-America over the last decade contributing strongly to the continuous global upward trend in activity. Both regions reported their major increases in non-malignant disorders, in the SEAR/WPR with allogeneic and in Latin America with autologous HCT. AFR/EMR had a remarkable growth of 80% in comparison to 65% in North America and 54% in Europe. Despite the differential increase, TR still vary by more than 10-fold from region to region. A deeper analysis of the number of HCT teams in haematologica | 2022; 107(5)

HCT worldwide access

Figure 2. Number of hematopoietic cell transplant (HCT) teams, HCT team density and HCT per team according to world regions. EUR: Europe; EMR: East Mediterranean Region; AFR: Africa; SEAR/WPR: South East Asia Region/Western Pacific Region; TD: team density.

Figure 3. Change in activity in percent from 2006 and 2016 according to disease indication and world region. HCT: hematopoietic cell transplantation; EUR: Europe; LABM: Latin American Blood and Marrow Transplantation Group; TAFR/EMR: Africa/East Mediterranean Region;; SEAR/WPR: South East Asia Region/Western Pacific Region.

haematologica | 2022; 107(5)

1051

D. Niederwieser et al.

the different regions showed three patterns: high (>500 teams; SEAR/WPR and EUR), intermediate (100-500 teams; North America) and low (<100 teams; AFR, EMR, Latin America). TD was similar in North America and Europe (6.0 and 7.7 respectively), between 1-2 in SEAR/WPR and Latin America and <1 in all other regions. While North America and EMR had most HCT per HCT team (93.7 and 86.6, respectively), Europe and Latin America had intermediate (57.0 and 48.2, respectively), SEAR/WPR and AFR had the lowest (27.9 and 32.7, respectively) HCT per team. It is encouraging that HCT and TR in developing regions are increasing steadily, although there are differences in indications and diseases, and are closing the HCT access gap. However, the disparities in access between regions remains substantial. Although the WBMT has established a global coverage of international transplant societies with an estimated >90% reporting, at least one very large country had incomplete reporting and lacked a national registry. With the aim of establishing such a registry, the WBMT organized a successful workshop to highlight the importance of coordinated reporting. Increased numbers of teams and activities in that country are now expected starting in 2016 (Xu L-P, Lu P_H, Wu D-P et al. Hematopoietic stem cell transplantation activity in China 2019: a report from the Chinese Blood and Marrow Transplantation Registry Group. Bone marrow transplantation; 2021 [in press]). Therefore, the data presented here are certainly an underestimate but are expected to increase further with greater completeness of reporting.14 The considerable increase seen in the SEAR/WPR and Latin American region in non-malignant disorders, hemoglobinopathies (allogeneic HCT) and AID (autologous HCT, predominantly in multiple sclerosis), is of special interest. It is reassuring to see a worldwide trend of more patients to be transplanted in CR1 rather than later in their disease course. This phenomenon is much more evident in AML and remains constant in ALL. Lymphoma accounts for less of the increase in autologous HCT than do PCD. The indications for CML decreased only slightly in the 10 year period following the availability of successive tyrosine kinase inhibitor pharmaceuticals. However, it should be noted here that more patients were transplanted beyond chronic phase CML, and not in line with ELN recommendations.15 The changes in donor type and graft source confirm trends previously described. Frequencies of haploidentical HCT increased considerably to reach >10% of all allogeneic HCT throughout all regions. As a consequence, more related (matched and haploidentical) HCT were performed than were unrelated HCT, and the later appears to have reached a plateau. The increase of haploidentical related donors might have several reasons including lack of donors for Latin-American patients in international registries, lack of local national donor registries and economic. The availability of new technologies like the postcyclophosphamid protocol and the so called ‘Beijing Protocol’ using G-CSF/anti-thymocyte globulin and multiple stem cell sources (BM and PB stem cells)14,16 may have contributed to this development. Overall, CB HCT decreased consistent with the pattern seen in Europe. There are a few limitations in this analysis. The first one relates to the reporting lag of the analysis. Logistics in obtaining data from 1,660 transplant centers worldwide are the main cause for this delay. While some of the regions are reporting in real time, emerging regions have not a central national or regional reporting system. In these regions, WBMT collected the information from individual transplant 1052

centers returning the results to the regional societies for completeness checking. An internet-based reporting system was developed by the WBMT and is expected to gradually bring up the survey reports to real time. An additional limitation of this analysis is the inability to provide information on utilization of HCT for specific illnesses and on information restricted to first HCT. Utilization is currently being analyzed worldwide for myeloma and AML by the WBMT, but should be extended to major indications. Furthermore, information on second or third HCT is not currently available, although these account for approximately 10% of HCT in developed countries, but will be implemented with the new reporting system. One of the challenges that the WBMT faces, is how to go beyond calculations of global HCT activity and accelerate global equity of access to HCT. The most efficient method may be to increase HCT activity/team. In fact, transplant teams have already increased their annual HCT/team by a median of 14 HCT/team since 2006. As shown in this study, 50 HCT/team are feasible even in developing countries and almost 100 HCT/team are currently being performed in North America. Increasing team numbers might be more difficult. This is due in part to the funds allocated to local health expenditure, but also due to limitations in the infrastructure required including blood banks, intensive care units, multidisciplinary teams and microbiology expertise. Shortages and unavailability of medicines and lack of trained biomechanical/biotechnical technicians have also hindered HCT activities in developing countries. The WBMT has the capacity to review and analyze global HCT activity data and apply this data to support HCT activities globally. Global activity survey data has been instrumental in informing and shaping HCT support workshops in the different regions, which have been successful in fostering collaboration amongst international societies and supportive expert HCT networks globally. The WBMT, on its site, has already prepared a variety of documents to facilitate the establishment of new transplant programs (requirements for establishing a program, list of essential medicines, use of biosimilars to reduce costs and establishing unrelated donor registries).17–21 In addition, supervisory telemedicine is an evolving and potentially powerful tool to overcome lack of experience with collateral benefits for conventional hematology, blood banking, microbiology and virology. Devoted physicians and willing health authorities are essential for the application of such technologies and successful collaborations accompanied by demonstration of compliance with international standards to provide reassurance to internal and external stakeholders.22 The achievements obtained in the last decade should be an incentive to continue and even increase the common efforts to improve access and close the gap worldwide faster. This is a common effort of professional organizations, WHO, politicians and Health Authorities. The role of the WHO is essential in coordinating this process among their member states.

Disclosures No conflicts of interest to disclose. Contributions DN, HB, MA, KB, KF, and FRA designed the study; HB, NB, CB, NC, SC, AE,CF, SG, NH, AAH, SH, AH, MH, MI, GJ, AK, JK, NK, RPL, JWL, JMR, MP, JP, KP, AS, JAS, AS, JS, DW, NW, MK, MA, HG, YA, WS contributed data haematologica | 2022; 107(5)

HCT worldwide access

and assured quality of the data given to the analysis; HB, DN, MP, and LF analysed data; DN, HB, WS, and NH drafted the manuscript; DN, HB, NB, CB, NC, SC, AE, CF, SG, NH, AAH, SH, AH, MH, MI, GJ, AK, JK, NK, RPL, JWL, JMR, MP, JP, KP, AS, JAS, AS, JS, DW, NW, MK, MA, HG, YA processed the manuscript. European data were derived from the European Society for Blood and Marrow Transplantation (EBMT) database for the years 1965–89 and from the EBMT annual activity survey office since 1990. Noneuropean data were initially provided by the Center for International Blood and Marrow Transplant Research (CIBMTR) since 1964. They were supplemented or replaced by the activity surveys of the Asian Pacific Blood and Marrow Transplantation Group (APBMT) since 1974, the Australasian Bone Marrow Transplant Recipient Registry (ABMTRR) since 1992, the Eastern Mediterranean Blood and Marrow Transplantation Group (EMBMT) since 1984, the Cell Therapy Transplant Canada (CTTC) since 2002, the Latin American Bone Marrow Transplantation Group (LABMT) since 2009, and the African Blood and Marrow Transplant Group (AFBMT) since 2010. Unrelated donor and cord blood information were derived from the World Marrow Donor Association (WMDA) and Bone Marrow Donors Worldwide (BMDW).

References 1. Thomas ED, Lochte HL, Lu WC, Ferrebee JW. Intravenous infusion of bone marrow in patients receiving radiation and chemotherapy. New Engl J Med. 1957 (257):491-496. 2. Gratwohl A, Hermans J, Goldman J, et al. Risk assessment for patients with chronic myeloid leukaemia before allogeneic blood or marrow transplantation. Lancet. 1998; 352(9134):1087-1092. 3. McSweeney PA, Niederwieser D, Shizuru JA, et al. Hematopoietic cell transplantation in older patients with hematologic malignancies: replacing high-dose cytotoxic therapy with graft-versus-tumor effects. Blood. 2001;97(11):3390-3400. 4. Bacigalupo A. Second EBMT Workshop on reduced intensity allogeneic hemopoietic stem cell transplants (RI-HSCT). Bone Marrow Transplant. 2002;29(3):191-195. 5. Giralt S, Ballen K, Rizzo D, et al. Reducedintensity conditioning regimen workshop: defining the dose spectrum. Report of a workshop convened by the center for international blood and marrow transplant research. Biol Blood Marrow Transplant. 2009;15(3):367-369. 6. Attal G, Harousseau JL, Stoppa AM, et al. A prospective, randomized trial of autologous bone marrow transplantation and chemotherapy in multiple myeloma. N Engl J Med. 1996;335(2):91-97. 7. Snowden JA, Badoglio M, Labopin M, et al. Evolution, trends, outcomes, and economics of hematopoietic stem cell transplantation in severe autoimmune diseases 4. Blood Adv. 2017;1(27):2742-2755. 8. Gratwohl A, Baldomero H, Aljurf M, et al. Hematopoietic stem cell transplantation: a global perspective. JAMA. 2010; 303(16): 1617-1624. 9. Gratwohl A, Baldomero H, Gratwohl M, et

haematologica | 2022; 107(5)

Acknowledgments The cooperation of all participating teams, countries and organizations with their staff is greatly appreciated, specifically the following: ABMTRR: St. Vincent´s Hospital Sydney, APBMT: Aichi Medical University, CTTC: Centre Tecnològic de Telecomunicacions de Catalunya; CIBMTR: Medical College of Wisconsin, EBMT: Coordination Offices in Barcelona, Paris and London and the Austrian Registry (ASCTR), the Czech BMT Registry, the French Registry (SFGM), the German Registry (DRST), the Italian Registry (GITMO), the Dutch Registry (HOVON), the Spanish BMT Registry (GETH), the Swiss Registry (SBST), the Turkish BMT Registry and the British Registry (BSBMT), EMBMT, SBTMO, LABMT, AFBMT, WMDA, and Eurocord. The authors thank Michael Cross for editorial work on the manuscript. Funding Funding for this study was indirectly provided by support of the WBMT. Funding was solely to support the study; no individual payment was made to any of the persons involved in the study. The Activity Survey Office is partly supported by the University Hospital of Basel. MG was supported by SCCER CREST. The Online Supplementary Appendix shows details about funding of the participating institutions.

al. Quantitative and qualitative differences in use and trends of hematopoietic stem cell transplantation: a Global Observational Study. Haematologica. 2013;98(8):1282-290. 10. Gratwohl A, Pasquini MC, Aljurf M, et al. One million haemopoietic stem-cell transplants: a retrospective observational study. Lancet Haematol. 2015;2(3):e91-100. 11. Niederwieser D, Baldomero H, Szer J, Gratwohl M, Aljurf M, Atsuta Y, et al. Hematopoietic stem cell transplantation activity worldwide in 2012 and a SWOT analysis of the Worldwide Network for Blood and Marrow Transplantation Group including the global survey. Bone Marrow Transplant. 2016;51(6):778-785. 12. Baldomero H, Aljurf M, Zaidi SZA, et al. Narrowing the gap for hematopoietic stem cell transplantation in the EastMediterranean/African region: comparison with global HSCT indications and trends. Bone Marrow Transplant. 2019;54(3):402417. 13. Iida M, Dodds A, Akter MR, et al. The 2016 APBMT Activity Survey Report: trends in haploidentical and cord blood transplantation in the Asia-Pacific region. Blood Cell Therapy. 2021(4):20-28. 14. Niederwieser D. The Chinese HCT survey: a non-manipulated haploidentical transplantation procedure makes a novel contribution to data sharing within the regional and global transplant registries and to worldwide knowledge. Bone Marrow Transplant. 2021;56(6):1229-1231. 15. Hochhaus A, Baccarani M, Silver RT, et al. European LeukemiaNet 2020 recommendations for treating chronic myeloid leukemia. Leukemia. 2020;34(4):966-984. 16. Xu L-P, Wu D-P, Han M-Z, et al. A review of hematopoietic cell transplantation in China: data and trends during 2008-2016 [eng]. Bone Marrow Transplant. 2017; 52(11): 1512-1518.

17. Fakih RE, Greinix H, Koh M, et al. Worldwide Network for Blood and Marrow Transplantation (WBMT) recommendations regarding essential medications required to establish an early stage hematopoietic cell transplantation program. Transplant Cell Ther. 2021;27(3):267.e1-267.e5. 18. Muhsen IN, Hashmi SK, Niederwieser D, et al. Worldwide Network for Blood and Marrow Transplantation (WBMT) perspective: the role of biosimilars in hematopoietic cell transplant: current opportunities and challenges in low- and lower-middle income countries. Bone Marrow Transplant. 2020;55(4):698-707. 19. Pasquini MC, Srivastava A, Ahmed SO, et al. Worldwide Network for Blood and Marrow Transplantation (WBMT) recommendations for establishing a hematopoietic cell transplantation program (part I): minimum requirements and beyond. Hematol Oncol Stem Cell Ther. 2020;13(3):131-142. 20. Aljurf M, Weisdorf D, Hashmi SK, et al. Worldwide Network for Blood and Marrow Transplantation (WBMT) recommendations for establishing a hematopoietic stem cell transplantation program in countries with limited resources (part II): clinical, technical and socio-economic considerations. Hematol Oncol Stem Cell Ther. 2020;13(1):7-16. 21. Aljurf M, Weisdorf D, Alfraih F, et al. "Worldwide Network for Blood & Marrow Transplantation (WBMT) special article, challenges facing emerging alternate donor registries". Bone Marrow Transplantat. 2019;54(8):1179-1188. 22. Frutos C, Enciso ME, Glasenapp A von, Quiroz A, Batista J, Niederwieser D. Bridging the gap using telemedicine: optimizing an existing autologous hematopoietic SCT unit into an allogeneic hematopoietic SCT unit in Paraguay with the help of the WBMT. Blood Adv. 2019;3(Suppl 1):S45–7.

1053

ARTICLE Ferrata Storti Foundation

Haematologica 2022 Volume 107(5):1054-1063

Cell Therapy & Immunotherapy

Improved outcome of patients with graft-versus-host disease after allogeneic hematopoietic cell transplantation for hematologic malignancies over time: an EBMT mega-file study Hildegard T. Greinix,1 Dirk-Jan Eikema,2 Linda Koster,3 Olaf Penack,4 Ibrahim Yakoub-Agha,5 Silvia Montoto,6 Christian Chabannon,7 Jan Styczynski,8 Arnon Nagler,9 Marie Robin,10 Stephen Robinson,11 Yves Chalandon,12 Malgorzata Mikulska,13 Stefan Schönland,14 Zinaida Peric,15 Annalisa Ruggeri,16 Francesco Lanza,17 Liesbeth C. de Wreede,18 Mohamad Mohty,19 Grzegorz W. Basak20 and Nicolaus Kröger21 1 Department of Medicine, Division of Haematology, Medical University of Graz, Graz, Austria; 2EBMT Statistical Unit, Leiden, the Netherlands; 3EBMT Data Office Leiden, Leiden, the Netherlands; 4Medical Department of Hematology, Oncology and Tumor Immunology, Charité Universitaetsmedizin Berlin, Berlin, Germany; 5Department of Haematology, CHU de Lille, University Lille, INSERM U1286, Infinite, Lille, France; 6 St. Bartholomew`s hospital, Barts Health NHS Trust, London, UK; 7Institut Paoli Calmettes, Marseille, France; 8Collegium Medicum, Nicolaus Copernicus University Torun, Bydgoszcz, Poland; 9Hematology and BMT Division, Chaim Sheba Medical Center, Tel-Hashomer, Tel-Aviv University, Ramat-Gan, Israel; 10Hematology / Transplantation, Hôpital Saint-Louis, Paris Cedex 10, France; 11Bristol Oncology Centre, Bristol, UK; 12 Département d’Oncologie, Service d’Hématologie, Hôpitaux Universitaires de Genève and Faculty of Medicine, University of Geneva, Geneva, Switzerland; 13Division of Infectious Diseases, University of Genova, Ospedale Policlinico San Martino, Genova, Italy; 14University Hospital of Heidelberg, Heidelberg, Germany; 15University Hospital Center Rebro, Zagreb, Croatia; 16Hematology and Bone Marrow Transplant Unit, IRCCS San Raffaele Scientific Institute, Milan, Italy; 17Ravenna Hospital, Ravenna, Italy; 18 Department of Biomedical Data Sciences, LUMC Leiden, Leiden, the Netherlands; 19 Hopital Saint Antoine, Paris Cedex 12, Paris, France; 20Department of Hematology, Oncology and Internal Medicine, Medical University of Warsaw, Warsaw, Poland and 21 University Hospital Eppendorf, Hamburg, Germany

ABSTRACT

Correspondence: HILDEGARD GREINIX hildegard.greinix@medunigraz.at Received: July 3, 2020. Accepted: June 11, 2021. Pre-published: June 24, 2021. https://doi.org/10.3324/haematol.2020.265769

1054

cute graft-versus-host disease (aGvHD) remains a major threat to successful outcome following allogeneic hematopoietic cell transplantation though advances in prophylaxis and supportive care have been made. The aim of this study is to test whether the incidence and mortality of aGvHD have decreased over time. 102,557 patients with a median age of 47.6 years and with malignancies after first allogeneic sibling or unrelated donor (URD) transplant were studied in the following periods: 1990-1995, 1996-2000, 2001-2005, 2006-2010 and 2011-2015. Findings: 100-day incidences of aGvHD grades II-IV decreased from 40% to 38%, 32%, 29% and 28%, respectively, over calendar time (P<0.001). In multivariate analysis URD, not in complete remission (CR) at transplant or untreated, and female donor for male recipient were factors associated with increased risk whereas the use of ATG/alemtuzumab decreased aGvHD incidence. Median follow-up was 214, 169, 127, 81 and 30 months, respectively, for the periods analyzed. Three-year-survival after aGvHD grades II-IV increased significantly from 38% to 40%, 43%, 44%, and 45%, respectively. In multivariate analysis URD, not in CR at transplant, peripheral blood as stem cell source, female donor for male recipient, and the use of ATG/alemtuzumab were associated with increased mortality whereas reduced-intensity conditioning was linked to lower mortality. Mortality increased with increasing patient age but decreased in the recent cohorts. Our analysis demonstrates that aGvHD has decreased over recent decades and also that the survival rates of patients affected with aGvHD has improved.

haematologica | 2022; 107(5)

Improved survival after acute GvHD

Introduction Allogeneic hematopoietic stem cell transplantation (HSCT) has been increasingly used to cure malignant and benign hematologic diseases.1 Transplanted T-cells from the donor can recognize and eradicate hematologic malignancies through the immunologic graft-versus-leukemia effect. Unfortunately, donor conventional T-cells recognize normal recipient tissues and attack them, causing graft-versus-host disease (GvHD); this is the major cause of non-relapse mortality (NRM) following HSCT: GvHD is commonly reported in 40-60% of patients following HSCT.2,3 Despite advances in GvHD prophylaxis, current pharmacological approaches fail to completely prevent acute GvHD (aGvHD).4-6 Corticosteroids remain the established first-line therapy, however only around 50% of patients with aGvHD achieve complete responses7 and prognosis of steroid-refractory patients is dismal.8 Therefore, aGvHD and its associated infectious complications and organ toxicities contribute substantially to morbidity and early mortality following HSCT. Over recent decades, transplant practices have changed markedly with recipient age increasing, the use of unrelated and mismatched donors, reduced-intensity conditioning (RIC) and peripheral blood stem cells (PBSCs) as the predominant graft source.9,10 Furthermore, improvements in supportive care measures such as novel antimicrobial agents and diagnostic procedures have also had an impact on HSCT outcome over time. Retrospective analyses have revealed an improvement in the survival of more recent transplant recipients.9,11 Gooley et al., observed reduced incidences of aGvHD grades III-IV, less major organ injury and a reduction in life-threatening infections in the early stages following HSCT in recentlytransplanted HSCT patients.11 Khoury et al. reported improved survival over time for transplant recipients after myeloablative conditioning with aGvHD given tacrolimusbased GvHD prophylaxis. This was most evident for patients with grade II aGvHD.12 To date, analyses on changes in survival outcome over time in an aGvHD-affected patient cohort given various conditioning regimen intensities are lacking. Therefore, we conducted a large registry study to assess whether outcome of HSCT patients experiencing severe aGvHD has improved over time.

Methods Data collection This was a multicenter, retrospective study performed by the Transplant Complications and Chronic Malignancies Working Parties of the European Society for Blood and Marrow Transplantation (EBMT). Data on transplantations were obtained from the EBMT registry. A total of 590 transplant centers contributed patients. The study was approved by the scientific board of both EBMT working parties.

Study endpoints The primary endpoints were aGvHD grades II-IV following HSCT and overall survival (OS) after aGvHD grades II-IV, with events defined as death from any cause after experiencing aGvHD grades II-IV. Secondary endpoints were incidence of aGvHD grades III-IV following HSCT; OS after aGvHD grades III-IV; NRM after aGvHD grades II-IV, defined as death occurring before signs of progression or relapse; relapse and progression after aGvHD IIIV, defined as recurrence and continuation of the original disease, respectively, following aGvHD II-IV and disease-free survival (DFS) after aGvHD II-IV, and defined as survival after aGvHD IIIV in the absence of signs of progression or relapse. For outcomes following aGvHD, only the subset of patients experiencing aGvHD were analyzed. In this subset, the starting point was the date of aGvHD onset. A clock-back approach was used, i.e., the timescale starts from the date of aGvHD onset rather than the date of HSCT.

Statistical analysis The primary comparison concerned outcomes in transplantation periods 1990-1995, 1996-2000, 2001-2005, 2006-2010, and 2011-2015. Cumulative incidence estimates were calculated for aGvHD, relapse and NRM, in a competing risks framework. NRM was a competing risk in the estimation of malignancy relapse, and relapse was a competing risk for estimation of NRM. For aGvHD, only mortality was considered a competing event. Baseline characteristics were compared across these transplant periods and tested by means of c2 tests for categorical variables and Kruskal-Wallis tests for continuous variables. Univariate analyses compared the outcomes OS, DFS, NRM, and relapse incidence between the five time cohorts. The probabilities of OS and DFS for all patients were calculated using the Kaplan Meier estimator. Group differences were tested by means of log rank test and Gray’s test. The median follow-up was estimated by reverse Kaplan-Meier estimator. Multivariable analyses were performed using Cox proportional hazards regression models for OS and RFS. Cause-specific hazards models were applied in the analysis of aGvHD, relapse and NRM. Each of the outcomes was analyzed using the same covariate structure. Covariates included were age at transplant (continuous in decades), transplantation year (in decades), conditioning intensity (reduced vs myeloablative), donor type (unrelated vs HLA identical), graft source (bone marrow (BM) vs PBSC), patientdonor gender match (male/female, female/male, and female/female vs male/male), antithymocyte globulin (ATG) and alemtuzumab. Models were stratified by malignant disease (Acute Lymphocytic Leukemia (ALL), Acute Myeloid Leukemia (AML), Myelodysplastic Syndrome (MDS), Multiple Myeloma (MM), Myeloproliferative Disorder (MPE) and Other. Hazard ratios including 95% confidence intervals and P-values are provided. Analyses were done in R, version 3.6.0, using survival, prodlim and cmprsk packages.

Results Patients

Patient selection Patients were 18 to 80 years of age with hematological malignant disease and received a first allograft between 1990 and 2015 from a HLA-identical sibling or matched or mismatched unrelated donor. Both myeloablative and RIC regimens, and any disease stage at HSCT, were included. Recipients of in-vitro T-cell depleted grafts, haplo-identical transplants and cord-blood transplants were excluded. The maximal severity of aGvHD for each patient was used in all analyses.13

haematologica | 2022; 107(5)

Between 1990 and 2015, 102,557 patients with a median age (range) of 47.6 (18-83.8) years with malignant disease received a first allogeneic HSCT. Conditioning was myeloablative in 55.8% and RIC in 44.2% of patients. Of note is that 50.6% of patients had an HLA-identical sibling and 49.4% a matched or mismatched unrelated donor. The stem cell source in 79.2% of patients was PBSC and BM in 20.8% of patients. Cyclosporine-based GvHD prophylaxis was given in 77.2% of patients; of these patients, 1055

H.T. Greinix et al.

57.9% also were given methotrexate (MTX) while 25.3% received mycophenolate mofetil (MMF). Donor, recipient and transplant characteristics are summarized in Table 1. The median follow-up for all patients was 214 (207.4220.3) months, 169.3 (166.5.7-172.1) months, 127.7 (125.9-129.4) months, 81 (80.3-81.8) months and 29.7 (29.2-30.2) months for 1990-1995, 1996-2000, 2001-2005, 2006-2010, and 2011-2015, respectively.

Characteristics of aGvHD Incidences (95% CI) of aGvHD grades II-IV by 100 days significantly (p<0.001) decreased from 40% (3842%), to 38% (37-39%), 32% (31-33%), 29% (2829%), and 28% (27-28%) for the periods 1990-1995, 1996-2000, 2001-2005, 2006-2010, and 2011-2015, respectively (Figure 1A). Of the patients who developed aGvHD grades II-IV, the median days to aGvHD

(interquartile range, IQR) was 20 (14-31), 22 (15-35), 25 (16-39), 25 (16-40), and 25 (16-40) in the periods 19901995, 1996-2000, 2001-2005, 2006-2010, and 20112015, respectively. Involvement of the gastrointestinal (GI) tract was observed in 50.3%, 54.3%, 55.1%, 59% and 66.3% of all patients experiencing aGvHD, respectively. First-line therapy consisted of corticosteroids in 55.5%, 87.1%, 78.7%, 79.1%, and 72.3% of patients, respectively. In multivariate analysis (Table 2) URD (HR 1.61 (1.541.67); P<0.001), not in CR at HSCT (HR 1.25 (1.2-1.3); P<0.001) or untreated (HR 1.11 (1.02-1.2); P=0.02), BM stem cell source (HR 1.2 (1.15-1.25); P<0.001), and a female donor for a male recipient (HR 1.16; 1.11-1.21; P<0.001) were all associated with increased risk for aGvHD grades II-IV whereas the use of ATG or alemtuzumab (HR 0.79 (0.74-0.84); P<0.001) was associated

Table 1. Transplant characteristics.

Number of patients Age, median (range) Patient sex Diagnosis

Disease status at HSCT

Donor/recipient relationship Donor/recipient sex match

Time from diagnosis to HSCT, median (range) Conditioning intensity Stem cell source GvHD prophylaxis

ATG/Campath

Male Female MM MPE MDS AML ALL Other CR noCR Untreated Related Unrelated MM MF FM FF

standard reduced BM PB CsA alone CsA + MTX CsA + MMF Tacrolimus + MTX Tacrolimus + MMF Other no yes

1990-1995 N (%)

1996-2000 N (%)

2001-2005 N (%)

2006-2010 N (%)

2011-2015 N (%)

3512 (100) 35.4 (18-75.8) 2074 (59.1) 1438 (40.9) 154 (4.4 %) 1006 (28.6) 280 (8) 1109 (31.6) 422 (12) 541 (15.4) 2399 (69.6) 940 (27.3) 107 (3.1) 2954 (84.1) 558 (15.9) 1195 (34.3) 862 (24.7) 765 (21.9) 664 (19) 8.6 (0-237.2)

9521 (100) 39.4 (18-71.7) 5598 (58.8) 3923 (41.2) 460 (4.8) 3140 (33) 797 (8.4) 2437 (25.6) 1106 (11.6) 1581 (16.6) 5928 (64) 3004 (32.4) 333 (3.6 %) 6776 (71.2) 2745 (28.8) 3389 (36.2) 2117 (22.6) 2040 (21.8) 1803 (19.3) 9.8 (0-238.2)

19865 (100) 44.5 (18-77.3) 11555 (58.2) 8310 (41.8) 1912 (9.6) 3071 (15.5) 1935 (9.7) 5754 (29) 2320 (11.7) 4873 (24.5) 11022 (57.3) 7178 (37.3) 1042 (5.4) 12406 (62.5) 7459 (37.5) 7184 (36.8) 4178 (21.4) 4511 (23.1) 3657 (18.7) 10.6 (0.1-239.8)

30194 (100) 48.9 (18-83.8) 17633 (58.4) 12561 (41.6) 2111 (7) 2768 (9.2) 3800 (12.6) 10358 (34.3) 3726 (12.3) 7431 (24.6) 17573 (60) 10071 (34.4) 1668 (5.7) 14486 (48) 15708 (52) 11413 (38.5) 5887 (19.9) 7020 (23.7) 5303 (17.9) 10.2 (0-238.3)

39465 (100) 52.4 (18-79.7) 23481 (59.5) 15984 (40.5) 2382 (6) 3306 (8.4) 6258 (15.9) 13799 (35) 4550 (11.5) 9170 (23.2) 24068 (62.7) 12274 (32) 2016 (5.3) 15262 (38.7) 24203 (61.3) 15851 (41) 7176 (18.6) 9227 (23.9) 6424 (16.6) 9 (0-239.9)

3335 (100)

6187 (83) 1263 (17) 5411 (56.8) 4110 (43.2) 811 (12.4) 4244 (64.6) 167 (2.5)

11324 (58.6) 8009 (41.4) 4685 (23.6 %) 15180 (76.4) 1808 (15.6) 5703 (49.1) 1805 (15.5) 26 (0.2) 59 (0.5) 2207 (19) 7680 (59.8) 5169 (40.2)

15610 (52.1) 14324 (47.9) 4094 (13.6) 26100 (86.4) 3627 (13.5) 11574 (43.1) 6009 (22.4) 301 (1.1) 782 (2.9) 4581 (17) 13531 (48.8) 14194 (51.2)

18806 (48.3) 20123 (51.7) 3816 (9.7) 35649 (90.3) 4682 (12.1) 14962 (38.6) 9053 (23.4) 659 (1.7) 1327 (3.4) 8062 (20.8) 15420 (39.2) 23956 (60.8)

3369 (95.9) 143 (4.1) 276 (9) 2340 (76.3)

451 (14.7) 2821 (89.5) 331 (10.5)

1 (0 .0) 1342 (20.4) 5150 (74.1) 1798 (25.9)

<0.001 0.01 <0.001

<0.001

<0.001 <0.001

<0.001 <0.001 <0.001 <0.001

<0.001

AML: acute myeloid leukemia; ALL: acute lymphoblastic leukemia; MDS: myelodysplastic syndrome; MPE: myeloproliferative disease; MM: myeloma; HSCT: hematopoietic stem cell transplantation; CR: complete remission; URD: unrelated donor; BM: bone marrow; PBSC: peripheral blood stem cells; CsA: cyclosporine A; MTX: methotrexate; MMF: mycophenolate mofetil; ATG: antithymocyte globulin.

1056

haematologica | 2022; 107(5)

Improved survival after acute GvHD

with a lower incidence. Of note is that aGvHD grades II-IV decreased with increasing transplant year (HR 0.57 (0.55-0.59); P<0.001), even in patients treated without ATG or alemtuzumab. The increased risk of aGvHD due to an increase in age was minimally reduced in more recent transplant years (per year: HR = 0.99 (0.99-1.0), P=0.05). Incidences of aGvHD grades III-IV by 100 days signifi-

cantly decreased from 19% (18-20%), to 16% (16-17%), 13% (13-14%), 11% (11-12%), and 11% (11-11%) for the periods 1990-1995, 1996-2000, 2001-2005, 2006-2010, and 2011-2015, respectively (Figure 1B). In multivariate analysis, URD (HR 1.52 (1.43-1.62); P<0.001), not in CR at HSCT (HR 1.5 (1.41-1.6); P<0.001) or untreated (HR 1.29 (1.13-1.46); P<0.001), use of BM instead of PBSC as stem cell source (HR 1.08; 1.01-1.16;

Figure 1. Incidences of aGvHD grades II-IV and grades III-IV over time. (A) Incidences of aGvHD grades II-IV over time. Cumulative incidence estimates were calculated for aGvHD grades II-IV with mortality considered as a competing event. The endpoint aGvHD grades II-IV was estimated from the date of first transplantation. Data are shown according to transplantation periods 1990-1995, 1996-2000, 2001-2005, 2006-2010 and 2011-2015. (B) Incidences of aGvHD grades III-IV over time. Cumulative incidence estimates were calculated for aGvHD grades III-IV with mortality considered as a competing event. The endpoint aGvHD grades III-IV was estimated from the date of first transplantation. Data are shown according to transplantation periods 1990-1995, 1996-2000, 2001-2005, 2006-2010 and 2011-2015.

Table 2. Multivariable Cox regression analysis regarding acute GvHD grades II-IV.

aGvHD II-IV Age at HSCT (dec) Conditioning intensity Donor/recipient relationship Disease status at HSCT

Stem cell source Recipient/donor sex match

ATG/Alemtuzumab

Standard Reduced Related Unrelated CR noCR untreated PB BM MM MF FM FF No Yes

HSCT year (dec) ATG/Alemtuzumab x HSCT year (dec)

aGvHD III-IV

HR (95% CI)

1.02 (1-1.03)

0.022

1.02 (1-1.04)

0.11

0.93 (0.9-0.97)

0.001

0.88 (0.83-0.94)

<0.001

1.61 (1.54-1.67)

<0.001

1.52 (1.43-1.62)

<0.001

1.25 (1.2-1.3) 1.11 (1.02-1.2)

<0.001 0.02

1.5 (1.41-1.6) 1.29 (1.13-1.46)

<0.001 <0.001

1.2 (1.15-1.25)

<0.001

1.08 (1.01-1.16)

0.029

1.16 (1.11-1.21) 0.93 (0.89-0.98) 1.03 (0.98-1.08)

<0.001 0.002 0.3

1.2 (1.12-1.29) 0.87 (0.81-0.94) 0.89 (0.82-0.96)

<0.001 <0.001 0.004

0.79 (0.74-0.84) 0.57 (0.55-0.59) 0.97 (0.92-1.03)

<0.001 <0.001 0.4

1.03 (0.95-1.12) 0.56 (0.53-0.6) 0.73 (0.67-0.8)

0.5 <0.001 <0.001

OS: overall survival; DFS: disease-free survival; NRM: non-relapse mortality; CR: complete remission; BM: bone marrow; PBSC: peripheral blood stem cells; ATG: antithymocyte globulin; HSCT: hematopoietic stem cell transplantation.

haematologica | 2022; 107(5)

1057

H.T. Greinix et al.

P<0.001), and a female donor for a male recipient (HR 1.2; 1.12-1.28; P<0.001) were all associated with increased risk for aGvHD grades III-IV whereas RIC (HR 0.88 (0.830.94); P<0.001), male donor (HR 0.87 (0.81-0.94); P<0.001) or female donor for a female recipient (HR 0.89 (0.82-0.96); P=0.004) were associated with a lower incidence, respectively. Use of ATG/alemtuzumab reduced risk for aGvHD grades III-IV per decade (HR 0.73 (0.670.8), P<0.001). Of note is that aGvHD grades III-IV also

decreased per decade in patients treated without ATG or alemtuzumab (HR 0.56 (0.53-0.6); P<0.001).

Outcomes For the total study population (n=102,557), three-year OS significantly increased from 49% (48-51%), to 51% (50-52%), 52% (52-53%), 53% (53-54%), and 54% (5354%) for the periods 1990-1995, 1996-2000, 2001-2005, 2006-2010, and 2011-2015, respectively (Figure 2).

Figure 2. Three-year overall survival over time for the whole patient cohort. The probability of OS for all patients was calculated using the Kaplan Meier estimator. Data are shown according to transplantation periods 1990-1995, 1996-2000, 2001-2005, 2006-2010 and 2011-2015.

Figure 3. Three-year overall survival over time. (A) Three-year overall survival after aGvHD grades II-IV over time. The probability of OS for patients experiencing aGvHD grades II-IV was calculated using the Kaplan Meier estimator. Data are shown according to transplantation periods 1990-1995, 1996-2000, 2001-2005, 2006-2010 and 2011-2015. (B) Three-year overall survival after aGvHD grades III-IV over time. The probability of OS for patients experiencing aGvHD grades III-IV was calculated using the Kaplan Meier estimator. Data are shown according to transplantation periods 1990-1995, 1996-2000, 2001-2005, 2006-2010 and 2011-2015.

1058

haematologica | 2022; 107(5)

Improved survival after acute GvHD

Outcomes after aGvHD grades II-IV Survival at 36 months after aGvHD grades II-IV increased significantly (P<0.001) from 38% (36-41%) to 40% (38-42%), 43% (42-45%), 44% (43-45%), and 45% (44-46%), respectively, for the times periods studied (Figure 3A). Causes of death are shown in Online Supplementary Table 2: though GvHD- and infection-associated deaths decreased over time, mortality due to relapse/progression of underlying malignant disease increased. Results of multivariate analysis are shown in Table 3. In multivariate analysis, URD (HR 1.14; 1.09-1.2; P<0.001), not in CR at HSCT (HR 1.47; 1.4-1.55; P<0.001), PBSC as stem cell source (HR 1.09 (1.03-1.15); P=0.004), a female donor for a male recipient (HR 1.11; 1.05-1.18; P<0.001), and use of ATG/alemtuzumab (HR 1.27 (1.18-1.36); P<0.001) were all associated with increased mortality after aGvHD grades II-IV whereas RIC (HR 0.92; 0.87-0.97; P=0.004) was associated with lower mortality, respectively. Patients transplanted more recently and given BM had improved outcomes after experiencing aGvHD: in earlier HSCT periods, BM as stem cell source had no significant impact on mortality (HR 0.95 (0.89-1.02); P=0.2) per decade but, in later periods, BM was associated with significantly reduced mortality after aGvHD (HR 0.89 (0.820.96); P=0.003). Of note is mortality increasing with increasing patient age (HR 1.15 per decade; 1.13-1.18; P<0.001) e.g., three-year OS was 60% (57-63%) in patients aged 20 years and 40% (37-44%) in patients aged 70 years, respectively. Mortality decreased in the more recent transplant years (HR 0.73; 0.7-0.78; P<0.001). Three-year NRM after experiencing aGvHD grades II-IV significantly (P<0.001) decreased from 47% (45-50%), to 42% (40-44%), 35% (34-37%), 37% (36-38%), and 36% (35-37%) for the periods 1990-1995, 1996-2000, 2001-

2005, 2006-2010, and 2011-2015, respectively (Figure 4). In multivariate analysis (Table 3), age (HR 1.22 (1.181.25); P<0.001), URD (HR 1.23 (1.15-1.31); P<0.001), not in CR at HSCT (HR 1.21 (1.13-1.29); P<0.001), and female donor for a male recipient (HR 1.22; 1.14-1.31; P<0.001) were associated with increased NRM whereas more recent HSCT year (HR 0.72 (0.67-0.77); p<0.001) and RIC (HR 0.89 (0.83-0.95); P<0.001) were associated with decreased NRM. Three-year DFS after aGvHD grades II-IV significantly (P<0.001) increased from 34% (32-37%), to 35% (3336%), 38% (36-39%), 39% (38-40%), and 40% (38-41%), respectively. In multivariate analysis, use of RIC (HR 0.95 (0.9-1); P=0.04) was associated with improved DFS after experiencing aGvHD grades II-IV whereas URD (HR 1.09 (1.041.14); P<0.001) and not in CR at HSCT (HR 1.48 (1.421.56); P<0.001) had reduced DFS, respectively. Of note is that DFS reduced with increasing patient age (HR 1.12; 1.09-1.14; P<0.001), but improved in more recent transplants in patients not treated with ATG/alemtuzumab (HR 0.8 (0.76-0.84); P<0.001) and in patients treated with ATG/alemtuzumab (HR 0.79 (0.74-0.85); P<0.001). Three-year relapse incidence after aGvHD grades II-IV significantly (P<0.001) increased from 19% (17-21%), to 23% (22-24%), 27% (26-28%), 25% (24-26%), and 25% (24-26%), respectively (Figure 5). In multivariate analysis, use of URD (HR 0.9; 0.84-0.98; P=0.01) and female donor for a male recipient (HR 0.83 (0.76-0.9); P<0.001) and use of BM as stem cell source (HR 0.87 (0.8-0.95); P=0.002) were associated with reduced relapse incidence whereas not being in CR at HSCT (HR 2.02 (1.87-2.18); P<0.001), and use of PBSC as stem cell source (HR 1.15; 1.05-1.25; P=0.002) were associated with increased relapse risk. Although conditioning intensity

Table 3. Multivariable Cox regression analysis regarding outcome after aGvHD grades II-IV.

OS HR (95% CI) Age at HSCT (dec) Conditioning intensity Donor/recipient relationship Disease status at HSCT

Stem cell source Recipient/donor sex match

ATG/Alemtuzumab HSCT year (dec) ATG/Alemtuzumab x HSCT year (dec)

1.16 (1.13-1.18) <0.001 standard reduced related unrelated CR noCR untreated PB BM MM MF FM FF no yes

0.92 (0.87-0.97)

0.004

DFS HR (95% CI) P

Relapse HR (95% CI) P

1.12 (1.09-1.14) <0.001 0.99 (0.96-1.02) 0.7 0.95 (0.9-1)

0.041

NRM HR (95% CI) P 1.22 (1.18-1.25) <0.001

1.03 (0.95-1.13) 0.4

0.89 (0.83-0.95) <0.001

0.9 (0.84-0.98) 0.011

1.23 (1.15-1.31) <0.001

1.14 (1.09-1.2) <0.001

1.09 (1.04-1.14) <0.001

1.47 (1.4-1.55) <0.001 1.05 (0.94-1.17) 0.4

1.48 (1.42-1.56) <0.001 2.02 (1.87-2.18)<0.001 1.21 (1.13-1.29) <0.001 1.04 (0.93-1.16) 0.5 0.94 (0.77-1.15) 0.5 1.07 (0.94-1.22) 0.3

0.92 (0.87-0.97)

0.004

0.92 (0.87-0.97) 0.002

0.87 (0.8-0.95) 0.002

1.11 (1.05-1.18) <0.001 0.96 (0.91-1.02) 0.2 0.95 (0.89-1.01) 0.08

1.05 (1-1.11) 0.07 0.96 (0.91-1.01) 0.14 0.94 (0.89-1) 0.047

0.83 (0.76-0.9) <0.001 1.22 (1.14-1.31) <0.001 0.92 (0.84-1) 0.05 0.99 (0.92-1.07) 0.8 0.93 (0.85-1.02) 0.11 0.95 (0.88-1.03) 0.2

1.27 (1.18-1.36) <0.001 0.76 (0.72-0.8) <0.001 0.81 (0.75-0.87) <0.001

1.34 (1.25-1.43) <0.001 1.3 (1.16-1.45) <0.001 1.34 (1.23-1.45) <0.001 0.8 (0.76-0.84) <0.001 0.93 (0.86-1) 0.06 0.72 (0.67-0.77) <0.001 0.79 (0.74-0.85) <0.001 0.88 (0.78-0.99) 0.03 0.75 (0.69-0.82) <0.001

0.94 (0.88-1.01)

0.12

haematologica | 2022; 107(5)

1059

H.T. Greinix et al.

Figure 4. Three-year non-relapse mortality after aGvHD grades II-IV over time. Cumulative incidence estimates were calculated for NRM in a competing risks framework with relapse as a competing risk for estimation of NRM. Data are shown according to transplantation periods 1990-1995, 1996-2000, 2001-2005, 2006-2010 and 2011-2015.

had no significant impact on relapse incidence for the cohort as a whole, patients with AML (HR 1.22 (1.081.37); P<0.001) and ALL (HR 1.53 (1.21-1.94), P<0.001) given RIC had a significant increased relapse risk while this was not the case in patients with MDS, respectively.

Outcomes after aGvHD grades III-IV Survival at 36 months after aGvHD grades III-IV increased significantly (P<0.001) from 22% (19-25%) to 23% (21-25%), 28% (27-30%), 28% (27-30%), and 29% (27-31%) for the periods 1990-1995, 1996-2000, 20012005, 2006-2010, and 2011-2015, respectively (Figure 3B). In multivariate analysis, URD (HR 1.18 (1.1-1.27); P<0.001), not in CR at HSCT (HR 1.25 (1.17-1.34); P<0.001) and use of ATG/alemtuzumab (HR 1.28 (1.171.4); P<0.001) were all associated with increased mortality after aGvHD grades III-IV. Of note is that mortality decreased in the more recent transplant cohorts treated without ATG/Alemtuzumab (HR 0.74 (0.69-0.8); P<0.001) and treated with ATG/Alemtuzumab (HR 0.86 (0.78-0.95); P=0.002).

Discussion In retrospective analyses, improvement over time in survival outcome for patients given allogeneic HSCT has been reported in parallel with changes of transplant practices. These include a spectrum of diseases treated with allografting, more frequent use of PBSC rather than BM, administration of more unrelated instead of related donor grafts, and older patient and donor age.11 Whether outcome improvement over time is also found in patients experiencing severe aGvHD is less well known. Therefore, we ana1060

lyzed the outcome of patients experiencing severe aGvHD over time in a large patient cohort reported to the EBMT. We observed a significant decrease of aGvHD grades II to IV and grades III to IV over time that was most pronounced in the more recent transplant cohort. Since having an URD compared to a related donor significantly increased the risk for aGvHD in multivariate analysis of our patient cohort, improvements in HLA typing in the more recent transplant years could have an impact on this finding. Over recent years and based on the outcome of numerous studies, the identity of ten alleles in five HLA loci, namely HLA-A, -B, -C, -DRB1, and -DQB1, and using high-resolution typing instead of serologic typing has become the gold standard of URD matching. Several studies have shown an association between allelic mismatches in HLA-A, -B, -C and -DRB1 and higher rates of aGvHD.14-16 Recent developments in clinical diagnosis, improved understanding of pathophysiological features, the use of both standard and experimental options for prevention, and the use of biomarkers to tailor treatment to individual patients could lead to a further reduction in aGvHD rates in the future. After adjusting for significant patient-, disease-, and transplant-related variables, patients with aGvHD grades II-IV, in the more recent cohort, had significantly lower NRM and better DFS and OS compared with those in the earlier years. Three-year survival of patients experiencing aGvHD grades II-IV improved significantly over time, reaching 45% for patients given HSCT between 2011 and 2015. Whereas Khoury et al., observed significant improvements in OS over time that was limited to patients treated with tacrolimus-based GvHD prophylaxis, and mostly in those with overall grade II aGvHD, GvHD prophylaxis had no significant impact on OS in our cohorts, in multivariate analysis.12 Of note is that threehaematologica | 2022; 107(5)

Improved survival after acute GvHD

Figure 5. Three-year relapse incidence after aGvHD grades II-IV over time. Cumulative incidence estimates were calculated for relapse in a competing risks framework with NRM as a competing risk in the estimation of malignancy relapse.

year OS also increased over time in patients experiencing aGvHD grades III-IV, reaching 29% in the most recent HSCT cohort, independent of GvHD prophylaxis. Previous studies reported an overall long-term survival of 10% to 25% in patients with severe aGvHD, defined as overall grades III to IV.17-20 Our survival rates compare favorably to Khoury et al. as well as El-Jawahri et al., who reported that longer time to aGvHD onset and younger recipient age were associated with improved OS,12,21 respectively. In our study, recipients of URD, patients not in CR at HSCT, and use of a female donor for a male patient were associated with increased mortality after aGvHD grades III-IV whereas recipients of RIC had a significantly lower mortality. The significant improvement in outcome of patients over the past 25 years can also be seen in the decrease of three-year NRM from 47% in 1990-1995 to 36% in 20112015 in patients experiencing aGvHD grades II-IV. The reduction in NRM over time is even more impressive when considering the significant changes in patient characteristics over recent decades, with an increase of median recipient age from 35.4 to 52.3 years, more frequent use of URD (from 15.8% to 61.3%) and an increase in patients not in CR at HSCT from 27.3% to 32%, respectively. The reasons behind these impressive improvements are likely multifactorial, including improved prevention and treatment of infectious complications that are a main cause of morbidity and mortality of GvHD patients under long-lasting immunosuppressive treatment and improved supportive care practices.22 Of note is that infectious death in our study declined from 30.1% to 23.4% over time. McDonald et al., recently reported a significant reduction of NRM in a patient cohort undergoing HSCT between 2013 and 2017, compared to previous haematologica | 2022; 107(5)

years.23 No change in overall cytomegalovirus (CMV) infection, but a substantial reduction in higher level CMV DNAemia and in gram-negative bacteremia and invasive mold infections was observed. This supports the notion that the use of less intensive conditioning regimens and the availability and use of improved antifungal drugs as prophylaxis for high-risk patients may have contributed to this reduction. Of note, in our study, RIC administration was also significantly associated with a lower risk for NRM. In recent years, clinicians have applied lower steroid doses for front-line therapy of aGvHD compared to the 1990s23,24 due to an increased awareness of steroid toxicity and increased NRM without improved response rates resulting from high-dose steroid treatment.25,26 Preclinical studies have revealed major pathophysiological pathways driving aGvHD and including tissue damage due to the administration of conditioning regimens or infection, alloreactivity seen as the body’s recognition of foreign major and minor histocompatibility antigens, and altered mechanisms of tissue repair and protection, such as microbiome dysregulation with a decline in protective microbial-derived metabolites.27 In recent years, more therapeutic strategies regarding aGvHD have become available in clinic, including the administration of costimulatory pathway blockade, targeted anti–interleukin-6 monoclonal antibodies, histone deacetylase inhibitors, kinase inhibitors and proteasome inhibitors, the antiinflammatory protease inhibitor alpha-1-antitrypsin, CTLA-4 antagonism, CCR5 blockade, and adoptive regulatory T cell transfer.27-34 These and other new strategies that are being developed will have a positive impact not only on response rates of aGvHD but also on the overall outcome of patients afflicted by this serious HSCT complication. 1061

H.T. Greinix et al.

Interestingly, the use of BM as stem cell source was associated with an increased risk for aGvHD grades II-IV and grades III-IV but also with a significantly higher DFS and overall survival. In a phase III, multicenter, randomized study of transplantation of PBSC versus BM from URD, the rates of aGvHD were similar in the two groups.35 However, these incidences were around 50% and, thus, markedly higher than in our cohorts where, in the most recent patient group, incidence was 28%. Further, Anasetti et al., did not observe survival differences between their study cohorts. In a retrospective analysis including 2463 recipients of PBSC and 1713 of BM from URD, no significant differences in the three-year probabilities of TRM, relapse, leukemia-free survival, and OS between the groups were observed in patients with leukemia and MDS.36 In a long-term follow-up report of the randomized study, recipients of URD BM had better psychological well-being, less burdensome cGvHD symptoms, and were more likely to return to work than recipients of PBSC at five years after HSCT.37 In our study, three-year relapse incidence after aGvHD grades II-IV significantly increased from 19% to 25% over time and was associated with a lack of CR at HSCT. It is important to acknowledge that more patients with MDS have been given HSCT in recent years and, of these, it is probable that patients were either not in CR or were untreated due to the fact that it still controversial as to whether pretreatment of MDS patients prior to HSCT is of clinical benefit and thus, should be recommended.38 Furthermore, in recent years, AML patients referred to HSCT in first CR have intermediate and adverse risk disease, as defined by the European Leukemia Net criteria39 and, especially, the later patient category is known to have a higher relapse risk after HSCT.39 Three-year relapse incidence in our cohort was also significantly increased after the use of RIC in patients with AML and MM. In line with our findings, substantially higher relapse rates after RIC compared to myeloablative conditioning have been reported in patients with AML and MDS.40,41 The strength of our study is the large sample size and long time period for comparison of HSCT outcome, as well as the participation of many transplant centers reporting consecutive patients to the EBMT registry and thus, providing real world data for detailed analysis over time. We would like to acknowledge the following limitations to this analysis. First, we cannot distinguish between mismatched and matched unrelated donors (10/10).

References 1. Passweg JR, Baldomero H, Chabannon C, et al. The EBMT activity survey on hematopoietic-cell transplantation and cellular therapy 2018: CAR-T's come into focus. Bone Marrow Transplant. 2020; 55(8):1604-1613. 2. Ferrara JL, Reddy P. Pathophysiology of graft-versus-host disease. Semin Hematol. 2006;43(1):3-10. 3. Jagasia M, Arora M, Flowers ME, et al. Risk factors for acute GVHD and survival after hematopoietic cell transplantation. Blood. 2012;119(1):296-307. 4. Cutler C, Logan B, Nakamura R, et al. Tacrolimus/sirolimus vs

1062

Second, limitations in data on antimicrobial agents and other supportive care measures do not allow an analysis of changes in these practices over time as a potential factor for improved outcome. Thirdly, insufficient data on steroid dose, and type/duration of salvage immunosuppressive therapy do not allow detailed analyses of treatment intensity on outcome. In addition, detailed aGvHD treatment response data are not available, therefore we cannot characterize the burden of steroid-refractory aGvHD across cohorts. In conclusion, our findings demonstrate that the advances and changes in allogeneic HSCT practices over the past 2.5 decades have led to significantly improved outcomes in patients experiencing severe aGvHD. Although incidences of aGvHD have significantly declined and the OS of patients experiencing aGvHD has improved, there is still a need for further progress. Increasing use of posttransplant cyclophosphamide for GvHD prophylaxis, not only after haploidentical but also related and URD transplants, could have an impact on incidences of both acute as well as chronic GvHD, as previously reported.42,43 The administration of BM rather than PBSC as stem cell source reportedly reduces the incidence of cGvHD and the use of myeloablative conditioning regimens in patients with aggressive malignant disease is another option for improvement of HSCT outcome. More efficient and less toxic front-line immunosuppressive therapies for treatment of aGvHD, including treatments without the administration of corticosteroids, would have the potential to further reduce NRM and improve survival of patients following allogeneic HSCT. Disclosures No conflicts of interest to disclose. Contributions HTG designed the study and wrote the manuscript; LK prepared the dataset and DJE LK performed statistical analyses; OP, IYA, SM, CC, JS, AN, MR, SR, YC, MM, SS, ZP, AR, FL, MM, GWB, and NK interpreted the data. All authors critically reviewed the manuscript and approved the final version for submission. Acknowledgments The authors would like to thank all patients and their families. They would also like to thank all data managers and transplant centers for contributing and collecting their data for the EBMT registry.

tacrolimus/methotrexate as GVHD prophylaxis after matched, related donor allogeneic HCT. Blood. 2014;124(8):1372-1377. 5. Nash RA, Antin JH, Karanes C, et al. Phase 3 study comparing methotrexate and tacrolimus with methotrexate and cyclosporine for prophylaxis of acute graftversus-host disease after marrow transplantation from unrelated donors. Blood. 2000; 96(6):2062-2068. 6. Finke J, Bethge WA, Schmoor C, et al. Standard graft-versus-host disease prophylaxis with or without anti-T-cell globulin in haematopoietic cell transplantation from matched unrelated donors: a randomised, open-label, multicentre phase 3 trial. Lancet Oncol. 2009;10(9):855-864. 7. MacMillan ML, DeFor TE, Weisdorf DJ.

The best endpoint for acute GVHD treatment trials. Blood. 2010;115(26):5412-5417. 8. Levine JE, Logan B, Wu J, et al. Graft-versus-host disease treatment: predictors of survival. Biol Blood Marrow Transplant. 2010;16(12):1693-1699. 9. Hahn T, McCarthy PL Jr, Hassebroek A, et al. Significant improvement in survival after allogeneic hematopoietic cell transplantation during a period of significantly increased use, older recipient age, and use of unrelated donors. J Clin Oncol. 2013; 31(19):2437-2449. 10. Juric MK, Ghimire S, Ogonek J, et al. Milestones of hematopoietic stem cell transplantation - from first human studies to current developments. Front Immunol. 2016;7:470.

haematologica | 2022; 107(5)

Improved survival after acute GvHD

11. Gooley TA, Chien JW, Pergam SA, et al. Reduced mortality after allogeneic hematopoietic-cell transplantation. N Engl J Med. 2010;363(22):2091-2101. 12. Khoury HJ, Wang T, Hemmer MT, et al. Improved survival after acute graft-versushost disease diagnosis in the modern era. Haematologica. 2017;102(5):958-966. 13. Przepiorka D, Weisdorf D, Martin P, et al. 1994 Consensus Conference on acute GVHD grading. Bone Marrow Transplant. 1995;15(6):825-828. 14. Morishima Y, Sasazuki T, Inoko H, et al. The clinical significance of human leukocyte antigen (HLA) allele compatibility in patients receiving a marrow transplant from serologically HLA-A, HLA-B, and HLA-DR matched unrelated donors. Blood. 2002;99(11):4200-4206. 15. Petersdorf EW, Gooley TA, Anasetti C, et al. Optimizing outcome after unrelated marrow transplantation by comprehensive matching of HLA class I and II alleles in the donor and recipient. Blood. 1998; 92(10):3515-3520. 16. Sasazuki T, Juji T, Morishima Y, et al. Effect of matching of class I HLA alleles on clinical outcome after transplantation of hematopoietic stem cells from an unrelated donor. Japan Marrow Donor Program. N Engl J Med. 1998;339(17):1177-1185. 17. Klein SA, Bug G, Mousset S, Hofmann WK, Hoelzer D, Martin H. Long term outcome of patients with steroid-refractory acute intestinal graft versus host disease after treatment with pentostatin. Br J Haematol. 2011;154(1):143-146. 18. Schub N, Günther A, Schrauder A, et al. Therapy of steroid-refractory acute GVHD with CD52 antibody alemtuzumab is effective. Bone Marrow Transplant. 2011; 46(1):143-147. 19. Schmitt T, Luft T, Hegenbart U, Tran TH, Ho AD, Dreger P. Pentostatin for treatment of steroid-refractory acute GVHD: a retrospective single-center analysis. Bone Marrow Transplant. 2011;46(4):580-585. 20. Jamani K, Russell JA, Daly A, et al. Prognosis of grade 3-4 acute GVHD continues to be dismal. Bone Marrow Transplant. 2013;48(10):1359-1361. 21. El-Jawahri A, Li S, Antin JH, et al. Improved treatment-related mortality and overall survival of patients with grade IV acute GVHD in the modern years. Biol Blood Marrow Transplant. 2016;22(5):910-918. 22. Horan JT, Logan BR, Agovi-Johnson MA, et al. Reducing the risk for transplantationrelated mortality after allogeneic hematopoietic cell transplantation: how much progress has been made? J Clin Oncol. 2011;29(7):805-813.

haematologica | 2022; 107(5)

23. McDonald GB, Sandmaier BM, Mielcarek M, et al. Survival, nonrelapse mortality, and relapse-related mortality after allogeneic hematopoietic cell transplantation: comparing 2003-2007 cersus 2013-2017 cohorts. Ann Intern Med. 2020;172(4):229-239. 24. Wolff D, Ayuk F, Elmaagacli A, et al. Current practice in diagnosis and treatment of acute graft-versus-host disease: results from a survey among German-AustrianSwiss hematopoietic stem cell transplant centers. Biol Blood Marrow Transplant. 2013;19(5):767-776. 25. Van Lint MT, Uderzo C, Locasciulli A, et al. Early treatment of acute graft-versus-host disease with high- or low-dose 6-methylprednisolone: a multicenter randomized trial from the Italian Group for Bone Marrow Transplantation. Blood. 1998; 92(7):2288-2293. 26. Mielcarek M, Furlong T, Storer BE, et al. Effectiveness and safety of lower dose prednisone for initial treatment of acute graft-versus-host disease: a randomized controlled trial. Haematologica. 2015; 100(6):842-848. 27. Zeiser R, Blazar BR. Acute graft-versus-host disease - biologic process, prevention, and therapy. N Engl J Med. 2017; 377(22):21672179. 28. Zeiser R, Burchert A, Lengerke C, et al. Ruxolitinib in corticosteroid-refractory graft-versus-host disease after allogeneic stem cell transplantation: a multicenter survey. Leukemia. 2015;29(10):2062-2068. 29. Kennedy GA, Varelias A, Vuckovic S, et al. Addition of interleukin-6 inhibition with tocilizumab to standard graft-versus-host disease prophylaxis after allogeneic stemcell transplantation: a phase 1/2 trial. Lancet Oncol. 2014;15(13):1451-1459. 30. Koreth J, Stevenson KE, Kim HT, et al. Bortezomib, tacrolimus, and methotrexate for prophylaxis of graft-versus-host disease after reduced-intensity conditioning allogeneic stem cell transplantation from HLAmismatched unrelated donors. Blood. 2009;114(18):3956-3959. 31. Tawara I, Sun Y, Lewis EC, et al. Alpha-1antitrypsin monotherapy reduces graftversus-host disease after experimental allogeneic bone marrow transplantation. Proc Natl Acad Sci U S A. 2012;109(2):564569. 32. Koura DT, Horan JT, Langston AA, et al. In vivo T cell costimulation blockade with abatacept for acute graft-versus-host disease prevention: a first-in-disease trial. Biol Blood Marrow Transplant. 2013;19(11): 1638-1649. 33. Reshef R, Luger SM, Hexner EO, et al. Blockade of lymphocyte chemotaxis in vis-

ceral graft-versus-host disease. N Engl J Med. 2012;367(2):135-145. 34. Martelli MF, Di Ianni M, Ruggeri L, et al. HLA-haploidentical transplantation with regulatory and conventional T-cell adoptive immunotherapy prevents acute leukemia relapse. Blood. 2014;124(4):638-644. 35. Anasetti C, Logan BR, Lee SJ, et al. Peripheral-blood stem cells versus bone marrow from unrelated donors. N Engl J Med. 2012;367(16):1487-1496. 36. Eapen M, Logan BR, Appelbaum FR, et al. Long-term survival after transplantation of unrelated donor peripheral blood or bone marrow hematopoietic cells for hematologic malignancy. Biol Blood Marrow Transplant. 2015;21(1):55-59. 37. Lee SJ, Logan B, Westervelt P, et al. Comparison of patient-reported outcomes in 5-Year survivors who received bone marrow vs peripheral blood unrelated donor transplantation: long-term follow-up of a randomized clinical trial. JAMA Oncol. 2016;2(12):1583-1589. 38. Kröger N. Induction, bridging, or straight ahead: the ongoing dilemma of allografting in advanced myelodysplastic syndrome. Biol Blood Marrow Transplant. 2019; 25(8):e247-e249. 39. Döhner H, Estey E, Grimwade D, et al. Diagnosis and management of AML in adults: 2017 ELN recommendations from an international expert panel. Blood. 2017; 129(4):424-447. 40. Scott BL, Pasquini MC, Logan BR, et al. Myeloablative versus reduced-intensity hematopoietic cell transplantation for acute myeloid leukemia and myelodysplastic syndromes. J Clin Oncol. 2017;35(11):11541161. 41. Festuccia M, Deeg HJ, Gooley TA, et al. Minimal identifiable disease and the role of conditioning intensity in hematopoietic cell transplantation for myelodysplastic syndrome and acute myelogenous leukemia evolving from myelodysplastic syndrome. Biol Blood Marrow Transplant. 2016; 22(7):1227-1233. 42. 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. 43. Nagler A, Labopin M, Dholaria B, et al. Comparison of haploidentical bone marrow versus matched unrelated donor peripheral blood stem cell transplantation with post-transplant cyclophosphamide in patients with acute leukemia. Clin Cancer Res. 2021;27(3):843-851.

1063

ARTICLE Ferrata Storti Foundation

Haematologica 2022 Volume 107(5):1064-1071

Coagulation & its Disorders

A homozygous duplication of the FGG exon 8-intron 8 junction causes congenital afibrinogenemia. Lessons learned from the study of a large consanguineous Turkish family Michel Guipponi,1,2 Frédéric Masclaux,1 Frédérique Sloan-Béna,1 Corinne Di Sanza,2 Namik Özbek,3 Flora Peyvandi,4,5 Marzia Menegatti,4 Alessandro Casini,6 Baris Malbora7 and Marguerite Neerman-Arbez2 Medical Genetics Service, University Hospitals of Geneva, Geneva, Switzerland; Department of Genetic Medicine and Development, Faculty of Medicine, University of Geneva, Geneva, Switzerland; 3Department of Pediatric Hematology, Ankara City Hospital, Ankara, Turkey; 4Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico, Angelo Bianchi Bonomi Hemophilia and Thrombosis Center and Fondazione Luigi Villa, Milan, Italy; 5Università degli Studi di Milano, Department of Pathophysiology and Transplantation, Milan, Italy; 6Division of Angiology and Hemostasis, University Hospitals of Geneva, Geneva, Switzerland and 7Department of Pediatric Hematology and Oncology, Istanbul Yeni Yuzyil University, Istanbul, Turkey 1 2

ABSTRACT

Correspondence: MARGUERITE NEERMAN-ARBEZ Marguerite.Neerman-Arbez@unige.ch Received: April 9, 2021. Accepted: June 21, 2021. Pre-published: July 1, 2021. https://doi.org/10.3324/haematol.2021.278945

ongenital afibrinogenemia is the most severe congenital fibrinogen disorder, characterized by undetectable fibrinogen in circulation. Causative mutations can be divided into two main classes: null mutations with no protein production at all and missense mutations producing abnormal protein chains that are retained inside the cell. The vast majority of cases are due to single base pair mutations or small insertions or deletions in the coding regions or intron-exon junctions of FGB, FGA and FGG. Only a few large rearrangements have been described, all deletions involving FGA. Here we report the characterization of a 403 bp duplication of the FGG exon 8-intron 8 junction accounting for congenital afibrinogenemia in a large consanguineous family from Turkey. This mutation, which had escaped detection by Sanger sequencing of short polymerase chain reaction (PCR) amplicons of coding sequences and splice sites, was identified by studying multiple alignments of reads obtained from whole exome sequencing of a heterozygous individual followed by PCR amplification and sequencing of a larger portion of FGG. Because the mutation duplicates the donor splice site of intron 8, we predicted that the impact of the mutation would be on FGG transcript splicing. Analysis of mRNA produced by cells transiently transfected with normal or mutant minigene constructs showed that the duplication causes production of several aberrant FGG transcripts generating premature truncating codons.

1064

Introduction The ultimate goal of the coagulation cascade is the controlled conversion by thrombin of fibrinogen into fibrin which forms a polymer to give stability, strength and adhesive surfaces to growing blood clots containing platelets and red blood cells. Human fibrinogen1,2 is produced in the liver from three homologous polypeptide chains, Bβ, Aa and g encoded by the fibrinogen gene cluster FGB, FGA and FGG, on human chromosome 4. Two copies of each polypeptide chain assemble to form a hexamer (AaBβg) held together by disulphide bonds. Alternative spliced transcripts are produced for FGA and FGG, these are AaE and g’ respectively. While AaE chains are present in only 1-2% of circulating fibrinogen, g’ chains are present in 8-15% of circulating fibrinogen, in het2

haematologica | 2022; 107(5)

FGG mutation in congenital afibrinogenemia

erodimeric or homodimeric form. Normal plasma fibrinogen levels vary between 2 and 4 g L-1. Variation in fibrinogen levels is a complex trait, influenced by both the environment and genotype. Inherited disorders of fibrinogen include Type I disorders (afibrinogenemia and hypofibrinogenemia) which affect the quantity of fibrinogen in circulation and Type II disorders (dysfibrinogenemia and hypodysfibrinogenemia) which affect the quality of circulating fibrinogen.3,4 Congenital afibrinogenemia is the most severe disorder, characterized by undetectable fibrinogen in circulation.5,6 While the first dysfibrinogenemia mutation was identified as early as 19687 the molecular basis of afibrinogenemia was elucidated much later.8 This disorder is characterized by autosomal recessive inheritance and the complete absence of fibrinogen in plasma. In populations where consanguineous marriages are common, the prevalence of afibrinogenemia is increased.9 We identified the first causative mutation for congenital afibrinogenemia, a large, recurrent deletion in FGA in 19998 identified in homozygosity in four members of a Swiss family. Since then, the underlying molecular pathophysiology of numerous causative mutations leading to fibrinogen deficiency has been determined by our group and many others (reviewed in4).10-12 Causative mutations can be divided into two main classes: null mutations with no protein production at all and missense mutations producing abnormal protein chains that are retained inside the cell. The vast majority of cases are due to single base pair mutations or small insertions of deletions in the coding regions or intron-exon junctions of FGB, FGA and FGG. These can easily be identified by polymerase chain reaction (PCR) amplification followed by Sanger sequencing or by next-generation sequencing in particular whole exome sequencing (WES). Only a few large rearrangements have been described. In addition to the recurrent deletion we identified with breakpoints in FGA intron 1 and the FGA–FGB intergenic region, while three other large deletions in the fibrinogen gene cluster have been reported by others, all involving part of the FGA gene. These are a deletion of 1.2 kb eliminating the entire FGA exon 4 in a Japanese patient;13 a deletion of 15 kb, with breakpoints situated in FGA intron 4 and in the FGA–FGB intergenic region in a Thai patient;14 and a 4.1-kb deletion encompassing FGA exon 1 in an Italian patient.15 All patients were homozygous for the identified deletions except for the Thai patient, for whom complete maternal uniparental disomy was confirmed for the deleted chromosome 4.14 Rearrangements of this type cannot be identified by simple PCR analysis of coding regions. Other techniques such as array comparative genomic hybridization (CGH) can be useful in some cases, however the resolution of commercial arrays limits the discovery of rearrangements i.e., deletions, duplications greater than 15 kb. Consequently mutations less than 15 kb will escape detection using this technique in most diagnostic settings. Here we report the characterization of a 403 bp duplication of the FGG exon 8-intron 8 junction accounting for congenital afibrinogenemia in a large consanguineous family from Turkey. This mutation, which had escaped detection by Sanger sequencing of short PCR amplicons of coding sequences and splice sites, was identified by studying multiple alignments of reads obtained from WES of a heterozygous individual followed by PCR haematologica | 2022; 107(5)

amplification and sequencing of à larger portion of FGG. The mutation duplicates the donor splice site of intron 8 which leads to aberrant splicing of both the major g transcript and the minor g’ transcript.

Methods Patient samples This study was performed with Institutional Review Board approval and with written informed consent from all patients, in accordance with the Declaration of Helsinki. Platelet-poor plasma samples were obtained from citrated venous blood and analyzed as described in the Online Supplementary Appendix.

Polymerase chain reaction and Sanger sequencing Genomic DNA was extracted from whole blood-EDTA according to standard protocols. PCR amplifications of all FGB, FGA and FGG coding regions and intron-exon junctions were performed as previously described.16 Standard primer sequences and PCR protocols are available on demand. Specific primer sequences for this study are available in the Online Supplementary Appendix. Sanger sequencing of purified PCR products was performed by Fasteris AG, Geneva, Switzerland.

Array comparative genomic hybridization analysis The array CGH analysis was performed using Human Genome CGH Microarray Kit G3 1 M (Agilent Technologies, Palo Alto, USA) with ~2.4 kb overall median probe spacing according to protocols provided by the manufacturers. Copy number variant analysis was done using the Agilent Genomic Workbench Software 7.0.4.0. and UCSC Genome Browser Human Genome GRCh37/hg19.

Next-generation sequencing WES was performed at the Health 2030 Genome Center at Campus Biotech, Geneva using IDT Research Exome reagents. Read mapping and variant calling were performed using BWA 0.7.13, Picard 2.9.0, GATK HaplotypeCaller 3.7, aligned to the GRCh37/hg19 reference genome and annotated with Annovar 2017/07/17 and UCSC RefSeq (refGene) downloaded on 2018/08/10.

Minigene constructs and transfections PCR products including intronic and exonic sequences from FGG intron 7 to FGG exon 10 were amplified from the genomic DNA of one homozygous affected individual and one normal individual and cloned into the pcDNA3.1 V5His TOPO-TA eukaryotic expression vector (Invitrogen) to obtain mutant and normal minigenes. The presence of the 403 bp duplicated fragment in the mutant clone was confirmed by Sanger sequencing. Transient transfections of HEK-293T cells (105 cells/ condition) were performed in 6-well plates using Lipofectamine 2000 (Invitrogen) in OptiMEM (Gibco Invitrogen) and 2 μg of normal or mutant construct. Two days post-transfection, cells were lyzed in Trizol for RNA extraction using the Turbo DNA free kit (Invitrogen). Reverse transcription and PCR amplification of cDNA for analysis of splicing variants are described in the Online Supplementary Appendix.

Results The members of the large consanguineous family (Figure 1) all originate from a village in Turkey which 1065

M. Guipponi et al.

keeps a religious faith different to the surrounding villages in the area. While the surrounding villages have a 'Sunni' faith, the village where the patients live has an Alavian-Bektashi faith. Since there are no marriages between these two religious groups, all marriages are between individuals from the same village. Interestingly, the original founders migrated from the Iranian province of Mazandaran, formerly known as Taberistan, during the 9th and 10th centuries. Current inhabitants of Maznandar are also of Alavian-Bektashi faith. Bleeding severity was assessed according to the score from the EN-RBD study.17 Patients were divided into clinical bleeding severity categories (asymptomatic and grade 1, 2, or 3 bleeding). Category 1 refers to provoked bleeding episodes, category 2 refers to spontaneous minor bleeding episodes (e.g., bruising), and category 3 refers to spontaneous major bleeding episodes (e.g., cerebral bleeds or hemarthrosis).17 Of the eight patients diagnosed with afibrinogenemia (Figure 1), four were available for genetic analysis. Fibrinogen measurements, both antigenic and coagulable, were performed for 41 additional family members (Table I). Three patients available for study were male (patient ID: 1283, 1316, and 1317), and one was female (patient ID: 1314). All patients with afibrinogenemia have a severe clinical phenotype (grade 3) while most heterozygous patients with hypofibrinogenemia have a grade 0 (mean, 0.5). No patient received fibrinogen on prophylaxis. Two male patients with afibrinogenemia (1283 and

1317), for which no additional clinical information is available, have experienced a thrombotic event. This is not unusual; afibrinogenemia is associated with an increased thrombotic risk even in the absence of additional thrombotic risk factors, whether genetic or environmental. In a recent study of 204 afibrinogenemic patients, 37 (18 %) reported a thrombotic event.18

Identification of a duplication of the FGG exon 8intron 8 junction We aimed to identify the causative mutation by first screening three affected patients (1314, 1317 and 1283) and one heterozygous carrier (1288) by PCR amplification of all FGB, FGA and FGG coding regions and intron-exon junctions followed by Sanger sequencing as previously described.16 This approach was unsuccessful, no causative mutation was identified. We then performed array-CGH analysis and identified 23 variants not listed in the database of Genomic Variants (http://dgv.tcag.ca/dgv/app/home). However none of these variants was a candidate for the afibrinogenemia phenotype so this approach was also unsuccessful. Finally, as part of our ongoing research project determining the causative mutations and genetic modifiers of congenital fibrinogen disorders which uses WES followed by variant calling in a panel of selected genes including the fibrinogen genes, we included one heterozygous carrier, 1292, the mother of affected patient 1283, in the study. A detailed analysis of the reads suggested the pres-

Figure 1. Family tree of the large consanguineous family. (A) Main family tree. (B) Other family members including relatives of individual 1305 (indicated with an asterisk).

1066

haematologica | 2022; 107(5)

FGG mutation in congenital afibrinogenemia

ence of an insertion of 36 bp in exon 8 of the FGG gene NM_021870.3,exon8,c.959_960insAATCCACCTGCTG CAAAATATCCAGTAGTTTGGCAT. This abnormality was present in only 14.2% of the reads (14/98) covering this position. The apparent localization of the insertion in FGG exon 8 was in contradiction with our previous results demonstrating the absence of a causative mutation in any fibrinogen gene coding sequence. BLAT analysis revealed that the 36 bp-long sequence was in fact in intron 8 of FGG suggesting the presence of a more complex local rearrangement. The visual inspection of the reads aligned to FGG exon 8 further supported this hypothesis with the presence of hard clipped and outward-facing read pairs spanning far apart from that expected based on the library insert size (Figure 2). This configuration of the read pairs was highly suggestive of the presence of a heterozygous duplication that could not be fully characterized by our exon-centered analysis. We therefore amplified by PCR a larger portion of FGG from intron 7 to exon 10 which yielded a 4 kb product corresponding to the normal sequence and a 4.4 kb product for affected individuals. Heterozygous individuals showed both bands, as expected (Figure 3A). Sequencing of the larger band revealed the presence of a duplicated sequence of 403 bp (out of 404, one base in a stretch of 4 intronic “A”s is missing) containing the last 169 bases from FGG exon 8, the donor splice site GT, and 232 additional bases of FGG intron 8 (Figure 3B). Genotypes for all family members are indicated in Table I. The presence of the duplication in homozygosity was associated with absence of fibrinogen in circulation and afibrinogenemia in all four affected individuals, while heterozygosity was associated with decreased fibrinogen levels (Table I).

Duplication causes aberrant splicing of both FGG transcripts Because the mutation duplicates the donor splice site of intron 8, we predicted that the impact of the mutation would be on FGG transcript splicing. Interestingly, two FGG transcripts are normally produced which differ at the 3’end: the major g chain mRNA has ten exons while in the minor g’ chain isoform intron 9 is not spliced out, substituting the four amino acids encoded by exon 10 with twenty g’ COOH-terminal residues.19-21 The presence of the duplication was thus anticipated to affect splicing of both isoforms. HEK-293T cells were transiently transfected with minigene constructs encompassing intron 7 to exon 10 with and without the duplication (Figure 4A). RNA produced were reverse transcribed to cDNA which were used in two different PCR reactions to amplify transcripts containing exon 10, present in the major g transcript, and transcripts containing the last bases of exon 9, retained in the minor g’ transcript but spliced out in the major g transcript. The results obtained for transcripts containing exon 10 (Figure 4B) showed one major product, indicated by an asterisk, for the normal minigene which was shown by sequencing to correspond to the correctly spliced mRNA containing exons 8, 9 and 10 encoding the major g chain. Sequencing of clones of individual PCR products showed that a transcript retaining intron 9 was also produced, thus encoding the minor g‘ transcript even though exon 10 is present. For the mutant, the major product obtained, indicated by an asterisk, retained intron 8 with the duplication, resulting haematologica | 2022; 107(5)

in a transcript with a frameshift and a premature truncating codon 13 codons downstream. Cloning of the PCR products allowed the identification of additional minor aberrant transcripts resulting in frameshifts and premature truncating codons. One includTable 1. Patient symptoms, fibrinogen measurements and genotypes.

Patient ID

1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321

Fibrinogen clauss (mg/dL)

Fibrinogen antigen (mg/dL)

197 273 282 169 214 231 <20 213 105 152 85 152 226 291 170 103 115 130 273 144 186 317 246 n.d. 272 334 90 152 269 108 441 266 203 389 222 123 149 < 20 127 < 20 < 20 338 121 179 165

193 204 254 215 218 206 <2 209 150 177 126 170 260 322 161 122 110 145 320 142 175 334 267 n.d. 260 342 86 157 267 132 448 248 192 450 216 161 187 <2 157 <2 <2 350 140 220 171

Bleeding Thrombosis Genotype score 0 0 0 0 0 0 3 0 0 0 0 1 0 0 0 0 2 1 0 0 0 0 0 0 0 0 2 0 0 1 0 0 0 0 0 0 0 3 0 3 3 0 1 1 0

Yes

normal normal normal heterozygous normal normal homozygous normal heterozygous heterozygous heterozygous heterozygous normal normal normal heterozygous heterozygous heterozygous heterozygous heterozygous normal normal normal normal normal normal heterozygous heterozygous normal heterozygous normal normal normal normal normal heterozygous heterozygous homzygous heterozygous homzygous homzygous normal heterozygous normal heterozygous 1067

M. Guipponi et al.

Figure 2. Whole exome sequencing read alignments. (A) IGV (Interactive Genome Viewer) screen capture of FGG gene showing reads between exon 7 and exon 10 (gene located on reverse strand, transcript NM_000509.5) aligned to the GRCh37/hg19 reference genome. Reads are colored by pair orientation as defined by standard IGV settings. Green color defines paired-end reads orientation inconsistencies which can delineate tandem duplication with respect to the reference genome. (B) Zoom on the reads that span the internal junction of the tandem duplication and only partially align to the reference genome (reads referred to as “hardclipped” at position 155’527’621).

Figure 3. Identification of a duplicated segment at the FGG exon 7-intron 8 junction. (A) PCR amplification of a portion of FGG from intron 7 to exon 10 yields a 4 kb product corresponding to the normal sequence and a 4.4 kb product for affected individuals. Heterozygous individuals show both bands. (B) Partial sequence of the 4.4 kb band. The acceptor “AG” site at the end of intron 7 and duplicated donor “GT” splice sites at the beginning of intron 8 are highlighted in yellow. Duplicated sequences are shown in blue. The duplicated sequence contains 403 bp out of 404 bp of the normal sequence, one base in a stretch of 4 intronic “A”s (shown in red) is missing.

1068

haematologica | 2022; 107(5)

FGG mutation in congenital afibrinogenemia

Figure 4. Minigene constructs and reverse transcription polymerase chain reaction analysis of splicing variants. (A) Minigene constructs for the normal (top) and duplicated (bottom) FGG sequence between intron 7 and the 3’UTR of exon 10. HEK-293T cells transiently transfected with these constructs can produce transcripts for both the major g transcript, containing exon 10, and the minor g’ transcript, which does not splice out the 3’ portion of exon 9. (B) cDNA obtained for transcripts containing exon 10. The major product for the normal minigene, indicated by an asterisk, corresponds to the correctly spliced mRNA containing exons 8, 9 and 10 encoding the major g chain. For the mutant, the major product, indicated by an asterisk, retains intron 8 with the duplication, resulting in a transcript with a frameshift and a premature truncating codon. A correctly spliced transcript was also identified for the mutant minigene, but this is expected to be a rare event based on the intensity of the corresponding band indicated by an arrow. (C) cDNA obtained for transcripts containing the last bases of exon 9 retained in the minor g’ transcript. The major band for the normal minigene, shown with an asterisk, corresponds to the normal g’ transcript while the major transcript produced from the mutant minigene, retained intron 8 resulting in a frameshift as described above. Again, normal splicing is observed (indicated by an arrow) but at low levels.

ed part of the duplicated exonic segment due to utilisation of the normal donor splice site of intron 8 and a cryptic acceptor site situated in the duplicated sequence, followed by normal splicing of intron 9 using the duplicated donor site. Another transcript showed skipping of exon 9, joining together exon 8 and exon 10. The latter product was also identified in a clone of the normal minigene. Finally, a normal correctly spliced transcript was also identified for the mutant minigene, but this is expected to be a rare event based on the intensity of the corresponding band indicated by an arrow (Figure 4B). In order to have a complete picture of the different transcripts produced by alternative splicing of exon 9, we performed a second PCR amplification using a reverse primer localized on the last bases of exon 9 which are retained in the minor g’ transcript but spliced out in the g transcript. Again, one major product was identified for the normal construct and one for the mutant construct (Figure 4C). Sequencing confirmed production of the normal g’ transcript for cells transfected with the normal minigene while those transfected with the mutant minigene retained intron 8 resulting in a frameshift as described above.

Discussion We describe here the identification of a homozygous duplication at the FGG exon 8-intron 8 junction accounthaematologica | 2022; 107(5)

ing for congenital afibrinogenemia in a large consanguineous family from Turkey. In principle identification of the mutation causing the complete deficiency of fibrinogen in the affected individuals, necessarily homozygous given the structure of the family pedigree, should be relatively straightforward for a laboratory equipped for standard genetic screening methods i.e., PCR amplification of the three fibrinogen encoding genes followed by Sanger sequencing. Indeed, we previously studied the molecular epidemiology of causative mutations for congenital fibrinogen disorders22 with the aim to design a cost- and time-effective screening strategy based on the genetic data from 266 unrelated patients genotyped in our laboratory. When we prospectively tested our strategy on 32 consecutive new probands we found that screening of FGA exons 2, 4, 5 and FGG exon 8 combined with the search for the 11 kb deletion of FGA led to the identification of approximately 80% of mutated alleles, including 15 new mutations.22 In this case the size of the duplicated sequence (403 bp) was too large to be detected by standard PCR amplification approaches, which often involve amplifying single exons with only the immediate intronic sequences, and too small to be detected by array CGH. Identification of the mutation was only possible following a deep investigation of an aberrant sequence picked up by WES analysis but not initially confirmed by Sanger sequencing, an approach which is unlikely to be undertaken by routine diagnostic labora1069

M. Guipponi et al.

tories. One lesson learned from this study is that amplifying larger overlapping portions of the genes of interest may allow the identification of similar mutations involving large insertions and duplications. In our case, amplifying for example FGG exons 7 to 10, around 4,400 bp in the normal sequence, and allowing sufficient elongation time to amplify the larger mutant band would have enabled the identification of the mutation sooner. Recent technology such as Nanopore DNA sequencing allowing sequencing of DNA fragments up to a few Mbp may be useful to identify such mutations, allowing for example the sequencing of the entire fibrinogen gene cluster with haplotype phasing of variant sites along the sequence. The duplication is only one of five mutations involving rearrangements of more than 100 bp of the fibrinogen genes, and the first identified in the FGG gene. As previously mentioned four deletions of several kilobases have been identified in FGA including the recurrent deletion we identified in a Swiss family.8,13-15 In addition, an inframe duplication of 117 bp “Fibrinogen Champagne au Mont d’Or” leads to duplication of 39 amino acids within a repetitive sequence of 13 amino acids in the connector portion of the aC domain23 a mutation predicted to cause an extension of the coiled coil. In FGB, while fibrinogen New York24 is described as a deletion of the amino acids encoded by FGB exon 2 the mutation is not characterized at the DNA level and as mentioned by the authors is most probably due to a splice-site mutation leading to exon 2 skipping rather than deletion of exon 2. It is likely that many more mutations of this sort have remained elusive even for laboratories specialized in mutation identification. We hypothesized that the same duplication was likely to be found in other afibrinogenemic patients and their family members from the same geographical region or ethnic group. We therefore screened 10 unrelated patients from Turkey by PCR amplification and identified one additional afibrinogenemic patient who was homozygous for the same duplication. His parents, his grandmother and his two sisters were all heterozygous (data not shown). Enquiries into the geographical origin of this family revealed that they were originally from Kumbet village, Ortaköy, Aksaray, i.e., the same region as the first family. While the nature of the mutation and the clear association with the phenotype i.e., complete fibrinogen deficiency with a homozygous genotype and partial fibrinogen deficiency with a heterozygous phenotype did not allow any reasonable doubt that we had identified the causative mutation we wished to identify the underlying molecular mechanism. Since the mutation duplicates the donor splice site of intron 8 we predicted that the mutation would impact FGG transcript splicing, of both

References 1. Vilar R, Fish RJ, Casini A, Neerman-Arbez M. Fibrin(ogen) in human disease: both friend and foe. Haematologica. 2020;105(2): 284-296. 2. Pieters M, Wolberg AS. Fibrinogen and fibrin: an illustrated review. Res Pract Thromb Haemost. 2019;3(2):161-172. 3. Casini A, Brungs T, Lavenu-Bombled C, Vilar R, Neerman-Arbez M, de Moerloose P. Genetics, diagnosis and clinical features of congenital hypodysfibrinogenemia: a sys-

1070

the major g transcript and the minor g’ transcript. Analysis of RNA produced by cells transiently transfected with normal or duplicated minigene constructs showed that in this model system the duplication causes production of several different aberrant transcripts of both isoforms leading to frameshifts and premature truncating codons. Regarding the clinical significance of our findings, genotyping patients with fibrinogen disorders is now recommended in guidelines from the International Society of Thrombosis and Hemostasis.25 In quantitative fibrinogen disorders, the identification of the causative mutation(s) can help to distinguish between afibrinogenemia and severe hypofibrinogenemia. Providing an accurate diagnosis for these patients is important since in specific clinical settings such as pregnancy or surgery, patient management could be different. In conclusion, we have identified a large duplication of several hundreds of basepairs at the FGG exon 8-intron 8 junction accounting for congenital afibrinogenemia in a large consanguineous Turkish family. The nature and size of the duplication can explain why this mutation was not identified using a standard PCR approach. It is highly likely that other patients with inherited quantitative fibrinogen disorders for whom no causative mutation has been identified harbor similar rearrangements. Disclosures MG, FM, FS-B, MM, NÖ, BM and MN-A have no conflicts of interest to disclose; FP declares that she is a speaker at educational meetings and a member of advisory boards for Roche, Sanofi, SOBI and Takeda; AC declares grants and fees paid to his institution from Takeda and travel support from SOBI. Contributions MG, FM, FSB and CDS performed genetic experiments and interpreted the results; NO, FP, MM, AC and BM collected patient samples and clinical information and performed fibrinogen measurements; MNA directed the study and wrote the first draft of the manuscript. All authors contributed to writing and editing the final manuscript. Acknowledgements The authors thank Dr. Cédric Howald and Dr. Keith Harshman at the Health 2030 Genome Centre at Campus Biotech, Geneva for whole exome sequencing and initial processing of the data and Dr. Nermin Keni and Dr. Ekrem Unal for providing patient samples from the second Turkish family. Funding This study was funded by a grant from the Swiss National Science Foundation (grant # 31003A_172864) to MNA.

tematic literature review and report of a novel mutation. J Thromb Haemost. 2017;15(5):876-888. 4. Neerman-Arbez M, de Moerloose P, Casini A. Laboratory and genetic investigation of mutations accounting for congenital fibrinogen disorders. Semin Thromb Hemost. 2016;42(4):356-365. 5. Menegatti M, Peyvandi F. Treatment of rare factor deficiencies other than hemophilia. Blood. 2019;133(5):415-424. 6. Casini A, Neerman-Arbez M, de Moerloose P. Heterogeneity of congenital afibrinogene-

mia, from epidemiology to clinical consequences and management. Blood Rev. 2020;26:100793. 7. Blomback M, Blomback B, Mammen EF, Prasad AS. Fibrinogen Detroit--a molecular defect in the N-terminal disulphide knot of human fibrinogen? Nature. 1968;218(5137): 134-137. 8. Neerman-Arbez M, Honsberger A, Antonarakis SE, Morris MA. Deletion of the fibrinogen [correction of fibrogen] alphachain gene (FGA) causes congenital afibrinogenemia. J Clin Invest. 1999;103(2):215-218.

haematologica | 2022; 107(5)

FGG mutation in congenital afibrinogenemia

9. Peyvandi F, Mannucci PM. Rare coagulation disorders. Thromb Haemost. 1999;82(4): 1207-1214. 10. Casini A, de Moerloose P, Neerman-Arbez M. Clinical features and management of congenital fibrinogen deficiencies. Semin Thromb Hemost. 2016;42(4):366-374. 11. Casini A, Neerman-Arbez M, Ariens RA, de Moerloose P. Dysfibrinogenemia: from molecular anomalies to clinical manifestations and management. J Thromb Haemost. 2015;13(6):909-919. 12. de Moerloose P, Casini A, Neerman-Arbez M. Congenital fibrinogen disorders: an update. Semin Thromb Hemost. 2013;39(6): 585-595. 13. Watanabe K, Shibuya A, Ishii E, et al. Identification of simultaneous mutation of fibrinogen alpha chain and protein C genes in a Japanese kindred. Br J Haematol. 2003;120(1):101-108. 14. Spena S, Duga S, Asselta R, et al. Congenital afibrinogenaemia caused by uniparental isodisomy of chromosome 4 containing a novel 15-kb deletion involving fibrinogen alpha-chain gene. Eur J Hum Genet. 2004;12(11):891-898.

haematologica | 2022; 107(5)

15. Monaldini L, Asselta R, Duga S, et al. Mutational screening of six afibrinogenemic patients: identification and characterization of four novel molecular defects. Thromb Haemost. 2007;97(4):546-551. 16. Neerman-Arbez M, de Moerloose P, Bridel C, et al. Mutations in the fibrinogen aalpha gene account for the majority of cases of congenital afibrinogenemia. Blood. 2000;96 (1):149-152. 17. PeyvandiI F, Palla R, MenegattiI M, et al. Coagulation factor activity and clinical bleeding severity in rare bleeding disorders: results from the European Network of Rare Bleeding Disorders. J Thromb Haemost. 2012;10(4):615-621. 18. Casini A, von Mackensen S, Santoro C, et al. Clinical phenotype, fibrinogen supplementation and health-related quality of life in patients with afibrinogenemia. Blood. 2021; 137(22):3127-3136. 19. Mosesson MW, Siebenlist KR, Meh DA. The structure and biological features of fibrinogen and fibrin. Ann N Y Acad Sci. 2001;936:11-30. 20. Standeven KF, Ariens RA, Grant PJ. The molecular physiology and pathology of fib-

rin structure/function. Blood Rev. 2005;19 (5):275-288. 21. de Maat MP, Verschuur M. Fibrinogen heterogeneity: inherited and noninherited. Curr Opin Hematol. 2005;12(5):377-383. 22. Casini A, Blondon M, Tintillier V, et al. Mutational epidemiology of congenital fibrinogen disorders. Thromb Haemost. 2018;118(11):1867-1874. 23. Hanss MM, Ffrench PO, Mornex JF, et al. Two novel fibrinogen variants found in patients with pulmonary embolism and their families. J Thromb Haemost. 2003;1(6): 1251-1257. 24. Liu CY, Koehn JA, Morgan FJ. Characterization of fibrinogen New York 1. A dysfunctional fibrinogen with a deletion of B beta(9-72) corresponding exactly to exon 2 of the gene. J Biol Chem. 1985;260 (7):4390-4396. 25. Casini A, Undas A, Palla R, Thachil J, de Moerloose P, Subcommittee on Factor X, et al. Diagnosis and classification of congenital fibrinogen disorders: communication from the SSC of the ISTH. J Thromb Haemost. 2018;16(9):1887-1890.

1071

ARTICLE Ferrata Storti Foundation

Hemostasis

Germline GATA2 variant disrupting endothelial eNOS function and angiogenesis can be restored by c-Jun/AP-1 upregulation Giulio Purgatorio,1 Elisa Piselli,1 Giuseppe Guglielmini,1 Emanuela Falcinelli,1 Loredana Bury,1 Valeria Di Battista,2 Fabrizia Pellanera,2 Francesca Milano,2 Caterina Matteucci,2 Cristina Mecucci2# and Paolo Gresele1# Section of Internal and Cardiovascular Medicine, Department of Medicine and Surgery, University of Perugia and 2Hematology and Bone Marrow Transplantation Unit, Department of Medicine and Surgery, University of Perugia, Perugia, Italy. 1

Haematologica 2022 Volume 107(5):1072-1085

PG and CM contributed equally as co-senior authors.

ABSTRACT

Correspondence: PAOLO GRESELE paolo.gresele@unipg.it Received: February 1, 2021. Accepted: June 16, 2021. Pre-published: July 8, 2021.

ATA2 is a transcription factor with key roles in hematopoiesis. Germline GATA2 gene variants have been associated with several inherited and acquired hematologic disorders, including myelodysplastic syndromes. Among the spectrum of GATA2 deficiency-associated manifestations thrombosis has been reported in 25% of patients, but the mechanisms are unknown. GATA2 was shown to be involved in endothelial nitric oxide synthase (eNOS) regulation and vascular development. We assessed eNOS expression and angiogenesis in patients with GATA2 deficiency. Platelets and blood outgrowth endothelial cells (BOEC) from GATA2 variant carriers showed impaired NO production and reduction of eNOS mRNA and protein expression and of eNOS activity. GATA2 binding to the eNOS gene was impaired in BOEC from GATA2-deficient patients, differently from control BOEC. GATA2 deficiency BOEC showed also defective angiogenesis, which was completely restored by treatment with the NO-donor Snitroso-N-acetylpenicillamine (SNAP). Atorvastatin, but not resveratrol, largely restored eNOS expression, NO biosynthesis and neoangiogenesis in GATA2-deficient BOEC by a mechanism involving increased expression of the eNOS transcription factor AP-1/c-JUN, replacing GATA2 when the latter is inactive. Our results unravel a possible thrombogenic mechanism of GATA2 mutations, definitely establish the regulation of eNOS by GATA2 in endothelial cells and show that endothelial angiogenesis is strictly dependent on the eNOS/NO axis. Given the ability of atorvastatin to restore NO production and angiogenesis by GATA2-deficient endothelial cells, the preventive effect of atorvastatin on thrombotic events and possibly on other clinical manifestations of the syndrome related to deranged angiogenesis should be explored in patients with GATA2 deficiency in an ad hoc designed clinical trial.

https://doi.org/10.3324/haematol.2021.278450

1072

Introduction GATA2 is a zinc finger transcription factor with key roles in the regulation of gene expression in hematopoietic cells.1 GATA2 binds consensus sequences in promoter/enhancer regions of target genes to regulate endothelial to hematopoietic transition in the embryo and to maintain the stem cell pool regulating hematopoietic stem cell (HSC) survival and self-renewal in the adult.2,3 Less than a decade ago heterozygous variants of GATA2 were first identified as the cause of four previously described hematologic syndromes, later recognized as different manifestations of a single genetic disorder.4-7 Heterozygous germline variants in GATA2 lead to what is now referred to as GATA2 deficiency, a mutable disorder with remarkable clinical heterogeneity involving hematopoiesis, immunity and the lymphatic system.2,8 A number of non-hematological and non-infectious complications have also

haematologica | 2022; 107(5)

Defective eNOS and angiogenesis in GATA2 R398W

been reported in patients with GATA2 deficiency, and among these thrombosis appears to be a rather frequent occurrence.2,8,9 Different types of thrombotic events have been reported, ranging from atherothrombotic or embolic stroke to recurrent venous thromboembolism (VTE), retinal vein thrombosis or catheter-related thrombosis, amounting to 25% of patients with GATA2 deficiency suffering thrombotic episodes, half of whom with multiple events.2,8,10,11 The reason of the high incidence of thrombosis in GATA2 deficiency is unknown, but it has been suggested to be multi-factorial since patients carrying GATA2 variants often have thrombotic risk factors, including infection, malignancy, bone marrow transplantation and central venous catheters.2 GATA2 is involved in vascular development12,13 and the knockdown of GATA2 in human endothelial cells led to vascular abnormalities,14 suggesting a role of endothelial rather than coagulation abnormalities in the pathogenesis of thrombosis associated with GATA2 deficiency. Interestingly, GATA2 was found to act as a promoter of the endothelial nitric oxide synthase gene (eNOS), the enzyme producing nitric oxide (NO), in bovine aortic endothelial cells and in the airway epithelium.15,16 NO plays a crucial role in the cardiovascular system acting as a powerful antithrombotic agent through its vaso-dilatatory, platelet inhibitory and anti atherosclerotic activities17-19 and recent data show that its deficiency is associated with enhanced risk of VTE too.20,21 However, no studies so far have explored NO production in patients with GATA2 deficiency. Here we show for the first time that platelets and endothelial cells from patients with GATA2 deficiency due to the R398W GATA2 variant exhibit impaired NO production and defective angiogenesis which can largely be corrected by pharmacologically-induced recovery of eNOS mRNA expression.

Chromatin immunoprecipitation Chromatin immunoprecipitation (ChiP) assay was performed using the SimpleChiP® enzymatic Chromatin IP Kit #9002 (Cell Signaling, Danver, MA, USA), according to the manufacturer’s instructions.24,25 Purified DNA was used as template for quantitative polymerase chain reaction (qPCR), using primers amplifying the eNOS promoter regions that are recognized by the AP-1 (forward: CTCAGCCCTAGTCTCTCTGC; revere: GGTTCTTGGGGATAGAGGCC) and GATA2 (forward: GGTGCCACATCACAGAAGGA; reverse: CACAATGGGACAGGAACAAGC) transcription factors, as previously described15. For details see the Online Supplemental Appendix.

Protein expression: western blotting Proteins were extracted from BOEC, quantified using the Bradford method, and Western blotting was carried out as described.26,27 For details see the Online Supplementary Appendix.

Immunofluorescence analysis of GATA2 GATA2 distribution was evaluated by immunofluorescence analysis with confocal microscopy, as previously described.28 For details see the Online Supplementary Appendix.

Nitric oxide production NO generation in BOEC and platelets was studied by flow cytometry using a specific fluorescent probe (4-amino-5-methylamino-2’,7’-Difluorofluorescein diacetate, DAF-FM diacetate, Invitrogen). BOEC were stimulated with acetylcholine 10 μM or acetylcholine 10 μM plus N5-(1-Iminoethyl)-L-ornithine dihydrochloride (L-NIO), a NOS inhibitor29 100 μM. Platelets were stimulated with type I collagen (Mascia Brunelli, Milan, Italy) at increasing concentrations (1-10 μg/mL) and NO-generated fluorescence was analyzed as previously described.30 For details see the Online Supplementary Appendix.

eNOS activity assay Methods Germline GATA2 mutation patients The proband (II2) was a 22 years old girl with mild anemia, reduction of monocytes, B and NK cells, recurrent otitis and papilloma virus infections, compatible with the MonoMac syndrome.4 Her mother (I1, 65 years old) and her sister (II1, 30 years old) did not show any clinical phenotype, and her father (I2, 66 years old) had a past clinical history of stroke and venous thrombosis. The father had suffered a first ischemic stroke at the age of 59, with no evidence of a cardioembolic source, and a recurrent ischemic stroke at the age of 66 while on aspirin. He later developed an unprovoked subclavian vein thrombosis and was then put under permanent oral anticoagulation with a direct oral anti-Xa drug. A thorough assessment for inherited or acquired thrombophilic conditions was negative. The study was approved by Comitato Universitario di Bioetica, University of Perugia (date of approval: 01/07/2019; approval number: 2019-21). The study was conducted according to the Declaration of Helsinki. Informed consent was obtained from the proband and all family members.

eNOS activity was measured by assessing the enzymatic conversion of (H3)L-arginine to (H3)L-citrulline. A standard curve with increasing (H3)L-arginine concentrations was built for each assay.31 For details see the Online Supplementary Appendix.

Treatment of blood outgrowth endothelial cells with eNOS inducers BOEC from healthy controls and from GATA2-deficient patients at passage 5 were seeded in 24-well plates (150×103 cells/well) in serum-free EBM2 medium. Cells were incubated with atorvastatin at a concentration of 50 μM,32 or with resveratrol 40 μM,33 or with their vehicle (dimethyl sulfoxide [DMSO]) for 24 hours (h) in serum-free medium. DMSO final concentration never exceeded 0.5%.

In vitro Matrigel angiogenesis assay Angiogenesis was estimated by measuring total tube length and by counting tubule number and branching points, as previously described.34 For details see the Online Supplementary Appendix.

Statistical analysis Blood outgrowth endothelial cells Blood outgrowth endothelial cells (BOEC) were isolated from peripheral blood of the proband and her family and from age- and sex-matched healthy controls and cultured as previously described.22,23 For details see the Online Supplementary Appendix.

haematologica | 2022; 107(5)

Data are presented as means ± standard error of the mean. The t-test for unpaired data was used to analyze results with a significant difference set at P<0.05. For details see the Online Supplementary Appendix. Further details on materials and methods are available in the Online Supplementary Appendix.

1073

G. Purgatorio et al.

Results Variant identification and blood outgrowth endothelial cells characterization The proband showed a heterozygous c.1192 C>T variant in GATA2 causing an arginine to tryptophan substitution (p.R398W). This is a known mutation affecting the zinc finger 2 (ZF2) domain of GATA2 and has been first reported as a recurrent missense variant in patients with the autosomal dominant monocytopenia and mycobacterial avium complex infection syndrome (MonoMAC).4 The same heterozygous variant was found in the father and sister, who were also both affected by mild monocytopenia and reduction of B and NK cells, although to a lesser extent compared to the proband (data not shown), but not in the mother (Figure 1A). BOEC derived from GATA2-mutated patients and healthy controls showed all the typical endothelial surface markers (CD31, CD146 and CD309) (Online Supplementary Figure S1A), and expressed von Willebrand

factor (Online Supplementary Figure S1AB). No difference was observed between healthy control and GATA2-mutated BOEC for viability, assessed by the fluorescein diacetate/propidium iodide (FDA/PI) assay (Online Supplementary Figure S2A) or by annexin-V/PI staining (Online Supplementary Figure S2B). Proliferation of BOEC derived from GATA2-mutated patients, assessed by bromodeoxyuridine (BrdU) incorporation, did not differ from that of healthy control BOEC (Online Supplementary Figure S3). Cells incubated with BrdU vehicle (EBM2 medium) were used as negative control.

GATA2 expression and subcellular distribution Real time PCR did not show differences in the expression of GATA2 mRNA between BOEC from healthy controls and those from the proband and familial carriers (Figure 1B). Moreover, there was no evidence of GATA2 protein decrease compared to controls (Figure 1C). Confocal microscopy analysis of BOEC from healthy controls and

Figure 1. Normal GATA2 mRNA and protein expression and impaired GATA2 binding to DNA in blood outgrowth endothelial cells from GATA2-deficient patients. (A) Pedigree showing the family with p.R398W GATA2 variant. Square denotes male and circle females. In gray the father (I2) and the sister (II1) asymptomatic carriers; the black circle indicates the proband (II2) symptomatic carrier. The white circle indicates the wild type mother (I1). (B) Real time polymerase chain reaction (PCR) of GATA2 mRNA expression in blood outgrowth endothelial cells (BOEC) from healthy controls, the unaffected family member (I1), and the GATA2-deficient family members. Expression of GATA2 mRNA is reported as fold change versus control BOEC normalized to a housekeeping mRNA (GAPDH). Values represent mean ± standard error of the mean (SEM) of 6 repeated measures from 6 controls and 3 different preparations from the GATA2-deficient family members (one-way ANOVA followed by Dunnett’s multiple comparison test). (C) Western blotting of GATA2 protein in BOEC from healthy controls, the unaffected family member (I1) and the GATA2-deficient family members. Actin was used as loading control. Optical densitometric analysis was performed using ImageJ software and results are expressed in arbitrary units. Values represent mean ± SEM of 8 repeated measures from 6 controls and 3 different preparations from the GATA2-deficient family members (one-way ANOVA followed by Dunnett’s multiple comparison test). (D) Chromatin immunoprecipitation (ChiP) quantitative PCR using primers amplifying the endothelial nitric oxide synthase gene (eNOS) promoter regions that are recognized by GATA2 performed using BOEC from healthy controls, the unaffected family member (I1) and the GATA2deficient family members. Aspecific binding of chromatin was excluded using immunoglobulin G (IgG)-immunoprecipitated chromatin as a negative control. Values represent mean ± SEM of 6 repeated measures from 6 controls and 3 different preparations from the GATA2 deficiency family members (*P<0.001 vs. controls, #P<0.001 vs. I1; one-way ANOVA followed by Dunnett’s multiple comparison test). Data are shown as fold change over IgG (n=3).

1074

haematologica | 2022; 107(5)

Defective eNOS and angiogenesis in GATA2 R398W

from GATA2-deficient patients showed the same subcellular distribution of GATA2, both in the nucleus and cytoplasm, as shown by green fluorescent protein reporter expression, in agreement with previous data14,35 (Online Supplementary Figure S4).

GATA2-binding to the eNOS gene and eNOS expression are impaired in GATA2 deficiency patients ChIP-qPCR of GATA2-bound chromatin showed that while GATA2 interacts directly with the human eNOS promoter in BOEC from healthy controls and from the unaffected family member (I1), this interaction is significantly reduced in all GATA2-mutated patients (Figure 1D). The eNOS (205 bp) signal was found in INPUT chromatin and in GATA2-immunoprecipitated chromatin from BOEC from healthy controls, while no eNOS DNA band was observed in GATA2-immunoprecipitated chromatin from BOEC from the GATA2-mutated patients (Online Supplementary Figure S5). In fact, patients with GATA2 deficiency showed a significant and striking reduction of eNOS mRNA and eNOS protein, compared to healthy controls (Figure 2A and B).

Silencing GATA2 reduces eNOS expression in control endothelial cells Transfection of healthy control BOEC with GATA2-targeted double-stranded small interfering RNA (siRNA) induced a significant reduction of GATA2 mRNA, attaining almost total suppression with the combined treatment with three different siRNA (Figure 2C). GATA2 protein expression was also significantly suppressed by GATA2-targeted siRNA, starting 36 h after transfection and maximally at 48 h (Figure 2D). Thirtysix h after transfection, when GATA2 protein was strikingly decreased, a significant reduction of eNOS mRNA was observed (Figure 2E), while complete suppression of eNOS protein expression was evident 48 h after GATA2-targeted siRNA transfection (Figure 2F). Treatment of healthy control BOEC with siRNA GATA2 or scramble siRNA (NC) did not affect cell viability (Online Supplementary Figure S2A and B).

Platelets and blood outgrowth endothelial cells from GATA2 deficiency patients show impaired nitric oxide production Platelets and BOEC from patients with GATA2 deficiency showed defective NO production, as assessed by the diaminofluorescein-2 diacetate (DAF) fluorescence assay, compared with healthy controls. In particular, while platelets from healthy controls stimulated with increasing concentrations of collagen showed a dose-dependent rise of DAF-fluorescence, platelets from the GATA2-mutanted subjects did not. Collagen-triggered DAF fluorescence in platelets from GATA2-mutanted patients was significantly lower than that in platelets from healthy controls, at all concentrations of collagen used (Figure 3A). Similarly, BOEC from healthy controls showed a striking increase of DAF-fluorescence upon stimulation with acetylcholine (10 μM), while BOEC from GATA2 deficiency subjects showed a significantly lower increase of DAF fluorescence. Acetylcholine-triggered DAF fluorescence in BOEC was completely abolished by preincubation with the eNOS inhibitor L-NIO (Figure 3B). The impairment of eNOS activity in BOEC from patients with GATA2 deficiency subjects was confirmed by the significantly reduced conversion of (H3)L-arginine to (H3)L-citrulline in haematologica | 2022; 107(5)

resting and acetylcholine-stimulated cells. Here too L-NIO suppressed H3-L-citrulline generation (Figure 3C). Moreover, the amount of the NO metabolites [NO -/NO ](NOx) released in the supernatant by acetylcholine-stimulated BOEC from GATA2-deficient subjects was significantly reduced compared to BOEC from healthy subjects (Figure 3D). Finally, levels of NOx in circulating blood from patients with GATA2 deficiency were also significantly reduced compared to healthy controls (Online Supplementary Figure S6). Acetylcholine-triggered NO release by control BOEC transfected with GATA2-targeted siRNA, as assessed by the DAF assay, was strikingly reduced compared with BOEC transfected with scramble siRNA (Figure 3E). 2

GATA2 and eNOS are required for angiogenesis BOEC from patients with GATA2 deficiency showed impaired angiogenesis in a tube formation assay compared with BOEC from healthy controls. Indeed tube length, branching points and the number of tubes were significantly reduced (Figure 4). In order to confirm the role of GATA2 and eNOS in angiogenesis, microvessel sprouting was evaluated in GATA2-silenced control BOEC. Tube formation was normal 36 h after GATA2 silencing, when GATA2 protein is suppressed but eNOS protein is not yet reduced (Online Supplementary Figure S7), while 48 h after transfection, when both GATA2 and eNOS were suppressed, angiogenesis was impaired with alterations similar to those found in BOEC from patients with GATA2 deficiency (Figure 4). In order to further clarify whether impaired eNOS expression, and not GATA2 suppression, was responsible of altered angiogenesis in patients with GATA2 deficiency, we treated their BOEC with the NO donor SNAP (100 μM) for 24 h and complete restoration of angiogenesis was observed. Similarly, treatment of GATA2-silenced control BOEC with the NO donor SNAP (100 μM) restored angiogenesis by significantly increasing tube number, branching points and tube lenght (Figure 4). Concordantly, treatment of control BOEC with the eNOS inhibitor L-NIO (100 μM) impaired tube formation (Online Supplementary Figure S8).

The eNOS inducer atorvastatin restores eNOS expression and nitric oxide production in blood outgrowth endothelial cells from GATA2-deficient patients Incubation with atorvastatin of BOEC from GATA2-deficient patients and from healthy controls for 24 h significantly increased eNOS mRNA expression compared with vehicle. On the contrary, treatment with resveratrol enhanced mRNA expression in healthy control BOEC but not in GATA2-mutated BOEC (Figure 5A). These data were confirmed by western blotting of eNOS protein, showing a significant increase of eNOS protein expression with atorvastatin in healthy control and GATA2-mutated BOEC but only in healthy control BOEC resveratrol (Online Supplementary Figure S9A and B). BOEC derived from the unaffected family member (I1) showed the same behavior of healthy control BOEC (Online Supplementary Figure S9C). Increased eNOS expression induced by preincubation with atorvastatin was associated with enhanced NO production in both control and GATA2-deficient BOEC, as assessed by the DAF assay and by the measurement of NOx in cell supernatant (Figure 5B). On the contrary, preincubation with resveratrol enhanced the release of NO from control 1075

G. Purgatorio et al.

Figure 2. Impaired GATA2 activity and expression alter eNOS mRNA and protein expression. (A) Real time polymerase chain reaction (PCR) of the endothelial nitric oxide synthase gene (eNOS) mRNA expression in blood outgrowth endothelial cells (BOEC) from healthy controls, the unaffected family member (I1) and the GATA2-deficient family members. eNOS mRNA expression is reported as fold change versus healthy control BOEC normalized to a housekeeping mRNA (GAPDH). Values represent mean ± standard error of the mean (SEM) of 8 repeated measures from 6 controls and 3 different preparations from the family members (**P<0.001 vs. controls; one-way ANOVA followed by Dunnett’s multiple comparison test). (B) Western blotting of eNOS protein in BOEC from healthy controls, the unaffected family member (I1) and the GATA2 deficiency family members. Actin was used as loading control. Optical densitometric analysis was performed using ImageJ software and results are expressed in arbitrary units. Values represent mean ± SEM of 6 repeated measures from 6 controls and 3 different preparations from the patients (one-way ANOVA followed by Dunnett’s multiple comparison test, *P<0.005, **P<0.001 vs. controls). (C) Real time PCR of GATA2-coding mRNA extracted from BOEC after incubation with three different small interfering RNA (siRNA) directed against GATA2 (25 nM) and their combinations. GATA2 expression was normalized to that of GAPDH. Values represent mean ± SEM of 4 repeated measures (*P<0.05 vs. controls, one-way ANOVA followed by Bonferroni’s multiple comparison test). (D) Western blotting of GATA2 protein extracted from BOEC after incubation with siRNA directed against GATA2 (25 nM) for 24-48 hours (h). HPRT was used as a loading control. Optical densitometric analysis was performed using ImageJ software and results are expressed in arbitrary units. Values represent mean ± SEM of 4 repeated measures (**P<0.001 vs. untreated BOEC, two-way ANOVA followed by Dunnett’s multiple comparison test). (E) Expression of eNOS mRNA after 24 h and 36 h of incubation of control BOEC with siRNA directed against GATA2 (25 nM) measured by real time PCR. The expression of eNOS is reported as fold change versus untreated BOEC and normalized to a housekeeping mRNA (GAPDH). Values are means ± SEM of 6 different transfection experiments (**P<0.005 vs. control BOEC, one way-ANOVA followed by Bonferroni’s multiple comparison test). (F) Western blotting of eNOS protein expression in BOEC treated with siRNA directed against GATA2 (25 nM) for 24, 36 and 48 h. HPRT was used as loading control. Optical densitometric analysis was performed using ImageJ software and results are expressed in arbitrary units. Values are mean ± SEM of 6 different transfection experiments (**P<0.001 vs. control BOEC, two-way ANOVA followed by Dunnett’s multiple comparison test).

1076

haematologica | 2022; 107(5)

Defective eNOS and angiogenesis in GATA2 R398W

Figure 3. Impaired nitric oxide production from platelets and blood outgrowth endothelial cells from GATA2-mutated patients. (A) 4-amino-5-methylamino-2',7'-difluorofluorescein (DAF-FM)–loaded platelets from healthy controls, the unaffected family member (I1) and the GATA2-deficient family members were stimulated with collagen (1-10 μg/mL) and DAF-FM fluorescence intensity was recorded. Values are means ± standard error of the mean (SEM) of 6 repeated measures (**P<0.001 vs. controls; two-way ANOVA followed by Tukey’s multiple comparison test). (B) DAF-FM-loaded blood outgrowth endothelial cells (BOEC) from healthy controls, the unaffected family member (I1) and the GATA2-deficient family members were stimulated with acetylcholine (ACh) 10 μM or ACh10 μM plus L-NIO 100 μM. Values are means ± SEM of 6 repeated measures (*P<0.0001 vs. baseline, #P<0.005 vs. control, +P<0.005 vs. ACh 10 μM; two-way ANOVA followed by Tukey’s multiple comparison test). (C) Endothelial nitric oxide synthase gene (eNOS) activity ((3H)L-Citrulline production) of acetylcholine- and ACh plus L-NIO-stimulated BOEC from healthy controls, the unaffected family member (I1) and the GATA2-deficient family members. Values are means ± SEM of 6 repeated measures (*P<0.001 vs. controls, # P<0.001 vs. unstimulated, two-way ANOVA followed by Tukey’s multiple comparison test). (D) Nitrite and nitrate (NOx) released in the supernatant from healthy controls, the unaffected family member (I1) and the GATA2-deficient family members under resting conditions and after stimulation with ACh 10 μM for 30 minutes. Concentration is expressed in μM /L/150x103 cells. Values are means ± SEM of 6 repeated measures (+P<0.001 vs. Baseline, *P<0.001 vs. controls; two-way ANOVA followed by Tukey’s multiple comparison test). (E) DAF-FM-loaded BOEC silenced with small interfering RNA (siRNA) directed against GATA2 or scramble siRNA were stimulated with acetylcholine 10 μM or acetylcholine 10 μM plus L-NIO 100 μM and fluorescence was recorded. Values are means ± SEM of 6 repeated measures (*P<0.001 vs. unstimulated, #P<0.001 vs. controls and scramble, two-way ANOVA followed by Tukey’s multiple comparison test).

haematologica | 2022; 107(5)

1077

Figure 4 Impaired angiogenesis in blood outgrowth endothelial cells (BOEC) from GATA2-mutant patients and in GATA2-silenced control blood outgrowth endothelial cells is restored by exogenous supplementation with nitric oxide. Impaired capillary tube formation on matrigel-coated wells by blood outgrowth endothelial cells (BOEC) from healthy controls, the unaffected family member (I1) and the GATA2-deficient family members and GATA2-silenced BOEC in comparison with healthy control BOEC and BOEC transfected with a scramble small interfering RNA (siRNA) (NC), respectively. Treatment with the nitric oxide (NO) donor SNAP (100 µM for 24 hours restored tube formation in BOEC from GATA2mutated patients and in GATA2-silenced BOEC. Angiogenesis was quantified with ImageJ software (AngioTool 64) by measuring total tube length, tube number and branching points. Values are means ± standard error of the mean (SEM) of 6 repeated measures (*P<0.0001 vs. control and #P<0.0001 vs. vehicle, two-way ANOVA followed by Tukey’s multiple comparison test). Specimens were analyzed at room temperature by a Carl Zeiss Axio Observer.A1 microscope (Carl Zeiss Inc, Oberkochen, Germany) using a 2.5X Plan-Apochromat objective and images acquired using the AxioVision software (Carl Zeiss Inc).

G. Purgatorio et al.

1078

haematologica | 2022; 107(5)

Defective eNOS and angiogenesis in GATA2 R398W

Figure 5. Atorvastatin, but not resveratrol, increases eNOS expression and nitric oxide production of blood outgrowth endothelial cells from GATA2-mutant patients. (A) Real time polymerase chain reaction (PCR) of endothelial nitric oxide synthase gene (eNOS)-coding mRNA of blood outgrowth endothelial cells (BOEC) from healthy controls, the unaffected family member (I1) and the GATA2-deficient family members after preincubation with atorvastatin (50 μM) or resveratrol (40 μM) for 24 hours. The expression of eNOS mRNA is reported as fold change versus vehicle and normalized to a housekeeping mRNA (GAPDH). Values are means ± standard error of the mean (SEM) of 4 repeated measures (#P<0.001 vs. vehicle,two-way ANOVA followed by Tukey’s multiple comparison test). (B) Nitrite and nitrate released (NOx) in the supernatant of BOEC from healthy controls, the unaffected family member (I1) and the GATA2-deficient family members after incubation with atorvastatin (50 μM) and resveratrol (40 μM) for 24 h. Values are means ± SEM of 4 repeated measures (+P<0.0001 vs. resting, *P<0.001 vs. control, #P<0.0001 vs. acetylcholine (ACh), two-way ANOVA followed by Tukey’s multiple comparison test). (C) Nitric oxide (NO) production by BOEC from healthy controls, the unaffected family member (I1) and the GATA2 deficiency family members. 4-amino-5-methylamino-2',7'-difluorofluorescein (DAF-FM)-loaded BOEC, after treatment with atorvastatin (50 μM) and resveratrol (40 μM) for 24 h, were stimulated with ACh 10 μM or ACh 10 μM plus L-NIO 100 μM and fluorescence was recorded. Values are means ± SEM of 4 repeated measures (*P<0.001 vs. control, #P<0.001 vs. untreated, two-way ANOVA followed by Tukey’s multiple comparison test).

haematologica | 2022; 107(5)

1079

G. Purgatorio et al.

Figure 6. Legend on following page.

1080

haematologica | 2022; 107(5)

Defective eNOS and angiogenesis in GATA2 R398W

Figure 6 Atorvastatin restores eNOS expression and endothelial tube formation via AP1/c-Jun signaling. (A) Restored capillary tube formation in blood outgrowth endothelial cells (BOEC) from healthy controls (CTRL), the unaffected family member (I1) and the GATA2-deficient family members on matrigel-coated wells in comparison with healthy controls BOEC, after treatment with atorvastatin (50 μM) for 24 hours. Angiogenesis was quantified with ImageJ (AngioTool64) software by measuring total tube length, tube number and branching points. Values are means ± standard error of the mean (SEM) of 6 repeated measures (*P<0.001 vs. controls, # P<0.001 vs. Vehicle, two-way ANOVA followed by Tukey’s multiple comparison test). Specimens were analyzed at room temperature by a Carl Zeiss Axio Observer.A1 microscope (Carl Zeiss Inc, Oberkochen, Germany) using a 2.5X Plan-Apochromat objective and images acquired using the AxioVision software (Carl Zeiss Inc). (B) Real time polymerase chain reaction (PCR) of c-Jun/AP-1 mRNA of BOEC from healthy controls, the unaffected family member (I1) and the GATA2-deficient family members after preincubation with atorvastatin (50 μM) or resveratrol (40 μM) for 24 h. The expression of c-Jun/AP-1 mRNA is reported as fold change versus healthy control BOEC and normalized to a housekeeping mRNA (GAPDH). Values are means ± SEM of 4 repeated measurements (*P<0.001 vs. vehicle and resv, two-way ANOVA followed by Tukey’s multiple comparison test). (C) Western blotting of c-Jun/AP-1 protein in BOEC from healthy controls and the GATA2-deficient family members after incubation with resveratrol (40 μM) for 24 h. β-actin was used as loading control. Optical densitometric analysis was performed using ImageJ software and results are expressed in arbitrary units. Values represent mean ± SEM of 6 repeated measures (two-way ANOVA followed by Tukey’s multiple comparison test). Results from the unaffected family member (I1) are shown in the Online Supplementary Figure S11. (D) Western blotting of c-Jun/AP-1 protein in BOEC from healthy controls and the GATA2-deficient family members after incubation with atorvastatin (50 μM) for 24 h. β-actin was used as loading control. Optical densitometric analysis was performed using ImageJ software and results are expressed in arbitrary units. Values represent mean ± SEM of 6 repeated measures (*P<0.001 vs. vehicles, twoway ANOVA followed by Tukey’s multiple comparison test). Results from the unaffected family member (I1) are shown in the Online Supplementary Figure S11. (E) Chromatin immunoprecipitation (ChiP) quantitative PCR (qPCR) using primers amplifying the endothelial nitric oxide synthase gene (eNOS) promoter regions that are recognized by AP-1. The figure shows the real time PCR of AP-1-bound chromatin of BOEC from healthy controls, the unaffected family member (I1) and the GATA2deficient family members in resting conditions and after treatment with the eNOS inducers atorvastatin and resveratrol. Values represent mean ± SEM of 6 repeated measures (*P 0.001 vs. resveratrol; two-way ANOVA followed by Dunnett’s multiple comparison test). Data are shown as fold change over immunoglobulin g (IgG). (F) ChiP qPCR using primers amplifying the eNOS promoter regions that are recognized by GATA-2. The figure shows the real time PCR of GATA2-bound chromatin of BOEC from healthy controls, the unaffected family member (I1) and the GATA2-deficient family members in resting conditions and after treatment with the eNOS inducers atorvastatin and resveratrol. Values represent mean ± SEM of 6 repeated measures (*P<0.001 vs. GATA2-mutated patients, #P<0.001 vs. Atorva; two-way ANOVA followed by Dunnett’s multiple comparison test). Data are shown as fold change over IgG.

BOEC but not from BOEC of patients with GATA2 deficiency (Figure 5B). These results were confirmed by measuring NO production in DAF-loaded BOEC (Figure 5C).

The eNOS inducer atorvastatin restores angiogenesis in blood outgrowth endothelial cells from GATA2-deficient patients by upregulating the expression of AP-1/c-Jun Increased expression of eNOS induced by atorvastatin was associated with enhanced angiogenesis in both BOEC from healthy controls and from patients with GATA2 deficiency as shown by significantly increased tube number, branching points and tube length, while resveratrol was ineffective (Figure 6A). Treatment with atorvastatin or resveratrol did not increase GATA2 mRNA and protein expression either in patient or in control BOEC (Online Supplementary Figure S10A to C), including BOEC from the unaffected family member (I1) (Online Supplementary Figure S10D). On the other hand, atorvastatin increased the expression of another eNOS transcription factor, c-Jun/AP-1. In fact after 24 hours of incubation with atorvastatin BOEC from both healthy controls and GATA2-mutanted patients showed a significant increase of c-Jun/AP-1 mRNA expression, an effect not observed with resveratrol (Figure 6B). Western blotting confirmed a significant increase of c-Jun/AP-1 protein expression in BOEC from healthy controls and GATA2-deficient patients treated with atorvastatin, but not with resveratrol (Figure 6C and D). BOEC derived from the unaffected family member (I1) showed the same behavior of healthy control BOEC (Online Supplementary Figure S11). In order to confirm that increased eNOS expression after treatment with atorvastatin was due to enhanced c-Jun/AP-1 binding to DNA, we carried out qPCR of c-Jun/AP-1-bound chromatin which was significantly increased in both BOEC from GATA2-deficient patients and from healthy controls treated with atorvastatin, but not with resveratrol (Figure 6E). GATA2-bound chromatin was significantly increased in healthy control BOEC and the unaffected family member, compared to BOEC of GATA2-mutated patients (Figure 6F) both under resting conditions and after treatment with the eNOS inducers atorvastatin and resveratrol. Interestingly, preincubation with resveratrol haematologica | 2022; 107(5)

increased GATA2-bound chromatin in healthy control BOEC significantly more than atorvastatin (Figure 6F). qPCR was confirmed by PCR followed by DNA electrophoresis (Online Supplementary Figure S12).

The eNOS inducer resveratrol enhances eNOS expression by upregulating Runx1 and its interaction with GATA2 Resveratrol, but not atorvastatin, enhanced Runx1 protein expression both in healthy control and GATA2-mutated BOEC (Figure 7A and B). However, despite increased Runx1 expression by resveratrol, GATA2-Runx1 binding was significantly lower in GATA2 deficiency patients compared to healthy controls (Figure 7C), including the unaffected family member (I1) (Online Supplementary Figure S13A and B).

Discussion Our study shows that i) GATA2 deficiency is associated with defective expression of eNOS by platelets and endothelial cells with impaired NO production, that ii) reduced eNOS expression in turn causes an altered angiogenic activity of endothelial cells, and that iii) treatment of endothelial cells from GATA2-deficient patients with the eNOS mRNA inducer atorvastatin restores eNOS expression, NO production and angiogenesis. NO is a mediator released by platelets and endothelial cells which plays an important antithrombotic role by preventing almost all aspects of platelet activation, by displaying anti atherosclerotic effects, by reducing leukocyte adhesion and activation and by regulating vascular tone.17-19,36 An impaired effectiveness of NO or its rapid inactivation due to gene variants affecting guanylylcyclase or glutathione peroxidase have been shown to be responsible of an increased tendency to ischemic stroke and myocardial infarction37,38 and acquired endothelial dysfunction and NO deficiency are associated with enhanced risk of cardiovascular disease and venous thromboembolism.20,21,36,39 We therefore hypothesize that the high incidence of thrombotic events among patients with GATA2 deficiency, so far largely unexplained,2 may depend to a great extent on the inability of their platelets and endothelial cells to produce NO. 1081

G. Purgatorio et al.

C Figure 7. Resveratrol restores eNOS expression and endothelial tube formation via Runx1 signaling and its interaction with GATA2. (A) Western blotting of RUNX1 protein in blood outgrowth endothelial cells (BOEC) from healthy controls and from GATA2-mutated patients after incubation with atorvastatin (50 μM) for 24 hours (h). β-actin was used as loading control. Optical densitometric analysis was performed using ImageJ software and results are expressed in arbitrary units. Values represent mean ± standard error of the mean (SEM) of 6 repeated measures (two-way ANOVA followed by Tukey’s multiple comparison test). Results from the unaffected family member (I1) are shown in the Online Supplementary Figure S13A. (B) Western blotting of RUNX1 protein in BOEC from healthy controls and from GATA2-mutated patients after stimulus with resveratrol (40 μM) for 24h. β-actin was used as loading control. Optical densitometric analysis was performed using ImageJ software and results are expressed in arbitrary units. Values represent mean ± SEM of 6 repeated measures (*P<0.001 vs. vehicle; two-way ANOVA followed by Tukey’s multiple comparison test). Results from the unaffected family member (I1) are shown in the Online Supplementary Figure 13A. (C) Co-immunoprecipitation (Co-IP) of RUNX1-GATA2 complex in BOEC from healthy controls and from GATA2-mutated patients after incubation with resveratrol (40 μM) for 24 h. GATA2 protein was immunoprecipitated and western blotting was evaluated on RUNX1 protein. RUNX1 total lysate was used as INPUT for quantification. Optical densitometric analysis was performed using ImageJ software and results are expressed in arbitrary units. In patients with GATA2 deficiency the ratio Co-IP/lysates is reduced due to the increased Runx1 expression in lysates in response to resveratrol. Values represent mean ± SEM of 6 repeated measures (*P<0.001 vs. vehicle, #P<0.001 vs. controls, two-way ANOVA followed by Tukey’s multiple comparison test). Results from the unaffected family member (I1) are shown in the Online Supplementary Figure S13B. eNOS: endothelial nitric oxide synthase gene.

Figure 8. Mechanism of action of atorvastatin and resveratrol on eNOS expression and neoangiogenesis in patients with GATA2 deficiency. In R398W variant-carrying blood outgrowth endothelial cells (BOEC) the binding of GATA2 to DNA is reduced. Treatment with atorvastatin upregulates c-Jun/AP-1 which, when GATA2 is inactive, restores endothelial nitric oxide synthase gene (eNOS) mRNA and protein expression and thus nitric oxide (NO) production and neoangiogenesis. Resveratrol acts increasing RUNX1 expression and its binding to GATA2, therefore in GATA2 deficiency BOEC impaired eNOS transcription factor activity is not overcome.

1082

haematologica | 2022; 107(5)

Defective eNOS and angiogenesis in GATA2 R398W

The role of GATA2 as a promoter of the eNOS gene in bovine aortic endothelial cells and in airway epithelial cells was previously reported,15,16 but so far no studies had analyzed the impact that GATA2 variants which are associated with hematopoietic impairment have on its transcription factor activity for eNOS and on eNOS expression and its function in human endothelial cells. Here we show that the R398W GATA2 variant, a germline mutation frequently found in patients with the GATA2 deficiency syndrome,40 impairs GATA2 binding to the eNOS gene in patientderived endothelial cells reducing the transcription of eNOS mRNA and consequently decreasing NO production. Concordantly, the silencing of GATA2 mRNA in BOEC from healthy controls, using a combination of siRNA targeting different sequences of the transcript, generated a NO production defect identical to that of GATA2-deficient patients. The role of GATA2 in vascular development was also previously reported using human endothelial cells,14 but no studies had explored angiogenesis in GATA2-deficient patients. Here we show that BOEC from patients with GATA2 deficiency display a striking impairment of endothelial tube formation and that this impairment is strictly dependent on NO insufficiency, in fact the supplementation of both GATA2-deficient BOEC and GATA2silenced control BOEC with exogenous NO restored a normal angiogenetic profile, while the treatment of control BOEC with the eNOS antagonist L-NIO generated an angiogenesis defect identical to that of GATA2-deficient BOEC. Several of the hematological and non-hematological manifestations of GATA2 deficiency have been associated with alterations in angiogenesis, like leukemia, solid organ tumors, lymphedema, venous thrombosis and stroke8,41,42 and GATA2-dependent pro- or anti-angiogenic microRNA regulation has been shown to have an important impact on endothelial biology and potentially on vascular disease.14 Therefore, our observation that defective eNOS provokes impaired angiogenesis in GATA2-deficient BOEC might be of relevance not only for the thrombotic but also for other clinical manifestations of the GATA2 deficiency syndrome. Given the striking impairment of NO production by GATA2-deficient BOEC, we explored whether some known inducers of eNOS expression or activity might restore NO production.19,31,43,44 We show that treatment with the hydroxymethylglutaryl-CoenzymeA inhibitor atorvastatin, a known enhancer of eNOS mRNA expression, but not with the natural polyphenol resveratrol, an agent preventing oxidation-triggered eNOS uncoupling, largely restored eNOS expression and NO biosynthesis in GATA2deficient BOEC. Interestingly, atorvastatin almost completely restored also the angiogenic activity of BOEC from GATA2-deficient patients. eNOS is an enzyme with a complex regulatory pattern, both at the transcriptional and post-transcriptional level, and sterol-regulatory cis-elements are present in the 5’ regulatory region of its gene suggesting that intracellular cholesterol levels are modulators of its expression.45,46 The cholesterol-lowering agent atorvastatin increases the expression of eNOS mRNA and protein by mechanisms not completely understood involving the blockade of Rho geranylgeranylation and the regulation of endoglin expression, and also enhances post-translationally eNOS activity by favoring its phosphorylation.19,47,50 It is thus likely that, despite the failure of the transcriptional function of GATA2, atorvastatin may restore eNOS mRNA transcription in BOEC from haematologica | 2022; 107(5)

patients with GATA2 deficiency by stimulating the activity of other transcription factors involved in the regulation of eNOS expression.47 Conversely resveratrol, a polyphenol with antioxidant effects and multiple beneficial activities on vascular function,43 enhanced eNOS expression in control BOEC but did not restore it in GATA2-deficient BOEC. Resveratrol upregulates eNOS expression acting through transcriptional and posttranscriptional (stabilization of mRNA) mechanisms.33 The discrepancy between the effects of atorvastatin and resveratrol in GATA2 deficiency BOEC may be due to their effect at different positions of the promoter sequence. Atorvastatin increases eNOS mRNA by enhancing the activity of transcription factors that bind the eNOS promoter in the 5’ region,32 including the transcription factor c-Jun/AP-115 which was in fact significantly increased by this drug in both healthy control and GATA2-mutated BOEC. Differently, resveratrol enhances the activity of transcription factors that bind the eNOS promoter in the proximal 263 bp region33 that involves the GATA2 binding site (254-279 bp).16 Interestingly, we found that resveratrol enhances eNOS expression in normal endothelial cells by increasing RUNX1. RUNX1 displays its transcription factor activity function by interacting with GATA251,52 and in the absence of RUNX1, GATA2 may rescue its transcriptional activity supporting hematopoiesis, which explains why the knockout of GATA2 reduces survival of RUNX1-/- zebrafish.53 The dependency of RUNX1 transcription factor activity on GATA2 to display its function explains why a ZNF2 GATA2 mutation, reducing GATA2 binding to DNA, prevents eNOS upregulation despite increased RUNX1 expression by resveratrol (Figure 8). Statins significantly reduce the risk of ischemic cardioand cerebro-vascular events and of venous thromboembolism, in part due to their ability to restore NO production,54,56 with little side effects and no hemorrhagic risk. Based on our results, the preventive effect of atorvastatin on thrombotic events in patients with GATA2 deficiency, and possibly its beneficial effects on other clinical manifestations of the syndrome related to deranged angiogenesis, should be explored in an ad hoc designed clinical trial. It should be considered that our results were obtained from studies of a single family with a specific in variant of the GATA2 gene and, thus, given the variant heterogeneity on GATA2 deficiencies, they might not be relevant for all cases of GATA2 deficiency. Our study reports the first observation, to our knowledge, of a germline gene variant-induced alteration of eNOS expression in humans and unravels the cause of altered angiogenesis in GATA2-deficient patients, suggesting that this may represent an important thrombogenic mechanism in these patients. We also identified a therapeutic option, through atorvastatin, able to restore eNOS expression and angiogenesis in GATA2 deficiency which deserves to be tested in vivo. Disclosures No conflicts of interest to disclose. Contributions GP, EP, GG, EF, LB, VB, FP, FM and CMa performed experiments, analyzed and interpreted data; CMe and PG designed and supervised the study; CM contributed the patient for the study; GP and PG wrote the manuscript; all authors critically revised the manuscript. 1083

G. Purgatorio et al.

Acknowledgments We thank M.P. Martelli for help with confocal microscopy studies.

http://www.progettomynerva.it/). EF and LB were supported by a fellowship from Fondazione Umberto Veronesi.

Fundings This study was supported by AIRC 5 × 1000, MYNERVA project, #21267 (Myeloid Neoplasms Research Venture Airc,

Data sharing statement All data generated and analyzed during this study are included in this published article and its additional file.

References 1. Bresnick EH, Katsumura KR, Lee HY, Johnson KD, Perkins AS. Master regulatory GATA transcription factors: mechanistic principles and emerging links to hematologic malignancies. Nucleic Acids Res. 2012;40(13):5819-5831. 2. McReynolds LJ, Calvo KR, Holland SM. Germline GATA2 mutation and bone marrow failure. Hematol Oncol Clin North Am. 2018;32(4):713-728. 3. Collin M, Dickinson R, Bigley V. Haematopoietic and immune defects associated with GATA2 mutation. Br J Haematol. 2015;169(2):173-187. 4. Hsu AP, Sampaio EP, Khan J, et al. Mutations in GATA2 are associated with the autosomal dominant and sporadic monocytopenia and mycobacterial infection (MonoMAC) syndrome. Blood. 2011;118(10):2653-2655. 5. Dickinson RE, Griffin H, Bigley V, et al. Exome sequencing identifies GATA-2 mutation as the cause of dendritic cell, monocyte, B and NK lymphoid deficiency. Blood. 2011;118(10):2656-2658. 6. Hahn CN, Chong CE, Carmichael CL, et al. Heritable GATA2 mutations associated with familial myelodysplastic syndrome and acute myeloid leukemia. Nat Genet. 2011;43(10):1012-1017. 7. Ostergaard P, Simpson MA, Connell FC, et al. Mutations in GATA2 cause primary lymphedema associated with a predisposition to acute myeloid leukemia (Emberger syndrome). Nat Genet. 2011;43(10):929931. 8. Spinner MA, Sanchez LA, Hsu AP, et al. GATA2 deficiency: a protean disorder of hematopoiesis, lymphatics, and immunity. Blood. 2014;123(6):809-821. 9. Crispino JD, Horwitz MS. GATA factor mutations in hematologic disease. Blood. 2017;129(15):2103-2110. 10. Zolton JR, Parikh TP, Hickstein DD, et al. Oocyte cryopreservation for women with GATA2 deficiency. J Assist Reprod Genet. 2018;35(7):1201-1207. 11. Berry D, Fekrat S. Central retinal vein occlusion in GATA2 deficiency. Retin Cases Brief Rep. 2019;13(2):181-184. 12. Johnson KD, Hsu AP, Ryu MJ, et al. Cis-element mutated in GATA2-dependent immunodeficiency governs hematopoiesis and vascular integrity. J Clin Invest. 2012;122(10):3692-3704. 13. Linneman AK, O’Geen H, Keles S, Farnham PJ, Bresnick EH. Genetic framework for GATA factor function in vascular biology. Proc Natl Acad Sci USA. 2011;108(33): 13641-13646. 14. Hartmann D, Fiedler J, Sonnenschein K, et al. MicroRNA-based therapy of GATA2deficient vascular disease. Circulation. 2016;134(24):1973-1990. 15. Zhang R, Min W, Sessa WC. Functional analysis of the human endothelial nitric oxide synthase promoter. Sp1 and GATA

1084

factors are necessary for basal transcription in endothelial cells. J Biol Chem. 1995;270(25):15320-15326. 16. German Z, Chambliss KL, Pace MC, Arnet UA, Lowenstein CJ, Shaul PW. Molecular basis of cell-specific endothelial nitric-oxide synthase expression in airway epithelium. J Biol Chem. 2000;275(11):8183-8189. 17. Reichenbach G, Momi S, Gresele P. Nitric oxide and its antithrombotic action in the cardiovascular system. Curr Drug Targets Cardiovasc Haematol Disord. 2005;5(1): 65-74. 18. Tziros C, Freedman JE. The many antithrombotic actions of nitric oxide. Curr Drug Targets. 2006;7(10):1243-1251. 19. Gresele P, Momi S, Guglielmini G. Nitric oxide-enhancing or –releasing agents as antithrombotic drugs. Biochem Pharmacol. 2019;166:300-312. 20. Migliacci R, Becattini C, Pesavento R, et al. Endothelial dysfunction in patients with spontaneous venous thromboembolism. Haematologica. 2007;92(6):812-818. 21. Momi S, Falcinelli E, Evangelista S, Gresele P. Antithrombotic and in vivo antiplatelet activity of Nebivolol are mediated by increased platelet-derived nitric oxide. Arterioscler Thromb Vasc Biol. 2014;34(4): 820-829. 22. Martin-Ramirez J, Hofman M, Van Den Biggelaar M, Hebbel RP, Voorberg J. Establishment of outgrowth endothelial cells from peripheral blood. Nat Protocols. 2012;7(9):1709-1715. 23. Cecchetti L, Tolley ND, Michetti N, Bury L, Weyrich AS, Gresele P. Megakaryocytes differentially sort mRNAs for matrix metalloproteinases and their inhibitors into platelets: a mechanism for regulating synthetic events. Blood. 2011;118(7):19031911. 24. Mikkelsen TS, Ku M, Jaffe DB, et al. Genome-wide maps of chromatin state in pluripotent and lineage-committed cells. Nature. 2007;448(7153):553-560. 25. He S, Wang F, Yang L, et al. Expression of DNMT1 and DNMT3a are regulated by GLI1 in human pancreatic cancer. PLoS One. 2011;6(11):e27684. 26. Sebastiano M, Momi S, Falcinelli E, Bury L, Hoylaerts MF, Gresele P. A novel mechanism regulating human platelet activation by MMP-2-mediated PAR1 biased signaling. Blood. 2017;129(7),883-895. 27. Amison RT, Momi S, Morris A, et al. RhoA signaling through platelet P2Y₁ receptor controls leukocyte recruitment in allergic mice. J Allergy Clin Immunol. 2015; 135(2):528-538. 28. Bury L, Malara A, Momi S, Petito E, Balduini A, Gresele P. Mechanisms of thrombocytopenia in platelet-type von Willebrand disease. Haematologica. 2019;104(7):1473-1481. 29. Balaguer S, Diaz L, Gomes A, et al. RealTime cytometric assay of nitric oxide and superoxide interaction in peripheral blood monocytes: a no-wash, no-lyse kinetic method. Cytometry Part B Clinical

Cytometry. 2015;92(3):211-217. 30. Cozzi MR, Guglielmini G, Battiston M, et al. Visualization of nitric oxide production by individual platelets during adhesion in flowing blood. Blood. 2015;125(4):697-705. 31. Gresele P, Pignatelli P, Guglielmini G, et al. Resveratrol, at concentrations attainable with moderate wine consumption, stimulates human platelet nitric oxide production. J Nutr. 2008;138(9):1602-1608. 32. Mashimo Y, Ishikawa T, Numakura M, Kinoshita M, Teramoto T. Critical promoter region for statin-induced human endothelial nitric oxide synthase (eNOS) transcription in EA.hy926 cells. J Atheroscler Thromb. 2013; 20(4):321-329. 33. Wallerath T, Deckert G, Ternes T, et al. Resveratrol, a polyphenolic phytoalexin present in red wine, enhances expression and activity of endothelial nitric oxide synthase. Circulation. 2002;106(13):1652-1658. 34. Khoo CP, Micklem K, Watt SM. A comparison of methods for quantifying angiogenesis in the matrigel assay in vitro. Tissue Eng Part C Methods. 2011;17(9):895-906. 35. Frye M, Taddei A, Dierkes C, et al. Matrix stiffness controls lymphatic vessel formation through regulation of a GATA2-dependent transcriptional program. Nat Commun. 2018;9(1):1511. 36. Gresele P, Momi S, Migliacci R. Endothelium, venous thromboembolism and ischaemic cardiovascular events. Thromb Haemost. 2010;103(1):56-61. 37. Freedman JE, Loscalzo J, Benoit SE, Valeri CR, Barnard MR, Michelson AD. Decreased platelet inhibition by nitric oxide in two brothers with a history of arterial thrombosis. J Clin Invest 1996;97(4);979-987. 38. Erdmann J, Stark K, Esslinger UB, et al. Dysfunctional nitric oxide signalling increases risk of myocardial infarction. Nature. 2013;504(7480)432-436. 39. Suzuki H, Matsuzawa Y, Konishi M, et al. Utility of noninvasive endothelial function test for prediction of deep vein thrombosis after total hip or knee arthroplasty. Circ J. 2014;78(7):1723-1732. 40. Chong CE, Venugopal P, Stokes PH, et al. Differential effects on gene transcription and hematopoietic differentiation correlate with GATA2 mutant disease phenotypes. Leukemia. 2018;32(1):194-202. 41. Rodríguez-Caso L, Reyes-Palomares A, Sánchez-Jiménez F, Quesada AR, Medina MÁ. What is known on angiogenesis-related rare diseases? A systematic review of literature. J Cell Mol Med. 2012;16(12):28722893. 42. Sun LL, Li WD, Lei FR, Li XQ. The regulatory role of microRNAs in angiogenesisrelated diseases. J Cell Mol Med. 2018;22(10):4568-4587. 43. Gresele P, Cerletti C, Guglielmini G, Pignatelli P, De Gaetano G, Violi F. Effects of resveratrol and other wine polyphenols on vascular function: an update. J Nutr Biochem. 2011;22(3):201-211. 44. Forstermann U, Li H. Therapeutic effect of

haematologica | 2022; 107(5)

Defective eNOS and angiogenesis in GATA2 R398W

enhancing endothelial nitric oxide syntase (eNOS) expression and preventing eNOS uncoupling. Br J Pharmacol. 2011;164(2): 213-223. 45. Marsden PA, Heng HH, Scherer SW, et al. Structure and chromosomal localization of the human constitutive endothelial nitric oxide synthase gene. J Biol Chem. 1993;268(23):17478-17488. 46. Robinson LJ, Weremowicz S, Morton CC, Michel T. Isolation and chromosomal localization of the human endothelial nitric oxide synthase (NOS3) gene. Genomics. 1994;19(2):350-357. 47. Garcia V, Sessa WC. Endothelial NOS: perspective and recent developments. Br J Pharmacol. 2019;176(2):189-196. 48. Zemankova L, Varejckova M, Dolezalova E, et al. Atorvastatin-induced endothelial nitric oxide synthase expression in

haematologica | 2022; 107(5)

endothelial cells is mediated by endoglin. J Physiol Pharmacol. 2015;66(3):403-413. 49. Laufs U, La Fata V, Plutzky J, Liao JK. Upregulation of endothelial nitric oxide synthase by HMG CoA reductase inhibitors. Circulation. 1998;97(12):11291135. 50. Momi S, Monopoli A, Alberti PF, et al. Nitric oxide enhances the anti-inflammatory and anti-atherogenic activity of atorvastatin in a mouse model of accelerated atherosclerosis. Cardiovasc Res. 2012;94(3): 428-438. 51. Kaur S, Rawal P, Siddiqui H, et al. Increased expression of RUNX1 in liver correlates with NASH activity score in patients with non-alcoholic steatohepatitis (NASH). Cells. 2019;8(10):1277. 52. Wilson NK, Foster SD, Wang X, et al. Combinatorial transcriptional control in

blood stem/progenitor cells: genome-wide analysis of ten major transcriptional regulators. Cell Stem Cell. 2010;7(4):532-544. 53. Bresciani E, Carrington B, Kwon Kim EM, et al. GATA2 is required for the runx1-independent hematopoiesis in Zebrafish. Blood 2019; 134 (Suppl 1):2463 54. Blum A, Shamburek R. The pleiotropic effects of statins on endothelial function, vascular inflammation, immunomodulation and thrombogenesis. Atherosclerosis. 2009;203(2);325-330. 55. Oesterle A, Laufs U, Liao JK. Pleiotropic effects of statins on the cardiovascular system. Circ Res. 2017;120(1):229-243. 56. Paciullo F, Momi S, Gresele P. PCSK9 in haemostasis and thrombosis: possible pleiotropic effects of PCSK9 inhibitors in cardiovascular prevention. Thromb Haemost. 2019;119(3):359-367.

1085

ARTICLE Ferrata Storti Foundation

Haematologica 2022 Volume 107(5):1086-1094

Hodgkin Lymphoma

Older patients (aged ≥60 years) with previously untreated advanced-stage classical Hodgkin lymphoma: a detailed analysis from the phase III ECHELON-1 study Andrew M. Evens,1 Joseph M. Connors,2 Anas Younes,3° Stephen M. Ansell,4 Won Seog Kim,5 John Radford,6 Tatyana Feldman,7 Joseph Tuscano,8 Kerry J. Savage,2 Yasuhiro Oki,9 Andrew Grigg,10 Christopher Pocock,11 Monika Dlugosz-Danecka,12 Keenan Fenton,13 Andres Forero-Torres,13 Rachael Liu,14 Hina Jolin,14 Ashish Gautam14 and Andrea Gallamini15 1

Division of Blood Disorders, Rutgers Cancer Institute of New Jersey, New Brunswick, NJ, USA; 2BC Cancer Center for Lymphoid Cancer and Department of Medical Oncology, Vancouver, BC, Canada; 3Memorial Sloan Kettering Cancer Center, New York, NY, USA; 4 Mayo Clinic, Rochester, NY, USA; 5Sungkyunkwan University School of Medicine, Samsung Medical Center, Seoul, South Korea; 6University of Manchester and the Christie NHS Foundation Trust Manchester Academic Health Science Center, Manchester, UK; 7John Theurer Cancer Center, Hackensack, NJ, USA; 8UC Davis Cancer Center, Sacramento, CA, USA; 9Genentech, South San Francisco, CA, USA; 10Olivia Newton-John Cancer Wellness and Research Center, Austin Health and Department of Clinical Haematology, Austin Hospital, Heidelberg, Australia; 11Haematology, East Kent Hospitals, Canterbury, UK; 12Maria Sklodowska-Curie National Research Institute of Oncology, Krakow, Poland; 13Seagen Inc., Bothell, WA, USA; 14Millennium Pharmaceuticals, Inc., Cambridge, MA, USA, a wholly owned subsidiary of Takeda Pharmaceutical Company Limited and 15Research and Innovation Department, A. Lacassagne Cancer Center, Nice, France °Current affiliation: Haematology (Early and Late Stage) Oncology R&D, AstraZeneca, New York, NY, USA

ABSTRACT

Correspondence: ANDREW M. EVENS ae378@cinj.rutgers.edu Received: February 1, 2021. Accepted: June 17, 2021. Pre-published: June 24, 2021. https://doi.org/10.3324/haematol.2021.278438

1086

ffective and tolerable treatments are needed for older patients with classical Hodgkin lymphoma. We report results for older patients with classical Hodgkin lymphoma treated in the large phase III ECHELON-1 study of frontline brentuximab vedotin plus doxorubicin, vinblastine, and dacarbazine (A+AVD) versus doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD). Modified progression-free survival per independent review facility for older versus younger patients (aged ≥60 vs. <60 years) was a pre-specified subgroup analysis; as the ECHELON-1 study was not powered for these analyses, reported P-values are descriptive. Of 1,334 enrolled patients, 186 (14%) were aged ≥60 years (A+AVD: n=84, ABVD: n=102); results below refer to this age group. Modified progression-free survival per independent review facility was similar in the two arms at 24 months (A+AVD: 70.3% [95% confidence interval (CI): 58.4–79.4], ABVD: 71.4% [95% CI: 60.5–79.8], hazard ratio (HR)=1.00 [95% CI: 0.58–1.72], P=0.993). After a median follow-up of 60.9 months, 5-year progression-free survival per investigator was 67.1% with A+AVD versus 61.6% with ABVD (HR=0.820 [95% CI: 0.494–1.362], P=0.443). Comparing A+AVD versus ABVD, grade 3/4 peripheral neuropathy occurred in 18% versus 3%; any-grade febrile neutropenia in 37% versus 17%; and any-grade pulmonary toxicity in 2% versus 13%, respectively, with three (3%) pulmonary toxicity-related deaths in patients receiving ABVD (none in those receiving A+AVD). Altogether, A+AVD showed overall similar efficacy to ABVD with survival rates in both arms comparing favorably to those of prior series in older patients with advanced-stage classical Hodgkin lymphoma. Compared to ABVD, A+AVD was associated with higher rates of neuropathy and neutropenia, but lower rates of pulmonary-related toxicity. Trials registered at ClinicalTrials.gov identifiers: NCT01712490; EudraCT number: 2011-005450-60.

haematologica | 2022; 107(5)

Analysis of older cHL patients in ECHELON-1

Introduction Older patients (aged ≥60 years) account for approximately 20-25% of cases of classical Hodgkin lymphoma (cHL) in population-based studies.1-3 While outcomes for younger patients with cHL have improved significantly in recent decades, similar progress has not been seen for older patients,4 in particular for those with advancedstage disease.2,3,5 This has been attributable to biological disease differences and co-morbidities associated with advanced age resulting in poor tolerance of chemotherapy and increased incidence of severe toxicities, including treatment-related deaths.4,6 Intensive regimens, such as bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, and prednisone (BEACOPP) are too toxic for older patients and may result in increased treatment-related mortality.7 In addition, bleomycin, a component of the doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD) regimen is associated with a significantly elevated risk of pulmonary toxicity in older patients,5,8-10 particularly in those aged ≥70 years.11-13 Brentuximab vedotin has been evaluated as an alternative treatment approach in older, less fit patients with previously untreated cHL, both as monotherapy14,15 and in combination regimens.16-18 Results from these earlyphase trials demonstrated tolerability and encouraging efficacy, with objective response rates of 98–100% and complete response rates of 44–87%.16-18 Sequential therapy in a phase II multicenter study with two cycles of brentuximab vedotin followed by six cycles of doxorubicin, vinblastine, and dacarbazine (AVD) yielded encouraging results.19 There has been a relative paucity of randomized phase III clinical trials in the frontline cHL setting that have included older patients in the contemporary era. In the primary analysis of the phase III ECHELON-1 study performed after a median follow-up of 24.6 months, frontline administration of brentuximab vedotin in combination with AVD (A+AVD) significantly improved the primary endpoint, modified progression-free survival (PFS) per independent review facility (IRF), compared with ABVD (hazard ratio [HR]=0.77 [95% confidence interval (CI): 0.60–0.98], P=0.035).20 Exploratory 3- and 5-year analyses reported continued provision of per-investigator PFS benefits for A+AVD compared with ABVD.21,22 Here we report the results of pre-specified analyses and post hoc analyses with extended follow-up of the efficacy and safety of A+AVD versus ABVD in 186 older cHL patients (aged ≥60 years).

reductions and modifications for brentuximab vedotin, including for the management of peripheral neuropathy, have been described previously.20 Patients were assessed for response to study treatment per IRF in accordance with the 2007 Revised Response Criteria for Malignant Lymphoma.23 Computed tomography scans were performed at screening, at the end of cycle 2, after administration of the last dose of frontline therapy, and during follow-up (every 3 months in the first year, and every 6 months thereafter). Positron emission tomography (PET) scans were performed at screening, the end of cycle 2, and the end of treatment. Adverse events were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 4.03. ECHELON-1 was conducted in accordance with regulatory requirements; the protocol was approved by the institutional review boards and ethics committees at each registered site. Written informed consent, in accordance with local ethics committee instructions, was mandatory before enrollment. This study was conducted according to the guideline of the International Conference on Harmonization Good Clinical Practice.

Endpoints and analyses The primary endpoint for ECHELON-1 was modified PFS per IRF, defined as the time to progression, death, or evidence of non-complete response per IRF (Deauville score ≥3) after completion of frontline therapy, followed by subsequent anticancer therapy (chemotherapy and/or radiotherapy). Overall survival (OS) was defined as the time from randomization to death from any cause and was the key secondary endpoint. Here we report a pre-specified subgroup analysis of modified PFS per IRF in older patients (defined as ≥60 years of age), as well as exploratory analyses, including PFS per investigator assessment (the time from randomization to relapse/progression or death) and safety. Subgroups of patients for efficacy and safety analyses were derived from the intention-to-treat (all randomized patients enrolled in ECHELON-1) and safety (all patients who received at least one dose of trial drug) populations, respectively. Following the primary analysis, the protocol did not require investigators to submit further information to the IRF, thus extended follow-up for analysis of modified PFS or PFS by IRF was not conducted. Modified PFS and PFS were summarized using the Kaplan-Meier methodology. ECHELON-1 was not powered for age-based subgroup analyses, so reported P-values are descriptive and without multiplicity adjustment.

Results Patients

Methods Study design and assessments The study design and population of patients for the openlabel, global, randomized, phase III ECHELON-1 study have been described previously.20 Briefly, patients aged ≥18 years (no upper age limit) with histologically confirmed, advanced (Ann Arbor stage III/IV) cHL who had received no prior systemic chemotherapy or radiotherapy were randomized 1:1 to receive A+AVD (brentuximab vedotin 1.2 mg/kg, doxorubicin 25 mg/m2, vinblastine 6 mg/m2, and dacarbazine 375 mg/m2) or ABVD (doxorubicin 25 mg/m2, bleomycin 10 units/m2, vinblastine 6 mg/m2, and dacarbazine 375 mg/m2) intravenously on days 1 and 15 of each 28-day cycle for up to six cycles. Dose

haematologica | 2022; 107(5)

As reported previously, 1,334 patients were included in the intention-to-treat population;20 of whom 186/1,334 (14%; A+AVD: n=84, ABVD: n=102) were aged ≥60 years (A+AVD arm: median age 68 years [range, 60–82], ABVD arm: median age 66 years [range, 60–83]) and were included in these sub-analyses. Patients’ demographics and disease characteristics were well balanced across the treatment arms in both older and younger patients. Within both arms, older patients tended to have a poorer Eastern Cooperative Oncology performance status than younger patients (Table 1).

Efficacy in older patients At the time of the primary analysis the median followup for older patients was 25 months (range, 24.2–25.8). 1087

A.M. Evens et al. Table 1. Baseline characteristics of the patients.

Median age, years (range) Male, n (%) White, n (%) Ann Arbor stage, n (%)* III IV ECOG PS score, n (%)† 0 1 2

Patients aged ≥60 years A+AVD ABVD Total (n=84) (n=102) (n=186)

Patients aged <60 years A+AVD ABVD Total (n=580) (n=568) (n=1,148)

ITT population (all ages)38 A+AVD ABVD Total (n=664) (n=670) (N=1,334)

68 (60–82) 55 (65) 76 (90)

66 (60–83) 64 (63) 82 (80)

67 (60–83) 119 (64) 158 (85)

33 (18–59) 323 (56) 484 (83)

33 (18–59) 334 (59) 472 (83)

33 (18–59) 657 (57) 956 (83)

35 (18–82) 378 (57) 560 (84)

37 (18–83) 398 (59) 554 (83)

36 (18–83) 776 (58) 1,114 (84)

31 (37) 51 (61)

34 (34) 67 (66)

65 (35) 118 (64)

206 (36) 374 (64)

212 (37) 354 (62)

418 (36) 728 (63)

237 (36) 425 (64)

246 (37) 421 (63)

483 (36) 846 (64)

30 (36) 44 (52) 10 (12)

36 (36) 55 (54) 10 (10)

66 (36) 99 (54) 20 (11)

346 (60) 216 (37) 18 (3)

342 (60) 208 (37) 17 (3)

688 (60) 424 (37) 35 (3)

376 (57) 260 (39) 28 (4)

378 (57) 263 (39) 27 (4)

754 (57) 523 (39) 55 (4)

A+AVD: brentuximab vedotin plus doxorubicin, vinblastine, and dacarbazine; ABVD: doxorubicin, bleomycin, vinblastine, and dacarbazine; ECOG PS: Eastern Cooperative Oncology Group performance status; ITT: intention-to-treat. *Ann Arbor stage at initial diagnosis was not applicable or missing for four patients; one patient had Ann Arbor stage II disease (major protocol violation). †ECOG PS score was not obtained or missing for two patients.

Modified PFS per IRF was similar in the two treatment arms at 24 months (A+AVD: 70.3% [95% CI: 58.4–79.4], ABVD: 71.4% [95% CI: 60.5–79.8], HR=1.00 [95% CI: 0.58–1.72], P=0.993) (Figure 1A, Table 2). At the end of randomized treatment, the complete response rate per IRF in older patients was 61% in both arms (difference [A+AVD - ABVD]: -0.1% [95% CI: -14.5–14.3]) (Online Supplementary Table S1). After a median follow-up of 60.9 months’ (95% CI: 60.6–61.7), 5-year PFS per investigator assessment for older cHL patients treated on ECHELON-1 was 67.1% (95% CI: 55.1–76.5) with A+AVD versus 61.6% (95% CI: 50.9–70.7) with ABVD (HR=0.820 [95% CI: 0.494–1.362], P=0.443) (Figure 1B; Table 2). Among younger patients, 5year PFS per investigator assessment was 84.3% (95% CI: 81.0–87.1) and 77.8% (95% CI: 74.0–81.1), respectively (HR=0.665 [95% CI: 0.51–0.88], P=0.003) (Table 2, Online Supplementary Figure S1). For older patients, the per investigator PFS was similar in both arms in patients with stage III disease (HR=1.051 [95% CI: 0.42–2.66], P=0.917) or stage IV disease (HR=0.722 [95% CI: 0.39–1.33], P=0.291) (Table 2). In exploratory analyses by interim PET scan status after two cycles (PET2), 5-year PFS per investigator assessment for older cHL patients in the A+AVD versus ABVD arm was 71.9% versus 64.9% in PET2-negative patients (HR=0.720 [95% CI: 0.40–1.29], P=0.268), and 40.0% versus 25.0% in PET2-positive patients (HR=0.923 [95% CI: 0.23–3.72], P=0.910); however, numbers of patients were low in the PET2-positive, aged ≥60 years subgroup in the A+AVD arm (n=5) and the ABVD arm (n=8) (Online Supplementary Table S2). For both older and younger cHL patients, PFS rates were higher in PET2-negative versus PET2-positive patients within each study arm (Online Supplementary Table S2). Per protocol, OS was assessed at the time of the primary analysis (median follow-up 28 months) and the final analysis will be performed once 112 events have occurred in the entire study. Among older patients, 15 patients in the A+AVD arm and 17 in the ABVD arm had died as of the April 20, 2017 data cut. Data on salvage therapy are not available. 1088

Safety A total of 181 older patients were evaluable for safety (A+AVD: n=83, ABVD: n=98). Older patients received a median of six cycles of treatment across both treatment arms. In the A+AVD arm, 80% of older patients required one or more dose modification of brentuximab vedotin: dose reduction, 31%; dose held, 5%; dose delayed, 61%; brentuximab vedotin discontinued, 20%. The mean relative dose intensity in older patients for brentuximab vedotin was 92%; relative dose intensities in the A+AVD versus ABVD arms for doxorubicin were 97% versus 97%; for vinblastine 93% versus 93%; and for dacarbazine 98% versus 96% (Online Supplementary Table S3). In the ABVD arm, 71% of older patients required one or more dose modification of bleomycin: dose reduction, 9%; dose held, 4%; dose interrupted, 1%; dose delayed, 49%; bleomycin discontinued, 28%. The mean relative dose intensity for bleomycin was 88.7% (Online Supplementary Table S3). Overall, the incidences of grade ≥3 treatment-emergent adverse events were higher in older patients than in younger patients (Table 3). Within both age groups, there was a higher incidence of any-grade pulmonary-related events in the ABVD arm than in the A+AVD arm. In older patients, a total of eight deaths occurred on-study (within 30 days of the last dose of frontline treatment), which yielded a treatment-related mortality rate of 4.4% (8/181; 3/83 [3.6%] in the A+AVD arm and 5/98 [5.1%] in the ABVD arm). Of these eight deaths, three occurred in the A+AVD arm (due to hemophagocytic lymphohistiocytosis, multiple organ dysfunction syndrome, and myocardial infarction [each, n=1]), none of which was associated with pulmonary toxicity (Online Supplementary Table S4). The remaining five deaths occurred in the ABVD arm (due to pneumonia [n=2], interstitial lung disease [n=1], respiratory disorder [n=1], and cardiac arrest [n=1]). Treatment-related pulmonary-related toxicity was associated with three of these five deaths in the ABVD arm, occurring in patients aged 78, 80, and 83 years, and could not be ruled out as having a causal relationship with the other two deaths. The incidence of grade ≥3 neutropenia was higher in the A+AVD arm than in the ABVD arm in older patients (70% vs. 59%). The incidence of any-grade febrile neutropenia haematologica | 2022; 107(5)

Analysis of older cHL patients in ECHELON-1

Figure 1. Progression-free survival in patients aged ≥60 years. (A) Modified progression-free survival (PFS) per independent review facility after a median follow-up of 25 months. (B) PFS per investigator after a median follow-up of 60.9 months. A+AVD: brentuximab vedotin plus doxorubicin, vinblastine, and dacarbazine; ABVD: doxorubicin, bleomycin, vinblastine, and dacarbazine; CI: confidence interval; HR: hazard ratio; INV: investigator; IRF: independent review facility.

was higher in the A+AVD arm than in the ABVD arm in both older patients (37% vs. 17%) and younger patients (17% vs. 6%) (Table 3). In the A+AVD arm, the use of granulocyte colony-stimulating factor (G-CSF) primary prophylaxis, given per institutional guidelines, was associated with a lower incidence of neutropenia (40% with vs. 78% without primary prophylaxis) and febrile neutropenia (30% with vs. 38% without primary prophylaxis) in older patients (Table 4). The incidence of any-grade peripheral neuropathy was higher in the A+AVD arm than in the ABVD arm in both older (65% vs. 43%) and younger patients (67% vs. 43%) (Table 5). Furthermore, the rate of severe grade 3/4 peripheral neuropathy was higher in older patients who received A+AVD than in haematologica | 2022; 107(5)

those who received ABVD (18% vs. 3%). Rates of resolution or improvement in peripheral neuropathy appeared similar in older cHL patients treated with A+AVD and ABVD (80% vs. 83%; respectively). In older patients, 24 and 12 patients had residual peripheral neuropathy, which was grade 1 (n=14 and n=6), grade 2 (n=7 and n=4), and grade 3 (n=3 and n=2) in severity in the A+AVD and ABVD arms, respectively.

Discussion Outcomes for older patients with cHL, particularly those with advanced disease, have historically been poor 1089

A.M. Evens et al.

Table 2. Summary of modified progression-free survival per independent review facility and per investigator.

Aged ≥60 years (n=186) A+AVD (n=84) 24-month modified PFS† per IRF, % (95% CI)20 24-month PFS‡ per INV, % (95% CI) 60-month PFS‡ per INV, % (95% CI)

ABVD (n=102)

Aged ≥60 years with stage III disease (n=65)* A+AVD ABVD (n=31) (n=34)

Aged ≥60 years with stage IV disease (n=118)* A+AVD ABVD (n=51) (n=67)

Aged <60 years (n=1,148) A+AVD (n=580)

ABVD (n=568)

ITT population (n=1,334) A+AVD (n=664)

ABVD (n=670)

70.3 71.4 67.7 80.9 71.3 66.1 83.7 78.2 82.1 77.2 (58.4-79.4) (60.5-79.8) (44.9-82.6) (66.2-90.9) (56.3-81.9) (51.8-77.1) (80.2-86.6) (74.4-81.6) (78.8-85.0) (73.7-80.4) 74.4 70.8 74.8 85.3 74.1 62.7 86.5 80.4 84.5 78.3 (62.2-82.7) (60.6-78.8) (54.2-87.1) (68.2-93.6) (59.6-84.1) (49.5-73.5) (83.4-89.1) (76.8-83.5) (81.4-87.1) (74.9-81.4) 67.1 61.6 70.1 69.9 65.1 57.0 84.3 77.8 80.7 73.1 (55.1-76.5) (50.9-70.7) (48.7-83.9) (51.3-82.6) (49.9-76.8) (43.5-68.5) (81.0-87.1) (74.0-81.1) (77.1-83.8) (69.0-76.7)

A+AVD: brentuximab vedotin plus doxorubicin, vinblastine, and dacarbazine; ABVD: doxorubicin, bleomycin, vinblastine, and dacarbazine; CI: confidence interval; INV: investigator; IRF: independent review facility; ITT: intention-to-treat; PFS: progression-free survival. *Three patients aged ≥60 years were excluded from analysis by disease stage due to missing data (n=2) or stage II disease (n=1). †2-year modified PFS per IRF based on the primary analysis. ‡2- and 5-year PFS per INV based on a median of 60.9 months’ extended follow-up in patients aged ≥60 years and 60.8 months in patients aged <60 years.

Table 3. Safety summary.

Grade ≥3 AE, n (%) On-study deaths,† n (%) Grade ≥3 neutropenia,‡ n (%) Any-grade FN on study, n (%) Any-grade pulmonary AE, n (%)

Patients aged ≥60 years evaluable for safety* (n=181) A+AVD ABVD (n=83) (n=98)

Patients aged <60 years evaluable for safety* (n=1,140) A+AVD ABVD (n=579) (n=561)

Safety population*,38 (n=1,321) A+AVD ABVD (n=662) (n=659)

73 (88) 3 (4) 58 (70) 31 (37) 2 (2)

476 (82) 6 (1) 372 (64) 97 (17) 10 (2)

549 (83) 9 (1) 430 (65) 128 (19) 12 (2)

78 (80) 5 (5) 58 (59) 17 (17) 13 (13)

356 (63) 8 (1) 259 (46) 35 (6) 31 (6)

434 (66) 13 (2) 317 (48) 52 (8) 44 (7)

A+AVD: brentuximab vedotin plus doxorubicin, vinblastine, and dacarbazine; ABVD: doxorubicin, bleomycin, vinblastine, and dacarbazine; AE: adverse events; FN: febrile neutropenia. *Received ≥1 dose of study therapy. †Within 30 days of the last dose of frontline treatment. ‡Neutropenia includes preferred terms of 'neutropenia' and 'neutrophil count decreased'.

compared with those of younger patients.2-4,6 We report here one of the largest prospective, randomized clinical trials in cHL completed in the contemporary era that have included and analyzed the outcomes of older patients. Among older patients ≥60 years treated in ECHELON-1, we report that modified PFS per IRF was statistically similar overall for patients treated with A+AVD or ABVD, being approximately 70% at 2 years in both arms. After a median follow-up of approximately 5 years, A+AVD demonstrated an apparent treatment benefit, although the numerical improvement in PFS over that with ABVD was not statistically significant. A+AVD was associated with more frequent neuropathy and febrile neutropenia, but less frequent pulmonary toxicity than ABVD. Additionally, older cHL patients had higher rates of febrile neutropenia and neuropathy compared with younger patients treated in ECHELON-1. In interpreting these observations, several factors should be considered. As older adults may often have multiple comorbidities that pose challenges to the use of traditional multi-agent treatment options, there is a need to identify tolerable and effective treatment regimens. This may reflect improvements in supportive care as well as patient selection. Several recent phase II studies have assessed the efficacy of multiple brentuximab vedotin-based regimens in the frontline cHL setting in older patients. In a phase II study, 1090

a sequential administration approach was assessed, in which patients with unfavorable stage II (IIB, IIAX, or IIBX) to stage IV disease received two lead-in doses of single-agent brentuximab vedotin (1.8 mg/kg once every 3 weeks), followed by six cycles of AVD. Patients who responded then received four consolidative doses of brentuximab vedotin.19 This regimen was well tolerated, with lower rates of grade ≥3 neutropenia (44%) and peripheral sensory neuropathy (4%) compared with those seen in older patients in the A+AVD arm in ECHELON-1, suggesting potentially better tolerability of sequential treatment.19 An objective response rate of 95% (complete responses: 93%) and 2-year PFS and OS rates of 84% and 93%, respectively, were also reported.19 Furthermore, survival rates varied based on patients’ fitness in this study with superior PFS and OS being observed among fit older cHL patients with lower Cumulative Illness Rating ScaleGeriatric co-morbidity scores and those without loss of instrumental activities of daily living, the latter of which persisted on multivariate analyses. Unfortunately, baseline or prospective geriatric assessments were not performed in ECHELON-1. Brentuximab vedotin has also been assessed as monotherapy and in combination with bendamustine, dacarbazine, or nivolumab.14,16,24,25 Importantly, patients enrolled in this study were ineligible for conventional haematologica | 2022; 107(5)

Analysis of older cHL patients in ECHELON-1

Table 4. Safety profile according to receipt of granulocyte colony-stimulating factor primary prophylaxis during days 1–5 of cycle 1.

G-CSF received† Any-grade neutropenia, n (%) FN in cycle 1, n (%) Any-grade FN on study, n (%) Infections and Infestations System Organ Class, n (%) Any SAE on study, n (%)

Patients aged ≥60 years evaluable for safety* (n=181) A+AVD (n=83) ABVD (n=98) Yes No Yes No (n=10) (n=73) (n=9) (n=89)

Patients aged <60 years evaluable for safety* (n=1,140) A+AVD (n=579) ABVD (n=561) Yes No Yes No (n=73) (n=506) (n=34) (n=527)

4 (40) 1 (10) 3 (30) 8 (80)

57 (78) 20 (27) 28 (38) 43 (59)

1 (11) 2 (22) 2 (22) 5 (56)

64 (72) 8 (9) 15 (17) 60 (67)

25 (34) 0 6 (8) 31 (42)

368 (73) 41 (8) 91 (18) 279 (55)

8 (24) 0 1 (3) 14 (41)

288 (55) 16 (3) 34 (6) 252 (48)

5 (50)

53 (73)

2 (22)

44 (49)

22 (30)

204 (40)

5 (15)

127 (24)

A+AVD: brentuximab vedotin plus doxorubicin, vinblastine, and dacarbazine; ABVD: doxorubicin, bleomycin, vinblastine, and dacarbazine; FN: febrile neutropenia; G-CSF: granulocyte colony-stimulating factor; SAE: serious adverse event. *Received ≥1 dose of study therapy. †G-CSF was given per institutional practice.

Table 5. Peripheral neuropathy: incidence and resolution.

Patients aged ≥60 years evaluable for safety* (n=181) A+AVD ABVD (n=83) (n=98) Any-grade PN, n/N (%) Grade 1 PN, n/N (%) Grade 2 PN, n/N (%) Grade 3/4 PN,† n/N (%) Patients with PN and complete resolution/improvement, n/N (%) PN complete resolution, n/N (%) PN improvement, n/N (%)

Patients aged <60 years evaluable for safety* (n=1,140) A+AVD ABVD (n=579) (n=561)

54/83 (65) 23/83 (28) 16/83 (19) 15/83 (18) 43/54 (80)

42/98 (43) 26/98 (27) 13/98 (13) 3/98 (3) 35/42 (83)

389/579 (67) 219/579 (38) 114/579 (20) 56/579 (9) 332/389 (85)

244/561 (43) 192/561 (34) 44/561 (8) 8/561 (1) 210/244 (86)

30/54 (56) 13/54 (24)

30/42 (71) 5/42 (12)

286/389 (74) 46/389 (12)

197/244 (81) 13/244 (5)

A+AVD: brentuximab vedotin plus doxorubicin, vinblastine, and dacarbazine; ABVD: doxorubicin, bleomycin, vinblastine, and dacarbazine; PN: peripheral neuropathy. *Received ≥1 dose of study therapy. †Among all patients evaluable for safety (n=1,321), only one case of grade 4 PN was reported, and this event occurred in a patient aged <60 years in the A+AVD arm.

frontline chemotherapy combinations (according to the investigator’s judgement). Initial assessment of brentuximab vedotin monotherapy demonstrated promising efficacy with 92% of patients (median age, 78 years) achieving an objective response (complete response rate: 73%).14 The combinations of brentuximab vedotin with dacarbazine, bendamustine, or nivolumab produced 100% objective response rates with each regimen (complete response rates: 62%, 88%, and 72%, respectively).16,26 Enrollment to the bendamustine combination was discontinued because of 65% of patients experiencing serious adverse events.16 Updated analyses with median followups of 59.4 and 58.6 months in the brentuximab vedotin monotherapy and dacarbazine and nivolumab combination therapy arms, respectively, showed median PFS of 10.5 and 46.8 months, and OS of 77.5 and 64.0 months, respectively. The median PFS and OS had not been reached in the nivolumab arm, with a median follow-up of 19.4 months.25 The authors concluded that brentuximab vedotin plus dacarbazine or nivolumab were reasonable combinations in this more unfit/frail population of patients. The nivolumab combination was associated with a higher rate of grade ≥3 treatment-related adverse events compared with the dacarbazine combination (60% vs. 37%), including peripheral neuropathy (35% vs. 26%), but a lower rate of serious treatment-related adverse events (5% vs. 11%) and treatment discontinuations due haematologica | 2022; 107(5)

to adverse events (30% vs. 42%).25 Another phase II study of brentuximab vedotin plus nivolumab in previously untreated older patients (≥60 years) suggested a lower objective response rate of 64%, including 52% with complete responses, at an interim analysis, which indicated that the combination was active in this population but did not meet predefined criteria that required a higher level of activity for further enrollment in the trial to proceed.27 The overall incidence of treatment-emergent adverse events in ECHELON-1 was comparable in the A+AVD and ABVD arms. A lower incidence of pulmonary-related toxicity was observed in the A+AVD arm than in the ABVD arm for both older and younger patients, with this difference being more marked in older patients. In older patients, three out of five on-study deaths in the ABVD arm were associated with pulmonary toxicity compared with none in the A+AVD arm (Online Supplementary Table S4), with 28% of patients in the ABVD treatment arm requiring bleomycin discontinuation. Since this study was initiated, it has been shown that pulmonary toxicity with ABVD can be reduced without reducing efficacy by omitting bleomycin from the regimen after two cycles in PET2-negative patients.28 The decision over whether to use this risk-adapted approach over A+AVD requires an assessment of the efficacy benefits, safety, and treatment costs for each individual patient. For older patients, the increased risk of toxicity, during the first two cycles of 1091

A.M. Evens et al.

treatment in those receiving PET2-adapted therapy and for PET2-positive patients (who continue on more intensive therapy), must be considered. However, treatment intensification is not recommended for older patients because of poor tolerance of BEACOPP.29 In the randomized HD9elderly study comparing baseline-BEACOPP regimen with cyclophosphamide, vincristine, procarbazine, prednisone + ABVD (COPP-ABVD), the treatment-related mortality rates among 75 patients with advanced-stage HL aged 66–75 years were 21% and 8%, respectively.7 A modified regimen incorporating brentuximab vedotin, dacarbazine, and dexamethasone (BrECADD) in place of bleomycin, vincristine, procarbazine, and prednisone (as used in BEACOPP) is being investigated in a phase III trial (HD21; NCT02661503)30 after a phase II study found that this regimen was associated with a relatively favorable toxicity profile while maintaining a complete response rate of 88%.31 Microtubule inhibitors, such as the vinca alkaloids (e.g., vinblastine and vincristine) and the monomethyl auristatin E component of brentuximab vedotin are associated with occurrence of peripheral neuropathy.32-34 The incidence of any-grade peripheral neuropathy in older patients was higher with A+AVD than with ABVD (65% vs. 43%), especially grade 3/4 peripheral neuropathy (18% vs. 3%). Severe peripheral neuropathy was also more frequently seen in older than younger cHL patients treated with A+AVD. In the A+AVD arm, approximately fourfifths of older patients with peripheral neuropathy experienced improvement or resolution, a rate similar to that observed in the ABVD arm. With longer follow-up, residual peripheral neuropathy continues to improve and resolve.21,35 These findings highlight the importance of appropriate screening, monitoring, and active clinical management of peripheral neuropathy in patients treated with A+AVD (including potential dose reductions particularly in older patients who frequently present with multiple comorbidities). In the current analyses, the rates of neutropenia and febrile neutropenia were higher in the A+AVD arm overall and, moreover, higher in older than younger patients in both the A+AVD and ABVD arms. Although the use of primary prophylaxis with G-CSF was not mandated in ECHELON-1 and the cohort of older patients who received G-CSF primary prophylaxis was small (n=10), post-protocol amendment use of G-CSF primary prophylaxis was associated with reduced rates of neutropenia and febrile neutropenia in patients treated with A+AVD. Similar effects of primary prophylaxis with G-CSF on rates of neutropenia and febrile neutropenia were observed in patients treated with A+AVD in the overall ECHELON-1 study population.36 Consequently, G-CSF primary prophylaxis is recommended for all patients who receive A+AVD.37 As the optimal dosing schedule has not been established, G-CSF should be administered with each cycle, starting at cycle 1, as recommended in the US prescribing information and EU Summary of Product Characteristics. Taken together, these data showed overall similar efficacy for A+AVD and ABVD in older patients with stage III/IV cHL. A+AVD was associated with increased neuropathy and neutropenia but with less pulmonary-related toxicity compared with ABVD. Thus, A+AVD represents a treatment option (with primary prophylaxis with G-CSF) for selected fit, older patients with cHL overall, 1092

and especially for patients in whom pulmonary toxicity is a concern. Moreover, outcomes reported here set a new benchmark for older patients with untreated cHL when treated with A+AVD or ABVD. However, continued study of new therapeutic regimens is needed to improve outcomes and to decrease toxicity for older cHL patients. This includes continued examination of PET responseadapted strategies, which may be prognostic in brentuximab vedotin-based treatment for older cHL patients,19,21 as well as analysis of timing of brentuximab vedotin relative to chemotherapy (i.e., sequential vs. concurrent), integration of other targeted therapeutic agents (e.g., NCT03907488), and via the incorporation of objective geriatric assessments for prediction of tolerable and individualized therapy. Disclosures AME reports consultancy with honoraria for MorphoSys, Miltenyi, Seagen, and Epizyme; membership on an advisory committee with honoraria for Pharmacyclics and Novartis, Inc.; and research funding from Tesaro. JMC reports research funding from Merck, Amgen, Roche Canada, NanoString Technologies, Seagen, Janssen, F. Hoffmann-La Roche, Bayer Healthcare, Cephalon, Bristol Myers-Squibb, Lilly, Genentech, and Millennium Pharmaceuticals Inc., a wholly owned subsidiary of Takeda Pharmaceutical Company Limited; being a named inventor on a patent licensed to NanoString Technologies; and honoraria from Seagen. AY reports research funding from J&J, Bristol Myers-Squibb, Curis, Genentech, Pharmacyclics, Janssen, Novartis, Roche, Abbvie, and Astra Zeneca; and honoraria from Celgene, Bristol Myers-Squibb, Sanofi, Abbvie, Merck, Bayer, Incyte, Seagen, Roche, and Takeda. SMA reports research funding from LAM Therapeutics, Regeneron, Pfizer, Bristol-Myers Squibb, Merck & Co, Trillium, Seagen, Celldex, Takeda, and Affimed. JR reports research funding from Celgene, ADC Therapeutics, Pfizer, and Millennium Pharmaceuticals Inc., a wholly owned subsidiary of Takeda Pharmaceutical Company Limited; consultancy for ADC Therapeutics, Novartis, Takeda, Seagen, and Bristol Myers-Squibb; equity ownership in GlaxoSmithKline and AstraZeneca; and speakers bureau for Novartis, Takeda, Seagen, and Bristol Myers-Squibb. TF reports research funding from Seagen and Portola; and speakers bureau for Seagen, Janssen, Pharmacyclics, J&J, Celgene, and KITE. JT reports research funding from Millennium Pharmaceuticals Inc., a wholly owned subsidiary of Takeda Pharmaceutical Company Limited, Genentech, Celgene, and Pharmacyclics; speakers bureau for Amgen, Seagen, and Celgene; and honoraria from Amgen, Seagen, and Celgene. KJS reports honoraria from and consulting for Seagen, Merck, Bristol Myers-Squibb, Abbvie, Astra Zeneca, Verastem, and Gilead Consulting Servier; and research funding from Seagen, Millennium Pharmaceuticals Inc., a wholly owned subsidiary of Takeda Pharmaceutical Company, Merck, Bristol MyersSquibb, Beigene, and Roche Canada. YO reports employment with Genentech and research funding from Seagen and Millennium Pharmaceuticals Inc., a wholly owned subsidiary of Takeda Pharmaceutical Company Limited; honoraria from Takeda Millennium; and employment with Jazz Pharmaceuticals. AGri reports membership on an entity's Board of Directors or advisory committees for Roche, Gilead, Bristol Myers-Squibb, and Takeda. CP reports employment with Kent & Canterbury Hospital. MD-D reports consultancy for Servier and Roche. KF reports employment with and equity ownership in Seagen, Inc. AF-T reports employment with Seagen, Inc. RL, HJ, and AGau report employment with Millennium Pharmaceuticals, haematologica | 2022; 107(5)

Analysis of older cHL patients in ECHELON-1

Inc., a wholly owned subsidiary of Takeda Pharmaceutical Company Limited. AGal reports consultancy and speakers bureau for Takeda. W-SK has declared no conflicts of interest. Contributions Data were verified by the sponsor, analyzed by sponsor statisticians, and interpreted by academic authors and sponsor representatives. The manuscript was prepared by the authors with the assistance of a medical writer funded by the sponsor. Data were collected, and study procedures were overseen by AME, JMC, AY, SMA, WSK, JR, TF, JT, KJS, YO, AG, CP, MD-D, and AGal; data were verified by KF, AF-T, RL, HT, and AGal, analyzed by RL, and interpreted by all authors. All authors had full access to the data during development of the manuscript. All authors vouch for completeness and accuracy of the data and had final responsibility for the manuscript content and decision to submit. Acknowledgments The authors would like to thank the patients who participated in this study and their families. They would also like to acknowledge other investigators and staff at all ECHELON-1 clinical sites and the members of the Independent Data Monitoring Committee and Independent Review Committee. The authors

References 1. Thyss A, Saada E, Gastaud L, Peyrade F, Re D. Hodgkin's lymphoma in older patients: an orphan disease? Mediterr J Hematol Infect Dis. 2014;6(1):e2014050. 2. Proctor SJ, Wilkinson J, Jones G, et al. Evaluation of treatment outcome in 175 patients with Hodgkin lymphoma aged 60 years or over: the SHIELD study. Blood. 2012;119(25):6005-6015. 3. Stark GL, Wood KM, Jack F, et al. Hodgkin's disease in the elderly: a population-based study. Br J Haematol. 2002;119(2):432-440. 4. Engert A, Ballova V, Haverkamp H, et al. Hodgkin's lymphoma in elderly patients: a comprehensive retrospective analysis from the German Hodgkin's Study Group. J Clin Oncol. 2005;23(22):5052-5060. 5. Evens AM, Hong F, Gordon LI, et al. The efficacy and tolerability of adriamycin, bleomycin, vinblastine, dacarbazine and Stanford V in older Hodgkin lymphoma patients: a comprehensive analysis from the North American intergroup trial E2496. Br J Haematol. 2013;161(1):76-86. 6. Evens AM, Hong F. How can outcomes be improved for older patients with Hodgkin lymphoma? J Clin Oncol. 2013;31(12):15021505. 7. Ballova V, Ruffer JU, Haverkamp H, et al. A prospectively randomized trial carried out by the German Hodgkin Study Group (GHSG) for elderly patients with advanced Hodgkin's disease comparing BEACOPP baseline and COPP-ABVD (study HD9elderly). Ann Oncol. 2005;16(1):124-131. 8. Evens AM, Helenowski I, Ramsdale E, et al. A retrospective multicenter analysis of elderly Hodgkin lymphoma: outcomes and prognostic factors in the modern era. Blood. 2012;119(3):692-695. 9. Boll B, Goergen H, Behringer K, et al. Bleomycin in older early-stage favorable Hodgkin lymphoma patients: analysis of the German Hodgkin Study Group (GHSG) HD10 and HD13 trials. Blood. 2016; 127(18):2189-2192.

haematologica | 2022; 107(5)

acknowledge the writing assistance of Laura Webb and Hedley Coppock of Ashfield MedComms, an Ashfield Health company, part of UDG Healthcare plc, during the development of this manuscript, which was funded by Millennium Pharmaceuticals, Inc., and complied with the Good Publication Practice 3 ethical guidelines.39 Funding This work was supported by Millennium Pharmaceuticals, Inc., Cambridge, MA, USA, a wholly owned subsidiary of Takeda Pharmaceutical Company Limited (grant number not applicable); and Seagen, Inc., Bothell, WA, USA (grant number not applicable). Data-sharing statement The datasets, including the redacted study protocol, redacted statistical analysis plan, and individual participant’s data supporting the results reported in this article, will be made available within 3 months from initial request, to researchers who provide a methodologically sound proposal. The data will be provided after de-identification, in compliance with applicable privacy laws, data protection, and requirements for consent and anonymization.

10. Stamatoullas A, Brice P, Bouabdallah R, et al. Outcome of patients older than 60 years with classical Hodgkin lymphoma treated with front line ABVD chemotherapy: frequent pulmonary events suggest limiting the use of bleomycin in the elderly. Br J Haematol. 2015;170(2):179-184. 11. Sleijfer S. Bleomycin-induced pneumonitis. Chest. 2001;120(2):617-624. 12. Martin WG, Ristow KM, Habermann TM, et al. Bleomycin pulmonary toxicity has a negative impact on the outcome of patients with Hodgkin's lymphoma. J Clin Oncol. 2005;23(30):7614-7620. 13. Thomas TS, Luo S, Reagan PM, et al. Advancing age and the risk of bleomycin pulmonary toxicity in a largely older cohort of patients with newly diagnosed Hodgkin lymphoma. J Geriatr Oncol. 2020;11(1):69-74. 14. Forero-Torres A, Holkova B, Goldschmidt J, et al. Phase 2 study of frontline brentuximab vedotin monotherapy in Hodgkin lymphoma patients aged 60 years and older. Blood. 2015;126(26):2798-2804. 15. Gibb A, Pirrie SJ, Linton K, et al. Results of a UK National Cancer Research Institute phase II study of brentuximab vedotin using a response-adapted design in the first-line treatment of patients with classical Hodgkin lymphoma unsuitable for chemotherapy due to age, frailty or comorbidity (BREVITY). Br J Haematol. 2020;193(1):63-71. 16. Friedberg JW, Forero-Torres A, Bordoni RE, et al. Frontline brentuximab vedotin in combination with dacarbazine or bendamustine in patients aged ≥60 years with HL. Blood. 2017;130(26):2829-2837. 17. Gallamini A, Bijou F, Viotti J, et al. Brentuximab-vedotin and bendamustine is a feasible and effective drug combination as first-line treatment of Hodgkin lymphoma in the elderly (HALO trial). Hematol Oncol. 2017;35(Suppl 2):170. 18. Fosså A, Böll B, Goergen H, et al. T021 (0147) B-CAP (brentuximab vedotin, cyclophosphamide, doxorubicin and predniso(lo)Ne) in older patients with advancedstage Hodgkin lymphoma: results of a phase

II intergroup trial by the German Hodgkin Study Group (GHSG) and the Nordic Lymphoma Group (NLG). HemaSphere. 2018;2:27. 19. Evens AM, Advani RH, Helenowski IB, et al. Multicenter phase II study of sequential brentuximab vedotin and doxorubicin, vinblastine, and dacarbazine chemotherapy for older patients with untreated classical Hodgkin lymphoma. J Clin Oncol. 2018; 36(30):3015-3022. 20. Connors JM, Jurczak W, Straus DJ, et al. Brentuximab vedotin with chemotherapy for stage III or IV Hodgkin's lymphoma. N Engl J Med. 2018;378(4):331-344. 21. Straus DJ, Diugosz-Danecka M, Alekseev S, et al. Brentuximab vedotin with chemotherapy for stage III/IV classical Hodgkin lymphoma: 3-year update of the ECHELON-1 study. Blood. 2020;135(10):735-742. 22. Straus DJ, Długosz-Danecka M, Connors JM, et al. Brentuximab vedotin with chemotherapy for stage III or IV classical Hodgkin lymphoma (ECHELON-1): 5-year update of an international, open-label, randomised, phase 3 trial. Lancet Haematol. 2021;8(6):e410-e421. 23. Cheson BD, Pfistner B, Juweid ME, et al. Revised response criteria for malignant lymphoma. J Clin Oncol. 2007;25(5):579-586. 24. Friedberg JW, Forero-Torres A, Holkova B, et al. Long-term follow-up of brentuximab vedotin ± dacarbazine as first line therapy in elderly patients with Hodgkin lymphoma. J Clin Oncol. 2018;36(Suppl 15):7542. 25. Yasenchak CA, Bordoni R, Patel-Donnelly D, et al. Frontline brentuximab vedotin as monotherapy or in combination for older Hodgkin lymphoma patients. Blood. 2020; 136(Suppl 1):18-19. 26. Yasenchak CA, Bordoni R, Yazbeck V, et al. Phase 2 study of frontline brentuximab vedotin plus nivolumab in patients with Hodgkin lymphoma aged ≥60 years. Blood. 2019;134(Suppl 1):237. 27. Cheson BD, Bartlett NL, LaPlant B, et al. Brentuximab vedotin plus nivolumab as first-line therapy in older or chemotherapy-

1093

A.M. Evens et al. ineligible patients with Hodgkin lymphoma (ACCRU): a multicentre, single-arm, phase 2 trial. Lancet Haematol. 2020;7(11):e808e815. 28. Johnson P, Federico M, Kirkwood A, et al. Adapted treatment guided by interim PETCT scan in advanced Hodgkin's lymphoma. N Engl J Med. 2016;374(25):2419-2429. 29. Spinner MA, Advani RH. Risk-adapted therapy for advanced-stage Hodgkin lymphoma. Hematology Am Soc Hematol Educ Program. 2018;2018(1):200-206. 30. ClinicalTrials.gov. HD21 for Advanced Stages. NCT02661503. [last accessed May 18, 2021]; Available from: https:// clinicaltrials.gov/ct2/show/NCT02661503?term=HD 21&cond=Hodgkin+Lymphoma&draw=2& rank=1. 31. Eichenauer DA, Plütschow A, Kreissl S, et al. Incorporation of brentuximab vedotin into first-line treatment of advanced classical

1094

Hodgkin's lymphoma: final analysis of a phase 2 randomised trial by the German Hodgkin Study Group. Lancet Oncol. 2017; 18(12):1680-1687. 32. van de Donk NW, Dhimolea E. Brentuximab vedotin. MAbs. 2012;4(4):458-465. 33. Donaghy H. Effects of antibody, drug and linker on the preclinical and clinical toxicities of antibody-drug conjugates. MAbs. 2016; 8(4):659-671. 34. Addington J, Freimer M. Chemotherapyinduced peripheral neuropathy: an update on the current understanding. F1000Res. 2016;5:F1000 Faculty Rev-1466. 35. Radford J, Connors JM, Younes A, et al. Resolution of peripheral neuropathy (PN) in patients who received A+AVD or ABVD in the phase 3 ECHELON-1 trial. Blood. 2018; 132(Suppl 1):2921. 36. Straus D, Collins G, Walewski J, et al. Primary prophylaxis with G-CSF may

improve outcomes in patients with newly diagnosed stage III/IV Hodgkin lymphoma treated with brentuximab vedotin plus chemotherapy. Leuk Lymphoma. 2020; 61(12):2931-2938. 37. Sureda A, Domingo-Domenech E, Gautam A. Neutropenia during frontline treatment of advanced Hodgkin lymphoma: incidence, risk factors, and management. Crit Rev Oncol Hematol. 2019;138:1-5. 38. Straus DJ, Collins GP, Walewski JA, et al. Improving outcomes with brentuximab vedotin (BV) plus chemotherapy in patients with newly diagnosed advanced stage Hodgkin lymphoma. J Clin Oncol. 2018; 36(Suppl 15):7534. 39. Battisti WP, Wager E, Baltzer L, et al. Good Publication practice for communicating company-sponsored medical research: GPP3. Ann Intern Med. 2015;163(6):461464.

haematologica | 2022; 107(5)

ARTICLE

Immunologic Dysregulation

Complement dysregulation is associated with severe COVID-19 illness

Ferrata Storti Foundation

Jia Yu, Gloria F. Gerber, Hang Chen, Xuan Yuan, Shruti Chaturvedi, Evan M. Braunstein and Robert A. Brodsky Division of Hematology, Department of Medicine, Johns Hopkins School of Medicine, Baltimore, MD, USA

ABSTRACT

evere acute respiratory syndrome coronavirus-2 (SARS-CoV-2) may manifest as thrombosis, stroke, renal failure, myocardial infarction, and thrombocytopenia, reminiscent of other complement-mediated diseases. Multiple clinical and preclinical studies have implicated complement in the pathogenesis of COVID-19 illness. We previously found that the SARS-CoV-2 spike protein activates the alternative pathway of complement (APC) in vitro through interfering with the function of complement factor H, a key negative regulator of APC. Here, we demonstrated that serum from 58 COVID-19 patients (32 patients with minimal oxygen requirement, 7 on high flow oxygen, 17 requiring mechanical ventilation and 2 deaths) can induce complementmediated cell death in a functional assay (the modified Ham test) and increase membrane attack complex (C5b-9) deposition on the cell surface. A positive modified Ham assay (>20% cell-killing) was present in 41.2% COVID-19 patients requiring intubation (n=7/17) and only 6.3% in COVID-19 patients requiring minimal oxygen support (n=2/32). C5 and factor D inhibition effectively mitigated the complement amplification induced by COVID-19 patient serum. Increased serum factor Bb level was associated with disease severity in COVID-19 patients, suggesting that APC dysregulation plays an important role. Moreover, SARS-CoV-2 spike proteins directly block complement factor H from binding to heparin, which may lead to complement dysregulation on the cell surface. Taken together, our data suggest that complement dysregulation contributes to the pathogenesis of COVID-19 and may be a marker of disease severity.

Introduction Complement has emerged as a potential driver of the pathogenesis of severe coronavirus disease 2019 (COVID-19).1–3 Clinically, endothelial damage, systemic microvascular thrombosis and recalcitrant hypercoagulability observed in COVID19 mirror other disorders of complement regulation, such as atypical hemolytic uremic syndrome (aHUS) and catastrophic antiphospholipid antibody syndrome (CAPS).1–6 Autopsy studies showed microvascular thrombi and complement deposition in the lungs and kidneys of patients with severe COVID-19.7,8 Several studies have demonstrated elevated levels of soluble C5a and C5b-9 in COVID-19 patients.9,10 Further, plasma C3a levels are higher in COVID-19 patients admitted to the intensive care unit (ICU) compared to non-ICU patients.11 A small number of patients treated with eculizumab, a monoclonal anti-C5 antibody, experienced rapid improvement in hypoxia and inflammatory markers.12,13 Three patients treated with an upstream complement C3 inhibitor demonstrated reduction in inflammatory markers and improved disease severity within 48 hours of treatment.14 Clinical trials of complement inhibitors in COVID-19 are ongoing (clinicaltrials gov. Identifier: NCT04355494, NCT04390464). The complement system, a key component of the human innate immune response, is a collection of proteases, receptors and inhibitors that functions to eliminate cellular debris, promote inflammation and defend against pathogens.15 The

haematologica | 2022; 107(5)

Haematologica 2022 Volume 107(5):1095-1105

Correspondence: ROBERT A. BRODSKY brodsro@jhmi.edu Received: May 17, 2021. Accepted: July 14, 2021. Pre-published: July 22, 2021. https://doi.org/10.3324/haematol.2021.279155

1095

J. Yu et al.

system consists of three major pathways – the classical, lectin and alternative pathways (APC). The three pathways converge on the cleavage of C3 and subsequently C5, which produces anaphylatoxins (C3a and C5a, respectively) that are associated with the acute inflammatory response and thrombosis.16 The membrane attack complex (C5b-9) is the terminal product of complement activation.15 Evidence for the role of complement in COVID-19 continues to emerge, but understanding of the underlying mechanisms remains incomplete. Recently, we demonstrated that the SARS-CoV-2 spike protein (subunits 1 and 2) converts inactivator surfaces into activator surfaces through interference with the function of complement factor H (CFH), a critical negative regulator of APC on host cells.17 These studies were performed by adding recombinant spike protein to normal human serum (NHS) and measuring complement-mediated killing on the surface of nucleated cells. Complement inhibition with C5 and factor D inhibitors effectively prevented C5b-9 accumulation induced by the SARS-CoV-2 spike proteins.17 In order to confirm that complement is a rational target for treating COVID-19, it is important to show that serum from patients infected with SARS-CoV-2 also displays complement dysregulation. In this study, we evaluated cell surface C5b-9 deposition and complement-mediated cell killing induced by COVID-19 patient sera in the modified Ham (mHam) test5,18,19 and showed that impaired complement regulation correlates with COVID-19 disease severity and can be mitigated with complement inhibitors.

Methods Patients and samples collection Serum samples and a limited clinical data set were obtained for 58 COVID-19 patients from the Clinical Characterization Protocol for Severe Infectious Diseases repository (clinicaltrials gov. Identifier: NCT04496466), approved by the Johns Hopkins Institutional Review Board. Samples were collected between April and May 2020. One sample collected outside of the patient’s hospitalization for COVID-19 was excluded. We also recruited one COVID-19 patient requiring mechanical ventilation and five healthy individuals who received the Pfizer-BioNTech (BTN162b2) COVID-19 vaccine between November and December 2020. Informed written consent for sample use from the Complement Associated Disorders Registry study was obtained. Blood was collected by venipuncture in serum separator tubes. Date of COVID-19 diagnosis was based on the earliest known positive SARS-CoV-2 nucleic acid amplification test from a nasopharyngeal swab or time an infection flag was placed on the patient’s chart. Severity of COVID-19 was graded based on World Health Organization (WHO) eight-point ordinal outcome scale.20

The modified Ham test The mHam test was used to assess complement-mediated cell killing of TF1PIGAnull cells induced by patient serum as previously described.5,17–19,21 The assay is detailed in the Online Supplementary Appendix.

Detection of complement activity by flow cytometry C5b-9 and C3c deposition after incubation of TF1PIGAnull cells with COVID-19 patient serum was measured by flow cytometry as previously described.5,17,21 The assay is detailed in Online Supplementary Appendix.

1096

Complement inhibition in patient serum Patient serum was incubated with 1 μM factor D inhibitor (ACH145951, Achillion Pharmaceuticals) or 50 μg anti-C5 monoclonal antibody (anti-C5Ab, Alexion pharmaceuticals) and then added to cells. C5b-9 and C3c deposition was measured by flow cytometry.

Quantification of serum factor Bb by enzyme-linked immunosorbant assay Factor Bb level in patient serum was measured using MicroVue Bb Plus EIA kit (Quidel). NHS preincubated with 20 μg/mL cobra venom factor (Complement Technology) served as a positive control.

Competitive heparin binding assay Immunoprecipitation of His-tagged SARS-CoV-2 spike protein subunit 1 (S1, RayBiotech) and subunit 2 (S2, RayBiotech) was performed using Heparin-Sepharose beads (BioVision). We incubated 0.2 M S1 or S2 overnight at 4°C with 80 μL beads in the presence and absence of 0.2 M CFH protein (Complement Technology) in phosphate buffered saline. After washing, the beads were denatured with LDS buffer (Invitrogen) and reducing agent (Invitrogen) at 70°C for 15 minutes and then centrifuged. The supernatants were collected for western blotting. We used 10 μL of input protein solution for western blotting. We loaded 10 μL reduced proteins to mini-PROTEIN TGX Gels (Bio-Rad Laboratories) and then transferred to PVDF membranes. The membrane was probed with anti-CFH antibody (1:1,500, Cat. Sc-53073, Santa Cruz Biotechnology) or anti-His antibody (1:2,000, Cat. Sc-53073, Santa Cruz Biotechnology). The blot was incubated with HRP-linked anti-mouse immunoglobulin (Ig) G (1:10,000, Cat. 7076S, Cell Signaling Technology) and imaged with SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Fisher).

Statistical analysis Data was summarized as mean ± standard error (SE). c2 test was used to compare the rates of mHam positivity across WHO groups. Student’s t-test was used to evaluate differences between unpaired groups. P<0.05 was considered statistically significant.

Results Patient characteristics All 58 patients studied required hospitalization for their COVID-19 illness (WHO score 4-8). Thirty-two patients required minimal oxygen support (WHO score 4), seven required high flow nasal cannula oxygen therapy (WHO score 5), 17 received mechanical ventilation and additional organ support (WHO score 7), and two died (WHO score 8) (Table 1). Five healthy individuals were recruited at the same hospital system and their serum samples post COVID-19 vaccination were tested for complement activation.

COVID-19 patient serum induced complement-mediated cell killing The mHam measures the ability of a nucleated cell to protect itself from complement-mediated cell killing in vitro in the absence of downstream cell-surface complement regulators, CD55 and CD59.18,21 Thus, mHam exposes defects in complement regulation in patient serum. This assay has been validated in complement-mediated disorders such as aHUS, HELLP and CAPS.5,18,21,22 The mHam haematologica | 2022; 107(5)

Complement dysregulation & COVID-19

Table 1. COVID-19 patient characteristics.

Age (years), median Sex Female Male Race Black White Other Ethnicity Hispanic Non-hispanic BMI, median Days from COVID-19 diagnosis to serum sample collection, median Days from hospital admission to serum sample collection, median Length of hospitalization (days), median

Oxygen (N=32)

Hi-Flow Oxygen (N=7)

Ventilation (N=17)

Death (N=2)

55 (26-78)

58 (49-69)

59 (27-80)

57 (56-57)

18 14

3 4

7 10

1 1

11 11 10

4 2 1

9 4 4

1 0 1

23 9 33.6 (20.5-58.2) 4 (1-13)

6 1 34.2 (27.1-51.9) 8 (2-16)

14 3 30.4 (16.2-46.3) 13 (3-34)

1 1 65.3 18

2 (1-13)

6 (1-10)

13 (2-32)

9 (2-45)

11 (7-15)

28 (3-70)

Figure 1. Complement-mediated cell killing induced by COVID-19 patient serum is associated with disease severity. TF1PIGAnull cells were treated with 20% COVID19 patient serum, and cell killing was measured using the modified Ham (mHam) test. Complement-mediated cell killing (%) was significantly elevated in COVID-19 patients requiring mechanical ventilation (vent), as compared to pooled normal human serum (NHS) and COVID-19 patients who needed minimal oxygen support (P<0.01). The dotted line at 20% non-viable cells represented the threshold for a positive mHam based on a previously established receiver operator curve. All experiments were run in triplicate. Sia: sialidase (used as a positive control); stx1: Shiga toxin subunit 1 (used as a positive control). Hi-Flow: High-Flow Oxygen.

test was positive (> 20% cell killing) in 41.2% (7 of 17) of patients who required intubation (WHO score 7), compared to 6.3% (2 of 32) of those who only needed minimal oxygen support (WHO score 4) (P=0.002) (Figure 1). Serum from COVID-19 patients who required mechanical ventilation induced significantly higher cell killing compared to those on minimal oxygen support (mean 17.3% vs. 7.7%, P<0.01), suggesting that SARS-CoV-2 infection impairs the ability to regulate complement on cells.

COVID-19 patient serum increased C5b-9 and C3c deposition on cell surface C3c and C5b-9 deposition, biomarkers for complement activation, on the surface of TF1PIGAnull cells after exposure to COVID-19 patient serum was evaluated. All comhaematologica | 2022; 107(5)

plement pathway buffer (GVB++) and alternative pathway specific buffer (GVB0 MgEGTA) were used. In all pathway buffer, C5b-9 deposition was increased compared to control normal human serum in virtually all COVID-19 patients regardless of WHO score (Figure 2A). In APC specific buffer, the deposition of C5b-9 was significantly elevated in patients requiring minimal oxygen support, intubation and those who died (Figure 2B) compared to control serum. C3c deposition was significantly increased in all patient groups in APC specific buffer (Figure 2D) but not in all complement pathway buffer (Figure 2C). We selected patient samples with positive mHam and increased C5b-9 deposition to perform blocking experiments. Representative examples of blocking with the complement inhibitors in two patient sera are shown in 1097

J. Yu et al.

Figure 3. C5b-9 deposition was completely inhibited by blocking the terminal complement pathway with an antiC5 antibody. The factor D inhibitor (ACH145951), an APC-specific inhibitor, partially reduced C5b-9 deposition induced by serum from patient 2 and achieved complete

inhibition in patient 1 (Figure 3A). In addition, the factor D inhibitor was more effective than anti-C5 antibody in inhibiting C3c deposition triggered by COVID-19 patient sera (Figure 3B), since it targets upstream of the anti-C5 monoclonal antibody.

Figure 2. COVID-19 patient serum induces C5b-9 and C3c deposition on the cell surface. COVID-19 patient serum led to increased cell surface C5b-9 deposition in all complement pathway buffer (A) and alternative pathway specific buffer (B). C3c deposition was increased compared to control normal human serum (NHS) in all complement pathway buffer (C) and alternative pathway specific buffer (D). Statistical significance was calculated between each disease group and the NHS control group. EDTA: ethylenediaminetetraacetic (used as a negative control); Sia: sialidase (used as a positive control for alternative pathway of complement [APC] activation); stx1: Shiga toxin subunit 1 (used as a positive control for all complement pathways activation). Vent: ventilation.

1098

haematologica | 2022; 107(5)

Complement dysregulation & COVID-19

Figure 3. C5 and factor D inhibition block complement activation induced by COVID-19 patient serum. Flow cytometry demonstrated increased C5b-9 (A) and C3c (B) deposition in two representative COVID-19 patients. C5b-9 deposition was completely blocked in the presence of 50 μg anti-C5 antibody. 1 μM factor D inhibitor (ACH145951) partially reduced the C5b-9 deposition in patient 2 and achieved complete inhibition in patient 1 (A). Factor D inhibitor also effectively decreased C3c accumulation induced by both patients’ sera, whereas anti-C5 antibody did not appreciably prevent C3c deposition (B). SSCH: side scatter; ACH145951: factor D inhibitor; AntiC5 Ab: anti-C5 monoclonal antibody. NHS: normal human serum; EDTA: ethylenediaminetetraacetic.

haematologica | 2022; 107(5)

1099

J. Yu et al.

Figure 4. The Bb level is increased in COVID-19 patient serum and associated with disease severity. Serum Bb level was significantly elevated in COVID-19 patients who required minimal oxygen support, high flow nasal cannula oxygen therapy (Hi-Flow oxygen) and mechanical ventilation (Vent) as compared to pooled normal human serum (NHS) and healthy controls (healthy ctrl). Patients requiring mechanical ventilation also had significantly higher serum Bb level than patients with minimal oxygen support (P<0.05), suggesting greater alternative pathway of complement dysregulation. All experiments were run in duplicate. CVF: cobra venom factor (used as a positive control).

Increased alternative pathway of complement activation is associated with COVID-19 disease severity We utilized standard enzyme-linked immunosorbant assay (ELISA) to measure the factor Bb level in COVID-19 patient serum. Factor Bb, which results from cleavage of factor B by factor D, is a biomarker of APC activation. Regardless of disease severity, the serum level of Bb was significantly higher in COVID-19 patients compared to the healthy controls. We also found that COVID-19 patients requiring intubation (WHO score 7) had significantly higher Bb levels than those requiring minimal oxygen support (WHO score 4) (Figure 4). These observations suggested that increased APC activation is associated with disease severity in COVID-19.

SARS-CoV-2 spike proteins compete with complement factor H for cell surface heparan sulfate binding We previously demonstrated that SARS-CoV-2 spike proteins (both subunit 1 and 2) bind heparan sulfate on the cell surface.17 Heparan sulfate also serves as a necessary cofactor for binding of SARS-CoV-2 spike proteins to the angiotensin receptor 2 (ACE2).23 CFH, a negative regulator of APC, also utilizes glycosaminoglycans, such as heparan sulfate, and a2,3 N-linked sialic acid residues for binding to nucleated cells; thus, we hypothesized that SARS-CoV-2 spike protein competes with CFH for binding to heparan sulfate and its tissue specific, more highly-sulfated variant, heparin. In order to evaluate whether SARS-CoV-2 spike proteins block CFH from binding to heparan sulfate, we compared the heparin-binding activity of CFH in the presence and absence of the SARS-CoV-2 spike proteins using heparin-linked beads. CFH alone bound to the heparinbeads with high affinity (Figure 5A, lane 1). In the presence of SARS-CoV-2 S1 and S2, binding of CFH to the heparinbeads was markedly reduced (Figure 5A, lanes 2 and 3). We also compared the heparin-binding ability of SARSCoV-2 spike proteins in the presence and absence of CFH. S1 alone showed strong binding to heparin-beads and retained high heparin-binding activity in the presence of CFH (Figure 5B). S2 bound to heparin-beads with similarly high efficiency under both conditions (Figure 5C). These results indicated the SARS-CoV-2 spike proteins have high1100

er binding affinity for heparin than CFH and interfere with the binding of CFH to heparin.

SARS-CoV-2 mRNA vaccine does not markedly increase complement activity in healthy individuals The mRNA COVID-19 vaccines employ the SARS-CoV2 spike protein as an immunogenic target. Given that SARS-CoV-2 spike proteins activate complement in vitro, concern arises whether the COVID-19 vaccine could also trigger transient complement dysregulation in vivo through generation of the spike protein. In order to test this, we obtained serum samples from five healthy individuals who received both doses of the Pfizer-BioNTech (BTN162b2) COVID-19 vaccine at three time points: before the vaccine (baseline), 24 to 48 hours after receiving the first vaccine dose, and 24 to 48 hours after the second dose. We measured the serum level of factor Bb in these individuals pre- and post-COVID-19 vaccination. Two of five individuals showed significantly higher serum Bb levels from their baseline after receiving the COVID-19 vaccine (Figure 6). Notably, these two individuals experienced side effects post vaccination including fever, headache and fatigue. We next performed functional assays to assess for cell surface complement amplification post-vaccination. On average, serum collected after the first vaccine dose did not lead to increased C5b-9 deposition on the surface of TF1PIGAnull cells compared to the individual’s serum pre-vaccination. The second dose of the vaccine led to an 11% increase in the C5b-9 deposition. All individuals had negative mHam results at baseline, which remained negative after receiving the COVID-19 vaccine (data not shown). These results from the functional assays demonstrate that the Pfizer-BioNTech SARS-CoV-2 mRNA vaccine does not sufficiently alter complement regulation in healthy individuals.

Discussion Previously, we showed that the SARS-CoV-2 spike protein dysregulates the alternative complement pathway in vitro by interfering with CFH function on the cell.17 Here, haematologica | 2022; 107(5)

Complement dysregulation & COVID-19

we extend these findings and demonstrate that the SARSCoV-2 spike protein directly competes with CFH for binding to heparan sulfate. In addition, we demonstrate, for the first time, that COVID-19 patient serum induces complement dysregulation on the cell surface using a functional assay (the mHam). Moreover, a positive mHam is highly associated with disease severity in COVID-19 patients, and C5b-9 cell deposition induced by patient sera is blocked by factor D and C5 inhibitors. Our finding that complement dysregulation is inherent to the pathogenesis of COVID-19 is in agreement with existing preclinical and clinical data. Autopsy studies revealed depositions of complement proteins in lung and other tissues co-localize with the SARS-CoV-2 spike proteins.8,24 Markers of complement activation in sera from COVID-19 hospitalized patients are associated with respiratory failure. Specifically, markers of classical/lectin (C4d), alternative (C3bBbP) and common pathway (C3bc, sC5b-9) amplification were increased in COVID-19 patient sera throughout hospitalization, indicative of sustained activation of all complement pathways.9–11,25 Sinkovits et al.26 showed that complement overactivation and consumption is predictive of in-hospital mortality in SARS-CoV-2 infection. Specifically, these authors reported consumption of C3 in the serum of patients with severe COVID-19 disease. This is consistent with our finding of increased C3c deposition on TF1PIGAnull cells. In transcriptome analysis, expression of multiple complement genes such as C2, C3, CFB and CFH were upregulated in COVID-19 patients.27,28 Gavriilaki et al.29 analyzed genetic

Figure 5 (right). SARS-CoV-2 spike proteins compete with complement factor H for the same binding sites on heparin. (A) The binding of complement factor H (CFH) to heparin-linked beads was markedly reduced in the presence of SARSCoV-2 spike protein subunit 1 (S1) and subunit 2 (S2). The western blot was performed using anti-CFH antibody. (B) S1 and S2 retained similarly high binding affinity to heparin-linked beads in the presence and absence of CFH. The western blot was performed using anti-His antibody. The input is the protein solution that is used to incubate with the beads. S: SARS-CoV-2 spike protein; S1: SARSCoV-2 spike protein subunit 1; S2: SARS-CoV-2 spike protein subunit 2.

Figure 6. SARS-CoV-2 mRNA vaccine minimally induces cell surface C5b-9 deposition but increases serum Bb level in healthy individuals. Sera from five healthy individuals were collected before they received the COVID-19 mRNA vaccine, 24-48 hours after their first vaccine dose and 24-48 hours after their second dose. (A) On average, the first vaccine dose did not lead to elevation in C5b-9 deposition from the individual’s baseline, whereas the second dose increased cell surface C5b-9 deposition by 11%. (B) Two of five healthy individuals (Ctrl) had markedly increased serum Bb level post vaccination as compared to baseline, which correlated with their vaccine side effects including headache, fatigue and fever.

haematologica | 2022; 107(5)

1101

J. Yu et al.

and clinical data from 97 patients hospitalized with COVID-19 and found an increase in rare variants associated with thrombotic microangiopathies (several involving the alternative pathway of complement) in patients with severe COVID-19 disease. Complement amplification in hospitalized patients with COVID-19 is multifactorial. Our previous data adding

recombinant spike protein to normal human serum showed that complement-mediated cell killing was almost entirely through the alternative complement pathway.17 Here, using serum from patients with COVID-19, we find additional contributions from the non-APC pathways. This may be because serum samples in this study were obtained a median of 7 days after testing positive for

Figure 7. Proposed model for complement dysregulation in SARS-CoV-2 infection. Early in infection, the SARS-CoV-2 spike protein binds heparan sulfate on the endothelial cell surface and interferes with the inhibitory function of complement factor H (CFH), leading to alternative pathway of complement dysregulation. Suppression of CFH binding results in increased cleavage of factor B by factor D and generation of Bb. Factor Bb binds to C3b to form the alternative pathway C3 convertase (C3bBb), leading to the cleavage of C3 and generation of the C5 convertase (C4b2a3b or C3bBb3b). The C5 convertase cleaves C5 to generate C5a and C5b, which complexes with C6-9 to form the membrane attack complex (C5b-9). C3a and C5a are anaphylatoxins that recruit inflammatory cells and upregulate the expression of acute phase proteins, such as C-reactive protein. Complement amplification from the classical and lectin pathways follows subsequent tissue damage, secondary infections and thromboses. Formation of antibody-antigen complexes can activate the classical complement pathway by binding to the C1 complex and cleaving C4 and C2 to form the classical C3 convertase (C4b2a). In the lectin pathway, MBL-MASP binds to carbohydrates on the surface of microbes and mediates the cleavage of C2 and C4, to generate the C3 convertase (C4b2a). Preconditions that enhance inflammation (e.g., obesity, diabetes, and vascular disease) or contribute to complement activation (e.g., third-trimester pregnancy) or dysregulation (age-related macular degeneration and other germline complement mutations) may contribute to a severe phenotype through upregulation of these pathways. HS: heparan sulfate; MBL-MASP: mannose-binding protein-mannose-binding lectin serine protease; CRP: C-reactive protein.

1102

haematologica | 2022; 107(5)

Complement dysregulation & COVID-19

COVID-19, at a time when the SARS-CoV-2 viral load is decreasing,3,31 as opposed to measuring cell surface C5b-9 deposition within minutes after adding spike protein to normal human serum. We also demonstrated that mHam positivity (measuring the activation of all complement pathways) from SARS-CoV-2 infected patients is associated with the need for mechanical ventilation; however, we did not find as strong a correlation between cell surface C5b-9 and the mHam as we did after supplementing normal human serum with the spike protein. Virtually all patients had elevated levels of Bb in their serum, even two of the healthy subjects after COVID-19 mRNA vaccination, suggesting that APC activation is an early event in the pathogenesis of SARS-CoV-2 infection. This is in agreement with the prominent role for APC found in proteomics studies, which showed increased complement factor B (CFB) levels in serum from severe COVID-19 patients.32 CFB deposition was also observed in the lung tissue of COVID-19 patients, and a CFB inhibitor blocked the C3a generated by infection of respiratory epithelial cells with SARS-CoV-2.33,34 Pekayvaz et al.35 further showed upregulation of complement factor D (CFD), produced mainly in adipocytes, in monocytes of severe COVID-19 patients. Ma et al.36 demonstrated that enhanced activation of the APC is associated with markers of endothelial injury and hypercoagulability in severe COVID-19 patients as compared to other non-COVID-19 patients admitted to the intensive care unit with acute respiratory failure. C5b-9 deposition induced by COVID-19 patient serum is blocked by both a terminal complement inhibitor (antiC5 antibody) and an alternative pathway specific inhibitor (factor D inhibitor, ACH145951). The factor D inhibitor was more effective in blocking C3c deposition induced by COVID-19 patient serum as compared to the anti-C5 antibody. These results are supported by observations from case series of eculizumab, a monoclonal anti-C5 antibody, in which treated COVID-19 patients showed significant improvements in clinical parameters.12,13 However, the phase III trial of eculizumab (clinicaltrials gov. Identifier: NCT04355494) in COVID-19 patients on mechanical ventilation was paused due to interim analysis of 122 patients showing that the drug did not meet its prespecified efficacy outcome of survival on day 29. Final results from this trial are eagerly anticipated as are those for COVID-19 patients who are hospitalized but not on mechanical ventilation. COVID-19 patients may derive benefit from complement inhibition early in their disease course or from more proximal complement inhibition. A comparative study of eculizumab versus AMY-101, an upstream C3 inhibitor, in a small number of patients showed that both decreased inflammatory markers and led to improvements in lung functions. The three patients who received AMY-101 demonstrated greater reduction in plasma levels of C3a, sC5b-9 and CFB as compared to patients who received eculizumab.37 This limited clinical data in addition to our in vitro results suggests that proximal complement inhibitors may be more effective than terminal inhibitors in reducing COVID-19 disease severity. Notably, treatment of six severely ill COVID-19 patients with Narsoplimab, a monoclonal antibody against MASP-2 inhibiting lectin-pathway activation, showed rapid reduction in serum inflammatory markers and survival in all patients.38 In addition to the COVID-19 infection itself, there are likely multifactorial contributions haematologica | 2022; 107(5)

from tissue damage, secondary infections, and thrombosis, leading to complement activation from all pathways. Further studies comparing different complement inhibitors would be valuable to identify the most appropriate therapeutic targets. The SARS-CoV-2 spike protein interferes with the function of CFH and this is likely an early event in the pathogenesis of COVID-19. In our prior work, we found that addition of purified CFH protein to serum treated with the SARS-CoV-2 spike protein decreased C3c and C5b-9 deposition on the cell surface.17 Here, we used competitive immunoprecipitation experiments to show that the SARSCoV-2 spike protein directly blocks CFH from binding to heparin, which may explain the APC dysregulation observed in COVID-19 infection (Figure 5). When binding to glycosaminoglycans on cell surface, including heparan sulfate, and a2-3 N-linked sialic acid residues, CFH achieves a more active conformation that allows for C3b binding.39 Interestingly, genetic variants in CFH, that occur in the same region where factor H binds heparan sulfate, have been identified as an important risk factor for morbidity and mortality from COVID-19.27 COVID-19 vaccines lead to the transient expression of the SARS-CoV-2 spike protein and are effective in preventing severe infection.40,41 Our vaccine studies (Figure 6) are reassuring that mRNA vaccines should not induce clinically significant complement amplification in healthy individuals, as suggested by the negative results from our functional assays; however, more data is necessary in patients with disorders of complement regulation, such as paroxysmal nocturnal hemoglobinuria (PNH), aHUS, CAPS, HELLP and cold agglutinin disease.42 Relapse of aHUS has been reported in patients with COVID-19 infection,43 and PNH patients have experienced adverse reactions to COVID-19 vaccines including severe hemolysis and need for blood transfusions even while on a C5 inhibitor.44 This evidence suggests that although complement activation induced by COVID-19 vaccines is wellcontrolled in healthy individuals, patients with disorders of complement regulation could be at higher risk for adverse reactions to vaccine. Limitations of our study are that we received a limited amounts of patient serum for experiments and were unable to test other complement markers such as C4d deposition on the cell surface. We also had limited access to clinical information from which to draw robust conclusions regarding the association of complement activation with clinical parameters. Further, serum sample collection was not standardized and occurred at different time points from the initial diagnosis of COVID-19 and hospital admission. For example, in one of the two patients who died due to multiorgan failure from COVID-19, serum was collected near the end of his clinical course, at which point peak amplification of complement may have passed. Finally, we do not have serial samples from patients to estimate the persistence of complement activation over time. In future studies, it will be important to do serial mHam, Bb, and surface C5b-9 deposition studies starting soon after infection and correlating with SARS-CoV-2 viral load. In summary, we showed that COVID-19 patient serum can induce complement dysregulation on cell surfaces that tracks with disease severity. Our previous data showed that the SARS-CoV-2 spike proteins convert inactivator surfaces to activator surfaces. Taken together, we postu1103

J. Yu et al.

late that dysregulation of the APC is likely an early event after SARS-CoV-2 infection. Complement amplification from classical and lectin pathways following tissue damage, secondary infections, and thrombosis likely exacerbate end-organ damage similar to severe forms of aHUS and CAPS (Figure 7). Preconditions that lead to inflammation (e.g., obesity, diabetes and vascular disease) or contribute to complement activation (e.g., third-trimester pregnancy) or dysregulation (age-related macular degeneration and other germline complement mutations) may contribute to a severe phenotype. Indeed, components of the APC (CFB, CFD, and C3) are elevated in patients with obesity and insulin resistance.45 Prospective studies correlating SARS-CoV-2 viral load to complement-mediated cell damage over the course of infection and additional genetic studies probing for rare variants in complement regulatory genes are needed. Our data also suggest that for complement inhibitors to be most effective, they should be initiated early in the disease process, but this too requires prospective study, as is the subject of the ongoing TACTIC-R study (clinical trials gov. Identifier: NCT04390464).46

References 1. Java A, Apicelli AJ, Liszewski MK, et al. The complement system in COVID-19: friend and foe? JCI Insight. 2020;5(15):e140711. 2. Gavriilaki E, Brodsky RA. Severe COVID19 infection and thrombotic microangiopathy: success does not come easily. Br J Haematol. 2020;189(6):e227-e230. 3. Risitano AM, Mastellos DC, Huber-Lang M, et al. Complement as a target in COVID-19? Nat Rev Immunol. 2020; 20(6):343-344. 4. Campbell CM, Kahwash R. Will complement inhibition be the new target in treating COVID-19-related systemic thrombosis? Circulation. 2020;141(22):1739-1741. 5. Chaturvedi S, Braunstein EM, Yuan X, et al. Complement activity and complement regulatory gene mutations are associated with thrombosis in APS and CAPS. Blood. 2020;135(4):239-251. 6. Baines AC, Brodsky RA. Complementopathies. Blood Rev. 2017; 31(4):213-223. 7. Wichmann D, Sperhake JP, Lütgehetmann M, et al. Autopsy findings and venous thromboembolism in patients with COVID-19: a prospective cohort study. Ann Intern Med. 2020;173(4):268-277. 8. Diao B, Wang C, Wang R, et al. Human kidney is a target for novel severe acute respiratory syndrome coronavirus 2 infection. Nat Commun. 2021;12(1):1-9. 9. Carvelli J, Demaria O, Vély F, et al. Association of COVID-19 inflammation with activation of the C5a–C5aR1 axis. Nature. 2020;588(7836):146-150. 10. Cugno M, Meroni PL, Gualtierotti R, et al. Complement activation in patients with COVID-19: a novel therapeutic target. J Allergy Clin Immunol. 2020;146(1):215217. 11. de Nooijer AH, Grondman I, Janssen NAF, et al. Complement activation in the disease course of coronavirus disease 2019 and its effects on clinical outcomes. J Infect Dis. 2021;223(2):214-224. 12. Diurno F, Numis FG, Porta G, et al.

1104

Disclosures RAB has served on the advisory board for Alexion Pharmaceutical Inc.; SC has served on boards for Alexion and Sanofi-Genzyme, and her institution has received research funding on her behalf from Takeda. Contributions JY designed and performed experiments, analyzed the data, enrolled patients and wrote the first draft of the manuscript; GFG enrolled patients, analyzed the data, and edited the manuscript; HC designed and performed the heparin binding assay, analyzed the data, and edited the manuscript; XY, SC and EMB interpreted the data and edited the manuscript; RAB designed the study, supervised the experiments, interpreted the data and wrote portions of the manuscript. Funding This work was supported by grants from National Institutes of Health (NIH), National Heart, Lung, and Blood Institute (NHLBI) grant R01 HL 133113 (RAB); NIH grant K08 HL138142 (EMB); NHLBI T32 HL 007525 (GFG).

Eculizumab treatment in patients with COVID-19: preliminary results from real life ASL Napoli 2 Nord experience. Eur Rev Med Pharmacol Sci. 2020;24(7):4040-4047. 13. Annane D, Heming N, Grimaldi-Bensouda L, et al. Eculizumab as an emergency treatment for adult patients with severe COVID-19 in the intensive care unit: A proof-of-concept study. EClinicalMedicine. 2020;28:100590. 14. Mastaglio S, Ruggeri A, Risitano AM, et al. The first case of COVID-19 treated with the complement C3 inhibitor AMY-101. Clin Immunol. 2020;215:108450. 15. Ling M, Murali M. Analysis of the complement system in the clinical immunology laboratory. Clin Lab Med. 2019;39(4):579590. 16. Kourtzelis I, Markiewski MM, Doumas M, et al. Complement anaphylatoxin C5a contributes to hemodialysis-associated thrombosis. Blood. 2010;116(4):631-639. 17. Yu J, Yuan X, Chen H, Chaturvedi S, Braunstein EM, Brodsky RA. Direct activation of the alternative complement pathway by SARS-CoV-2 spike proteins is blocked by factor D Inhibition. Blood. 2020;136(18):2080-2089. 18. Gavriilaki E, Yuan X, Ye Z, et al. Modified Ham test for atypical hemolytic uremic syndrome. Blood. 2015;125(23):3637-3646. 19. Vaught AJ, Braunstein EM, Jasem J, et al. Germline mutations in the alternative pathway of complement predispose to HELLP syndrome. JCI Insight. 2018;3(6):5-7. 20. World Health Organization. COVID-19 Therapeutic Trial Synopsis. World Health Organization. https://cdn. who.int/media/ docs/default-source/blue-print/covid-19therapeutic-trial-synopsis. pdf?sfvrsn=44b83344_1&download=true 2020, Accessed May 12, 2021. 21. Yuan X, Yu J, Gerber G, et al. Ex vivo assays to detect complement activation in complementopathies. Clin Immunol. 2020; 221: 108616. 22. Vaught AJ, Braunstein EM, Jasem J, et al. Germline mutations in the alternative pathway of complement predispose to HELLP syndrome. JCI Insight. 2018;3(6):e99128. 23. Clausen TM, Sandoval DR, Spliid CB, et al.

SARS-CoV-2 infection depends on cellular heparan sulfate and ACE2. Cell. 2020;183(4):1043-1057.e15. 24. Magro C, Mulvey JJ, Berlin D, et al. Complement associated microvascular injury and thrombosis in the pathogenesis of severe COVID-19 infection: a report of five cases. Transl Res. 2020;220:1-13. 25. Holter JC, Pischke SE, de Boer E, et al. Systemic complement activation is associated with respiratory failure in COVID-19 hospitalized patients. Proc Natl Acad Sci U S A. 2020;117(40):25018-25025. 26. Sinkovits G, Mező B, Réti M, et al. Complement overactivation and consumption predicts in-hospital mortality in SARSCoV-2 infection. Front Immunol. 2021;12: 663187. 27. Ramlall V, Thangaraj PM, Meydan C, et al. Immune complement and coagulation dysfunction in adverse outcomes of SARSCoV-2 infection. Nat Med. 2020; 26(10):1609-1615. 28. Valenti L, Griffini S, Lamorte G, et al. Chromosome 3 cluster rs11385942 variant links complement activation with severe COVID-19. J Autoimmun. 2021;117: 102595. 29. Gavriilaki E, Asteris PG, Touloumenidou T, et al. Genetic justification of severe COVID-19 using a rigorous algorithm. Clin Immunol. 2021;226:108726. 30. Jang S, Rhee JY, Wi YM, Jung BK. Viral kinetics of SARS-CoV-2 over the preclinical, clinical, and postclinical period. Int J Infect Dis. 2021;102:561-565. 31. Sun J, Tang X, Bai R, et al. The kinetics of viral load and antibodies to SARS-CoV-2. Clin Microbiol Infect. 2020;26(12):1690.e11690.e4. 32. Messner CB, Demichev V, Wendisch D, et al. Ultra-high-throughput clinical proteomics reveals classifiers of COVID-19 infection. Cell Syst. 2020;11(1):11-24.e4. 33. Yan B, Freiwald T, Chauss D, et al. SARSCoV-2 drives JAK1/2-dependent local complement hyperactivation. Sci Immunol. 2021;6(58):1-20. 34. Macor P, Durigutto P, Mangogna A, et al. Multi-organ complement deposition in COVID-19 patients. medRxiv. 2021 Jan 8.

haematologica | 2022; 107(5)

Complement dysregulation & COVID-19

doi: 10.1101/2021.01.07.21249116 [preprint, not peer-reviewed]. 35. Pekayvaz K, Leunig A, Kaiser R, et al. Protective immune trajectories in early viral containment of non-pneumonic SARSCoV-2 infection. bioRxiv. doi: https://doi.org/10.1101/2021.02.03.429351 [preprint, not peer-reviewed]. 36. Ma L, Sahu SK, Cano M, et al. Increased complement activation is a distinctive feature of severe SARS-CoV-2 infection. Sci Immunol. 2021;6(59):1-18. 37. Mastellos DC, Pires da Silva BGP, Fonseca BAL, et al. Complement C3 vs C5 inhibition in severe COVID-19: early clinical findings reveal differential biological efficacy. Clin Immunol. 2020;220:108598. 38. Rambaldi A, Gritti G, Micò MC, et al. Endothelial injury and thrombotic microan-

haematologica | 2022; 107(5)

giopathy in COVID-19: treatment with the lectin-pathway inhibitor narsoplimab. Immunobiology. 2020;225(6):152001. 39. Perkins SJ, Fung KW, Khan S. Molecular interactions between complement factor H and its heparin and heparan sulfate ligands. Front Immunol. 2014;5:1-14. 40. Polack FP, Thomas SJ, Kitchin N, et al. Safety and Efficacy of the BNT162b2 mRNA Covid-19 Vaccine. N Engl J Med. 2020;383(27):2603-2615. 41. Baden LR, El Sahly HM, Essink B, et al. Efficacy and safety of the mRNA-1273 SARS-CoV-2 Vaccine. N Engl J Med. 2021;384(5):403-416. 42. Gavriilaki E, Brodsky RA. Complementopathies and precision medicine. J Clin Invest. 2020;130(5):2152-2163. 43. Ville S, Le Bot S, Chapelet-Debout A, et al.

Atypical HUS relapse triggered by COVID19. Kidney Int. 2021;99(1):267-268. 44. Gerber GF, Yuan X, Yu J, et al. COVID-19 Vaccines induce severe hemolysis in paroxysmal nocturnal hemoglobinuria. Blood. 2021;137(26):3670-3673. 45. Shim K, Begum R, Yang C, Wang H. Complement activation in obesity, insulin resistance, and type 2 diabetes mellitus. World J Diabetes. 2020;11(1):1-12. 46. Kulkarni S, Fisk M, Kostapanos M, et al. Repurposed immunomodulatory drugs for Covid-19 in pre-ICU patients - multi-arm Therapeutic study in pre-ICU patients admitted with Covid-19 - repurposed drugs (TACTIC-R): a structured summary of a study protocol for a randomised controlled trial. Trials. 2020;21 (1):20-22.

1105

ARTICLE Ferrata Storti Foundation

Myeloproliferative Disorders

Hemorrhage in patients with polycythemia vera receiving aspirin with an anticoagulant: a prospective, observational study Jeffrey I. Zwicker,1 Dilan Paranagama,2 David S. Lessen,3 Philomena M. Colucci,2 and Michael R. Grunwald4 Beth Israel Deaconess Medical Center, Division of Hematology, Harvard Medical School, Boston, MA; 2Incyte Corporation, Wilmington, DE; 3Division of Hematology, Sylvester Comprehensive Cancer Center, University of Miami Miller School of Medicine, Miami, FL and 4Department of Hematologic Oncology and Blood Disorders, Levine Cancer Institute, Atrium Health, Charlotte, NC, USA 1

Haematologica 2022 Volume 107(5):1106-1110

ABSTRACT

Correspondence: JEFFREY ZWICKER jzwicker@bidmc.harvard.edu Received: April 21, 2021. Accepted: June 16, 2021. Pre-published: June 24, 2021. https://doi.org/10.3324/haematol.2021.279032

1106

olycythemia vera (PV) is associated with increased risk of thrombosis and hemorrhage. Aspirin, recommended for primary thromboprophylaxis, is often combined with anticoagulants during management of acute thrombotic events. The safety of dual antiplatelet and anticoagulant therapy is not established in PV. In a prospective, observational study, 2,510 patients with PV were enrolled at 227 sites in the United States. Patients were monitored for the development of hemorrhage and thrombosis after enrollment. A total of 1,602 patients with PV received aspirin with median follow-up of 2.4 years (range, 0-3.6 years). The exposure-adjusted rate of all hemorrhages in patients receiving aspirin alone was 1.40 per 100 patient-years (95% confidence interval [CI]: 0.99-1.82). The combination of aspirin plus anticoagulant was associated with an incidence of hemorrhage of 6.75 per 100 patient-years (95% CI: 3.04-10.46). The risk of hemorrhage was significantly greater in patients receiving the combination of aspirin and anticoagulant compared with aspirin alone (total hemorrhages, hazard ratio [HR]: 5.83; 95% CI: 3.36-10.11; P<0.001; severe hemorrhage, HR: 7.49; 95% CI: 3.02-18.62; P<0.001). Periods of thrombocytosis (>600×109/L) were associated with an increased risk of hemorrhage (HR: 2.25; 95% CI: 1.164.38; P=0.02). Rates of hemorrhage were similar for aspirin in combination with warfarin or direct-acting oral anticoagulants. We conclude that the combination of aspirin and anticoagulants is associated with significantly increased risk of hemorrhage in patients with PV (clinicaltrials gov. Identifier: NCT02252159).

Introduction Polycythemia vera (PV) is a myeloproliferative neoplasm (MPN) usually involving a mutation of the Janus kinase 2 gene and characterized phenotypically by an increased risk of thrombosis and hemorrhage.1 Low-dose aspirin effectively reduces the rate of arterial and venous thrombosis and is recommended for primary thromboprophylaxis in patients diagnosed with PV.2,3 Despite the administration of antiplatelet and cytoreductive therapy, approximately 25% of patients with PV ultimately develop thrombotic complications.4 The optimal approach to the management of thrombosis in patients with MPN is undefined. Anticoagulation is recommended for the acute treatment of thrombosis3 but is associated with a recurrent thrombosis incidence of 5-6 events per 100 patient-years5,6 and major hemorrhage incidence of 1-3 events per 100 patient-years.5-7 Due to the high rates of recurrent thrombotic events, aspirin is often continued along with therapeutic anticoagulation, but the relative safety of combined anticoagulant-antiplatelet therapy is not established. We evaluated the rates of hemorrhage among patients with PV receiving anticoagulation with or without aspirin who were enrolled in a multicenter, prospective, observational study.

haematologica | 2022; 107(5)

Anticoagulants and hemorrhage in PV

Methods Study design REVEAL is a multicenter, non-interventional, prospective, observational study of patients with PV (clinicaltrials gov. Identifier: NCT02252159). Details regarding study design, patient eligibility and study conduct were previously described.8 The study enrolled 2,510 patients with a diagnosis of PV from 227 sites in the United States over a 24-month period (July 22, 2014 to August 3, 2016) followed by an observation period up to 3 years from the date of last patient enrollment. Eligibility criteria included age ≥18 years old and a clinical diagnosis of PV. Patients were excluded if they had a life expectancy <6 months; diagnosis of myelofibrosis, acute myeloid leukemia or myelodysplastic syndrome; underwent (or were planning to undergo) an allogeneic transplant; or were status post-splenectomy. Data pertaining to PV-directed treatment, concomitant medications, adverse events, laboratory values and other relevant clinical measures were collected for at least 6 months before enrollment and during the follow-up period. The study was conducted in accordance with the Declaration of Helsinki. Approval of all study materials was obtained by a central Institutional Review Board (Sterling IRB, Atlanta, GA, USA) and local Institutional Review Boards of participating centers. All patients provided written informed consent.

Cox proportional hazards model adjusting for age, sex, duration of PV, history of hemorrhage and platelet count ranges. Platelet counts were included in the models as three-level categorical variables: ≤100×109/L; >100×109/L and ≤600×109/L; >600×109/L based on the estimated thresholds for increased hemorrhage in the Primary Thrombocythemia study.10 For the time-dependent covariate analysis, linear interpolation was used to determine laboratory values at time points between observed values. Missing data were not imputed. Observed P-values were assessed at a=0.05 level for statistical significance. The P-values were not adjusted for multiple tests.

Results Of the 2,510 enrolled patients, 1,602 patients with PV received aspirin and were included in the analysis (Table 1). The median time from PV diagnosis to enrollment was 4 years (range, 0-39.2 years). At the time of enrollment, 103 patients (6.4%) were receiving the combination of aspirin plus anticoagulant and 1,499 patients (93.6%) were receiving aspirin alone. Median follow-up of patients from enrollment to the time of analysis was 2.4 years (range, 0-3.6 years).

Bleeding events

Overall incidence of hemorrhage

Hemorrhagic events occurring before enrollment were identified from the patient’s medical history collected at the time of enrollment. Hemorrhagic events during the follow-up period were determined from prospectively collected adverse events. Postenrollment events were graded according to Common Terminology Criteria for Adverse Events (CTCAE) version 4.03 by treating physicians.9 Grade 3 or 4 hemorrhagic events were classified as severe hemorrhages. Use of anticoagulant medications and aspirin was determined based on concomitant medications data collected at baseline and prospectively throughout the study. Medication start dates and end dates were used to determine whether patients were receiving an anticoagulant concomitantly with aspirin. Rivaroxaban, apixaban and dabigatran were combined into a direct-acting oral anticoagulants (DOAC) group.

Sixty-nine patients developed one or more hemorrhagic events after enrollment. The exposure-adjusted rate of all hemorrhages was 1.71 per 100 patient-years (95% confidence interval [CI]: 1.27-2.15), and the overall cumulative incidence of hemorrhage at 3 years was 4.70% (95% CI: 3.47-6.15). Among those with hemorrhage, the most common sites were gastrointestinal (n=30, 43.5%), cutaneous (n=19, 27.5%), central nervous system (n=10, 14.5%) and genitourinary (n=7, 10.1%). Severe hemorrhagic events occurred in 25 patients with an exposure-adjusted rate of 0.55 events per 100 patientyears (95% CI: 0.30-0.80). The most common sites of severe hemorrhage were gastrointestinal (n=11, 44%) and central nervous system (n=9, 36%). Of those with severe hemorrhagic events, there were five fatal events with all but one involving a central nervous system bleed.

Statistical analysis In order to assess cumulative incidence rates and exposureadjusted event rates, patients were assigned to groups based on their treatment at enrollment (aspirin or aspirin plus anticoagulant). If a patient changed their treatment group, then time was censored on the last day in the original treatment group. If a patient died or discontinued the study, time was censored at date of death or discontinuation, respectively. Time was censored at the last visit for patients who continued to be enrolled in the study without any events. The risk of hemorrhage associated with aspirin plus anticoagulant versus aspirin alone was assessed with a

Table 1. Baseline characteristics.

Characteristic Median (range) age, years Women, n (%) Median (range) disease duration, years History of hypertension, n (%) History of hemorrhagic events, n (%) History of thrombotic events, n (%) Arterial Venous

haematologica | 2022; 107(5)

N=1,602 67 (22-95) 734 (45.8) 4.0 (0.0-39.2) 916 (57.2) 103 (6.4) 300 (18.7) 157 (9.8) 164 (10.2)

Risk of hemorrhage associated with aspirin alone or with anticoagulants Aspirin alone was associated with 1.4 hemorrhagic events per 100 patient-years (95% CI: 0.99-1.82), whereas the combination of aspirin with anticoagulation was associated with 6.75 hemorrhagic events per 100 patient-years (95% CI: 3.04-10.46). The cumulative incidence of hemorrhage at 3 years in patients receiving aspirin alone was 3.6% (95% CI: 2.64-4.84) and 19.8% (95% CI: 9.78-32.45) for those receiving aspirin with an anticoagulant as shown in Figure 1A. The rate of severe hemorrhage was relatively higher among those who received aspirin with an anticoagulant (1.46 events per 100 patient-years; 95% CI: 0-3.21) compared with those who received aspirin alone (0.49 events per 100 patient-years; 95% CI: 0.25-0.74). The cumulative incidence rate of severe hemorrhage at 3 years was 3.7% (95% CI: 0.96-9.50) for patients who had aspirin with an anticoagulant and 1.2% (95% CI: 0.74-2.00) for patients who had aspirin alone (Figure 1B). In a Cox proportional hazards model, the use of aspirin plus anticoagulant was associated with a greater than 5-fold increased risk of hemorrhage compared with aspirin alone 1107

J.I. Zwicker et al.

(Table 2; hazard ratio [HR]: 5.83; 95% CI: 3.36-10.11; P<0.001). The risk of severe hemorrhage was significantly increased with the combination of aspirin plus anticoagulant compared with aspirin alone (HR: 7.49; 95% CI: 3.0218.62; P<0.001).

Influence of platelet count on rates of hemorrhage In a time-dependent exposure Cox proportional hazards model, thrombocytosis was associated with an increased risk of total and severe hemorrhages. The cumulative exposure time for platelet count elevation (≥600×109/L) was 305.5 patient-years, controlled platelet count (100600×109/L) was 2,443.7 patient-years and thrombocytopenia (below 100×109/L) was 21.7 patient-years. As shown in Table 2, thrombocytopenia (platelet count <100×109/L) was not statistically associated with an increased risk of hemorrhage (HR: 1.72; 95% CI: 0.23-12.71; P=0.60) or severe hemorrhage (HR: 5.15; 95% CI: 0.65-40.75; P=0.12) when compared with periods when platelet count was between 100×109/L and 600×109/L. Thrombocytosis was associated with an increased risk of any hemorrhage (HR: 2.25; 95% CI: 1.16-4.38; P=0.02) but not severe hemorrhage (HR: 1.03; 95% CI, 0.24-4.52; P=0.97). Periods of modest thrombocytosis (i.e., platelet counts between 400x109/L to 600x109/L) were not associated with increased risk of hemorrhage compared to periods of platelet counts below 400x109/L (HR: 1.10; 95% CI: 0.59-2.05).

Comparison of aspirin with direct-acting oral anticoagulant or warfarin Of the patients who received an anticoagulant, 73 received warfarin and 72 received a direct-acting oral anticoagulant (DOAC). The DOAC group included rivaroxaban (n=50), apixaban (n=31) and dabigatran (n=6). In a time-dependent exposure model, the risk of severe hemorrhage was significantly greater with warfarin with aspirin compared with aspirin alone (Table 2). The risk of severe hemorrhage was not statistically different with the use of aspirin plus DOACs versus aspirin plus warfarin.

Incidence of thrombosis At the time of enrollment, 300 patients (18.7%) had a history of thrombosis with 157 arterial events and 164 venous thromboembolic events. During the follow-up period, 61 patients experienced a thrombotic event. A numerically higher proportion of patients receiving anticoagulant plus aspirin experienced a thrombotic event compared with those who were on aspirin alone (Figure 2). The 3-year cumulative incidence of thrombosis (arterial or venous) was 4.8% (95% CI: 3.57-6.20). In a Cox proportional hazards model with treatment as a two-level, time-dependent covariate (adjusting for age, sex, disease duration, history of thrombosis before enrollment), the use of aspirin with an anticoagulant was not associated with a lower rate of thrombosis compared with aspirin alone (HR: 1.97; 95% CI: 0.57-6.78; P=0.28).

Figure 1. Cumulative incidence of hemorrhage in patients with polycythemia vera receiving aspirin with or without anticoagulants. The cumulative incidence of (A) any hemorrhage or (B) severe hemorrhage shown for patients receiving aspirin plus anticoagulant (blue) and aspirin alone (green).

Table 2. Cox proportional hazard ratios for risk factors contributing to hemorrhage in polycythemia vera.

Variable Models 1 and 2* AC + ASA Models 3 and 4* Warfarin + ASA DOAC + ASA Other anticoagulants + ASA Models 5 and 6* Platelet ≤100×109/L Platelet >600×109/L

All hemorrhages HR (95% CI)

P-value

Severe hemorrhage HR (95% CI)

P-value

5.83 (3.36-10.11)

<0.001

7.49 (3.02-18.62)

<0.001

4.44 (2.06-9.54) 3.97 (1.65-9.60) 9.26 (4.13-20.75)

<0.001 0.002 <0.0001

5.48 (1.81-16.61) 2.23 (0.50-9.96) 3.79 (0.86-16.68)

0.003 0.29 0.078

1.72 (0.23-12.71) 2.25 (1.16-4.38)

0.60 0.02

5.15 (0.65-40.75) 1.03 (0.24-4.52)

0.12 0.97

AC: anticoagulant; ASA: aspirin; CI: confidence interval; DOAC: direct-acting oral anticoagulant; HR: hazard ratio. *Models for all hemorrhage and severe hemorrhage were run separately, adjusting for variables of age, sex, disease duration and history of bleeding. Models 1 and 2: HR compared with aspirin alone. Models 3 and 4: all HR compared with aspirin alone. Models 5 and 6: HR compared with normal platelet range (100-600×109/L).

1108

haematologica | 2022; 107(5)

Anticoagulants and hemorrhage in PV

Discussion Management of thrombotic events in patients with MPN is challenging. Although low-dose aspirin can effectively reduce the rate of thrombotic complications in PV, the optimal antithrombotic approach to secondary prevention is less clear. There are no randomized studies evaluating different antithrombotic regimens in patients with MPN, and recurrent thrombosis is common after initiation of therapeutic anticoagulation.5,6 Combined antiplatelet and anticoagulant therapy is often prescribed in patients with MPN without knowledge of hemorrhagic risks. In this large, prospective multicenter study, we evaluated whether the addition of aspirin to an anticoagulant increased the risk of hemorrhage with PV. We observed that the combination of anticoagulant with aspirin was associated with a greater than seven-fold increased risk of severe hemorrhage compared with aspirin alone. Our observation that the combination of aspirin and an anticoagulant significantly increased the risk of hemorrhage is in keeping with the larger published experience evaluating the combination for cardiovascular indications.11,12 The Augustus trial recently reported rates of hemorrhagic outcomes in more than 4,600 patients with atrial fibrillation undergoing percutaneous coronary interventions who were randomized to warfarin or apixaban plus aspirin or placebo.12 The risk of major hemorrhage was significantly higher in patients receiving aspirin plus an anticoagulant compared with an anticoagulant without aspirin (HR: 1.89; 95% CI: 1.59-2.24; P<0.001).12 There are limited published data on the safety of combined antiplatelet and antithrombotic therapy in patients with MPN. In a recent Italian cohort that included 155 patients with an MPN and history of thrombosis, the 3-year cumulative incidence of recurrent thrombosis was 18.0% accompanied by a 6.5% incidence of major hemorrhage.5 Of the 19 patients (12%) in this cohort who received combined vitamin K antagonist with aspirin, the incidence of major hemorrhage was 3.8 per 100 patient-years (95% CI: 0.4-13.8), which was statistically similar to those who received vitamin K antagonist alone (2.2 per 100 patient-years; 95% CI: 0.9-4.4; P=0.50).5 Older series included even fewer patients (less than 10) exposed to combined antithrombotic therapy.7 There is limited evidence regarding the safety of DOAC in combination with aspirin in patients with MPN.13,14 The administration of a DOAC with aspirin did not appear to increase the risk of hemorrhage compared with warfarin and aspirin. There is evidence in cardiology cohorts that the combination of a DOAC with aspirin may be safer than warfarin combined with aspirin.11,12 We observed a similar trend for severe hemorrhages, but the difference was not statistically significant. Although the combination of aspirin and an anticoagulant significantly increased the risk of severe hemorrhage relative to aspirin alone, whether the combination therapy offers therapeutic benefit over an anticoagulant alone was not established. The indication for anticoagulant therapy was not captured, and the absence of this information precludes such an efficacy analysis. Due to its over-thecounter availability, the use of aspirin may not be fully captured in the medical records; therefore, the analysis was limited to patients with documented aspirin use. We also note that patients were monitored for the development of hemorrhage, which was graded by CTCAE criteria. Alternative definitions of severe hemorrhage such as those established by the International Society of Thrombosis and haematologica | 2022; 107(5)

Figure 2. Kaplan-Meier estimate of thrombotic event-free survival by medication group at enrollment. Kaplan-Meier estimate of thrombotic event (TE)-free survival by medication groups of aspirin plus anticoagulant (blue) and aspirin alone (green).

Hemostasis (ISTH)15 were not utilized. Nevertheless, the vast majority of the hemorrhages described as severe in this cohort would be expected to meet the ISTH definition of major hemorrhage considering that grade 3 or 4 hemorrhages per CTCAE criteria require procedural intervention or are considered life-threatening. Other limitations of this analysis include the non-randomized nature of the study, the relatively short follow-up period, and the differences in the disease and/or treatment duration. In summary, this prospective, multicenter, observational study revealed a significant increase in the risk of hemorrhage and severe hemorrhage attributed to the combination of aspirin and anticoagulants compared with aspirin alone. In patients previously taking aspirin for cardiovascular risk modification, the American Society of Hematology advises against the continuation of aspirin following the initiation of anticoagulation for treatment of VTE.16 Whether the combination of aspirin and anticoagulants offers therapeutic benefit in the management of thromboembolic disease in PV is not established. Disclosures JIZ reports research funding from Incyte Corporation and Quercegen Pharmaceuticals; consultancy fees from CSL, Merck, Parexel and Sanofi; and honoraria/advisory fees from Daiichi Sankyo, Pfizer/BMS and Portola. DP and PMC are employees and shareholders of Incyte Corporation. DSL reports research funding from Astellas Pharma, Exact Sciences, Incyte Corporation and Pfizer; speakers bureau and consulting honoraria from Bayer and Incyte Corporation; and speakers bureau participation for Amgen. MRG has provided consultancy to AbbVie, Agios, Amgen, Astellas, Cardinal Health, Bristol Myers Squibb, Daiichi Sankyo, Gilead, Incyte Corporation, Karius, Merck, Pfizer, Premier Inc., Sierra Oncology, Stemline, and Trovagene; and has received research funding from Forma Therapeutics, Genentech/Roche, Incyte Corporation and Janssen. Contributions PMC, DSL, MRG, DP, and JIZ participated in the clinical study design and conduct; JIZ and DP performed data analyses; 1109

J.I. Zwicker et al.

JIZ and DP drafted the manuscript. All authors reviewed, provided substantive comments and approved the manuscript.

Corporation); formatting and copyediting assistance was provided by Envision Pharma Group and funded by Incyte Corporation.

Acknowledgments The authors would like to thank the patients and their families, the investigators, and the site personnel who participated in this study. Editorial support was provided by Amanda M. Kelly (Incyte

Data sharing statement De-identified patient-level data pertaining to these analyses are available upon reasonable request. Proposals for data access should be sent to datasharing@incyte.com.

References 1. Tefferi A, Pardanani A. Myeloproliferative neoplasms: a contemporary review. JAMA Oncol. 2015;1(1):97-105. 2. Landolfi R, Marchioli R, Kutti J, et al. Efficacy and safety of low-dose aspirin in polycythemia vera. N Engl J Med. 2004;350 (2):114-124. 3. Kreher S, Ochsenreither S, Trappe RU, et al. Prophylaxis and management of venous thromboembolism in patients with myeloproliferative neoplasms: consensus statement of the Haemostasis Working Party of the German Society of Hematology and Oncology (DGHO), the Austrian Society of Hematology and Oncology (ÖGHO) and Society of Thrombosis and Haemostasis Research (GTH e.V.). Ann Hematol. 2014;93(12):1953-1963. 4. Szuber N, Mudireddy M, Nicolosi M, et al. 3023 Mayo Clinic patients with myeloproliferative neoplasms: risk-stratified comparison of survival and outcomes data among disease subgroups. Mayo Clin Proc. 2019;94(4):599-610. 5. De Stefano V, Ruggeri M, Cervantes F, et al. High rate of recurrent venous thromboembolism in patients with myeloproliferative neoplasms and effect of prophylaxis with vitamin K antagonists. Leukemia. 2016;30 (10):2032-2038. 6. Hernández-Boluda J, Arellano-Rodrigo E,

1110

Cervantes F, et al. Oral anticoagulation to prevent thrombosis recurrence in polycythemia vera and essential thrombocythemia. Ann Hematol. 2015;94(6):911918. 7. De Stefano V, Za T, Rossi E, et al. Recurrent thrombosis in patients with polycythemia vera and essential thrombocythemia: incidence, risk factors, and effect of treatments. Haematologica. 2008;93(3):372-380. 8. Grunwald MR, Stein BL, Boccia RV, et al. Clinical and disease characteristics from REVEAL at time of enrollment (baseline): prospective observational study of patients with polycythemia vera in the United States. Clin Lymphoma Myeloma Leuk. 2018;18(12):788-795.e2. 9. US Department of Health and Human Services. Common Terminology Criteria for Adverse Events (CTCAE) v4.0. 2010. https://evs.nci.nih.gov/ftp1/CTCAE/CTC AE_4.03/CTCAE_4.03_2010-0614_QuickReference_8.5x11.pdf. Accessed 25 June 2020. 10. Campbell PJ, MacLean C, Beer PA, et al. Correlation of blood counts with vascular complications in essential thrombocythemia: analysis of the prospective PT1 cohort. Blood. 2012;120(7):1409-1411. 11. Flaker GC, Gruber M, Connolly SJ, et al. Risks and benefits of combining aspirin with anticoagulant therapy in patients with atrial fibrillation: an exploratory analysis of

stroke prevention using an oral thrombin inhibitor in atrial fibrillation (SPORTIF) trials. Am Heart J. 2006;152(5):967-973. 12. Lopes RD, Heizer G, Aronson R, et al. Antithrombotic therapy after acute coronary syndrome or PCI in atrial fibrillation. N Engl J Med. 2019;380(16):1509-1524. 13. Kaifie A, Kirschner M, Wolf D, et al. Bleeding, thrombosis, and anticoagulation in myeloproliferative neoplasms (MPN): analysis from the German SAL-MPN-registry. J Hematol Oncol. 2016;9:18. 14. Barbui T, De Stefano V, Falanga A, et al. Addressing and proposing solutions for unmet clinical needs in the management of myeloproliferative neoplasm-associated thrombosis: a consensus-based position paper. Blood Cancer J. 2019;9(8):61. 15. Schulman S and Kearon C, on behallf of the Subcommittee on Control of Anticoagulation of the Scientific and Standardization Committee of the International Society on Thrombosis and Haemostasis. Definition of major bleeding in clinical investigations of antihemostatic medicinal products in non-surgical patients. J Thromb Haemost. 2005;3(4):692-694. 16. Ortel TL, Neumann I, Ageno W, et al. American Society of Hematology 2020 guidelines for management of venous thromboembolism: treatment of deep vein thrombosis and pulmonary embolism. Blood Adv. 2020;4(19):4693-4738.

haematologica | 2022; 107(5)

ARTICLE

Non-Hodgkin Lymphoma

Toxicity and efficacy of chimeric antigen receptor T-cell therapy in patients with diffuse large B-cell lymphoma above the age of 70 years compared to younger patients – a matched control multicenter cohort study Ron Ram,1,2 Sigal Grisariu,3 Liat Shargian-Alon,2,4 Odelia Amit,1,2 Yaeli Bar-On,1,2 Polina Stepensky,3 Moshe Yeshurun,2,4 Batia Avni,3 David Hagin,2,5 Chava Perry,1,2 Ronit Gurion,2,4 Nadav Sarid,1,2 Yair Herishanu,1,2 Ronit Gold,1 Chen Glait-Santar,1,2 Sigi Kay1,2 and Irit Avivi1,2

Ferrata Storti Foundation

Haematologica 2022 Volume 107(5):1111-1118

BMT Unit, Tel Aviv (Sourasky) Medical Center, Tel Aviv; 2Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv; 3Department of Bone Marrow Transplantation Hadassah University Hospital and The Hebrew University of Jerusalem, Faculty of Medicine, Jerusalem; 4Institute of Hematology, Rabin Medical Center, Petach Tikva and 5 Immunology Unit, Tel Aviv (Sourasky) Medical Center, Tel Aviv, Israel 1

ABSTRACT

ata regarding efficacy and toxicity of chimeric antigen receptor T (CAR-T) cell therapy in the elderly, geriatric population are insufficient. In 2019, tisagenlecleucel and axicabtagene-ciloleucel were commercially approved for relapsed/refractory diffuse large B-cell lymphoma. From May 2019 onwards, 47 relapsed/refractory diffuse large Bcell lymphoma patients, ≥70 years underwent lymphopharesis in three Israeli centers. Elderly (n=41, mean age 76.2 years) and young (n=41, mean age 55.4 years) patients were matched based on ECOG performance status and lactose dehydrogenase levels. There were no differences in CD4/CD8 ratio (P=0.94), %CD4 naive (P=0.92), %CD8 naive (P=0.44) and exhaustion markers (both HLA-DR and PD-1) between CAR-T cell products in both cohorts. Forty-one elderly patients (87%) received CAR-T cell infusion. There were no differences in the incidence of grade ≥3 cytokine-release-syndrome (P=0.29), grade≥3 neurotoxicity (P=0.54), and duration of hospitalization (P=0.55) between elderly and younger patients. There was no difference in median D7-CAR-T cell expansion (P=0.145). Response rates were similar between the two groups (complete response 46% and partial response 17% in the elderly group, P=0.337). Non-relapse mortality at 1 and 3 months was 0 in both groups. With a median follow-up of 7 months (range, 1.3-17.2 months), 6- and 12-months progression-free and overall survival in elderly patients were 39% and 32%, and 74% and 69%, respectively. EORTC QLQ-C30 questionnaires, obtained at 1 month, showed worsening of disability and cancer-related-symptoms in elderly versus younger patients. We conclude that outcomes of CAR-T cell therapy are comparable between elderly, geriatric and younger patients, indicating that age as per se should not preclude CAR-T cell administration. Longer rehabilitation therapy is essential to improve disabilities and long-term symptoms.

Introduction The median age of diagnosis for diffuse large cell B-cell lymphoma (DLBCL), the most common subtype of aggressive lymphoma, is 66 years with approximately 40% of the patients above the age of 70 years (SEER cancer statistics). Advanced age is a major risk factor for relapse and death in patients with DLBCL.1 A recent study, evaluating the real-world outcome of DLBCL patients, reported a 52% complete remission rate in patients older than 65 years. However 22% of these complete responders subsequently experienced disease relapse, indicating that almost two thirds of elderly DLBCL patients will eventually require salvage therapy.2 Although

haematologica | 2022; 107(5)

Correspondence: RON RAM ronram73@gmail.com Received: January 8, 2021. Accepted: July 1, 2021. Pre-published: July 8, 2021. https://doi.org/10.3324/haematol.2021.278288

1111

R. Ram et al.

for selected patients both allogeneic and autologous hematopoietic cell transplantations are curative therapeutic options, for elderly patients who are non-transplant eligible options are limited, and those who fail to respond/relapse following first line therapy, die of the disease.3,4 The introduction of chimeric antigen receptor T (CAR-T) cell therapy, providing long-term remission in 30-40% of patients,5,6 appears to be the most powerful, if not the only potentially curative therapy for these relapsed/refractory (R/R) elderly DLBCL patients. Nevertheless, elderly patients (>65-70 years old) are often considered ineligible for CAR-T cell therapy; having a relatively poor performance status (PS) and concomitant comorbidities, making them more susceptible for treatment-related adverse events , particularly cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS).7 A recent analysis, evaluating the outcome of young versus elderly patients (age >65 years, n=27), included in the ZUMA-1 trial after fulfilling the inclusion criteria (e.g., Eastern Cooperative Oncology Group performace status [ECOG PS]= 0-1, left ventricular ejection fraction [LVEF] >50%), reported similar rates of CRS, complete remission (CR) and long-term remissions in both age groups. However, the incidence of ICANS grade 3, was higher in elderly patients.8 In addition, encouraging response rates were also reported in several small retrospective studies, though toxicity profile remained debatable.9-11 Hence, we aimed to retrospectively analyze the realworld data of the efficacy and toxicity profile of a nonselective population of elderly DLBCL patients treated with CAR-T cell therapy, compared with matched younger patients. In addition, we focused on both clinical parameters and quality of life domains, as well as on T-cell subpopulations and fitness of elderly patients, compared to younger counterparts.

Methods Since April 2019 tisagenlecleucel has been commercially available in Israel, while axicabtagene ciloleucel has been available since April 2020. The national infrastructure of eligibility for CART cell therapy does not require a centralized committee to approve the treatment. Each center has been approved by the relevant pharmaceutical companies, the MOH, and in some cases by JACIE for the infusion of CAR-T cells. Among all three accredited centers in Israel, we retrospectively searched the DLBCL CAR-T surveillance database for all patients referred for CD19-directed CAR-T. The study was performed in accordance with the Declaration of Helsinki and was approved by the institutional ethics committee. For referral and eligibility, lymphopheresis, bridging therapy, and preparative regimen and supportive care sections see the Online Supplementary Appendix.

Definitions of endpoints Microbiological and clinical documented infections (MDI and CDI, respectively) were defined according to the European Conference on Infections in Leukemia (ECIL) guidelines12 and organ dysfunction was defined as either congestive heart failure, acute kidney injury, or atrial fibrillation. Adverse events were graded according to the National Cancer Institute Common Toxicity Criteria Version 5.0. Patients were monitored daily for the occurrence of CRS and ICANS. Grading and treatment followed the American Society

1112

for Transplantation and Cellular Therapy (ASTCT) and European Society for Blood and Marrow Transplantation (EBMT) recommendations.7,13 Briefly, tocilizumab was given in the context of fluid-resistant hypotension grade 2 CRS or low saturation and steroids were started in cases of tocilizumabrefractory CRS or ICANS grade 2 or higher. Disease status was evaluated by Positron emission tomography/computed tomography (PET-CT) scan, performed within 7 days prior to admission for CAR-T cell therapy, and on day 30 and 90 post CAR-T cell infusion. Following white blood cell count recovery, patients carried out a weekly full blood count, and monthly cytomegalo virus (CMV) and Kaposi's sarcoma-associated herpes virus 6 (HHV6), and immunoglobulins status for the first year. The cell therapy coordinator nurse assessed quality of life prior to, 30 days after and 90 days after infusion using the EORTC QLQC30 (version 3) questionnaires (including the following domains – disability assessment, cancer/toxicity-associated symptoms, emotional symptoms, overall health self-assessment, and overall quality of life self-assessment). For evaluation of pretreatment T-cell compartment and assessment of CAR-T cell product and persistency see the Online Supplementary Appendix.

Statistical analysis All consecutive patients ≥70 years (study cohort) were matched with patients younger than 70 years (control group). Patients were identified from a surveillance database of the participating centers. Matching was performed according to ECOG PS at screening (0-1 vs. 2-3) and lactose dehydrogenase (LDH) blood levels prior the infusion of CAR-T cell product (high vs. normal). Selection of these two parameters was based on previous real-word data, published by two different groups, confirming these parameters to predict patient's outcome.14,15 Continuous variables were described as the mean, median, standard deviation and range of number observations, as applicable. Categorical data were described with contingency tables including frequency and percent. Comparison between the different baseline domains of the study cohort and counterpart control cohort was performed using wither Pearson Chi-Square or non-parametric Student t-test, as appropriate. One-way ANOVA test with F calculation was performed to compare the quality-of-life questionnaire domains between base line, 30 days and 3 months values. A linear regression was performed for the association between baseline characteristics and response to CAR-T cell infusion. Cox proportional-hazards model was performed to identify parameters associated with progression-free survival (PFS) or overall survival (OS). Status of disease at 1 month post CAR-T cell therapy was analyzed as a time-dependent covariate. A two-sided P-value of <0.05 was considered as statistically significant.

Results Between April 2019 and October 2020, 49 patients, ≥70 years of age were screened for eligibility in three CAR-T cell therapy centers in Israel. Two patients were ineligible for CAR-T cell collection (ECOG=4, n=1; active hepatitis B virus infection, n=1). Forty-seven patients (96%) were eligible and all underwent successful apheresis. Median time from referral to apheresis was 11 days (range, 1-71 days), with no difference in time to apheresis between elderly and younger cohorts (mean days from referral to apheresis 18.8 vs. 15.4, respectively, P=0.453). In six patients (12.7%), CAR-T manufactured cell product was not eventually infused (out of specification and terminahaematologica | 2022; 107(5)

CAR-T for elderly patients with DLBCL

tion, n=3; out of specification and progression of secondary malignancy, n=1; progression of DLBCL, n=2). One patient had an out of specification product due to low viability of CAR-T cells, however the product was infused. Forty-one patients were given a commercial CAR-T cell product (87% out of all patients that underwent apheresis) and encompassed the study population. Forty-one matched patients aged <70 years were included in the control group. Table 1 depicts the characteristics of the two cohorts. The mean age of the study and the control cohorts were 76.2 (±4.4) and 55.4 (±15) years, respectively and the median follow-up of surviving patients was 7

months (range, 1.3-17.2 months) and 7 month (range, 1.316.7months), respectively. Similar percentage of patients in both cohorts had >2 lines of therapy prior to CAR-T cell therapy (47.3% and 51.2%, respectively). Percentage of patients that had a previous autologous hematopoietic cell transplantation (HCT) and CR state at CAR-T cell infusion was higher in the control group compared to the older aged group. While, overall, the percentage of patients with ECOG PS 2-3 was similar between the two groups (61%), there was a lower percentage of patients with ECOG PS 3 in the study cohort compared to the control group (12.2% vs. 39%), Table 1.

Table 1. Characteristics of patients

Domain Age in years, mean (± S.D.) Sex – Female Product Tisa-cel Axi-cel Transformed indolent lymphoma Non-GCB subtype ECOG performance status 0 1 2 3 Specific comorbidities Ischemic heart disease Hypertension Diabetes mellitus Smoker Chronic kidney disease Cerebrovascular attack Dementia N lines prior to CAR-T 2 3 4 5 >5 Previous autologous transplantation Days from referring to collection, mean (±S.D.) N cycles of collection, mean (±S.D.) Collection efficiency, mean (±S.D.) Days from collection to pick-up Bridging to CAR-T infusion None/steroids only Chemotherapy ± radiation Radiation Novel agents* Status prior to CAR-T infusion Complete remission Partial remission Stable disease Progressive disease Not evaluated Days from apheresis to CAR-T infusion, mean (±S.D.) High LDH prior to CAR-T infusion LDH (in patients with elevated values) - median, range (U/L)

Study Cohort (n=41)

Control (n=41)

P-value

76.2 (4.4) 24, 61%

55.4 (15) 23, 54%

33 (80.5%) 8 (19.5%) 11 (26.8%) 25 (61%)

34 (82.9%) 7 (17.1%) 8 (19.5%) 21 (51%)

<0.001 0.674 0.775

3 (7.3%) 13 (31.7%) 20 (48.8%) 5 (12.2%)

1 (2.4%) 15 (36.6%) 9 (22%) 16 (39%)

10 (24%) 14 (34%) 7 (17%) 3 (7.3%) 4 (9.7%) 3 (7.3%) 1 (2.4%)

7 (17%) 8 (20%) 9 (22%) 4 (9.7%) 2 (4.9%) 1 (2.4%) 1(2.4%)

22 (53.7%) 13 (31.7%) 5 (12.2%) 0 (0%) 1 (2.4%) 3 (7.3%) 18.8 (11.3) 2.46 (0.64) 52.9 (15) 4.1 (3.6)

20 (48.8%) 11 (26.8%) 4 (9.8%) 2 (4.9%) 4 (9.8%) 14 (34.1%) 15.4 (8.9) 2.4 (0.75) 55.3 (13) 5.5 (4.5)

7 (17.1%) 29 (70.7%) 3 (7.3%) 2 (4.9%)

12 (29%) 23 (56%) 5 (12.2%) 1 (2.8%)

7 (8.5%) 30 (34.1%) 3 (7.3%) 20 (48.8%) 3 (7.3%) 36.5 (12) 18 (43.9%) 548 (380-2,041)

6 (14.6%) 6 (14.6%) 11 (26.8%) 15 (36.6%) 3 (7.3%) 38.7 (12) 18 (43.9%) 575 (382-1,891)

0.432 0.29 0.187

0.41 0.14 0.58 0.69 0.39 0.61 0.61 0.453

0.003 0.453 0.696 0.530 0.232 0.417

0.017

0.453 1 0.65

GCB: germinal center B cell; LDH: lactate dehydrogenase; S.D.: standard deviation.

haematologica | 2022; 107(5)

1113

R. Ram et al.

Apheresis There was no difference in the mean number of collection cycles and in the collection efficiency between the two groups (2.46 vs. 2.4, P=0.696 and 52.9 vs. 55.3, P=0.53, respectively). Age (as a continuous variable), advanced ECOG PS, number of lines of therapy prior to lymphopheresis, and previous autologous HCT did not impact the required number of collection cycles (β=0.03, 95% confidence interval[CI]: -0.28 to 0.36, P=0.801; β=0.18, 95% CI: -0.06 to 0.54, P=0.114; β=-0.16, 95% CI: -0.02 to 0.36, P=0.172; and β=0.13, 95% CI: -0.2 to 0.64, P=0.305, respectively). Data of T-cell subpopulations were available in 19 (46%) patients and were compared to results obtained in younger patients (n=16, 39%), (Figure 1A to D; Online Supplementary Figure S1). There was no difference in the CD4/CD8 ratio between the two groups (P=0.94). There were no differences in percentages of naïve, TCM, TEM,

and TEMRA-CD4 subpopulations in the apheresis product, between the two groups (P=0.92, P=0.35, P=0.45, and P=0.16, respectively). This was also true for the counterparts, CD8 subpopulations (P=0.44, P=0.35, P=0.33, and P=0.47, respectively), Figure 1A and B. There was also no difference in the expression of exhaustion markers by both CD4 and CD8 cells between the apheresis products of aged versus younger patients (P=0.172 for CD4-HLA-DR, P=0.244 for CD4-PD-1, P=0.06 for CD8-HLA-DR, and P=0.354 for CD8-PD-1).

Hospitalization and early toxicity Mean days from apheresis to CAR-T cell infusion was 36.5 (standard deviation [SD] ±12) , with no difference between the elderly and the control group (P=0.453), Table 2. Analysis of the CAR-T cell product showed that there were no differences in the percentages of CD4 naive, TCM,

1114

Figure 1. Subpopulations of T cells in the apheresis and in the CAR-T cell therapy product from a portion of the study group (n=19) and control group (n=16). (A) Subpopulations of CD4 T cells in apheresis product; (B) subpopulations of CD8 T cells in apheresis product; (C) subpopulations of CD4 T cells in the CAR-T cell therapy product; (D) subpopulations of CD8 T cells in the CAR-T cell therapy product.

haematologica | 2022; 107(5)

CAR-T for elderly patients with DLBCL

TEM, and TEMRA subpopulations between the elderly and the younger-patients’ group (0.48 vs. 0.29, P=0.59; 15.9 vs. 31.5, P=0.39; 82.3 vs. 67.8, P=0.42; and 1.18 vs. 0.42, P=0.42, respectively). This was also true for the comparable CD8 subpopulations (0.67 vs. 0.39, P=0.37; 13 vs. 30.2, P=0.24; 84.7 vs. 70, P=0.28; and 1.75 vs. 0.49, P=0.32, respectively), Figure 1C and D. There was also no difference in the expression of exhaustion markers expressed by CD4 and CD8 cells (69.6 vs. 83.4, P=0.39 for CD4-HLA-DR, 81.1 vs. 90.6, P=0.22 for CD4-PD-1, 86.1 vs. 92.5, P=0.30 for CD8HLA-DR, and 63.7 vs. 72.4, P=0.36 for CD8-PD-1). Among the elderly group, there were six cases (15% of patients) of CDI (pneumonia, n=3; cellulitis, n=2, and lineassociated tunnel infection, n=1). There were four cases (10%) of MDI (gram negative bacteremia, n=2 and staphylococcus hominis bacteremia, n=2). Incidence of overall CDI and MDI infections were similar between the elderly and the control groups (26.8% and 19.5%, respectively, P=0.301). There were no cases of exacerbation of congestive heart failure. There were three cases of acute kidney injury (2 associated with preparative regimen and 1 associated with sepsis) and three cases of atrial fibrillation (all associated with ongoing CRS). Incidence of organ failure was also similar between the two groups (P=1). Median time to CRS was in the elderly cohort was 3 days (range, 0-6 days), similar to the control group (median 3 days [range, 1-6 days], P=0.65). Overall, there were 28 cases (68%) of CRS (grade 1, n=9; grade 2, n=15, and grade 3, n=4). There was no difference between the elderly and the control group in overall CRS (69.3% in both groups, P=0.88) and grade 3-4 CRS (9.8% vs. 7.3%, respectively, P=0.29). Median days to ICANS was 4 days (range, 2-8 days). Overall, there were 11 cases (27%) of ICANS (grade 1, n=5; grade 2, n=5, and grade 3, n=1). There was no difference between the elderly and the control group in overall ICANS (27.5% vs. 17.1%. respectively, P=0.48) and grade 34 ICANS (2.5% vs. 4.9%, respectively, P=0.54). History of vascular disease or dementia did not predict occurrence of ICANS in the elderly group (overall response [OR] 1.2, 95% CI: 0.78-1.81, P=0.45 and OR=1.4, 95% CI: 0.81-1.63, P=0.38, respectively). Mean doses of tocilizumab/patient was in the two groups (1.5 vs. 0.9, respectively, P=0.484). Similarly, the percentage of patients given steroids was similar (32.5% vs. 24.4%, respectively, P=0.596). Nine patients (22%) required granulocyte colony-stimulating factor (GCSF) on day 14 post CAR-T cell infusion due to delayed count recovery, no difference was found in with the control group (P=0.15). Mean days of admission was 23.4 (±8) days, compared to 24.6 (±9.6) in the control group (P=0.55).

Late toxicity Late pancytopenia occurred in nine patients (absolute neutrophil count [ANC] only, n=2 [4.9%], platelet count only, n=2 [4.9%], and ≥ 2 cytopenia, n=5 [12%]). No difference was found in the incidence of late cytopenia between the two groups (P=0.399). There were five (15.2% of 32 patients with available data) patients with CMV reactivation, none had CMV disease. Two patients with continuous CMV viremia eventually received an anti-CMV therapy (1 valgancyclovir and 1 foscarnet) with no subsequent reappearance of CMV. There were three (15% of 20 patients with available data) patients that developed reactivation of HHV6. Of them, one patient had ongoing grade 3 ICANS. In this patient, HHV6-polymerase chain reaction (PCR) obtained from the cerebrospinal fluid was positive, howevhaematologica | 2022; 107(5)

er both electroencephalogram and magnetic resonance imaging were not suggestive of limbic encephalitis. The patient was treated with a 3-week course of foscarnet and steroids and subsequently recovered, however we were unable to conclude if this was a definite HHV6 encephalitis. The other two patients did not have clinical symptoms of HHV6 systemic infection or encephalitis, therefore, were only monitored until HHV6-PCR levels became negative. Out of 21 patients with available immunoglobulin G (IgG) levels 1 month post CAR-T cell infusion, there were five cases (24%) with IgG levels below 400 mg/L, similar to the control group (P=0.398), Table 2.

Quality of life In 23 patients (56%) EORTC QLQ-C30 questionnaires were available. Thirty-day questionnaire, compared to the baseline questionnaire, showed increased disability in four of five domains, increase in cancer/treatment-related symptoms in six of 11 domains and worsening of emotional Table 2. Toxicity and response to CAR-T cell therapy

Domain

Study Cohort (n=41)

Cytokine release syndrome 0 13 (31.7%) 1 9 (22%) 2 15 (36.6%) 3 4 (9.8%) 4 0 (0%) Immune effector cell-associated neurotoxicity syndrome 0 29 (72.5%) 1 5 (12.5%) 2 5 (12.5%) 3 1 (2.5%) 4 0 (0%) N Tocilizumab (dose/patient) 1.5 Patients given steroids 14 (32.5%) Need for GCSF on day 14 9 (22%) Early infections CDI 5 (12%) MDI 6 (14.8%) Organ dysfunction Congestive heart failure 0 (0%) Atrial fibrillation 3 (7.3%) Acute kidney injury 3 (7.3%) Days of hospitalization 23.4 (8) Late pancytopenia 9 (22%) IgG levels < 4 gr/L* 5 (23.8%) Reactivation of CMV** 6 (18.8%) Reactivation of HHV6*** 3 (15%) 1- month PET/CT results CR 19 (46%) PR 7 (17%) PD 13 (32%) Progression-free survival (6 months) 39% Overall survival (6 months) 74%

Control (n=41) 13 (31.7%) 7 (17.1%) 18 (43.9%) 3 (7.3%) 0 (0%)

34 (82.9%) 3 (7.3%) 2 (4.9%) 2 (4.9%) 0 (0%) 0.9 10 (24.4%) 15 (36.9%) 0 8 (19.5%) 0 (0%) 3 (7.3%) 3 (7.3%) 24.6 (9.6) 11 (26.8%) 5 (33%) 5 (15.2%) 1 (4.3%) 24 (59%) 8 (19%) 9 (22%) 54% 76%

P-value 0.881

0.475

0.484 0.258 0.112 0.301

0.547 0.399 0.398 0.234 0.09 0.337

0.209 0.792

GCSF: granulocyte colony stimulating factor; CDI: clinical documented infections; MDI: microbiology documented infections; Ig: immunoglobulin; CMV: cytomegalovirus; HHV6: human herpes virus 6; PET/CT: positron emmission tomography/computerized tomography; CR: complete remission; PR: partial remission; PD: disease progression. * Out of 21 patients in the study group and 15 patients in the control group. ** Out of 32 patients in the study group and 33 patients in the control group. *** Out of 20 patients in the study group and 23 patients in the control group.

1115

R. Ram et al.

symptoms in four of 12 domains, while there was no change in both overall health perception and overall quality of life. The 3-month questionnaire, when compared to the base-line questionnaire, showed improvement in disability in all five domains, improvement in all cancer/treatmentrelated symptoms, improvement of emotional symptoms in ten of 12 domains and improvement both overall health perception (mean baseline 3.83 vs. mean 3 months 5.6, F6.007, P=0.005) and overall quality of life (mean baseline 3.87 vs. mean 3 months 5.4, F-2.68, P=0.081).

Efficacy At date of analysis there were 31 patients (76%) alive and 16 patients (39%) in an ongoing CR state. Non-relapse mor-

tality at 1 and 3 months was 0. Expansion of CAR-T cells on day 7 was available in 19 (46%) patients. There was no difference in the CAR-T cell blood levels between elderly and the control group (P=0.145). PET-CT at 1 month post CAR-T cell infusion demonstrated CR, partial remission (PR) and progressive disease (PD) in 19 (46%), 7 (17 %), and 13 (32%) patients, respectively. In two patients the results of PET-CT are still pending. There was no difference in the overall response rate (ORR) between the elderly and the control group (63% vs. 78%, respectively, P=0.337). Multivariate binary logistic model identified that high LDH prior to admission for CAR-T cell infusion was associated with lower chances of achieving CR state at day 30 post CAR-T cell infusion (OR: -0.8, 95% CI: 0.48-0.97,

Figure 2. Progression-free survival. Progressionfree survival after CAR-T infusion of elderly versus young patients with diffuse large cell B-cell lymphoma.

Figure 3. Overall survival. Overall survival after CAR-T infusion of elderly versus young patients with diffuse large cell B-cell lymphoma.

1116

haematologica | 2022; 107(5)

CAR-T for elderly patients with DLBCL

P=0.048). Age, sex, ECOG PS, administration of bridging therapy and the occurrence of CRS had no impact on CR rate. At 3, 6, and 12 months, the projected PFS was 51%, 39%, and 32%, respectively in the elderly group, and approached 67%, 54%, and 54%, respectively in the younger group). Median PFS was 3.6 (95% CI: 1.6-5.6) months in the elderly patients, while not reached in the younger group of patients (P=0.209), Figure 2. Multivariate regression model identified prior autologous HCT (hazard ratio [HR]: 0.208, 95% CI: 0.05-0.87, P=0.032) and a CR state 1 month post CAR-T cell therapy (HR: -0.076, 95% CI: 0.025-0.23, P<0.001) to be associated with better PFS, while both high LDH prior infusion of CAR-T cells (HR: -2.7, 95% CI: 1.1-6.7, P=0.03) and non-germinal center B-cell immunophenotype (HR: 2.29, 95% CI: 0.98-5.4, P=0.057) were associated with shorter PFS. Age and ECOG PS had no impact on PFS. At 3, 6, and 12 months, projected OS was 84%, 74%, and 69%, respectively in the elderly group, 87%, 76%, and 53% in the younger group. Median OS was not reached in both groups (P=0.792), Figure 3. Multivariate regression model for OS identified a CR state 1 month post CAR-T cell therapy (HR: 0.017, 95% CI: 0.003-0.11, P<0.001) to be associated with increased OS, while both receiving bridging therapy and high LDH prior to CAR-T cell nfusion (HR: 5.6, 95% CI: 0.99-32.3, P=0.05 and HR: 4.9, 95% CI: 1.5-16.1, P=0.008, respectively) were associated with shorter OS. Age and ECOG PS had no impact on OS.

Discussion This study explored the characteristics and outcome of all consecutive patients, 70 years or older, referred for CAR-T cells during the study period, and matched these patients with younger individuals, based on PS and LDH prior to admission for infusion. We showed that both toxicity and efficacy are similar in elderly compared to the younger patients. Despite the improvement in outcomes of elderly DLBCL patients over the last two decades, the prognosis remains disappointing, with 2-year event-free survival of 57%.16 Unfortunately, response rates achieved with second line salvage therapy in these non-transplant eligibility patients are generally dismal and most patients die of their disease in less than a year.17,18 Therefore, CAR-T cell therapy, providing the only curative option for these elderly patients, should be strongly considered. At present, many countries employ very restrictive criteria for selecting patients to CAR-T cell therapy, considering CAR-T cell therapy as a "highly toxic" treatment.15,19,20 According to their policies, patients that present with significant comorbidities (i.e., <45% EF, PS>1, chronic renal disease etc.) are ineligible for CAR-T cell therapy. Thus, a substantial proportion of elderly patients, "fulfilling these criteria", would be ineligible for CAR-T cell therapy. Indeed, the French group excluded 41% of all referred DLBCL patients due to employment of these restrictive standards. As such, the median age of CAR-T cell therapy patients in some centers in Europe is lower than 60 years, indicating that the majority of elderly DLBCL patients that experienced disease relapse and required CAR-T cell therapy, were not considered for this life saving treatment.20,21 According to our data, almost all (96%) elderly patients that were referred to CAR-T cell therapy, were found to be haematologica | 2022; 107(5)

eligible and underwent successful apheresis. although analyzed in only a portion of the patients in the elderly group," lymphocyte fitness", reflected by T-cell subsets and exhaustion markers, was non-inferior in elderly compared with younger patients, and translated into comparable CAR-T cell products with no increase in production failure and similar expansion of CAR-T cells. These results are analogous to those reported in the ZUMA-1 trial.8 Despite the fact that two thirds of our patients had a poor PS and almost half entered CAR-T cell therapy with increased LDH levels (reflecting highly proliferative disease), 84% were eventually transfused. This transfusion rate is not inferior to that reported in other real life series, including highly selective CAR-T cell programs, emphasizing the importance of logistical factors and decreasing the time from enrollment to infusion.14,20 Acute treatment related toxicities, including CRS and ICANS, were mainly grades 1-2, with no differences between the cohorts as previously proposed by others.10 The relatively low incidence of high grade ICANS in both young and elderly patients, might reflect the predominant employment of tisagenlecleucel in our cohort of patients. In contrast, studies focusing on early toxicity of axicabtagene, have generally reported higher rates of high grades ICANS, especially in older aged patients.8,22 Of note, prior cerebral vascular disease nor dementia were found to be associated with increased risk for ICANS in our cohort of patients. Long-term toxicity was not higher in elderly versus younger patients. Interestingly, a substantial number of patients developed reactivation of CMV, and rarely, HHV6, with no need for active intervention in the majority, and with no infection-related sequels. This manifestation has not been reported yet and we continue to perform routine surveillance monitoring to further investigate post CAR-T cell therapy virus reactivation. The risk for significant hypogammaglobulinemia and persistent cytopenia were not higher in the aged patient and were in line with those reported in prior real-life series.6-8 An important finding of our study is the dual effect of CAR-T cell therapy on patients’ quality of life, during admission and after returning home. During hospitalization, patients reported on increase disability, aggravation in cancer/treatment-related symptoms and worsening emotional distress. Considering that patients are usually hospitalized for 3-4 weeks, direct intervention during that period, employing intensive physical and emotional rehabilitation programs may help alleviating symptoms. Thereafter, those patients who respond to CAR-T cell therapy may expect to gain a long-term clinically meaningful improvements in daily functioning.23 Indeed, in our cohort of patients, all symptoms improved after several months, and patients reported a significant progress in their quality of life. Despite the poor features of our elderly cohort of patients (poor PS in 66%, high LDH in approximately 40%, progressive disease and ≥3 prior lines in almost half), the CR rate was 46.3%, similar to the CR rate reported in the Juliet study and in other real-world series.6,8,10 CR rate was affected by LDH level only, and was not adversely affected by age, supporting prior smaller unmatched studies, that reported similar overall and complete response rates in elderly versus younger patients.8 PFS in these elderly patients was also not statistically shorter than in their younger counterparts and was similar to that reported in the Juliet study.6 In line with previous reports, cell of origin 1117

R. Ram et al.

(non-GCB vs. GCB ) and high LDH level, being surrogates for a more aggressive disease, were found to be associated with shorter PFS.14,15 In contrast, prior autograft (indicating the lack of primary refectory disease) and the achievement of CR at 30 days post CAR-T cell therapy predicted durable responses, supporting prior reports.14 The OS in our cohort of patients was also encouraging, approaching 69% at 12 months post CAR-T transfusion. In consensus with prior reports, administration of bridging therapy and high LDH levels were both found to be associated with shorter survival, whereas an achievement of CR at days 30 post CAR-T cell therapy predicted a longer survival.14,15 Our study has several limitations, mainly due to the retrospective nature and the relatively small number of patients. In several analyses, data on potential significant factors were missing for all or for some of the patients, e.g., molecular characteristics and T-cell subsets in apheresis bags. Moreover, there were other imbalances between the two groups (i.e., higher percentage of patients with ECOG PS of 3 in the control group and the number of lines prior to CAR-T cell therapy ), which might affect our results and make it difficult to conclude if elderly patients have a worse or a similar outcome to their younger counterparts. Although this cohort represent non-selected real-life elderly lymphoma patients, it is possible that presumably “ineligible” patients were not referred for CAR-T cell therapy and thus our population is still relatively “selective”.

References 1. Hedstrom G, Hagberg O, Jerkeman M, Enblad G. The impact of age on survival of diffuse large B-cell lymphoma - a populationbased study. Acta Oncol. 2015;54(6):916-923. 2. Adiyaman SC, Alacacioglu A, Ersen Danyeli A, et al. Prognostic factors in elderly patients with diffuse large B-cell lymphoma and their treatment results. Turk J Haematol. 2019;36(2):81-87. 3. Lazarus HM, Carreras J, Boudreau C, et al. Influence of age and histology on outcome in adult non-Hodgkin lymphoma patients undergoing autologous hematopoietic cell transplantation (HCT): a report from the Center For International Blood & Marrow Transplant Research (CIBMTR). Biol Blood Marrow Transplant. 2008;14(12):1323-1333. 4. Shah NN, Ahn KW, Litovich C, et al. Outcomes of Medicare-age eligible NHL patients receiving RIC allogeneic transplantation: a CIBMTR analysis. Blood Adv. 2018;2(8):933-940. 5. Locke FL, Ghobadi A, Jacobson CA, et al. Long-term safety and activity of axicabtagene ciloleucel in refractory large B-cell lymphoma (ZUMA-1): a single-arm, multicentre, phase 1-2 trial. Lancet Oncol. 2019;20(1):3142. 6. Schuster SJ, Bishop MR, Tam CS, et al. Tisagenlecleucel in adult relapsed or refractory diffuse large B-cell lymphoma. N Engl J Med. 2019;380(1):45-56. 7. Yakoub-Agha I, Chabannon C, Bader P, et al. Management of adults and children undergoing chimeric antigen receptor T-cell therapy: best practice recommendations of the European Society for Blood and Marrow Transplantation (EBMT) and the Joint Accreditation Committee of ISCT and EBMT (JACIE). Haematologica. 2020;105

1118

Nevertheless, in order to overcome the potential bias of age we performed matched control cohort analysis with younger counterparts and indeed showed that age in itself should not preclude elderly patients from undergoing CAR-T cell therapy. Ideally, a geriatric scoring assessment should be employed and help to establish eligibility criteria for CAR-T cell therapy in this aged group pf patients. Future studies should focus on improving the long-term efficacy of the product as well as earlier intervention in cases of progression and active rehabilitation to improve quality of life post therapy in this population. Disclosures RR has received honoraria from Novartis and Gilead and has participated in advisory board meeting of Novartis. Contributions RR and IA conceived the presented research; RR, SG, LSS, OA, YBO, PS, MY, BA, CP, RG, NS, YH, RG and IA performed the clinical activity and collected the clinical data; RR, SG and LSA verified the data; DH, CGS and SK performed the laboratory tests and analyses. All authors discussed the results and contributed to the final manuscript. Funding None of the authors received financial support for the research, authorship, and/or publication of this article.

(2):297-316. 8. Neelapu SS, Jacobson CA, Oluwole OO, et al. Outcomes of older patients in ZUMA-1, a pivotal study of axicabtagene ciloleucel in refractory large B-cell lymphoma. Blood. 2020;135(23):2106-2109. 9. Gajra A, Zettler ME, Phillips EG, Jr., et al. Neurological adverse events following CAR T-cell therapy: a real-world analysis. Immunotherapy. 2020;12(14):1077-1082. 10. Lin RJ, Lobaugh SM, Pennisi M, et al. Impact and safety of chimeric antigen receptor T cell therapy in older, vulnerable patients with relapsed/refractory large b-cell lymphoma. Haematologica. 2020;106(1):255-258. 11. Zettler ME, Feinberg BA, Phillips EG, Jr., Klink AJ, Mehta S, Gajra A. Real-world adverse events associated with CAR T-cell therapy among adults age ≥65 years. J Geriatr Oncol. 2021;12(2):239-242. 12. Maertens J, Marchetti O, Herbrecht R, et al. European guidelines for antifungal management in leukemia and hematopoietic stem cell transplant recipients: summary of the ECIL 3--2009 update. Bone Marrow Transplant. 2011;46(5):709-718. 13. Lee DW, Santomasso BD, Locke FL, et al. ASTCT consensus grading for cytokine release syndrome and neurologic toxicity associated with immune effector cells. Biol Blood Marrow Transplant. 2019;25(4):625638. 14. Nastoupil LJ, Jain MD, Feng L, et al. Standard-of-care axicabtagene ciloleucel for relapsed or refractory large B-cell lymphoma: results from the US Lymphoma CAR T Consortium. J Clin Oncol. 2020;38(27):31193128. 15. Vercellino L, Di Blasi R, Kanoun S, et al. Predictive factors of early progression after CAR T-cell therapy in relapsed/refractory diffuse large B-cell lymphoma. Blood Adv.

2020;4(22):5607-5615. 16. Coiffier B, Lepage E, Briere J, et al. CHOP chemotherapy plus rituximab compared with CHOP alone in elderly patients with diffuse large-B-cell lymphoma. N Engl J Med. 2002;346(4):235-242. 17. Crump M, Neelapu SS, Farooq U, et al. Outcomes in refractory diffuse large B-cell lymphoma: results from the international SCHOLAR-1 study. Blood. 2017;130(16): 1800-1808. 18. Gisselbrecht C, Van Den Neste E. How I manage patients with relapsed/refractory diffuse large B cell lymphoma. Br J Haematol. 2018;182(5):633-643. 19. Greinix HT, Attarbaschi A, Girschikofsky M, et al. Ensuring center quality, proper patient selection and fair access to chimeric antigen receptor T-cell therapy: position statement of the Austrian CAR-T Cell Network. memo. 2020;13(1):27-31. 20. Kuhnl A, Roddie C, Martinez-Cibrian N, et al. Real-world data of high-grade lymphoma patients treated with CD19 CAR-T in England. Blood. 2019;134(Suppl 1):S767. 21. Paillassa J, Di Blasi R, Chevret S, et al. CD19 CAR T-cell therapy in patients with relapse/refractory DLBCL: retrospective analysis of the eligibility criteria. Blood. 2019;134(Suppl 1):S2887. 22. Riedell PA, Walling C, Nastoupil LJ, et al. A multicenter retrospective analysis of clinical outcomes, toxicities, and patterns of use in institutions utilizing commercial axicabtagene ciloleucel and tisagenlecleucel for relapsed/refractory aggressive B-cell lymphomas. Blood. 2019;134(Suppl 1):s1599. 23. Maziarz RT, Waller EK, Jaeger U, et al. Patient-reported long-term quality of life after tisagenlecleucel in relapsed/refractory diffuse large B-cell lymphoma. Blood Adv. 2020;4(4):629-637.

haematologica | 2022; 107(5)

ARTICLE

Non-Hodgkin Lymphoma

Response and resistance to CDK12 inhibition in aggressive B-cell lymphomas

Ferrata Storti Foundation

Jing Gao,1,* Michelle Y. Wang,1,* Yuan Ren,1 Tint Lwin,1 Tao Li,1 Joy C. Yan,1 Eduardo M. Sotomayor,2 Derek R. Duckett,1 Bijal D. Shah,3 Kenneth H. Shain,3 Xiaohong Zhao1 and Jianguo Tao4 1

Chemical Biology and Molecular Medicine Program, H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL; 2Cancer Institute, Tampa General Hospital,Tampa, FL; 3 Departments of Malignant Hematology, H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL and 4(Previous affiliation) Department of Hematopathology and Laboratory Medicine, H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL, USA *JG and MYW contributed equally as co-first authors.

Haematologica 2022 Volume 107(5):1119-1130

ABSTRACT

espite significant progress in the treatment of patients with diffuse large B-cell lymphoma (DLBCL) and mantle cell lymphoma (MCL), the prognosis of patients with relapsed disease remains poor due to the emergence of drug resistance and subsequent disease progression. Identification of novel targets and therapeutic strategies for these diseases represents an urgent need. Here, we report that both MCL and DLBCL are exquisitely sensitive to transcription-targeting drugs, in particular THZ531, a covalent inhibitor of cyclin-dependent kinase 12 (CDK12). By implementing pharmacogenomics and a cell-based drug screen, we found that THZ531 leads to inhibition of oncogenic transcriptional programs, especially the DNA damage response pathway, MYC target genes and the mTOR-4EBP1-MCL-1 axis, contributing to dramatic lymphoma suppression in vitro. We also identified de novo and established acquired THZ531-resistant lymphoma cells conferred by over-activation of the MEK-ERK and PI3K-AKT-mTOR pathways and upregulation of multidrug resistance-1 (MDR1) protein. Of note, EZH2 inhibitors reversed resistance to THZ531 by competitive inhibition of MDR1 and, in combination with THZ531, synergistically inhibited MCL and DLBCL growth in vitro. Our study indicates that CDK12 inhibitors, alone or together with EZH2 inhibitors, offer promise as novel effective approaches for difficult-to-treat DLBCL and MCL.

Introduction MYC is a transcription factor that promotes oncogenesis by activating and repressing its target genes that control cell growth and proliferation.1 While MYC has been described as a defining feature and the driving oncogene for Burkitt lymphoma, the significance of MYC has also been recognized in other non-Hodgkin lymphomas.2 MYC, which has been detected in 15-20% of diffuse large B-cell lymphomas (DLBCL), is associated with an adverse prognosis as a result of chemoresistance and, shortened survival. In mantle cell lymphoma (MCL), increased expression of MYC has been found to be associated with poor prognosis and more aggressive disease.3 MYC overexpression has also been implicated in high-grade large cell transformation.4 Despite current modes of intensive chemotherapy and immunotherapy, MCL and other MYC-associated lymphomas are aggressive diseases that respond poorly to chemoimmunotherapy and affected patients have a dismal survival. Identification of effective strategies to target these aggressive lymphomas represents an urgent need. Recently, we integrated data from our unbiased activity-based proteomic profiling, RNA-sequencing, and chromatin immunoprecipitation (ChIP) assays with sequencing studies. We found that the aggressive progression and development of drug resistance in MCL and MYC-associated lymphomas require complex transcriptome and kinome remodeling of cellular signaling networks, positive feedback loops that amplify pro-survival and growth signals.5 Rather than there being a single mechanism

haematologica | 2022; 107(5)

Correspondence: JIANGUO TAO dtep642021@gmail.com Received: March 10, 2021. Accepted: June 10, 2021. Pre-published: June 24, 2021. https://doi.org/10.3324/haematol.2021.278743

1119

J. Gao et al.

of driving lymphoma progression, networks are rewired and, consequently, the signal rewiring must be targeted as a whole to obtain durable and improved clinical responses. In support of this concept, increasing evidence suggests that super-enhancers are required to maintain the expression of genes critical for cancer cell survival and proliferation. Given the biological importance of super-enhancers in the regulation of global transcriptome landscaping, superenhancers have emerged as enticing therapeutic targets. Indeed, blocking super-enhancer-driven transcription resulted in the concurrent disruption of multiple oncogenic machineries to which aggressive lymphoma cells are addicted, regardless of their mutational profiles.6,7 Thus, small-molecule inhibitors targeting transcriptional regulators and downstream global transcriptome and kinome reprogramming represent a promising approach to the treatment of aggressive lymphomas. Transcriptional cyclin-dependent kinase 7 (CDK7) and 9 (CDK9), catalytic subunits of the transcription elongation factor P-TEFb, are considered to be gatekeepers of the transcriptional machinery,8 and clinical trials with small-molecule inhibitors of these kinases are underway.9 CDK7- and CDK9-mediated phosphorylation of serine on the C-terminal domain of RNA polymerase II (RNAPII) is known to be linked to transcriptional initiation and elongation. Phosphorylation of Ser2 on the C-terminal domain is catalyzed by P-TEFb and is associated with a large complex of proteins coined the super elongation complex.10,11 Moreover, CDK7 directly phosphorylates Ser5 and Ser7 of the C-terminal domain of RNAPII for RNAPII activity. Interestingly, ChIP-sequencing data demonstrated that CDK7 densely occupies super-enhancers that drive high levels of transcription of oncogenes, such as MYC, in a wide variety of cancers, including T-cell acute lymphoblastic leukemia, nonsmall cell lung cancer, neuroblastoma and triple-negative breast cancer.12-14 Importantly, treatment of cancer with a selective, small-molecule CDK7 inhibitor, THZ1, leads to rapid loss of these super-enhancer-driven oncogenic transcripts,15 suggesting that super-enhancer-driven genes are especially vulnerable to inhibition of transcriptional machinery such as through inhibition of CDK7.16 CDK12 belongs to the transcriptional CDK family of serine/threonine protein kinases that regulate transcriptional and post-transcriptional processes, thereby modulating multiple cellular functions. CDK12 modulates transcription elongation by phosphorylating the Ser2 on the C-terminal domain of RNAPII and was demonstrated to specifically upregulate the expression of genes involved in the DNA damage response (DDR), mRNA processing, stress and heat shock.17 Other studies indicated that CDK12 phosphorylates pre-mRNA processing factors directly, thereby inducing premature cleavage and polyadenylation and a loss of expression of long genes (>45 kb), a substantial proportion of which participate in the DDR.18,19 In addition, an increasing number of studies point to CDK12 inhibition as an effective strategy to inhibit tumor growth, and synthetic lethal interactions have been described with MYC, EWS/FLI and PARP/CHK1 inhibition.20,21 Therefore, CDK12 has emerged as an appealing therapeutic target. MCL is a B-cell malignancy in which the disruption of the DDR pathway and activation of cell survival mechanisms contribute to oncogenesis.22 MYC activation engages the replication stress and DNA damage pathway to allow not only robust cellular proliferation, but also limited clonal expansion and avoidance of cytotoxic DNA damage accu1120

mulation in aggressive B-cell lymphomas.23,24 These data implicate CDK12 as a potential novel vulnerability for MCL and MYC-associated large B-cell lymphomas. In this study, we determined the role of CDK12-mediated transcriptional activation and its associated pathway in cell survival and growth in MCL and MYC-associated large Bcell lymphomas. We defined the molecular mechanism for inhibiting CDK12 using THZ531 in these aggressive lymphomas. Importantly, we investigated the molecular mechanism conferring resistance to THZ531 and examined whether combined inhibitors of CDK12 and EZH2 cooperatively reprogram transcriptional repression to overcome THZ531 resistance, and, ultimately, inhibit lymphoma growth and survival in these difficult-to-treat, aggressive Bcell malignancies.

Methods Patients and tumor specimens The primary samples from MCL patients were obtained from fresh biopsy-derived lymphoma tissues (lymph nodes) and from peripheral blood following informed consent from patients and approval by the Moffitt Cancer Center/University of South Florida Institutional Review Board. For preparation of viable, sterile, single-cell suspensions, the lymph node tissue was diced and forced through a cell strainer into RPMI-1640 tissue culture medium. Cells, obtained after low-speed centrifugation were re-suspended in medium. Lymphoma cells from peripheral blood were isolated by Ficoll-Plaque purification, and only lymphoma samples that had more than 80% tumor cells were used for experiments.

Image-based cell-viability assay Cells were seeded in a 384-well plates of a reconstructed lymphoma tumor microenvironemnt, including high physiological densities (1-10x106 cells/mL), extracellular matrix (collagen, Advanced BioMatrix, #5005-B), and lymphoma stromal cells. A panel of drugs at five serial diluted concentrations was added to the medium, and plates were continuously imaged every 30 min for 96 h (cell lines) or 144 h (primary samples). All images were analyzed using a digital imaging analysis algorithm to detect cell viability based on membrane motion (pseudo-colored in green), and changes in viability were quantified by the area under the curve (AUC) as described elsewhere.25-27

RNA-sequencing All samples were prepared in biological triplicates. Cells (10x106) were treated with 100 nM THZ531 or dimethylsulfoxide (vehicle control) for 6 and 24 h. Total RNA was isolated using an RNA isolation kit, RNeasy Plus Mini (Qiagen Cat# 74134). Libraries were prepared using a TruSeq Stranded mRNA Library Prep Kit (Illumina Cat #RS-122-2101/2) according to the manufacturer’s instructions. RNA sequencing was performed on a HiSeq 2500v4 high output (50 bp, single-end reads). Tophat2 was used to align the Fastq files. Transcripts per kilobase million (TPM) values were calculated and normalized using Cuffnorm. Genes that had a P less than 0.05 and at least a 1.5-fold change were considered to be significantly altered between sensitive and resistant phenotypes. The cutoff value for expressed genes was a TPM value greater than or equal to 1.

Statistics Unless otherwise stated, comparisons and statistical significance between two groups in this paper are based on a two-sided Student t-test. P values of less than 0.05 were considered statisti-

haematologica | 2022; 107(5)

CDK12-transcription reprogramming in MCL and DLBCL

cally significant. Data are shown with the mean ± standard deviation of at least three experiments. Analysis of variance (ANOVA) or the Kruskal-Wallis test was used for comparing data from multiple groups.

Results Mantle cell lymphoma and other B-cell lymphoma cell lines and primary samples are exquisitely sensitive to CDK12 inhibition regardless of genetic background and drug resistance status We performed cell viability assays on over 40 B-cell lymphoma lines for their response to the CDK12 inhibitor THZ531. These cell lines included MCL, DLBCL, double hit lymphoma and Burkitt lymphoma lines with a variety of genetic backgrounds. Established ibrutinib-resistant lines were also included for their vulnerability to CDK12 inhibition. As shown in Figure 1A and Online Supplementary Figure S1A, most of these cell lines exhibited high sensitivity to THZ531 regardless of lymphoma type, genetic background (11q, p53 status) and drug resistance with all IC50 values in the nanomolar range except for REC-1. In parallel, we also performed western blotting on all tested lines to determine the relative protein abundances of MCL-1, 4EBP1 and its phosphorylation (Online Supplementary Figure S1B) and the correlation of protein abundances with THZ531 sensitivity. As shown in Figure 1B, THZ531 IC50 values were most correlated to the protein levels of MCL-1 and phosphorylated 4EBP1. Furthermore, primary MCL samples were exquisitely sensitive to THZ531 (Figure 1C), implying that CDK12 has a functional role in aggressive B-cell malignancies. We therefore queried CDK12 gene expression from RNAsequencing performed on 40 primary samples. Notably, the expression level of MCL-1 and DDR-related genes correlated positively with CDK12 mRNA expression in B-cell lymphomas (Figure 1D, Online Supplementary Figure S1C). Therefore, CDK12 expression is frequently elevated and is essential for cell growth and survival in MCL and aggressive B-cell lymphomas.

CDK12 sustains cell growth and survival through transcriptional activation of MYC, the mTOR-4EBP1-MCL-1 axis and DNA damage response pathway in mantle cell lymphoma and MYC-associated B-cell lymphomas The strong correlations of sensitivity to THZ531 and MCL-1 protein level and 4EBP1 phosphorylation indicated the functional role of the mTOR-4EBP1-MCL-1 axis in the growth and survival of these aggressive lymphomas. We next examined the effect of CDK12 inhibition with THZ531 on cellular signaling molecules such as those involved in the PI3K-AKT-mTOR pathway in addition to MCL-1 in MCL (Z138, Jeko-1), double-hit lymphoma (DOHH-2), DLBCL (Val) cell lines and primary MCL samples. Given that CDK12 is one of the transcriptional CDK considered to be a gatekeeper of transcriptional elongation, the effects of THZ531 on phosphorylation of Ser2 on the C-terminal domain of RNAPII, MYC, the AKT-mTOR4EBP1 pathway, BCL-2 family proteins, and apoptotic PARP cleavage were assessed by western blot. As shown in Figure 2A and Online Supplementary Figure S2A, THZ531 induced a dose-dependent inhibition of RNAPII-Ser2 phosphorylation and variable degrees of inhibition of RNAPIISer5/7 phosphorylation at higher doses with associated haematologica | 2022; 107(5)

dramatic reductions of MYC and MCL-1 protein levels and triggered significant PARP cleavage in all the tested lymphoma lines. In addition, phosphorylation of p70S6K and 4EBP1, the downstream targets of mTOR, were also markedly decreased after THZ531 treatment (Figure 2A, Online Supplementary Figure S2A). Together, these results, in conjunction with the correlative studies shown in Figure 1B, imply that the mTOR-4EBP1 pathway and subsequent cap-dependent translation targets, MCL-1 and MYC, mediate CDK12 function in these aggressive B-cell lymphomas. To confirm this notion, we employed the 4EBP1 mutant, 4EBP1, as a dominant cap-dependent translation initiation inhibitor.28 4EBP1 carries alanine substitutions at four serine/threonine residues and thereby prevents dissociation of 4EBP1 from eIF4E induced by mTORC1. The abundance of MYC and MCL-1 proteins was decreased upon ectopic expression of doxycycline-induced 4EBP1 in HBL-2 cells (Figure 2B). Interestingly, when cells with ectopic expression of doxycycline-induced 4EBP1 were treated with THZ531, we observed more dramatic reductions in MCL1, MYC, and cell viability with associated increased PARP cleavage relative to non-induced cells (Figure 2C, Online Supplementary Figure S2B). This enhanced effect of THZ531 is likely attributable to endogenous translation activity. Moreover, in line with these results, increased ectopic expression of MCL-1 attenuated THZ531-induced MCL-1 downregulation, apoptosis and colony formation (Figure 2D, E, Online Supplementary Figure S2C). Together, these results indicate that 4EBP1 and MCL-1 mediate, at least partially, the biological function of CDK12. To determine the molecular mechanism and cellular pathways responsible for the activity of THZ531, RNAsequencing was performed on THZ531-sensitive MCL lines (Z138, Jeko-1) that were treated or not with 100 nM THZ531 for 6 and 24 h. Transcriptomic analysis of these THZ531-treated cells revealed alteration of expression of a large set of genes, with a total of 2,405 genes downregulated in common in the two cell lines after 6 h and 2,914 genes downregulated in common after 24 h, implying alteration of similar pathways (Figure 3A, Online Supplementary Figure S3A). Gene set enrichment analysis revealed that genes altered by THZ531 treatment for 24 h are associated with proliferation and survival pathways as well as both the MYC and mTOR-AKT pathways (Figure 3B, Online Supplementary Figure S3B). Further analysis using the Kyoto Encyclopedia of Genes and Genomes (KEGG) revealed several DDR pathways significantly negatively enriched by 24 h of treatment, including homologous recombination and the mismatch-repair pathways (Figure 3C, Online Supplementary Figure S3C). Overall, these data support THZ531-mediated transcriptional suppression. In parallel, we performed drug screening in cells sensitive to THZ531 treatment and determined their drug response profiles. As shown in Figure 3D, the MCL and aggressive B-cell lymphoma lines were uniformly sensitive to inhibitors of the transcriptional machinery apparatus (CDK7, CDK9 and CDK12), AKT, mTOR, and PLK, as well as the chemotherapeutic drug doxorubicin. Intriguingly, each of these inhibitors and chemotherapeutic drugs is known to function by downregulating MCL-1, and ectopic overexpression of MCL-1 restored the potency of these agents (Figure 3E), thus supporting the role of MCL-1 in cell survival and drug response to CDK12 inhibition. In line with these cell line results, when we performed RNAsequencing on primary MCL samples after treatment with 1121

J. Gao et al.

Figure 1. Mantle cell lymphoma and other B-cell lymphoma cell lines and primary samples are exquisitely sensitive to CDK12 inhibition with THZ531, regardless of genetic background and drug resistance status. (A) Bar plot of log10 of 50% inhibitory concentration (IC50) of B-cell lymphoma cell lines treated with THZ531 for 72 h. (B) Correlation of MCL-1 (left) or phosphor-4EBP1 protein level (right) with THZ531 IC50 in B-cell lymphoma cell lines. (C) Image-based cell-viability assays of primary mantle cell lymphoma samples; 3x106 cells were seeded in a 384-well plate with extracellular matrix and lymphoma stromal cells. THZ531 at five serial diluted concentrations was added to the medium, and the plate was continuously imaged every 30 min for 144 h. All images were analyzed using a digital imaging analysis algorithm to detect cell viability based on membrane motion, and changes in viability were quantified by area under the curve (AUC). (D) Correlation of MCL-1 (left), ATM (middle) and ATR (right) gene expression (log2TPM) with CDK12 gene expression (log2TPM) in primary patients’ samples. Data shown in (A) are representative of at least three independent experiments. TPM: transcripts per kilobase million.

1122

haematologica | 2022; 107(5)

CDK12-transcription reprogramming in MCL and DLBCL

100 or 500 nM THZ531 for 12 h, we observed that gene signatures related to the MYC, mTOR and DDR pathways were significantly downregulated (Figure 3F, Online Supplementary Figure S3D). These results support the functional role of CDK12 in MCL survival and proliferation.

MDR1 upregulation drives resistance to THZ531 in mantle cell lymphoma and other aggressive B-cell lymphomas Despite the potency shown by THZ531 and other transcriptional CDK inhibitors in preclinical studies, resistance

Figure 2. CDK12 sustains cell growth and survival through transcriptional activation of MYC and the mTOR-4EBP1-MCL-1 axis in mantle cell lymphoma and MYCassociated B-cell lymphomas. (A) Western blot analysis of phosphorylation of RNAPII in Ser2, 5 and 7, RNAPII, cleaved PARP, phosphor-p70S6K, p70S6K, phosphor4EBP1, 4EBP1, MCL-1, BCL-XL, BCL-2 and MYC in THZ531 sensitive cell lines and mantle cell lymphoma primary samples treated with the indicated doses of THZ531 at different time points. (B) Western blot analysis of MCL-1, MYC, phosphor-4EBP1 and 4EBP1 protein levels in HBL-2 cells with and without doxycycline-induced overexpression of 4EBP1. (C) Western blot analysis of cleaved PARP, MCL-1, BCL-XL, BCL-2, MYC and 4EBP1 protein expression in HBL-2 cells with and without doxycycline-induced overexpression of 4EBP1 treated with the indicated doses of THZ531 for 24 h. (D) Image-based cell-viability assays of HBL-2 cells with and without overexpression of MCL-1. Cells (1x106) were seeded in a 384-well plate with extracellular matrix. THZ531 was added at five serial diluted concentrations to the medium, and the plate was continuously imaged every 30 min for 96 h. All images were analyzed using a digital imaging analysis algorithm to detect cell viability based on membrane motion, and changes in viability were quantified by the area under the curve. (E) Western blot analysis of cleaved PARP, MYC, MCL-1 and BCL-XL in HBL-2 cells with and without MCL-1 overexpression treated with the indicated doses of THZ531 for 24 h. DOX: doxycycline; DMSO: dimethylsulfoxide.

haematologica | 2022; 107(5)

1123

J. Gao et al.

Figure 3. Transcriptomic analysis of THZ531-sensitive cells. (A) Transcriptome changes occurring in common in Z138 and Jeko-1 cells treated with THZ531 after 6 h (left) and 24 h (right). Red: upregulated genes, black: downregulated genes. LFC: log2 fold change cut-off of log2 (1.5), P-value cutoff of 0.05. Three biologically independent samples. (B) Dot plot of hallmark pathways that, from gene set enrichment analysis (GSEA), are negatively enriched in both Z138 and Jeko-1 cell lines treated with THZ531 after 24 h. Larger dot sizes indicate a more negative enrichment score. Color scale represents significance. Genes are ranked according to their expression fold change after treatment. NES: normalized enrichment score; FDR: false discovery rate. (C) GSEA enrichment curves of KEGG pathways negatively enriched in both Z138 and Jeko-1 after THZ531 treatment. (D) Drug sensitivity shown as a heatmap of areas under the curves (AUC) calculated from dose-response curves obtained from cell viability assays performed in Z138, Jeko-1, DOHH2 and Val cell lines for each indicated drug administered for 72 h. (E ) Drug sensitivity shown as a heatmap of AUC calculated from dose-response curves obtained from image-based cell viability assays performed in HBL-2 cells with and without MCL1 overexpression for each indicated drug after 72 h of treatment. (F) Venn diagram of significantly negatively enriched GSEA hallmarks gene sets (NES <0, FDR <0.25) in mantle cell lymphoma primary patients’ samples treated with 500 nM THZ531 for 12 h. Pathways negatively enriched in common in all four samples are listed. Data shown in (D and E) are representative of at least three independent experiments.

1124

haematologica | 2022; 107(5)

CDK12-transcription reprogramming in MCL and DLBCL

Figure 4. MDR1 upregulation drives resistance to THZ531 in mantle cell lymphoma and other aggressive B-cell lymphomas. (A) Dose-response curves of Z138, Jeko-1 and REC-1 cell lines after 72 h of treatment with THZ531. Data are shown as mean ± standard deviation (SD) of three technical replicates for each cell line. (B) Western blot analysis of phosphorylation of RNAPII in Ser2, 5 and 7, RNAPII, cleaved PARP, phosphor-p70S6K, p70S6K, phosphor-4EBP1, 4EBP1, MCL-1, BCLXL, BCL-2 and MYC protein expression in REC-1 cells treated with the indicated doses of THZ531 at different time points. (C) Volcano plot showing no changes at the mRNA level after THZ531 treatment for 24 h relative to treatment with dimethylsulfoxide (DMSO) in REC-1 cells. Red: upregulated genes, black: downregulated genes. LFC: log2 fold change cutoff of log2 (1.5), P-value cutoff of 0.05. Three biologically independent samples. (D) Bar plot of differential expression of drug pump genes in the THZ531-resistant cell line REC-1 versus the THZ531-sensitive cell lines Z138 and Jeko-1. Bar length represents log2 fold change. Color represents direction of expressions change where red represents genes with log2 fold change greater than 1.5 and blue represents genes with log2 fold change less than 1.5. (E) Doseresponse curves of REC-1 cells after 72 h of treatment with different doses of THZ531, tariquidar, or THZ531+tarquidar. Data are shown as mean ± SD of three technical replicates for each cell line. (F) Dose-response curves of the KARPAS-422-THZ-R and Maver-1-THZ-R cell lines after 72 h of treatment with different doses of THZ531, tariquidar, or THZ531+tariquidar. Data are shown as mean ± SD of three technical replicates for each cell line. Data shown in (A, E and F) are representative of at least three independent experiments.

haematologica | 2022; 107(5)

1125

J. Gao et al.

to these agents inevitably arises. Given that a recent study showed that ABC transporters mediate resistance to a THZ series of transcriptional CDK inhibitors,29,30 we next determined the role of MDR1 in de novo and acquired THZ531 resistance. Although sensitive cells had IC50 values in the

nanomolar range, the de novo resistant line REC-1 had an IC50 of 1.63 μM (Figure 4A). Using the same functional biochemical and pharmacogenomic approaches as those used for the THZ531-sensitive lines, we investigated the molecular pathways and mechanisms responsible for resistance

Figure 5. EZH2 inhibitors restored sensitivity to THZ531 in THZ531-resistant cells by competing with THZ531 for MDR1. (A). Z-scores of normalized areas under the curve (AUC) of drug response curves from a high-throughput small-molecule drug screen performed in REC-1 cells for each indicated drug alone or combined with 500 nM THZ531. Selected compounds that had more potent effects upon combination with THZ531 are highlighted in red. (B) Left: dose-response curves of REC-1 cells after 72 h of treatment with different doses of THZ531, GSK343, or THZ531+GSK343. Right: dose-response curves of REC-1 cells after 72 h of treatment with different doses of THZ531, UNC1999, or THZ531+UNC1999. Data are shown as mean ± standard deviation of three technical replicates for each cell line. (C) Western blot analysis of phosphorylation of RNAPII in Ser2, RNAPII, cleaved PARP, phosphor-p70S6K, p70S6K, phosphor-4EBP1, 4EBP1, MCL-1, BCL-XL, BCL-2 and MYC protein expression in REC-1 cells after 24 h of treatment with dimethylsulfoxide (DMSO) or indicated doses of THZ531 and/or GSK343. (D) Western blot analysis of MDR1 and cleaved PARP protein expression in REC-1 (left) and Maver-1-THZ-R (right) cell lines after treatment with DMSO or the indicated doses of THZ531 and/or GSK343 at different time points. (E) REC-1 (left), KARPAS-422-THZ-R (middle) and Maver-1-THZ-R (right) cells were treated with different drugs at the indicated doses for 12 h. P-gp activity was determined using a fluorimetric MDR assay kit. The P-gp inhibitor verapamil (30 µM) served as a positive control. Results are expressed as mean ± standard error of mean. Data shown in (B) are representative of at least three independent experiments.

1126

haematologica | 2022; 107(5)

CDK12-transcription reprogramming in MCL and DLBCL

Figure 6. Activation of MEK-ERK and PI3K-AKT-mTOR pathways contributes to MDR1 upregulation in THZ531-resistant cells. (A) Image-based cell-viability assays of primary MCL samples, cells (3x106) cells were seeded in a 384-well plate with extracellular matrix and lymphoma stromal cells. THZ531 and GSK343 at five serial diluted concentrations were added to the medium, and the plate was continuously imaged every 30 min for 144 h. All images were analyzed using a digital imaging analysis algorithm to detect cell viability based on membrane motion, and changes in viability were quantified by area under the curve (AUC). (B) Western blot of MDR1 and cleaved PARP protein expression in REC-1 cells after 24 h or 48 h of treatment with dimethylsulfoxide or indicated doses of THZ531 and/or AZD8055, BEZ235, or trametinib. (C) Dose-response curves of the REC-1 cells after treatment for 72 h with different doses of THZ531 and/or AZD8055, BEZ235 and trametinib. Data are shown as mean ± standard deviation of three technical replicates for each cell line. Data shown in (C) are representative of at least three independent experiments.

haematologica | 2022; 107(5)

1127

J. Gao et al.

to THZ531. Western blotting revealed that, in contrast to THZ531-sensitive lines, the resistant REC-1 cells were refractory to THZ531 treatment and THZ531 failed to change the expression of MCL-1, MYC, and 4EBP1, PARP cleavage as well as all Ser2 phosphorylation of RNAPII (Figure 4B). Next, to investigate the cellular pathways responsible for THZ531 resistance, we analyzed the transcriptome changes after THZ531 treatment in REC-1. As illustrated in Figure 4C, in contrast to cells that were responsive to THZ531, THZ531 treatment for 24 h at a dose (100 nM) that was highly potent in THZ531-sensitive lines did not induce gene expression changes in REC1. We then compared the differential gene expression between THZ531-sensitive cell lines (Z138, Jeko-1) and a THZ531-resistant one (REC-1). Among drug transport genes identified by RNA-sequencing, the level of MDR1 transcript was dramatically higher in REC-1 cells than in Z138 and Jeko-1 cells (Figure 4D). Moreover, quantitative reverse transcriptase polymerase chain reaction analysis revealed greater MDR1 abundance in REC-1 cells (Online Supplementary Figure S4A). To study the functional role of MDR1 in resistance to THZ531, we tested whether inhibition of MDR1 by its specific inhibitor (tariquidar) could overcome THZ531 resistance. As shown in Figure 4E, tariquidar overcame THZ531 resistance and synergized with THZ531 in the suppression of cell viability in REC1 cells. In addition, we developed two cell lines, KARPAS422-THZ-R and Maver-1-THZ-R (Online Supplementary Figure S4B), with resistance to THZ531: “acquired” by chronic exposure to escalating doses of THZ531 over 3 months. In line with cells with de novo drug resistance, the THZ531-resistant KARPAS-422 and Maver-1 cells exhibited higher MDR1 expression at both the mRNA and protein levels (Online Supplementary Figure S4C). Cell viability of both cell lines was dramatically reduced when exposed to the combination of THZ531 with tariquidar (Figure 4F). Together, these data support that MDR1 upregulation drives THZ531 resistance in MCL and other aggressive Bcell lymphomas.

EZH2 inhibitors restored sensitivity to THZ531 in THZ531-resistant cells by competing with THZ531 for MDR1 To define the molecular determinants that drive resistance to THZ531, reverse this resistance and enhance THZ531 activity in THZ531-resistant cell lines, we performed drug screen assays on REC-1 against a set of epigenetic modifiers and kinase inhibitors in the presence or absence of THZ531 as an “anchor” in our drug screen platform. Using AUC to quantify the effect and potency of each single and combination treatment, we identified the inhibitors that enhanced THZ531 activity and ranked the combination potency by the differential killing effect between drug combinations and single inhibitors. Inhibitors found to enhance the potency of THZ531 included MEK, BTK, mTOR and EZH2 inhibitors. Among them, the drug screen assay identified that inhibitors targeting EZH2 in combination with THZ531 had the greatest increased potency against REC-1 cells (Figure 5A, Online Supplementary Figure S5A). Indeed, cell viability assays revealed that the combination of EZH2 inhibitors (GSK343 or UNC1999) with THZ531 had synergistic effects against cell survival in REC-1 cells (Figure 5B). Mechanistically, western blots revealed that neither EZH2 inhibitor nor THZ531 alone had an effect on Ser2 1128

phosphorylation of RNAPII, p-p70S6K, p-4EBP1, MCL-1 or PARP cleavage (Figure 5C, Online Supplementary Figure S5B). However, the combination of either EZH2 inhibitor with THZ531 triggered dramatic decreases of Ser2 phosphorylation of RNAPII, p-p70S6K, p-4EBP1 and MCL-1 levels, and significantly increased PARP cleavage (Figure 5C, Online Supplementary Figure S5B). Consistently, synergistic effects of THZ531 combined with GSK343 or UNC1999 on cell viability were also observed in the THZ531-resistant cell lines KARPAS-422-THZ-R and Maver-1-THZ-R (Online Supplementary Figure S5C). Next, to investigate the mechanisms by which EZH2 inhibition reinstated THZ531 sensitivity in resistant cells, we examined changes in MDR1 protein expression upon EZH2 inhibitor treatment in THZ531-resistant cells. As shown in Figure 5D, although treatments with any single agent had no effect on MDR1 protein expression, EZH2 inhibitor significantly sensitized REC-1 cells to THZ531 treatment, as measured by PARP cleavage. In contrast, EZH2 knockdown through shRNA showed that EZH2 downregulation had no effect on basal MDR1 expression and failed to reverse THZ531-induced downstream signaling via RNAPII-Ser2, MCL-1 and MYC in REC-1 cells (Online Supplementary Figure S5D). These data suggest that the potency enhancing effects of GSK343 and UNC1999 were not due to biological EZH2 regulation on either MDR1 or THZ531 signaling. Thus, to further reveal the underlying mechanism of MDR1 modulation by GSK343 and UNC1999, we performed P-gP-Glo assays, which served to elucidate whether GSK343 and UNC1999 are direct MDR1 inhibitors or substrates in REC-1, KARPAS422-THZ-R and Maver-1-THZ-R cells. GSK343 and UNC1999 were confirmed not to be direct inhibitors, but strong MDR1 substrates, comparable to verapamil (Figure 5E). Thus, the data support that EZH2 inhibitors GSK343 and UNC1999 function by competitively binding to MDR1.

Activation of MEK-ERK and PI3K-AKT-mTOR pathways contribute to MDR1 upregulation in THZ531-resistant cells We next systematically tested for synergistic combinations of THZ531 with EZH2 inhibitors in the sensitive as well as resistant cells to validate the combination as an effective strategy to overcome and block emergent resistance in primary samples. Importantly, in the innately THZ531-resistant primary samples, treatment with EZH2 inhibitors sensitized the cells to THZ531 (Figure 6A), strengthening the rationale to combine THZ531 and EZH2 inhibitors in the upfront setting. Given that the drug screen on REC-1 cells showed inhibitors of MEK and AKT sensitized REC-1 to THZ531 treatment, we hypothesized that activation of the MEKERK and PI3-AKT-mTOR pathways in REC-1 regulated MDR1 expression and affected the efficacy of THZ531. As shown in Figure 6B, inhibition of MEK with trametinib and PI3K-AKT with either AZD8055 or BEZ235 induced a significant decrease of the abundance of MDR1 protein in REC-1 cells. Accordingly, combined treatment of THZ531 with trametinib, AZD8055 or BEZ235 all induced a synergistic effect on REC-1 cell viability, as measured by MTT assays (Figure 6C). Overall, these data imply that activation of MEK-ERK and PI3K-AKT-mTOR pathways contributes to MDR1 upregulation in THZ531resistant cells. haematologica | 2022; 107(5)

CDK12-transcription reprogramming in MCL and DLBCL

Discussion MCL is an incurable form of lymphoma and the median overall survival of patients with this malignancy is only 4 to 5 years. Its clinical course usually involves sequential relapses, with no response to current standard therapies. MYCassociated large B-cell lymphoma is one of the most aggressive lymphomas. Patients often experience relapsed or refractory disease after an initial response to first-line therapy and the majority of those with MYC-associated B-cell lymphomas succumb to their disease. CDK12 is a transcriptional CDK that complexes with cyclin K to mediate gene transcription by phosphorylating RNAPII. CDK12 has been demonstrated to specifically upregulate the expression of genes involved in response to DNA damage, stress and proliferation, as well as mRNA processing and cell survival17 by directly phosphorylating pre-mRNA processing factors, which induces premature cleavage and polyadenylation and a loss of expression of long genes that participate in the DDR.18,19 An increasing number of studies have highlighted CDK12 as a therapeutic target for cancer. Inhibition of transcriptional CDK could be an effective strategy to overcome resistance to targeted therapies, including erlotinib and crizotinib. Numerous other studies have identified specific genetic or cellular contexts that confer enhanced sensitivity to CDK12 inhibition, including MYC dependency. In this study, we aimed to explore CDK12 as a novel vulnerability for MCL and MYCassociated B-cell lymphomas. We further determined the role of CDK12-mediated transcription activation and associated pathways in cell survival and growth of MCL and MYC-associated B-cell lymphomas. Here, we report that the aggressive B-cell lymphomas are exquisitely sensitive to transcription-targeting drugs, particular to the covalent CDK12 inhibitor THZ531. By implementing pharmacogenomics and a cell-based drug screen, we found that THZ531 leads to inhibition of oncogenic transcriptional programs, especially the DDR pathway, MYC target genes and the mTOR-4EBP1-MCL-1 axis, contributing to lymphoma suppression ex vivo. Importantly, we investigated molecular mechanisms involved in conferring resistance to THZ531 and examined whether combined inhibitors of CDK12 and EZH2 cooperatively reprogram transcription repression to overcome resistance to THZ531, and, ultimately, inhibit lymphoma growth and survival in aggressive B-cell malignancies. We demonstrate that MDR1 is overexpressed in cell lines with de novo and acquired resistance to THZ531, and that MDR1 overexpression results in THZ531 resistance via MDR1mediated export of THZ531. Our data implicate MDR1 as a therapeutic target to overcome THZ531 resistance. Additional testing with the chemically distinct CDK12 inhibitor SR-4835 revealed that MDR1 upregulation as an avenue of resistance in CDK12 inhibition is specific to THZ531 (data not shown). However, CDK12 inhibitors are

References 1. Nilsson JA, Cleveland JL. Myc pathways provoking cell suicide and cancer. Oncogene. 2003;22(56):9007-9021. 2. Dave SS, Fu K, Wright GW, et al. Molecular diagnosis of Burkitt's lymphoma. N Engl J Med. 2006;354(23):2431-2442. 3. Hartmann EM, Ott G, Rosenwald A.

haematologica | 2022; 107(5)

known functional substrates for drug transport molecules,31 and accordingly, it has been reported that MDR1 upregulation contributes to resistance against other compounds with the THZ chemical backbone, including the CDK7 inhibitor THZ1.29 We then identified GSK343 and UNC1999, two EZH2 inhibitors, as potent drugs in preventing MDR1-mediated export of THZ531 in lymphoma cells. The combination of GSK343 or UNC1999 with THZ531 exhibited synergistic efficacy against MCL and aggressive B-cell lymphomas. We further showed that GSK343 and UNC1999 inhibited MDR1 efflux by competing with THZ531 for MDR1, consistent with previous reports suggesting that GSK343 and UNC1999 are potential MDR1 substrates.32 Thus, GSK343 and UNC1999 have a broad drug-sensitizing potential for lymphoma therapy. Furthermore, we also demonstrated that activation of MEK-ERK and PI3K-AKT-mTOR pathways contributes to MDR1 upregulation in THZ531-resistant cells. We identified that de novo and established acquired THZ531-resistant lymphoma cells are associated with over-activation of MEK-ERK and PI3K-AKT-mTOR pathways, contributing to upregulation of MDR1 protein. Intriguingly, EZH2 inhibitors reversed THZ531 resistance by competitive inhibition of MDR1 and, in combination with THZ531, synergistically inhibited MCL and MYC-associated lymphoma growth in vitro. Our study indicates that CDK12 inhibitors, both alone and together with EZH2 inhibitors, offer promise as a novel treatment strategy that can be an effective approach for MYC-dependent lymphomas and MCL. Furthermore, GSK343 and UNC1999 can also reverse drug resistance that has developed from MDR1 upregulation. Thus, GSK343 and UNC1999 can be used together with other MDR1-induced drugs such as doxorubicin and carfilzomib to overcome drug resistance for greater and longerlasting efficacy in patients with aggressive B-cell lymphomas. Disclosures No conflicts of interest to disclose. Contributions JT and XZ conceived and designed the study; JG, MYW, YR, TLi, JCY and TLw performed experiments, and collected and assembled the data; JG, MYW, XZ, RY and JT analyzed and interpreted the data; JT, JG and MW wrote , reviewed and/or revised the manuscript; EMS, KHS, BDS, DRD and JT provided administrative, technical or material support. Funding This work was supported in part by grants from the National Cancer Institute CA241713, CA233601, and CA234519 (to JT), a grant from the Lymphoma Research Foundation (to JT) and by funds from Florida State Live Like Bella Pediatric Cancer Research Initiative to the H. Lee Moffitt Cancer Center & Research Institute.

Molecular biology and genetics of lymphomas. Hematol Oncol Clin North Am. 2008;22(5):807-823 4. Slack GW, Gascoyne RD. MYC and aggressive B-cell lymphomas. Adv Anat Pathol. 2011;8(3):219-228. 5. Ren Y, Bi C, Zhao X, et al. PLK1 stabilizes a MYC-dependent kinase network in aggressive B cell lymphomas. J Clin Invest.

2018;128(12):5517-5530. 6. Hnisz D, Schuijers J, Lin CY, et al. Convergence of developmental and oncogenic signaling pathways at transcriptional super-enhancers. Mol Cell. 2015;58(2):362370. 7. Loven J, Hoke HA, Lin CY, et al. Selective inhibition of tumor oncogenes by disruption of super-enhancers. Cell. 2013;153(2):320-

1129

J. Gao et al. 334. 8. Winter GE, Mayer A, Buckley DL, et al. BET bromodomain proteins function as master transcription elongation factors independent of CDK9 recruitment. Mol Cell. 2017; 67(1):5-18. 9. Ferguson FM, Gray NS. Kinase inhibitors: the road ahead. Nat Rev Drug Discov. 2018; 17(5):353-377. 10. Smith E, Lin C, Shilatifard A. The super elongation complex (SEC) and MLL in development and disease. Genes Dev. 2011; 25(7):661-672. 11. He N, Liu M, Hsu J, et al. HIV-1 Tat and host AFF4 recruit two transcription elongation factors into a bifunctional complex for coordinated activation of HIV-1 transcription. Mol Cell. 2010; 38(3):428-438. 12. Chipumuro E, Marco E, Christensen CL, et al. CDK7 inhibition suppresses superenhancer-linked oncogenic transcription in MYCN-driven cancer. Cell. 2014; 159(5): 1126-1139. 13. Christensen CL, Kwiatkowski N, Abraham BJ, et al. Targeting transcriptional addictions in small cell lung cancer with a covalent CDK7 inhibitor. Cancer Cell. 2014; 26(6):909-922. 14. Wang Y, Zhang T, Kwiatkowski N, et al. CDK7-dependent transcriptional addiction in triple-negative breast cancer. Cell. 2015; 163(1):174-186. 15. Pelish HE, Liau BB, Nitulescu II, et al. Mediator kinase inhibition further activates super-enhancer-associated genes in AML. Nature. 2015;526(7572):273-276. 16. Zhao X, Ren Y, Lawlor M, et al. BCL2

1130

amplicon loss and transcriptional remodeling drives ABT-199 resistance in B cell lymphoma models. Cancer Cell. 2019;35(5):752766. 17. Lui GYL, Grandori C, Kemp CJ. CDK12: an emerging therapeutic target for cancer. J Clin Pathol. 2018;71(11):957-962. 18. Dubbury SJ, Boutz PL, Sharp PA. CDK12 regulates DNA repair genes by suppressing intronic polyadenylation. Nature. 2018; 564(7734):141-145. 19. Krajewska M, Dries R, Grassetti AV, et al. CDK12 loss in cancer cells affects DNA damage response genes through premature cleavage and polyadenylation. Nat Commun. 2019;10(1):1757. 20. Iniguez AB, Stolte B, Wang EJ, et al. EWS/FLI confers tumor cell synthetic lethality to CDK12 inhibition in Ewing sarcoma. Cancer Cell. 2018;33(2):202-216. 21. Johnson SF, Cruz C, Greifenberg AK, et al. CDK12 inhibition reverses de novo and acquired PARP inhibitor resistance in BRCA wild-type and mutated models of triple-negative breast cancer. Cell Rep. 2016; 17(9):2367-2381. 22. Jares P, Colomer D, Campo E. Molecular pathogenesis of mantle cell lymphoma. J Clin Invest. 2012;122(10):3416-3423. 23. Campaner S, Amati B. Two sides of the Myc-induced DNA damage response: from tumor suppression to tumor maintenance. Cell Div. 2012;7(1):6. 24. Rohban S, Campaner S. Myc induced replicative stress response: how to cope with it and exploit it. Biochim Biophys Acta. 2015;1849(5):517-524

25. Silva A, Jacobson T, Meads M, et al. An organotypic high throughput system for characterization of drug sensitivity of primary multiple myeloma cells. J Vis Exp. 2015;(101):e53070. 26. Silva A, Silva MC, Sudalagunta P, et al. An ex vivo platform for the prediction of clinical response in multiple myeloma. Cancer Res. 2017;77(12):3336-3351. 27. Zhao X, Lwin T, Silva A, et al. Unification of de novo and acquired ibrutinib resistance in mantle cell lymphoma. Nat Commun. 2017; 8:14920. 28. Wang J, Ye Q, She QB. New insights into 4EBP1-regulated translation in cancer progression and metastasis. Cancer Cell Microenviron. 2014;1(5):e331. 29. Gao Y, Zhang T, Terai H, et al. Overcoming resistance to the THZ series of covalent transcriptional CDK inhibitors. Cell Chem Biol. 2018;25(2):135-142. 30. Olson CM, Liang Y, Leggett A, et al. Development of a selective CDK7 covalent inhibitor reveals predominant cell-cycle phenotype. Cell Chem Biol. 2019;26(6):792-803. 31. Cihalova D, Staud F, Ceckova M. Interactions of cyclin-dependent kinase inhibitors AT-7519, flavopiridol and SNS032 with ABCB1, ABCG2 and ABCC1 transporters and their potential to overcome multidrug resistance in vitro. Cancer Chemother Pharmacol. 2015;76(1):105116. 32. Zhang P, de Gooijer MC, Buil LC, et al. ABCB1 and ABCG2 restrict the brain penetration of a panel of novel EZH2-Inhibitors. Int J Cancer. 2015;137(8):2007-2018.

haematologica | 2022; 107(5)

ARTICLE

Non-Hodgkin Lymphoma

Characterization of GECPAR, a noncoding RNA that regulates the transcriptional program of diffuse large B-cell lymphoma

Ferrata Storti Foundation

Sara Napoli,1 Luciano Cascione,1,2 Andrea Rinaldi,1 Filippo Spriano,1 Francesca Guidetti,1 Fangwen Zhang,1 Maria Teresa Cacciapuoti,3 Afua Adjeiwaa Mensah,1 Giulio Sartori,1 Nicolas Munz,1 Mattia Forcato,4 Silvio Bicciato4, Annalisa Chiappella,5 Paola Ghione,6 Olivier Elemento,7,8 Leandro Cerchietti, 6 Giorgio Inghirami3 and Francesco Bertoni1,9 Institute of Oncology Research, Faculty of Biomedical Sciences, USI, Bellinzona, Switzerland; 2SIB Swiss Institute of Bioinformatics, Lausanne, Switzerland; 3Pathology and Laboratory Medicine Department, Weill Cornell Medicine, New York, NY, USA; 4 Center for Genome Research, Department of Life Sciences University of Modena and Reggio, Modena, Italy; 5Ematologia, A.O.U. Città della Salute e della Scienza di Torino, Turin, Italy; 6Department of Medicine, Division of Hematology and Medical Oncology, Weill Cornell Medicine, New York, NY, USA; 7Institute for Computational Biomedicine, Department of Physiology and Biophysics, Weill Cornell Medicine, New York, NY, USA; 8 Caryl and Israel Englander Institute for Precision Medicine, Weill Cornell Medicine, New York, NY, USA and 9Oncology Institute of Southern Switzerland, Bellinzona, Switzerland. 1

Haematologica 2022 Volume 107(5):1131-1143

ABSTRACT

nhancers are regulatory regions of DNA, which play a key role in cell-type specific differentiation and development. Most active enhancers are transcribed into enhancer RNA (eRNA) that can regulate transcription of target genes by means of in cis as well as in trans action. eRNA stabilize contacts between distal genomic regions and mediate the interaction of DNA with master transcription factors. Here, we characterized an enhancer eRNA, GECPAR (germinal center proliferative adapter RNA), which is specifically transcribed in normal and neoplastic germinal center B cells from the super-enhancer of POU2AF1, a key regulatory gene of the germinal center reaction. Using diffuse large B-cell lymphoma cell line models, we demonstrated the tumor suppressor activity of GECPAR, which is mediated via its transcriptional regulation of proliferation and differentiation genes, particularly MYC and the Wnt pathway.

Introduction Enhancers are regulatory DNA regions that positively drive gene transcription across neighboring genomic regions spanning many megabases and are characterized by distinct epigenetic features:1,2 a high ratio of H3K4me1 to H3K4me3; enrichment of H3K27ac, which is deposited by the CREBBP/p300 complex;3 high accessibility to chromatin readers such as bromodomain and extraterminal domain (BET) proteins and transcription factors (TF). Some enhancers are actively transcribed giving rise to noncoding RNA called enhancer RNA (eRNA).4 Transcribed enhancers are more acetylated, more enriched of TF and co-activators, and are also more active in the transactivation of promoters, with which they interact inside 3D structures called enhancer-promoter loops.5 Clusters of enhancers, called super-enhancers (SE), are strongly transcribed and produce several eRNA controlling key genes, which regulate cellular development and differentiation.6,7 eRNA are crucial components of the regulatory chromatin machinery that controls the expression of key context-specific, protein-coding genes. They usually stabilize multiprotein complexes and constitute a scaffold for DNA loops by enforcing interactions between distant DNA regions, including those located on different chromosomes.8-11 As they lack a poly A tail, their activity is restrained to the site of transcription and they undergo rapid decay. However, polyadenylated long intergenic non-coding RNA (lincRNA) also comprise enhancer-derived

haematologica | 2022; 107(5)

Correspondence: FRANCESCO BERTONI francesco.bertoni@ior.usi.ch SARA NAPOLI sara.napoli@ior.usi.ch

Received: July 29, 2020. Accepted: June 16, 2021. Pre-published: June 24, 2021. https://doi.org/10.3324/haematol.2020.267096

1131

S. Napoli et al.

non-coding transcripts (e-lncRNA),12 and the stabilization of these eRNA confers to them the capability to act in trans, regulating several distant targets.13 Individual eRNA are expressed in a tissue-specific manner. In normal B cells at various stages of differentiation, the expression of non-coding RNA can more precisely define cellular subsets than protein-coding transcripts.14,15 In particular, eRNA are differentially expressed during Bcell development and they are associated with proteincoding genes that play an essential role in B-cell differentiation. Diffuse large B-cell lymphoma (DLBCL) derives from germinal center (GC) B cells. DLBCL is typically divided into two main subtypes: GC B cell-like (GCB-DLBCL), whose transcriptional profile resembles that of light zone GC B cells, and activated B cell-like (ABC-DLBCL), whose transcriptome resembles that of plasmablasts.16 However, DLBCL within each of these subgroups exhibit biological, genetic and transcriptional heterogeneity.17-19 Lineage-specific and growth-dependent transcription factors like BCL6, Myc, NF-kB, p53, and E2F1 can activate specific genetic signatures, depending on the activation of unique subsets of enhancers20,21 and contribute to disease heterogeneity. Here, we studied a unique eRNA associated with the POU2AF1 gene, that we termed GECPAR, for germinal center proliferative adapter RNA. POU2AF1 encodes the protein OCA-B, co-activator of OCT2, a B-cell specific transcription factor which plays a pivotal role in the regulation of normal and neoplastic GC B cells.22,23 The SE proximal to POU2AF1 is the most activated SE in GCBDLBCL.23 Loss of GECPAR correlated with reduced transcription of TLE4, which is a negative regulator of LEF1, a Wnt pathway effector protein that in turn regulates also NF-kB. GECPAR loss also increased MYC expression and proliferation of DLBCL cell lines. Conversely, its overexpression impaired cell proliferation. Collectively, our data provides evidence of the nodal role of GECPAR in the regulatory network modulating B-cell differentiation and proliferation.

Methods Detailed descriptions of the experimental methods are included in the Online Supplementary Appendix.

Human samples, cell lines, small interfering RNA transfection Established human DLBCL cell lines and patient-derived tumor xenograft cell lines (PDTX-CL) were grown as previously described.24 All patients providing samples gave written informed consent. Molecular and clinical data acquisition and PDTX establishment were approved and carried out in accordance with Declaration of Helsinki and were approved by Institutional Review Boards of the New York Presbyterian Hospital, Weill Cornell Medicine (WCM) and the Ospedale San Giovanni Battista delle Molinette. Cell lines were checked for their identity.24 Cells were transfected with small interfering RNA (siRNA) or lockednucleic acid (LNA) using the 4D Nucleofector.

GECPAR cloning and infection into lymphoma cells, RNA sequencing Cellular lysates were fractionated as previously described.25 For strand-specific quantitative reverse transcription polymerase chain reaction (qRT-PCR), only the forward primer was used to

1132

amplify the antisense strand and only the reverse primer to amplify the sense strand. 5’ and 3’ rapid amplification of cDNA ends (RACE) was done using Invitrogen RACE System kits. GECPAR was cloned into the pGEM-T vector and subcloned in pCDH-CMV-MCS-EF1-copGFP. pCDH empty backbone or pCDH_GECPAR were transfected in HEK293T, and viral supernatant was then used to infect lymphoma cells. RNA sequencing (RNA-Seq) in cell lines was performed using the NEBNext Ultra II Directional RNA Library Prep.

Capture hybridization analysis of RNA targets (CHART) sequencing CHART enrichment and ribonuclease H (RNAseH) mapping experiments were performed following previously reported protocols.26,27 The enrichment of CHART signals was determined relative to the oligo controls. Conservative enrichment profiles were determined using the SPP package28 and MACS,29 as described by Vance and colleagues.30

Results The super-enhancer associated with the POU2AF1 gene locus is transcribed in normal B cells and diffuse large B-cell lymphoma cell lines Analysis of publicly available RNA-Seq data on RNA polyA+ or polyA-31 showed that CD20+ cells express a non-polyadenylated portion of the LOC100132078 transcript and also two isoforms of a more abundant antisense transcript (Figure 1A; Online Supplementary Figure S1A). Due to its proximity to the POU2AF1 gene and its localization in a genomic region with characteristic SE features (highly acetylated, enriched in H3K4me1 but not H3K4me3, based on ENCODE ChIP-Seq data), we hypothesized that it could be an eRNA with particular relevance for GC B cells In order to confirm the eRNA length reconstructed in CD20+ cells, we performed 5’ and 3’ RACE in the DLBCL cell line OCI-LY1. For the 3’-end detection we ran two reactions, with or without the addition of an artificial polyA tail. We identified a transcript lacking a polyA tail and another that was 400 bases longer and naturally polyadenylated. Similarly to the aforementioned polyAtranscript reported in CD20+ normal B cells, neither of the transcripts identified in DLBCL cells extended beyond the annotated first exon. The 5’ RACE reaction reverse transcribed from exon 4 did not identify a specific 5’-end for exon 1, indicating that the long annotated transcript, LOC100132078, was likely not stable in our model. Conversely, reverse transcribing from exon 1, we identified a 5’-end located at nucleotide +366, mirroring our in silico observations for CD20+ normal B cells (Figure 1A and B). We renamed the stabilized portion of LOC100132078 we had sequenced in the OCI-LY1 model as “GECPAR”.

GECPAR is mainly chromatin associated and partially polyadenylated In order to further characterize the physical characteristics of GECPAR, RNA was extracted from the cytoplasm, nucleoplasm and chromatin fractions. In GCB-DLBCL (OCI-LY1 and Karpas422) and ABC-DLBCL (HBL1, U2932) cell lines, GECPAR was transcribed but mostly retained on chromatin, in accordance with reported features of eRNA.6,7 It was also clearly detected in the nuclehaematologica | 2022; 107(5)

GECPAR, eRNA master regulator of germinal center

oplasm and cytoplasm of OCI-LY1, a cell line with 5-fold higher levels of chromatin-associated GECPAR than the other cell lines (Figure 1C). Semi-quantitative directional RT-PCR showed that chromatin association was particular to GECPAR since its antisense transcript, when

expressed, was more ubiquitously distributed (Online Supplementary Figure S1B). Quantification of KCNQ1OT1, MALAT1 and β-actin mRNA served as a control for chromatin-associated, nuclear and cytosolic RNA, respectively (Online Supplementary Figure S1C).

Figure 1. Figure continued on following page.

haematologica | 2022; 107(5)

1133

S. Napoli et al. B

D Figure 1. POU2AF1 super-enhancer derived transcript in normal B cells and diffuse large B-cell lymphoma cell lines. (A) Top: schematic representation of transcripts annotated in chromosome 11q23, between POU2AF1 and BTG genes, according UCSC Genome Browser. Bottom: close-up of LOC100132078 annotated transcript, aligned with CAGE signals on strand plus and minus, transcripts sequenced and reconstructed in RNA polyA+ or polyA- from CD20+ cells, and histone marks from ENCODE project. Red lines show positions of exact 5' and 3' ends of GECPAR determined by rapid amplification of cDNA ends (RACE) in OCI-LY1. Arrows indicate position of primers used for 5' and 3' RACE, in particular red arrows primers used for the retrotranscription step. (B) 5’ (left) and 3’ (right) RACE performed in OCI-LY1. Numbers on the right of the bands indicate the exact nucleotides corresponding to 5’ and 3’ends of GECPAR respect to nucleotide +1, the TSS of annotated LOC100132078. (C) GECPAR level measured by quatitative reverse transcription (qRT) in subcellular compartments in four diffuse large B-cell lymphoma (DLBCL) cell lines, two germinal center B cell-like (GCB) and two activated B cell-like (ABC)-DLBCL. (D) GECPAR level measured by qRT in total RNA transcripts or polyadenylated only, in four DLBCL cell lines. Data are mean ± standard deviation of independent determinations. *P<0.05.

Strong association of a transcript to chromatin usually correlates with its lack of polyadenylation consequent rapid degradation by the RNA exosome.32 In order to determine if these features were applicable to GECPAR, we assessed its polyadenylation status. The latter was abundant in total transcripts reverse-transcribed using random hexamers, especially in the two GCB-DLBCL cell lines. Conversely, when oligo-dT was used for reverse transcription, GECPAR was clearly detectable in only OCI-LY1, in agreement with the higher abundance of GECPAR in this cell line. (Figure 1D).

GECPAR is predominantly transcribed in germinal center diffuse large B-cell lymphoma cell lines and patients We measured GECPAR transcription by directional qRT-PCR in 22 DLBCL cell lines (GCB, n=16; ABC, n=8). The overlapping antisense transcript was evaluated in parallel as a control. GECPAR was more frequently expressed in GCB- than ABC-DLBCL cell lines (11/16 vs. 0/8; P=0.001). In particular, it was expressed at high levels in five (OCI-LY1, OCI-LY1b, OCI-LY8, OCILY18, VAL), and at lower levels in six (SU-DHL-4, SUDHL-6, SU-DHL-16, SU-DHL-8, SU-DHL-10, TOLEDO) GCB-DLBCL cell lines. The transcript was barely detectable in the remaining five GCB and in all the eight ABC-DLBCL cell lines, while the antisense transcript was more broadly expressed in all cell lines (Figure 2A). We also evaluated GECPAR level in a total RNA-Seq 1134

dataset33 obtained from specimens derived from normal tonsil (n=31) and DLBCL patients (GCB, n=16; ABC, n=18). The transcript was significantly more expressed in normal cells compared to tumor cells, and, in accordance with our cell lines data, it was generally more abundant in GCB- than in ABC-DLBCL (Figure 2B). The higher GECPAR expression in GCB-DLBCL was confirmed in a validation cohort of 74 patients (GCB, n=31; non-GCB, n=43) (GSE145043) (Online Supplementary Figure S2A) and in a second one of 350 patients (GCB, n=183; ABC, n=167) (GSE10846). Variation of GECPAR expression in DLBCL cell lines and patients might be partially explained by its unstable genomic locus.34-36 A focal deletion of the chromosomal region containing the eRNA was observed in three of 737 mature lymphoid tumors37-41 (Online Supplementary Figure S2B). The normal tonsil derived cells were then subdivided according to B-cell maturation stage.42 GECPAR was most highly expressed by centroblasts while naïve B cells expressed the lowest levels. This observation further underlined the specific transcription of GECPAR in GC-derived cells. We also analyzed a catalog of murine lncRNA expressed in different developmental stages of B-cell maturation.43 Similar to our observations in humans, the murine GECPAR orthologue was mainly expressed in GC B cells, confirming the specific and conserved association of GECPAR with the GC B-cell transcriptional program (Online Supplementary Figure S2C). haematologica | 2022; 107(5)

GECPAR, eRNA master regulator of germinal center

Figure 2. Legend on following page.

haematologica | 2022; 107(5)

1135

S. Napoli et al.

Figure 2. GECPAR specific expression in germinal center B cell-like cells and correlation with essential genes. (A) Top: GECPAR expression in a panel of 22 diffuse large B-cell lymphoma (DLBCL) cell lines, 16 germinal center B cell-like (GCB) and 8 activated B cell-like, bottom, expression level of GECPAR antisense transcript, measured as control. (B) Top: box plots of GECPAR expression quantified by total RNA sequencing in normal individuals or GCB- or ABC-DLBCL patients. Bottom: box plots of GECPAR expression in normal individuals stratified for cell of origin. (C) Heat map of differential gene expression, in GCB-DLBCL cell lines dichotomized for GECPAR expression. (D) Preranked gene set enrichment analysis, in GCB-DLBCL cell lines classified for GECPAR expression. (E) Heat map of differential gene expression, in 16 GCB-DLBCL patients, classified for GECPAR expression. (F) Preranked gene set enrichment analysis, in DLBCL patients classified for GECPAR expression

Figure 3. Legend on following page.

1136

haematologica | 2022; 107(5)

GECPAR, eRNA master regulator of germinal center

Figure 3. GECPAR antiproliferative activity and activation of germinal center B cell-like transcriptional program. (A) Proliferation assay after interference with GECPAR by 4 different LNA antisense oligonucleotides in U2932, VAL and OCI-LY18. Average of 3 independent experiments, *P<0.05, **P<0.01. (B) Growth curve of SUDHL2 GFP+ and SUDHL2 Gecpar- GFP+, left, or OCI-Ly10 GFPbright and SUDHL2 Gecpar- GFPbright, right, measured by Incucyte. Average of 3 independent experiments, *P<0.05, **P<0.01. (C) Preranked gene set enrichment analysis (GSEA) of RNA sequencing data after GECPAR knockdown (KD) in U2932. (D) Preranked GSEA of RNA sequencing data in GECPAR overexpressing SUDHL2 respect to control. (E) GECPAR expression in 4 patient-derived tumor xenograft models (PDTX) derived from 2 activated B cell-like diffuse large B-cell lymphoma (ABC-DLBCL) and 2 germinal center B cell-like (GCB)-DLBCL patients. (F) Left: proliferation assay in PDTX-RN 5 days after GECPAR KD. Right: proliferation assay in PDTX-KD 9 days after GECPAR infection.

GECPAR expression correlates with cell cycle genes and the germinal center diffuse large B-cell lymphoma oncogenic signature In order to identify a gene expression signature associated with GECPAR, we focused on the 16 GCB-DLBCL cell lines with available expression profiling data44 and split them in two groups based on the median GECPAR expression. We identified 122 significantly upregulated and 73 downregulated genes (absolute log fold change ≥0.59 and P≤0.05), that could divide GCB-DLBCL cell lines into high and low GECPAR expressers (Figure 2C; Online Supplementary Table S1). Transcripts that were more expressed in GECPAR- high than in GEPCAR-low expressers showed a significant enrichment of cell cycle genes and essential cell survival genes, while genes involved in MAPK and PI3K pathways, as well as LEF1 targets were comparatively less enriched (Figure 2D). When we divided the 16 GCB-DLBCL patient specimens according to GECPAR expression GECPAR-high specimens showed an enrichment of cell cycle genes, particularly the G2M checkpoint as well as genes essential for cell survival (Figure 2E; Online Supplementary Table S2). Conversely, LEF1 targets and genes downstream of TGFβ and ATF2 were downregulated in DLBCL with high GECPAR expression (Figure 2F). Comparison of the genes associated with differential GECPAR expression in cell lines and clinical specimens (Online Supplementary Table S3; Online Supplementary Figure S2D and E) revealed that common genes were mainly involved in negative regulation of the cell cycle. Due to these observations, we hypothesized that GECPAR had an antiproliferative function. 2

GECPAR exhibits antiproliferative activity in diffuse large B-cell lymphoma cells In order to investigate the putative antiproliferative role of GECPAR we induced degradation of GECPAR using LNA oligonucleotides in VAL, OCI-LY18 and OCI-LY1, three GCB DLBCL cell lines with high level of GECPAR and U2932, an ABC-DLBCL with moderate GECPAR expression (Online Supplementary Figure S3A). After 24 hours we measured POU2AF1 mRNA and observed a negligible effect on its expression (Online Supplementary Figure S3B).Therefore, despite GECPAR transcription being dependent on activation of the same superenhancer (Online Supplementary Figure S3C and D) needed for POU2AF1 transcription (Online Supplementary Figure S3E), GECPAR itself was not essential for POU2AF1 transcription. Degradation of GECPAR led to an increase in cell proliferation in all the tested cell lines, suggesting a tumor suppressor function of GECPAR (Figures 3A; Online Supplementary Figure S3F to G). In order to further confirm the antiproliferative activity of GECPAR, we then overexpressed GECPAR in SUDHL2 and OCI-Ly10, two ABC cell lines with low GECPAR levels. The growth of stable GFP-positive GECPAR-expressing cells (Online Supplementary Figure S3H and I) was followed by imaging haematologica | 2022; 107(5)

in real time for 5 days. In both cell models, we measured a significant reduction in proliferation of cells overexpressing GECPAR compared to control infected cells (Figure 3B; Online Supplementary Figure S3J). In particular, OCI-Ly10 expressed very intense GFP fluorescence (Online Supplementary Figure S3I) and could grow as a monolayer on L-poly-ornithin-coated surface allowing monitoring the growth of cells with specific green fluorescence intensity. On the contrary, SUDHL2 tended to form clusters, despite of the L-poly-ornithin coating, and the instrument could hardly discriminate fluorescence from single cells over time. In that case, we could measure the cell growth by phase contrast image analysis, more accurately. The number of total cells and of GFP expressing cells counted at time 0 are reported in the Online Supplementary Figure S3K. As further confirmation, we analyzed GECPAR function also in two ABC- (PDTX-KD and PDTX-RRR) and two GCB- (PDTX-SS and PDTX-RN) DLBCL PDX models. We confirmed that GECPAR was higher in the two GCB than ABC cases (Figure 3E). Furthermore, we selected the PDX cells with the highest GECPAR expression (PDTX-RN) and we silenced GECPAR by LNA antisense oligonucleotides (Online Supplementary Figure S4A). GECPAR silencing increased the proliferation rate also in this model (Figure 3F; Online Supplementary Figure S4B). In addition, we overexpressed GECPAR in PDTX-KD cells, which had a very low amount of the transcript. We seeded the cells 24 hours (h) alter transduction and we monitored them (Online Supplementary Figure S4C). As for SUDHL2, although we could not monitor their growth along the whole experiment due to their tendency to form clusters, we measured GFP expression by fluorescence-activated cell sorting (FACS) (Online Supplementary Figure S4D), GECPAR expression by qRT-PCR (Online Supplementary Figure S4E) and cell viability by MTT assay (Figure 3F) after 9 days. As observed with ABC-DLBCL cell lines, also PDX cells, derived from an ABC-DBCL with low GECPAR expression, reduced their proliferation rate after GECPAR overexpression.

GECPAR polarizes cells towards a germinal center B cell-like transcriptional program We performed transcriptional analysis after GECPAR knockdown (KD) and overexpression in U2932 and SUDHL2 cells, respectively. Knockdown of GECPAR resulted in 1,099 significantly downregulated and 528 upregulated genes (Online Supplementary Table S4), while overexpression of GECPAR led to significant upregulation of 3,152 genes and downregulation of 787 genes (Online Supplementary Table S5). Genes upregulated after GECPAR silencing comprised proliferation genes, which were conversely downregulated in GECPAR-overexpressing cells. Further, while U2932, an ABC-DLBCL with moderate basal GECPAR expression still presented an enrichment of oncogenic genes typical of ABC-DLBCL after GECPAR knockdown (Figure 3D; Online Supplementary 1137

S. Napoli et al.

Figure 4. GECPAR has a favorable impact on the outcome of germinal center B cell-like diffuse large B-cell lymphoma patients. (A) Kaplan-Meier curves of diffuse large B-cell lymphoma (DLBCL) patients treated with R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, prednisone) and stratified for GECPAR expression. (B) Kaplan-Meier curves of germinal center B cell-like (GCB)- and activated B cell-like (ABC)-DLBCL patients treated with R-CHOP and stratified for GECPAR expression.

Table S6, left), the other ABC-DLBCL SUDHL2, showed an enrichment of GCB-DLBCL genes (Figure 3E; Online Supplementary Table S6, right), after GECPAR overexpression. Finally, GECPAR transcription was strongly induced by anti-IgM stimulation of the BCR (Online Supplementary Figure S3L). Together, these observations provided further support of GECPAR’s role in maintaining the GC transcriptional program.

GECPAR expression has favorable prognostic impact in germinal center diffuse large B-cell lymphoma patients We assessed the expression of GECPAR in 91 DLBCL patients treated with R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, prednisone) and its potential impact on the clinical outcome. We classified patients in three subgroups: low expressor (below the 15th percentile of the whole population), high expressor (over the 70th percentile), and neutral (in between). High 1138

expressor patients had a higher survival probability than low expressor (P=0.01) (Figure 4A). Then, we looked at the high and low expressors based on their cell of origin. As expected by previous analysis (Online Supplementary Figure S2A), the high expressors were mainly GCB patients. However, among GCB-DLBCL patients, cases with low GECPAR expression had the same risk of death as ABC patients, while the high expressors showed a better outcome (P=0.03). All together, these observations further sustain the tumor suppressor role we attributed to GECPAR based on our in vitro experiments.

GECPAR acts in trans regulating cell growth and differentiation by means of Wnt pathway In order to identify the genes directly regulated by GECPAR, we performed CHART-Seq in OCI-LY1 and U2932. We identified 4,172 peaks in OCI-LY1 and 692 peaks in U2932 (Figure 5A; Online Supplementary Tables S7 and S8). haematologica | 2022; 107(5)

GECPAR, eRNA master regulator of germinal center

Figure 5 GECPAR in trans transcriptional regulatory function. (A) Pipeline of CHART experiment and analysis. (B) Panther gene ontology classification of 325 GECPAR target genes identified both in OCI-LY1 and U2932 by CHART. (C) Preranked gene set enrichment analysis (GSEA) of RNA sequencing (RNA-Seq) data after GECPAR knockdown (KD) in U2932. (D) Preranked GSEA of RNA-Seq data after GECPAR overexpression in SUDHL2. (E) Top: Venn diagram crossing genes with GECPAR binding detected by CHART sequencing and significant expression modulation after GECPAR KD in U2932. Direct downregulated (left) and upregulated (right) GECPAR targets are listed. Bottom, preranked GSEA of direct GECPAR positively regulated targets, in germinal center B cell-like diffuse large B-cell lymphoma patients dichotomized for GECPAR expression

The most prominent peaks were validated in an independent CHART experiment by qRT, confirming the robustness of both the enrichment experiment and downstream analysis (Online Supplementary Figure S5B). As an additional control, we measured the levels of transcripts associated with GECPAR binding including CREBBP, CREB5, TLE4 and CYLD. After 24 h of GECPAR silencing with LNA oligonucleotides in U2932, the levels of these transcripts were reduced by 50-80% (Online Supplementary Figure S5C) and after 72 h we noticed a reduction of 50% also in the level of CYLD and TLE4 proteins (Online Supplementary Figure S5D, top). We also measured the increase in protein levels in SUDHL2 and OCI-Ly10 stably overexpressing GECPAR, for TLE4 and CYLD, or CREBBP and CYLD, respectively (Online Supplementary Figure S5D, bottom). haematologica | 2022; 107(5)

GECPAR capture was done with a set of probes, selected after RNAseH sensitivity assay (Online Supplementary Figure S5A). Only peaks called by two different algorithms (MACS and SPP) were taken in account: 4,172 in OCI-LY1 and 692 in U2932 (Figure 5A, Online Supplementary Tables S7 and S8). We identified a putative GECPAR binding motif. Among 78 CHARTseq peaks that fell within an interval of 10 kb in both cell lines there was a significant putative GECPAR binding motif (13 matches, P-values between 2.15x10-7 and 1.9x10-9) (Online Supplementary Figure S5E), In order to identify biological processes directly influenced by GECPAR independently of the cell of origin, we analysed 325 genes bound by the eRNA in both OCI-LY1 and U2932. The most significantly enriched classes of 1139

S. Napoli et al. A

Figure 6. Wnt inhibitor sensitivity in activated B cell-like diffuse large Bcell lymphoma cell lines in dependence of GECPAR expression. (A) Anticorrelation between AZ6102 log half-maximal inhibitory concentration (IC ) and GECPAR expression, left, or tankirase protein level, right, in activated B cell-like diffuse large B-cell lymphoma cell lines. (B) Cell cycle analysis in two different GECPAR overexpressing SUDHL2 clones and relative controls, exposed to dimethyl sulfoxide (DMSO) or 5 μM of AZ6102 for 48 hours. 50

genes belonged to the Wnt signaling pathway, cell growth and differentiation (Figure 5B). RNA-Seq data after GECPAR knockdown showed modulation of three pathways associated with development, differentiation and proliferation and known to cross-talk with the Wnt pathway, such as TGF β, NF-kB and MAPK (Figure 5C).4547 Negative regulators of TGF-β pathways including SMAD7, SMURF1 and SMURF2 (Online Supplementary Figure S6A) and negative regulators of MAPK signaling, DUSP1, DUSP8 and DUSP10 (Online Supplementary Figure S6B), were downregulated, after GECPAR silencing. Some of the downregulated genes belonging to the aforementioned pathways are also negatively regulated by NFkB (Online Supplementary Figure S6C). Notably, WNT and MAPK pathways were also affected in SUDHL2 cells overexpressing GECPAR (Figure 5D). Intersection of CHARTseq and RNA-Seq data for U2932 cells with GECPAR knockdown identified MYC and PRDM1 among seven genes negatively regulated by GECPAR, indicating that the eRNA influenced both the proliferative capability, reducing MYC, and the terminal differ1140

entiation to plasma cells, reducing PRDM1, the genes coding for BLIMP1. Interestingly, 21 direct GECPAR upregulated targets were positively correlated with GECPAR expression also in GCB-DLBCL specimens (Figure 5E). Among them there were KLF6, NOTCH2, components of BMP, cAMP and TNF-a pathways. Strikingly, we also identified TLE4 (Groucho), which forms a corepressor complex with TCF/LEF1 and recruits HDAC to inhibit transactivation of TCF/LEF1 target genes.48 Our identification of GECPAR involvement in Wnt signaling prompted us to evaluate the activity of the tankyrase 1/2 (TNKS1/2) inhibitor, AZ6102, that prevents nuclear translocation of β-catenin.49 For the four ABC-DLBCL cell lines we tested, GECPAR expression and sensitivity to AZ6102 were significantly anticorrelated (Figure 6A), suggesting that expression of GECPAR sensitized cells to Wnt pathway inhibition. All seven GCB-DLBCL cell lines tested where equally sensitive to Wnt pathway inhibition (Online Supplementary Figure S7). The differential sensitivity to AZ6102 in ABC-DLBCL was not related to tankyrase expression, since protein haematologica | 2022; 107(5)

GECPAR, eRNA master regulator of germinal center

levels were similar for the four cell lines (Figure 6A). Further, GECPAR overexpressing SUDHL2 cells were more sensitive to Wnt inhibition than the parental control, in terms of cell cycle perturbation. AZ6102 treatment more readily caused G2/M arrest, subG1 accumulation and decreased re-entry in G1 in GECPAR overexpressing cells (Figure 6B).

Discussion eRNA have recently started to be recognized as potent modulators of coding gene transcription.50,51 Here, we provide the first evidence of a lncRNA, transcribed in a SE specifically active during maturation of GC B cells, which plays an antiproliferative role in DLBCL models and is associated with favorable clinical outcome in GCBDLBCL patients. The lncRNA LOC100132078 was previously annotated as an unknown ncRNA, mainly expressed in lymph nodes and testis,52 and reported among p53-induced eRNA in breast cancer.53 Since it mapped inside a SE relevant for GC formation3,23,43 and in a site of recurrent genomic instability in lymphoid tumors,34-36 we elucidated its role in DLBCL, the neoplastic counterpart derived from GC B cells. We defined this lncRNA as eRNA according to the main features of this class of ncRNA: it was encoded within a SE; it was a non-polyA chromatin-associated transcript: its expression, highly cell type specific, was dependent on enhancer activation. We also identified a stabilized 970 nucleotide-long transcript, which, based on its expression pattern, we named GECPAR. It was less expressed in DLBCL samples than in normal tonsil B cells and in vitro experiments showed an inverse correlation with cell proliferation, suggesting an antitumoral function. The latter was further supported by the association between high GECPAR expression and favorable outcome in GCB DLBCL patients. GECPAR did not seem to act by in cis transactivation of the juxtaposed POU2AF1 gene, which is strongly expressed in GC-derived malignancies.22 Indeed, although GECPAR and POU2AF1 transcript levels were correlated in cell lines and in clinical specimens, silencing of the eRNA did not strongly impair expression of the coding gene. This is not uncommon and might be due to redundant functions of multiple enhancers that target a given promoter.54 On the contrary, GECPAR showed in trans activity and directly regulated the expression of several transcripts, mainly involved in cell growth and differentiation. These regulated genes were identified as common GECPAR targets in a GCBand an ABC- DLBCL cell line, both of which had constitutively high GECPAR expression. GECPAR expression was increased after BCR activation, an event that causes transcriptional reprograming of B cells. The exogenous overexpression of GECPAR in an ABC-DLBCL cell line confirmed its ability to switch the lymphoma cell towards the GCB-DLBCL transcriptional signature. Nuclear enriched lncRNA regulating transcription in trans have been described and they often modulate cell development.43,55 We propose that GECPAR is used by normal GC B cells to fine-tune the balance between proliferation and differentiation by directly repressing MYC and PRDM1 expression. MYC has a stage-specific role in the GC, particularly in light zone B cells, namely centrohaematologica | 2022; 107(5)

cytes, from which GCB-DLBCL tumor cells derive. After antigen-driven selection, B cells that still need to improve their antigen affinity can re-enter in the dark zone where they undergo additional cycles of somatic hypermutation. This so-called “cyclic re-entry” is critical for maintaining the GC and is induced by the re-expression of MYC via BCR activation through NF-kB and FOXO1.56,57 We propose GECPAR as a key surveillant of this process, as it directly represses MYC in that phase. Termination of the GC reaction is modulated by NF-kB activation downstream of the BCR. It induces IRF4, master regulator of terminal B-cell differentiation which in turn activates the plasma cell master regulator BLIMP1, encoded by PRDM1.58 GECPAR itself directly represses PRDM1, impeding terminal differentiation into plasma blast. In conclusion, GECPAR, which is induced by BCR activation, would retain B cells in the GC light zone, reducing the tendency to re-enter in the dark zone or to exit and differentiate to plasma cells. GECPAR also reduces B-cell proliferation rate and the tendency to differentiate, possibly by directly inducing TLE4, a negative repressor of TCF/LEF1. LEF1 is the key mediator of nuclear Wnt signaling and is important in lymphopoiesis. LEF1 is overexpressed in the nucleus of approximately 40% of DLBCL.59 MYC and Wnt pathway are connected in a positive feedback-loop involving LEF1.60 GECPAR, which directly inhibited MYC expression, indirectly enhanced its antiproliferative activity via TLE4 that contributed to the arrest of terminal differentiation induced by NF-kB. Indeed, GECPAR expression was inversely correlated with many LEF1 targets, in both DLBCL cell lines and specimens, and some of them were related to NF-kB regulation. Moreover, GECPAR silencing induced upregulation of important NF-kB genes, such as CARD11, REL and IKBKB, supporting the link between GECPAR and Wnt/NF-kB crosstalk. Several bidirectional connections between Wnt and NF-kB pathways45 have been reported in cancer and in particular, in DLBCL.61 We propose GECPAR as an additional layer of control of NFkB activation in GC B cells, pausing terminal differentiation to plasma blasts. The greater sensitivity of ABC-DLBCL with high GECPAR expression to pharmacological inhibition of Wnt further supports the relationship between GECPAR and Wnt pathway regulation and uncovers alternative therapeutic options for ABC-DLBCL patients. In conclusion, our work describes a novel mechanism of regulation of GC differentiation, which might contribute to DLBCL pathogenesis, and could help in understanding the heterogeneity of this disease. Disclosures LC received a travel grant from HTG. FB received institutional research funds from Acerta, ADC Therapeutics, Bayer AG, Cellestia, CTI Life Sciences, EMD Serono, Helsinn, ImmunoGen, Menarini Ricerche, NEOMED Therapeutics 1, Nordic Nanovector ASA, Oncology Therapeutic Development and PIQUR Therapeutics AG; received consultancy fees from Helsinn and Menarini; provided expert statements to HTG; received travel grants from Amgen, Astra Zeneca, Jazz Pharmaceuticals and PIQUR Therapeutics. All the other authors have nothing to disclose. Contributions SN performed experiments, analyzed and interpreted data, 1141

S. Napoli et al.

wrote the manuscript; SN and FB conceived and supervised the study, and extensively reviewed the manuscript; LC performed data mining; AR performed next generation sequencing; AAM reviewed the manuscript; FS, FG, FZ, GS performed experiments; MF, SB provided bioinformatic support; MTC, AC, PG

References 1. Loven J, Hoke HA, Lin CY, et al. Selective inhibition of tumor oncogenes by disruption of super-enhancers. Cell. 2013;153(2):320334. 2. Whyte WA, Orlando DA, Hnisz D, et al. Master transcription factors and mediator establish super-enhancers at key cell identity genes. Cell. 2013;153(2):307-319. 3. Zhang J, Vlasevska S, Wells VA, et al. The CREBBP acetyltransferase is a haploinsufficient tumor suppressor in B-cell lymphoma. Cancer Discov. 2017;7(3):322-337. 4. Kim TK, Hemberg M, Gray JM, et al. Widespread transcription at neuronal activity-regulated enhancers. Nature. 2010;465 (7295):182-187. 5. Arnold PR, Wells AD, Li XC. Diversity and emerging roles of enhancer RNA in regulation of gene expression and cell fate. Front Cell Dev Biol. 2019;7:377. 6. Li W, Notani D, Rosenfeld MG. Enhancers as non-coding RNA transcription units: recent insights and future perspectives. Nat Rev Genet. 2016;17(4):207-223. 7. Liu F. Enhancer-derived RNA: a primer. Genomics Proteomics Bioinformatics. 2017;15(3):196-200. 8. Jiao W, Chen Y, Song H, et al. HPSE enhancer RNA promotes cancer progression through driving chromatin looping and regulating hnRNPU/p300/EGR1/HPSE axis. Oncogene. 2018;37(20):2728-2745. 9. Li W, Notani D, Ma Q, et al. Functional roles of enhancer RNAs for oestrogen-dependent transcriptional activation. Nature. 2013;498 (7455):516-520. 10. Fanucchi S, Shibayama Y, Burd S, et al. Chromosomal contact permits transcription between coregulated genes. Cell. 2013;155 (3):606-620. 11. Tan SH, Leong WZ, Ngoc PCT, et al. The enhancer RNA ARIEL activates the oncogenic transcriptional program in T-cell acute lymphoblastic leukemia. Blood. 2019;134 (3):239-251. 12. Marques AC, Hughes J, Graham B, et al. Chromatin signatures at transcriptional start sites separate two equally populated yet distinct classes of intergenic long noncoding RNAs. Genome Biol. 2013;14(11):R131. 13. Alvarez-Dominguez JR, Knoll M, Gromatzky AA, et al. The super-enhancerderived alncRNA-EC7/Bloodlinc potentiates red blood cell development in trans. Cell Rep. 2017;19(12):2503-2514. 14. Agirre X, Meydan C, Jiang Y, et al. Long non-coding RNAs discriminate the stages and gene regulatory states of human humoral immune response. Nat Commun. 2019;10(1):821. 15. Verma A, Jiang Y, Du W, et al. Transcriptome sequencing reveals thousands of novel long non-coding RNAs in B cell lymphoma. Genome Med. 2015;7:110. 16. Mlynarczyk C, Fontan L, Melnick A. Germinal center-derived lymphomas: the darkest side of humoral immunity. Immunol Rev. 2019;288(1):214-239.

1142

and GI provided clinical samples data; OE and GI provided transcriptome data. All authors read and edited the manuscript. Funding Partially supported by funds from Gelu Foundation (to BF).

17. Pasqualucci L, Dalla-Favera R. Genetics of diffuse large B-cell lymphoma. Blood. 2018;131(21):2307-2319. 18. Wright GW, Huang DW, Phelan JD, et al. A probabilistic classification tool for genetic subtypes of diffuse large B cell lymphoma with therapeutic implications. Cancer Cell. 2020;37(4):551-568. 19. Chapuy B, Stewart C, Dunford AJ, et al. Molecular subtypes of diffuse large B cell lymphoma are associated with distinct pathogenic mechanisms and outcomes. Nat Med. 2018;24(5):679-690. 20. Ceribelli M, Kelly PN, Shaffer AL, et al. Blockade of oncogenic IkappaB kinase activity in diffuse large B-cell lymphoma by bromodomain and extraterminal domain protein inhibitors. Proc Natl Acad Sci U S A. 2014;111(31):11365-11370. 21. Reddy A, Zhang J, Davis NS, et al. Genetic and functional drivers of diffuse large B cell lymphoma. Cell. 2017;171(2):481-494. 22. Teitell MA. OCA-B regulation of B-cell development and function. Trends Immunol. 2003;24(10):546-553. 23. Chapuy B, McKeown MR, Lin CY, et al. Discovery and characterization of superenhancer-associated dependencies in diffuse large B cell lymphoma. Cancer Cell. 2013;24(6):777-790. 24. Gaudio E, Tarantelli C, Spriano F, et al. Targeting CD205 with the antibody drug conjugate MEN1309/OBT076 is an active new therapeutic strategy in lymphoma models. Haematologica. 2020;105(11):25842591. 25. Napoli S, Piccinelli V, Mapelli SN, et al. Natural antisense transcripts drive a regulatory cascade controlling c-MYC transcription. RNA Biol. 2017;14(12):1742-1755. 26. Vance KW. Mapping long noncoding RNA chromatin occupancy using capture hybridization analysis of RNA targets (CHART). Methods Mol Biol. 2017;1468:3950. 27. Sexton AN, Machyna M, Simon MD. Capture hybridization analysis of DNA targets. Methods Mol Biol. 2016;1480:87-97. 28. Kharchenko PV, Tolstorukov MY, Park PJ. Design and analysis of ChIP-seq experiments for DNA-binding proteins. Nat Biotechnol. 2008;26(12):1351-1359. 29. Zhang Y, Liu T, Meyer CA, et al. Modelbased analysis of ChIP-Seq (MACS). Genome Biol. 2008;9(9):R137. 30. Vance KW, Sansom SN, Lee S, et al. The long non-coding RNA Paupar regulates the expression of both local and distal genes. EMBO J. 2014;33(4):296-311. 31. Consortium EP. An integrated encyclopedia of DNA elements in the human genome. Nature. 2012;489(7414):57-74. 32. Pefanis E, Wang J, Rothschild G, et al. RNA exosome-regulated long non-coding RNA transcription controls super-enhancer activity. Cell. 2015;161(4):774-789. 33. Teater M, Dominguez PM, Redmond D, et al. AICDA drives epigenetic heterogeneity and accelerates germinal center-derived lymphomagenesis. Nat Commun. 2018;9(1): 222.

34. Poretti G, Kwee I, Bernasconi B, et al. Chromosome 11q23.1 is an unstable region in B-cell tumor cell lines [Research Support, Non-U.S. Gov't]. Leuk Res. 2011;35(6):808813. 35. Auer RL, Starczynski J, McElwaine S, et al. Identification of a potential role for POU2AF1 and BTG4 in the deletion of 11q23 in chronic lymphocytic leukemia. Genes Chromosomes Cancer. 2005;43(1):110. 36. Zhao C, Inoue J, Imoto I, et al. POU2AF1, an amplification target at 11q23, promotes growth of multiple myeloma cells by directly regulating expression of a B-cell maturation factor, TNFRSF17. Oncogene. 2008;27(1):63-75. 37. Chigrinova E, Rinaldi A, Kwee I, et al. Two main genetic pathways lead to the transformation of chronic lymphocytic leukemia to Richter syndrome. Blood. 2013;122(15): 2673-2682. 38. Boi M, Rinaldi A, Kwee I, et al. PRDM1/BLIMP1 is commonly inactivated in anaplastic large T-cell lymphoma. Blood. 2013;122(15):2683-2693. 39. Martinez N, Almaraz C, Vaque JP, et al. Whole-exome sequencing in splenic marginal zone lymphoma reveals mutations in genes involved in marginal zone differentiation. Leukemia. 2014;28(6):1334-1340. 40. Rinaldi A, Kwee I, Young KH, et al. Genome-wide high resolution DNA profiling of hairy cell leukaemia. Br J Haematol. 2013;162(4):566-569. 41. Rossi D, Trifonov V, Fangazio M, et al. The coding genome of splenic marginal zone lymphoma: activation of NOTCH2 and other pathways regulating marginal zone development. J Exp Med. 2012;209(9):15371551. 42. De Silva NS, Klein U. Dynamics of B cells in germinal centres. Nat Rev Immunol. 2015;15(3):137-148. 43. Brazao TF, Johnson JS, Muller J, et al. Long noncoding RNAs in B-cell development and activation. Blood. 2016;128(7):e10-19. 44. Tarantelli C, Gaudio E, Arribas AJ, et al. PQR309 is a novel dual PI3K/mTOR inhibitor with preclinical antitumor activity in lymphomas as a single agent and in combination therapy. Clin Cancer Res. 2018;24 (1):120-129. 45. Ma B, Hottiger MO. Crosstalk between Wnt/beta-Catenin and NF-kappaB signaling pathway during inflammation. Front Immunol. 2016;7:378. 46. Guo X, Wang XF. Signaling cross-talk between TGF-beta/BMP and other pathways. Cell Res. 2009;19(1):71-88. 47. Grumolato L, Liu G, Haremaki T, et al. betaCatenin-independent activation of TCF1/LEF1 in human hematopoietic tumor cells through interaction with ATF2 transcription factors. PLoS Genet. 2013;9(8): e1003603. 48. Daniels DL, Weis WI. Beta-catenin directly displaces Groucho/TLE repressors from Tcf/Lef in Wnt-mediated transcription activation. Nat Struct Mol Biol. 2005;12(4):364-371.

haematologica | 2022; 107(5)

GECPAR, eRNA master regulator of germinal center

49. Johannes JW, Almeida L, Barlaam B, et al. Pyrimidinone nicotinamide mimetics as selective tankyrase and wnt pathway inhibitors suitable for in vivo pharmacology. ACS Med Chem Lett. 2015;6(3):254-259. 50. Franco HL, Nagari A, Malladi VS, et al. Enhancer transcription reveals subtype-specific gene expression programs controlling breast cancer pathogenesis. Genome Res. 2018;28(2):159-170. 51. Zhang Z, Lee JH, Ruan H, et al. Transcriptional landscape and clinical utility of enhancer RNAs for eRNA-targeted therapy in cancer. Nat Commun. 2019;10(1): 4562. 52. Fagerberg L, Hallstrom BM, Oksvold P, et al. Analysis of the human tissue-specific expression by genome-wide integration of transcriptomics and antibody-based proteomics. Mol Cell Proteomics. 2014;13(2):

haematologica | 2022; 107(5)

397-406. 53. Leveille N, Melo CA, Rooijers K, et al. Genome-wide profiling of p53-regulated enhancer RNAs uncovers a subset of enhancers controlled by a lncRNA. Nat Commun. 2015;6:6520. 54. Li X, Fu XD. Chromatin-associated RNAs as facilitators of functional genomic interactions. Nat Rev Genet. 2019;20(9):503-519. 55. Guttman M, Donaghey J, Carey BW, et al. lincRNAs act in the circuitry controlling pluripotency and differentiation. Nature. 2011;477(7364):295-300. 56. Pasqualucci L. Molecular pathogenesis of germinal center-derived B cell lymphomas. Immunol Rev. 2019;288(1):240-261. 57. Luo W, Weisel F, Shlomchik MJ. B cell receptor and CD40 signaling are rewired for synergistic induction of the c-Myc transcription factor in germinal center B cells. Immunity.

2018;48(2):313-326. 58. Boi M, Zucca E, Inghirami G, et al. PRDM1/BLIMP1: a tumor suppressor gene in B and T cell lymphomas. Leuk Lymphoma. 2015;56(5):1223-1228. 59. Tandon B, Peterson L, Gao J, et al. Nuclear overexpression of lymphoid-enhancer-binding factor 1 identifies chronic lymphocytic leukemia/small lymphocytic lymphoma in small B-cell lymphomas. Mod Pathol. 2011;24(11):1433-1443. 60. Hao YH, Lafita-Navarro MC, Zacharias L, et al. Induction of LEF1 by MYC activates the WNT pathway and maintains cell proliferation. Cell Commun Signal. 2019;17(1):129. 61. Bognar MK, Vincendeau M, Erdmann T, et al. Oncogenic CARMA1 couples NF-kappaB and beta-catenin signaling in diffuse large Bcell lymphomas. Oncogene. 2016;35(32): 4269-4281.

1143

ARTICLE Ferrata Storti Foundation

Haematologica 2022 Volume 107(5):1144-1152

Non-Hodgkin Lymphoma

Prephase rituximab/prednisone therapy and aging-related, proinflammatory cytokine milieu in older, vulnerable patients with newly diagnosed diffuse large B-cell lymphoma Richard J. Lin,1,2* Colette N. Owens,1,2* Esther Drill,3 Augustine Iannotta,1 Mayan Oliveros,1 Dylan L. Schick,1 Ariela Noy,1,2 John F. Gerecitano,1,2 Pamela R. Drullinsky,1,2 Philip C. Caron,1,2 Anita Kumar,1,2 Matthew J. Matasar,1,2 Craig Moskowitz,1,2 Beatriz Korc-Grodzicki,2,4 Andrew D. Zelenetz,1,2 Gilles A. Salles,1,2 and Paul A. Hamlin1,2 1

Department of Medicine, Division of Hematologic Malignancies, Memorial Sloan Kettering Cancer Center; 2Department of Medicine, Weill Cornell Medical College; 3 Epidemiology and Biostatistics, Memorial Sloan Kettering Cancer Center and 4 Department of Medicine, Geriatrics Service, Memorial Sloan Kettering Cancer Center, New York, NY, USA *RJL and CNO contributed equally as co-first authors.

ABSTRACT

Correspondence: PAUL A. HAMLIN hamlinp@mskcc.org Received: March 6, 2021. Accepted: July 12, 2021. Pre-published: July 22, 2021. https://doi.org/10.3324/haematol.2021.278719

1144

iffuse large B-cell lymphoma (DLBCL) predominantly affects older adults with suboptimal therapeutic outcomes due to increased treatment-related mortality and toxicities in vulnerable patients, clinically defined by geriatric impairments such as functional limitation, multimorbidity, or cognitive deficits. In this prospective pilot study, we evaluated a rituximab/prednisone prephase treatment strategy in 33 older, vulnerable patients with newly diagnosed DLBCL, defined by either age ≥70 years or age 60-70 years with Karnofsky performance scale (KPS) <80. A single dose of rituximab 375 mg/m2 between 3-10 days and oral prednisone for at least 5 days prior to the first dose of chemoimmunotherapy was administered. All patients completed prephase treatment and all but one commenced anthracycline-based chemoimmunotherapy. Only one early cycle death occurred. Toxicity events, defined by either unplanned hospitalization, unplanned dose reduction/delay, or chemotherapy discontinuation, occurred in 22 patients (67%). Sixteen patients (48%) experienced grade 3 or higher non-hematologic toxicities and/or grade 4 or higher hematologic toxicities. With a median follow-up of 4.4 years, both 5-year progression-free survival and overall survival were at 81% (95% confidence interval: 69-96). Importantly, we found that phenotypic impairments in basic and instrumental activities of daily living, physical function, mobility, KPS, and Cancer and Aging Research Group chemotherapy toxicity risk score were significantly associated with senescence-associated, proinflammatory cytokine milieu which was readily reversed with prephase treatment, potentially explaining its clinical effectiveness. Prephase therapy with rituximab/prednisone should be considered for all older, vulnerable DLBCL patients prior to curative intent, anthracycline-based chemoimmunotherapy. This trial was registered as clinicaltrials gov. Identifier: NCT 89028394.

Introduction Diffuse large B-cell lymphoma (DLBCL) disproportionally affects older patients and improving their therapeutic outcomes remains an unmet medical need. Epidemiologic studies have shown that even in the rituximab era, many older patients either do not receive or receive suboptimal dose and/or duration of chemoimmunotherapy to achieve a curative intent.1,2 While the biology of disease may be more aggressive,3 older patients commonly have multimorbidity, functional and/or cognitive impairment, or overt frailty that limits the delivery of upfront, cur-

haematologica | 2022; 107(5)

Prephase therapy in older patients with DLBCL

ative chemoimmunotherapy.4,5 Moreover, if not adequately addressed, these non-oncologic geriatric issues may exacerbate treatment-related toxicities, trigger functional decline, and adversely impact subsequent therapies such as hematopoietic cell transplantation and cellular therapy.4-6 Therefore, it is essential that older patients with newly diagnosed DLBCL receive adequate assessment and management of their aging-related issues. Biologically, inferior outcomes in older non-Hodgkin lymphoma (NHL) patients may result from preexisting geriatric deficits, the lymphoma itself, or treatment-related toxicities. Induction chemoimmunotherapy may worsen health status when it results in toxicities or improve it by controlling disease related impairments. The German Non-Hodgkin Lymphoma Study Group (DSHNHL) found that after the initiation of the NHL-B2 trial, toxic mortality was most common in the first and second cycle of therapy.7 They subsequently pioneered a strategy of “prephase” therapy with prednisone 100 mg daily for 5-7 days with or without vincristine 1 mg single dose prior to the initiation of full dose combination chemotherapy. This was incorporated into the latter part of the NHL-B2 trial, RICOVER60 trial, and the LYSA group LNH097B trial with fewer toxic deaths reported.7-9 The mechanism of this effect has not been examined in detail, although it is thought that prephase therapy improves functional status and physiologic reserve by reducing tumor burden. However, although vincristine is delivered by a simple 10-minute intravenous push, it is among the more toxic agents with significant risks of neuropathy and constipation resulting in its frequent dose reduction or omission.10 Geriatric assessment (GA) is increasingly incorporated into the care of older cancer patients to help guide treatment decision-making, predict toxicities, and manage non-oncologic geriatric issues. Multidimensional GA identifies otherwise unrecognized health problems among unselected older patients beyond traditional Karnofsky performance scale (KPS) and includes function status, comorbidity, mobility, cognition, nutrition, and psychosocial status.11,12 A largely self-administered GA instrument, the Cancer Aging Research Group (CARG) chemotherapy toxicity risk score incorporates 11 variables to predict high-grade, chemotherapy-related toxicities for older solid tumor patients.13,14 Several GA domains have been examined in small cohort studies of lymphoma patients, yet it remains unclear how they could be integrated into and improve outcomes in the context of curative intent chemoimmunotherapy.15 It has been long postulated that the mechanism underlying phenotypic and functional aging is related to perturbations in several biochemical and cellular pathways.16 One of them is cellular senescence, a state of stable growth arrest once cells are subjected to significant stress and have accumulated enough DNA damage.17 Senescence cells create a highly dynamic and persistent program of senescence-associated secretory phenotype (SASP), consisting of abundant secretion of proinflammatory proteins into the tissue microenvironment that modulates cancer immune surveillance and therapeutic response.18 Identification of important SASP components may generate novel targets to restore immune therapy responsiveness and enhance treatment outcomes.19 In this prospective pilot study, we examine the feasibility and safety of a novel prephase treatment with rituximab/prednisone for older, vulnerable patients with newly diaghaematologica | 2022; 107(5)

nosed DLBCL and its impact on meaningful toxicity outcomes, CARG-based GA measures, and the SASP-related, proinflammatory cytokine milieu.

Methods Trial design and schema We conducted a pilot study of rituximab/prednisone prephase treatment for older patients with newly diagnosed DLBCL prior to curative intent, anthracycline-based chemoimmunotherapy within our larger prospective observational study (Figure 1). Patients were eligible if they were aged 70 years or older, or were 60-70 years old with KPS <80. The protocol was approved by the Institutional Review Board at the Memorial Sloan Kettering Cancer Center and conducted according to the Declaration of Helsinki. All patients received a single dose of rituximab 375 mg/m2 given between 3 and 10 days and oral prednisone for at least 5 days during the 14 days prior to the first cycle. The preferred prednisone dose was 100 mg daily for 7 days with minimal allowable dose of 50 mg daily for 5 days.

Geriatric assessment CARG-based GA was performed at baseline, following prephase treatment and before the first cycle, and prior to each subsequent cycle as previously described.13,14 The CARG chemotherapy risk calculator provided both a risk score (range, 0-19) and a corresponding absolute chemotherapy-toxicity risk percentage. The clinician portion of the GA included baseline patient and disease characteristics, clinician-rated KPS, memory screening test short Blessed Orientation-Memory-Concentration (BOMC),20 Timed Up and Go (TUG), and Mini-Nutrition Assessment (MNA).21

Outcome and toxicities assessment The primary outcome was the composite of toxicity or severe toxicity events (TE). TE was defined as any of the following: i) hospitalization during or within 30 days following chemotherapy; ii) dose delay or reduction to a dose intensity ≤80% of the planned dose intensity; iii) discontinuation of chemotherapy due to toxicity. Severe TE (STE) was defined as the occurrence of either i) or iii) above. Secondary toxicity endpoints were: i) grade 3 or higher non-hematologic toxicity; ii) grade 4 or higher hematologic toxicity. Toxicity was graded according to CTCAE v4.0.3.

Proinflammatory cytokine analysis Immunoassays of the levels of proinflammatory cytokines were batch performed in duplicate on singly thawed plasma samples on a commercial multiplex cytokine panel including IL6, IL-1b, IL-2, IL-4, IL-5, IL-8, IL-10, IL-12, TNF-a, and IFN-g (Miso Discovery Laboratory). These cytokines were chosen given commercial availabilities on multiplex plates and technical expertise at our institution.

Statistical analyses In order to compare pre- and posttreatment GA scales and cytokine (CK) levels, one-sided pairwise Wilcoxon tests were used. CK levels were log-transformed before analyses. Associations between baseline GA, ΔGA, baseline CK, ΔCK and toxicities were assessed with one-sided Wilcoxon rank sum tests. Associations between baseline GA, ΔGA, baseline CK, ΔCK and stage were assessed with two-sided Wilcoxon rank sum tests. Associations between toxicities and stage were assessed with Fishers’ Exact test. Correlations between baseline

1145

R.J. Lin et al.

GA measures and CK levels were assessed with the Spearman’s rank correlation coefficient. Progression-free survival (PFS) was calculated from treatment initiation to disease progression or death. Overall survival (OS) was calculated from diagnosis to death. All tests were corrected for multiple testing with the false discovery rate (q-value) at a significance level of 0.05. Statistical analyses were conducted in R Version 4.0.0.

Results Patient characteristics Baseline characteristics are summarized in Table 1. The cohort consisted of 33 patients with a median age of 75 years (range, 65–87) and ten patients (30%) were 80 years old or older. Nearly two thirds of patients were stage III and IV at diagnosis and more than half had intermediate2/high risk-disease based on age-adjusted International Prognostic Index (AA-IPI). All patients completed prephase treatment with rituximab and prednisone, with 26 patients full 100 mg dose and 7 patients 50 mg doses. One patient did not receive chemoimmunotherapy following prephase treatment due to rapid disease progression. For the 32 patients who initiated chemoimmunotherapy, 23 of 25 patients (92%) completed all six planned cycles; six of six completed all four planned cycles; and there was one early death due to rapid disease progression. Baseline GA revealed that 23 patients (70%) had intermediate to high comorbidity burden. There were significant functional limitations in ADL (median 65; range, 0– 100), IADL (median 14, range, 0–14), and social activities (median 50; range 25–75). The median TUG time was 11.31 seconds (range, 5–40), with <10 seconds denoting no mobility limitation. The median MNA score was 23 (range, 10.5–30), with 15 patients (45%) considered at risk for malnutrition or malnourished (score >24). The median CARG chemotherapy toxicity risk score was 10 (range, 6– 17), corresponding to an absolute, grade 3+ chemotherapy-related toxicity risk of 54% (range, 32–89), both in the high-risk category.

Outcomes and toxicities As shown in Figure 2, with a median follow-up of 4.4 (range, 0.4–5.7) years, both 5-year PFS and OS for the cohort were 81% (95% confidence interval [CI]: 69–96). PFS and OS according to AA-IPI is also shown in Figure 2. Among stage III/IV patients (n=20), the 5-year PFS and OS were both 74% (95% CI: 57–97, data not shown). Six patients died, including four from relapse/progression of disease and two while in remission. We summarized toxic events and high-grade CTATE toxicities through all treatment cycles in Table 2. TE and STE occurred in 22 patients (67%) and 12 patients (36%), respectively. The majority of TE was dose reduction/delay, occurring in 19 patients (58%). There were seven hospitalizations following cycle 1. Grade 3+ non-hematologic toxicities occurred in 16 patients (48%). The total number of grade 3+ nonhematologic toxicity was 26, including ten infections (3 following cycle 1), five cardiac toxicities, three electrolyte/metabolic toxicities, three gastrointestinal/nutritional toxicities, two hemorrhage/thrombosis, one infusion reaction, and one psychiatric illness. Grades 3+ and 4+ hematologic toxicities occurred in 23 patients (70%) and six patients (18%), respectively. Most toxicities, 17 (65%), was in cycle 1-3. In total, 16 patients (48%) had at least one grade 3+ non-hematological toxicity or grade 4+ hematologic toxicity. Toxicity events did not differ significantly by stage of disease at diagnosis (Online Supplementary Table S1).

Geriatric assessment and senescence-associated secretory phenotype-associated proinflammatory cytokines We examined the senescence-associated, proinflammatory cytokine milieu in our cohort of older patients and its relationship to clinical geriatric impairments. As shown in Figure 3 and the Online Supplementary Table S2, we found that levels of several proinflammatory cytokines were associated with individual geriatric impairment in older lymphoma patients. Specifically, elevated interleukin (IL)2, IL-6, IL-10, and TNF-a levels were significantly associated with reduced KPS, functional limitation measured by

Figure 1. Study design. Prephase pilot embedded within a prospective, large cohort study of geriatric assessment in older patients with newly diagnosed non-Hodgkin lymphoma. DLBCL: diffuse large B-cell lymphoma; GA: geriatric assessment; NHL: non-Hodgkin lymphoma; KPS: Karnofsky performance scale; R-CHOP: rituximab, cyclophosphamide, doxorubicin hydrochloride, vincristine sulfate, and prednisone.

1146

haematologica | 2022; 107(5)

Prephase therapy in older patients with DLBCL

ADL and IADL, and social activity limitation. They were also significantly associated with increased TUG time, CARG chemotherapy toxicity risk, and baseline LDH level. We next examined the impact of prephase treatment on GA measures and CK levels. As shown in Table 3, prephase therapy led to significant reduction in levels of IL-10 (q=0.033), IL-4 (q=0.04), IL-6 (q=0.01), and TNF-a (q=0.01). The magnitude of changes was most pronounced for TNF-a and IL-6 (Online Supplementary Figure S1). Prephase therapy did not lead to significant changes in GA measures within the 2-week period (Table 3),

Table 1. Baseline characteristics.

Total cohort (N=33) Age, years, median (range) Female sex, n (%) Stage, n (%) I/II III/IV Cell of origin (n=31), Hans, n (%) GCB Non-GCB Histology, n (%) De novo Transformed follicular Richter’s transformation Age-adjusted IPI, n (%) Low/Int-1 Int-2/High Induction regimen, n (%) R-CHOP R-mini-CHOP R-EPOCH No treatment Comorbidity, n (%) Low Intermediate High Clinician rated KPS, median (range) Patient rated KPS, median (range) ADL score, median (range) IADL score, median (range) Activity limitation score, median (range) Number of falls last 6 months, median (range) TUG in seconds, median (range) Cognition score (BOMC), median (range) Mini-nutritional assessment, median (range) CARG score, median (range) CARG % risk, median (range)

75 (65-87) 20 (61) 13 (39) 20 (61) 18 (58) 13 (42)

Association of geriatric assessment and cytokines with outcomes Finally, we examined the association of both baseline and changes in GA measures and CK levels following prephase therapy with outcomes. As shown in Table 4, we did not find a significant association of any baseline or changes in GA measures with the development of TE, although there was a trend toward significance for baseline CARG risk score (P=0.040) and the absolute toxicity risk percentage (P=0.019) prior to correction for multiple testing. Similarly, there was no significant association between baseline or changes in cytokine levels with the development of TE, although there was a trend toward significance for baseline IL-10 (P=0.012) and IL-13 (P=0.006) prior to correction for multiple testing (Online Supplementary Table S5). Due to the small number of PFS/OS events (n=6), we did not assess the association of GA measures and CK levels with these outcomes.

Discussion 29 (88) 3 (9) 1 (3) 14 (42) 19 (58) 29 (91) 1 (3) 2 (6) 1 10 (30) 17 (52) 6 (18) 80 (40-100) 80 (40-100) 65 (0-100) 14 (2-14) 50 (25-75) 0 (0-9) 11.31 (5-40) 2 (0-10) 23 (10.5-30) 10 (6-17) 54 (32-89)

GCB: germinal center B-cell type; IPI: international prognostic index; R-CHOP: Rituximab, Cyclophosphamide, Doxorubicin hydrochloride, Vincristine sulfate, Prednisone; R-EPOCH: Rituximab, Etoposide phosphate, Prednisone, Vincristine sulfate, Cyclophosphamide, Doxorubicin hydrochloride; KPS: Karnofsky performance scale; ADL: activities of daily living; IADL: instrumental activities of daily living; TUG: timed-get-up and go; BOMC: Blessed Orientation Memory Concentration; CARG: cancer and aging research group.

haematologica | 2022; 107(5)

although there was a trend of improvement in clinician rated KPS and CARG chemotherapy toxicity risk score prior to correction for multiple testing. Neither baseline or changes in GA measures or CK levels differed by early versus later stage of disease (Online Supplementary Tables S3 and S4).

Improving outcomes for older, vulnerable patients with aggressive lymphoma has remained a challenge over the last few decades. In this study, we examined a novel rituximab/prednisone prephase therapy for older, vulnerable DLBCL patients prior to curative, anthracycline-based chemoimmunotherapy. This prephase therapy was feasible with acceptable toxicity profiles. Most patients in this vulnerable cohort completed the planned 4-6 cycles of treatment, although dose reduction/delays were common. Although conclusions are limited by the pilot nature of the study with small sample size, the favorable 5-year survival of over 80% was comparable to selected historical Table 2. Toxicity events and high-grade toxicities.

Category of events Toxicity events (unplanned hospitalization, chemotherapy discontinuation, and/or dose reduction/delay) Severe toxicity events (unplanned hospitalization and/or chemotherapy discontinuation) Early induction death (prior to cycle 3) Category of toxicities Grade 3+ hematologic toxicities Grade 4+ hematologic toxicities Grade 3+ non-hematologic toxicities Grade 3+ non-hematologic toxicities plus Grade 4+ hematologic toxicities

Number of patients (%) 22 (67)

12 (36)

1 (3) 23 (70) 6 (18) 16 (48) 16 (48)

Total numbers of grade 3+ non-hematologic toxicities: infection (10); cardiac (5); electrolyte/metabolic (3); gastrointestinal/nutrition (3); bleeding/thrombosis (2); infusion reaction (1); psychiatric (1).

1147

R.J. Lin et al.

cohorts.7-9 It should also be noted that based on the CARG risk score, our cohort was much frailer than the original CARG patient populations with more than half in the high-risk group.13,14 Moreover, we found that a senescence-associated, proinflammatory CK milieu, which was

readily reversed by prephase treatment, was significantly associated with GA-defined, prognostically important geriatric impairments in function, mobility, and chemotherapy toxicity risk. While our sample size and event rate are too small to definitively associate GA meas-

Figure 2. Survival outcomes. (A) Kaplan-Meier survival estimate of progression-free survival (PFS) (red-line with shaded area denoted 95% confidence interval). Life table is listed below. (B) KaplanMeier survival estimate of overall survival (OS) (redline with shaded area denoted 95% confidence interval). Life table is listed below. (C) Kaplan-Meier survival estimate of PFS by AA-IPI (red-line denoted low/low-intermediate and blue-line denoted highintermediate/high categories). Life table is listed below. (D) Kaplan-Meier survival estimate of OS by AA-IPI (red-line denoted low/low-intermediate and blue-line denoted highintermediate/high categories). Life table is listed below.

1148

haematologica | 2022; 107(5)

Prephase therapy in older patients with DLBCL

ures and CK levels with survival, some of these measures may warrant additional investigation in larger cohorts of patients. Since the NHL-02 trial, prephase therapy with vincristine and prednisone has since been adopted in several trials.7-9 It has been speculated that prephase therapy

improves performance status thus reducing early cycle toxic death and allowing full-intensity chemoimmunotherapy as illustrated in the LYSA phase II trial.9 A recent prospective study also supported this notion by demonstrating an improved KPS and reduced incidence of febrile neutropenia with prephase therapy.22 However, the

Table 3. Impact of prephase therapy on geriatric assessment measures and proinflammatory cytokines.

Measures

LDH (log level) Clinician rated KPS Patient rated KPS ADL IADL Activity limitation TUG CARG score % risk Cytokines (log level) IFN-g IL-10 IL-12 IL-13 IL1-β IL-2 IL-4 IL-6 IL-8 TNF-a

30 32 32 31 31 31 29 32 32 N 30 30 30 30 30 29 30 30 28 30

Pre- to post-prephase therapy changes Changes (median, IQR) P-value q-value 0.01 (-0.24, 0.18) 0 (0, 10) 0 (0, 10) 0 (-10, 7.5) 0 (-0.5, 0) 0 (-12.5, 3.1) -1 (-2, 1.01) 0 (-2, 0) 0 (-4, 0) Changes (median, IQR) -0.06 (-1.61, 0.85) -0.5 (-1.59, 0.2) -0.09 (-0.38, 0.27) 0 (-1.19, 1.95) 0 (0, 1.19) 0.19 (-0.39, 1.68) -0.25 (-0.84, 0.16) -0.61 (-1.49, 0.05) -0.33 (-0.91, 0.76) -0.65 (-1.49, -0.03)

0.317 0.039 0.420 0.483 0.637 0.978 0.105 0.044 0.135

0.571 0.198 0.621 0.621 0.717 0.978 0.304 0.198 0.304

0.208 0.010 0.306 0.514 0.953 0.890 0.016 0.002 0.307 0.001

0.416 0.033 0.439 0.643 0.953 0.953 0.04 0.01 0.439 0.01

IQR: interquartile range; LDH: lactate dehydrogenase; KPS: Karnofsky performance scale; ADL: activities of daily living; IADL: instrumental activities of daily living; TUG: timed-getup and go; CARG: cancer and aging research group; IFN: interferon; IL: interleukin; TNF: tumor necrosis factor..

Table 4. Association of baseline and changes in geriatric assessment measures with toxicity events.

Characteristic (IQR)

At least 1 toxic event, N = 22

No toxic events, N = 11

P-value

q-value

Baseline LOG LDH Baseline cKPS Baseline pKPS Baseline ADL Baseline IADL Baseline ACTIVITY LIMIT Baseline TUG Baseline CARG score Baseline % risk Δ in LOG LDH Δ in cKPS Δ in pKPS Δ in ADL Δ in IADL Δ in ACTIVITY LIMIT Δ in TUG Δ in CARG score Δ in % risk

33 33 33 33 33 33 31 33 33 30 32 32 31 31 31 29 32 32

5.60 (5.51, 5.96) 75 (70, 90) 80 (62, 98) 48 (26, 84) 13.00 (9.50, 14.00) 47 (44, 56) 11 (10, 20) 10.50 (9.00, 13.00) 66 (54, 89) 0.01 (-0.17, 0.17) 0 (0, 10) 0 (0, 10) 5 (-5, 10) 0.00 (-1.00, 0.00) 0 (-14, 6) -1.1 (-4.2, 1.1) 0.00 (-3.00, 0.00) 0 (-12, 0)

5.40 (5.31, 5.89) 90 (70, 90) 90 (80, 90) 80 (32, 92) 14.00 (9.00, 14.00) 56 (38, 62) 11 (9, 13) 9.00 (7.00, 10.50) 54 (52, 54) -0.20 (-0.26, 0.17) 0 (0, 15) 0 (0, 5) 0 (-10, 0) 0.00 (0.00, 0.00) 0 (-12, 0) -0.3 (-1.4, 0.8) 0.00 (-1.00, 0.50) 0 (0, 0)

0.14 0.2 0.3 0.13 0.2 0.2 0.2 0.040 0.019 0.2 0.3 0.6 0.8 0.3 0.5 0.7 0.8 >0.9

0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.4 0.3 0.5 0.5 0.7 0.9 0.5 0.7 0.8 0.9 >0.9

IQR: interquartile range; LDH: lactate dehydrogenase; cKPS: clinician-rated karnofsky performance scale; pKPS: patient-rated Karnofsky performance scale; ADL: activities of daily living; IADL: instrumental activities of daily living; TUG: timed-get-up and go; CARG: cancer and aging group.

haematologica | 2022; 107(5)

1149

R.J. Lin et al.

Figure 3. Heatmap of significant correlations of baseline geriatric assessment measures with baseline proinflammatory cytokine levels (q<0.05) with blue color demonstrating significant positive association and red color demonstrating significant negative association. LDH: lactate dehydrogenase; KPS: Karnofsky performance scale; ADL: activities of daily living; IADL: instrumental activities of daily living; TUG: timed-get-up and go; CARG: cancer and aging group; IFN: interferon; IL: interleukin; TNF: tumor necrosis factor.

underlying mechanism has never been examined and vincristine has significant neurotoxicity especially for older adult thus making it a poor partner of prednisone.10 Rituximab is likely more effective as a cytoreductive and debulking agent than vincristine at a dose of 1 mg; concurrently, however, we acknowledge concern for rituximab toxicities which include infusion related reactions that at times can be severe. We showed that this regimen led to an acceptable rate of 48% grade 3+ non-hematologic plus grade 4+ hematologic toxicities in this vulnerable cohort of older patients who were at high risk for chemotherapyrelated toxicities. Importantly, most patients in our cohort were able to complete planned cycles of curative, anthracycline-based chemoimmunotherapy and there was one early cycle death. Therefore, it is not surprising that our cohort of patients had outstanding PFS and OS. However, while we had hoped to see a decrease in mortality/toxicity from the prephase treatment, we could not directly compare survival and toxicity results from this pilot study with small sample size to previous large trials and registry studies.7-9,22,23 In addition, with this design we cannot ascertain if rituximab’s addition to prednisone added benefit, although the additional dose of rituximab did not appear to add toxicity. Among hematologic malignancies including lymphoma, GA domains including function, mobility, cognition, and comorbidity have been consistently shown to be associated with survival and/or treatment-related toxicities.24-27 Our study is the first to examine specifically the CARG chemotherapy risk score in lymphoma patients and its 1150

longitudinal changes. The CARG score is easy to derive and well validated in solid tumors, however, its dynamic changes in response to therapy is unknown.13,14 In our study, we were unable to demonstrate a significant change in GA measures pre- and post-prephase therapy. It is possible that the short time interval, less than 2 weeks on average, is not adequate to detect a significant change in GA domains such as functional status. Indeed, a recent study in acute myeloid leukemia patients measured GA changes 8-12 weeks following the initial assessment.28 Alternatively, our sample size may be too small or that prephase treatment alone is inadequate to impact GA changes. Interestingly, there was a suggestion of an association between the baseline CARG risk score and the absolute chemotherapy toxicity risk with TE, which will be explored further. Nevertheless, given accumulating evidence of how geriatric frailty affects older lymphoma patients,29,30 our longitudinal GA data may allow in-depth examination of the prognostic impact of both baseline and changes in individual geriatric deficits, and may also be used prospectively to guide treatment-decision making in older, vulnerable lymphoma patients as shown in previous studies.31-34 Perhaps the most interesting aspect of our findings is the SASP-associated, proinflammatory cytokine milieu and its relationship to GA and prephase therapy. A few key cytokines in this age-related SASP such as IL-6, TNFa , IL-10, and IL-2 were strongly associated with geriatric impairments including functional limitations in haematologica | 2022; 107(5)

Prephase therapy in older patients with DLBCL

ADL/IADL, mobility impairment measured by TUG, and CARG chemotherapy toxicity risk score. This finding suggests a novel, chronic inflammation-based, biologic underpinning of geriatric frailty in older lymphoma patients and several potential mechanism-based intervention strategies. Most importantly, this proinflammatory cytokine milieu was readily reversed with prephase therapy which may account for its effectiveness. Some of these cytokines were also associated with baseline lactate dehydrogenase, suggesting that tumor burden may also contribute to geriatric frailty and the SASP, consistent with a previous study showing that tumor debulking could improve lymphoma patients’ KPS.35 We could not, however, ascertain the contribution of disease-related factors in our study. Nevertheless, there are several additional implications. First, if validated, these specific cytokines such as TNF-a, IL-6, and IL-10 could potentially serve as frailty biomarkers for older lymphoma patients. This notion is supported by previous findings where IL-6 and TNF-a were found to be associated with KPS, cytopenia, OS, and PFS.36-38 These markers are biologically and mechanistically more specific than c-reactive protein or albumin.9,39 It is also possible that elevated proinflammatory cytokines are caused by a combination of disease, comorbidities, KPS, and/or other host factors, in addition to cellular senescence. These possibilities will need to be examined in a large study. Second, the use of senolytics to deplete senescence cells is a potential strategy that could potentially reverse the SASP-associated frailty phenotype in the short term. Commercially available senolytics such as desatinib/quercetin have been studied in an early phase human trial.40,41 Finally, specific inhibitors of these frailty CK are available commercially. One of them, siltuximab, has been tested in a phase I trial for patients with hematologic malignancies including NHL. The drug was well tolerated with no dose-limiting toxicity and sustained suppression of CRP was observed.42 Although we examined most key inflammatory cytokines, not all SASP-associated proteins were analyzed in this study and thus our results may not be generalizable to other components of SASP. In conclusion, we show that rituximab/prednisone is a feasible and safe prephase regimen, which may have enhanced the delivery of chemoimmunotherapy for older, vulnerable patients with newly diagnosed DLBCL. Our findings of acceptable short-term toxicities and excellent long-term survival are hypothesis generating and mechanistically appealing since this strategy appears to target the aging-related SASP and the proinflammatory CK milieu. While requiring validation from a prospective randomized study which should also examine individual components of the prephase regimen, we propose that prephase therapy with rituximab/prednisone is considered for older, vulnerable NHL patients starting curative-intent chemoimmunotherapy. We also strongly advocate for incorporating GA into the care of older lym-

References 1. Hamlin PA, Satram-Hoang S, Reyes C, Hoang KQ, Guduru SR, Skettino S. Treatment patterns and comparative effectiveness in elderly diffuse large B-cell lymphoma patients: a surveillance, epidemiology, and end results-medicare analysis.

haematologica | 2022; 107(5)

phoma patients given its wealth of prognostic information and potential value in improving treatment tolerance. Disclosures RJL sits on the advisory board for Kite, a Gilead Company; PAH has received research support and consultancy fees from Portola Pharmaceuticals, Inc.; AN has received research funding from Pharmacyclics and Raphael, honoraria from Prime Oncology, Medscape and EUSA, and sits on the advisory Board for Janssen; ADZ has received consultancy fees from Genentech/Roche, Gilead, Celgene, Janssen, Amgen, Novartis and Adaptive Biotechnology, has received research funding from MEI Pharmaceuticals, Roche, Gilead and Beigene, sits on the advisory Board for MorphoSys, Gilead, Genentech, Abbvie and AstraZeneca, serves as DMC chair for Beigene and is DMC member for BMS/Celgene/Juno; MJM has receieved consultancy fees or sits on the advisory boards of Genentech, Roche, GlaxoSmithKline, Bayer, Merck, Pharmacyclics, Janssen, Seattle Genetics, Takeda, Teva, Juno Therapeutics, Rocket Medical, June Therapeutics, Immunovaccine Technologies, and Daiichi Sankyo, has received research funding from Genentech Roche, GlaxoSmithKline, IGM Biosciences, Bayer, Pharmacyclics, Janssen, Rocket Medical, Seattle Genetics, and Immunovaccine Technology; GAS acts as a consultant and sits on the advisory board for Abbvie, Allogene, Autolus, Beigene, BMS/Celgene, Debiopharm, Genmab, Kite/Gilead, Janssen, Milteniy, Morphosys, Norvartis, Roche, and Velosbio. All other authors declare no conflict of interests. Contributions PAH, CNO, RJL, and BK designed the research, interpreted the data, and wrote the paper; RJL, CNO, AI, MO, DLS, and ED collected and analyzed the data; PAH, CNO, ADZ, AN, DJS, SMH, AJM, PCC, AMH, AK, MJM, MLP, CLB, AY, JFG, and PRD contributed the data. All authors reviewed and approved the manuscript. Acknowledgements We acknowledge administrative assistance provided by Shreena Patel. Funding This research was supported in part by the NIH/NCI Cancer Center Support grant P30 CA008748 and the Lacher Lymphoma Foundation. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Data sharing statement Deidentified individual participant data that underlie the reported results will be made available 3 months after publication for a period of 5 years after the publication date upon request. The study protocol is included as a data supplement available with the online version of this article.

Oncologist. 2014;19(12):1249-1257. 2. Levin E, Peng Y, Ji Y, Gilbertson D, Morrison VA. Practice patterns in older patients with diffuse large B-cell lymphoma: a medicare analysis, 2007-2015. J Geriatr Oncol. 2020;11(8):1344-1348. 3. Mareschal S, Lanic H, Ruminy P, Bastard C, Tilly H, Jardin F. The proportion of activated B-cell like subtype among de novo dif-

fuse large B-cell lymphoma increases with age. Haematologica. 2011;96(12):18881890. 4. Soubeyran PL, Cordoba R. Approaches for vulnerable and frail older patients with diffuse large B-cell lymphomas. Curr Opin Oncol. 2019;31(5):369-373. 5. Khan Y, Brem EA. Considerations for the treatment of diffuse large B cell lymphoma

1151

R.J. Lin et al. in the elderly. Curr Hematol Malig Rep. 2019;14(4):228-238. 6. Morrison VA, Hamlin P, Soubeyran P, et al. Approach to therapy of diffuse large B-cell lymphoma in the elderly: the International Society of Geriatric Oncology (SIOG) expert position commentary. Ann Oncol. 2015;26(6):1058-1068. 7. Pfreundschuh M, Trumper L, Kloess M, et al. Two-weekly or 3-weekly CHOP chemotherapy with or without etoposide for the treatment of young patients with good-prognosis (normal LDH) aggressive lymphomas: results of the NHL-B1 trial of the DSHNHL. Blood. 2004;104(3):626-633. 8. Pfreundschuh M, Schubert J, Ziepert M, et al. Six versus eight cycles of bi-weekly CHOP-14 with or without rituximab in elderly patients with aggressive CD20+ Bcell lymphomas: a randomised controlled trial (RICOVER-60). Lancet Oncol. 2008; 9(2):105-116. 9. Peyrade F, Bologna S, Delwail V, et al. Combination of ofatumumab and reduceddose CHOP for diffuse large B-cell lymphomas in patients aged 80 years or older: an open-label, multicentre, single-arm, phase 2 trial from the LYSA group. Lancet Haematol. 2017;4(1):e46-e55. 10. Madsen ML, Due H, Ejskjaer N, Jensen P, Madsen J, Dybkaer K. Aspects of vincristine-induced neuropathy in hematologic malignancies: a systematic review. Cancer Chemother Pharmacol. 2019; 84(3):471-485. 11. Mohile SG, Dale W, Somerfield MR, et al. Practical assessment and management of vulnerabilities in older patients receiving chemotherapy: ASCO Guideline for Geriatric Oncology. J Clin Oncol. 2018; 36(22):2326-2347. 12. DuMontier C, Loh KP, Bain PA, et al. Defining undertreatment and overtreatment in older adults with cancer: a scoping literature review. J Clin Oncol. 2020; 38(22):2558-2569. 13. Hurria A, Togawa K, Mohile SG, et al. Predicting chemotherapy toxicity in older adults with cancer: a prospective multicenter study. J Clin Oncol. 2011;29(25):34573465. 14. Hurria A, Mohile S, Gajra A, et al. Validation of a prediction tool for chemotherapy toxicity in older adults with cancer. J Clin Oncol. 2016;34(20):23662371. 15. Lin RJ, Behera M, Diefenbach CS, Flowers CR. Role of anthracycline and comprehensive geriatric assessment for elderly patients with diffuse large B-cell lymphoma. Blood. 2017;130(20):2180-2185. 16. Lopez-Otin C, Blasco MA, Partridge L, Serrano M, Kroemer G. The hallmarks of aging. Cell. 2013;153(6):1194-1217. 17. Gorgoulis V, Adams PD, Alimonti A, et al. Cellular senescence: defining a path forward. Cell. 2019;179(4):813-827.

1152

18. Tchkonia T, Zhu Y, van Deursen J, Campisi J, Kirkland JL. Cellular senescence and the senescent secretory phenotype: therapeutic opportunities. J Clin Invest. 2013; 123(3):966-972. 19. Faget DV, Ren Q, Stewart SA. Unmasking senescence: context-dependent effects of SASP in cancer. Nat Rev Cancer. 2019; 19(8):439-453. 20. Katzman R, Brown T, Fuld P, Peck A, Schechter R, Schimmel H. Validation of a short Orientation-Memory-Concentration Test of cognitive impairment. Am J Psychiatry. 1983;140(6):734-739. 21. Guigoz Y, Vellas B. The Mini Nutritional Assessment (MNA) for grading the nutritional state of elderly patients: presentation of the MNA, history and validation. Nestle Nutr Workshop Ser Clin Perform Programme. 1999;1:3-11; discussion 11-12. 22. Lakshmaiah KC, Asati V, Babu KG, et al. Role of prephase treatment prior to definitive chemotherapy in patients with diffuse large B-cell lymphoma. Eur J Haematol. 2018;100(6):644-648. 23. Juul MB, Jensen PH, Engberg H, et al. Treatment strategies and outcomes in diffuse large B-cell lymphoma among 1011 patients aged 75 years or older: a Danish population-based cohort study. Eur J Cancer. 2018;99:86-96. 24. Klepin HD. Ready for prime time: role for geriatric assessment to improve quality of care in hematology practice. Blood. 2019; 134(23):2005-2012. 25. Abel GA, Klepin HD. Frailty and the management of hematologic malignancies. Blood. 2018;131(5):515-524. 26. Liu MA, DuMontier C, Murillo A, et al. Gait speed, grip strength, and clinical outcomes in older patients with hematologic malignancies. Blood. 2019;134(4):374-382. 27. Hshieh TT, Jung WF, Grande LJ, et al. Prevalence of cognitive impairment and association with survival among older patients with hematologic cancers. JAMA Oncol. 2018;4(5):686-693. 28. Saad M, Loh KP, Tooze JA, et al. Geriatric assessment and survival among older adults receiving postremission therapy for acute myeloid leukemia. Blood. 2020; 136(23):2715-2719. 29. Tucci A, Martelli M, Rigacci L, et al. Comprehensive geriatric assessment is an essential tool to support treatment decisions in elderly patients with diffuse large B-cell lymphoma: a prospective multicenter evaluation in 173 patients by the Lymphoma Italian Foundation (FIL). Leuk Lymphoma. 2015;56(4):921-926. 30. Merli F, Luminari S, Rossi G, et al. Outcome of frail elderly patients with diffuse large Bcell lymphoma prospectively identified by Comprehensive Geriatric Assessment: results from a study of the Fondazione Italiana Linfomi. Leuk Lymphoma. 2014; 55(1):38-43.

31. Bai JF, Han HX, Feng R, et al. Comprehensive geriatric assessment (CGA): a simple tool for guiding the treatment of older adults with diffuse large B-cell lymphoma in China. Oncologist. 2020;25(8):e1202-e1208. 32. Olivieri A, Gini G, Bocci C, et al. Tailored therapy in an unselected population of 91 elderly patients with DLBCL prospectively evaluated using a simplified CGA. Oncologist. 2012;17(5):663-672. 33. Spina M, Balzarotti M, Uziel L, et al. Modulated chemotherapy according to modified comprehensive geriatric assessment in 100 consecutive elderly patients with diffuse large B-cell lymphoma. Oncologist. 2012;17(6):838-846. 34. Merli F, Luminari S, Tucci A, et al. Simplified geriatric assessment in older patients with diffuse large B-cell lymphoma: the Prospective Elderly Project of the Fondazione Italiana Linfomi. J Clin Oncol. 2021;39(11):1214-1222. 35. Bowcock SJ, Fontana V, Patrick HE. Very poor performance status elderly patients with aggressive B cell lymphomas can benefit from intensive chemotherapy. Br J Haematol. 2012;157(3):391-393. 36. Voog E, Bienvenu J, Warzocha K, et al. Factors that predict chemotherapy-induced myelosuppression in lymphoma patients: role of the tumor necrosis factor ligandreceptor system. J Clin Oncol. 2000; 18(2):325-331. 37. Dlouhy I, Filella X, Rovira J, et al. High serum levels of soluble interleukin-2 receptor (sIL2-R), interleukin-6 (IL-6) and tumor necrosis factor alpha (TNF) are associated with adverse clinical features and predict poor outcome in diffuse large B-cell lymphoma. Leuk Res. 2017;59:20-25. 38. Zhong H, Chen J, Cheng S, et al. Prognostic nomogram incorporating inflammatory cytokines for overall survival in patients with aggressive non-Hodgkin's lymphoma. EBioMedicine. 2019;41:167-174. 39. Loh KP, Tooze JA, Nicklas BJ, et al. Inflammatory biomarkers, geriatric assessment, and treatment outcomes in acute myeloid leukemia. J Geriatr Oncol. 2020; 11(3):410-416. 40. Justice JN, Nambiar AM, Tchkonia T, et al. Senolytics in idiopathic pulmonary fibrosis: Results from a first-in-human, open-label, pilot study. EBioMedicine. 2019;40:554563. 41. Chen Y, Liu S, Leng SX. Chronic low-grade inflammatory phenotype (CLIP) and senescent immune dysregulation. Clin Ther. 2019;41(3):400-409. 42. Kurzrock R, Voorhees PM, Casper C, et al. A phase I, open-label study of siltuximab, an anti-IL-6 monoclonal antibody, in patients with B-cell non-Hodgkin lymphoma, multiple myeloma, or Castleman disease. Clin Cancer Res. 2013;19(13):36593670.

haematologica | 2022; 107(5)

ARTICLE

Non-Hodgkin Lymphoma

Dose-adjusted EPOCH and rituximab for the treatment of double expressor and double-hit diffuse large B-cell lymphoma: impact of TP53 mutations on clinical outcome Anna Dodero,1* Anna Guidetti,1* Fabrizio Marino,2 Alessandra Tucci,3 Francesco Barretta,4 Alessandro Re,3 Monica Balzarotti,5 Cristiana Carniti Chiara Monfrini,1 Annalisa Chiappella,1 Antonello Cabras,6 Fabio Facchetti,7 Martina Pennisi,1 Daoud Rahal,8 Valentina Monti,6 Liliana Devizzi,1 Rosalba Miceli,4 Federica Cocito,9 Lucia Farina,1 Francesca Ricci,5 Giuseppe Rossi,3 Carmelo Carlo-Stella2#and Paolo Corradini1,10#

Ferrata Storti Foundation

Haematologica 2022 Volume 107(5):1153-1162

Department of Hematology, Fondazione IRCCS Istituto Nazionale dei Tumori, Milano; Department of Biomedical Sciences, Humanitas University and Department of Oncology and Hematology, IRCCS Humanitas Research Hospital, Rozzano-Milano; 3 Department of Hematology, ASST Spedali Civili di Brescia, Brescia; 4Department of Clinical Epidemiology and Trial Organization, Fondazione IRCCS Istituto Nazionale dei Tumori, Milano; 5Department of Oncology and Hematology, IRCCS Humanitas Research Hospital, Rozzano-Milano; 6Department of Pathology, Fondazione IRCCS Istituto Nazionale dei Tumori, Milano; 7Department of Pathology, ASST Spedali Civili di Brescia, Brescia; 8Department of Pathology, IRCCS Humanitas Research Hospital, RozzanoMilano; 9Department of Hematology, Ospedale San Gerardo, Monza and 10Chair of Hematology, University of Milan, Milano, Italy 2

*AD and AG contributed equally as co-first authors. # CC-S and PC contributed equally as co-senior authors.

ABSTRACT

iffuse large B-cell lymphoma (DLBCL) is a heterogeneous disease, including one-third of cases overexpressing MYC and BCL2 proteins (double expressor lymphoma, DEL) and 5-10% of patients with chromosomal rearrangements of MYC, BCL2 and/or BCL-6 (double/triple-hit lymphomas, DH/TH). TP53 mutations are detected in 2025% of DEL. We report the efficacy of dose-adjusted EPOCH and rituximab (DA-EPOCH-R) in a series of 122 consecutive patients, including DEL (n=81, 66%), DEL-MYC (n=9, 7%), DEL-BCL2 (n=13, 11%), or high-grade lymphomas (DH/TH) (n=19, 16%). Central nervous system (CNS) prophylaxis included intravenous methotrexate (n=66), intrathecal chemotherapy (IT) (n=40) or no prophylaxis (n=16). Sixty-seven patients (55%) had highintermediate or high International Prognostic Index (IPI) and 30 (25%) had high CNS-IPI. The 2-year progression-free survival (PFS) and overall survival (OS) for the entire study population were 74% and 84%, respectively. There was a trend for inferior OS for DH/TH (2-year OS: 66%, P=0.058) as compared to all the others. The outcome was significantly better for the IPI 0-2 versus IPI 3-5 (OS: 98% vs. 72%, P=0.002). DA-EPOCH-R did not overcome the negative prognostic value of TP53 mutations: 2-year OS of 62% versus 88% (P=0.036) were observed for mutated as compared to wild-type cases, respectively. Systemic CNS prophylaxis conferred a better 2-year OS (94%) as compared to IT or no prophylaxis (76% and 65%, respectively; P=0.008). DA-EPOCH-R treatment resulted in a favorable outcome in patients with DEL and DEL with single rearrangement, whereas those with multiple genetic alterations such as DEL-DH/TH and TP53 mutated cases still have an inferior outcome.

Introduction Diffuse large B-cell lymphoma (DLBCL) is a clinically and biologically heterogeneous disease. Historically DLBCL patients have been uniformly treated with RCHOP chemoimmunotherapy regimen (rituximab, cyclophosphamide, adri-

haematologica | 2022; 107(5)

Correspondence: ANNA DODERO anna.dodero@istitutotumori.mi.it Received: February 26, 2021. Accepted: July 13 2021. Pre-published: July 22, 2021. https://doi.org/10.3324/haematol.2021.278638

1153

A. Dodero et al.

amycin, vincristine, prednisone), leading to a long-lasting complete remission in approximately 60% of cases.1,2 For these patients, the International Prognostic Index (IPI), is an easy and valid tool for prognostic stratification.3 In 2000 Alizadeh et al. used gene expression profiling (GEP) to identify the cell of origin (COO) and described prognostic subgroups according to diversity in gene expression patterns indicative of different stages of B-cell differentiation.4 Immunohistochemical algorithms, used as surrogates of GEP for the identification of COO subgroups, have been extensively used but have a limited concordance with GEP when applied to patients with DLBCL treated with R-CHOP.5 Among other possible prognostic indicators for patients with DLBCL, tumor protein p53 (TP53) mutations seem to represent simple and attractive biomarkers to be used in the daily routine clinical practice. The TP53 gene is involved in maintaining genomic stability in response to DNA damage by activating DNA repair programs and by triggering cell cycle arrest. The loss of TP53 is associated with lymphomagenesis and resistance to chemotherapy.6 TP53 mutations are present in 10% of DLBCL patients and confer a poor prognosis in patients treated with RCHOP.7,8 Recently, Chapuy et al. performed whole genomic sequencing in 304 primary DLBCL patients and identified genetically different subgroups among germinal center B-cell like (GCB) and activated B-cell (ABC) lymphoma patients. Patients with TP53 mutations were described as a distinct cluster with a very poor prognosis.9 MYC and BCL2 represent other possible poor prognostic markers when rearranged or overexpressed in DLBCL and lymphomas harboring the double or triple translocation (DH or TH) represent new entities in the recently updated 2016 World Health Organization Lymphoma Classification. Most DH lymphomas are in the GCB category, they present with advanced stages and have a very poor prognosis when treated with the R-CHOP regimen.10-12 Also, the BCL2 translocation alone has been shown to play a significant prognostic role in GCB DLBCL patients treated with R-CHOP.13,14 Additionally, the presence of MYC rearrangements seems to be a poor prognostic marker, although its role in the absence of either BCL2 or p53 alterations remains controversial.15 The concomitant overexpression of MYC and BCL2 on the tumor cell surface observed in double expressor lymphomas (DEL), could be the result of other mechanisms, different from translocation such as copy gain/amplification. The prognosis of DEL patients treated with standard R-CHOP is worse than that of non-DEL patients,16,17 and these individuals have a risk of central nervous system (CNS) relapse of 9.7% at 2 years.18 Moreover, a consensus on the optimal treatment for these patients has yet to be established. Since 2013 at our Institution, DEL patients have been treated with the intensive chemotherapy regimen doseadjusted EPOCH and rituximab (DA-EPOCH-R), achieving a promising 2-year progression-free survival (PFS) and overall survival (OS) of 62% and 85%, respectively. These survival rates were better than those reported with R-CHOP in a historical cohort.19 The aim of the present study was to assess the incidence and prognostic role of TP53 mutations and BCL2, MYC translocations in a large cohort of DEL patients consecutively treated with DA-EPOCH-R. The analysis of

1154

the cumulative risk of CNS relapse as well as the effect of the different CNS prophylaxis applied in this selected cohort of DEL patients was also assessed.

Methods Patients All consecutive patients with a diagnosis of DEL treated since 2013 with DA-EPOCH-R at four hematological divisions in northern Italy were retrospectively identified. For inclusion in this observational study, the following inclusion criteria were required: i) histologically proven diagnosis of DEL with an age ≥18 years; ii) availability of formalin-fixed, paraffin-embedded (FFPE) samples; iii) no exposure to previous therapy except for the first cycle of R-CHOP that was allowed while waiting for immunohistochemistry (IHC) results and cytogenetic characterization. Exclusion criteria were HIV positivity, CNS involvement at diagnosis, and histology other than DEL. The ethics committees of the participating centers approved the study (INT 35/17). Written informed consent was obtained from all patients.

Immunohistochemistry and fluorescence in situ hybridization analysis FFPE tissue samples were sectioned at 3-μm thickness. IHC was performed using the EnVision FLEX+, mouse, high pH method (Dako Denmark A/S Produktionsvej 42 DK-2600 Glostrup, Denmark) and a Dako Autostainer Link48 (Dako, Italia, SPA, Milano, Italy). Slides were stained with monoclonal antibodies against CD19, CD20, CD10, BCL2, BCL6, MUM1, MYC, and Ki67. Fluorescence in situ hybridization (FISH) analyses for BCL2, BCL6, and MYC rearrangements were performed in all patients using “LSI BCL2 LSI BCL6 (ABR), C-MYC “break apart” probes (Vysis/Abbott Molecular, Illinois, USA) according to the manufacturer’s instructions (detailed in the Online Supplementary Appendix).

TP53 mutations TP53 is located on chromosome 17p13.1 and consists of 14 exons (1-11), 10 of which are coding sequences for the p53 protein.20,21 TP53 mutations were assessed according to Institutional practice, Sanger sequencing was used for samples collected from patients at Humanitas Cancer Center while next-generation sequencing followed by direct sequencing was used for all other samples (procedures described in the Online Supplementary Appendix).

Treatment and central nervous system prophylaxis Patients received the DA-EPOCH-R regimen therapy every 21 days for six cycles. In all patients dose adjustment based on cell counts between cycles according to the NCI algorithm was applied.22 At diagnosis, all the patients performed lumbar puncture for cytology and flow cytometry analyses of cerebrospinal fluid (CSF). During this first procedure, all patients received intrathecal chemotherapy with methotrexate (MTX) and cytarabine. Neuro-imaging was considered only in case of the presence of neurological signs or symptoms. In absence of definitive clinical guidelines on CNS prophylaxis, the choice was determined by the treating physician. High-dose MTX (HD-MTX: 3 g/m2) was used as prophylaxis in 66 patients (for the majority at the end of the treatment). Intrathecal chemotherapy at day 1 of every cycle, including MTX/cytarabine/dexamethasone, was given to 40 patients.

haematologica | 2022; 107(5)

DA-EPOCH-R in double expressor lymphoma

Aim of the study and statistical analysis Objectives of the study were assessment of the effect of biological variables (TP53 mutations and BCL2 and MYC translocations) on survival analyses at 2-years in a large cohort of DEL patients consecutively treated with DA-EPOCH-R. Analysis of the cumulative risk of CNS recurrence in the

whole population and PFS and OS impact of different CNS prophylaxis were also evaluated. Fisher’s exact test was performed to assess the association between TP53 mutation status and patients’ clinical characteristics. Statistical analyses were performed using R (version 3.5.0) (detailed in the Online Supplementary Appendix).

Table 1. Clinical characteristics of 122 consecutive diffuse large B-cell lymphoma patients overall and according to rearrangements.

Age, continuous Median (years) third/first quartile Age, categorical ≤60 years old >60 years old Rearrangements None DEL-BCL2 DEL-MYC DEL-DH/TH Ki67 (%)* Median third/first quartile Sex Male Female Cell of origin GCB Non-GCB Not assessed Staging I-II III-IV IPI 0-2 3-5 CNS-IPI 0-1 2-3 4-6 Extranodal site (at risk for CNS) Yes No CNS prophylaxis None IT MTX IV MTX Autologous SCT Yes No

Overall N=122

DEL N=81

BCL2 N=13

MYC N=9

HG-BCL N=19

59.0 (49.0; 65.0)

57.0 (46.0; 64.0)

64.0 (55.0; 67.0)

57.0 (37.0; 61.0)

62.0 (57.0; 67.0)

66 (54.1) 56 (45.9)

48 (59.3) 33 (40.7)

5 (38.5) 8 (61.5)

5 (55.6) 4 (44.4)

8 (42.1) 11 (57.9 )

81 (66.4) 13 (10.7) 9 (7.4) 19 (15.6)

81 (100) -

13 (100) -

9 (100) -

19 (100)

90.0 (75.0; 90.0)

90.0 (80.0; 90.0)

90.0 (85.0; 95.0)

90.0 (81,3; 91.3)

75.0 (65.0; 87.5)

75 (61.5) 47 (38.5)

49 (60.5) 32 (39.5)

7 (53.8) 6 (46.2)

6 (66.7) 3 (33.3)

13 (68.4) 6 (31.6)

55 (45.1) 60 (49.2) 7 (5.7)

22 (27.2) 52 (64.2) 7 (8.6)

10 (76.9) 3 (23.1) 0 (0.0)

5 (55.6) 4 (44.4) 0 (0.0)

18 (94.7) 1 (5.3) 0 (0.0)

27 (22.1) 95 (77.9)

18 (22.2) 63 (77.8)

2 (15.4) 11 (84.6)

4 (44.4) 5 (55.6)

3 (15.8) 16 (84.2)

55 (45.1) 67 (54.9)

40 (49.4) 41 (50.6)

4 (30.8) 9 (69.2)

6 (66.7) 3 (33.3)

5 (26.3) 14 (73.7)

31 (25.4) 61 (50.0) 30 (24.6)

21 (25.9) 47 (58.0) 13 (16.0)

3 (23.1) 5 (38.5) 5 (38.5)

5 (55.6) 2 (22.2) 2 (22.2)

2 (10.5) 7 (36.8) 10 (52.6)

16 (13.1) 106 (86.9)

10 (12.3) 71 (87.7)

0 (0.0) 13 (100.0)

4 (44.4) 5 (55.6)

2 (10 .5) 17 (89.5)

16 (13.1) 40 (32.8) 66 (54.1)

12 (14.8) 24 (29.6) 45 (55.6)

0 (0.0) 5 (38.5) 8 (61.5)

1 (11.1) 3 (33.3) 5 (55.6)

3 (15.8) 8 (42.1) 8 (42.1)

22 (18.0) 100 (82.0)

12 (14.8) 69 (85.2)

3 (23.1) 10 (76.9)

1 (11.1) 8 (88.9)

6 (31.6) 13 (68.4)

DEL: double expressor lymphomas; HG-BCL: high-grade B-cell lymphomas; DH/TH: double-hit/triple-hit; GCB: germinal center lymphomas; Non-GCB: non-germinal center lymphomas; IPI: International prognostic index; CNS: central nervous system; IT: intrathecal; MTX: methotrexate; IV: intravenous; SCT: stem cell transplantation. *10 missing values, 8 in DEL group and 1 each in DEL/BCL2 and DEL/MYC rearrangements group.

haematologica | 2022; 107(5)

1155

A. Dodero et al.

Table 2. Kaplan-Meier estimates of 2-year progression-free and overall survival according to patients and disease characteristics

2-year PFS (95% CI) Age ≤60 >60 Sex Male Female Cell of origin** Germinal central B cell Non-Germinal central B cell Rearrangements DEL DEL-BCL2 DEL-MYC DEL-DH/TH Staging I-II III-IV International prognostic index 0-2 3-5 CNS prophylaxis None Intrathecal methotrexate Intravenous methotrexate TP53 mutation*** Wild-type Mutated

2-year OS (95% CI)

0.138 80.0 (70.4; 90.9) 67.7 (55.6; 82.4)

P* 0.064

88.4 (80.0; 97.7) 78.2 (66.9; 91.3) 0.012

66.3 (55.9; 78.7) 87.1 (77.0; 98.7)

0.094 80.7 (71.3; 91.3) 88.9 (79.0; 100)

0.544 70.6 (59.1; 84.4) 74.6 (63.2; 87.9)

0.907 81.0 (69.9; 93.9) 84.7 (75.4; 95.2)

0.203 74.8 (64.9; 86.3) 69.2 (48.2; 99.5) 100 63.2 (44.8; 89.0)

0.058 85.9 (77.6; 95.2) 90.9 (75.4; 100) 100 66.2 (47.4; 92.5)

0.048 91.8 (81.6; 100) 69.8 (60.6; 80.4)

0.081 95.0 (85.9; 100) 81.0 (72.6; 90.3)

0.002 88.1 (79.6; 97.6) 62.2 (50.7; 76.3)

0.002 97.8 (93.7; 100) 71.8 (60.5; 85.2)

0.027 50.8 (28.6; 90.2) 69.5 (56.5; 85.5) 81.8 (72.0; 93.0)

0.008 64.9 (41.7; 100) 75.8 (63.2; 91.0) 94.3 (88.2; 100)

0.033 79.9 (68.9; 92.8) 58.3 (37.3; 91.1)

0.036 87.7 (78.0; 98.5) 61.7 (39.8; 95.6)

PFS: progression-free survival; OS: overall survival; CI: confidence interval; DEL: double expressor lymphomas; DH/TH: double-hit/triple-hit; CNS: central nervous system. *Logrank test P-value; **Excluding 7 not-assessed patients; ***Excluding 53 not-assessed patients.

Results Patients’ characteristics and treatment A total of 122 patients affected by DEL were consecutively treated with DA-EPOCH-R between November 2015 and March 2020. Patients’ characteristics are summarized in Table 1. The median age was 59 years (range, 24-79 years), and 62% were male. MYC and BCL2 rearrangements by FISH were evaluated in all cases. BCL6 rearrangement could be done for 95 out of 122 patients (78%). According to IHC and FISH testing, the study population was divided into three subgroups: i) DEL only without any rearrangements (n=81, 66%); ii) DEL single-hit (DEL-SH) with either MYC or BCL2 rearrangement (n=9 MYC, n=13 BCL2); and iii) high-grade lymphomas (double-hit/triple-hit [DH/TH]) (n=10 MYC/BCL2, n=3 MYC/BCL6, n=6 MYC/BCL2/BCL6). The COO assignment according to Hans algorithm was analyzed in 115 of 122 patients (94%) of whom 55 (45%) were GCB and 60 (49%) non-GCB. Ninety-five (78%) patients had an advanced stage and 67 (55%) presented an IPI score of 3-5. Moreover, the number of patients with limited disease was low (n=27) with only four cases with stage I (all these cases presented with bulky extranodal disease) (Online Supplementary Table S1). 1156

The median number of chemotherapy cycles was 6 (range, 1-6 cycles). Due to the aggressive clinical presentation requiring urgent treatment, 19 (15%) patients received the first cycle of R-CHOP while waiting for complete FISH analyses. Response assessment following DA-EPOCH-R treatment was feasible in 117 patients: of these 84 (72%) and 16 (14%) achieved a complete or partial remission, respectively. Seventeen patients (14%) showed progressive disease. Five patients were not evaluable for a response after DA-EPOCH-R for toxicity (n=1, death of pneumonia), de-escalation therapy (n=2), consolidation with high-dose therapy before the sixth cycle (n=2). Overall, 22 of 122 (18%) patients (n=12 DE, n=3 SH-BCL2, n=1 SH-MYC, n=6 DH/TH) underwent autologous stem cell transplantation, in clinical remission, during treatment (n=2) and as consolidation after six cycles of DA-EPOCH-R (n=20) (Online Supplementary Table S2). With the exclusion of one patient who died of pneumonia, other adverse events were manageable. We observed febrile neutropenia in 16 of 122 (13%) and infections requiring hospital admission (n=6 pneumonia, n=1 sepsis) in seven patients (6%).

Survival outcome After a median follow-up of 24 months (interquartile range [IQR], 14-38 months), 110 patients were alive and 22 haematologica | 2022; 107(5)

DA-EPOCH-R in double expressor lymphoma

died (n=20 for disease progression, n=1 for toxicity, n=1 suicide). PFS (95% confidence interval [CI]) and OS at 2-year were 74% (66-83%) and 84% (77-91%), respectively (Figure 1A and B). The 2-year OS and PFS were not signifi-

cantly different between DEL, DEL-MYC, DEL-BCL2 and, DEL-DH/TH, with a trend for inferior survival in this last subgroup (OS: 66% [range, 47-92%], P=0,058) (Figure 1C and D).

Figure 1. Kaplan-Meier estimates of progression-free survival and overall survival. Progression-free survival (A, C and E) and overall survival (B, D and F) for the whole cohort (A and B) and according to rearrangements (C and D). DEL: double expressor lymphomas only; DEL-MYC, DEL-BCL2: high-grade lymphomas (doublehit/triple-hit [DH/TH]) and International Prognostic Index (panels E and F).

haematologica | 2022; 107(5)

1157

A. Dodero et al.

Table 3. Table 3. Clinical characteristics of 69 patients evaluated for TP53 mutation status.

Rearrangements None DEL-BCL2 DEL-MYC DEL-DH/TH Cell of origin** Germinal central B cell Non-Germinal central B cell Staging I-II III-IV International prognostic index 0-2 3-5 Systemic CNS therapy None Intrathecal methotrexate Intravenous methotrexate Autologous stem cell transplantation Yes No Progression-free survival 2-year estimate (95% CI) Events CNS relapse-free probability 2-year estimate (95% CI) Events

Mutated TP53 N=16

Wild-type TP53 N=53

8 (50.0) 3 (18.8) 3 (18.8) 2 (12.5)

36 (67.9) 6 (11.3) 3 (5.7) 8 (15.1)

8 (53.3) 7 (46.7)

25 (48.1) 27 (51.9)

2 (12.5) 14 (87.5)

18 (34.0) 35 (66.0)

5 (31.2) 11 (68.8)

28 (52.8) 25 (47.2)

P* 0.258

0.776

0.124

0.161

0.241 0 (0.0) 7 (43.8) 9 (56.2)

5 (9.4) 14 (26.4) 34 (64.2) 0.334

1 (6.2) 15 (93.8)

11 (20.8) 42 (79.2)

58.3 (37.3; 91.1) 6 (37.5)

79.9 (68.9; 92.8) 9 (17.0)

90.0 (73.2; 100) 1 (6.3)

92.5 (84.6; 100) 3 (5.7)

0.033***

0.782***

DH/TH: double/triple hit; CNS: central nervous system. *Fisher Exact test P-value; **Excluding 7 not-assessed patients; ***Log-rank test P-value.

Age above 60 years did not affect outcome whereas the male sex was associated with a significantly shorter PFS (Online Supplementary Table S3). The COO did not show a significant impact either on PFS or OS. Isolated MYC (≥70%) or BCL2 (≥80%) as assessed by IHC did not impact PFS and OS (data not shown). As expected, patients with a limited disease had a significantly higher 2-year PFS 92% (range, 81-100%) as compared to advanced stages (70% [range, 61-80%], P=0,048) (Table 2). The analysis of outcome by IPI score showed a 2-year PFS of 62% (range, 51-76%) and OS of 71% (range, 61-85%) for high-intermediate and high IPI score that was inferior compared to low-intermediate and low cases (88%, [range, 79-98%] and 98% [range, 94-100%]; P=0,002 and P=0,002 respectively) (Figure 1E and F). In those achieving a response, we did observe a significant difference in outcome between patients who did or did not receive autologous transplantation (Online Supplementary Figure S1). Complete results of univariable Cox models for PFS and OS according to the patients and disease characteristics are reported in the Online Supplementary Table S4.

Evaluation of TP53 mutation The TP53 mutation could be retrospectively evaluated in 69 of 122 (57%) patients due to the absence of sufficient residual archival material or to poor quality of the genomic 1158

DNA extracted from paraffin-embedded tissues. The TP53 mutation status was assessed in 44 DEL (64%), six DELMYC (9%), nine DEL-BCL2 (13%), and ten DEL-DH/TH (15%). Overall a pathogenic TP53 mutation (as defined by the IARC TP53 database) was present in 16 patients (23%). We evaluated the outcome according to the presence or absence of TP53 mutation. The two groups were not statistically different for the main clinical characteristics (Table 3). The 2-year PFS was 58% (range, 37-91%) and 80% (range, 70-93%; P=0.033) and the 2-year OS was 62% (range, 4096%) and 88% (range, 78-99%; P=0.036), for mutated and wild-type cases respectively (Figure 2A and B).

Multivariable analysis Cox multivariate models concerning the PFS and OS of the patients were performed as summarized in Table 4. TP53 mutation, IPI 3-5, and absence of CNS prophylaxis had a negative prognostic impact on OS whereas the female sex was associated with a significantly improved PFS.

Outcome of relapsed patients Among 28 patients who relapsed (n=17 DEL, n=4 DELSH [only with BCL2 translocation], n= 7 DEL-DH/TH), 19 (68%) died of lymphoma whereas nine patients are still alive (n=6 DEL, n=3 DEL/BCL2). Five of nine patients are in complete remission after receiving different salvage haematologica | 2022; 107(5)

DA-EPOCH-R in double expressor lymphoma Table 4. Results of the multivariable Cox models for progression-free and overall survival.

Model

Rearrangements DEL-BCL2 vs. DEL DEL-MYC vs. DEL** DEL-DH/TH vs. DEL TP53 mutation Mutated vs. wild-type Not performed vs. wild-type International prognostic index 0-2 vs. 3-5 Systemic CNS therapy None vs. intravenous MTX Intrathecal MTX vs. intravenous MTX Staging III-IV vs. I-II Sex Female vs. male Age*** 65 vs. 49

Progression-free survival Hazard ratio P* (95% CI)

Overall survival Hazard ratio (95% CI)

0.880 1.49 (0.48; 4.64) 1.15 (0.46; 2.87)

P* 0.408

0.16 (0.02; 1.48) 1.00 (0.37; 2.75) 0.072

3.13 (1.04; 9.40) 0.98 (0.41; 2.35)

0.002 8.90 (2.14; 36.99) 0.75 (0.22; 2.53)

0.063 0.36 (0.12; 1.06)

0.018 0.18 (0.04; 7.4)

0.062 3.74 (1.22; 11.41) 1.99 (0.83; 4.76)

0.019 8.49 (1.82; 39.57) 4.25 (1.2; 15.02)

0.847 1.19 (0.21; 6.69)

0.045 0.36 (0.14; 0.98)

0.752

1.52 (0.51; 4.56)

CI: confidence interval; DEL: double expressor lymphomas; DH/TH: double-hit/triple-hit; CNS: central nervous system; MTX, methotrexate. *Wald test P-value; **No survival or progression events observed; ***Modeled as restricted cubic spline and reporting result of 65 vs. 49 years comparison.

therapies (n=2 auto-stem cell transplantation [auto-SCT], n=2 allo-SCT, n=1 lenalidomide in combination with radiotherapy).

Central nervous system prophylaxis and central nervous system relapse At diagnosis, only two patients had cerebrospinal fluid involvement: one died early of systemic progressive disease and the other is still alive after therapy including high-dose methotrexate. The CNS prophylaxis was chosen at the discretion of the treating physician. Sixty-six (54%) patients received systemic HD-MTX, 40 (33%) underwent intrathecal chemotherapy with methotrexate and cytarabine and 16 (13%) did not receive any CNS prophylaxis at all. In particular, patients not receiving CNS prophylaxis had less extranodal involvement at risk for CNS relapse. All characteristics are detailed in the Online Supplementary Table S5. Systemic methotrexate-based CNS prophylaxis conferred a better 2-year OS (94%, [range, 88-100%]) as compared to intrathecal or no CNS prophylaxis (75%, [range, 63-91%] and 65%, [range, 42-100%] respectively; P=0.008) (Figure 2C and D). A significant advantage in OS was observed even after exclusion of DH/TH patients (2-year percentage OS 96% vs. 81%, vs. 63%, respectively, [P<0.001; Figure 2E and F]) that was the subgroup with the worst outcome. Overall, we observed five CNS relapses, and the cumulative incidence of relapse at 1- and 2-year was 2% (range, 19%) and 5% (range, 2-13%) respectively in the entire cohort. All patients with CNS relapse were DEL only, all but one were non-GCB. Four of five patients died of CNS lymphomas. The CNS relapse occurred even in patients who received CNS prophylaxis (3 of 5 patients) and in four out five with low CNS-IPI. haematologica | 2022; 107(5)

Discussion In the present retrospective study, we collected a large number of consecutive DEL patients (n=122) who were treated with the DA-EPOCH-R regimen to test the hypothesis that an intensive regimen could overcome poor clinical, and biological prognostic factors. To the best of our knowledge, this study includes the largest series of DEL patients exposed to an intensified regimen. All DEL patients were analyzed for MYC and/or BCL2 rearrangements and, partly, for TP53 mutational status. Indeed, the 2-year PFS and OS of 74% and 84%, respectively, for the entire cohort seem promising. Further, in patients characterized by IPI score of 3-5, the results are comparable to those achieved with other intensified treatments (R-CODOX/IVAC and R-ACVB).23,24 Currently, the treatment of DEL without any gene rearrangement remains an unmet clinical need. The phase III trial performed by the Alliance group comparing the RCHOP and DA-EPOCH-R regimens in newly diagnosed DLBCL included only a limited number of DEL, thus preventing the possibility of drawing definitive conclusions on the role of DA-EPOCH-R in this subtype.25 Our series of 81 DEL patients without any rearrangement showed a 2-year PFS and OS of 75% and 86%, respectively, suggesting a potential role of the intensive regimen. Recently, Morschhauser et al. evaluated the combination of venetoclax with standard R-CHOP or obinutuzumab-CHOP in DEL patients or those with high expression of BCL2, showing a similar 2-year PFS of 72% and 77%, respectively. Patients with a DEL-DH/TH status showed mainly an intermediate and high-risk IPI (75%) and had a trend for inferior OS at 2 years (66%) as compared to DEL only (86%), whereas the observed 2-year PFS of 63% was not 1159

A. Dodero et al.

statistically different among these three subgroups. In addition, we have to consider that the prognosis of highgrade B-cell lymphomas (HGBCL) with overexpression of MYC or BCL2 proteins seems poor as compared to DH/TH, not DEL.27 This finding suggests that more than 50% of DEL-DH/TH could be cured with an intensive reg-

imen but that in case of relapse they have a trend for poor overall survival. Likely, the recent Food and Drug Administration and European Medicines Agency approval of new therapies such as CAR-T cells will have an impact on the OS of relapsed DEL-DH/TH HGBCL.28,29 A total of 27 patients with limited-stage disease (stage I,

Figure 2. Kaplan-Meier estimates of progression-free survival and overall survival according to TP53 mutation and central nervous prophylaxis. Progression-free survival (A, C and E) and overall survival (B, D and F) according to TP53 mutational status (A and B) and central nervous system prophylaxis (none, intrathecal, intravenous methotrexate) including (C and-D) and excluding (E and F) high-grade B-cell lymphomas (double-hit/triple-hit [DH/TH]).

1160

haematologica | 2022; 107(5)

DA-EPOCH-R in double expressor lymphoma

n=4 with extensive extranodal localization; stage II, n=23 (DEL, n=18; SH, n=6; DH/TH, n=3) were treated with DA-EPOCH-R and experienced an impressive 2-year PFS and OS of 92% and 95%, respectively. Interestingly, the majority of limited disease patients did not have TP53 mutation (mutated, n=2; wild-type, n=18; not tested, n=7). Only a handful of studies, mainly of retrospective nature, have been focused on the prognosis of single hit or DH/TH lymphomas with limited-stage disease. Torka and colleagues analyzed the outcome of 81 patients carrying MYC rearrangement, including 40 DH, who received the standard R-CHOP regimen or intensive chemotherapy.30 The authors did not find any statistical difference in PFS and OS across treatment strategies but in the subgroup of DH, the intensive therapy was associated with a higher CR rate. Over the past years, the combination of CNS-IPI and biological factors has been considered the best way to estimate the CNS relapse risk. In a retrospective study of newly diagnosed DLBCL patients treated with R-CHOP, without any CNS systemic prophylaxis, Savage and colleagues reported that the DEL/non-GCB subgroup had a significantly higher risk of CNS relapse as compared to the non-DEL/non-GCB subgroup (15% vs. 3%).31 In contrast, in the prospective Goya trial, COO and not MYC/BCL2 double expression impacted the risk of CNS relapse.32 More recently, the Nordic group has suggested the efficacy of the early administration of HD-MTX in newly diagnosed DLBCL at high risk of CNS relapse.33 Our study is the first in which the risk of CNS relapse has been evaluated following the administration of a DA-EPOCH-R regimen and HD-MTX in 66 of 122 (54%) patients. Our population included 24% of patients with high-risk CNS-IPI. Despite the high-risk population analyzed, systemic CNS prophylaxis was associated with a very low cumulative incidence of CNS relapse (5%) and a significant advantage in PFS and OS. The advantage on the outcome can be influenced by other untested factors and should be confirmed in a larger trial. Xu-Monette et al.6 reported a 21% incidence of TP53 mutations in a large population of DLBCL patients treated with R-CHOP and demonstrated that TP53 disruption was associated to poor PFS and OS in both GCB and non-GCB subtypes. Following a better characterization of DLBCL beyond the COO, the frequency of this mutation was tested in DEL and did not result to be different from non DEL (25% vs. 22%, respectively). The role of genomic alterations, including TP53 mutations, has recently been investigated in several studies by applying novel molecular tech-

References 1. Coiffier B, Lepage E, Briere J, et al. CHOP chemotherapy plus rituximab compared with CHOP alone in elderly patients with diffuse large-B-cell lymphoma. N Engl J Med. 2002;346(4):235-242. 2. Sehn LH, Donaldson J, Chhanabhai M, et al. Introduction of combined CHOP plus rituximab therapy dramatically improved outcome of diffuse large B-cell lymphoma in British Columbia. J Clin Oncol. 2005; 23(22):5027-5033. 3. Project IN-HsLPF. A predictive model for aggressive non-Hodgkin's lymphoma. N Engl J Med. 1993;329(14):987-994. 4. Alizadeh AA, Eisen MB, Davis RE, et al.

haematologica | 2022; 107(5)

niques.9 Interestingly, Meriranta et al. found a frequent association between TP53 alterations and MYC overexpression or translocations and, in those with TP53 mutation, a worse outcome in DEL as compared to non-DEL was reported.34 Recently, Song et al. reported a poor prognosis in DLBCL carrying double-hit signature and TP53 inactivation.35 In our study, the prevalence of the TP53 mutation was in keeping with previous studies, and the PFS/OS following DA-EPOCH-R treatment in TP53-mutated patients remained significantly lower than that observed in non-mutated patients suggesting the failure of intensive therapy to overcome the mutation adverse effect. The observed 2-year OS of 62% seems better when compared to the OS of 48% reported by Xu Monette6 with R-CHOP and of 27% described by Chiappella et al.36 following intensified therapy. These are indirect comparisons requiring a clinical study to be validated. The intensified treatment with DA-EPOCH-R was well tolerated considering that 56% of patients were older than 60 years. We observed a limited incidence of severe toxicities, with one death for pneumonia and two patients deescalated to R-CHOP for repeated infections. Other observed adverse events were manageable and did not compromise the completion of the therapeutic program. Our data are promising, but we have to consider some limitations including: i) the retrospective nature; ii) the absence of information about MYC translocation partners (IG vs. non-IG); iii) the determination of cell of origin performed according to the Hans algorithm and not by the nanostring technology; iv) the absence of a control series of non-DEL patients with a single rearrangement or with DH/TH genotype. In conclusion, we show a good outcome for DAEPOCH-R in combination with HD-MTX in DEL and DEL-SH lymphomas without TP53 mutations, but the lower survival of patients DEL-DH/TH subtype or DEL with TP53 mutations requires further clinical studies aimed at testing novel agents combined with chemotherapy. Disclosures No conflicts of interests to disclose. Contributions AD, AG designed research, analyzed data, and wrote the manuscript; FM, AT, AR, MB, AC, MP, LD, FR, FC and LF collected data; FB and RM performed statistical analyses; CC, CM and DR performed TP53 mutations; AC, DR, VM and FF performed histological diagnosis and FISH analysis; PC and CC-S supervised the study.

Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature. 2000;403(6769):503-511. 5. Hans CP, Weisenburger DD, Greiner TC, et al. Confirmation of the molecular classification of diffuse large B-cell lymphoma by immunohistochemistry using a tissue microarray. Blood. 2004;103(1):275-282. 6. Xu-Monette ZY, Medeiros LJ, Li Y, et al. Dysfunction of the TP53 tumor suppressor gene in lymphoid malignancies. Blood. 2012;119(16):3668-3683. 7. Xu-Monette ZY, Wu L, Visco C, et al. Mutational profile and prognostic significance of TP53 in diffuse large B-cell lymphoma patients treated with R-CHOP: report from an International DLBCL

Rituximab-CHOP Consortium Program Study. Blood. 2012;120(19):3986-3996. 8. Young KH, Leroy K, Møller MB, et al. Structural profiles of TP53 gene mutations predict clinical outcome in diffuse large Bcell lymphoma: an international collaborative study. Blood. 2008;112(8):3088-3098. 9. Chapuy B, Stewart C, Dunford AJ, et al. Molecular subtypes of diffuse large B cell lymphoma are associated with distinct pathogenic mechanisms and outcomes. Nat Med. 2018;24(5):679-690. 10. Johnson NA, Savage KJ, Ludkovski O, et al. Lymphomas with concurrent BCL2 and MYC translocations: the critical factors associated with survival. Blood. 2009;114(11):2273-2279.

1161

A. Dodero et al. 11. Le Gouill S, Talmant P, Touzeau C, et al. The clinical presentation and prognosis of diffuse large B-cell lymphoma with t(14;18) and 8q24/c-MYC rearrangement. Haematologica. 2007;92(10):1335-1342. 12. Niitsu N, Okamoto M, Miura I, Hirano M. Clinical features and prognosis of de novo diffuse large B-cell lymphoma with t(14;18) and 8q24/c-MYC translocations. Leukemia. 2009;23(4):777-783. 13. Barrans SL, Evans PA, O'Connor SJ, et al. The t(14;18) is associated with germinal center-derived diffuse large B-cell lymphoma and is a strong predictor of outcome. Clin Cancer Res. 2003;9(6):2133-2139. 14. Visco C, Tzankov A, Xu-Monette ZY, et al. Patients with diffuse large B-cell lymphoma of germinal center origin with BCL2 translocations have poor outcome, irrespective of MYC status: a report from an International DLBCL rituximab-CHOP Consortium Program Study. Haematologica. 2013;98(2):255-263.ì 15. Ennishi D, Mottok A, Ben-Neriah S, et al. Genetic profiling of Myc and BCL2 in diffuse large B-cell lymphomas determines cell of origin specific clinical impact. Blood. 2017;129(20):2760-2770. 16. Hu S, Xu-Monette ZY, Tzankov A, et al. MYC/BCL2 protein coexpression contributes to the inferior survival of activated B-cell subtype of diffuse large B-cell lymphoma and demonstrates high-risk gene expression signatures: a report from The International DLBCL Rituximab-CHOP Consortium Program. Blood. 2013; 121(20):4021-4031. 17. Perry AM, Alvarado-Bernal Y, Laurini JA, et al. MYC and BCL2 protein expression predicts survival in patients with diffuse large B-cell lymphoma treated with rituximab. Br J Haematol. 2014;165(3):382-391. 18. Savage KJ, Slack GW, Mottok A, et al. Impact of dual expression of MYC and BCL2 by immunohistochemistry on the risk of CNS relapse in DLBCL. Blood. 2016;127(18):2182-2188. 19. Dodero A, Guidetti A, Tucci A, et al. Doseadjusted EPOCH plus rituximab improves

1162

the clinical outcome of young patients affected by double expressor diffuse large B-cell lymphoma. Leukemia. 2019; 33(4):1047-1051. 20. Hainaut P, Pfeifer GP. Somatic TP53 Mutations in the Era of Genome Sequencing. Cold Spring Harb Perspect Med. 2016;6(11):a026179. 21. Pospisilova S, Gonzalez D, Malcikova J, et al. ERIC recommendations on TP53 mutation analysis in chronic lymphocytic leukemia. Leukemia. 2012;26(7):1458-1461. 22. Wilson WH, Grossbard ML, Pittaluga S, et al. Dose-adjusted EPOCH chemotherapy for untreated large B-cell lymphomas: a pharmacodynamic approach with high efficacy. Blood. 2002;99(8):2685-2693. 23. McMillan AK, Phillips EH, Kirkwood AA, et al. Favourable outcomes for high-risk diffuse large B-cell lymphoma (IPI 3-5) treated with front-line R-CODOX-M/R-IVAC chemotherapy: results of a phase 2 UK NCRI trial. Ann Oncol. 2020;31(9):1251-1259. 24. Molina TJ, Canioni D, Copie-Bergman C, et al. Young patients with non-germinal center B-cell-like diffuse large B-cell lymphoma benefit from intensified chemotherapy with ACVBP plus rituximab compared with CHOP plus rituximab: analysis of data from the Groupe d'Etudes des Lymphomes de l'Adulte/lymphoma study association phase III trial LNH 03-2B. J Clin Oncol. 2014;32(35):3996-4003. 25. Bartlett NL, Wilson WH, Jung SH, et al. Dose-adjusted EPOCH-R compared with R-CHOP as frontline therapy for diffuse large B-cell lymphoma: clinical outcomes of the Phase III Intergroup Trial Alliance/CALGB 50303. J Clin Oncol. 2019;37(21):1790-1799. 26. Morschhauser F, Feugier P, Flinn IW, et al. A phase 2 study of venetoclax plus RCHOP as first-line treatment for patients with diffuse large B-cell lymphoma. Blood. 2021;137(5):600-609. 27. Sesques P, Johnson NA. Approach to the diagnosis and treatment of high-grade Bcell lymphomas with MYC and BCL2 and/or BCL6 rearrangements. Blood. 2017;

129(3):280-288. 28. Schuster SJ, Svoboda J, Chong EA, et al. Chimeric antigen receptor T cells in refractory B-cell lymphomas. N Engl J Med. 2017;377(26):2545-2554. 29. Locke FL, Ghobadi A, Jacobson CA, et al. Long-term safety and activity of axicabtagene ciloleucel in refractory large B-cell lymphoma (ZUMA-1): a single-arm, multicentre, phase 1-2 trial. Lancet Oncol. 2019;20(1):31-42. 30. Torka P, Kothari SK, Sundaram S, et al. Outcomes of patients with limited-stage aggressive large B-cell lymphoma with high-risk cytogenetics. Blood Adv. 2020;4(2):253-262. 31. Savage KJ, Slack GW, Mottok A, et al. Impact of dual expressionof MYC and BCL2 by immunohistochemistryon the risk of CNS relapse in DLBCL. Blood. 2016;127(18):2182-2188. 32. Klanova M, Sehn LH, Bence-Bruckler I, et al. Integration of cell of origin into the clinical CNS International Prognostic Index improves CNS relapse prediction in DLBCL. Blood. 2019;133(9):919-926. 33. Leppä S, Jørgensen J, Tierens A, et al. Patients with high-risk DLBCL benefit from dose-dense immunochemotherapy combined with early systemic CNS prophylaxis. Blood Adv. 2020;4(9):1906-1915. 34. Meriranta L, Pasanen A, Alkodsi A, Haukka J, Karjalainen-Lindsberg ML, Leppä S. Molecular background delineates outcome of double protein expressor diffuse large Bcell lymphoma. Blood Adv. 2020; 4(15):3742-3753. 35. Song JY, Perry AM, Herrera AF, et al. Double-hit signature with TP53 abnormalities predicts poor survival in patients with germinal center type diffuse large B-cell lymphoma treated with R-CHOP. Clin Cancer Res. 2021;27(6):1671-1680. 36. Chiappella A, Diop F, Agostinelli C, et al. TP53 mutation had a negative prognostic impact in untreated young patients with diffuse large B-cell lymphoma at high-risk: a sub-analysis of FIL-DLCL04 study. Hemasphere. 2018;2(S1):711-712.

haematologica | 2022; 107(5)

ARTICLE

Plasma Cell Disorders

Natural history of Waldenström macroglobulinemia following acquired resistance to ibrutinib monotherapy

Ferrata Storti Foundation

Joshua N. Gustine,1,2 Shayna Sarosiek,1,3 Catherine A. Flynn,1 Kirsten Meid,1 Carly Leventoff,1 Timothy White,1 Maria Luisa Guerrera,1 Lian Xu,1 Amanda Kofides,1 Nicholas Tsakmaklis,1 Manit Munshi,1 Maria Demos,1 Christopher J. Patterson,1 Xia Liu,1 Guang Yang,1,3 Zachary R. Hunter,1,3 Andrew R. Branagan,3,4 Steven P. Treon1,3 and Jorge J. Castillo1,3 1

Bing Center for Waldenström’s Macroglobulinemia, Dana-Farber Cancer Institute; Boston University School of Medicine; 3Department of Medicine, Harvard Medical School and 4Division of Medical Oncology, Massachusetts General Hospital, Boston, MA, USA

Haematologica 2022 Volume 107(5):1163-1171

ABSTRACT

brutinib is highly active and produces long-term responses in patients with Waldenström macroglobulinemia (WM), but acquired resistance can occur with prolonged treatment. We therefore evaluated the natural history and treatment outcomes in 51 WM patients with acquired resistance to ibrutinib monotherapy. The median time between ibrutinib initiation and discontinuation was 2 years (range, 0.4-6.5 years). Following discontinuation of ibrutinib, a rapid increase in serum immunoglobulin M level was observed in 60% (29/48) of evaluable patients, of whom ten acutely developed symptomatic hyperviscosity. Forty-eight patients (94%) received salvage therapy after ibrutinib. The median time to salvage therapy after ibrutinib cessation was 18 days (95% confidence interval [CI]: 13-27). The overall and major response rates to salvage therapy were 56% and 44%, respectively, and the median duration of response was 48 months (95% CI: 34-not reached). Quadruple-class (rituximab, alkylator, proteasome inhibitor, ibrutinib) exposed disease (odds ratio [OR] 0.20, 95% CI: 0.05-0.73) and salvage therapy ≤7 days after discontinuing ibrutinib (OR 4.12, 95% CI: 1.0718.9) were identified as independent predictors of a response to salvage therapy. The 5-year overall survival (OS) following discontinuation of ibrutinib was 44% (95% CI: 26-75). Response to salvage therapy was associated with better OS after ibrutinib (hazard ratio 0.08, 95% CI: 0.02-0.38). TP53 mutations were associated with shorter OS, while acquired BTK C481S mutations had no impact. Our findings reveal that continuation of ibrutinib until subsequent treatment is associated with improved disease control and clinical outcomes.

Introduction Waldenström macroglobulinemia (WM) is an immunglobulin M (IgM)-secreting lymphoplasmacytic lymphoma.1 Whole-genome sequencing has identified highly recurrent somatic mutations in MYD88 (95-97%) and CXCR4 (30-40%) in WM patients.2,3 Mutated MYD88 triggers NF-kB pro-survival signaling via Bruton’s tyrosine kinase (BTK) and interleukin-1 receptor-associated kinase 1 (IRAK1)/IRAK4, and transactivates hematopoietic cell kinase (HCK).4,5 Both BTK and HCK are targeted by ibrutinib.4,5 Mutations in the C-terminal domain of CXCR4 are typically subclonal and support intrinsic ibrutinib resistance through upregulation of AKT and ERK1/2 signaling.6-9 In 2015, ibrutinib became the first approved agent by the United States Food and Drug Administration and European Medicines Agency for the treatment of symptomatic WM patients. The regulatory approval of ibrutinib was based on the results from a multi-center, single-arm, phase II trial of 63 previously treated WM patients.10 Ibrutinib monotherapy was highly active with an overall response rate

haematologica | 2022; 107(5)

Correspondence: JORGE J. CASTILLO jorgej_castillo@dfci.harvard.edu Received: April 28, 2021. Accepted: June 16, 2021. Pre-published: June 24, 2021. https://doi.org/10.3324/haematol.2021.279112

1163

J.N. Gustine et al.

(ORR) of 91%, major response rate (MRR) of 79%, and very good partial response rate (VGPR) of 30% with prolonged follow-up.10,11 Responses to ibrutinib were durable with an estimated 5-year progression-free survival (PFS) and overall survival (OS) of 54% and 87%, respectively. A notable finding was the impact of MYD88 and CXCR4 mutations on ibrutinib outcomes. Patients wild-type (WT) for both MYD88 and CXCR4 had no major responses and a median PFS of 5 months to ibrutinib.10-12 Among patients with mutated MYD88, the concurrent presence of a CXCR4 mutation adversely impacted response rates, response kinetics, and 5-year PFS (38% vs. 70%).10,11 Similar outcomes to ibrutinib monotherapy have been reported in phase II trials of treatment-naïve (n=30) and rituximab-refractory WM patients (n=31), as well as in the recent phase III ASPEN trial (n=199) comparing ibrutinib to zanubrutinib.13-16 Despite the high response rates and durable remissions, acquired ibrutinib resistance is increasingly being observed in WM patients. Approximately half of WM patients who progress on ibrutinib acquire BTK mutations at the binding site of ibrutinib (BTK C481S) or its downstream mediator PLCg2.17 BTK C481S mutations are highly subclonal and confer protection to BTK WT clones via a paracrine mechanism.17,18 Acquired deletions in 6q and 8p that contain regulators of BTK, MYD88/NF-kB, and apoptotic signaling also occur.19 However, data on the clinical outcomes of WM patients who progress while on active ibrutinib therapy are limited. Preliminary studies have described an abrupt increase in serum IgM level (i.e., IgM rebound) in some WM patients who discontinue ibrutinib.20,21 We sought to further characterize the clinical presentation, management, and outcomes of WM patients with acquired ibrutinib resistance, as well as the impact of BTK C481S mutations.

lymphoma (DLBCL).23,24 The ORR and MRR were assessed for each regimen used after T0. Duration of response (DOR) was defined as the length of time between response attainment and progression, death, or last follow-up. Survival after disease progression on ibrutinib was defined as the length of time between T0 and the date of death or last follow-up.

Tumor genotyping The presence of MYD88, CXCR4, and BTK mutations was detected by allele-specific polymerase chain reaction (AS-PCR) and Sanger sequencing methods, as previously described.6,17,25 A clinically validated next-generation sequencing (NGS) assay was also performed in a subset of patients on unselected bone marrow (BM) aspirate samples to identify TP53 mutations.26

Statistical analyses Patient characteristics were summarized using descriptive statistics. Continuous variables were dichotomized using standard WM cutoffs to facilitate analysis, and comparisons were made using the c2 test or Fischer exact test depending on the number of observations. Univariate and multivariate logistic regression models were utilized to identify predictive factors for an IgM rebound and response to salvage therapy; the outcome measure was odds ratio (OR) with 95% confidence interval (CI). Time to events was estimated using the Kaplan-Meier method, and comparisons between groups were made using the log-rank test. The Cox-proportional hazard regression method was used to fit univariate and multivariate models for OS; the outcome measure was hazard ratio (HR) with 95% CI. P-values were two-sided and considered statistically significant if <0.05. All calculations and graphs were obtained using R (R Foundation for Statistical Computing, Vienna, Austria).

Results Patient characteristics

Methods Study design and patient selection We reviewed a prospectively maintained database of 362 patients seen at our institution between January 2012 and October 2020 who met clinicopathological criteria for WM and received ibrutinib monotherapy.1 Patients who had disease progression on active ibrutinib therapy per consensus guidelines were identified and included in this study.22 A transient increase in serum IgM level associated with a temporary hold of ibrutinib was not considered disease progression. The date a patient discontinued ibrutinib because of disease progression was defined as time-zero (T0). Pertinent clinical and pathological data were gathered for all patients at the time of T0 until the last follow-up or death. The Dana-Farber/Harvard Cancer Center Institutional Review Board approved this study, and all patients provided written consent.

Response and outcome definitions We defined an IgM rebound as a ≥25% increase in serum IgM level following T0, with an absolute increase of at least 500 mg/dL, consistent with previous studies.20,21 Response assessment to salvage therapy was performed according to consensus guidelines from the 6th International Workshop on WM.22 The ORR was defined as a minor response or better (≥25% reduction in serum IgM level), and the MRR was defined as a partial response or better (≥50% reduction in serum IgM level). Consensus guidelines were also utilized to assess response to salvage therapy for patients with light chain (AL) amyloidosis and diffuse large B cell

1164

We identified 51 WM patients with acquired resistance to ibrutinib monotherapy whose findings are included in this study. The baseline clinical characteristics at T0 are summarized in Table 1. The median duration between WM diagnosis and study entry (T0) was 8.2 years (range, 0.5-24 years). The median treatment duration with ibrutinib before T0 was 2.0 years (range, 0.4-6.5 years). The median time between disease progression on ibrutinib and T0 was 25 days (range, 0-426 days); seven patients (14%) deriving clinical benefit continued on ibrutinib for >90 days after meeting criteria for disease progression before discontinuing therapy. Forty-three patients (84%) had received ibrutinib in the relapsed or refractory setting, and eight (16%) in the frontline setting. The median number of treatment lines including ibrutinib before T0 was four (range, 1-9). Twenty patients (39%) were previously exposed to the major drug classes during their disease course, including rituximab, proteasome inhibitors, alkylators, and ibrutinib (i.e., “quadruple-class exposed”). MYD88 and CXCR4 mutations were present in 93% and 58% of genotyped patients, respectively, and the majority (87%) of CXCR4 mutations were nonsense variants. The clinical manifestations at the time of disease progression on ibrutinib showed considerable heterogeneity and are presented in Table 2.

Serum immunoglobulin M rebound The peak absolute change in serum IgM level following haematologica | 2022; 107(5)

Acquired ibrutinib resistance in WM

T0 for each patient is shown in Figure 1. An IgM rebound occurred in 29 of 48 (60%) evaluable patients following T0. Three patients who developed symptomatic hyperviscosity while progressing on ibrutinib received plasmapheresis immediately before and after T0 and were deemed non-evaluable for an IgM rebound. The median time to an IgM rebound was 27 days (95% CI: 24-33; Table 1. Baseline characteristics at time of ibrutinib discontinuation (T0).

Patient characteristic

All patients (n=51)

Median age (range) – yrs WM diagnosis 59 (40-91) Ibrutinib initiation 66 (43-93) Ibrutinib discontinuation 69 (43-93) Time from WM diagnosis Median (range) – yrs 8.2 (0.5-24) >10 yrs – no. (%) 20 (39) Time from ibrutinib initiation Median (range) – yrs 2 (0.4-6.5) >2 yrs – no. (%) 25 (49) Time from disease progression on ibrutinib Median (range) – days 25 (0-426) >90 days – no. (%) 7 (14) Male sex – no. (%) 33 (65) Hemoglobin level Median (range) – g/dL 10.3 (7.3-16.6) <10 mg/dL – no. (%) 20 (39) Platelet count Median (range) – K/uL 171 (7-463) <100 K/uL – no. (%) 16 (31) Serum IgM level Median (range) – mg/dL 1,567 (97-7,935) >4000 mg/dL – no. (%) 9 (18) Bone marrow involvement Median (range) - % 70 (5-90) >50% – no. (%) 16 (70) Previous therapy (including ibrutinib) Median no. of treatment lines (range) 4 (1-9) Treatment lines – no. (%) 1-2 21 (42) 3-4 15 (29) 5+ 15 (29) Types of therapy – no. (%) IB 8 (16) IB + R 5 (10) IB + R + PI 10 (20) IB + R + alkylator 8 (16) IB + R + PI + alkylator 20 (39) MYD88 mutation – no. (%) 43 (93) CXCR4 mutation – no. (%) 23 (58) Nonsense 20 (87%) Frameshift 3 (13%) Data on bone marrow involvement at the time of ibrutinib relapse was available for 23 patients. MYD88 and CXCR4 mutation status was available for 46 and 40 patients, respectively. IB: ibrutinib; R: rituximab; PI: proteasome inhibitor. WM: Waldenström macroglobulinemia.

haematologica | 2022; 107(5)

Figure 2A). The cumulative incidence of an IgM rebound following T0 increased over time: 7 days (9%); 14 days (13%); 21 days (25%); 28 days (46%); and 35 days (65%). Patients with an IgM rebound had a peak median absolute and relative increase in serum IgM level of 1,405 mg/dL (range, 571-7,820 mg/dL) and 79% (range, 27-1,663%), respectively. The degree of BM involvement at T0 significantly correlated with both the absolute (r=0.44; P=0.047) and relative (r=0.45; P=0.04) changes in serum IgM level. Twenty-one patients (72%) had an increase in serum IgM level back to the pre-ibrutinib baseline or higher. Symptomatic hyperviscosity acutely developed after T0 in ten of 29 patients (34%) with an IgM rebound that prompted emergent plasmapheresis. The median time from T0 to the onset of symptomatic hyperviscosity was 29 days (range, 14-51 days). Serial IgM measurements were available for seven of ten (70%) patients that developed symptomatic hyperviscosity and are shown in Online Supplementary Figure S1. Seven patients (24%) had an IgM rebound present during the first cycle of salvage therapy; none of these patients were receiving rituximab concurrently. The timing of salvage therapy following T0 impacted the risk of an IgM rebound. Patients who received salvage therapy ≤7 versus >7 days from T0 had significantly lower odds of an IgM rebound (29% vs. 76%; OR 0.15, 95% CI: 0.03-0.67; P=0.005). Bridging ibrutinib with salvage therapy was also associated with significantly lower odds of an IgM rebound compared to no bridging (17% vs. 69%; OR 0.10, 95% CI: 0.01-0.97; P=0.03). There was a trend for lower odds of an IgM rebound when bridging ibrutinib versus starting salvage therapy within 7 days of T0 (17% vs. 43%; OR 0.11, 95% CI: 0.01-1.19; P=0.11). We were unable to identify any factor at T0 predictive of an IgM rebound. Age, time on ibrutinib, time from WM diagnosis, sex, hemoglobin level, platelet count, serum IgM level, number and type of previous therapies, and MYD88 and CXCR4 mutation status were not associated with higher or lower odds of an IgM rebound (P>0.05 for all comparisons; Online Supplementary Table S1).

Salvage therapy Forty-eight patients (94%) received salvage therapy following T0. The median time to salvage therapy was 18 Table 2. Clinical manifestations of disease progression on ibrutinib.

Feature Progressive cytopenias Lymphadenopathy and/or splenomegaly Malignant pleural effusion Cardiac AL amyloidosis Isolated serum IgM increase Symptomatic hyperviscosity Soft tissue mass Bing-Neel syndrome DLBCL transformation Malignant pericardial effusion Renal monoclonal IgM deposition

All patients (N=51) 22 (43) 12 (24) 6 (12) 5 (10) 4 (8) 4 (8) 4 (8) 3 (6) 2 (4) 1 (2) 1 (2)

Patients may have had more than one manifestation of disease progression on ibrutinib. Soft tissue masses developed in the palate, orbit, maxilla, and thoracic spine with cord displacement (n=1 for each). None of the patients with histological transformation were previously treated with a nucleoside analogue. AL: light chain amyloidosis; IgM: immunoglobulin M; DLBCL: diffuse large B-cell lymphoma.

1165

J.N. Gustine et al.

days (95% CI: 13-27); treatment was started within 4 and 8 weeks of T0 for 69% and 93% of patients, respectively (Figure 2B). Reasons for not receiving salvage therapy included patient choice of hospice care (n=2) and decompensated heart failure from cardiac AL amyloidosis (n=1). The ORR and MRR to the first salvage regimen following T0 were 56% (27/48) and 44% (21/48), respectively. Among patients who responded to salvage therapy, the median DOR was 48 months (34 months-NR), and the 3-year DOR was 61% (41-90%). Twenty patients were refractory (42%) to the first salvage regimen; 11 patients received subsequent treatment, and nine patients died from progressive disease before receiving additional treatment. The specific treatment regimens utilized for the first salvage regimen after T0 with the corresponding response rates and DOR are detailed in Table 3. We then performed additional analyses to identify factors at T0 predictive of a response to the first salvage regimen. Patients with quadruple-class (rituximab, proteasome inhibitor, alkylator, ibrutinib) exposed disease had significantly lower odds of a response to the first salvage regimen compared to those without (33% vs. 73%; OR 0.18, 95% CI: 0.04-0.76; P=0.01). Age, time on ibrutinib, time from WM diagnosis, sex, hemoglobin level, platelet count, serum IgM level, number of previous therapies, and MYD88 and CXCR4 mutation status were not associated with higher or lower odds of a response to the first salvage regimen following T0 (P>0.05 for all comparisons; Online Supplementary Table S2). The timing of salvage therapy following T0 also impacted the likelihood of a response to the first salvage regimen. Patients who received salvage therapy ≤7 versus >7 days from T0 had significantly higher odds of a response (75% vs. 45%; OR 4.47, 95% CI: 1.07-23.2; P=0.03). Bridging ibrutinib with salvage therapy was also associated with a significantly higher response rate (100% vs. 49%; P=0.01). There was a trend for a higher response rate with ibrutinib bridging versus initiating salvage therapy within 7 days of T0 (100% vs. 58%; P=0.054). In a multivariate model, we evaluated quadruple-class exposed disease against receiving salvage therapy ≤7 days after T0 for the odds of a response to salvage therapy. Both quadruple-class exposed disease (OR 0.20, 95% CI: 0.050.73; P=0.02) and receiving salvage therapy ≤7 days after T0 (OR 4.12, 95% CI: 1.07-18.9; P=0.048) remained inde-

pendently associated with the odds of attaining a response to salvage therapy. Eight patients bridged ibrutinib with the subsequent treatment. Ibrutinib overlapped with the salvage regimen for two cycles in six patients, and one cycle in two patients. The following treatment regimens were added while continuing ibrutinib: bendamustine and rituximab (Benda-R; n=3), bortezomib, dexamethasone, and rituximab (BDR; n=3), ixazomib, dexamethasone, and rituximab (IDR; n=1), and fludarabine and rituximab (Flu-R; n=1). The ORR and MRR to bridging ibrutinib with salvage therapy were both 100%. Six patients were evaluable for an IgM rebound; two patients had developed symptomatic hyperviscosity as part of clinical progression on ibrutinib and were deemed unevaluable for an IgM rebound. Only one patient (17%) had an asymptomatic IgM rebound after bridging ibrutinib with Benda-R for two cycles, which subsequently resolved with two additional treatment cycles. The two non-evaluable patients with symptomatic hyperviscosity were able to stop plasmapheresis after one cycle of bridging with BDR, and then discontinued ibrutinib without evidence of an IgM rebound following one additional cycle of bridging. Ibrutinib was bridged with Flu-R in one patient with BingNeel syndrome, and there was no evidence of an IgM rebound or worsening of neurological symptoms follow-

Figure 1. Peak absolute change in serum immunoglobulin M level following discontinuation of ibrutinib. Values are depicted for each of the 48 patients who were evaluable for an immunoglobulin M (IgM) rebound.

Table 3. Response outcomes according to each salvage regimen utilized following ibrutinib discontinuation.

Salvage Regimen

≥Minor Response (ORR)

≥Partial Response (MRR)

DOR (months)

Benda-R PI-Dex-R Flu-R CCD R-CHOP Ofatumumab Daratumumab Ibrutinib + PI Everolimus

22 8 4 2 2 1 2 1 1

14 (64) 5 (63) 4 (100) 1 (50) 1 (50) 1 (100) 0 (0) 0 (0) 0 (0)

11 (50) 4 (50) 3 (75) 1 (50) 1 (50) 0 (0) 0 (0) 0 (0) 0 (0)

3-yr: 57% 3-yr: 80% 3-yr: 67% 2.7+ 26+ 15 -

A total of 48 of 51 patients (94%) received at least one salvage therapy following time-zero (T0). Five patients received investigational agents and the responses are not included in the table. One patient with AL amyloidosis received consolidation with an autologous stem cell transplant following Benda-R. The following proteasome inhibitors were used as part of a PI-Dex-R regimen: bortezomib (n=5); carfilzomib (n=2); ixazomib (n=1). ORR: overall response rate; MRR: major response rate; DOR: duration of response; yr: years; PI-Dex-R: proteasome inhibitor, dexamethasone, rituximab; Flu-R: fludarabine, rituximab; CCD: carfilzomib, cyclophosphamide, dexamethasone; R-CHOP: rituximab, cyclophosphamide, doxorubicin, vincristine, prednisone.

1166

haematologica | 2022; 107(5)

Acquired ibrutinib resistance in WM

ing T0. Bridging ibrutinib with salvage therapy was well tolerated, and no unexpected toxicities were observed.

Survival outcomes

The median follow-up from T0 was 13 months (range, 0.2-75 months) for the entire cohort, and 20 patients (39%) had died at the time of this report. The median OS from T0 was 51 months (95% CI: 15.3-not reached [NR]), and the 5year OS was 44% (95% CI: 26-75) (Figure 3A). The median OS for the patients who received at least one salvage regimen after T0 was 51 months (95% CI: 21-NR). Patients who did not receive any salvage therapy following T0 (n=3) had a median OS of 0.4 months (95% CI: 0.20-NR), with survival times of 0.2, 0.8, and 0.4 months, respectively. The median OS from WM diagnosis for the entire cohort was 20.4 years (95% CI: 13.2-NR; Online Supplementary Figure S2). The prognostic factors identified in a univariate analysis that impact OS after T0 are shown in Table 4. Only the types of previous therapy received before T0 significantly impacted OS (P=0.018; Figure 3B). Quadruple-class (rituximab, proteasome inhibitor, alkylator, ibrutinib) exposed disease was significantly associated with a shorter OS following T0 (HR 8.08, 95% CI: 1.05-6.21; P=0.04). Among patients without quadruple-class exposed disease, there was no significant difference in OS between the different types of previous therapy (P=0.57). Patients with and without quadruple-class exposed disease had a median OS following T0 of 13.2 months and NR, respectively (P<0.001; Online Supplementary Figure S3). The 5-year OS for patients without quadruple-class exposed disease was 62% (95% CI: 38-98). OS was impacted by the attainment and depth of response to the first salvage regimen after T0. The median OS was significantly longer among patients who achieved a response to the first salvage regimen versus those patients who did not (NR vs. 10.8 months; 95% CI: 0.010.27; P<0.0001; Figure 3C). When stratified by the depth Table 4. Prognostic factors for overall survival at the time of ibrutinib discontinuation (T0).

Patient characteristic

Figure 2. Estimated cumulative incidence of an immunoglobulin M rebound (A) and salvage therapy (B) following discontinuation of ibrutinib. An immunoglobulin M (IgM) rebound occurred in 29 of 48 (60%) evaluable patients. The median time to an IgM rebound was 27 days (95% confidence interval [CI]: 24-33 days). Forty-eight patients (94%) received salvage therapy following time-zero (T0). The median time to salvage therapy was 18 days (95% CI: 13-27 days).

haematologica | 2022; 107(5)

Age >65 yrs. >10 yrs from WM diagnosis Male sex Hemoglobin level <10 mg/dL Platelet count <100,000/μL Serum IgM level >4000 mg/dL >4 prior therapies Type of previous therapy IB IB + R IB + R + PI IB + R + alkylator IB + R + PI + alkylator MYD88 mutation CXCR4 mutation

HR (95% CI)

1.26 (0.50-3.18) 1.60 (0.66-3.87) 1.53 (0.55-4.22) 2.12 (0.85-5.26) 1.92 (0.76-4.81) 0.51 (0.12-2.20) 2.08 (0.75-5.78)

0.62 0.30 0.41 0.11 0.17 0.37 0.16

Reference 1.45 (0.09-23.40) 3.29 (0.35-30.7) 1.20 (0.07-19.4) 8.08 (1.05-62.1) 0.55 (0.12-2.34) 1.36 (0.57-3.27)

0.80 0.30 0.90 0.04 0.41 0.49

The number of previous treatment lines includes ibrutinib monotherapy for all patients. The “types of previous therapy” variable summarizes the different classes of anti-neoplastic agents received throughout the Waldenström macroglobulinemia disease course for each patient. MYD88 and CXCR4 mutation status was available for 46 and 40 patients, respectively. IB: ibrutinib; R: rituximab; PI: proteasome inhibitor, yrs: years; HR: hazard ratio; CI: confidence interval.

1167

J.N. Gustine et al. A

Figure 3. Overall survival following ibrutinib discontinuation in resistant Waldenström macroglobulinemia patients. Kaplan-Meier overall survival curves following discontinuation of ibrutinib for the entire cohort (A) and stratified by types of previous therapy (B), response attainment to first salvage regimen (C), and depth of response to first salvage regimen (D). All patients were previously treated with ibrutinib monotherapy. The “types of previous therapy” variable summarizes the different classes of anti-neoplastic agents received throughout the Waldenström macroglobulinemia disease course for each patient. IB: ibrutinib; R: rituximab; PI: proteasome inhibitor.

of response, the median OS for patients who achieved a major response, minor response, and no response were NR (95% CI: NR-NR), 51.1 months (95% CI: 23-NR), and 10.8 months (95% CI: 6.4-NR), respectively (P<0.001; Figure 3D). The 5-year OS for patients who achieved a major response to the first salvage regimen was 100%. We then evaluated the presence of quadruple-class exposed disease against attaining a response to the first salvage regimen in a multivariate model for OS following T0. Only a response to salvage therapy remained independently associated with OS (HR 0.08, 95% CI: 0.02-0.38; P=0.002), whereas the presence of quadruple-class exposed disease had no impact (P=0.20).

Acquired BTK C481S mutations BTK mutation testing was performed in 21 patients. Seven patients (33%) had a BTK C481S mutation, including one patient with three different BTK C481S variants. There was no difference in the time to ibrutinib discontinuation (T0) between patients with BTK C481S and BTK WT (1.9 vs. 1.8 years; P=0.50; Online Supplementary Figure S4). There was also no difference in age, time from WM diagnosis, sex, hemoglobin level, platelet count, serum IgM level, number or type of prior therapies, and MYD88 and CXCR4 mutation status between patients with BTK 1168

C481S and BTK WT (P>0.05 for all comparisons; Online Supplementary Table S3). Likewise, BTK C481S was not associated with higher or lower odds of an IgM rebound (P=0.99) or response to the first salvage regimen after T0 (P=0.16). By univariate analysis, patients with BTK C481S had a significantly shorter median OS following T0 versus BTK WT (6.4 months vs. NR; P=0.026; Online Supplementary Figure S5). In an exploratory analysis, we evaluated the presence of BTK C481S against quadruple-class exposed disease for OS after T0. In this model, only quadruple-class exposed disease was significantly associated with worse OS (HR 5.50, 95% CI: 1.15-26.2; P=0.03). BTK C481S was not independently associated with OS after adjusting for quadruple-class exposed disease (P=0.09).

TP53 mutations Three of 20 patients (15%) had a TP53 mutation detected. Two TP53 mutations were detected in one patient, and all TP53 mutations localized to the DNA-binding domain. All three patients had mutated MYD88, and two patients had a CXCR4 mutation; no concurrent BTK mutations were identified in the two patients tested. All three patients with a TP53 mutation had an IgM rebound following T0. No patient with a TP53 mutation responded to haematologica | 2022; 107(5)

Acquired ibrutinib resistance in WM

salvage therapy, and all were quadruple-class exposed. Patients with a TP53 mutation had a significantly shorter median OS following T0 versus those without (0.5 vs. 21.3 months; P=0.02; Online Supplementary Figure S6).

Discussion In this study, we sought to describe the natural history of WM patients who acquired resistance to ibrutinib monotherapy. Despite the high response rates and durable remissions, acquired ibrutinib resistance represents an emerging problem in WM patients, and understanding the subsequent disease course may help direct management strategies. Central to our findings was that stopping ibrutinib in resistant WM patients heralded rapid disease progression, which prompted the need for salvage therapy to achieve disease control. This contrasts the indolent posttreatment course typically observed in WM patients following rituximab-based regimens.27-30 Withholding ibrutinib temporarily for adverse events or procedures can also lead to acute increases in serum IgM level, anemia, and constitutional symptoms, highlighting the capacity of tumoral cells to rapidly disseminate disease following ibrutinib withdrawal.10,13,20,31,32 The exact mechanism driving the rapid disease progression after ibrutinib cessation remains to be clarified. However, the BTK substrate STAT5A regulates IgM secretion in WM cells, and its selective reactivation following ibrutinib withdrawal likely contributes to the rapid increase in serum IgM level observed.33,34 In addition, acquired ibrutinib resistance is associated with the clonal expansion of BTK and PLCg2 mutations that trigger prosurvival ERK1/2 signaling and cytokine release, as well as deletions in 6q and 8p that contain regulators of BTK, MYD88/NF-kB, and apoptotic signaling.17-19 It is possible these molecular mechanisms mediating ibrutinib failure contribute to disease acceleration following ibrutinib withdrawal. Indeed, we previously observed a higher risk of rapid disease progression in WM patients discontinuing ibrutinib for acquired resistance versus intolerance, signifying differences in underlying disease biology.20 A similar observation has also been described in patients with chronic lymphocytic leukemia (CLL), wherein rapid increases in serum lymphocyte counts were reported after stopping ibrutinib (i.e., “CLL flare”).35,36 Additional investigation is needed to elucidate whether the rapid disease progression in WM patients is driven by a hypersecretory state, rapid tumor proliferation, or a combination of both. Evaluating both the BM tumor burden and transcriptional signature in WM cells before and after ibrutinib discontinuation would provide further mechanistic insights into this phenomenon. Akin to previous studies, we observed the occurrence of an IgM rebound following discontinuation of ibrutinib.20,21 Rapid increases in serum IgM level can exacerbate WMrelated morbidity caused by the IgM paraprotein, including hyperviscosity, peripheral neuropathy, cold agglutinemia, and cryoglobulinemia.37 In this study, approximately one in three patients with an IgM rebound acutely developed symptomatic hyperviscosity and required emergent plasmapheresis. These findings indicate that close monitoring of serum IgM levels is necessary in WM patients immediately after stopping ibrutinib. Hyperviscosity prophylaxis with plasmapheresis may also warrant considerhaematologica | 2022; 107(5)

ation in WM patients stopping ibrutinib with high serum IgM levels, as the risk of symptomatic hyperviscosity increases exponentially when the serum IgM level rises above 3,000 mg/dL.38 A similar approach is recommended in WM patients receiving rituximab-based therapy to mitigate the risk of hyperviscosity-related injury caused by an IgM flare.39,40 Our data suggest that early initiation of salvage therapy can forestall disease acceleration after stopping ibrutinib. This observation is clinically relevant given the impact of response attainment to salvage therapy on post-ibrutinib survival. Patients who received treatment within 1 week of ibrutinib discontinuation had a significantly lower risk of an IgM rebound, as well as higher response rates to salvage therapy. Notably, bridging ibrutinib in combination with the subsequent therapy for 1-2 cycles achieved an objective response in all patients, and may represent a strategy to maintain disease control in select patients. Similar efficacy with bridging has been reported in ibrutinib-resistant CLL patients who bridged ibrutinib with venetoclax.41 Taken together, these data support the recent consensus guidelines that recommend continuing ibrutinib until the subsequent therapy, plus consideration of bridging, in ibrutinib-resistant WM patients.42 Clinical trials should also consider allowing shorter wash-out periods or overlap of ibrutinib for WM patients in this clinical scenario. The optimal treatment regimen for WM patients after ibrutinib has yet to be established in prospective studies. Our findings demonstrate that standard WM regimens such as Benda-R and BDR are effective as salvage therapy, especially in patients naïve to these agents. Patients with quadruple-class exposed disease, by contrast, had inferior post-ibrutinib outcomes, likely reflecting the presence of a WM clone with little residual sensitivity to available therapies. Importantly, the BCL2 inhibitor venetoclax may represent a novel treatment option for WM patients. Preliminary results from a phase II trial evaluating venetoclax in relapsed or refractory WM patients reported an ORR of 87%, MRR of 81%, and 2-year PFS of 76%. Responses to venetoclax were attained in WM patients previously treated with ibrutinib, akin to studies evaluating venetoclax in ibrutinib-resistant CLL patients.43,44 Combination therapy with IDR or idelalisib plus obinutuzumab are alternative novel salvage regimens, but their activity following ibrutinib is currently unknown.45-48 Non-covalent BTK inhibitors, such as LOXO-305 (clinaltrials gov. Identifier: NCT03740529), vecabrutinib (clinaltrials gov. Identifier: NCT03037645), and ARQ-513 (clinaltrials gov. Identifier: NCT03162536), that bind to nonBTK C481S targets are also under investigation in WM patients. Lastly, a clinical trial is underway with the HCK inhibitor dasatinib for WM patients who are progressing on ibrutinib (clinaltrials gov. Identifier: NCT04115059). Clinical trials have shown CXCR4 mutations confer resistance to ibrutinib monotherapy in WM patients, characterized by lower response rates, delayed response attainment, and shorter PFS.10-15,49 Consistent with these findings, our cohort of ibrutinib-resistant WM patients was enriched for CXCR4 mutations relative to the established incidence (58% vs. 30-40%).3,6,50 Moreover, the majority of CXCR4 mutations were nonsense variants, supporting recent reports that this subtype of CXCR4 mutation shows greater resistance to ibrutinib monotherapy.11,51,52 Combination therapy with ibrutinib plus ritux1169

J.N. Gustine et al.

imab is also adversely impacted by CXCR4 mutations, with a shorter 36-month PFS in CXCR4 mutated versus CXCR4 WT WM patients (64% vs. 84%, respectively).53-55 Given the importance of CXCR4 mutations, clinical trials evaluating the CXCR4 inhibitors ulocuplumab (clinaltrials gov. Identifier: NCT03225716) and mavorixafor (clinaltrials gov. Identifier: NCT04274738) in combination with ibrutinib are currently ongoing in CXCR4-mutated WM patients. A notable finding was the similar disease course between BTK C481S and BTK WT ibrutinib-resistant WM patients. It is possible a shared ERK1/2 signature underlies this clinical observation. In WM patients with BTK WT, PLCg2 mutations and DOK2 deletions were identified as possible molecular mechanisms driving acquired ibrutinib resistance.17,19 Both are predicted to trigger ERK1/2 signaling similar to the effect of BTK C481S mutations,18,56 although studies are needed to confirm the functional significance of PLCg2 and DOK2 in WM. These studies may also inform the utility of ERK1/2 inhibitors as a strategy to overcome acquired ibrutinib resistance in WM patients with BTK WT. The use of an ERK1/2 inhibitor has previously been shown to abrogate the effects of BTK C481S in WM cells and restore sensitivity to ibrutinib.18 We also observed TP53 mutations were associated with refractory disease and shorter survival after acquiring resistance to ibrutinib. Although both preclinical and clinical data suggest ibrutinib has activity in TP53-mutated WM patients, additional work is needed to identify novel treatments for this high-risk group.57,58 A phase II trial evaluating ibrutinib in previously untreated WM patients with serial wholeexome sequencing is now complete and will provide additional insights into mechanisms of ibrutinib resistance, as well as the impact of ibrutinib on clonal evolution (clinicaltrials gov. Identifier: NCT02604511). Limitations of this study include the inherent selection bias associated with a retrospective study from a single tertiary referral center. Nevertheless, this study constitutes the largest clinical experience of WM patients with acquired ibrutinib resistance, and the patients included are

References 1. Owen RG, Treon SP, Al-Katib A, et al. Clinicopathological definition of Waldenstrom's macroglobulinemia: consensus panel recommendations from the Second International Workshop on Waldenstrom's Macroglobulinemia. Semin Oncol. 2003;30(2):110-115. 2. Treon SP, Xu L, Yang G, et al. MYD88 L265P somatic mutation in Waldenström's macroglobulinemia. N Engl J Med. 2012; 367(9):826-833. 3. Hunter ZR, Xu L, Yang G, et al. The genomic landscape of Waldenström macroglobulinemia is characterized by highly recurring MYD88 and WHIM-like CXCR4 mutations, and small somatic deletions associated with B-cell lymphomagenesis. Blood. 2014;123(11):1637-1646. 4. Yang G, Zhou Y, Liu X, et al. A mutation in MYD88 (L265P) supports the survival of lymphoplasmacytic cells by activation of Bruton tyrosine kinase in Waldenström macroglobulinemia. Blood. 2013; 122(7): 1222-1232. 5. Yang G, Buhrlage SJ, Tan L, et al. HCK is a

1170

representative of those who participate in clinical trials. This study can therefore serve as a “real-world” benchmark for assessing new drugs in WM patients with acquired ibrutinib resistance. In conclusion, our findings show that discontinuation of ibrutinib can herald rapid disease progression in WM patients with acquired ibrutinib resistance. A rapid rebound in serum IgM level frequently occurs and can cause symptomatic hyperviscosity. Continuing ibrutinib until the subsequent treatment, with consideration of bridging, may represent a reasonable strategy to maintain disease control. Prospective studies are needed to clarify the optimal management of WM patients with acquired ibrutinib resistance. Disclosures SPT, JJC, GY, and ZRH have received research funding and/or consulting fees from Pharmacyclics and Janssen Pharmaceuticals; SPT has received research funding from Bristol Myers Squibb, X4 Pharmaceuticals, and Beigene; JJC received research funding and/or consulting fees from Abbvie, Beigene, Roche, and TG Therapeutics. Contributions JNG, SS, SPT, and JJC designed the study and performed the data analysis; MLG, LX, AK, NT, MM, MD, XL, GY, and ZRH performed molecular testing on patient samples; SS, CAF, KM, CL, TW, CJP, ARB, SPT and JJC took care of the patients and collected the samples; JNG, SPT, and JJC drafted the manuscript. All authors critically reviewed and approved the manuscript. Funding JJC would like to thank the support of the WMR Fund. The authors would also like to thank the Siegel Research Fund for WM, the Orszag Family Fund for WM Research, the D’Amato Family Fund for WM Research, the International Waldenström’s Macroglobulinemia Foundation, and the Kerry Robertson Fund for WM. SPT, ZRH and GY are supported by an NIH SPORE in Multiple Myeloma (grant: 2P50CA100707-16A1).

survival determinant transactivated by mutated MYD88, and a direct target of ibrutinib. Blood. 2016;127(25):3237-3252. 6. Xu L, Hunter ZR, Tsakmaklis N, et al. Clonal architecture of CXCR4 WHIM-like mutations in Waldenström macroglobulinaemia. Br J Haematol. 2016; 172(5):735744. 7. Cao Y, Hunter ZR, Liu X, et al. The WHIMlike CXCR4S338X somatic mutation activates AKT and ERK, and promotes resistance to ibrutinib and other agents used in the treatment of Waldenstrom’s macroglobulinemia. Leukemia. 2015; 29(1): 169-176. 8. Cao Y, Hunter ZR, Liu X, et al. CXCR4 WHIM-like frameshift and nonsense mutations promote ibrutinib resistance but do not supplant MYD88L265P-directed survival signalling in Waldenström macroglobulinaemia cells. Br J Haematol. 2015; 168(5):701-707. 9. Roccaro AM, Sacco A, Jimenez C, et al. C1013G/CXCR4 acts as a driver mutation of tumor progression and modulator of drug resistance in lymphoplasmacytic lymphoma. Blood. 2014;123(26):4120-4131. 10. Treon SP, Tripsas CK, Meid K, et al.

Ibrutinib in previously treated Waldenström’s macroglobulinemia. N Engl J Med. 2015;372(15):1430-1440. 11. Treon SP, Meid K, Gustine J, et al. Longterm follow-up of ibrutinib monotherapy in symptomatic, previously treated patients with Waldenström macroglobulinemia. J Clin Oncol. 2021;39(6):565-575. 12. Treon SP, Xu L, Hunter Z. MYD88 mutations and response to ibrutinib in Waldenström's macroglobulinemia. N Engl J Med. 2015;373(6):584-586. 13. Treon SP, Gustine J, Meid K, et al. Ibrutinib monotherapy in symptomatic, treatmentnaïve patients with Waldenström macroglobulinemia. J Clin Oncol. 2018; 36(27):2755-2761. 14. Dimopoulos MA, Trotman J, Tedeschi A, et al. Ibrutinib for patients with rituximabrefractory Waldenstrom's macroglobulinaemia (iNNOVATE): an open-label substudy of an international, multicentre, phase 3 trial. Lancet Oncol. 2017;18(2):241250. 15. Trotman J, Buske C, Tedeschi A, et al. Long-term follow-up of ibrutinib treatment for rituximab-refractory Waldenström's macroglobulinemia: final analysis of the

haematologica | 2022; 107(5)

Acquired ibrutinib resistance in WM

open-label substudy of the phase 3 iNNOVATETM Trial. Blood. 2020; 136(Suppl 1):S38-39. 16. Tam CS, Opat S, D'Sa S, et al. A randomized phase 3 trial of zanubrutinib vs ibrutinib in symptomatic Waldenström macroglobulinemia: the ASPEN study. Blood. 2020;136(18):2038-2050. 17. Xu L, Tsakmaklis N, Yang G, et al. Acquired mutations associated with ibrutinib resistance in Waldenström macroglobulinemia. Blood. 2017;129(18):2519-2525. 18. Chen JG, Liu X, Munshi M, et al. BTKCys481Ser drives ibrutinib resistance via ERK1/2 and protects BTKwild-type MYD88-mutated cells by a paracrine mechanism. Blood. 2018;131(18):2047-2059. 19. Jiménez C, Chan GG, Xu L, et al. Genomic evolution of ibrutinib-resistant clones in Waldenström macroglobulinaemia. Br J Haematol. 2020;189(6):1165-1170. 20. Gustine JN, Meid K, Dubeau T, et al. Ibrutinib discontinuation in Waldenström macroglobulinemia: etiologies, outcomes, and IgM rebound. Am J Hematol. 2018; 93(4):511-517. 21. Abeykoon JP, Zanwar S, Ansell SM, et al. Ibrutinib monotherapy outside of clinical trial setting in Waldenström macroglobulinaemia: practice patterns, toxicities and outcomes. Br J Haematol. 2020;188(3):394403. 22. Owen RG, Kyle RA, Stone MJ, et al. Response assessment in Waldenström macroglobulinaemia: update from the VIth International Workshop. Br J Haematol. 2013;160(2):171-176. 23. Palladini G, Dispenzieri A, Gertz MA, et al. New criteria for response to treatment in immunoglobulin light chain amyloidosis based on free light chain measurement and cardiac biomarkers: impact on survival outcomes. J Clin Oncol. 2012;30(36):45414549. 24. Cheson BD, Fisher RI, Barrington SF, et al. Recommendations for initial evaluation, staging, and response assessment of Hodgkin and non-Hodgkin lymphoma: the Lugano Classification. J Clin Oncol. 2014; 32(27):3059-3067. 25. Xu L, Hunter ZR, Yang G, et al. MYD88 L265P in Waldenström macroglobulinemia, immunoglobulin M monoclonal gammopathy, and other B-cell lymphoproliferative disorders using conventional and quantitative allele-specific polymerase chain reaction. Blood. 2013;121(11):2051-2058. 26. Kluk MJ, Lindsley RC, Aster JC, et al. Validation and implementation of a custom next-generation sequencing clinical assay for hematologic malignancies. J Mol Diagn. 2016;18(4):507-515. 27. Treon SP, Ioakimidis L, Soumerai JD, et al. Primary therapy of Waldenström macroglobulinemia with bortezomib, dexamethasone, and rituximab: WMCTG clinical trial 05-180. J Clin Oncol. 2009; 27(23):3830-3835. 28. Dimopoulos MA, García-Sanz R, Gavriatopoulou M, et al. Primary therapy of Waldenstrom macroglobulinemia (WM) with weekly bortezomib, low-dose dexamethasone, and rituximab (BDR): long-term results of a phase 2 study of the European Myeloma Network (EMN). Blood. 2013; 122(19):3276-3282. 29. Rummel MJ, Niederle N, Maschmeyer G, et al. Bendamustine plus rituximab versus CHOP plus rituximab as first-line treatment for patients with indolent and mantle-cell lymphomas: an open-label, multi-

haematologica | 2022; 107(5)

centre, randomised, phase 3 non-inferiority trial. Lancet. 2013;381(9873):1203-1210. 30. Castillo JJ, Gustine JN, Meid K, et al. Response and survival for primary therapy combination regimens and maintenance rituximab in Waldenström macroglobulinaemia. Br J Haematol. 2018;181(1):77-85. 31. Castillo JJ, Gustine JN, Meid K, Dubeau T, Severns P, Treon SP. Ibrutinib withdrawal symptoms in patients with Waldenström macroglobulinemia. Haematologica. 2018; 103(7):e307-e310. 32. Castillo JJ, Gustine JN, Meid K, et al. Impact of ibrutinib dose intensity on patient outcomes in previously treated Waldenström macroglobulinemia. Haematologica. 2018; 103(10):e466-e468. 33. Hodge LS, Ziesmer SC, Yang Z-Z, Secreto FJ, Novak AJ, Ansell SM. Constitutive activation of STAT5A and STAT5B regulates IgM secretion in Waldenstrom's macroglobulinemia. Blood. 2014; 123(7): 1055-1058. 34. Mahajan S, Vassilev A, Sun N, Ozer Z, Mao C, Uckun FM. Transcription factor STAT5A is a substrate of Bruton's tyrosine kinase in B cells. J Biol Chem. 2001;276(33):3121631228. 35. Maddocks KJ, Ruppert AS, Lozanski G, et al. Etiology of ibrutinib therapy discontinuation and outcomes in patients with chronic lymphocytic leukemia. JAMA Oncol. 2015;1(1):80-87. 36. Hampel PJ, Ding W, Call TG, et al. Rapid disease progression following discontinuation of ibrutinib in patients with chronic lymphocytic leukemia treated in routine clinical practice. Leuk Lymphoma. 2019; 60(11):2712-2719. 37. Treon SP. How I treat Waldenström macroglobulinemia. Blood. 2015; 126(6): 721-732. 38. Gustine JN, Meid K, Dubeau T, et al. Serum IgM level as predictor of symptomatic hyperviscosity in patients with Waldenström macroglobulinaemia. Br J Haematol. 2017;177(5):717-725. 39. Treon SP, Branagan AR, Hunter Z, Santos D, Tournhilac O, Anderson KC. Paradoxical increases in serum IgM and viscosity levels following rituximab in Waldenstrom's macroglobulinemia. Ann Oncoly. 2004;15(10):1481-1483. 40. Ghobrial IM, Fonseca R, Greipp PR, et al. Initial immunoglobulin M ‘flare’ after rituximab therapy in patients diagnosed with Waldenstrom macroglobulinemia. Cancer. 2004;101(11):2593-2598. 41. Hampel PJ, Call TG, Ding W, et al. Addition of venetoclax at time of progression in ibrutinib-treated patients with chronic lymphocytic leukemia: Combination therapy to prevent ibrutinib flare. Am J Hematol. 2020;95(3):E57-e60. 42. Castillo JJ, Advani RH, Branagan AR, et al. Consensus treatment recommendations from the tenth International Workshop for Waldenstrom Macroglobulinaemia. Lancet Haematol. 2020;7(11):e827-e837. 43. Castillo J, Allan J, Siddiqi T, et al. Multicenter prospective phase II study of venetoclax in patients with previously treated Waldenstrom macroglobulinemia. Clin Lymphoma Myeloma Leuk. 2019;19(10, Supplement):e39-e40. 44. Jones JA, Mato AR, Wierda WG, et al. Venetoclax for chronic lymphocytic leukaemia progressing after ibrutinib: an interim analysis of a multicentre, openlabel, phase 2 trial. Lancet Oncol. 2018; 19(1):65-75.

45. Castillo JJ, Meid K, Gustine JN, et al. Prospective clinical trial of ixazomib, dexamethasone, and rituximab as primary therapy in Waldenström macroglobulinemia. Clin Cancer Res. 2018;24(14):32473252. 46. Castillo JJ, Meid K, Flynn CA, et al. Ixazomib, dexamethasone, and rituximab in treatment-naive patients with Waldenström macroglobulinemia: longterm follow-up. Blood Adv. 2020; 4(16):3952-3959. 47. Kersten MJ, Minnema MC, Vos JM, et al. Ixazomib, rituximab and dexamethasone (IRD) in patients with relapsed or progressive Waldenstrom's macroblobulinemia: results of the prospective phase I/II HOVON 124/Ecwm-R2 trial. Blood. 2019;134(Suppl 1):S344. 48. Tomowiak C, Desseaux K, Poulain S, et al. Open label non-randomized phase II study exploring «chemo-free » treatment association with idelalisib + obinutuzumab in patients with relapsed/refractory (R/R) Waldenstrom's macroglobulinemia (MW), a Filo trial: results of the intermediary analysis of the induction phase. Blood. 2019;134(Suppl 1):S346. 49. Castillo JJ, Gustine JN, Meid K, et al. Response and survival outcomes to ibrutinib monotherapy for patients with Waldenström macroglobulinemia on and off clinical trials. Hemasphere. 2020; 4(3):e363. 50. Treon SP, Cao Y, Xu L, Yang G, Liu X, Hunter ZR. Somatic mutations in MYD88 and CXCR4 are determinants of clinical presentation and overall survival in Waldenström macroglobulinemia. Blood. 2014;123(18):2791-2796. 51. Castillo JJ, Xu L, Gustine JN, et al. CXCR4 mutation subtypes impact response and survival outcomes in patients with Waldenström macroglobulinaemia treated with ibrutinib. Br J Haematol. 2019; 187(3):356-363 52. Gustine JN, Xu L, Tsakmaklis N, et al. CXCR4S338X clonality is an important determinant of ibrutinib outcomes in patients with Waldenström macroglobulinemia. Blood Adv. 2019;3(19):2800-2803. 53. Buske C, Tedeschi A, Trotman J, et al. Ibrutinib treatment in Waldenström’s macroglobulinemia: follow-up efficacy and safety from the iNNOVATE study. Blood. 2018;132(Supplement 1):149. 54. Dimopoulos MA, Tedeschi A, Trotman J, et al. Phase 3 trial of ibrutinib plus rituximab in Waldenström’s macroglobulinemia. N Engl J Med. 2018;378(25):2399-2410. 55. Buske C, Tedeschi A, Trotman J, et al. Fiveyear follow-up of Ibrutinib plus rituximab vs. placebo plus rituximab for Waldenstrom's macroglobulinemia: final analysis from the randomized Phase 3 iNNOVATETM Study. Blood. 2020; 136(Supplement 1):24-26. 56. Shinohara H, Inoue A, Toyama-Sorimachi N, et al. Dok-1 and Dok-2 are negative regulators of lipopolysaccharide-induced signaling. J Exp Med. 2005;201(3):333-339. 57. Poulain S, Roumier C, Bertrand E, et al. TP53 mutation and its prognostic significance in Waldenstrom's macroglobulinemia. Clin Cancer Res. 2017;23(20):6325-6335. 58. Gustine JN, Tsakmaklis N, Demos MG, et al. TP53 mutations are associated with mutated MYD88 and CXCR4, and confer an adverse outcome in Waldenström macroglobulinaemia. Br J Haematol. 2019; 184(2):242-245.

1171

ARTICLE Ferrata Storti Foundation

Quality of Life

Randomized controlled trial of geriatric consultation versus standard care in older adults with hematologic malignancies Clark DuMontier,1,2,3 Hajime Uno,3,4 Tammy Hshieh,1,3,5 Guohai Zhou,1,3 Richard Chen,5 Emily S. Magnavita,5 Lee Mozessohn,6 Houman Javedan,1,3 Richard M. Stone,3,5 Robert J. Soiffer,3,5 Jane A. Driver1,2,3# and Gregory A. Abel3,5# Division of Aging, Brigham and Women’s Hospital, Boston, MA, USA; 2New England Geriatric Research Education and Clinical Center, VA Boston Healthcare System, Boston, MA, USA; 3Harvard Medical School, Boston, MA, USA; 4Department of Data Sciences, Dana-Farber Cancer Institute, Boston, MA, USA; 5Department of Medical Oncology, DanaFarber Cancer Institute, Boston, MA, USA and 6Sunnybrook Odette Cancer Centre, Toronto, Ontario, Canada 1

Haematologica 2022 Volume 107(5):1172-1180

JAD and GAA contributed equally as co-senior authors.

ABSTRACT

Correspondence: GREGORY A. ABEL gregory_abel@dfci.harvard.edu Received: March 18, 2021. Accepted: September 3, 2021. Pre-published: September 23, 2021. https://doi.org/10.3324/haematol.2021.278802

1172

e conducted a randomized controlled trial in older adults with hematologic malignancies to determine the impact of geriatrician consultation embedded in our oncology clinic alongside standard care. From February 2015 to May 2018, transplant-ineligible patients aged ≥75 years who presented for initial consultation for lymphoma, leukemia, or multiple myeloma at Dana-Farber Cancer Institute (Boston, MA, USA) were eligible. Pre-frail and frail patients, classified based on phenotypic and deficit-accumulation approaches, were randomized to receive either standard oncologic care with or without consultation with a geriatrician. The primary outcome was 1-year overall survival. Secondary outcomes included unplanned care utilization within 6 months of follow-up and documented end-of-life (EOL) goals-of-care discussions. Clinicians were surveyed as to their impressions of geriatric consultation. One hundred sixty patients were randomized to either geriatric consultation plus standard care (n=60) or standard care alone (n=100). The median age of the patients was 80.4 years (standard deviation = 4.2). Of those randomized to geriatric consultation, 48 (80%) completed at least one visit with a geriatrician. Consultation did not improve survival at 1 year compared to standard care (difference: 2.9%, 95% confidence interval: -9.5% to 15.2%, P=0.65), and did not significantly reduce the incidence of emergency department visits, hospital admissions, or days in hospital. Consultation did improve the odds of having EOL goals-of-care discussions (odds ratio = 3.12, 95% confidence interval: 1.03 to 9.41) and was valued by surveyed hematologic-oncology clinicians, with 62.9%-88.2% of them rating consultation as useful in the management of several geriatric domains.

Introduction Older adults constitute the majority of patients with hematologic malignancies, as the median ages at diagnosis of non-Hodgkin lymphoma, leukemia, and multiple myeloma are 67, 67, and 69 years, respectively.1-3 Compared with younger patients, older patients with blood cancers often have age-related vulnerabilities that complicate their care.4 Cognitive impairment, functional dependency, and frailty are prevalent and associated with worse outcomes such as increased treatment toxicity, unplanned hospitalizations, and higher mortality.5-7 To manage this complexity, cancer organizations such as the American Society of Clinical Oncology (ASCO) recommend that all older adults with cancer treated with chemotherapy undergo a geriatric assessment (GA): a multidisciplinary evaluation of domains necessary for older adult health and well-being.8

haematologica | 2022; 107(5)

RCT of geriatric consultation in blood cancers

There is strong evidence from observational studies that, in comparison with standard oncologic assessment, GA better identifies age-related vulnerabilities, guides the care of these vulnerabilities, influences treatment decisions, and predicts outcomes in older patients with cancer.9-14 Moreover, recent randomized controlled trials suggest that GA-guided interventions may reduce treatment toxicity, improve quality of life, and improve communication with patients and caregivers.15-17 Unfortunately, these trials include mainly older patients with solid tumors. To our knowledge, no similar trials in patients with blood cancers have been reported. We thus leveraged the embedded geriatrics resources available in our outpatient blood cancer clinic to determine the impact of consultation with a geriatrician alongside standard oncologic care for patients aged 75 and older with hematologic malignancies.

Methods Patients and study design This randomized controlled trial enrolled patients from February 2015 to May 2018 (ClinicalTrials.gov identifier NCT02359838) (Online Supplementary File S1). Eligible patients included all patients aged 75 years and older who presented to Dana-Farber Cancer Institute (Boston, MA, USA) for initial consultation seeking management for newly diagnosed or previously diagnosed and treated lymphoma, leukemia, or multiple myeloma. Patients were ineligible if they were referred for consultation for stem cell transplantation or did not plan to continue their care at our institution. Eligible patients who consented to participate in the study underwent an in-person screening GA administered by a research assistant on the same day as their initial hematologic oncology consultation, as described previously.5 From this assessment, frailty status was derived using both the phenotypic and deficit-accumulation approaches - two of the most widely-studied approaches in aging research (see the protocol in Online Supplementary File S1 for further details regarding these approaches and their cut-off values that classified severity of frailty).18,19 In brief, the frailty phenotype uses five criteria to define a syndrome (slow gait, weakness [grip strength], self-reported exhaustion, low physical activity, and weight loss; average time to complete, 5-10 minutes). The deficit-accumulation method counts numerous aging-related health deficits across multiple domains from a GA to define frailty as the proportion of deficits present in an individual out of the total number of possible deficits measured (average time to complete, 15-20 minutes). We did not use a disease-specific frailty score such as the International Myeloma Working Group score. Patients classified as pre-frail or frail by either approach were randomized to either standard oncologic care as they would normally receive at Dana-Farber or standard care plus embedded consultation provided by a geriatrician. All oncologists were blinded to the initial geriatric screening and frailty classification, precluding an influence on initial treatment recommendations. Oncologists of patients in the intervention group may have become aware of patients’ frailty status later in the study after the patients had been assessed by the geriatrician. Randomization was stratified by disease type to minimize potential imbalances in blood cancers and treatments. Randomization was first conducted on a 1:1 ratio but was switched to a 2:1 ratio (standard care: standard care plus geriatric consultation) to increase enrollment, which was initially limited due to difficulties in scheduling patients assigned to the intervention arm to one of the twice-weekly geriatric clinic sessions. The study was powered to detect a difference in 1-year overall survival

haematologica | 2022; 107(5)

(the primary outcome) of 25% between study arms, hypothesizing that the geriatric consultation arm would demonstrate this percent improvement in survival. This effect size was estimated based on prior observational data regarding survival rates in similar patients treated at Dana-Farber and the survival benefit associated with integrated palliative care in patients with lung cancer.20 With 2:1 randomization, the sample size needed to detect the estimated effect size was calculated to be 160 - 107 in the standard care arm and 53 in the arm with geriatric consultation - assuming 80% power and a one-sided type I error rate of 5%. This sample size was recalculated from 152 (76 per group), which was originally calculated for 1:1 randomization. The study was approved by the Dana-Farber/Harvard Cancer Center Office for the Protection of Human Research Subjects.

Geriatric consultation intervention Patients who were assigned to the intervention received embedded geriatric consultation with a licensed geriatrician in addition to their standard oncologic care managed by their hematologic oncologist. The embedded geriatrics clinic is located within DanaFarber on the same floors as the hematologic malignancies clinics. Of note, embedded geriatric consultation had been available for patients referred from the leukemia clinic (without prior GA screening and randomization) for 2 years preceding the start of the trial. After assignment, patients from leukemia, lymphoma, and myeloma clinics were scheduled with a geriatrician either on the same day as their follow-up oncology consultation or at a different time in accordance with the patient’s schedule and appointment availability; we intended for the patients to be seen as early as possible but did not require a specific time period for the first visit. Consistent with other trial designs evaluating the effectiveness of integrated subspecialty care,20 the geriatrician provided further management and interventions individualized to the patient based on clinical judgment and best-available evidence; no pre-specified interventions were required. If indicated, geriatricians communicated with the patient’s primary care provider and utilized referral systems (e.g., physical therapy, psychiatry) already established at Dana-Farber. Follow-up appointments were encouraged, but not required. In keeping with routine care provided by geriatricians, the geriatrician conducted a GA for every patient encountered. To characterize the interventions recommended by the geriatrician, a content analysis of the geriatricians’ notes was conducted.21 For each patient we classified whether the geriatrician recommended an intervention targeting one or more domains described in ASCO’s Guideline for Geriatric Oncology: (i) function and falls; (ii) comorbidity and polypharmacy; (iii) cognition; (iv) depression/mood; and (v) nutrition.8 Recommended interventions could include counseling, recommendations for non-pharmacological interventions, pharmacological interventions, and referrals to other specialties or allied healthcare. For each patient, all geriatricians’ notes through the 1-year follow-up period were reviewed, and new interventions were only counted once.

Outcome measures The primary outcome of this study was 1-year overall survival from the time of initial hematologic oncology consultation. Vital status was confirmed by a combination of chart review and calls to patients’ primary care providers. Secondary outcomes were assessed via chart review and included the number of emergency department visits, the number of unplanned hospital admissions, and the number of days spent in the hospital22 within 6 months after patients’ initial consultations at Dana-Farber. Having any end-of-life (EOL) goals-of-care discus-

1173

C. DuMontier et al.

sions documented in the medical record during the 1-year followup period was also measured via chart review. EOL goals-of-care discussion was defined as a discussion regarding EOL preferences by any treating clinician including resuscitation/code status, hospice, and/or preferred location for dying.23 Finally, after completion of enrollment, a survey was administered to 65 Dana-Farber hematologic oncologists, physician assistants, and nurse practitioners who cared for patients in the geriatric consultation arm. This survey sought to gather clinicians’ opinions regarding the usefulness of consultation (on a Likert scale where 1 = “least” useful and 5 = “most” useful) in addressing geriatric domains of care and areas of management for specific age-related issues (see survey instrument, Online Supplementary Figure S2).

Statistical analysis For the primary analysis, the impact of geriatric consultation on 1-year overall survival was assessed using Kaplan-Meier analysis comparing the 1-year survival rate between patients receiving geriatric consultation plus standard oncologic care and patients receiving standard oncologic care alone. Differences in 1-year survival rate and a corresponding 95% confidence interval (95% CI) were calculated to summarize the effect of geriatric consultation on 1-year overall survival.24,25 Multivariable Cox regression and weighted logistic regression models26 were also used to estimate the treatment effect, adjusting for any potential remaining imbalances after randomization related to age, sex, disease aggressiveness (defined according to previous methods5-7), and frailty (prefrail versus frail). Aggressive diseases included diffuse large B-cell lymphoma, mantle cell lymphoma, multiple myeloma, acute myeloid leukemia, and indolent diseases included marginal zone lymphoma, follicular lymphoma, chronic lymphocytic leukemia, myelodysplastic syndrome, myeloproliferative neoplasm/myeloproliferative disease, and hairy-cell leukemia. For secondary analyses, the effects of geriatric consultation on emergency department visits, hospitalizations, and number of days in the hospital were assessed using separate negative binomial regression models, each adjusting for age, sex, disease aggressiveness, and frailty. The impact of geriatric consultation on the likelihood of having documented EOL goals-of-care discussions during the follow-up period was assessed using multivariable logistic regression, adjusting for the aforementioned covariates. Exploratory analyses investigated any association between number of geriatrician visits and mortality, as well as a subgroup analysis determining any difference in effect by frailty severity. All primary and secondary analyses were performed as intention-totreat analyses, followed by per-protocol analyses that excluded patients who, although assigned to the intervention, ended up not completing their geriatric visit. SAS (version 9.4, SAS Institute, Cary, NC, USA) and R (version 4.0,0, https://www.R-project.org, R foundation for Statistical Computing, Vienna, Austria) statistical software were used for all analyses.

Results Patients’ characteristics Between February 2015 and May 2018, 270 eligible patients with planned follow-up at Dana-Farber were approached for enrollment (Figure 1). Of these, 232 agreed to participate and underwent the screening GA, after which 72 patients were classified as robust and thus excluded from the trial. One hundred sixty pre-frail/frail patients were randomly assigned to receive geriatric consultation plus standard oncologic care (n=60) or standard care alone (n=100). One patient in the standard care arm was lost to follow-up 1174

because the patient never returned to Dana-Farber after initial consult and vital status could not be confirmed. This patient was assumed to be alive at the end of the study period and was included in the analyses. In the intervention arm, three patients died before receiving their consultation, three cancelled the consultation, and six did not return to Dana Farber (i.e., they continued their care at their local practice). The two study arms were overall balanced in terms of baseline characteristics (Table 1), with high rates of functional impairment (35.6% with dependency in instrumental activities of daily living [IADL]), cognitive impairment (39.5% with impairment in executive function), and mobility impairment (60.6% with gait speed <0.8 meters/second). Online Supplementary Table S1 lists the latest active treatment regimens within 3 months of initial consultation.

Uptake of the embedded geriatric consultation Of those randomized to geriatric consultation, 48 (80%) completed at least one visit with a geriatrician (95% CI: 68% to 88%). Of those 12 assigned to receive geriatric consultation who did not complete it, three died, three cancelled the consultation (although continued their cancer care at Dana-Farber), and six ended up not returning to Dana-Farber for further care. Among the 48 who completed at least one consultation, 26 completed one or more additional visits with a geriatrician (range of total visits per patient, 1-12). Patients enrolled toward the end of the study period tended to have more total visits than patients enrolled toward the beginning (Online Supplementary Figure S2).

Geriatric consultation and 1-year overall survival After being randomized to the geriatric consultation arm, time to the initial visit with a geriatrician varied across patients with a median of 36 days (range, 0-224 days; interquartile range, 76 days). The median follow-up extended beyond our outcome of 1-year survival. Among the 48 patients who were seen by the geriatrician in the consultation arm, the median number of interventions recommended for each patient was two, with a range of zero to four interventions. The most common interventions fell within the comorbidity/polypharmacy domain (39 [81.3%] patients receiving one or more interventions); followed by nutrition (26 [54.2%]); function/falls (23 [47.9%]); cognition (15 [31.3%]); and depression/mood (8 [16.7%]). Ninetyseven of these interventions were carried out by the geriatrician through counseling, non-pharmacological recommendations, or pharmacological prescriptions. Fourteen of these interventions were referrals or coordination with other disciplines, including physical therapists, social workers, and nutritionists. No control patients crossed over to the consultation arm in the 1-year follow-up period (i.e., no control patient received an embedded geriatric consultation). A cumulative total of 32 patients died in the year following their initial consultation, 11 (18.3%) in the geriatric consultation arm versus 21 (21.0%) in the standard care arm. Overall survival at 1 year was not significantly higher in patients receiving geriatric consultation (81.7%, 95% CI: 71.0% to 90.2%) in comparison with patients receiving standard care (78.8%, 95% CI: 69.7% to 85.7%; difference: 2.9%, 95% CI: -9.5% to 15.2%, P=0.65) (Figure 2A). Results were similar in the per-protocol analysis (Figure 2B), as were results after adjustment for covariates in the multihaematologica | 2022; 107(5)

RCT of geriatric consultation in blood cancers

variable analyses (Table 2, Online Supplementary Tables S2 and S3). Moreover, there was no significant association between the number of visits with a geriatrician and mortality (hazard ratio = 0.78, 95% CI: 0.43 to 1.39), and there was no difference in the effect of consultation on mortality among frail versus pre-frail patients (test for interaction P=0.41). Frail patients experienced higher mortality, independently of intervention or other covariates (Online Supplementary Table S2).

Geriatric consultation and acute care utilization Thirty-six of 160 patients (22.5%) experienced one or more unplanned hospitalizations during the first 6 months of follow-up, and the same number made one or more emergency department visits. In comparison with patients who received standard oncologic care, patients who received geriatric consultation did not have a significantly lower incidence of emergency department visits (incident rate ratio [IRR] = 0.89, 95% CI: 0.33 to 2.42), hospitalizations (IRR = 0.91, 95% CI: 0.30 to 2.71), or days spent in

hospital (IRR = 1.05, 95% CI: 0.29 to 3.79), adjusting for covariates (Table 2). Per-protocol analyses yielded similar results.

Geriatric consultation and end-of-life goals-of-care discussions Seventeen of 160 patients (10.6%) received one or more EOL goals-of-care discussions during follow-up. In comparison with patients who received standard oncologic care, patients who received geriatric consultation had an over three-fold higher odds of a documented goals-of-care discussion (odds ratio = 3.12, 95% CI: 1.03 to 9.41). Per-protocol analyses yielded similar results (odds ratio = 3.58, 95% CI: 1.13 to 11.35). Three patients in each arm received a palliative care consultation.

Hematologic oncologists’ and other clinicians’ perceived value of geriatric consultation Thirty-five of 65 (53.8%) hematologic oncologists, nurse practitioners, and physician assistants whose patients had

Figure 1. CONSORT flow diagram of trial enrollment and analysis. DFCI: Dana-Farber Cancer Institute; GA: geriatric assessment.

haematologica | 2022; 107(5)

1175

C. DuMontier et al.

received geriatric consultation responded to the survey evaluating the perceived value of geriatric recommendations. The majority found embedded geriatric consultation to be valuable in managing several age-related domains of care (Table 3). Domains of care in which consultation was found to be most valued included evaluation of cognition, connecting patients to resources, diagnosing frailty, and managing non-oncologic comorbidities. Specific areas of management found to be most useful included optimizing functional status, treating falls, and treatment of depression and other mood disorders.

Discussion We found that in pre-frail and frail older patients with hematologic malignancies, embedded geriatric consultation did not improve 1-year overall survival or acute care utilization. However, consultation significantly increased the likelihood of having EOL goals-of-care discussions. Moreover, hematologic oncology clinicians highly valued the services provided by geriatrics in the care of their older patients. Our trial addresses a critical gap regarding the effectiveness of geriatric-driven interventions in older patients with hematologic malignancies, complementing emerging evidence from other studies predominantly including older adults with solid tumors.15-17,27-31 Few prior studies have investigated GA-driven interventions for patients with blood cancers. Artz and colleagues recently reported that in patients with a median age of 67 years undergoing hematopoietic stem cell transplantation for blood cancers, GA-driven interventions implemented by a multidisciplinary geriatrics team improved 1-year overall survival in comparison with conducting a GA alone without a multidisciplinary team to manage any detected vulnerabilities.32 While provocative, a limitation of this nonrandomized study was its use of a historical control group for comparison. In our randomized trial of transplant-ineligible patients aged ≥75 years undergoing geriatric consultation versus standard oncologic care, we did not find evidence of an effect on 1-year overall survival, even in the frail subgroup. Care for frail older adults is often complex and fragmented.33 A significant strength of our consultative model is that our geriatricians - trained specialists in frailty and complex care - were embedded in our center, caring for patients alongside their hematologic oncologists in the same clinic. On the other hand, certain aspects of our model may have limited its effectiveness in both reducing mortality and acute care utilization. First, although 80% of patients assigned to receive consultation ended up completing the consultation, our challenges in enrolling and assigning patients to the consultation arm reflected the limited capacity of geriatricians in our clinic. Relatedly, we found the time to have the initial visit with a geriatrician varied across patients, with a median of 36 days. Delays were largely a function of the patients’ busy schedules and the fact that our geriatrics clinics only occurred twice per week. Second, the geriatricians worked within the established referral structures existing at Dana-Farber rather than with a dedicated multidisciplinary team (e.g., including pharmacists, social workers, and allied health specialties), which may have limited the breadth and timeliness of any geriatricsrecommended interventions.34,35 Lastly, the utilization of our geriatric consultation service evolved over the study period, 1176

with more clinicians requesting longitudinal co-management, rather than a single consultation, later in the study. We implicitly intended more longitudinal management for all patients assigned to the geriatrics intervention but found that just over half received additional follow-up visits. The effectiveness of longitudinal geriatric co-management models delivered earlier in follow-up, with or without multidisciplinary support, warrants further investigation in frail older adults with blood cancers. Preliminary findings from such a geriatrician-led model in older adults with mostly solid tumors are encouraging, showing - in contrast to our study - a significant reduction in emergency presentations and unplanned hospitalizations in comparison to usual care.36 Important distinctions between this model and ours are worth noting. Patients’ initial visit with the geriatrician occurred upon enrollment in the study, which ensured earlier delivery of any GA-driven interventions. Additionally, although the intervention design, like ours, allowed for individualized management tailored by the geriatrician, standardized interventions were provided to all patients assigned to the geriatrics arm that included supportive care information and optimization of physical activity and nutrition. Lastly, more longitudinal co-management occurred than in our study’s consultative model, with patients receiving reassessments at multiple points in their follow-up period. Although we did not find an association between number of visits and mortality, our study was underpowered to formally analyze this association. Moreover, the geriatrician may have elected to see sicker patients more often, confounding the association. Earlier delivery, more integration, and more longitudinal follow-up Table 1. Baseline characteristics of the study population.

Characteristic

All (n=160)

Age, mean (SD) 80.4 (4.2) Male, n. (%) 104 (65.0) Disease type, n. (%) Lymphoid 50 (31.3) Myeloid 48 (30.0) Myeloma 62 (38.8) Aggressive disease, n. (%) 60 (37.5) Frailty, n. (%) Pre-frail 124 (77.5) Frail 36 (22.5) Gait speed, n. < 0.8 m/s (%) 97 (60.6) Declined/missing 4 (2.5) Cognition, n. with impairmenta (%) Delayed recall 25 (16.0) Declined/missing 4 (2.5) Executive function 60 (39.5) Declined/missing 8 (5.0) Function, n. with impairmentb (%) ADL 27 (16.9) IADL 57 (35.6)

Standard oncologic care (n=100)

Geriatric consultation + standard care (n=60)

80.3 (3.9) 64 (64.0)

80.5 (4.7) 40 (66.7)

36 (36.0) 28 (28.0) 36 (36.0) 37 (37.0)

14 (23.3) 20 (33.3) 26 (43.3) 23 (38.3)

75 (75.0) 25 (25.0) 60 (60.0) 3 (3.0)

49 (81.7) 11 (18.3) 37 (61.7) 1 (1.7)

16 (16.7) 4 (4.0) 38 (40.9) 7 (7.0)

9 (15.0) 0 (0.0) 22 (37.3) 1 (1.7)

18 (18.0) 34 (34.0)

9 (15.0) 23 (38.3)

Delayed recall was assessed using the five-word delayed recall component of the Montreal Cognitive Assessment, with probable impairment defined as the ability to recall two or fewer words after 5 minutes.5 Executive function was assessed using the Clock-in-the-Box test, with probable impairment defined as scoring five or less. bImpairment of basic activities of daily living (ADL) and instrumental activities of daily living (IADL) was defined as patients reporting requiring assistance or being dependent on others to complete one or more of six ADL or five IADL, respectively. SD: standard deviation; ADL: basic activities of daily living; IADL: instrumental activities of daily living. a

haematologica | 2022; 107(5)

RCT of geriatric consultation in blood cancers

for all patients could have contributed to a more effective geriatrics-led intervention. Reducing mortality is not necessary to justify the integration of geriatrics into the care of patients with blood cancers. Emerging evidence from other trials of predominantly patients with solid tumors suggests that GA-driven interventions improve meaningful outcomes other than survival

in older patients with cancer, including decreased treatment toxicity and improvements across multiple domains of quality of life.16,17,27,28,36,37 Various models of GA-driven interventions were studied in these trials, ranging from a GA summary with recommended interventions carried out by the treating hematologic oncologist, to an embedded comanagement model led by a geriatrician (as described

Figure 2. One-year overall survival by standard oncologic care (control) versus geriatric consultation + standard care. (A) Intent-to-treat analysis. (B) Per-protocol analysis.

haematologica | 2022; 107(5)

1177

C. DuMontier et al.

Table 2. Multivariable analyses assessing effect of geriatric consultation on overall mortality rate through 1 year of follow-up, acute care utilization, and goals-of-care discussions.

Overall mortality rate through 1 year Intent-to-treat vs. control Per-protocol vs control

ED visits Intent-to-treat vs. control Per-protocol vs. control

Hospitalizations Intent-to-treat vs. control Per-protocol vs. control

Days in hospital Intent-to-treat vs. control Per-protocol vs. control

EOL GOC discussions Intent-to-treat vs. control Per-protocol vs. control

Hazard ratio (95% CI)

P value

0.93 (0.45 - 1.95) 0.70 (0.30 - 1.66)

0.85 0.42

Incidence rate ratio (95% CI)

P value

0.89 (0.33 - 2.42) 0.77 (0.26 - 2.23)

0.82 0.62

Incidence rate ratio (95% CI)

P value

0.91 (0.30 - 2.71) 0.74 (0.24 - 2.32)

0.86 0.61

Incidence rate ratio (95% CI)

P value

1.05 (0.29 - 3.79) 0.82 (0.20 - 3.38)

0.94 0.78

Odds ratio (95% CI)

P value

3.12 (1.03 - 9.41) 3.58 (1.13 - 11.35)

0.05 0.03

All models for the multivariable analyses were adjusted for age, sex, disease aggressiveness, and frailty. Separate models were run for per-protocol analysis. 95% CI: 95% confidence interval; ED: emergency department; EOL GOC: end-of-life goals of care.

above). Future trials in older adults with blood cancers should investigate not only the effectiveness of different models of geriatrics-driven interventions in terms of mortality, but also their impact on treatment decisions and patient-centered outcomes such as function and quality of life, which are outcomes often valued by older patients just as much as, if not more than, survival.38-41 Reducing toxicity and optimizing function and quality of life - all while maintaining similar survival in comparison with standard oncologic care - constitute a net benefit for complex older patients.42 To this end, our finding that geriatric consultation increased the likelihood of having documented EOL goalsof-care discussions is clinically relevant to pre-frail and frail older adults with blood cancers, many of whom have a high risk of death regardless of intervention.4,43 Discussing preferences regarding place of death and resuscitation status is of paramount importance in frail older patients with blood cancers; doing so early in the outpatient setting can reduce intensive care use in the days before death while increasing hospice enrollment.44 Moreover, a geriatrician’s evaluation of age-related vulnerabilities (e.g., functional and cognitive impairment) and their potential reversibility better informs these goals-of-care discussions.42 Many frail older patients may have other advanced conditions that limit their prognosis independently of their cancer or its treatment, diminishing the benefits and increasing the harms of intensive chemotherapy. Indeed, our trial population had high rates of cognitive, functional, and mobility impairment, more representative of patients aged ≥75 years treated in practice than the small number of patients in this age group enrolled in clinical trials.4,45 Beyond aligning EOL care with patients’ preferences, the geriatricians’ expertise in evaluation and management of age-related vulnerabilities was highly valued by surveyed hematologic oncologists and other clinicians at DanaFarber. Most rated geriatric consultation to be useful in the 1178

Table 3. Survey results of oncologists’ opinions regarding value of geriatric consultation.a

Domains of care Evaluating cognition Connecting patients to resources Diagnosing frailty Managing non-oncologic comorbidities Tailoring end-of-life care Informing treatment decisions

Management of age-related issues Functional status Falls Depression Mood disorders Insomnia Nutrition Pain

Number of responses

% who answered 4 or 5 (95% CI)

35 35 35 35 35 35

85.7 (69.7 - 95.2) 80.0 (63.1 - 91.6) 77.1 (59.9 - 89.6) 77.1 (59.9 - 89.6) 71.4 (53.7 - 85.4) 62.9 (44.9 - 78.5)

Number of responses

% who answered 4 or 5 (95% CI)

35 35 35 34 35 35 35

88.2 (72.6 - 96.7) 85.7 (69.7 - 95.2) 80.0 (63.1 - 91.6) 79.4 (62.1 - 91.3) 77.1 (59.9 - 89.6) 62.9 (44.9 - 78.5) 62.9 (44.9 - 78.5)

a For each question, responses were rated on a Likert scale ranging from 0 = not at all useful to 5 = very useful. CI: confidence interval.

evaluation of cognition, management of non-oncologic comorbidities, and management of functional status and falls. Fewer clinicians found geriatric consultation to be useful in informing oncologic treatment decisions and the management of nutrition and pain. The latter might in part be due to the comfort of hematologic oncology teams in treating these problems themselves, with support from nutritionists and other allied healthcare services. Our study has limitations other than those related to the geriatric consultation model listed above. Our study took place at a large, academic, tertiary care center that may limit generalizability of its findings to community practices. However, GA-driven interventions have been shown to be feasible and improve outcomes in other settings, including community hematologic oncology clinics.14,28 Competing risk of mortality may have hindered observation of hospitalizations and other secondary outcomes; three patients in the consultation arm died before they could even receive the intervention. Although we did not detect a difference in care utilization between study arms, the study may have been underpowered to investigate these secondary outcomes. Indeed, our overall event rates for deaths and care utilization were low, likely because many of our patients were on observation for less aggressive disease. Future trials could further minimize heterogeneity in patients’ characteristics by limiting enrollment to patients with one or two types of blood cancer on active treatment. Along with investigating patient-centered outcomes, future trials should also investigate the impact of GA-guided care on treatment toxicity, treatment discontinuation, and progression free survival. In conclusion, our randomized trial of embedded geriatric consultation for pre-frail and frail older patients with blood cancers did not show an improvement in survival or healthcare utilization, but did increase EOL goals of care discussions and was valued by hematologic oncology clinicians. Lessons learned from our trial complemented by the results emerging from others suggest that ensuring earlier delivery and more longitudinal co-management may be necessary to have an impact on outcomes such as survival and hospitalizations. Such models should be investigated in older adults with blood cancers, along with their impact on patient-centered outcomes such as function and quality of life. Future haematologica | 2022; 107(5)

RCT of geriatric consultation in blood cancers

studies should determine not only which models may be effective, but also whether certain components of GA-driven interventions are more effective than others. The benefits of different models must be balanced against their scalability, especially considering the current limitations in oncology practices’ access to geriatricians.46,47 Such information will help oncology practices with varying resources adapt models of geriatric care that are feasible, effective, and sustainable in improving the care of older patients with blood cancers. Disclosures RJS serves on the Board of Directors for Kiadis and Be The Match/National Marrow Donor Program; has provided consulting for Gilead, Rheos Therapeutics, VOR Biopharma, and Takeda; and served on a Data Safety Monitoring Board for Juno/Celgene. Contributions GAA, JAD, RMS, and RJS designed the trial and oversaw execution of the protocol, contributed to data analysis and interpretation, and to preparing the manuscript; CD contributed to data acquisition, analysis, and interpretation, and wrote manuscript; HU contributed to the trial design, data analysis and interpretation, and to preparing the manuscript; GZ analyzed data and contributed to interpretation and manuscript preparation; TH oversaw execution of the protocol, contributed to data acquisition and interpretation, and to preparing the manuscript;

References 1. Surveillance, Epidemiology, and End Results (SEER) (2013-2017). Cancer Stat Facts: Myeloma. National Cancer Institute. Accessed June, 2020. https://seer.cancer.gov/statfacts/html/mul my.html 2. Surveillance, Epidemiology, and End Results (SEER) (2013-2017). Cancer Stat Facts: Leukemia. National Cancer Institute. Accessed June, 2020. https://seer.cancer.gov/statfacts/html/mul my.html 3. Surveillance, Epidemiology, and End Results (SEER) (2013-2017). Cancer Stat Facts: Non-Hodgkin Lymphoma. National Cancer Institute. Accessed June, 2020. https://seer.cancer.gov/statfacts/html/mul my.html 4. Abel GA, Klepin HD. Frailty and the management of hematologic malignancies. Blood. 2018;131(5):515-524. 5. Hshieh TT, Jung WF, Grande LJ, et al. Prevalence of cognitive impairment and association with survival among older patients with hematologic cancers. JAMA Oncol. 2018;4(5):686-693. 6. DuMontier C, Liu MA, Murillo A, et al. Function, survival, and care utilization among older adults with hematologic malignancies. J Am Geriatr Soc. 2019;67 (5):889-897. 7. Liu MA, DuMontier C, Murillo A, et al. Gait speed, grip strength, and clinical outcomes in older patients with hematologic malignancies. Blood. 2019;134(4):374-382. 8. Mohile SG, Dale W, Somerfield MR, et al. Practical assessment and management of vulnerabilities in older patients receiving chemotherapy: ASCO guideline for geriatric oncology. J Clin Oncol. 2018;36(22):2326-2347. 9. Repetto L, Fratino L, Audisio RA, et al.

haematologica | 2022; 107(5)

RC and EM contributed to data acquisition and interpretation, and to preparing the manuscript; LM and HJ oversaw execution of the protocol, and contributed to data interpretation and preparing the manuscript. Funding This work was supported by the Harvard Translational Research in Aging Training Program (National Institute on Aging of the National Institutes of Health: T32AG023480) (CD); the Dana-Farber/Harvard Cancer Center SPORE in Multiple Myeloma (National Cancer Institute of the National Institutes of Health: P50 CA100707) (CD); the Harvard Catalyst (GZ); the Harvard Clinical and Translational Science Center (National Center for Advancing Translational Sciences, National Institutes of Health Award UL 1TR002541) and financial contributions from Harvard University and its affiliated academic healthcare centers. The content is solely the responsibility of the authors and does not necessarily represent the official views of Harvard Catalyst, Harvard University and its affiliated academic healthcare centers, or the National Institutes of Health; the Older Adult Hematologic Malignancy Program is supported by the Murphy Family Fund from the Dana-Farber Cancer Institute (GAA) Data-sharing statement Data and protocol requests will be considered on a case by case basis and in accordance with the regulations of the Dana-Farber Harvard Cancer Center Office for Human Research Studies.

Comprehensive geriatric assessment adds information to Eastern Cooperative Oncology Group performance status in elderly cancer patients: an Italian Group for Geriatric Oncology Study. J Clin Oncol. 2002;20(2):494-502. 10. Hurria A, Mohile S, Gajra A, et al. Validation of a prediction tool for chemotherapy toxicity in older adults with cancer. J Clin Oncol. 2016;34(20):23662371. 11. Extermann M, Boler I, Reich RR, et al. Predicting the risk of chemotherapy toxicity in older patients: the Chemotherapy Risk Assessment Scale for High-Age Patients (CRASH) score. Cancer. 2012;118(13):3377-3386. 12. Palumbo A, Bringhen S, Mateos MV, et al. Geriatric assessment predicts survival and toxicities in elderly myeloma patients: an International Myeloma Working Group report. Blood. 2015;125(13):2068-2074. 13. Mohile SG, Velarde C, Hurria A, et al. Geriatric assessment-guided care processes for older adults: a Delphi consensus of geriatric oncology experts. J Natl Compr Canc Netw. 2015;13(9):1120-1130. 14. Mohile SG, Magnuson A, Pandya C, et al. Community oncologists' decision-making for treatment of older patients with cancer. J Natl Compr Canc Netw. 2018;16(3):301309. 15. Mohile SG, Epstein RM, Hurria A, et al. Communication with older patients with cancer using geriatric assessment: a clusterrandomized clinical trial from the National Cancer Institute community oncology research program. JAMA Oncol. 2019;6(2): 196-204. 16. Corre R, Greillier L, Caër HL, et al. Use of a comprehensive geriatric assessment for the management of elderly patients with advanced non-small-cell lung cancer: the phase III randomized ESOGIA-GFPC-

GECP 08-02 study. J Clin Oncol. 2016;34 (13):1476-1483. 17. Nipp RD, Temel B, Fuh C-X, et al. Pilot randomized trial of a transdisciplinary geriatric and palliative care intervention for older adults with cancer. J Natl Compr Canc Netw. 2020;18(5):591. 18. Fried LP, Tangen CM, Walston J, et al. Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci. 2001;56(3):M146-156. 19. Rockwood K, Mitnitski A. Frailty in relation to the accumulation of deficits. J Gerontol A Biol Sci Med Sci. 2007;62(7): 722-727. 20. Temel JS, Greer JA, Muzikansky A, et al. Early palliative care for patients with metastatic non-small-cell lung cancer. N Engl J Med. 2010;363(8):733-742. 21. Abel GA, Neufeld EJ, Sorel M, Weeks JC. Direct-to-consumer advertising for bleeding disorders: a content analysis and expert evaluation of advertising claims. J Thromb Haemost. 2008;6(10):1680-1684. 22. El-Jawahri AR, Abel GA, Steensma DP, et al. Health care utilization and end-of-life care for older patients with acute myeloid leukemia. Cancer. 2015;121(16):2840-2848. 23. Mack JW, Cronin A, Taback N, et al. Endof-life care discussions among patients with advanced cancer: a cohort study. Ann Int Med. 2012;156(3):204-210. 24. Uno H, Claggett B, Tian L, et al. Moving beyond the hazard ratio in quantifying the between-group difference in survival analysis. J Clin Oncol. 2014;32(22):2380-2385. 25. surv2sampleComp: Inference for ModelFree Between-Group Parameters for Censored Survival Data. R package version 1.0-5. 2017. https://CRAN.Rproject.org/package=surv2sampleComp 26. Uno H, Cai T, Tian L, Wei LJ. Evaluating prediction rules for t-year survivors with censored regression models. J Am Stat Ass.

1179

C. DuMontier et al. 2007;102(478):527-537. 27. Li D, Sun C-L, Kim H, et al. Geriatric assessment-driven intervention (GAIN) on chemotherapy toxicity in older adults with cancer: A randomized controlled trial. J Clin Oncol. 2020;38(15_suppl):12010. 28. Mohile SG, Mohamed MR, Culakova E, et al. A geriatric assessment (GA) intervention to reduce treatment toxicity in older patients with advanced cancer: a University of Rochester Cancer Center NCI community oncology research program cluster randomized clinical trial (CRCT). J Clin Oncol. 2020;38(15_suppl):12009. 29. Soo WK, King M, Pope A, et al. The Elderly Functional Index (ELFI), a patient-reported outcome measure of functional status in patients with cancer: a multicentre, prospective validation study. Lancet Healthy Longev. 2021;2(1):e24-e33. 30. Shahrokni A, Tin AL, Sarraf S, et al. Association of geriatric comanagement and 90-day postoperative mortality among patients aged 75 years and older with cancer. JAMA Netw Open. 2020;3(8):e209265. 31. Lund CM, Vistisen KK, Olsen AP, et al. The effect of geriatric intervention in frail older patients receiving chemotherapy for colorectal cancer: a randomised trial (GERICO). Br J Cancer. 2021;124(12):1949-1958. 32. Derman BA, Kordas K, Ridgeway J, et al. Results from a multidisciplinary clinic guided by geriatric assessment before stem cell transplantation in older adults. Blood Adv. 2019;3(22):3488-3498. 33. Clarfield AM, Bergman H, Kane R.

1180

Fragmentation of care for frail older people-an international problem. Experience from three countries: Israel, Canada, and the United States. J Am Geriatr Soc. 2001;49 (12):1714-1721. 34. Dale W, Chow S, Sajid S. Socioeconomic considerations and shared-care models of cancer care for older adults. Clin Geriatr Med. 2016;32(1):35-44. 35. Presley CJ, Krok-Schoen JL, Wall SA, et al. Implementing a multidisciplinary approach for older adults with cancer: geriatric oncology in practice. BMC Geriatr. 2020;20(1):231. 36. Soo W-K, King M, Pope A, Parente P, Darzins P, Davis ID. Integrated geriatric assessment and treatment (INTEGERATE) in older people with cancer planned for systemic anticancer therapy. J Clin Oncol. 2020;38(15_suppl):12011. 37. Qian CL, Knight HP, Ferrone CR, et al. Randomized trial of a perioperative geriatric intervention for older adults with cancer. J Clin Oncol. 2020;38(15_suppl):12012. 38. Fried TR, Bradley EH, Towle VR, Allore H. Understanding the treatment preferences of seriously ill patients. N Engl J Med. 2002;346(14):1061-1066. 39. Loh KP, Mohile SG, Epstein RM, et al. Willingness to bear adversity and beliefs about the curability of advanced cancer in older adults. Cancer. 2019;125(14):25062513. 40. Celis ESPD, Li D, Sun C-L, et al. Patientdefined goals and preferences among older adults with cancer starting chemotherapy

(CT). J Clin Oncol. 2018;36(15_suppl): 10009. 41. Rosko AE, Wall S, Baiocchi R, et al. Aging phenotypes and restoring functional deficits in older adults with hematologic malignancy. J Natl Compr Canc Netw. 2021;19(9):1027-1036. 42. DuMontier C, Loh KP, Bain PA, et al. Defining undertreatment and overtreatment in older adults with cancer: a scoping literature review. J CLin Oncol. 2020;38(22):2558-2569. 43. Krok-Schoen JL, Fisher JL, Stephens JA, et al. Incidence and survival of hematological cancers among adults ages >/=75 years. Cancer Med. 2018;7(7):3425-3433. 44. Odejide OO, Uno H, Murillo A, Tulsky JA, Abel GA. Goals of care discussions for patients with blood cancers: association of person, place, and time with end-of-life care utilization. Cancer. 2020;126(3):515-522. 45. Ludmir EB, Mainwaring W, Lin TA, et al. Factors associated with age disparities among cancer clinical trial participants. JAMA Oncol. 2019;5(12):1769-1773. 46. Williams GR, Weaver KE, Lesser GJ, et al. Capacity to provide geriatric specialty Care for Older Adults in Community Oncology Practices. Oncologist. 2020;25(12):10321038. 47. Dale W, Williams GR, MacKenzie AR, et al. How is geriatric assessment used in clinical practice for older adults with cancer? A survey of cancer providers by the American Society of Clinical Oncology. JCO Oncol Pract. 2021;17(6):336-344.

haematologica | 2022; 107(5)

LETTERS TO THE EDITOR BNT162b2 COVID-19 and ChAdOx1 nCoV-19 vaccination in patients with myelodysplastic syndromes Many patients with hematological cancers are not completely protected after the initial dose or after both primary doses of the vaccines1,2 with most failing to seroconvert on completion of the two-dose vaccine schedule.2 These reports only included three patients with myelodysplastic syndrome (MDS). MDS represents a spectrum of clonal bone marrow neoplasms from lowrisk disease through to those transforming into acute myeloid leukemia. Patients with MDS, especially with lower-risk disease, many of whom are minimally treated and who might be expected to have a comparable immune response to healthy volunteers, and as such a better immune response to COVID-19 vaccines than other hematological cancers. Previous studies looking at the immune response to influenza vaccination in those with MDS had shown promising results with immune responses not differing from those of healthy family members.3 However, a recent study which included six MDS patients, reported poor seroconversion rates following a single dose of COVID-19 vaccine in a group of 60 myeloid cancer patients, including those who are not on cytoreductive treatments and those in complete hematological remission, suggesting a clear need for more detailed interrogation of COVID-19 vaccination in this group of patients.4 Here, we report the humoral and T-cell responses of 38 patients with MDS 2 weeks following completion of the second dose vaccine schedules of ChAdOx1 or BNT162b2 nCoV-19 vaccines. Following approval by the Institutional Review Boards, patients with MDS (n=38) vaccinated with either BNT162b2 mRNA or ChAdOx1 nCoV-19 COVID-19 vaccine provided written informed consent. Eligibility criteria for the study included diagnosis of MDS as per the World Health Organization classification5 and age ≥18 years. The study also included healthy volunteers (HV) (mainly healthcare workers, n=30) serving as a reference group, included principally to provide an experimental control for study assays and facilitate their comparison with results of other studies of BNT162b2 in healthy populations. Plasma samples were tested for immunoglobulin G (IgG) binding the SARS-CoV-2 spike (S) protein and nucleoprotein (N) and neutralization assays against HIV-1 based virus particles pseudotyped with SARS-CoV-2 Wuhan strain (WT), variant of concern (VOC)B.1.1.7 (a) or VOC.B.1.617.2 (δ) spike as previously described.1,2,6 Cellular responses were assessed using interferon g (IFNg) ELISPOT and flow cytometry (CD25 and CD69 expression) after 24 hours of peptide stimulation. IFNg ELISpot analysis was performed ex vivo for assessment of T-cell response following stimulation with SARS-CoV-2 peptide pools, Cytomegalovirus (CMV), Epstein-Barr virus (EBV), and influenza virus positive control (CEF) peptides for 24 hours. Thirty-eight MDS patients and 30 HV provided a blood sample 2 weeks following a second primary dose of their initial vaccine. Clinical characteristics along with median times to second dose are provided in Table 1. We observed significant differences between the ages of the HV and MDS cohorts (Student’s t-test, equal variance, P<0.001). 42% (n=16) of the MDS patients received BNT162b2 and 58% (n=22) received ChAdOx1 nCoV-19 vaccines. All HV received a delayed BNT162b2 second dose. As per UK government guidelines at the time of vaccination, individuals receiving BNT162b2 second doses received these between 8-12 weeks following the first

haematologica | 2022; 107(5)

dose, representing a delay compared to the licensed administration. Prior SARS-CoV-2 infection can influence the magnitude of the vaccine response,7 and as such we excluded two MDS and four HV based on being positive for nucleoprotein-specific IgG (IgG(N)) (representing response to prior infection) (Online Supplementary Figure S1A). We observed that the anti-S IgG titres at approximately 2 weeks following the second dose were within the upper quantile in these previously virus-exposed individuals (Online Supplementary Figure S1B, red dots). These were excluded from the overall immune efficacy analysis. In the remaining (HV BNT162b2, n=26; MDS BNT162b2, n=15 and MDS ChAdOx1, n=21) cohort; we assessed the anti-S IgG titres following their second primary dose. Overall serological responses were: HV BNT162b2 100% (26/26); MDS BNT162b2 100% (15/15) and MDS ChAdOx1 76.2% (16/21) (Figure 1A); notably, the MDS ChAdOx1 cohort demonstrated significantly decreased serological titres to the MDS BNT162b2 cohort (Figure 1A). It is noteworthy that the median titre for the MDS BNT162b2-vaccinated patients is higher (>103) compared to the median reported in a heterogenous BNT162b2-vaccinated hematological cancer population (<103) observed in McKenzie et al.2 Of the five nonresponders within the MDS ChAdOx1, three patients were on disease-modifying treatments (5-azacytidine, venetoclax and danazol), with the patient on venetoclax/rituximab having a concurrent diagnosis of chronic lymphocytic leukemia (CLL). None of these patients were noted to be on steroid therapy around the time of vaccination; and no differences in the clinical white blood cells were observed between serological responders or non-responders (Online Supplementary Figure 1C). Similar to our previous reports1,2 there was no significant correlation between spike IgG titres and age or the time between the first and second doses of the vaccine in the two MDS cohorts (Online Supplementary Figure S1D). Next, we assessed the functional implications of seroconversion by neutralization assays for SARS-CoV-2 WT and VOC a and δ (Figure 1B). All but four MDS patients (Figure 1B; colored dots) could neutralize all variant strains, but MDS cohorts showed significantly reduced median neutralizations for all three variant strains compared to HV (Figure 1B); importantly this was the case for both the MDS ChAdOx1 and MDS BNT162b2 cohorts. We acknowledge the younger age of the HV cohort may contribute to this reduction, although age was not a determinant of neutralization response in cancer patients in our previous reports.1,2 Review of the four MDS (2 BNT162b2 mRNA and two ChAdOx1 nCoV-19 COVID19-vaccinated) patients classified as non-responders by neutralization assay demonstrated that these patients were predominantly low risk MDS on no treatment, except one patient with excess of blasts on 5-azacytidine. These data clearly support the need for a third primary dose for this clinically vulnerable patient group irrespective of the seroconversion rates across cohorts. This is especially the case in those who have seroconverted but have a low anti-S IgG titre after the second dose. Third doses have demonstrated higher anti-S IgG titres in other hematological cohorts,8 and in keeping with our previous reports,1,2 anti-S IgG titres were highly correlated with neutralization among all cohorts (Figure 1C). In order to measure functional SARS-CoV-2 T-cell responses to vaccination, peripheral blood mononuclear cells (PBMC) from our study participants were assessed by ELISpot assays as described. It is noteworthy that no differences in the percentages of T cells amongst the 1181

Letters to the Editor

Table 1. Clinical characteristics of patients evaluable for analysis 2 weeks following two primary vaccine doses.

All MDS patients Total numbers Age Median (Q1-Q3) years Sex Male Female Race Caucasian BAME Median time from vaccine first dose to second dose Median (Q1-Q3) days Median time from vaccine second dose to blood sampling Median (Q1-Q3) days MDS WHO subtypes MDS with single lineage dysplasia MDS with ring sideroblasts MDS with isolated del5q MDS with multilineage dysplasia MDS with multilineage dysplasia (hypo) MDS with excess blasts Chronic myelomonocytic leukemia IPSS-R prognostic categories Low risk (low/very low/intermediate) High risk (high/very high) Treatment 15 days pre- and post-vaccination Transfusion support only or watch &wait Growth factors/ TPO mimetics Cyclosporin 5-Azacytidine Others*

BNT162b2 vaccinated MDS patients 16

ChAdOx1 vaccinated MDS 22

67 (63-72)

BNT162b2 vaccinated healthy volunteers 30

67.5 (59-73)

69 (60-73)

35 (27-49)

23/38 (61%) 15/38 (39%)

13 3

10 12

19 11

36/38 (95%) 2/38 (5%)

16 0

20 2

19 11

75 (68-80)

71 (68-77)

78 (70-80)

74 (61-78)

19 (16-28)

21 (18-30)

18 (15-24)

14 (13-17)

2/38 (5.2%) 5/38 (13.2%) 1/38 (2.6%) 20/38 (52.6%) 4/38 (10.5%) 5/38 (13.2%) 1/38 (2.6%)

0 1 0 9 2 3 1

2 4 1 11 2 2 0

30/38 (78.9%) 8/38 (21.1%)

11 5

19 3

22/38 (57.9%) 6/38 (15.8%) 3/38 (7.9%) 5/38 (13.2%) 1/38 (5.2%)

7/16 3/16 2/16 3/16 1/16

15/22 3/22 1/22 2/22 0/22

*This patient had concurrent chronic lymphocytic leukemia which was the indication for therapy with venetoclax and rituximab. MDS: myelodysplastic syndrome; WHO: World Health Organization; BAME: Black, Asian and minority ethnic; Q: quarter; TPO: thrombopoietin; IPSS-R: Revised International Prognostic Scoring System.

Figure 1. Humoral responses to BNT162b2 COVID-19 and ChAdOx1 nCoV-19 in patients with myelodysplastic syndromes. (A) Serum concentrations of immunoglobulin G (IgG) antibodies reactive to the spike protein of SARS-CoV-2 (S IgG) with cases positive for nucleorprotein N IgG removed. Healthy volunteer (HV; n=26), myelodysplastic syndrome (MDS) patients vaccinated with ChAdOx1 (MDS ChAdOx1; n=20), MDS patients vaccinated with BNT162b2 (MDS BNT162b2; n=15). Mean (95% confidence interval [CI]): healthy volunteers (HV) 3,611 (2,455-4,768), MDS ChAdOx1 360.9 (149.9-572.2) and MDS BNT162b2 3781 (523.9-7,037). Dashed line represents seroconversion threshold. Tukey’s multiple comparison’s test. (B) Neutralization of variants (as indicated in red) by plasma antibodies. Dashed line represents neutralization threshold. Individual cases on the threshold line are colored as indicated, as are their matched responses to other variants. HV (n=26); MDS ChAdOx1 (n=15); MDS BNT162b2 (n=15). Tukey’s multiple comparison’s test. (C) Correlation matrices showing serum S IgG 50% effective dose (ED50) (log) against neutralization for each indicated variant in the MDS ChAdOx1 (n=20) and MDS BNT162b2 (n=15) cohorts. Correlation coefficients (rho;r) and P-values are given. Dashed lines represent threshold as previously described. Pearson’s correlation test. WT: Wuhan strain.

1182

haematologica | 2022; 107(5)

Letters to the Editor

PBMC plated for ELISpot were observed across healthy and MDS cohorts (Online Supplementary Figure S1E). Using previously published thresholds for response,1,2 non-T cell responders were seen in all cohorts (Figure 2A; red dots). Specifically, SARS-CoV-2-specific IFNg T-cell responses against the δ variant were: HV BNT162b2 95% (20/21); MDS ChAdOx1 70.6% (12/17) and MDS BNT162b2 71.4% (10/14) (Figure 2A); in stark contrast to the comparable control CEF induced effector T-cell responses across healthy and MDS samples (Figure 2A). Interestingly, significantly reduced T-cell responses were seen in MDS BNT162b2-vaccinated patients when challenged with δ compared to wt variant strain (Figure 2B). Further, five MDS ChAdOx1 patients who did not have a serological response, were able to mount T-cell responses. Additionally, treatment with either azacytidine or calcineurin inhibitor cyclosporine did not impair appropriate T-cell responses. One high risk MDS BNT162b2 patient

on 5-azacytidine, who showed no neutralizing activity, showed significantly reduced T-cell response to WT and a, but not to δ variant. During the study period, the δ variant was the predominant VOC in the UK. We observed non-significant but positive correlations between serological and IFNg T-cell responses against the δ variant within the MDS vaccinated cohorts (Figure 2C). Numbers of individuals who were both serological and Tcell responders were as follows: HV 95% (20/21), MDS BNT162b2 71.4% (10/14) and MDS ChAdOx1 52.9% (9/17) (Figure 2C). In order to further investigate the cellular readout of vaccine efficacy, we assessed the activation state of SARS-CoV-2 stimulated CD8 T cells, by measuring activation markers CD25 and CD69 cell surface expression by flow cytometry before and after in vitro stimulation. Despite the poorer humoral response observed in MDS-ChAdOx1 vaccinated individuals, we found significantly higher activated CD25+ and CD69+

Figure 2. Cellular responses to BNT162b2 COVID-19 and ChAdOx1 nCoV-19 in patients with myelodysplastic syndromes. (A) Interferon g (IFNg) spot-forming units (SFU) formed after stimulation of peripheral blood mononuclear cells (PBMC) from indicated cohorts in response to indicated variants. Samples were classed as responders if >7 cytokine secreting cells/106 PBMC after correcting for background; as indicated by dashed line. Non-responders are colored as indicated. Wuhan strain (WT); (healthy volunteers [HV] [n=26]; MDS ChAdOx1 [n=20]; MDS BNT162b2 [n=15]); B.1.1.7; (HV [n=11]; MDS ChAdOx1 [n=11]; MDS BNT162b2 [n=15]); B.1.617.2; (HV [n=21]; MDS ChAdOx1 [n=17]; myelodysplastic syndrome (MDS) BNT162b2 [n=14]). Tukey’s multiple comparison’s test. Influenza virus positive control (CEF)= Cytomegalovirus (CMV), Epstein-Barr virus (EBV) and influenza virus positive control peptides: (B) IFNγ SFU formed after stimulation of PBMC from MDS BNT162b2 cases to indicated variants. WT (n=15); B.1.1.7 (n=11); B.1.617.2 (n=14). Tukey’s multiple comparison’s test. (C) Correlation matrices showing IFNg SFU formed after PBMC were stimulated with the B.1.617.2 variant and paired S IgG 50% effective dose (ED50) values for indicated cohorts. Correlation coefficients (rho;r), P-values, n numbers and % double positivity are given. Dashed lines represent thresholds as previously described. Pearson’s correlation test. E (i&ii). CD8+CD25+ cells (i) and CD8+CD69+ cells (ii) within the live CD3+ population after stimulation of PBMC from indicated cohorts in response to indicated variants. HV (n=26); MDS ChAdOx1 (n=20); MDS BNT162b2 (n=15). Tukey’s multiple comparison’s test.

haematologica | 2022; 107(5)

1183

Letters to the Editor

CD8 T cells across all variants in this group of patients compared to those vaccinated with BNT162b2 vaccine (Figure 2Di&ii). These data are compelling and warrant further investigation with one hypothesis being the ChAdOx1 vector reveals an innate weakness in this patient group inducing a hyper-stimulated but poorly efficacious effector T-cell response. In totality, although ChAdOx1-treated MDS patients do mount both humoral and cellular immune responses, they are weak in comparison to BNT162b2. The overall serological responses in the MDS cohorts were 100% for those who had completed the two-dose BNT162b2 vaccine schedule compared to 76.2% of patients vaccinated with the ChAdOx1 vaccine. As such, it may be pertinent to advise the clinical community to administer MDS patients with an mRNA-based vaccine to promote enhanced immunity. Finally, we observed that neutralization in seroconverted patients was significantly weaker for both the ChAdOx-1 and BNT162b2 MDS cohorts compared to HV, highlighting the potential benefit of a third primary dose for this clinically vulnerable patient group, in addition to subsequent booster doses. Sultan Abdul-Jawad,1* Richard Beatson,1* Thomas Lechmere,2* Rosalind Graham,1 Thanussuyah Alaguthurai,1,3 Carl Graham,2 Jennifer Vidler,4 Austin Kulasekararaj,4 Piers EM Patten,1,4 Katie J Doores2 and Sheeba Irshad1,3,5,6 1 Comprehensive Cancer Center, School of Cancer & Pharmaceutical Sciences, King's College London; 2Department of Infectious Diseases, School of Immunology & Microbial Sciences, King’s College London; 3 Breast Cancer Now Research Unit, King’s College London; 4 Department of Hematological Medicine, King’s College Hospital;5Guy’s and St Thomas’ NHS Foundation Trust; 6Cancer Research UK (CRUK) Clinician Scientist, London, UK *SAJ, RB and TL contributed equally as co-first authors Correspondence: SHEEBA IRSHAD - sheeba.irshad@kcl.ac.uk doi:10.3324/haematol.2021.280337 Received: November 15, 2021. Accepted: January 10, 2022. Pre-published: January 20, 2022. Disclosures: no conflicts of interest to disclose. Contributions: AK, PP, SI and KD developed the concept and study design; SAJ, RB, TA, RG and TL perfomed the investigation; TA, JV

1184

and AK recruited patients; TA, AK and JV searched the clinical database; SAJ and RB performed formal analysis; SAJ, RB and RG curated data; SAJ, RB, SI and KD visualized the research; SI wrote the original draft of the manuscript; RB, AK, PP, KD wrote, reviewed and edited the paper; SI, AK and PP acquired funding; SI and KD supervised the research. Acknowledgements: we thank patients and blood donors consenting in this. We thank members of the King’s College Hospital (KCH) trial teams who contributed to patient recruitment for the SOAP study at KCH hospitals. Funding: the SOAP study (IRAS 282337) is sponsored by King’s College London and GSTT Foundation NHS Trust. It is funded from grants from the KCL Charity funds to SI (PS10822), Cancer Research UK to S.I. (C56773/ A24869). This work was supported by Blood Cancer UK awarded to SI, AK and PP and the Leukaemia & Lymphoma Society (LLS) Award to SI and PP (6631-21). Data sharing statement: data from this study can be made available to other researchers in the field upon request and approval by the study management committee and subject to appropriate data transfer agreements. Requests should be directed to Dr Irshad.

References 1. Monin L, Laing AG, Munoz-Ruiz M, et al. Safety and immunogenicity of one versus two doses of the COVID-19 vaccine BNT162b2 for patients with cancer: interim analysis of a prospective observational study. Lancet Oncol. 2021;22(6):765-778. 2. McKenzie DR, Munoz-Ruiz M, Monin L, et al. Humoral and cellular immunity to delayed second dose of SARS-CoV-2 BNT162b2 mRNA vaccination in patients with cancer. Cancer Cell. 2021;39(11):1445-1447. 3. Vachhani P, Wiatrowski K, Srivastava P, et al. Quantification of humoral immune response to influenza vaccination in MDS. Blood. 2019;134(Suppl 1):S4756. 4. Chowdhury O, Bruguier H, Mallett G, et al. Impaired antibody response to COVID-19 vaccination in patients with chronic myeloid neoplasms. Br J Haematol. 2021;194(6):1010-1015. 5. Arber DA, Orazi A, Hasserjian R, et al. The 2016 revision to the World Health Organization classification of myeloid neoplasms and acute leukemia. Blood. 2016;127(20):2391-2405. 6. Dupont L, Snell LB, Graham C, et al. Neutralizing antibody activity in convalescent sera from infection in humans with SARS-CoV-2 and variants of concern. Nat Microbiol. 2021;6(11):1433-1442. 7. Anichini G, Terrosi C, Gandolfo C, et al. SARS-CoV-2 antibody response in persons with past natural infection. N Engl J Med. 2021;385(1):90-92. 8. Greenberger LM, Saltzman LA, Senefeld JW, Johnson PW, DeGennaro LJ, Nichols GL. Anti-spike antibody response to SARSCoV-2 booster vaccination in patients with B cell-derived hematologic malignancies. Cancer Cell. 2021;39(10):1297-1299.

haematologica | 2022; 107(5)

Letters to the Editor

A report from the Leukemia Electronic Abstraction of Records Network on risk of hepatotoxicity during pediatric acute lymphoblastic leukemia treatment The objective of this work was to identify determinants of treatment-associated hepatotoxicity (TAH) in a diverse population of 782 children with acute lymphoblastic leukemia (ALL). Based on extracted electronic medical record data, nearly all subjects experienced mildly elevated hepatic laboratory values (HL), particularly those given high-intensity treatment. Furthermore, 15.9% of subjects experienced TAH in at least one postinduction treatment phase, which was associated with increased body mass index, but did not affect relapse-free survival. While modern treatment for childhood ALL confers excellent survival,1 30-50% of children experience at least one serious adverse event during upfront ALL treatment.2 TAH may be related to a number of ALL therapeutics, e.g., asparaginase, antimetabolites, and anthracyclines. The reported incidence of TAH in pediatric ALL is highly variable, likely due to inconsistent defining criteria and data-capturing methods.3-10 To comprehensively characterize the impact of ALL therapy on HL and the treatment phase-specific incidence of TAH, we leveraged data from the Leukemia Electronic Abstraction of Records Network (LEARN). LEARN is a multi-institutional collaboration and childhood leukemia data repository that includes comprehensive demographic, anthropometric, diagnostic, treatment, laboratory, and outcome data. Given evidence for racial and ethnic disparities in childhood ALL outcomes and survival11 and the historic under-representation of children from minority groups in pediatric cancer trials,12 LEARN was constituted by institutions with highly diverse patient populations. LEARN relies on automated extraction of electronic medical record data after manual input of basic data, an ascertainment method which significantly improves reporting accuracy of laboratory adverse events.13 Here, we utilized LEARN data to assess HL changes by treatment phase and intensity, determining the incidence of TAH, its risk determinants, and its impact on patients’ outcomes. Our study utilized LEARN data from children (ages 121 years) diagnosed with ALL and treated at Texas Children’s Cancer and Hematology Centers (TXCH) or the Children’s Hospital of Philadelphia (CHOP) between 2006 and 2014. Children with infant ALL and Down syndrome were excluded, as were those who received part of their induction at another institution, did not complete induction, received non-standard agents or sequence of chemotherapy and/or a tyrosine kinase inhibitor, or underwent stem cell transplantation. Trained personnel manually populated a REDCap™ database with select data from the TXCH and CHOP electronic medical records, including date on diagnosis, dates of starting and ending chemotherapy courses, and risk stratification. Using the REDCap™ Application Programming Interface, we then auto-extracted demographic and laboratory data from each electronic medical record data warehouse. Manually-entered dates guided extraction by providing boundaries over which the data were extracted, enabling linkage of extracted data with a specific chemotherapy phase. Demographic data, disease characteristics, and HL were collected using a combination of targeted manual abstraction and extensive automated extraction from each institution’s electronic medical records. haematologica | 2022; 107(5)

HL included alanine aminotransaminase, aspartate aminotransferase, and total and conjugated bilirubin, normed to the age-based upper limit of normal (ULN). TAH was determined by the following criteria: (i) grade 4 transaminitis by the Common Toxicity Criteria of Adverse Events (CTCAE) v5.0, defined as alanine aminotransaminase or aspartate aminotransferase >20xULN; (ii) grade 3 hyperbilirubinemia by the CTCAE, defined as total bilirubin >3xULN, or (iii) conjugated bilirubin ≥1.2 mg/dL. TAH was defined based on established Children’s Oncology Group thresholds for dose modification considerations during ALL therapy. Each subject was categorized as having received high or standard-intensity treatment by phase, with high-intensity defined by inclusion of anthracycline (induction), cyclophosphamide (consolidation), and mercaptopurine (interim maintenance 1). Subjects were assigned final treatment intensity based on National Cancer Institute’s diagnostic criteria and interim maintenance 1 treatment assignment. Distributions of categorical characteristics and median age were compared by treatment intensity using c2 analyses and the Wilcoxon rank sum test, respectively. Median normed HL values were compared by treatment intensity for each treatment phase using the Wilcoxon rank sum test. Multivariable logistic regression models of factors influencing post-induction TAH and recurrent/persistent TAH (defined as TAH in 2 or more treatment phases) were performed. Cox regression models were used to calculate hazard ratios (HR) and 95% confidence intervals (95% CI) to compare overall and relapse-free survival in subjects with no TAH relative to those with any TAH, considered as a time-varying exposure introduced on the day of first documentation. All multivariable analyses were adjusted for treatment intensity, age at diagnosis, race/ethnicity, gender, body mass index, ALL immunophenotype, and end-induction minimal residual disease. Covariates were selected a priori, based on our hypotheses and clinical experience, and were included in all analyses. P-values <0.05 were considered statistically significant. Statistical analyses were performed using Stata 15.0 (StataCorp LP, College Station, TX, USA). Of 921 eligible patients, 782 met the inclusion criteria. Demographic, diagnostic, and disease characteristics of included subjects are shown in Table 1 by induction treatment intensity. Approximately one-third were Latino, 9% Black, 5% Asian, and the remainder were White. Subjects assigned to high-intensity induction were more likely to be overweight or obese (P<0.001), possibly reflecting older mean age (10.6 years vs. 4.6 years). Data on end-induction minimal residual disease were available for 681 of 782 subjects, of whom 22% (n=149) were positive for minimal residual disease, representing 17% (n=60) of subjects given standard-intensity treatment and 29% (n=89) of those given high-intensity treatment. A mean of 139.5 HL were obtained per subject. There were a greater number of HL for patients given highintensity treatment than for those given standard-intensity treatment (148.9 vs. 129.3, P<0.001). The number of subjects analyzed per phase varied over time and by treatment intensity: induction, n=782; consolidation, n=691; interim maintenance 1, n=643; delayed intensification, n=625; interim maintenance 2, n=235; and maintenance, n=571. Over 80% of subjects had a HL >ULN during at least one treatment phase, with the values being mostly 1-3xULN (Figure 1A-D). Alanine aminotransaminase was the most consistently and markedly elevated HL throughout all phases (Figure 1A). Total and conjugated bilirubin remained within normal limits for 1185

Letters to the Editor

Table 1. Demographics and disease characteristics of the study cohort.

Characteristic

All subjects (n=782)

Mean age (SD) Race/ethnicity Non-Latino-White Latino Non-Latino-Black Asian Other Gender Male Female BMI category* Not overweight or obese Overweight or obese ALL immunophenotype B-cell T-cell Minimal residual disease ^ Positive Negative

Assigned to standard intensity induction (n=419)

Assigned to high intensity induction (n=363)

P-value

10.6 (5.2)

<0.001

366 (46.8) 275 (35.2) 68 (8.7) 39 (5.0) 34 (4.3)

194 (46.3) 162 (38.7) 26 (6.2) 17 (4.1) 20 (4.8)

172 (47.4) 113 (31.1) 42 (11.6) 22 (6.1) 14 (3.9)

0.021

431 (55.1) 351 (44.9)

213 (50.8) 206 (49.2)

218 (60.1) 145 (39.9)

509 (70.2) 216 (29.8)

293 (76.5) 90 (23.5)

217 (63.4) 125 (36.6)

712 (91.1) 70 (8.9)

419 (100.0) 0 (0.0)

293 (80.7) 70 (19.3)

149 (21.9) 532 (78.1)

60 (16.4) 306 (83.6)

89 (28.3) 212 (71.7)

7.4 (4.9)

4.6 (2.3)

0.010

<0.001

*Subject numbers dependent on documented/abstracted height & weight from diagnosis. ^Subject numbers dependent on documented minimal residual disease at end of induction. Standard intensity: standard, three-drug induction versus high intensity: four-drug Induction. SD: standard deviation; BMI: body mass index; ALL: acute lymphoblastic leukemia.

Figure 1. Trends in hepatic laboratory values, including treatment-associated hepatotoxicity, during acute lymphoblastic leukemia therapy by treatment intensity. (A-D) The normed median hepatic laboratory value (HL) of subjects by each HL are represented by box and whisker plots, with outliers shown in the dots, subjects given standard intensity treatment in blue, and those given high intensity treatment in red. (A) Normed median alanine aminotransaminase (ALT, SGPT). (B) Normed median aspartate aminotransaminase (AST, SGOT). (C) Normed median total bilirubin (TBIL). (D) Normed median conjugated bilirubin (CBIL). Dashed lines indicate thresholds of CTCAE v5.0 grading for grade 3 or grade 4 ALT, AST, or TBIL as follows: ALT/AST: Grd 3= 5-20x upper limit of normal (ULN), Grd 4= >20x ULN. TBIL: Grd 3= 3-10x ULN, Grd 4= >10x ULN. (E) Percentage of patients with treatment-associated hepatotoxicity (TAH) by treatment intensity over all courses of therapy. *P<0.05, **P=0.001-<0.01, ***P<0.001, comparing standard vs. high intensity groups. For (E), comparisons were made between TAHALT/AST of each intensity group (hashed bars) and between TAH-TBIL/CBIL of each intensity group (open bars) for each treatment phase. Consol: consolidation; IM1: interim maintenance 1; DI: delayed intensification; IM2i: interim maintenance 2.

1186

haematologica | 2022; 107(5)

Letters to the Editor

nearly all subjects. Median HL were greater in the highintensity than standard-intensity groups across all phases, with the differences being statistically significant for all phases (P<0.01) except delayed intensification and maintenance. One hundred ten subjects (15.9%) experienced at least one episode of TAH after induction (Table 2). The majority of TAH events occurred during maintenance, and included both transaminitis and hyperbilirubinemia (Figure 1E). The presence of TAH was associated with being overweight/obese (odds ratio 1.7 [95% CI: 1.0-2.7], P=0.027). A minority of subjects (n=16, 2.3% overall) experienced recurrent/persistent TAH. Multivariable logistic regression did not identify patient characteristics associated with recurrent/persistent TAH in this small number of subjects (Table 2). Time-varying Cox regression analyses identified risk factors associated with overall and relapse-free survival. The median follow-up was 3.4 years for overall survival and 3.1 years for relapse-free survival. TAH was not associated with relapse-free survival in adjusted analysis (HR=0.7, 95% CI: 0.2-2.3, P=0.543), although the followup was relatively short. Older age and positive minimal residual disease were associated with poorer relapse-free survival, and non-Latino Black patients also experienced poorer relapse-free survival, consistent with prior reports.14 Because only three patients with TAH died, the study lacked power to assess the relationship between TAH and overall survival. Here, we report the landscape of HL and TAH by treatment phase in a large, diverse, contemporary cohort of children with uniformly treated ALL. We show that mild elevations of hepatic transaminase levels are common throughout ALL therapy, particularly with high-intensity treatment. TAH is a rare outcome that is most common during maintenance, when therapy includes continuous

antimetabolites. Our results provide reassurance that TAH-related dose modifications or delays in treatment are unlikely to have an impact on the risk of ALL relapse. To date, the largest study assessing HL during childhood ALL therapy (n=262) found a higher risk for TAH among children ≥10 years and obese children (body mass index ≥95th percentile).9 Our results confirm the association with body mass index in a larger, more diverse cohort but also provide reassurance that recurrent or persistent TAH is rare, and not predicted by known demographic or disease factors. Access to LEARN permitted novel examination of associations between TAH and treatment intensity by phase, rather than by protocol, showing a greater risk for TAH during the early phases of high-intensity treatment. While transaminitis is a recognized event during maintenance, we note that TAHhyperbilirubinemia is also frequent, suggesting that both should be monitored. Recent recommendations to cap the PEG-asparaginase dose in obese patients and for those ≥22 years old may lead to a decrease in TAH in these at-risk populations. The strengths of this study include the racial and ethnic diversity of the cohort, providing robust support that race and ethnicity do not independently predict TAH in childhood ALL. The use of automated HL extraction provides a more granular understanding of the impact of specific treatment blocks, minimizing abstraction error and reporting bias. Potential limitations include our inability to assess the impact on hepatic synthetic function and drug metabolism, as these assessments are not routinely obtained. Furthermore, determination of precise temporal trends with respect to the administration of specific chemotherapy agents and other concomitant medications was not possible from the data available. While our findings suggest an infrequent need for dose modifications due to TAH, dosing data were not available to con-

Table 2. Multiple logistic regression model for variables associated with treatment-associated hepatotoxicity.

Characteristic

Treatment intensity Standard intensity High intensity Mean age Race/ethnicity Non-Latino-White Latino Non-Latino-Black Asian Other Gender Male Female BMI category* Not overweight or obese Overweight or obese ALL immunophenotype B-cell T-cell Minimal residual disease^ Negative Positive

TAH in any phase (n=110) OR (95% CI) P-value

Recurrent/persistent TAH (n=16) OR (95% CI) P-value

1.0 (REF) 0.8 (0.5 -1.5) 1.0 (1.0-1.1)

0.515 0.358

1.0 (REF) 3.9 (0.7-22.7) 1.0 (0.9-1.2)

0.125 0.434

1.0 (REF) 1.1 (0.7-1.8) 0.6 (0.2-1.6) 1.4 (0.6-3.5) 0.4 (0.1-1.7)

0.771 0.321 0.470 0.211

1.0 (REF) 0.9 (0.3-3.1) --1.2 (0.1-10.5)

0.872 --0.878

1.0 (REF) 0.9 (0.6-1.5)

0.774

1.0 (REF) 1.2 (0.4-4.1)

0.755

1.0 (REF) 1.7 (1.1-2.8)

0.027

1.0 (REF) 1.6 (0.5-5.5)

0.461

1.0 (REF) 1.2 (0.5-2.6)

0.729

1.0 (REF) 2.3 (0.5-10.1)

0.253

1.0 (REF) 0.7 (0.4-1.3)

0.289

1.0 (REF) 0.4 (0.1-1.9)

0.237

*Subject numbers dependent on documented/abstracted height & weight from diagnosis. ^Subject numbers dependent on documented minimal residual disease at the end of induction. TAH: treatment-associated hepatotoxicity; OR: odds ratio; 95% CI: 95% confidence interval; REF: reference; BMI: body mass index; ALL: acute lymphoblastic leukemia.

haematologica | 2022; 107(5)

1187

Letters to the Editor

firm this observation. Last, the long-term impact of TAH on liver function could not be determined here.15 Despite these limitations, our results provide novel insights regarding the impact of ALL treatment on HL and the risk of TAH in overweight/obese patients, providing guidance for future study designs that integrate potentially hepatotoxic novel therapeutics, and supporting further investigation of underlying pharmacogenomic contributors to TAH. Joanna S. Yi,1 Tiffany M. Chambers,1 Kelly D. Getz,2 Tamara P. Miller,3 Evanette Burrows,2 Marla H. Daves,1 Philip J. Lupo,1 Michael E. Scheurer,1 Richard Aplenc,2 Karen R. Rabin1 and Maria M. Gramatges1 1 Baylor College of Medicine, Texas Children’s Cancer and Hematology Centers, Houston, TX; 2University of Pennsylvania Perelman School of Medicine, Children's Hospital of Philadelphia, Philadelphia, PA and 3Emory University School of Medicine, Children’s Healthcare of Atlanta, Atlanta, GA, USA Correspondence: JOANNA S. YI - Joanna.yi@bcm.edu doi:10.3324/haematol.2021.279805 Received: August 11, 2021. Accepted: January 12, 2022. Pre-published: January 27, 2022. Disclosures: no conflicts of interest to disclose Contributions: MMG, KRR, and RA conceptualized the study. KDG, TPM, EB, MHD, and RA abstracted the data. TMC and MES cleaned the data and performed all analyses with input from JSY, KDG, TPM, PJL, KRR, and MMG. JSY and MMG wrote the manuscript with significant editing from TMC, MES, and KRR. Funding: the authors would like to thank the following funding sources: Alex’s Lemonade Stand Foundation (ALSF) Epidemiology Award (to RA), ALSF Young Investigator Award (to KDG), St. Baldrick’s Consortium Grant with generous support from the Micaela’s Army Foundation (to PJL, KRR, and MES), Cancer Prevention and Research Institute of Texas #RP160771 (to MES), National Institutes of Health (NIH) National Cancer Institute (5K12CA090433-17, Principal investigator: Dr. Susan Blaney, support for JSY and 1P20CA262733-01, Principal investigators: Lupo and Rabin), NIH National Heart, Lung, and Blood Institute Career Development Award 5K01HL143153 (to KDG), and NIH National Cancer Institute Career Development Award K07CA211956 (to TPM)..

1188

Data-sharing statement: data from LEARN may be accessed by investigators who submit an application that is reviewed and approved by the study team.

References 1. Hunger SP, Mullighan CG. Acute lymphoblastic leukemia in children. N Engl J Med. 2015;373(16):1541-1552. 2. Hough R, Vora A. Crisis management in the treatment of childhood acute lymphoblastic leukemia: putting right what can go wrong (emergency complications of disease and treatment). Hematology Am Soc Hematol Educ Program. 2017;2017(1):251-258. 3. Schmiegelow K, Pulczynska M. Prognostic significance of hepatotoxicity during maintenance chemotherapy for childhood acute lymphoblastic leukaemia. Br J Cancer. 1990;61(5):767-772. 4. Farrow AC, Buchanan GR, Zwiener RJ, et al. Serum aminotransferase elevation during and following treatment of childhood acute lymphoblastic leukemia. J Clin Oncol. 1997;15(4):1560-1566. 5. Adam de Beaumais T, Dervieux T, Fakhoury M, et al. The impact of high-dose methotrexate on intracellular 6-mercaptopurine disposition during interval therapy of childhood acute lymphoblastic leukemia. Cancer Chemother Pharmacol. 2010;66(4):653-658. 6. Segal I, Rassekh SR, Bond MC, et al. Abnormal liver transaminases and conjugated hyperbilirubinemia at presentation of acute lymphoblastic leukemia. Pediatr Blood Cancer. 2010;55(3):434-439. 7. McAtee CL, Schneller N, Brackett J, et al. Treatment-related sinusoidal obstruction syndrome in children with de novo acute lymphoblastic leukemia during intensification. Cancer Chemother Pharmacol. 2017;80(6):1261-1264. 8. Ebbesen MS, Nygaard U, Rosthoj S, et al. Hepatotoxicity during maintenance therapy and prognosis in children with acute lymphoblastic leukemia. J Pediatr Hematol Oncol. 2017;39(3):161-166. 9. Denton CC, Rawlins YA, Oberley MJ, et al. Predictors of hepatotoxicity and pancreatitis in children and adolescents with acute lymphoblastic leukemia treated according to contemporary regimens. Pediatr Blood Cancer. 2018;65(3):e26891. 10. Hashmi SK, Navai SA, Chambers TM, et al. Incidence and predictors of treatment-related conjugated hyperbilirubinemia during early treatment phases for children with acute lymphoblastic leukemia. Pediatr Blood Cancer. 2019;67(2):e28063. 11. Goggins WB, Lo FF. Racial and ethnic disparities in survival of US children with acute lymphoblastic leukemia: evidence from the SEER database 1988-2008. Cancer Causes Control. 2012;23(5):737-743. 12. Faulk KE, Anderson-Mellies A, Cockburn M, et al. Assessment of enrollment characteristics for Children's Oncology Group (COG) upfront therapeutic clinical trials 2004-2015. PLoS One. 2020;15(4):e0230824. 13. Miller TP, Li Y, Getz KD, et al. Using electronic medical record data to report laboratory adverse events. Br J Haematol. 2017;177(2):283286. 14. Bhatia S, Sather HN, Heerema NA, et al. Racial and ethnic differences in survival of children with acute lymphoblastic leukemia. Blood. 2002;100(6):1957-1964. 15. Castellino S, Muir A, Shah A, et al. Hepato-biliary late effects in survivors of childhood and adolescent cancer: a report from the Children's Oncology Group. Pediatr Blood Cancer. 2010;54(5):663669.

haematologica | 2022; 107(5)

Letters to the Editor

SF3B1-mutant myelodysplastic syndrome/myeloproliferative neoplasms: a unique molecular and prognostic entity Molecular abnormalities are prognostically relevant in morphological subtypes of myelodysplastic/myeloproliferative neoplasms (MDS/MPN), giving rise to contemporary molecularly integrated prognostic models.1-3 Established adverse prognostic associations include truncating ASXL1 mutations in chronic myelomonocytic leukemia (CMML)1 and MDS/MPN with ring sideroblasts and thrombocytosis (MDS/MPN-RS-T),3 TP53 and CBL mutations in unclassifiable MDS/MPN (MDS/MPN-U),4 and TET2 mutations in BCR-ABL1-negative atypical chronic myeloid leukemia.2 Recently, molecular signatures have been used to further stratify MDS/MPN-U patients into CMML-like (ASXL1, SRSF2, RUNX1, and/or NRAS mutant), MDS/MPN-RS-T-like (JAK2 and/or SF3B1 mutant), atypical chronic myeloid leukemia-like (SETBP1 and/or ASXL1 mutant), TP53 mutant and an “others” category.5,6 Despite their prognostic impact, these mutations are not specific for underlying disease entities. Recently, SF3B1 mutations were shown to be disease-defining in a subset of patients with MDS7,8 and CMML.9 Whether SF3B1 mutations are similarly diseasedefining in other myeloid subgroups is not known. Given the relative rarity of MDS/MPN patients, we assembled a large, molecularly annotated cohort of MDS/MPN patients to assess the clinical and prognostic impact of SF3B1 mutations, agnostic of disease morphology. After Mayo Clinic institutional review board approval, clinical data from adult (age at diagnosis >18 years) patients with a World Health Organization (WHO)defined diagnosis of MDS/MPN (CMML, MDS/MPN-U and MDS/MPN-RS-T), from 1994 to 2020, were included in the analysis. Patients with atypical chronic myeloid leukemia were excluded due to lack of uniform genetic annotation, limited SF3B1 mutations (n=2), and patient numbers (n<50). A separate cohort of SF3B1-mutant MDS patients diagnosed between 1994 to 2017 was included for comparison. An external cohort of patients from H. Lee Moffitt Cancer Center (Tampa, FL, USA) was used for independent validation after institutional review board approval. Next-generation sequencing for myeloid relevant genes was done at diagnosis or first referral, using institutional or commercially available myeloid malignancy-specific gene panels according to previously published methods.4 The distribution of continuous variables was statistically compared using nonparametric (Mann-Whitney or Kruskal-Wallis) tests, while nominal variables were compared using the c2 test. Time-to-event analyses (for overall [OS] and acute myeloid leukemiafree survival [LFS]) were performed using the method of Kaplan-Meier, with death (for OS), transformation to acute myeloid leukemia (for LFS), and allogeneic hematopoietic stem cell transplantation (for both OS and LFS) used as censors. Overall, 778 consecutive WHO-defined MDS/MPN patients were included in the primary cohort (CMML, n=578 [74%]; MDS/MPN-RS-T, n=79 [10%] and MDS/MPN-U, n=121 [16%]). The median age was 72 (range, 18-95) years with 511 (66%) males (Table 1). Four (3%) patients in the MDS/MPN-U group met proposed criteria for oligomonocytic CMML and had an absolute monocyte count between 0.5 to 0.9 x 109/L, with monocytes constituting >10% of the total white blood cell count.10 Cytogenetic abnormalities (excluding sole -Y) were present in 197 (28%) of 695 assessable patients; 138 haematologica | 2022; 107(5)

(70%) patients with a single karyotypic abnormality, 35 (18%) with a complex karyotype (defined as ≥3 independent structural/numerical abnormalities, excluding autosomal monosomies) and 26 (13%) with monosomal karyotypes, with frequent cytogenetic abnormalities including 51 (26%) +8, 49 (25%) -7/7q-, 23 (12%) 20q-, 12 (6%) 5q- (2 as sole abnormalities, classified as MDS/MPN based on morphology), 11 (6%) 13q-, 6 (3%) inv(3)/3q26 (3 GATA2-EVI1 fusion), and 5 (3%) with-11/11q23 (KMT2A). Cytogenetic risk stratification as per the CMML-specific scoring system (CPSS) cytogenetic stratification11 was predictive of OS (P<0.0001) in our cohort with 498 (72%) in the low-risk category (median OS 41 months [95% CI: 32-50]), 120 (17%) in the intermediate-risk category (median OS 21 months [95% CI: 16-33]), and 77 (11%) in the high-risk category (median OS 16 months [95% CI: 11-23]). Next-generation sequencing information at diagnosis was available for 444 (57%) patients with frequent molecular abnormalities being ASXL1 (n=235; 45%), SRSF2 (n=179; 40%), TET2 (n=155; 39%), SF3B1 (n=78, 15%), and DNMT3A (n=30, 7%) mutations (Table 1). At last median follow-up of 44 (95% CI: 37-50) months, transformation to acute myeloid leukemia had occurred in 123 (16%) patients, and 414 (53%) deaths had been documented. The Kaplan-Meier estimate of median OS was 32 (95% CI: 28-38) months (CMML 31 [95% CI: 2737] months, MDS/MPN-RS-T 67 [95% CI: 43-101] months, and MDS/MPN-U 25 [95% CI 21-36] months), while the median was not reached for LFS. In the MDS/MPN cohort, there were 78 patients with SF3B1 mutations: 18 (23%) with CMML, 45 (58%) with MDS/MPN-RS-T, and 15 (19%) with MDS/MPN-U. There were 15 SF3B1 mutation hotspots (evaluable in 53 patients) with the most common abnormalities being K700E (n=24, 45%), H662Q (n=8, 15%), and K666R (n=6, 11%). The clinical and genomic characteristics are outlined in Online Supplementary Table S1. We then combined all SF3B1-mutant MDS/MPN patients into one category (n=78) and compared them to their wild-type counterparts (n=446) (Table 1). The two groups had significant differences in clinical and molecular features as highlighted in Table 1. The median variant allele frequency (VAF) of mutant SF3B1 was 43% (range, 8-65) overall, being 43% (range, 8-65) in CMML patients, 43% (range, 12-50) in MDS/MPN-RS-T patients, and 40% (range, 16-52) in MDS/MPN-U patients (P=0.9), and was comparable to the median variant allele frequency of mutant ASXL1 at 37% (range, 11-52): CMML 37% (range, 27-37), MDS/MPN-RS-T 32% (range, 18-52), and MDS/MPN-U 29% (range, 11-43). As expected, a higher frequency of SF3B1-mutant versus SF3B1-wild type MDS/MPN patients (21% vs. 2%, P<0.0001) were treated with lenalidomide and erythropoiesis-stimulating agents (64% vs. 39%, P<0.0001), but the frequency of hypomethylating agent therapy use was similar (21% vs. 32%, P=0.1) (Table 1). The SF3B1 mutant cohort had a lower rate of transformation to acute myeloid leukemia (5% vs. 18%, P=0.0006) in comparison to the SF3B1-wild type cohort. The Kaplan-Meier estimates of LFS (median not reached in both groups, P=0.0002) and OS (57 vs. 31 months, P=0.03) were higher in the SF3B1-mutant MDS/MPN patients (Table 1 and Figure 1). These findings were validated in an external MDS/MPN cohort from Moffitt Cancer Center comprising 380 patients, 253 with CMML, 80 with MDS/MPN-RS-T, and 47 with MDS/MPN-U. The validation cohort was similar to the Mayo Clinic cohort in terms of age (P=0.4) and median follow-up (P=0.1). Importantly, SF3B1-mutant VAF was not predictive of OS in either the Mayo Clinic cohort 1189

Letters to the Editor

Table 1. Comparison of the clinical and morphological characteristics of patients with SF3B1 mutant or wild-type myelodysplastic syndromes/myeloproliferative neoplasm compared with patients with SF3B1-mutant myelodysplastic syndromes.

Variable; N (%) or median (range) Age, years N. of males Hemoglobin, g/dL WBC count x 109/L ANC x 109/L AMC x 109/L Platelet count x 109/L BM RS, % PB blasts ≥1% BM blasts ≥5% Abnormal karyotype (except -Y, %), Evaluable=695

SF3B1 mutant MDS/MPN patients (n=78)

SF3B1 wild-type SF3B1 P value (SF3B1 P value (SF3B1 mutant MDS/MPN patients mutant MDS mutant vs. MDS/MPN vs. (n=446) (n=75) wild-type MDS/MPN) SF3B1 mutant MDS

74 (43-93) 42 (54) 9.4 (6.4-13.3) 7.6 (1.8-96.1) 4 (0.4-54.7) 0.7 (0.1-11.5) 521 (63-1243) 50 (0-90) 9 (12) 8 (10) 11 (15)

72 (18-95) 294 (66) 10.6 (4.2-16.9) 13 (1-265) 6.5 (0-151) 2.3 (0-40) 98 (8-1778) 0 (0-80) 130 (29) 150 (34) 123 (30)

74 (41-94) 48 (64) 9.5 (7-13.5) 5.2 (1.5-13.1) 2.9 (0.4-9.4) 0.4 (0.06-1) 268 (62-599) 40 (5-80) 11 (15)

0.3 <0.0001* <0.0001* <0.0001* 0.001* <0.0001* <0.0001* <0.0001* 0.0005* <0.0001* 0.0004*

0.8 <0.0001* 0.7 <0.0001* 0.0004* <0.0001* <0.0001* 0.4 0.0003* <0.0001* 0.7

33 (47) 46 (64) 15 (21) 15 (21) 2 (3)

170 (55) 120 (39) 6 (2) 94 (32) 25 (6) 27 (9)

1 (2) 52 (84) 9 (15) 10 (16) -

0.4 <0.0001* <0.0001* 0.1 0.01* 0.1

<0.0001* <0.0001* 0.006* 0.03* <0.0001* <0.0001*

4 (5) Median NR 57 (30-68)

79 (18) Median NR 31 (26-36)

2 (3) Median NR 65 (43-85)

0.4 0.0002* 0.03*

0.4 0.3 0.2

Treatment (total evaluable=619) Hydroxyurea ESA Lenalidomide HMA therapy Allogeneic HCT Investigational agents (clinical trial)

Outcomes Transformation to AML AML-free survival, months Overall survival, months; median (95% CI)

MDS: myelodysplastic syndromes; MPN: myeloproliferative neoplasms; WBC: white blood cell; PB: peripheral blood; BM: bone marrow; ANC: absolute neutrophil count; AMC: absolute monocyte count; RS: ring sideroblasts; BM: bone marrow; ESA: erythropoiesis-stimulating agent; HMA: hypomethylating agent; HCT: hematopoietic stem cell transplant; AML: acute myeloid leukemia; NR: not reached; 95% CI: 95% confidence interval.. *Statistically significant differences.

1190

Figure 1. Differences in outcomes of patients with myelodysplastic syndrome/myeloproliferative neoplasms with or without SF3B1 mutations. (A) KaplanMeier estimates of overall survival (OS) in myelodysplastic syndrome (MDS) /myeloproliferative neoplasms (MPN) patients with SF3B1 mutations compared to SF3B1 wild-type MDS/MPN patients (median: 57 vs. 31 months, P=0.03) in the Mayo Clinic cohort. (B) Leukemia-free survival in MDS/MPN patients with SF3B1 mutations compared to SF31B wild-type MDS/MPN patients (median not reached in either group, P=0.0002) in the Mayo Clinic cohort. (C) Overall survival in SF3B1 mutant MDS/MPN patients compared to SF3B1-wild type patients in the Moffitt Cancer Center cohort (median: 108 vs. 39 months, P<0.0001). (D) Leukemia-free survival in SF3B1 mutant MDS/MPN patients compared to SF3B1-wild-type patients in the Moffitt Cancer Center cohort (median not reached in either group, P=0.0001).

haematologica | 2022; 107(5)

Letters to the Editor

(P=0.3) or the Moffitt Cancer Center cohort (P=0.7). In addition, there were no differences in OS between patients with mutations in the K700E hotspot and nonK700E sites in either the Mayo Clinic cohort (median OS 49 [95% CI: 22-109] months vs. 67 [95% CI: 36-126] months, P=0.5) (Online Supplementary Figure S1A) or the Moffitt Cancer Center cohort (median OS 85 [95% CI) months vs. not reached, P=0.9) (Online Supplementary Figure S1B). We then compared SF3B1-mutant MDS/MPN patients (n=78) with SF3B1-mutant MDS patients (n=75) (Table 1). SF3B1-mutant MDS/MPN patients had a higher frequency of JAK2 V617F mutations (25% vs. 1%, P<0.0001; 10% vs. 1%, P=0.002 when MDS/MPN-RS-T patients were excluded) (Figure 2A, Table 1). When the SF3B1 mutation hotspots were compared between the two groups, SF3B1 K700E was the most common hotspot in both categories and was present in 24 (47%) MDS/MPN patients and 39 (53%) SF3B1-mutant MDS patients (P=0.5) (Online Supplementary Figure S1C, D). Overall, there were seven patients with co-occurring SF3B1 and SRSF2 mutations (4 with CMML, 3 with SF3B1-mutant MDS). Mutation details were available for two CMML patients; SF3B1 Y623C (42%)/SRSF2 P95H (45%) and SF3B1 K700E (45.2%)/SRSF2 P95H (2.8%), and two SF3B1-mutant MDS patients; SF3B1 K666Q (29%)/SRSF2 P95T (48%) and SF3B1 K700E (9%)/SRSF2 P95R (29%). At last median follow-up of 102 (95% CI: 63-141) months, there were no significant differences in rates of transformation to acute myeloid leukemia (5% vs. 3%, P=0.4), Kaplan-Meier estimates of median LFS

(median not reached, P=0.3) or median OS (median, 57 vs. 65 months, P=0.2) between the two cohorts (Figure 2B, Table 1). We then stratified SF3B1-mutant MDS/MPN patients by morphological features such as percentage of ring sideroblasts in bone marrow and percentages of blasts in peripheral blood and bone marrow. In a univariate survival analysis, percentage of peripheral blood blasts (P=0.1), percentage of bone marrow blasts ≥5 (P=0.4), abnormal karyotype (P=0.3), revised International Prognostic Scoring System score (P=0.8) and CPSS cytogenetic group (P=0.5) were not predictive of OS. Molecular abnormalities (overall frequency ≥5%) such as ASXL1 (P=0.3), TET2 (P=0.08), DNMT3A (P=0.6), JAK2 V617F (P=0.8), U2AF1 (P=0.2), SRSF2 (P=0.7), ZRSR2 (P=0.3), CBL (P=0.3) NRAS (P=0.8), or any RAS pathway mutation (KRAS/NRAS/CBL/PTPN11, P=0.3) did not affect OS (only 1 patient each had TP53 and RUNX1 mutations). Additionally, WHO criteria were unable to prognostically distinguish both Mayo Clinic (P=0.3) and combined (Mayo Clinic and Moffitt Cancer Center) cohorts of SF3B1-mutant MDS/MPN patients (P=0.7). Furthermore, neither the standard International Prognostic Scoring System (P=0.3), nor the revised version (P=0.7) was able to stratify SF3B1-mutant and MDS/MPN patients into prognostically relevant subtypes. Finally, we conducted a multivariate analysis in the combined Mayo Clinic cohort of MDS/MPN and MDS patients with known independent prognostic factors in myeloid malignancies such as hemoglobin <10 g/dL, age

Figure 2. Differences in genomic characteristics and outcomes among patients with SF3B1-mutant myelodysplastic syndrome/myeloproliferative neoplasms and SF3B1-mutant myelodysplastic syndromes. (A) Heatmap of molecular abnormalities in the two groups (mutations in genes with a frequency of 5% or higher are included in the figure). The only statistically significant difference between the SF3B1-mutant myelodysplastic syndrome (MDS) /myeloproliferative neoplasms (MPN) and MDS groups was the higher frequency of JAK2 in the former (25 vs. 1%, P<0.0001). (B) Kaplan-Meier estimate of overall survival between SF3B1-mutant MDS/MPN and SF3B1-mutant MDS (median, 57 months [95% confidence interval: 30-68] vs. 65 months [95% confidence interval: 43-85], P=0.2) patients in the Mayo Clinic cohort. (C) Kaplan-Meier estimate of overall survival between SF3B1-mutant MDS/MPN and SF3B1-mutant MDS (median, 85 months [95% confidence interval: 58-not reached] vs. 97 months [95% confidence interval: 55-118], P=0.7) patients in the Moffitt Cancer Center cohort.

haematologica | 2022; 107(5)

1191

Letters to the Editor

≥70 years, platelet count ≥450 x 109/L, cytogenetic subtypes (as per CPSS stratification), SF3B1 and ASXL1 mutations, and bone marrow blast percentage ≥5, and found that SF3B1 mutations retained their independent favorable prognostic impact (P=0.01) (Online Supplementary Table S2). In summary, our data indicate that SF3B1-mutant MDS/MPN is a clinically and genomically distinct category within overlap myeloid neoplasms and, pending further validation, should be considered as a unique prognostic entity. Additionally, patients with this condition have distinct clinical and molecular characteristics in comparison to SF3B1-mutant MDS patients, arguing against a uniform classification category of SF3B1mutant myeloid neoplasms. Limitations of our study include smaller numbers of patients in certain subgroup comparisons, differential follow-up times and therapy choices, and selection biases largely due to the retrospective nature of the analyses. Abhishek A. Mangaonkar,1 Terra L. Lasho,1 Christy Finke,1 Rhett P. Ketterling,2 Kaaren K. Reichard,2 Kristen McCullough,1 Naseema Gangat,1 Aref Al-Kali,1 Kebede H. Begna,1 William H. Hogan,1 Mark R. Litzow,1 Hassan Alkhateeb,1 Mithun Shah,1 Animesh Pardanani,1 Ayalew Tefferi,1 Najla H. Al Ali,3 Chetasi Talati,3 David Sallman,3 Eric Padron,3 Rami Komrokji3 and Mrinal M. Patnaik1 1 Division of Hematology, Department of Medicine, Mayo Clinic, Rochester, MN; 2Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, MN and 3Department of Malignant Hematology, H. Lee Moffitt Cancer Center, Tampa, FL, USA. Correspondence: MRINAL M. PATNAIK - patnaik.mrinal@mayo.edu RAMI KOMROKJI - Rami.Komrokji@moffitt.org doi:10.3324/haematol.2021.280463 Received: December 7, 2021. Accepted: January 28, 2022. Pre-published: February 10, 2022. Disclosures: AM has received research funding from BMS. EP has received honoraria from and/or serves on advisory boards for BluePrint, CTI, Stemline Therapeutics, Taiho and BMS; and has received research funding from Kura, Incyte, and BMS. MP has received research funding from Kura Oncology and Stemline Pharmaceuticals. All other authors declare that they have no conflicts of interest. Contributions: AM compiled the clinical and genomics data, analyzed data, and wrote all the drafts of the manuscript. TL and CF performed the genomics analysis. RPK and KKR reviewed the pathology information and edited all drafts of the manuscript. KM, NG, AAK,

1192

KHB, WHH, MRL, HA, MS, AP, and AT contributed patients and edited all drafts of the manuscript. NHA, CT, DS, EP and RK contributed data from the external (validation) cohort and edited all drafts of the manuscript. MP conceptualized the study and edited all drafts of the manuscript. All authors contributed to the writing of the manuscript. Funding: this publication was supported by grant #UL1 TR002377 from the National Center for Advancing Translational Sciences (NCATS); however, the findings do not necessarily represent official views of the National Institutes of Health. Data-sharing statement: for original data, please contact patnaik.mrinal@mayo.edu

References 1. Patnaik MM, Itzykson R, Lasho T, et al. ASXL1 and SETBP1 mutations and their prognostic contribution in chronic myelomonocytic leukemia: a two-center study of 466 patients. Leukemia. 2014;28(11):2206. 2. Patnaik MM, Barraco D, Lasho TL, et al. Targeted next generation sequencing and identification of risk factors in World Health Organization defined atypical chronic myeloid leukemia. Am J Hematol. 2017;92(6):542-548. 3. Patnaik MM, Lasho TL, Finke CM, et al. Predictors of survival in refractory anemia with ring sideroblasts and thrombocytosis (RARST) and the role of next-generation sequencing. Am J Hematol. 2016;91(5):492-498. 4. Mangaonkar AA, Swoboda DM, Coltro G, et al. Clinicopathologic characteristics, prognostication and treatment outcomes for myelodysplastic/myeloproliferative neoplasm, unclassifiable (MDS/MPN-U): Mayo Clinic-Moffitt Cancer Center study of 135 consecutive patients. Leukemia. 2020;34(2):656-661. 5. Palomo L, Meggendorfer M, Hutter S, et al. Molecular landscape and clonal architecture of adult myelodysplastic/myeloproliferative neoplasms. Blood. 2020;136(16):1851-1862. 6. Mangaonkar AA, Swoboda DM, Lasho TL, et al. Genomic stratification of myelodysplastic/myeloproliferative neoplasms, unclassifiable: sorting through the unsorted. Leukemia. 2021;35(11):33293333. 7. Malcovati L, Karimi M, Papaemmanuil E, et al. SF3B1 mutation identifies a distinct subset of myelodysplastic syndrome with ring sideroblasts. Blood. 2015;126(2):233-241. 8. Malcovati L, Stevenson K, Papaemmanuil E, et al. SF3B1-mutant MDS as a distinct disease subtype: a proposal from the International Working Group for the Prognosis of MDS. Blood. 2020;136(2):157170. 9. Wudhikarn K, Loghavi S, Mangaonkar AA, et al. SF3B1-mutant CMML defines a predominantly dysplastic CMML subtype with a superior acute leukemia-free survival. Blood Adv. 2020;4(22):57165721. 10. Valent P, Orazi A, Savona MR, et al. Proposed diagnostic criteria for classical chronic myelomonocytic leukemia (CMML), CMML variants and pre-CMML conditions. Haematologica. 2019;104(10):19351949. 11. Such E, Germing U, Malcovati L, Hrodek O. Development and validation of a prognostic scoring system for patients with chronic myelomonocytic leukemia. Transfuze a Hematologie Dnes. 2013;19(3):191.

haematologica | 2022; 107(5)

Letters to the Editor

Immune thrombocytopenia following vaccination during the COVID-19 pandemic To date, there have been over 3.2 million doses of ChAdOx1 nCoV-19 (ChAd) COVID-19 vaccine (AstraZeneca) and 1 million doses of BNT162b2 (BNT) COVID-19 vaccine (Pfizer-BioNTech) administered in Australia. Among the numerous safety signals that have been raised, we present our case series of immune thrombocytopenia (ITP) after COVID-19 vaccination.1-4 ITP following vaccination has been previously described in other settings and after mRNA-based COVID-19 vaccines.5-8 A Scottish National Registry study examined general practice data and identified a small increased incidence of ITP diagnoses between days 0-27 after vaccination with ChAD.9 We present the clinical characteristics and treatment outcomes of patients diagnosed with ITP following COVID vaccinations (ChAd or BNT) in Australia. After obtaining independent ethics committee approval, we contacted hemostasis hematologists across

Australia to participate in our comprehensive survey of clinical presentations of vaccine-associated ITP as defined by the temporal relationship of ITP within 42 days following COVID-19 vaccination, without an otherwise apparent alternative cause or thrombosis. Patients with thrombosis or elevated D-dimer levels were investigated and excluded for vaccine-induced immune-mediated thrombotic thrombocytopenia according to international guidelines.10 Response was defined as per international consensus guidelines as a platelet count ≥30x109/L, 2-fold increase over baseline and absence of bleeding. A complete response was defined as a platelet count ≥100x109/L and absence of bleeding.11 A total of 14 patients were diagnosed with ITP following vaccination. Twelve of these cases followed administration of the ChAd vaccine. Ten cases were de novo ITP, presented in Table 1. Four cases were relapses in patients with previously stable chronic ITP, presented in Table 2. None of the 14 patients had concurrent thrombosis. Among the 12 cases of ITP following administration of the ChAd vaccine, an enzyme-linked immunosorbent

Table 1. Demographics and clinical features of patients with newly diagnosed immune thrombocytopenia after COVID-19 vaccination.

Age Days COVID-19 Other Platelets at WHO Bleeding and after vaccine antecedent presentation bleeding gender vaccination vaccinations (and nadir score (30 days) if later) (x109/L)

First-line treatment

Second-line TTR TTCR Platelets Treatments treatment at day 30 at day 30 (x109/L)

52M

1st ChAd

None

Petechiae

Pred/IVIg

80F

1st ChAd

Influenza

82M

1st ChAd

None

22 (1)

Life threatening bleeding Petechiae

60F 83F

3 23

1st ChAd 1st ChAd

None None

3 10

1 1

Petechiae Petechiae, ecchymoses

Dex/IVIg Dex/IVIg

Pred None

61M

1st ChAd

None

Pred/MMF

82M

1st ChAd

None

Purpura

Dex

86M

1st ChAd

None

5 (3)

46M

1st BNT

None

5 (0)

22M

2nd ChAd

1st ChAd given 4 weeks prior

176

Pred/IVIg

Eltrombopag 18

157

Pred/IVIg

Eltrombopag 10

197

2 1

3 3

25 40

None

104

Pred

None

259

Pred 12.5 mg AML in daily remission Eltrombopag (not on chemo) 50 mg daily None None None None (Pulse Dex/IVIg repeated day 21) Pred 20 mg daily None MMF 500 mg BD Dex pulse Influenza repeated vaccination day 17 after presenting with ITP Pred 25 mg daily None

None

151

Pred 35 mg daily

None

N/A

AIHA

Major bleeding Pred/IVIg requiring hospitalization Mild blood Pred/IVIg loss Petechiae Pred/IVIg

None

Pred 5 mg daily Pred 75 mg daily

Other relevant history

None None

AIHA: autoimmune hemolytic anemia; AML: acute myeloid leukemia; BD: twice daily; BNT: BNT162b2 (Pfizer); ChAd: ChAdOx1 nCoV-19 (AstraZeneca); COVID-19: coronavirus disease 2019; Dex: dexamethasone; F: female; IVIg: intravenous immunoglobulin; ITP: immune thrombocytopenia; M: male; MMF: mycophenolate mofetil; Pred: prednisone; TTCR: time to complete response; TTR: time to response; WHO: World Health Organization.

haematologica | 2022; 107(5)

1193

Letters to the Editor

Table 2. Demographics and clinical features of patients with relapsed chronic immune thrombocytopenia after COVID-19 vaccination. Age Days COVID-19 Other Chronic Most Platelets WHO Bleeding FirstSecond- TTR TTCR Platelets Treatments Other and after vaccine antecedent ITP recent at bleeding line line at day 30 at day 30 relevant gender vaccination vaccinations treatments platelets presentation score treatment treatment (x109/L) history prior to (x109/L) vaccination (x109/L) 94F

1st BNT

77M

1st ChAd Influenza

73F

1st ChAd

None

73M

1st ChAd

None

Stable on romiplostim

Stable off treatment Stable off treatment Stable off treatment

188

255

120

None

No change None to ongoing romiplostim Petechiae Pred/IVIg None

None

144

None

Mild blood loss Petechiae

Pred 15 mg daily None

IVIg

None

215

Pred

None

234

None

Pred Positive 10 mg DAT and ANA

ANA: antinuclear antibodies; BNT: BNT162b2 (Pfizer); ChAd: ChAdOx1 nCoV-19 (AstraZeneca); COVID-19: coronavirus disease 2019; DAT: direct antiglobulin test; Dex: dexamethasone; F: female; IVIg: intravenous immunoglobulin; M: male; Pred: prednisone; TTCR: time to complete response; TTR: time to response; WHO: World Health Organization.

assay for platelet factor 4 (PF4) was performed in six and all of these tested negative. The median age of the patients was 75 years (range, 22-94), the median time to presentation after vaccination was 10 days (range, 2-31), and the platelet count at presentation was 7x109/L (range, 0-22x109/L). World Health Organization bleeding scores were mild: ten patients had grade 0 or 1, two patients had grade 2, and one patient each had grades 3 and 4.12 Ten cases had no prior history of ITP and all received treatment upfront: seven received prednisone, and three high-dose dexamethasone pulses. Eight patients also received between 1-2 g/kg intravenous immunoglobulins (IVIg) as part of first-line therapy. The median time to response was 3.5 days (range, 1-18). Ten evaluable patients achieved a complete response by a median of 9 days (range, 3-47). Day 30 data were available for nine of these ten patients without a prior history of ITP, as one left Australia: the median platelet count was 151x109/L (range, 8-259x109/L); eight were still on corticosteroids (median prednisone equivalent 20 mg daily), one was on eltrombopag (commenced as second-line treatment) and another was receiving mycophenolate mofetil that had been commenced in first-line treatment in combination with prednisone. One 80-year-old female presented with life-threatening bleeding (influenza vaccination 1 day prior and ChAd 21 days prior to presentation) and after no initial response to escalating prednisone doses and IVIg, eltrombopag was commenced on day 15. Platelets began to respond by day 18, and the platelet count rose to 157x109/L by day 30 after presentation while only on prednisone. One 82-year-old male presented with a platelet count of 3x109/L, and widespread bruising 9 days after his first ChAd vaccination. He was treated with high-dose dexamethasone and platelets responded, reaching 97x109/L by day 16 (Figure 1A). He received influenza vaccination the following day, but his ITP relapsed by day 32. He responded promptly to a second pulse of high-dose dexamethasone with a platelet ount of 65x109/L by day 36. He had never previously developed ITP despite numerous influenza vaccinations in the past. One 83-year-old female presented with a platelet count of 10x109/L, facial petechiae, and upper chest ecchy1194

moses 23 days after her first ChAd vaccination (Figure 1B). She responded promptly to a dexamethasone pulse 20 mg daily for 4 days and IVIg infusion 0.4 g/kg for 3 days. She relapsed on day 19 with platelets 23x109/L and new lower limb bruising, and was treated with another pulse of dexamethasone and IVIg 0.4 g/kg for 2 days. In total, there were four patients with chronic ITP who relapsed following COVID-19 vaccination. Three patients receiving ChAd had stable chronic ITP, and were off ITP-directed therapies at the time of COVID-19 vaccination. They were treated with standard first-line therapies and all responded within 3 days. IVIg monotherapy alone was successful in one 72-yearold female with chronic ITP who presented with a platelet count of 11x109/L but responded by day 3, achieving a complete response on day 5; her day 30 platelet count was 215x109/L (Figure 1C), and she had no need for steroids at any time despite having had refractory ITP requiring splenectomy in 1994. Her most recent prior platelet count was 255x109/L less than 3 weeks before vaccination. Her most recent prior ITP treatment had been rituximab monotherapy in 2011. A second chronic ITP patient, a 77-year-old male who received influenza vaccination prior to ChAd vaccination, presented with a platelet count of 2x109/L, achieved a response and complete response by days 3 and 8 respectively, had a day 30 platelet count of 144x109/L, and was on a weaning schedule of prednisone at day 30 after initially being treated with prednisone/IVIg upfront. The third patient with chronic ITP, a 73-year-old male with a pre-vaccination platelet count of 120x109/L, was thrombocytopenic (platelet count, 5x109/L) 31 days after ChAd vaccination. He was started on prednisone monotherapy and achieved a response within 2 days, a complete response by day 4, and a platelet count of 234x109/L by day 30 while on prednisone 10 mg daily. The fourth chronic ITP patient in this analysis was a 94-year-old female who received her first dose of BNT 9 days prior to presentation. She had previously enjoyed a stable platelet response on romiplostim for her chronic ITP with a recent platelet count of 86x109/L, falling to 12x109/L without any bleeding; her platelet count returned to baseline within 5 days of presentation. She proceeded to receive her second dose of BNT 21 days haematologica | 2022; 107(5)

Letters to the Editor

A Figure 1. Clinical course of three separate cases of immune thrombocytopenia following COVID-19 vaccination. (A) An 82-year-old male with newly diagnosed immune thomrbocytopenia (ITP) was treated with dexamethasone an initially responded, received influenza vaccination, relapsed, and responded again to another pulse of dexamethasone and weaning prednisone taper. (B) An 83-year-old female with newly diagnosed ITP was treated with two pulses of dexamethasone/IVIg. (C) A 73-year-old female had a relapse of chronic ITP after receiving ChAd vaccination, received IVIg 2 g/kg over 2 days as monotherapy. ChAd: ChAdOx1 nCoV-19 (AstraZeneca); IVIg: intravenous immunoglobulin.

after the first, relapsing again on day 15 with a platelet count of 14x109/L before returning to her stable baseline within a further 7 days. Our case series of vaccine-associated ITP comprises more cases of ITP following administration of the ChAd vaccine than after the BNT vaccine (12 from 3.2 million ChAd vaccinations vs. 2 from 1 million BNT), although there may be an ascertainment bias due to greater scrutiny of patients following ChAd vaccination, as suggested in a recent Scottish study even though this paper also concluded that there was an increased rate of ITP diagnoses of 1.13 per 100,000 doses.9 In contrast, a Scandinavian epidemiological study was unable to identify an increased rate of ITP diagnoses although rates of “unspecified thrombocytopenia” and bleeding events were increased significantly.13 Our study was not designed to address the questions of frequency or causality. Our designation of these cases as “vaccine-associated” ITP as opposed to co-incident ITP is based on the clinical diagnosis of ITP as one of exclusion. As vaccine association cannot be excluded, we cannot conclude that these patients have primary ITP, conceding that future outcomes may eventually justify revision of our diagnohaematologica | 2022; 107(5)

sis, which is common in ITP.14 Two of 14 cases are confounded at presentation by the recent administration of influenza vaccination, and another patient received influenza vaccination shortly after initial recovery from ITP before relapsing. However, these limitations reflect an unavoidable real-world dilemma as public health imperatives to protect populations at risk during a pandemic will likely outweigh the considerably smaller numerical risk of uncertain outcomes and vaccination side effects when immunization programs overlap. Most cases responded rapidly to first-line therapy although the majority remained on corticosteroids for at least 30 days (median prednisone equivalent dose 13.75 mg daily for all cases, 20 mg daily for those with newly diagnosed ITP). Patients whose chronic ITP relapsed after vaccination responded rapidly to first-line therapies, consistent with other observations,8 and reassuringly for those with underlying ITP who are at present hesitant to receive COVID-19 vaccination. So far, in three patients, a single pulse of high-dose dexamethasone was insufficient to maintain remission in this cohort, but repeat courses have been successful and well tolerated. Additional 1195

Letters to the Editor

strategies used successfully include eltrombopag and mycophenolate mofetil. Further data will be needed to understand the durability of these responses. We anticipate that there may be cases along a spectrum of clinical presentations between vaccine-induced immune-mediated thrombotic thrombocytopenia and vaccine-associated ITP, as have already been noted elsewhere.15 In our cohort, overlapping characteristics have not yet been identified, and all six patients with samples tested were negative for anti-PF4 antibodies. Both local and international registries are currently collecting data that will be useful for investigating treatment strategies and clinical outcomes for patients developing ITP following COVID-19 vaccination. Philip Young-Ill Choi,1,2 Danny Hsu,3 Huyen Anh Tran,4 Chee Wee Tan,5 Anoop Enjeti,6 Vivien Mun Yee Chen,7,8 Beng Hock Chong,9 Jennifer Curnow,10 Dominic Pepperell11 and Robert Bird12 1 The Canberra Hospital, Canberra, Australian Capital Territory; 2 John Curtin School of Medical Research, Australian National University, Canberra, Australian Capital Territory; 3Liverpool Hospital, Liverpool, New South Wales; 4The Alfred Hospital, Melbourne, Victoria; 5Royal Adelaide Hospital, Adelaide, South Australia; 6 Calvary Mater Hospital, Newcastle, New South Wales; 7ANZAC Research Institute, University of Sydney, Sydney, New South Wales; 8 Department of Haematology, Concord Repatriation and General Hospital, Sydney, New South Wales; 9NSW Health Pathology, St George Hospital, University NSW, Sydney, New South Wales; 10 Faculty of Medicine and Health, University of Sydney, Sydney, New South Wales; 11Fiona Stanley Hospital, Perth, Western Australia and 12 Princess Alexandra Hospital, Brisbane, Queensland, Australia Correspondence: PHILIP YOUNG-ILL CHOI - phil.choi@act.gov.au doi:10.3324/haematol.2021.279442 Received: June 15, 2021. Accepted: August 2, 2021. Pre-published: August 26, 2021. Disclosures: no conflicts of interest to disclose. Contributions: RB and PY-LC designed the study, analysed the data, and wrote the manuscript; DH, HAT, CWT, AE, VMYC, BHC, JC and DP reviewed and edited the manuscript. Acknowledgments: the authors would like to thank Drs Fernando

1196

Roncolato, Jock Simpson, Angeline Josie, Kate Melville, Susan Maccallum and Kim Cartwright for their contributions to this work.

References 1. Greinacher A, Thiele T, Warkentin TE, Weisser K, Kyrle PA, Eichinger S. Thrombotic thrombocytopenia after ChAdOx1 nCov-19 vaccination. N Engl J Med. 2021;384(22):2092-2101. 2. Gerber GF, Yuan X, Yu J, et al. COVID-19 vaccines induce severe hemolysis in paroxysmal nocturnal hemoglobinuria. Blood. 2021;137(26):36703673. 3. Patel SU, Khurram R, Lakhani A, Quirk B. Guillain-Barre syndrome following the first dose of the chimpanzee adenovirus-vectored COVID-19 vaccine, ChAdOx1. BMJ Case Rep. 2021;14(4):e242956. 4. Torjesen I. Covid-19: first UK vaccine safety data are “reassuring,” says regulator. BMJ. 2021;372:n363. 5. Miller E, Waight P, Farrington CP, Andrews N, Stowe J, Taylor B. Idiopathic thrombocytopenic purpura and MMR vaccine. Arch Dis Child. 2001;84(3):227-229. 6. Lee E-J, Cines DB, Gernsheimer T, et al. Thrombocytopenia following Pfizer and Moderna SARS-CoV-2 vaccination. Am J Hematol. 2021;96(5):534-537. 7. Cines DB, Bussel JB, Liebman HA, Luning Prak ET. The ITP syndrome: pathogenic and clinical diversity. Blood. 2009;113(26):6511-6521. 8. Kuter DJ. Exacerbation of immune thrombocytopenia following Covid19 vaccination. Br J Haematol. 2021;195(3):365-370. 9. Simpson CR, Shi T, Vasileiou E, et al. First-dose ChAdOx1 and BNT162b2 COVID-19 vaccines and thrombocytopenic, thromboembolic and hemorrhagic events in Scotland. Nat Med. 2021;27(7):1290-1297. 10. 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. 11. Rodeghiero F, Stasi R, Gernsheimer T, et al. Standardization of terminology, definitions and outcome criteria in immune thrombocytopenic purpura of adults and children: report from an international working group. Blood. 2009;113(11):2386-2393. 12. Neunert C, Terrell DR, Arnold DM, et al. American Society of Hematology 2019 guidelines for immune thrombocytopenia. Blood Adv. 2019;3(23):3829-3866. 13. Pottegård A, Lund LC, Karlstad Ø, et al. Arterial events, venous thromboembolism, thrombocytopenia, and bleeding after vaccination with Oxford-AstraZeneca ChAdOx1-S in Denmark and Norway: population based cohort study. BMJ. 2021;373:n1114. 14. 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. 15. Scully M, Singh D, Lown R, et al. Pathologic antibodies to platelet factor 4 after ChAdOx1 nCoV-19 vaccination. N Engl J Med. 2021; 384(23):2202-2211.

haematologica | 2022; 107(5)

Letters to the Editor

Impaired in vivo activated protein C response rates indicate a thrombophilic phenotype in inherited thrombophilia Venous thromboembolism (VTE) is a multifactorial disease. Hereditary risk factors include the common mutations factor V Leiden (FVL) and prothrombin (FII) 20210G>A, with a prevalence of 3–15% among whites, as well as deficiencies of the coagulation inhibitors antithrombin (AT), protein C (PC), and protein S.1 In the recent past, novel risk loci have been found by genomewide association studies.2,3 However, their consideration in addition to the classical thrombophilic defects results in an estimated heritability of VTE of only 15%, in contrast to 40–60% heritability observed in family-based studies.4 In order to identify further unknown genetic thrombophilic defects, consideration of the laboratory phenotype of increased thrombin formation in addition to the clinical phenotype of VTE has been proposed, based on the observation of elevated in vitro thrombin generation parameters in families with unexplained thrombophilia and in carriers of genetic variations in hemostasis-related genes other than FVL and FII

20210G>A.5 It remains unclear, however, if increased in vitro thrombin formation rates indeed reflect increased in vivo thrombin formation. In order to investigate this, we comparatively analyzed in vitro and in vivo thrombin formation in a cohort of healthy individuals and in thrombophilic patients. In vivo coagulation activation was induced by low-dose recombinant activated factor VII (rFVIIa). Subsequent hemostasis biomarker-monitoring included measurement of activated PC (APC) as a measure of the endothelial-dependent anticoagulant response. Recently, using this stimulated hemostasis activity pattern evaluation (SHAPE) approach, we were able to show increased in vivo thrombin generation rates and a comparable APC response in FVL and FII 20210G>A carriers.6,7 Moreover, we found that APC response rates correlated with the thrombotic risk in FVL carriers.7 The study population consisted of 30 healthy individuals and 51 patients with a history of VTE, thereof 28 FVL or FII 20210G>A carriers (FVL/FII 20210G>A cohort), and 23 unrelated subjects with unexplained familial VTE (FH cohort). A diagram of patient recruitment and selection criteria is shown in Figure 1, along with a description of study procedures. Blood samples were drawn before

Figure 1. Eligibility criteria and study procedures. Healthy individuals were recruited from blood donors. Patients with a history of venous thromboembolism (VTE) were recruited from the thrombophilia outpatient clinic of our hospital. The study proposal was approved by the Ethics Committee of the Medical Faculty of the University Bonn (reference number 016/16). Written informed consent was received prior to participation. All finally included study participants (n=81) received morning administration of 15 µg/kg recombinant activated factor VII (rFVIIa) as single intravenous bolus injection after overnight fast. Blood samples were drawn immediately before and 10 minutes (min), 30 min, 1, 2, 3, 5, and 8 hours after administration, each from a new venipuncture. After discarding the first 2 mL, blood was drawn into citrate tubes (10.5 mmol/L, Sarstedt, Nümbrecht, DE). Citrate tubes were supplemented with aprotinin (10 µmol/L) and bivalirudin (250 µg/mL) for activated protein C (APC) measurement. Plasma samples were obtained by centrifugation (2,600 x g, 10 min) within 30 min and stored at less than -70°C until assayed. All finally included study participants completed rFVIIa administration and follow-up blood sampling. All collected samples were analyzed. *Surgery, trauma, immobilization, pregnancy, and puerperium. #Transaminases, g-glutamyl transferase, urea, creatinine in serum. ** Decreased plasma levels of antithrombin, protein C, protein S, anti-cardiolipin and anti-β2 glycoprotein I immunoglobulin G (IgG) and IgM, functional lupus anticoagulants (activated partial thromboplastin time, dilute Russell viper venom time), and factor V Leiden (FVL) and prothrombin (FII) 20210G>A mutation (except for inclusion into the FVL/FII 20210G>A cohort).

haematologica | 2022; 107(5)

1197

Letters to the Editor

Table 1. Baseline characteristics and rFVIIa-induced biomarker changes.

Age, years (range) Sex (male/female) BMI, kg/m2 (range) DVT / PE / both, n Fibrinogen, g/L FII, % Factor XI, % Antithrombin, % sTM, ng/mL sEPCR, ng/mL PC, % F1+2, nmol/L AUC, nmol·h/L TAT, ng/mL AUC, pmol·h/L APC, pmol/L AUC, pmol·h/L

Healthy controls, N = 30

VTE, FVL or FII 20210G>A, N = 28*

VTE, family history of VTE, no RF, N = 23

35 (21-60) 12 / 18 23 (18-27) 252 (221-284) 103 (98-116) 102 (90-115) 107 (100-111) 1.62 (1.30-2.15) 45.6 (26.0-81.6) 106 (97-118) 0.16 (0.12-0.21) 0.29 (0.16-0.45) <21.2 (<21.3-<21.3) 35.9 (0.81-109.5) 0.68 (0.40-1.11) 6.55 (5.22-8.82)

41 (18-60) 12 / 16 24 (18-27) 15 / 3 / 10 262 (250-309) 124 (115-135) 101 (95-107) 98 (93-104) 1.61 (1.47-2.19) 57.0 (35.4-91.0) 112 (103-122) 0.25 (0.17-0.30) 0.34 (0.23-0.49) <21.3 (<21.3-29.7) 123.9 (45.1-188.2) 1.13 (0.75-1.43) 15.1 (10.7-22.7)

10-5 0.002 0.008 0.022 <10-4

38 (20-53) 9 / 14 24 (19-27) 10 / 6 / 7 267 (256-331) 114 (103-120) 115 (100-127) 100 (98-106) 1.62 (1.19-1.86) 72.5 (46,4-108.0) 105 (97-116) 0.15 (0.12-0.20) 0.42 (0.19-0.73) <21.3 (<21.3-24.7) 141.6 (12.8-332.3) 0.79 (0.39-1.11) 9.46 (5.50-14.55)

0.044 0.021 -

Age and body mass index (BMI) are shown as mean (range), all other variables as median (interquartile range). The area under the curve (AUC) quantifies changes of prothrombin activation fragment F1+2 (F1+2), thrombin-antithrombin complex (TAT), and activated protein C (APC) over 8 hours after intravenous injection of recombinant activated factor VII (rFVIIa). P describes significant (<0.05) differences to healthy controls. P was calculated using the unpaired Student t-test (prothrombin, FII; protein C, PC) or the Mann-Whitney test (all other parameters) and corrected for multiple testing using the Bonferroni method. DVT: deep vein thrombosis; FVL: factor V Leiden; PE: pulmonary embolism; sEPCR: soluble endothelial PC receptor; sTM: soluble thrombomodulin; VTE: venous thromboembolism. *14 heterozygous FII 20210G>A carriers, 1 homozygous and 13 heterozygous FVL carriers, thereof 2 with HR2 haplotype..

and during 8 hours after administration of 15 µg/kg rFVIIa. No adverse events were observed. APC was measured using an oligonucleotide-based enzyme capture assay (OECA).8 The thrombin biomarkers prothrombin activation fragment 1+2 (F1+2), thrombin-antithrombin complex (TAT), and other hemostasis parameters were determined using commercially available assays. In vitro thrombin generation was assessed before rFVIIa administration, using the calibrated automated thrombogram (CAT) assay (Thrombinoscope, Maastricht, NL). Table 1 lists demographic features and measurement results of hemostasis parameters in the three cohorts at baseline, and rFVIIa-induced changes of F1+2, TAT, and APC over time, expressed as area under the curve (AUC). Hemostasis parameters at baseline were comparable in FVL and FII 20210G>A carriers (Online Supplementary Table S1). In vitro thrombin formation kinetics were higher in the FH cohort than in FVL/FII 20210G>A carriers and healthy controls, indicated by an elevated endogenous thrombin potential (ETP) (Figure 2A), Additionally, peak thrombin concentration was increased compared with FVL/FII 20210G>A carriers, whereas lag time and time-to-peak did not differ significantly (Online Supplementary Figure S1A to C). The difference in the ETP was more pronounced at 1 pmol/L tissue factor (TF) concentration. This could be explained by higher FXI levels in the FH cohort, which have been shown to affect in vitro thrombin generation at a greater extent at lower TF concentrations.9 In the resting state, plasma levels of F1+2 were slightly increased in the FVL/FII 20210G>A cohort, giving additional evidence of increased thrombin formation. After infusion of rFVIIa, plasma levels of F1+2 (Figure 2B) and TAT (Figure 2C) increased significantly in all three cohorts (peak vs. baseline values, Wilcoxon signedrank test P<0.05 after Bonferroni correction). F1+2 increased in every participant, indicating that rFVIIa acti1198

vates the clotting cascade, resulting in thrombin formation. Every FVL/FII 20210G>A carrier showed an increase of F1+2 and TAT, whereas four subjects in the FH group and seven healthy controls showed an isolated increase of F1+2. This absence of a TAT increase could indicate a comparably lower thrombin formation rate. The most probable explanation of this discrepancy is the longer F1+2 half-life of approximately 2 hours in comparison to the TAT half-life of 44 minutes,10 making F1+2 a more sensitive thrombin generation marker. The in vivo thrombin generation parameters F1+2 AUC and TAT AUC correlated with each other in healthy controls and patients with a history of VTE (Figure 2D). However, they did not correlate with in vitro thrombin generation (representatively shown for ETP and TAT AUC, Online Supplementary Figure S1D and E), suggesting that different factors determine and interfere with the outcome in both distinct and complex methodological approaches. In addition, compared with FVL and FII 20210G>A carriers, a more heterogenous risk profile can be expected in the FH cohort. If the endothelium is intact, the thrombin formation capacity is effectively controlled by APC formation. The extent to which thrombin formation induces an increase in APC might therefore indicate the functionality of the APC-generating pathway in an individual patient and, moreover, modulate the thrombotic potential of increased thrombin formation rates. In order to investigate the reactivity of the PC system to thrombin formation we measured plasma levels of APC. After infusion of rFVIIa, APC increased significantly in all cohorts (Wilcoxon signed-rank test, P<0.05 after Bonferroni correction). Changes in APC (and thrombin biomarkers) did not differ in FVL and FII 20210G>A carriers (Online Supplementary Figure S2). In contrast to thrombin formation rates the APC response was significantly lower in the FH cohort than in the FVL/FII 20210G>A cohort and haematologica | 2022; 107(5)

Letters to the Editor

Figure 2. In vitro thrombin generation and in vivo thrombin-activated protein C response to rFVIIa. In vitro thrombin generation was measured by the calibrated automated thrombogram (CAT, Thrombinoscope, Maastricht, NL) in healthy controls (n=30) and in patients with venous thromboembolism (VTE) with factor V Leiden (FVL) or prothrombin (FII) 20210G>A mutation (n=28), or a family history of VTE without an established risk factor (RF, n=23). Plasma levels of prothrombin activation fragment 1+2 (F1+2), thrombin-antithrombin complex (TAT), and activated protein C (APC) were measured in the same population before (t=0) and after intravenous injection of 15 µg/kg recombinant activated factor VII (rFVIIa). (A) Endogenous thrombin potential (ETP) measured by CAT, presented as median and interquartile range (IQR, boxes), 1.5-fold IQR (whiskers), and outliers (circles). P-values <0.05 (Mann-Whitney test) are shown. (B) F1+2 and (C) TAT in plasma (median, IQR). (D) Area under the F1+2 generation curve (F1+2 AUC) in comparison to TAT AUC. Dotted lines indicate 90th percentiles of F1+2 AUC and TAT AUC, and 10th percentile of F1+2 AUC in healthy controls. (E) APC in plasma (median, IQR). (F) TAT AUC in comparison to APC AUC. Dotted lines indicate 90th percentiles of TAT AUC and APC AUC, and 10th percentile of APC AUC in healthy controls. The red area highlights the absence of a thrombin-related increase of APC in patients with unexplained familial thrombophilia (blue symbols). r: Pearson’s correlation coefficient.

did not differ from healthy controls (Figure 2E). As the APC response is a direct marker of the APC formation capacity of the endothelium, the disproportionately low APC response in relation to the thrombin formation rate indicates an impaired endothelial APC-generating activity in the FH cohort. This relative APC deficiency after coagulation activation would consecutively result in increased thrombin formation. Several data support this conclusion: i) previously, reciprocal and opposite changes of indirect thrombin and PC activation markers were observed in patients with abnormalities of the PC pathway in a basal state;11 ii) in a previous study, asymptomatic FVL carriers showed a higher APC response in the SHAPE approach than those with prior VTE;7 iii) in the present study, thrombin and APC formation rates (TAT AUC and APC AUC) correlated with each other in both FVL/FII 20210G>A carriers and patients with unexplained familial thrombosis, but not in healthy controls (Figure 2F). With seven subjects (25%) in the FVL/FII 20210G>A cohort and six subjects (26%) in the FH cohort, both TAT AUC and APC AUC lay above the 90th percentiles of the healthy controls in similar rates of patients. However, only two individuals (9%) in the FH cohort showed a disproportionately high APC formation rate, as evidenced by an APC AUC (slightly) above and TAT AUC within the 90th percentiles of the healthy controls. In the FVL/FII 20210G>A cohort such a pattern was observed more often (29%), and more distinctively haematologica | 2022; 107(5)

(Figure 2F). Thrombomodulin (TM) and endothelial PC receptor (EPCR) are two main factors that determine the APC formation capacity of the endothelium and variants in both genes have been suggested as thrombotic risk factors.12,13 In order to assess interindividual variations in TM and EPCR, we measured plasma levels of soluble EPCR and TM but did not find significant differences between cohorts. Potential sources of bias or imprecision include the size of the study population, the precision of rFVIIa dosing and times of blood draw, and laboratory analysis. In order to account for these issues, sample size, rFVIIa dosage and blood sampling times were chosen in orientation to previous pharmacokinetic studies on rFVIIa, yielding expected pharmacokinetic results (Online Supplementary Figure S1F).14 The OECA for APC measurement has been extensively assessed.8 Except for sECPR and sTM the other assays were covered by accreditation with the national accreditation body and were performed according to ISO standards. Moreover, the age and sex distribution, and the body mass index were similar in the different subgroups, ruling out a potential confounding effect of these variables. Finally, one might argue that instead of assessing a genetic hypercoagulable state in patients with unexplained familial thrombosis, an effect of the previous VTE may have been measured, as we did not include asymptomatic family members. In conclusion, the data indicate that a dysbalanced 1199

Letters to the Editor

APC response characterized by increased thrombin formation rates and simultaneously decreased APC formation rates contributes to the increased thrombotic risk of patients with familial thrombosis. Further studies are now warranted to elucidate the pathophysiological and genetic basis of the described phenotype. Moreover, the data show that the SHAPE procedure is a useful tool to measure the functionality of the PC pathway, which is helpful to investigate prothrombotic mechanisms in patients with thrombophilia without an established risk factor. Sara Reda,* Nadine Schwarz,* Jens Müller, Johannes Oldenburg, Bernd Pötzsch# and Heiko Rühl# Institute of Experimental Hematology and Transfusion Medicine, University Hospital Bonn, Bonn, Germany * SR and NS contributed equally as co-first authors # BP and HR contributed equally as co-senior authors Correspondence: HEIKO RÜHL - Heiko.Ruehl@ukbonn.de doi:10.3324/haematol.2021.280573 Received: December 21, 2021. Accepted: January 31, 2022. Pre-published: February 10, 2022. Disclosures: BP and JM have a patent DE102007063902B3 including the aptamer HS02-52G binding to APC. An assay for the quantification of APC levels in human plasma, based on this aptamer, has been licensed to ImmBioMed, Pfungstadt, Germany. All other authors declare no conflicts of interest. Contributions: HR, JM, and BP designed the experiments. SR, NS, and HR collected the data, SR and HR analyzed the data. SR, NS, JM, JO, BP, and HR drafted and edited the manuscript. Funding: this work was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – 419450023. HR is recipient of a fellowship from the Stiftung Hämotherapie-Forschung (Hemotherapy Research Foundation). Data-sharing statement: the datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. All authors have complete and on-going access to the study data.

References 1. Reitsma PH, Versteeg HH, Middeldorp S. Mechanistic view of risk

1200

factors for venous thromboembolism. Arterioscler Thromb Vasc Biol. 2012;32(3):563-568. 2. Klarin D, Busenkell E, Judy R, et al. Genome-wide association analysis of venous thromboembolism identifies new risk loci and genetic overlap with arterial vascular disease. Nat Genet. 2019;51(11):15741579. 3. Lindström S, Wang L, Smith EN, et al. Genomic and transcriptomic association studies identify 16 novel susceptibility loci for venous thromboembolism. Blood. 2019;134(19):1645-1657. 4. Zöller B. Genetics of venous thromboembolism revised. Blood. 2019;134(19):1568-1570. 5. Segers O, van Oerle R, ten Cate H, Rosing J, Castoldi E. Thrombin generation as an intermediate phenotype for venous thrombosis. Thromb Haemost. 2010;103(1):114-122. 6. Rühl H, Winterhagen FI, Berens C, Müller J, Oldenburg J, Pötzsch B. In vivo thrombin generation and subsequent APC formation are increased in factor V Leiden carriers. Blood. 2018;131(13):1489-1492. 7. Rühl H, Berens C, Winterhagen FI, et al. Increased activated protein C response rates reduce the thrombotic risk of factor V Leiden carriers but not of prothrombin 20210G>A Carriers. Circ Res. 2019;125(5):523-534. 8. Müller J, Friedrich M, Becher T, et al. Monitoring of plasma levels of activated protein C using a clinically applicable oligonucleotidebased enzyme capture assay. J Thromb Haemost. 2012;10(3):390398. 9. Keularts IM, Zivelin A, Seligsohn U, Hemker HC, Béguin S. The role of factor XI in thrombin generation induced by low concentrations of tissue factor. Thromb Haemost. 2001;85(6):1060-1065. 10. Rühl H, Berens C, Winterhagen A, Müller J, Oldenburg J, Pötzsch B. Label-free kinetic studies of hemostasis-related biomarkers including D-dimer using autologous serum transfusion. PLoS One. 2015;10(12):e0145012. 11. Bauer KA, Broekmans AW, Bertina RM, et al. Hemostatic enzyme generation in the blood of patients with hereditary protein C deficiency. Blood. 1988;71(5):1418-1426. 12. Medina P, Navarro S, Estellés A, et al. Contribution of polymorphisms in the endothelial protein C receptor gene to soluble endothelial protein C receptor and circulating activated protein C levels, and thrombotic risk. Thromb Haemost. 2004;91(5):905-911. 13. Navarro S, Medina P, Bonet E, et al. Association of the thrombomodulin gene c.1418C>T polymorphism with thrombomodulin levels and with venous thrombosis risk. Arterioscler Thromb Vasc Biol. 2013;33(6):1435-1440. 14. Fridberg MJ, Hedner U, Roberts HR, Erhardtsen E. A study of the pharmacokinetics and safety of recombinant activated factor VII in healthy Caucasian and Japanese subjects. Blood Coagul Fibrinolysis. 2005;16(4):259-266.

haematologica | 2022; 107(5)

Letters to the Editor

Erythrocytosis associated with EPAS1(HIF2A), EGLN1(PHD2), VHL, EPOR or BPGM mutations: the Mayo Clinic experience Germline mutations in the oxygen-sensing pathway (VHL-HIF2A-PHD2) or erythropoietin (EPO) signaling (EPOR) are relatively rare but may result in erythrocytosis with normal p50 measurement (oxygen tension at which hemoglobin is 50% saturated) accompanied by either an elevated or inappropriately normal EPO (VHL-HIF2APHD2) or subnormal EPO (EPOR).1 On the other hand, a left shift of the oxygen dissociation curve, with venous p50 <24 mmHg may result from high-oxygen affinity (HOA) hemoglobin variants, defective 2,3-bisphosphoglycerate mutase (BPGM) causing 2,3-BPG deficiency or methemoglobinemia.1 The incidence, clinical course and management of hereditary erythrocytosis has not been well-characterized due to its rare occurrence. In that regard, we recently reported on 41 patients with HOA variant associated erythrocytosis; over half of the patients manifested one or more symptoms thought to be related to increased hematocrit while thrombosis was documented in a quarter of the patients.2 Neither hematocrit level nor active phlebotomy showed significant correlation with either thrombotic or non-thrombotic symptoms, which might have resulted from the limited sample size.2 In a recent study which included 270 patients with idiopathic erythrocytosis, 1.1% harbored EPOR mutations, while pathogenic variants involving genes in the hypoxia pathway were identified in 23% of patients.3 Accordingly, we share the Mayo Clinic clinical and laboratory experience with hereditary erythrocytosis resulting from genetic alterations in the oxygen-sensing pathway (VHL-HIF2A-PHD2), EPOR or BPGM. All patients that underwent hereditary erythrocytosis evaluation at the Mayo Clinic over the last 10 years (2012-2021), were retrospectively recruited after obtaining Institutional Review Board approval. Polycythemia vera was excluded with JAK2 exon 12-15 sequencing. Hereditary erythrocytosis testing was pursued at the Mayo Clinic laboratory utilizing an algorithmic approach which included p50 measurement, serum EPO level (Epo), and DNA sequencing by polymerase chain reaction (PCR) of EPOR (exon 8), hypoxia-inducible factor 2 a (HIF2A) encoded by endothelial PASS domain protein 1 (EPAS1) (exons 9 and 12), prolyl hydroxylase 2(PHD2) encoded by EGL-9 family hypoxia inducible factor 1(EGLN1)(exons 1-5), von Hippel Lindau (VHL) (three coding exons and intron/exon boundaries) and BPGM (exons 1-4) as detailed in our prior work.4 Of 592 patients tested at the Mayo Clinic for HIF2A/PHD2/EPOR alterations, 14 pathogenic variants were identified in HIF2A (n=6, 1%), PHD2 (n=3, 0.5%), EPOR (n=2, 0.3%), while two of 421 (0.5%) and one of 446 (0.2%) patients harbored BPGM and VHL variants, respectively. In addition, 22 variants of uncertain significance (VUS) were reported; EPOR (n=1), HIF2A (n=3), PHD2 (n=10), BPGM (n=2), VHL (n=6), resulting in combined (pathogenic + VUS) Mayo Clinic incidence rates of 0.5%, 1.5%, 2.2%, 1% and 1.6% for EPOR, HIF2A, PHD2, BPGM, and VHL aberrations, respectively. Table 1 summarizes oxygen-sensing pathway (PHD2/HIF2A/VHL) pathogenic variants including clinical course of ten patients with median follow-up of 2 years, (range, 0.2-10 years). HIF2A pathogenic variants were noted in six patients; four harbored the heterozygous HIF2A c.1121T>A, p.(Phe374Tyr) alteration in exon 9, previously reported in association with neuroenhaematologica | 2022; 107(5)

docrine tumors with or without erythrocytosis.5 A 57year-old male with heterozygous HIF2A c.1121T>A mutation presented with a hemoglobin (Hb)/hematocrit (Hct)/Epo of 17.9 g/dL/54.4%/93.4 mIU/mL, diabetes mellitus and prior cerebrovascular accident (CVA)/ left ventricular thrombus, was started on phlebotomy, continued aspirin with anticoagulation and did not experience additional thromboses. The second case was a 56year-old female with heterozygous HIF2A c.1121T>A mutation. Hb/Hct/Epo at presentation; 19.1 g/dL/57%/40.8 mIU/mL, with hypertension and hyperlipidemia, developed multiple thromboses; myocardial infarction, followed by CVA, inferior vena caval thrombus post-diagnosis, the latter occurred despite ongoing phlebotomy and aspirin/clopidogrel. The remainder two patients with heterozygous HIF2A c.1121T>A mutations were 68- and 71-year-old males with hypertension and hyperlipidemia respectively, Hb/Hct/Epo at diagnosis were 19.1/57.2/20.7 and 17.2/52/7.7, both did not experience thrombosis with the former receiving phlebotomy and the latter low dose aspirin. Additionally, a 61-year-old female harbored a heterozygous missense alteration in HIF2A c.1620C>A, resulting in amino acid substitution p.Phe540Leu (F540L) previously reported by our group.4 She had a history of hypertension, presented with Hb/Hct/Epo of 16.1/47.8/7.3 and did not experience thrombosis while on low-dose aspirin. On the other hand, a 69-year-old hypertensive male with heterozygous HIF2A c.1609G>A, mutation with Hb/Hct/Epo of 23/58.7/175 at diagnosis, developed a CVA with ongoing phlebotomy. An elevated Epo level (range, 20.7-175, reference range; 2.6-18.5 mIU/mL) was noted in four of six patients with HIF2A pathogenic variants, which in all instances was accompanied by phlebotomy. All patients had one or more cardiovascular risk factors, with three patients (50%) experiencing thrombosis, two of which occurred with ongoing phlebotomy, suggesting the lack of benefit of phlebotomy. Of three patients with PHD2 pathogenic variants; a 35year-old female with family history of erythrocytosis, current smoker, without history of prior thrombosis, and Hb/Hct/Epo 17.2/52.6/11.2, demonstrated a PHD2 c.1111C>T, p.(Arg371Cys) missense variant. This variant has been reported in the human gene mutation database,6,7 and involves a highly conserved amino acid in the Fe(2+) 2-oxoglutarate dioxygenase domain, critical for hydroxylation of HIF; functional studies have not been performed but studies involving (Arg371His) have shown decreased ability of PHD2 to bind and hydroxylate HIF. On the other hand, two patients harbored previously reported PHD2 c.461C>A, p.(S154*) and c.1030C>T, p.(Arg344*) nonsense variants predicted to result in a premature stop codon in exon 1 and 3, respectively, and expected to be loss of function mutations.4 This included a 67-year-old male with PHD2 c.461C>A and a 60-year-old female with PHD2 c.1030C>T mutation, Hb/Hct/Epo at diagnosis were 17.8/50.7/10.3 and 17/not available/30, both did not experience thrombosis, former had known coronary artery disease and was on low-dose aspirin while the latter was hypertensive and receiving phlebotomy along with aspirin. A pathogenic variant in VHL was detected in a 19-yearold male, compound heterozygous (L188V and R200W) for the previously described VHL mutations,8 who presented with erythrocytosis (Hb/Hct 19/57) and a markedly elevated EPO level at 1465 mIU/mL. He was managed with phlebotomy every 4 weeks, in addition to aspirin and did not experience thrombosis. 1201

Letters to the Editor

Table 1. Clinical features and management of ten patients with EGLN1(PHD2)/ EPAS1(HIF2A)/VHL pathogenic variant associated erythrocytosis.

Patient n/ age at diagnosis/ sex #1 35/F

#2 60/F

#3 67/M

#4 69/M

#5 57/M

#6 56/F

#7 68/M

#8 71/M

#9 61/F

#10 19/M

Gene mutation

Family history

Hb/HcT

EPO

EGLN1(PHD2) Heterozygous c.1111C>T, p.(Arg371Cys)6,7 EGLN1 (PHD2) Heterozygous c.1030C>T, p.(Arg344*)4 EGLN1 (PHD2) Heterozygous c.461C>A, p.(Ser154*)4 EPAS1 (HIF2A) Heterozygous c.1609G>A, p.(Gly537Arg)19 EPAS1 (HIF2A) Heterozygous c.1121T>A, p.(Phe374Tyr)5 EPAS1(HIF2A) Heterozygous c.1121T>A, p.(Phe374Tyr)5 EPAS1(HIF2A) Heterozygous c.1121T>A, p.(Phe374Tyr)5 EPAS1(HIF2A) Heterozygous c.1121T>A, p.(Phe374Tyr)5 EPAS1(HIF2A) Heterozygous c.1620C>A, p.(Phe540Leu)20 VHL Heterozygous 562C>G, p.(Leu188Val) c.598C > T, p.(Arg200Trp)8

Sister

17.2/52.6

11.2

none

17/

Brother x2

17.8/50.7

10.3

none

23/58.7

175

none

17.9/54.4

93.4

none

19.1/57

40.8

none

19.1/57.2

20.7

none

17.2/52

7.7

none

16.1/47.8

7.3

none

19/57

1465

p50

CV risks

Thrombosis (Therapy at event)

Pregnancy

Phlebotomy

Aspirin

Anticoagulation

Smoking

none

2 live births

none

81 mg

none

HTN

none

2 live births

HcT< 42

81 mg

none

CAD

none

81 mg

none

HTN

CVA after diagnosis (phlebotomy)

Every 3 to 4 months HcT < 50

325 mg

none

HcT<45

325 mg

Enoxaparin apixaban

yes

Aspirin 81 mg Plavix 75 mg

warfarin

HcT<50

none

CVA prior to diagnosis LV thrombus (none) 26 HTN MI CVA hyperlipidemia IVC thrombus after diagnosis (phlebotomy, aspirin, Plavix) HTN none

hyperlipidemia

none

81 mg

none

HTN

none

81 mg

none

Every 4 weeks HcT <45

81 mg

none

Hb: hemoglobin; HcT: hematocrit; HTN: hypertension; DM: diabetes mellitus; CVA: cerebrovascular accident; LV: left ventricle; IVC: inferior vena cava; EPO: erythropoietin; p50: oxygen tension at which hemoglobin is 50% saturated.

Canonical exon 8 EPOR c.1316G>A mutations,9 occurred in two patients, 48- and 69-year-old females, with a family history of erythrocytosis, and Hb/Hct/Epo levels of 19.4/56.6/1.1 and 14.6/44.3/<1, respectively, underscoring the suppressed Epo levels with gain of function EPOR mutations (Table 2). Both patients underwent intermittent phlebotomy and had an uncomplicated course in terms of thrombosis and pregnancies. Two patients harbored BPGM pathogenic variants (Table 2) which included a 25-year-old male with hypertension who presented with Hb/Hct/Epo/p50 of 20/58/17.7/31, found to have a heterozygous missense alteration in BPGM at c.184C>T resulting in amino acid substitution p.Arg62Trp (R62W). While this specific

1202

amino acid change is novel, (p.Arg62Gln) has been reported in association with erythrocytosis in patients homozygous for the variant10 and compound heterozygous for Arg62Gln and another BPGM pathogenic variant.11 The second case was a 25-year-old male, current smoker with Hb/Hct/Epo of 17/49.1/5.1, who harbored a previously unreported BPGM c.258dup, p.(Leu87Serfs*3) frameshift variant in the first coding exon, predicted to result in a premature stop codon. Similar nonsense mutations leading to a predicted premature stop codon have been reported.10,12,13 Both patients had an uneventful clinical course, the first patient was receiving phlebotomy and aspirin while the second case was observed. Among 22 VUS that were reported, PHD2 was most

haematologica | 2022; 107(5)

Letters to the Editor

Table 2. Clinical features and management of four patients with EPOR/BPGM pathogenic variant associated erythrocytosis.

Patient n/ Gene age at mutation diagnosis/ sex #1 48/F

#2 69/F

#3 25/M

#4 25/M

Family history

Hb/HcT

EPOR Mother 19.4/56.6 Heterozygous Brother c.1316G>A, p.(Trp439*)9 EPOR Father 14.6/44.3 Heterozygous Son c.1316G>A, Daughter p.(Trp439*)9 BPGM Unknown 20/58 Heterozygous c.184C>T, p.(Arg62Trp)10,11 BPGM none 17/49.1 Heterozygous c.258dup, p.(Leu87Serfs*3)10,12,13

EPO

p50

1.1

CV risks

Thrombosis (therapy at event)

Pregnancy

Phlebotomy

Aspirin Anticoagulation

HTN

none

2 live births

Intermittent HcT< 50

none

2 live births

Every 4 to 8 weeks HcT< 43

81 mg

none

17.7

HTN

none

Every 6 weeks Hb< 14.5

81 mg

none

Smoking

none

5.1

frequently involved (Table 3). The majority (n=17, 77%) of cases were males with median age at diagnosis of 50 years (range, 16-73 years). All patients had normal p50 testing, whereas EPO levels were highly variable, median 8 mIU/mL (range, 3.8-47.7 mIU/mL). A family history of erythrocytosis was known in five patients (23%) and thrombosis occurred in two (9%) of patients; the majority were managed with phlebotomy/blood donation (n=16, 73%) and/or antiplatelet therapy (n=12, 55%). In the current series, we share our decades worth of hereditary erythrocytosis testing experience from the Mayo Clinic in order to define the incidence of alterations involving the hypoxia sensing pathway, in addition to EPOR and BPGM, providing a clinical perspective on the likelihood of encountering such abnormalities during the course of erythrocytosis evaluation. We limited the above series to the hypoxia sensing pathway genes, EPOR, and BPGM, since we have recently published on HOA variant associated erythrocytosis. Of the hypoxia sensing pathway alterations, homozygous VHL (598C>T) mutation Chuvash polycythemia [CP] is phenotypically well-characterized by an unusual propensity for vascular events leading to early mortality.14 In a prospective, age, sex-matched controlled study on the subject matter, age and prior thrombotic events emerged as independent predictors of thrombosis; moreover, phlebotomy was associated with an increased incidence of thrombosis.15 Similarly, among eight patients harboring the HIF2A p.M535V variant, five experienced thrombotic events versus none in 17 HIF2A wild-type patients.15 Furthermore, thrombotic events occurred despite phlebotomy and in the absence of cardiovascular risks.15 In our series, all three thrombotic events occurred in patients harboring HIF2A pathogenic variants, two of which were receiving phlebotomy, in addition to dual antiplatelet therapy in one patient. Of note, HIF2A alterations may be associated with neuroendocrine tumors such as pheochromocytoma, paraganglioma, somatostatinoma;16 however, none of our patients with HIF2A alterations developed tumors. Limitations of our study

haematologica | 2022; 107(5)

Table 3. Clinical features and management of 22 patients with variants of uncertain significance involving EPOR/EGLN1(PHD2)/ EPAS1(HIF2A)/VHL/BPGM and associated erythrocytosis.

Variables Gene, n EPOR EGLN1 (PHD2) EPAS1 (HIF2A) VHL BPGM Age in years, median (range) Male sex, n (%) Hemoglobin g/dL, median (range) Hematocrit %, median (range) Serum erythropoietin mIU/mL, median (range) Reference range, 2.6-18.5 mIU/mL p50 mm Hg, median (range) Cardiovascular risk factors, n (%) Family history, n (%) Thrombosisa, n (%) Major arterial thrombosis Major venous thrombosis Treatment, n (%) Phlebotomy/blood donation Antiplatelet therapy (aspirin or clopidogrel) Anticoagulation

N=22 1 10 3 6 2 50 (16-73) 17 (77) 18.2 (16-20.7) 53.4 (48.5-60) 8 (3.8-47.7) n=19 25 (24-29) 16 (73) 5(23) 2(9) 0 2 16 (73) 12 (55) 4 (18)

a. Major venous thrombosis included deep vein thrombosis, pulmonary embolism. EPOR- c.1310G>A, p.Arg437His, EGLN1 (PHD2)- c.826A>G, p.Met276Val, c.165G>C, p.Lys55Asn, c.709G>C, p.Asp237His, c.280A>G, p.Arg94Gly, c.1016G>C, p.Ser339Thr, c.112A>G, p.Ser38Gly , c.289G>A, p.Ala97Thr, c.586G>C, p.Glu196Gln, c.604A>G, p.Met202Val, c.*92G>A, single nucleotide substitution, EPAS1 (HIF2A)- c.1958C>T, p.Ala653Val, c.1556C>T, p.Thr519Met, c.1834G>A, p.Gly612Arg VHL- c.-61_51het_dup11 (g.10183420), c.241C>T, p.P81S, c.134C>G ,p.45R, c.167C>G, p.Ala56Gly, c.345C>T, p.H155H, c.599G>A, p.Arg200Gln (heterozygous), BPGM- c.115C>T, p.Arg39Trp, c.289G>C, p.Gly97Arg. EPOR: erythropoietin receptor; EGLN1 (PHD2): EGL-9 family hypoxia inducible factor 1 (prolyl hydroxylase 2); EPAS1 (HIF2A): endothelial PASS domain protein 1(hypoxia-inducible factor 2 ); VHL: von Hippel Lindau; BPGM: 2,3-bisphosphoglycerate mutase.

1203

Letters to the Editor

include the retrospective design, and heterogeneity in clinical practice in regard to diagnosis and management. In summary, we confirm the infrequent (0.5-2.2%) occurrence of genetic alterations involving the hypoxia pathway, EPOR and BPGM among patients undergoing hereditary erythrocytosis evaluation at the Mayo Clinic which includes testing for all congenital mutations except recently described EPO and iron-responsive element binding protein 1 (IRP1) mutations.17,18 Additionally, phenotypic correlations and management details are provided which may serve as a useful guide for clinicians. Naseema Gangat,1 Jennifer L. Oliveira,2 Tavanna R Porter,2 James D. Hoyer,2 Aref Al-Kali,1 Mrinal M Patnaik,1 Animesh Pardanani1 and Ayalew Tefferi1 1 Division of Hematology, Department of Internal Medicine, Mayo Clinic, Rochester and 2Division of Hematopathology, Department of Laboratory Medicine, Mayo Clinic, Rochester, MN, USA. Correspondence: NASEEMA GANGAT - gangat.naseema@mayo.edu AYALEW TEFFERI - tefferi.ayalew@mayo.edu doi:10.3324/haematol.2021.280516 Received: December 14, 2021. Accepted: February 2, 2022. Pre-published: February 10, 2022. Disclosures: no conflicts of interest to disclose. Contributions: NG and AT designed the study, collected data, performed analyses and wrote the letter; JLO, TRP and JDH reviewed and interpreted all sequencing data; AA, MMP and AP contributed patients. Data-sharing statement: data can be obtained by email (gangat.naseema@mayo.edu)

References 1. Gangat N, Szuber N, Pardanani A, Tefferi A. JAK2 unmutated erythrocytosis: current diagnostic approach and therapeutic views. Leukemia. 2021;35(8):2166-2181. 2. Gangat N, Oliveira JL, Hoyer JD, Patnaik MM, Pardanani A, Tefferi A. High-oxygen-affinity hemoglobinopathy-associated erythrocytosis: Clinical outcomes and impact of therapy in 41 cases. Am J Hematol. 2021;96(12):1647-1654. 3. Filser M, Aral B, Airaud F, et al. Low incidence of EPOR mutations in idiopathic erythrocytosis. Haematologica. 2021;106(1):299-301.

1204

4. Oliveira JL, Coon LM, Frederick LA, et al. Genotype-phenotype correlation of hereditary erythrocytosis mutations, a single center experience. Am J Hematol. 2018;93(8):1029-1041. 5. Lorenzo FR, Yang C, Ng Tang Fui M, et al. A novel EPAS1/HIF2A germline mutation in a congenital polycythemia with paraganglioma. J Mol Med (Berl). 2013;91(4):507-512. 6. Wilson R, Syed N, Shah P. Erythrocytosis due to PHD2 mutations: a review of clinical presentation, diagnosis, and genetics. Case Rep Hematol. 2016;2016:6373706. 7. Gardie B, Percy MJ, Hoogewijs D, et al. The role of PHD2 mutations in the pathogenesis of erythrocytosis. Hypoxia (Auckl). 2014;2:7190. 8. Bento C, Percy MJ, Gardie B, et al. Genetic basis of congenital erythrocytosis: mutation update and online databases. Hum Mutat. 2014;35(1):15-26. 9. de la Chapelle A, Traskelin AL, Juvonen E. Truncated erythropoietin receptor causes dominantly inherited benign human erythrocytosis. Proc Natl Acad Sci U S A. 1993;90(10):4495-4499. 10. Hoyer JD, Allen SL, Beutler E, Kubik K, West C, Fairbanks VF. Erythrocytosis due to bisphosphoglycerate mutase deficiency with concurrent glucose-6-phosphate dehydrogenase (G-6-PD) deficiency. Am J Hematol. 2004;75(4):205-208. 11. Lemarchandel V, Joulin V, Valentin C, et al. Compound heterozygosity in a complete erythrocyte bisphosphoglycerate mutase deficiency. Blood. 1992;80(10):2643-2649. 12. Petousi N, Copley RR, Lappin TR, et al. Erythrocytosis associated with a novel missense mutation in the BPGM gene. Haematologica. 2014;99(10):e201-204. 13. Camps C, Petousi N, Bento C, et al. Gene panel sequencing improves the diagnostic work-up of patients with idiopathic erythrocytosis and identifies new mutations. Haematologica. 2016;101(11):13061318. 14. Gordeuk VR, Sergueeva AI, Miasnikova GY, et al. Congenital disorder of oxygen sensing: association of the homozygous Chuvash polycythemia VHL mutation with thrombosis and vascular abnormalities but not tumors. Blood. 2004;103(10):3924-3932. 15. Gordeuk VR, Miasnikova GY, Sergueeva AI, et al. Thrombotic risk in congenital erythrocytosis due to up-regulated hypoxia sensing is not associated with elevated hematocrit. Haematologica. 2020;105(3):e87-e90. 16. Tarade D, Robinson CM, Lee JE, Ohh M. HIF-2alpha-pVHL complex reveals broad genotype-phenotype correlations in HIF-2alpha-driven disease. Nat Commun. 2018;9(1):3359. 17. Zmajkovic J, Lundberg P, Nienhold R, et al. A gain-of-function mutation in EPO in familial erythrocytosis. N Engl J Med. 2018;378(10):924-930. 18. Oskarsson GR, Oddsson A, Magnusson MK, et al. Predicted loss and gain of function mutations in ACO1 are associated with erythropoiesis. Commun Biol. 2020;3(1):189. 19. Percy MJ, Beer PA, Campbell G, et al. Novel exon 12 mutations in the HIF2A gene associated with erythrocytosis. Blood. 2008;111(11):5400-5402. 20. Percy MJ, Chung YJ, Harrison C, et al. Two new mutations in the HIF2A gene associated with erythrocytosis. Am J Hematol. 2012;87(4):439-442.

haematologica | 2022; 107(5)

Letters to the Editor

trials gov. Identifier: NCT00557193) was a phase III clinical trial for infants with newly diagnosed ALL with or without KMT2A-r.8 The trial was approved by Institutional Review Boards at participating COG member institutions and conducted in accordance with the Declaration of Helsinki. Informed consent was obtained from the parents or guardians according to federal and local regulations. The primary aim of AALL0631 was to test the safety and efficacy of the addition of the FLT3 inhibitor, lestaurtinib, to chemotherapy for infants with KMT2A-r.8 Infants with KMT2A-g were treated on a chemotherapy only arm, with a modified Interfant-99 based induction3,9 followed by modified COG P94079 therapy, with continuation therapy extending to 2 years from diagnosis (as compared to 46 weeks in the predecessor P9407 trial). The trial enrolled 64 infants with KMT2A-g and resulted in survival outcomes superior to those reported for infants with KMT2A-g in all prior COG and Interfant trials. Infants less than 366 days of age with a new diagnosis of B- or T-cell lineage ALL, acute undifferentiated leukemia (AUL), or mixed phenotype acute leukemia (MPAL) with predominantly lymphoid morphology and immunophenotype were eligible to enroll. Neonates less than 4 weeks old and at least 36 weeks gestational age at the time of diagnosis were eligible. Patients with Down syndrome, mature B-cell leukemia, acute myeloid

Outstanding outcomes in infants with KMT2Agermline acute lymphoblastic leukemia treated with chemotherapy alone: results of the Children’s Oncology Group AALL0631 trial Among infants with acute lymphoblastic leukemia (ALL), approximately 70-75% have KMT2A rearrangement (KMT2A-r) and 25-30% have KMT2A-germline (KMT2A-g) leukemia. Event-free (EFS) and overall survival (OS) for KMT2A-g infant ALL are significantly better than those of KMT2A-r infant ALL, but inferior to outcomes in older children with ALL. Aside from the absence of KMT2A-r, the well-defined prognostic factors in older children with B-ALL (age, initial white blood cell [WBC] count, cytogenetics) are not clearly established, as KMT2A-g infant ALL accounts for only ~1% of childhood ALL. Pediatric cooperative group trials conducted between 1996 and 2016 shown that the 4-6-year EFS/OS for infants with KMT2A-g ALL have ranged from 60-74% and 75-87%, respectively.1-4 Although the Japanese Infant Leukemia Study Group and the Japanese Pediatric Leukemia/Lymphoma Study Group reported remarkable 5-year EFS of 96% and 93% in KMT2A-g infant ALL in MLL96/98 and MLL-10, respectively, the cohort sizes were small, the sex ratios were skewed and the results have not been replicated in other cooperative groups.5-7 Children’s Oncology Group (COG) AALL0631 (clinical-

Table 1. Patient characteristics.

KMT2A-g <1 yr AALL0631 Total 64 Age at diagnosis Median age (range) 281 d (54 to 363 d) Sex Male:Female 1:1.1 Race White 53 (90%) Black or African American 5 (8%) Asian 1 (2%) American Indian 0 Native Hawaiian 0 Unknown 5 Ethnicity Hispanic or Latino 15 (24%) Not Hispanic or Latino 47 (76%) Unknown 2 WBC count at diagnosis, x109/L Median (range) 38 (0.6 to 918.2) Diagnosis B-lymphoblastic leukemia 58 (91%) T-lymphoblastic leukemia 6 (9%) AUL or indeterminate 0 MPAL 0 Unknown 0 CNS status CNS 1 41 (65%) CNS 2 16 (25%) CNS 3 6 (10%) Unknown 1

KMT2A-r <1 yr AALL0631

146

KMT2A-g <1 yr P9407

KMT2A-g 1-9 yr AALL03B1 + AALL08B1

19,047

169 d (0 to 360 d)

<.0001

291 d (17 to 365 d)

0.77

4.08 yr (1 to 9.98 yr)

1:1.4

0.45

1:0.4

0.03

1:0.8

0.31

108 (83%) 11 (8%) 8 (6%) 3 (2%) 0 16

0.45

25 (81%) 4 (13%) 2 (6%) 0 0 4

0.34

13,986 (85.6%) 1,174 (7.2%) 908 (5.6%) 167 (1.0%) 93 (0.6%) 2719

0.58

34 (24%) 106 (76%) 6

1.00

6 (19%) 26 (81%) 3

0.61

4,132 (22.8%) 13,989 (77.2%) 926

0.76

160 (1.6 to 4,334.5)

<.0001

87.1 (2.5 to 540)

0.28

9.9 (0.1 to 5,784)

<.0001

139 (96%) 0 1 (1%) 5 (3%) 1

0.0007

32 (91%) 3 (9%) 0 0 0

1.0

15,424 (92.1%) 1,282 (7.7%) 15 (0.1%) 32 (0.2%) 2,294

0.73

61 (42%) 58 (40%) 25 (17%) 2

0.01

20 (57%) 13 (37%) 2 (6%) 0

0.51

16,426 (88.3%) 1,895 (10.2%) 280 (1.5%) 446

<.0001

KMT2A-g, KMT2A-germline; KMT2A-r: KMT2A-rearranged; yr: year; d: days; WBC: white blood cell; L: liter; AUL: acute undifferentiated leukemia; MPAL: mixed phenotype acute leukemia; CNS: central nervous system.

haematologica | 2022; 107(5)

1205

Letters to the Editor

leukemia, or who had received prior anti-leukemic therapy (with the exception of limited corticosteroids or intrathecal chemotherapy) were ineligible. All patients had karyotypes and fluorescence in situ hybridization (FISH) to determine KMT2A status performed in COGapproved laboratories, with central review of results (AJC and NAH). Informative KMT2A FISH data were required to continue on AALL0631 post-induction therapy. AALL0631 opened to accrual in January 2008 and the original COG P9407-based induction regimen (cohort 1) resulted in excessive toxic mortality due to infections (4 of 26 patients, 15.4%).9 The study was temporarily closed to accrual in November 2008 and amended to substitute an Interfant-99 based induction3 with additional supportive care guidelines (cohort 2). This led to significantly less induction mortality and maintained complete remission rates.9 Post-induction, infants with KMT2A-g were non-randomly assigned to the Standard Risk (SR) arm (Online Supplementary Table S1). The study met accrual goals and closed to enrollment in June 2014. Data as of March 31, 2019, are included in this report. Median follow-up was 6.3 years. EFS was defined as the time from study entry to first event (treatment failure, relapse, second malignant neoplasm [SMN], or death) or censored at date of last contact. OS was defined as the time from study entry to death or censored at last contact. Estimates of EFS and OS were calculated using the Kaplan-Meier method with standard errors (SE) using Peto’s formula.10,11 Two-sided log-rank tests were used to compare survival between curves. Fisher’s exact tests were used to compare proportions and Wilcoxon ranksum tests were used to compare distributions of continuous measures. Statistical significance was defined as Pvalue less than 0.05. AALL0631 enrolled 210 eligible patients, including 64 (30.5%) with KMT2A-g (4 in cohort 1 and 60 in cohort 2). Patient characteristics were compared to infants with KMT2A-r in AALL0631, infants with KMT2A-g in P9407, and children age 1-9 years with KMT2A-g ALL and without Down Syndrome or Philadelphia chromosome-positive ALL, enrolled on the COG ALL classification studies AALL03B1 and AALL08B1 from March 1, 2004 to July 20, 2018, using frozen data from December 31, 2020 (Table 1). Notable differences in comparison to infants with KMT2A-r include older age and lower WBC count at diagnosis. Among infants with KMT2A-g, the proportion of females was higher in AALL0631 than in P9407. Central nervous system (CNS) leukemia was more common in infants than in children age 1-9 years with KMT2A-g, but less frequent than in infants with KMT2Ar (Table 2). The cytogenetic findings for infants with KMT2A-g are listed in the Online Supplementary Table S2. Among 64 infants with KMT2A-g, 62 were evaluable for morphologic remission at the end of induction (week 6). One patient was removed from protocol prior to postinduction evaluation of remission due to withdrawal of consent and one did not have bone marrow morphology evaluated due to an administrative error. Of the 62 patients with marrow assessed, 55 (89%) achieved remission and seven (11%) did not achieve remission (all had ≥5% marrow blasts). Four patients went off protocol prior to post-induction therapy, one each for: withdrawal of consent, physician preference, family preference, and severe adverse event (cerebral edema). All remaining 60 patients (100%) achieved remission by the end of induction intensification (week 10). There were no treatment failure events. The 5-year EFS (±SE) was 87.3 ± 4.7% and 5-year OS (±SE) was 93.6 ± 3.5% for infants with KMT2A-g. There 1206

Table 2. Univariate analysis of prognostic factors in infants with KMT2A-g leukemia.

Sex Female Male Age at diagnosis <6 months ≥6 months WBC count (cells per L) <50x109 ≥50x109 WBC count (cells per L) <300x109 ≥300x109 Diagnosis B-lymphoblastic leukemia T-lymphoblastic leukemia Cytogenetics Normal diploid Abnormal Unknown CNS status CNS 1 CNS 2 CNS 3 Unknown

5-year EFS (SE)

Estimated hazard ratio

33 31

96.9% (3.3%) 77.4% (8.9%)

ref 4.39

0.045

12 52

81.8% (13.2%) 88.5% (4.9%)

ref 0.72

0.68

37 27

91.7% (5.1%) 81.5% (8.5%)

ref 1.95

0.32

57 7

85.7% (5.3%) 100%

0.29

58 6

86.0% (5.2%) 100%

0.31

17 36 11

87.5% (9.8%) 91.7% (5.1%)

ref 0.48

0.36

41 16 6 1

92.5% (4.6%) 75.0% (11.3%) 83.3% (19.6%)

ref 4.23 2.13

0.10

KMT2A-g: KMT2A-germline; EFS: event-free survival; SE: standard error; WBC: white blood cell; L: liter; CNS: central nervous system; MRD: minimal residual disease.

were no deaths as first events. Eight infants relapsed (5 bone marrow and 3 isolated CNS). All relapses occurred within the first 3 years after diagnosis; five during continuation chemotherapy and three within 12 months after completion of continuation therapy. One infant developed a SMN (mucoepidermoid carcinoma) during the 5year follow-up period after the completion of chemotherapy. The relapse pattern was similar to that of P9407, which recorded five relapses, two marrow and three isolated extramedullary (1 subcutaneous, 1 CNS, 1 testicular), in 35 infants with KMT2A-g. The Kaplan-Meier survival curves for OS and EFS for the overall cohort and EFS curves for subgroups by sex, age < or ≥6 months at diagnosis, and WBC count < or ≥50,000 cells/µL at diagnosis are shown in Figure 1. The 5-year EFS among girls was superior to that of boys (96.9 ± 3.3% vs. 77.4 ± 8.9%, P=0.045; estimated hazard ratio: 4.4). In univariate analyses, age <6 months, WBC count ≥50,000 cells/µL, WBC count ≥300,000 cells/µL, B-cell vs. T-cell phenotype, normal diploid vs. abnormal cytogenetics, and CNS classification were not prognostic of 5-year EFS (Table 2). In cases with karyotypic data (n=53), the most frequent recurrent cytogenetic abnormalities involved chromosome 9p (10 patients, 19%) and t(1;19)(q23;p13.3) or 19p13.3 variant (5 patients, 9%). The recurrent chromosome abnormality dic(9;20)(p13.2;q11) was identified in five of 53 cases (9%). Translocation of chromosome 5p15 with chromosome 15 was observed in two cases. Abnormalities of chromosome bands 15q11-15 have previously been identified in cases of infant ALL, occur in 1% of pediatric ALL overall, and may indicate a favorable prognosis, if NUTM1 fusion is involved.12-14 Molecular studies for PAX5 and NUTM1 rearrangements were not performed in AALL0631, and the prognostic significance haematologica | 2022; 107(5)

Letters to the Editor

Figure 1. Kaplan-Meier survival curves for infants with KMT2A-germline leukemia. (A) Event-free survival (EFS) and overall survival (OS) for the standard risk group, KMT2A-germline. (B) EFS was lower for males vs. females. (C) There were no differences in EFS for infants < vs. ≥6 months at diagnosis or (D) with white blood cell (WBC) count < vs. ≥50,000/µL at diagnosis.

of the cytogenetic findings in KMT2A-g cases could not be determined, given the sample size and rarity of events. The most common reported toxicities were infectious, gastrointestinal, metabolism/nutrition disorders and neutropenia. Grade 3 or 4 infections were reported for 20% or greater of infants during each chemotherapy course and were observed in approximately 40% of infants during each of the continuation phases (Online Supplementary Table S3). Gastrointestinal toxicities were reported most often in the induction intensification and consolidation phases, the two phases containing high-dose methotrexate. Neurologic, respiratory, skin, cardiac, and vascular toxicities were less commonly reported. The toxicities were comparable to those observed in prior infant ALL trials, with notably fewer toxic deaths than P9407, which resulted in five deaths as first events.1,3,5 The high dose intensity of AALL0631, similar to that of P9407, and the extended duration of AALL0631 therapy may both have contributed to the observed excellent outcomes for infants with KMT2A-g. AALL0631, in comparison to the standard arm of the contemporary Interfant06 trial, gave considerably higher doses of chemotherapy. Considering age-based dose reductions in Interfant-06, the differences in cumulative chemotherapy doses were greatest in the youngest infants. Though well tolerated, the chemotherapy intensity of AALL0631 is also higher than that given to older children with KMT2A-g ALL on COG trials. The optimal therapy that will minimize toxicity risks and achieve superior survival for this very rare subset of pediatric ALL patients has yet to be defined. Future trials could consider the incorporation of targeted haematologica | 2022; 107(5)

immunotherapy agents and prioritize the identification of prognostic factors that will enable some infants with KMT2A-g ALL to be treated less intensively. Erin M. Guest,1 John A. Kairalla,2 Joanne M. Hilden,3 ZoAnn E. Dreyer,4 Andrew J. Carroll,5 Nyla A. Heerema,6 Cindy Y. Wang,2 Meenakshi Devidas,7 Lia Gore,3 Wanda L. Salzer,8 Naomi J. Winick,9 William L. Carroll,10 Elizabeth A. Raetz,10 Michael Borowitz,11 Mignon L. Loh,12 Stephen P. Hunger13 and Patrick A. Brown14 1 Division of Hematology/Oncology/Blood and Marrow Transplantation, Children’s Mercy Kansas City, Kansas City, MO; 2 Department of Biostatistics, Colleges of Medicine, Public Health & Health Professions, University of Florida, Gainesville, FL; 3Center for Cancer and Blood Disorders, Children’s Hospital Colorado, Aurora, CO; 4Texas Children’s Hospital, Houston, TX; 5Department of Genetics, University of Alabama at Birmingham, Birmingham, AL; 6 The Ohio State University Comprehensive Cancer Center, Columbus, OH; 7Department of Global Pediatric Medicine, St. Jude Children's Research Hospital, Memphis, TN; 8U.S. Army Medical Research and Materiel Command, Fort Detrick, MD; 9Division of Pediatric Hematology/Oncology, University of Texas Southwestern School of Medicine, Dallas, TX; 10Department of Pediatrics and Perlmutter Cancer Center, NYU Langone Health, New York, NY; 11 Departments of Pathology and Oncology, Johns Hopkins University, Baltimore, MD; 12Department of Pediatrics, Benioff Children's Hospital in the Helen Diller Family Comprehensive Cancer Center, University of California, San Francisco, CA; 13Department of Pediatrics and the Center for Childhood Cancer Research, Children's Hospital of Philadelphia and the Perelman School of Medicine at the University of 1207

Letters to the Editor

Pennsylvania, Philadelphia, PA and 14Division of Pediatric Oncology, Sidney Kimmel Comprehensive Cancer Center, Johns Hopkins University, Baltimore, MD, USA Correspondence: ERIN GUEST - eguest@cmh.edu doi:10.3324/haematol.2021.280146 Received: October 4, 2021. Accepted: February 4, 2022. Pre-published: February 17, 2022. Disclosures: the authors declare no competing financial interests in relation to the work described. EMG received consulting fees from Syndax and is a member of a Speakers Bureau for Jazz. LG provides unpaid Scientific Advisory Board service to Amgen, Janssen, Kura, OnKure, Pfizer, and Syndax, and owns common stock in Amgen, Deciphera, ITOS, Mirati, and Sanofi Paris. EAR received research funding from Pfizer (institutional) and serves on a DSMB for Celgene. SPH received consulting fees from Novartis, honoraria from Amgen, Jazz, and Servier, and owns common stock in Amgen. Contributions: the study was designed by EMG, JAK, MLL, SPH, and PAB; the statistical design and analyses were performed by JAK, CYW, and MD; the cytogenetics data was provided by AJC and NAH; EMG wrote the manuscript, with contributions from all authors. All authors gave final approval of the manuscript. Acknowledgments: LG is the Ergen Family Chair in Pediatric Oncology at Children’s Hospital Colorado. MLL is the Benioff Chair of Children’s Health and the Deborah and Arthur Ablin Endowed Chair for Pediatric Molecular Oncology at Benioff Children’s Hospital. EAR is a KiDS of NYU Foundation Professor at NYU Langone Health. SPH is the Jeffrey E. Perelman Distinguished Chair in Pediatrics at The Children’s Hospital of Philadelphia. Funding: this study was supported by NIH grants U10 CA98543 (COG Chair’s Grant), U10 CA180886 (NCTN Operations Center Grant), U10 CA98413 and U10 CA180899 (COG Statistics and Data Center Grants), and St. Baldrick’s Foundation. Data sharing statement: requests for access to COG protocol research data should be sent to: datarequest@childrensoncologygroup.org Disclaimer: the content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

1208

References 1. Dreyer ZE, Hilden JM, Jones TL, et al. Intensified chemotherapy without SCT in infant ALL: results from COG P9407 (Cohort 3). Pediatr Blood Cancer. 2015;62(3):419-426. 2. Pieters R, De Lorenzo P, Ancliffe P, et al. Outcome of infants younger than 1 year with acute lymphoblastic leukemia treated with the Interfant-06 protocol: results from an international phase III randomized study. J Clin Oncol. 2019;37(25):2246-2256. 3. Pieters R, Schrappe M, De Lorenzo P, et al. A treatment protocol for infants younger than 1 year with acute lymphoblastic leukaemia (Interfant-99): an observational study and a multicentre randomised trial. Lancet. 2007;370(9583):240-250. 4. Hilden JM, Dinndorf PA, Meerbaum SO, et al. Analysis of prognostic factors of acute lymphoblastic leukemia in infants: report on CCG 1953 from the Children's Oncology Group. Blood. 2006;108(2):441451. 5. Tomizawa D, Miyamura T, Imamura T, et al. A risk-stratified therapy for infants with acute lymphoblastic leukemia: a report from the JPLSG MLL-10 trial. Blood. 2020;136(16):1813-1823. 6. Tomizawa D, Koh K, Sato T, et al. Outcome of risk-based therapy for infant acute lymphoblastic leukemia with or without an MLL gene rearrangement, with emphasis on late effects: a final report of two consecutive studies, MLL96 and MLL98, of the Japan Infant Leukemia Study Group. Leukemia. 2007;21(11):2258-2263. 7. Nagayama J, Tomizawa D, Koh K, et al. Infants with acute lymphoblastic leukemia and a germline MLL gene are highly curable with use of chemotherapy alone: results from the Japan Infant Leukemia Study Group. Blood. 2006;107(12):4663-4665. 8. Brown PA, Kairalla JA, Hilden JM, et al. FLT3 inhibitor lestaurtinib plus chemotherapy for newly diagnosed KMT2A-rearranged infant acute lymphoblastic leukemia: Children's Oncology Group trial AALL0631. Leukemia. 2021;35(5):1279-1290. 9. Salzer WL, Jones TL, Devidas M, et al. Decreased induction morbidity and mortality following modification to induction therapy in infants with acute lymphoblastic leukemia enrolled on AALL0631: a report from the Children's Oncology Group. Pediatr Blood Cancer. 2015;62(3):414-418. 10. Kaplan E, Meier, P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53:457-481. 11. Peto R, Pike MC, Armitage P, et al. Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II. analysis and examples. Br J Cancer. 1977;35(1):1-39. 12. Fazio G, Bardini M, De Lorenzo P, et al. Recurrent genetic fusions redefine MLL germ line acute lymphoblastic leukemia in infants. Blood. 2021;137(14):1980-1984. 13. Boer JM, Valsecchi MG, Hormann FM, et al. Favorable outcome of NUTM1-rearranged infant and pediatric B cell precursor acute lymphoblastic leukemia in a collaborative international study. Leukemia. 2021;35(10):2978-2982. 14. De Lorenzo P, Moorman AV, Pieters R, et al. Cytogenetics and outcome of infants with acute lymphoblastic leukemia and absence of MLL rearrangements. Leukemia. 2014;28(2):428-430.

haematologica | 2022; 107(5)

Letters to the Editor

Mitochondrial ATP generation in stimulated platelets is essential for granule secretion but dispensable for aggregation and procoagulant activity Platelets, although anucleate, contain well-coupled and functional mitochondria1,2 and are capable of mitochondrial oxidative phosphorylation (OXPHOS).3 Simultaneous inhibition of glycolysis and OXPHOS abolishes agonistinduced aggregation and granule secretion.3 Recent studies have shown that there is considerable plasticity in energy metabolism of platelets.1 Yet, we demonstrated a switch to aerobic glycolysis from OXPHOS in stimulated platelets.2 Small-molecules that divert the flux from aerobic glycolysis towards OXPHOS prevent platelet activation.2 In addition, several studies including our own have established that inhibitors of OXPHOS alone do not compromise platelet aggregation.1,2,4 These observations led us to question whether mitochondrial ATP generation is dispensable for platelet function. We undertook a comprehensive study of the effect of mitochondrial inhibitors (antimycin, oligomycin) and an uncoupler (CCCP; carbonyl cyanide 3-chlorophenylhydrazone) on platelet activity. We also established the probable mechanistic basis underlying their effects on platelet function. The underpinning modes of action for these molecules differ considerably. Antimycin acts by inhibiting complex III of the electron transport chain, thereby preventing both oxidation of energy substrates (Online Supplementary Figure S1A) as well as ATP synthesis. Oligomycin targets FoF1 ATPase (complex V) leading to a block in ATP generation although oxidation of fuels continues to the extent allowed by proton leak (Online Supplementary Figure S1A). CCCP uncouples mitochondrial respiration and OXPHOS by dissipating the proton gradient (Online Supplementary Figure S1B, C) thus permitting maximal oxidation of fuels (Online Supplementary Figure S1A) albeit compromising formation of ATP. Hence, it is reasonable to expect that the effects of these molecules will not be identical. Since only oligomycin specifically and uniquely inhibits mitochondrial ATP generation, our inferences about the role of mitochondrial ATP in platelet function are based primarily on observations with oligomycin. Platelet aggregation induced by either thrombin (Figure 1A, B) or collagen (Online Supplementary Figure S1D, G) was profoundly impaired in the presence of CCCP while the presence of antimycin or oligomycin did not appear to have a significant effect, which was consistent with earlier reports.1,2,4 Furthermore, we found similar results with aspirinated platelets (Online Supplementary Figure S1E, I), suggesting that the effect of CCCP is likely independent of thromboxane A2 generation. Aggregation is mediated by fibrinogen which binds with high affinity to platelet surface integrin aIIbβ3 in an active conformation. In keeping with our observation on platelet aggregation, CCCP brought about a significant drop in thrombininduced aIIbβ3 integrin activation while antimycin and oligomycin had no such influence (Figure 1D, G). A fraction of stimulated platelets, characterized by expression of phosphatidylserine on the outer leaflet of the plasma membrane, provide a procoagulant surface for generation of a fibrin clot at the site of vessel injury.5 The extent of phosphatidylserine exposure (Figure 1E, H) upon thrombin challenge was significantly retarded by CCCP but unaffected by either antimycin or oligomycin. These observations led us to deduce that mitochondrial ATP generation is dispensable for platelet integrin activation, aggregation and procoagulant activity. Nevertheless, haematologica | 2022; 107(5)

disruption of mitochondrial membrane potential by CCCP impeded these processes. Platelet responses to agonist stimulation are energydemanding.3,6 However, the energy requirements differ considerably, with shape change being the least energyrequiring process, followed by aggregation, dense/a granule release and acid hydrolase secretion, in increasing order.3,7 We reasoned that inhibition of mitochondrial ATP generation could affect greater energy-intensive processes, such as exocytosis of granule contents. Release of adenine nucleotides and surface expression of P-selectin are markers of dense and a granule secretion, respectively. ATP release from platelets stimulated with either thrombin (Figure 1A, C) or collagen (Online Supplementary Figure S1D, H), as well as P-selectin externalization induced by thrombin (Figure 1F, I) were significantly compromised in the presence of oligomycin and CCCP. Antimycin had no significant effect on either dense or a granule secretion (Figure 1A, C, F, I; Online Supplementary Figure S1D, H), which was indicative of greater compensation by glycolytic ATP generated in the presence of antimycin (Figure 3C). Glycolysis rate is enhanced in antimycin-treated platelets in order to compensate for significant inhibition in oxidation of fuels.4 Contrasting this, either residual oxidation allowed in mitochondria of oligomycin-treated platelets, or maximal oxidation provoked by CCCP, could restrict the compensatory increase in glycolysis and ensuing ATP availability, thus failing to sustain platelet functions with higher energy cost, such as granule secretion. Expression of P-selectin on the surface of activated platelets triggers its interaction with the counter-ligand Pselectin glycoprotein ligand-1 (PSGL-1) present on leukocytes.8 Hence, we measured platelet-leukocyte aggregates in whole blood challenged with thrombin receptor-activating peptide (TRAP)-6 by flow cytometry. Both oligomycin and CCCP inhibited TRAP-stimulated platelet-neutrophil interaction (Online Supplementary Figure S2A-E, K), which could be attributable to impaired P-selectin externalization in these platelets. However, only CCCP was able to prevent TRAP-induced plateletmonocyte interaction (Online Supplementary Figure S2F-J, L). Given that thrombin-induced phosphatidylserine exposure on platelets was unaffected by oligomycin, a phosphotidylserine-Tim4 interaction9 could be mediating the platelet-monocyte interaction in oligomycin-treated platelets despite compromised P-selectin-facilitated interactions. This hypothesis was supported by our observation that oligomycin inhibited platelet-monocyte interactions (Online Supplementary Figure S1L) upon stimulation with ADP which, unlike TRAP, does not induce phosphatidylserine exposure. Several prothrombotic factors are released from platelet granules upon stimulation, including ADP and fibrinogen.10 ADP, in particular, plays a pivotal role in recruiting platelets to the ‘shell’ region of a growing thrombus.11 As we found that mitochondrial ATP sustained granule release, we hypothesized that it could be essential for the process of thrombosis. We studied platelet thrombus formation on immobilized collagen under arterial shear (1500 s-1). Washed human platelets were allowed to perfuse over the collagen-coated surface for 5 min. Total surface area covered by platelet thrombi was significantly diminished in the presence of oligomycin (Figure 2A, E) and CCCP (Figure 2A, F) but unaffected by antimycin (Figure 2A, D), which could be attributable to impaired ADP release consequent to oligomycin and CCCP treatment. This hypothesis was supported by the observation that, neither oligomycin 1209

Letters to the Editor

nor CCCP had any impact on thrombus growth in the presence of ticagrelor, an ADP receptor antagonist (Figure 2B, G-I). Furthermore, adhesion and spreading of platelets pretreated with tirofiban, an aIIbβ3 integrin blocker, on collagen matrix under shear were unaffected in the presence of antimycin, oligomycin or CCCP (Figure

2C, J-L). Thus, ATP sourced from mitochondria was dispensable for platelet aggregation, adhesion and procoagulant activity but was essential for granule secretion, platelet-neutrophil interactions, as well as thrombus growth. We asked whether differential modulation of platelet

Figure 1. Effects of pre-treatment with mitochondrial inhibitors (antimycin, oligomycin) or an uncoupler (CCCP) on human platelet function in vitro. Washed human platelets were pre-treated with vehicle (control), antimycin (2 μg/mL), oligomycin (10 μg/mL) or CCCP (100 μM) for 15 min at room temperature, followed by addition of human thrombin (0.5 U/mL) for 30 min, or as indicated. (A) Representative tracings showing thrombin (0.2 U/mL)-induced aggregation (1-4) or ATP release (1’-4’) of platelets pre-treated with vehicle (1,1’), antimycin (2,2’), oligomycin (3,3’) or CCCP (4,4’). (B, C) Bar diagram quantifying mean platelet aggregation and platelet dense granule secretion in different samples, respectively. (D-F) Representative histogram overlay plots showing binding of FITC-PAC1 (BD Biosciences #340507), PE-annexin V (Biolegend #640908), and PE-anti-CD62P (BD Biosciences #550561), respectively. (G-I) Corresponding bar diagrams quantifying aIIbβ3 integrin activation, phosphatidylserine exposure and P-selectin expression, respectively. Each dot represents an independent observation. Data are presented as mean ± standard error of mean. *P<0.05 with respect to vehicle-treated thrombin-stimulated platelets. #P<0.05 with respect to vehicletreated unstimulated platelets. Significance in difference of means was tested by repeated measures analysis of variance and the Dunnett multiple comparison test. PS: phosphatidylserine; Thr: thrombin.

1210

haematologica | 2022; 107(5)

Letters to the Editor

bioenergetics by the uncoupler and inhibitors accounts for differences in their impact on platelet responses. Concurrent inhibition of both mitochondrial respiration and glycolysis has been found to impede platelet aggregation.1,2,4 This prompted us to hypothesize that, CCCP

could be compromising both mitochondrial and glycolytic ATP generation, thereby leading to impaired platelet responses. Intriguingly, neither antimycin/oligomycin nor CCCP had any significant effect on cellular ATP level in platelets (Figure 3A). Sustained ATP levels suggest

Figure 2. Effects of pre-treatment with mitochondrial inhibitors (antimycin, oligomycin) or an uncoupler (CCCP) on platelet thrombus formation. Washed human platelets pretreated with antimycin (2 μg/mL), oligomycin (10 μg/mL), CCCP (100 µM) or vehicle (control) were labeled with calcein-AM (Molecular Probes #C3100MP) and perfused over an immobilized collagen matrix for 5 min in a microfluidics flow chamber at a shear rate of 1500 s-1. (A-C) Representative images of platelet accumulation after 5 min of perfusion of human platelets with treatments as indicated in the presence of vehicle (A), ticagrelor (1 μM) (B) or tirofiban (100 ng/mL) (C). (D-L) Corresponding bar diagrams representing total surface area covered by platelet thrombi after 5 min of perfusion on the collagen matrix. Each dot represents an independent observation. Data are presented as mean ± standard error of mean. *P<0.05 with respect to vehicle-treated sample. Statistical significance of difference of means was tested by a multiple paired Student t-test.

haematologica | 2022; 107(5)

1211

Letters to the Editor

compensatory augmentation in glycolysis. In agreement, treatment with either antimycin/oligomycin or CCCP was associated with enhanced lactate release from thrombin-stimulated platelets (Figure 3C), with antimycin being the most effective among them. These findings are strongly suggestive of a compensatory surge in glycolytic rate in platelets upon mitochondrial inhibition which sustains aIIbβ3 platelet integrin activation,

aggregation and phosphatidylserine exposure. Aggregation induced by threshold concentrations of thrombin and collagen was unaffected by oligomycin pre-treatment (Online Supplementary Figure S1F, J, K), which further bolsters the hypothesis that glycolytic ATP is necessary and sufficient to fuel platelet aggregation. The above observations led us to consider that the molecular events underpinning the effects of CCCP on

Figure 3. Effects of pre-treatment with mitochondrial inhibitors (antimycin, oligomycin) or an uncoupler (CCCP) on platelet ATP, reactive oxygen species, calcium and lactate release rates. Washed human platelets were pre-treated with vehicle (control), antimycin (2 μg/mL), oligomycin (10 μg/mL) or CCCP (100 μM) for 15 min at room temperature, followed by addition of thrombin (0.5 U/mL) for 5 min (mitochondrial calcium), 15 min (cellular ATP), or 30 min (lactate release rates, mitochondrial reactive oxygen species [ROS] and intracellular ROS). (A-C) Bar diagrams indicating cellular ATP levels, mitochondrial calcium and lactate release rates, respectively. (F, G) Histogram overlay plots representing MitoSOX (Invitrogen #M36008) and dichlorofluorescein (DCF) (Sigma #D6883) fluorescence intensities, respectively. (D, E) Corresponding bar diagrams indicating mitochondrial and intracellular ROS levels, respectively. Each dot represents an independent observation. Data are presented as mean ± standard error of mean *P<0.05 with respect to vehicle-treated thrombin-stimulated platelets. #P<0.05 with respect to vehicle-treated unstimulated platelets. Statistical significance of difference of means was tested by repeated measures analysis of variance and the Dunnett multiple comparison test. RP: resting platelets; Thr, thrombin.

1212

haematologica | 2022; 107(5)

Letters to the Editor

platelets could be mediated through mitochondrial functions unrelated to ATP generation. Reactive oxygen species (ROS) have now been established as important signaling molecules responsible for platelet activation.2,12 Mitochondria are a significant source of ROS in platelets12 and selectively scavenging mitochondrial superoxide with mitoTEMPO prevents platelet aggregation.13 ROS release from mitochondria is dependent on inner mitochondrial membrane polarization, and uncoupling is known to reduce mitochondrial ROS.14 In concurrence, CCCP but not antimycin or oligomycin brought about a significant drop in mitochondrial ROS in thrombin-stimulated platelets (Figure 3D, F). These changes were also reflected in intracellular ROS levels, which were significantly curbed in the presence of CCCP but remained unaffected by antimycin or oligomycin (Figure 3E, G). Phosphatidylserine exposure by platelets is dependent on cyclophilin D-dependent formation of mitochondrial permeability transition pores, which is triggered by calcium entry into mitochondria through a mitochondrial calcium uniporter along the electrical gradient across the inner mitochondrial membrane.15 We could find inhibition of thrombin-induced mitochondrial calcium transients in the presence of CCCP but not in the presence of antimycin or oligomycin (Figure 3B; Online Supplementary Figure S3A-D). Hence, it was fairly reasonable to posit that CCCP restrains platelet aggregation and procoagulant activity through abrogation of mitochondrial ROS and calcium transients, respectively. In summary, mitochondrial ATP was found to be dispensable for platelet aggregation and procoagulant activity, which are fueled by glycolytic ATP. However, maintenance of a proton gradient across inner mitochondrial membrane plays a vital role in these processes by supporting ROS generation and mitochondrial calcium influx, respectively. We discovered that mitochondrial ATP is critical for sustaining platelet granule secretion, platelet-neutrophil interactions and thrombus growth, especially when inadequately compensated by glycolytic ATP (Online Supplementary Figure S3E). This knowledge should have important implications for the development of anti-thrombotic strategies that selectively target platelet granule release in the treatment of thromboinflammatory diseases such as acute myocardial infarction, ischemic stroke, deep vein thrombosis and pulmonary embolism. Paresh P. Kulkarni,1 Mohammad Ekhlak,1 Vijay K. Sonkar2 and Debabrata Dash1 1 Center for Advanced Research on Platelet Signaling and Thrombosis Biology, Department of Biochemistry, Institute of Medical Sciences, Banaras Hindu University, Varanasi, Uttar Pradesh and 2 Department of Molecular and Human Genetics, Institute of Science, Banaras Hindu University, Varanasi, Uttar Pradesh, India Correspondence: DEBABRATA DASH: ddash.biochem@gmail.com doi:10.3324/haematol.2021.279847 Received: August 19, 2021.

haematologica | 2022; 107(5)

Accepted: December 16, 2021. Pre-published: December 23, 2021. Disclosures: this research was supported by a J.C. Bose National Fellowship and grants received by DD from the Indian Council of Medical Research (ICMR) under CAR, Department of Biotechnology (DBT) and Science and Engineering Research Board (SERB), Government of India. DD also acknowledges assistance from the Humboldt Foundation, Germany. ME is a recipient of a CSIR-JRF. Contributions: DD supervised the entire work; DD and PPK designed the research; PPK, ME and VKS performed experiments and analyzed results; DD and PPK wrote the manuscript.

References 1. Ravi S, Chacko B, Sawada H, et al. Metabolic plasticity in resting and thrombin activated platelets. PLoS One. 2015;10(4):e0123597. 2. Kulkarni PP, Tiwari A, Singh N, et al. Aerobic glycolysis fuels platelet activation: small-molecule modulators of platelet metabolism as anti-thrombotic agents. Haematologica. 2019;104(4):806818. 3. Holmsen H. Energy metabolism and platelet responses. Vox Sang. 1981;40(1):1-7. 4. Kaczara P, Sitek B, Przyborowski K, et al. Antiplatelet effect of carbon monoxide is mediated by NAD+and ATP depletion. Arterioscler Thromb Vasc Biol. 2020;40:2376-2390. 5. Nechipurenko DY, Receveur N, Yakimenko AO, et al. Clot contraction drives the translocation of procoagulant platelets to thrombus surface. Arterioscler Thromb Vasc Biol. 2019;39(1):3747. 6. Verhoeven AJM, Mommersteeg ME, Willem J, Akkerman N. Quantification of energy consumption in platelets during thrombin-induced aggregation and secretion. Tight coupling between platelet responses and the increment in energy consumption. Biochem J. 1984;221(3):777-787. 7. Holmsen H, Kaplan KL, Dangelmaier CA. Differential energy requirements for platelet responses. A simultaneous study of aggregation, three secretory processes, arachidonate liberation, phosphatidylinositol breakdown and phosphatidate production. Biochem J. 1982;208(1):9-18. 8. Kral JB, Schrottmaier WC, Salzmann M, Assinger A. Platelet interaction with innate immune cells. Transfus Med Hemother. 2016;43(2):78-88. 9. Segawa K, Nagata S. An apoptotic ‘eat me’ signal: phosphatidylserine exposure. Trends Cell Biol. 2015;25(11):639-650. 10. Golebiewska EM, Poole AW. Platelet secretion: from haemostasis to wound healing and beyond. Blood Rev. 2015;29(3):153-162. 11. Welsh JD, Stalker TJ, Voronov R, et al. A systems approach to hemostasis: 1. The interdependence of thrombus architecture and agonist movements in the gaps between platelets. Blood. 2014;124(11):1808-1815. 12. Masselli E, Pozzi G, Vaccarezza M, et al. ROS in platelet biology: functional aspects and methodological insights. Inty J Mol Sci. 2020;21(14):1-35. 13. Sonkar VK, Kumar R, Jensen M, et al. Nox2 NADPH oxidase is dispensable for platelet activation or arterial thrombosis in mice. Blood Adv. 2019;3(8):1272-1284. 14. Cadenas S. Mitochondrial uncoupling, ROS generation and cardioprotection. Biochim Biophys Acta Bioenerg. 2018;1859(9):940950. 15. Kholmukhamedov A, Janecke R, Choo HJ, Jobe SM. The mitochondrial calcium uniporter regulates procoagulant platelet formation. J Thromb Haemost. 2018;16(11):2315-2321.

1213

Letters to the Editor

Autologous hematopoietic cell transplantation in diffuse large B-cell lymphoma after three or more lines of prior therapy: evidence of durable benefit While most patients with diffuse large B-cell lymphoma (DLBCL) are cured with initial chemoimmunotherapy, one-third of patients will have relapsed and/or refractory (r/r) disease after frontline treatment. Salvage combination chemoimmunotherapy followed by autologous hematopoietic cell transplantation (autoHCT), in patients achieving an objective response to cures less than half of such patients.1,2 Most patients who undergo autoHCT do so after second line (2L) therapy, but some do so after having received three or more lines (3L+) of prior therapy. Data are lacking on the outcomes after autoHCT in patients with DLBCL in the 3L+ setting. Although CD19-directed chimeric antigen receptor T-cell (CAR-T) therapy is being increasingly used in the 3L+ setting with curative intent,3-5 this topic remains relevant given issues with access to CAR-T both in the US6 and worldwide, particularly in low and middle income countries.7 Here we report outcomes after autoHCT in the subset of patients with DLBCL who received 3L+ of systemic therapy in a Center for International Blood and Marrow Transplant Research (CIBMTR) registry analysis. The CIBMTR is a collaborative research program managed by the Medical College of Wisconsin and the National Marrow Donor Program that collects data from more than 380 transplant centers worldwide. Participating sites are required to report detailed data on both autologous and allogeneic HCT with frequent updates gathered during the longitudinal follow-up of transplant patients, and the compliance is monitored by on-site audits. Computerized checks for discrepancies, physicians’ review of submitted data, and on-site audits of participating centers ensure data quality. Observational studies conducted by the CIBMTR are performed in compliance with all applicable federal regulations pertaining to the protection of human research participants. The Medical College of Wisconsin and National Marrow Donor Program Institutional Review Boards approved this study. Patients with DLBCL (aged ≥18 years) who received autoHCT between 2003 and 2017 with a preparative regimen of either BEAM (carmustine, etoposide, cytarabine, melphalan) or R-BEAM (rituximab with BEAM) conditioning after 3L+ therapy were included in this analysis. All patients received rituximab-containing, anthracycline-based frontline therapy. Patients who received a bone marrow graft, with chemorefractory disease after salvage therapy, and with active central nervous system involvement prior to autoHCT were excluded. Patients with transformed DLBCL evolving from prior indolent lymphoma were also excluded. Chemosensitive disease was defined as achieving either a complete remission (CR) or partial remission (PR) to salvage treatment. Response to frontline chemoimmunotherapy and disease status at autoHCT were determined by each center using the International Working Group criteria.8,9 Early chemoimmunotherapy failure was defined as not achieving a CR after frontline chemoimmunotherapy or relapse/progression within 1 year of initial diagnosis.10 The primary endpoint was OS. Death from any cause was considered an event and surviving patients were censored at last follow-up. Secondary outcomes included non-relapse mortality (NRM), relapse/progression, 1214

and progression-free survival (PFS). NRM was defined as death without preceding evidence of lymphoma progression/relapse; relapse was considered a competing risk. Relapse/progression was defined as progressive lymphoma after autoHCT or lymphoma recurrence after a CR; NRM was considered a competing risk. For PFS, a patient was considered a treatment failure at the time of progression/relapse or death from any cause. Patients alive without evidence of disease relapse or progression were censored at last follow-up. All outcomes were calculated relative to the autoHCT date. The study cohort was divided according to remission status at the time of autoHCT (CR vs. PR). Patient-, disease- and transplant-related variables were compared between the two cohorts using the Chi-square test for categorical variables and the Wilcoxon two-sample test for continuous variables. The distribution of OS and PFS were estimated using the Kaplan-Meier method. Cumulative incidence method was used to estimate NRM, relapse/progression while accounting for competing events. The Cox proportional hazards model for PFS and OS and the cause-specific hazards model for relapse and NRM were used to identify prognostic factors using forward stepwise variable selection. No covariates violated the proportional hazards assumption. No significant interactions between the main effect and significant covariates were found. No center effects were found based on the score test of homogeneity.11 Results were reported as hazard ratio (HR), 95% confidence interval (CI) for HR and P-value. The adjusted probabilities for each outcome were calculated based on the final regression model. Covariates with a P-value <0.05 were considered statistically significant. All statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC). A total of 285 patients met the inclusion criteria; over the same interval, 577 patients in the dataset who otherwise met the inclusion criteria had undergo autoHCT after receiving two or fewer lines of prior therapy. Median age was 60 years (range, 19-80 years), 60% were male, 80% were Caucasian, and 63% had early chemoimmunotherapy failure. Eighty percent received BEAM conditioning and 20% received R-BEAM. Details regarding the 3L treatment regimen are included in the Online Supplementary Table S3. Baseline characteristics are shown in Table 1. 5-year OS and PFS were 51% (95% CI: 44-57) and 38% (95% CI: 32-55), respectively. Adjusted 1-year, 3-year, and 5-year PFS and OS are shown in Table 2. Patients in CR at autoHCT had a higher 1-year OS (84% vs. 63%, P<0.001) and PFS (69% vs. 48%, P<0.001) in contrast to patients in PR, with the difference in OS persisting at 3 years (OS 64% vs. 50%, P=0.02). There was a trend towards improved 5-year OS for patients in CR (56% vs. 45%, P=0.06), whereas 5year PFS did not differ significantly between the two cohorts (42% vs. 34%, P=0.18). The 1- and 5-year incidence of relapse/progression for all patients was 36% (95% CI: 31-42) and 50% (95% CI: 43-56), respectively. Patients in CR had a significantly lower 1-year incidence of relapse/progression (26% vs. 47%, P<0.001); however, there was no difference found in relapse/progression at 5 years between the two cohorts (45% vs. 54%, P=0.14). The 1-year and 5-year NRM were 5% (95% CI: 3-8) and 12% (95% CI: 9-17), respectively with no difference identified between patients in CR and those in PR. A graph of outcomes stratified by disease status at autoHCT is provided in Figure 1. A multivariable regression model was constructed to haematologica | 2022; 107(5)

Letters to the Editor

evaluate for association between disease status at autoHCT (CR vs. PR), baseline covariates, and NRM, relapse, PFS, and OS; covariates are listed in the Online Supplementary Table S1. PR at autoHCT was associated with significantly increased risk of relapse (HR 1.59, 95% CI: 1.13-2.24, P=0.008), inferior PFS (HR 1.46, 95% CI: 1.08-1.97, P=0.01), and OS (HR 1.55, 95% CI: 1.122.15, P=0.009, Online Supplementary Table S2). Causes of death were analyzed for the 144 patients who died. Overall, 61% (n=88) died from DLBCL, whereas the second leading cause of death was secondary malignancy (10%).

We have found in our registry analysis that autoHCT performed in the 3L+ setting for DLBCL is feasible and effective with a 5-year PFS of 41% and 35% in patients who achieved CR and PR, respectively, prior to autoHCT. Multivariate regression analyses demonstrated that CR at the time of autoHCT was associated with less relapse and improved PFS and OS. These data suggest that autoHCT still may play a role in the 3L+ setting in DLBCL for patients who demonstrate an objective response to a second salvage. In fact, a substantial percentage of 3L+ patients in PR at autoHCT experienced durable disease control. This finding is in keeping with a

Table 1. Baseline characteristics of patients receiving BEAM conditioning regimen and autologous hematopoietic cell transplantation for diffuse large B-cell lymphoma during 2003-2017 (>=three prior lines of treatment) Number of patients Patient age Median (range), y ≥ 65 y, n (%) Males Patient race Caucasian African American Other* Missing Karnofsky Performance Score ≥ 90 Missing Stage at diagnosis Stage III-IV Missing LDH Elevated at diagnosis Missing Bone marrow involvement at diagnosis No Missing Extranodal involvement at diagnosis Yes Missing Time from diagnosis to HCT, median (range), mo Early chemoimmunotherapy failure Yes Missing Primary refractory after first line of therapy Yes Missing Number of prior lines of therapy 3 >3 Conditioning regimen BEAM Rituximab-BEAM Year of transplant 2003-2007 2008-2012 2013-2017 Median follow-up of survivors (range), mo

All patients

P-value

285

154

131

60 (19-80) 85 (30) 170 (60)

60 (20-80) 46 (30) 92 (60)

59 (19-77) 39 (30) 78 (60)

229 (80) 28 (10) 18 (6) 10 (4)

124 (80) 15 (10) 11 (7) 4 (3)

105 (80) 13 (10) 7 (5) 6 (5)

140 (49) 12 (4)

87 (57) 7 (4)

53 (40) 5 (4)

199 (70) 17 (6)

106 (69) 11 (7)

93 (71) 6 (5)

0.66

39 (14) 182 (64)

25 (16) 95 (62)

14 (11) 87 (66)

0.39

202 (71) 15 (5)

102 (66) 7 (5)

100 (76) 8 (6)

181 (64) 15 (5) (5-172)

105 (68) 7 (5) 23 (6-140)

76 (58) 8 (6) 17 (5-172)

0.22

179 (63) 6 (2)

86 (56) 5 (3)

93 (71) 1 (1)

0.02

119 (42) 19 (7)

46 (30) 14 (9)

73 (56) 5 (4)

< 0.001

217 68

126 28

91 40

0.01

227 (80) 58 (20)

126 (82) 28 (18)

101 (77) 30 (23)

98 (34) 99 (35) 88 (31) 72 (4-145)

49 (32) 56 (36) 49 (32) 72 (6-145)

49 (37) 43 (33) 39 (30) 72 (4-143)

0.28 0.22 0.97 0.76

0.02

0.07 0.22

0.01

0.32

0.61

Unless otherwise noted, data are n (%). BEAM: carmustine, etoposide, cytarabine, and melphalan; HCT: hematopoietic stem cell transplantation; DLBCL: diffuse large B-cell lymphoma; LDH: lactate dehydrogenase; CR: complete remission; PR: partial remission; y:years, mo: months. *Other race: CR: 11 Asian; PR: 5 Asian; 1 American Indian or Alaska Native; 1 Native Hawaiian or Other Pacific Islander.

haematologica | 2022; 107(5)

1215

Letters to the Editor

recent CIBMTR analysis that patients with a PR prior to autoHCT had a 5-year PFS of 41%.12 As such, patients with chemosensitive disease, particularly those who attain a CR, should not be denied the opportunity for curative intent treatment with autoHCT solely due to the number of prior lines of therapy. We acknowledge a number of limitations with this analysis including its retrospective design as well as that these data pertain only to patients who respond to salvage therapy, and many patients do not.13 Registry data show that the number of patients who undergo autoHCT in the 3L+ setting is less than those who do so

after two lines of therapy.10,14 Of the patients in the CORAL study who did not respond to second-line therapy, only 39% responded to 3L therapy and 28% of patients ultimately proceeded to autoHCT.15 Furthermore, many novel therapies for DLBCL have been approved recently including multiple targeted treatments as well as three separate CAR-T products.3-5 CAR-T therapy has profoundly impacted the care of DLBCL given its ability to induce durable remissions even in the setting of chemorefractory disease. Although only approved in the third line setting at present, randomized trials of CAR-T compared to salvage chemoim-

Table 2. Adjusted outcomes.

Outcomes Non-relapse mortality (range) 1-year 3-year 5-year Relapse/progression (range) 1-year 3-year 5-year Progression-free survival (range) 1-year 3-year 5-year Overall survival (range) 1-year 3-year 5-year

All patients (N = 285)

CR (N = 154)

PR (N = 131)

P-Value

5 (3-8) 11 (8-15) 12 (9-17)

5 (2-9)% 12 (7-18)% 13 (8-19)%

5 (1-8)% 9 (4-14)% 12 (6-17)%

0.68 0.40 0.65

36 (31-42) 44 (38-50) 50 (43-56)

26 (19-33)% 38 (30-46)% 45 (37-53)%

47 (39-56)% 51 (42-59)% 54 (45-63)%

<0.001 0.03 0.14

59 (53-64) 45 (39-51) 38 (32-44)

69 (61-76)% 50 (42-59)% 42 (34-51)%

48 (39-56)% 40 (31-48)% 34 (26-43)%

<0.001 0.08 0.18

74 (69-79) 57 (51-63) 51 (44-57)

84 (79-90)% 64 (56-72)% 56 (48-65)%

63 (55-71)% 50 (42-59)% 45 (36-54)%

<0.001 0.02 0.06

Data are percentage probability (95% confidence interval). CR: complete remission; PR: partial remission.

Figure 1. Post- autologous hematopoietic cell transplantation (autoHCT) outcomes stratified by pre-autoHCT disease status (complete response vs. partial remission). (A) Non-relapse mortality, (B) relapse/progression, (C) progression-free survival and (D) overall survival. CR: complete remission; PR: partial remission.

1216

haematologica | 2022; 107(5)

Letters to the Editor

munotherapy/autoHCT as second line therapy are being conducted (TRANSFORM, clinicaltrials gov. Identifier: NCT03575351; ZUMA-7, clinicaltrials gov. Identifier: NCT03391466, and BELINDA, clinicaltrials gov. Identifier: NCT03570892) with potential practice-changing implications. The number of patients who received novel therapies including CAR-T prior to autoHCT in this analysis is likely low as the first commercial CAR-T product was approved in late 2017 and the first targeted therapy for relapsed DLBCL in 2019 (polatuzumab vedotin-piiq). Nonetheless, the retrospective data presented here suggest that autoHCT still has a role in r/r chemosensitive DLBCL even in later lines of therapy. If CAR-T ultimately becomes standard second line therapy, these data may serve as a benchmark for autoHCT outcomes in patients with 3+ prior lines of chemotherapy. Additionally, it would support offering an autoHCT in patients in the 3L+ setting in countries where CAR-T may not be available. Matthew Mei,1 Mehdi Hamadani,2 Kwang W. Ahn,3,4 Yue Chen,4 Mohamed A. Kharfan-Dabaja,5 Craig Sauter2 and Alex F. Herrera1 1 Department of Hematology and Hematopoietic Cell Transplantation, City of Hope, Duarte, CA; 2BMT & Cellular Therapy Program, Department of Medicine, Medical College of Wisconsin, Milwaukee, WI; 3Division of Biostatistics, Medical College of Wisconsin, Milwaukee, WI; 4Center for International Blood and Marrow Transplant Research, Milwaukee, WI and 5Division of Hematology-Oncology and Blood and Marrow Transplantation Program, Mayo Clinic, Jacksonville, FL, USA Correspondence: MEHDI HAMADANI mhamadani@mcw.edu doi:10.3324/haematol.2021.279999 Received: September 9, 2021. Accepted: December 24, 2021. Pre-published: February 3, 2022. Disclosures: MH reports research support/funding from Spectrum Pharmaceuticals, Astellas Pharma; acts as a consultant for Janssen R &D, Incyte Corporation ,ADC Therapeutics, Verastem, Kite and is part of the speaker’s bureau of Sanofi Genzyme, AstraZeneca, BeiGene. MM reports research support/funding from TG Therapeutics, Epizyme, Inc., Bristol Meyers Squibb, BeiGene, Morphosys Agand has received honoraria from EUSA, Janssen Pharmaceuticals and Sanofi-Genzyme. AH reports research funding/consultancy fees from Bristol Myers-Squibb, Genentech, Merck, Seattle Genetics, AstraZena, ADC Therapeutics, KiTE Pharma, Gilead Sciences, Karyopharm, Takeda, Tubulis. Contributions: MM and AH developed the concept and design of the study; YC and MH collected and assembled the data; KWA, YC and MH analyzed the data; MM, AH and MH prepared the first draft and wrote the manuscript. All authors interpreted data and helped to revise the manuscript. Funding: the CIBMTR is supported primarily by Public Health Service U24CA076518 from the National Cancer Institute (NCI), the National Heart, Lung and Blood Institute (NHLBI) and the National Institute of Allergy and Infectious Diseases (NIAID); HHSH250201700006C from the Health Resources and Services Administration (HRSA); and N00014-20-1-2705 and N00014-201-2832 from the Office of Naval Research; support is also provided by Be the Match Foundation, the Medical College of Wisconsin, the National Marrow Donor Program, and from the following commercial

haematologica | 2022; 107(5)

entities: AbbVie; Accenture; Actinium Pharmaceuticals, Inc.; Adaptive Biotechnologies Corporation; Adienne SA; Allovir, Inc.; Amgen, Inc.; Astellas Pharma US; bluebird bio, inc.; Bristol Myers Squibb Co.; CareDx; CSL Behring; CytoSen Therapeutics, Inc.; Daiichi Sankyo Co., Ltd.; Eurofins Viracor; ExcellThera; Fate Therapeutics; GamidaCell, Ltd.; Genentech Inc; Gilead; GlaxoSmithKline; Incyte Corporation; Janssen/Johnson & Johnson; Jasper Therapeutics; Jazz Pharmaceuticals, Inc.; Karyopharm Therapeutics; Kiadis Pharma; Kite, a Gilead Company; Kyowa Kirin; Magenta Therapeutics; Medac GmbH; Merck & Co.; Millennium, the Takeda Oncology Co.; Miltenyi Biotec, Inc.; MorphoSys; Novartis Pharmaceuticals Corporation; Omeros Corporation; Oncopeptides, Inc.; Orca Biosystems, Inc.; Pfizer, Inc.; Pharmacyclics, LLC; Sanofi Genzyme; Seagen, Inc.; Stemcyte; Takeda Pharmaceuticals; Tscan; Vertex; Vor Biopharma; Xenikos BV. Data sharing statement: CIBMTR supports accessibility of research in accord with the National Institutes of Health (NIH) Data Sharing Policy and the National Cancer Institute (NCI) Cancer Moonshot Public Access and Data Sharing Policy. The CIBMTR only releases de-identified datasets that comply with all relevant global regulations regarding privacy and confidentiality.

References 1. Philip T, Guglielmi C, Hagenbeek A, et al. Autologous bone marrow transplantation as compared with salvage chemotherapy in relapses of chemotherapy-sensitive non-Hodgkin's lymphoma. N Engl J Med. 1995;333(23):1540-1545. 2. Gisselbrecht C, Glass B, Mounier N, et al. Salvage regimens with autologous transplantation for relapsed large B-cell lymphoma in the rituximab era. J Clin Oncol. 2010;28(27):4184-4190. 3. Schuster SJ, Bishop MR, Tam CS, et al. Tisagenlecleucel in adult relapsed or refractory diffuse large B-cell lymphoma. N Engl J Med. 2018;380(1):45-56. 4. Abramson JS, Palomba ML, Gordon LI, et al. Lisocabtagene maraleucel for patients with relapsed or refractory large B-cell lymphomas (TRANSCEND NHL 001): a multicentre seamless design study. Lancet. 2020;396(10254):839-852. 5. Neelapu SS, Locke FL, Bartlett NL, et al. Axicabtagene ciloleucel CAR T-cell therapy in refractory large B-bell lymphoma. N Engl J Med. 2017;377(26):2531-2544. 6. Kansagra A, Farnia S, Majhail N. Expanding access to chimeric antigen receptor T-cell therapies: challenges and opportunities. Am Soc Clin Oncol Educ Book. 2020;40:1-8. 7. Burki TK. CAR T-cell therapy roll-out in low-income and middleincome countries. Lancet Haematol. 2021;8(4):e252-e253. 8. Cheson BD, Pfistner B, Juweid ME, et al. Revised response criteria for malignant lymphoma. J Clin Oncol. 2007;25(5):579-586. 9. Cheson BD, Fisher RI, Barrington SF, et al. Recommendations for initial evaluation, staging, and response assessment of Hodgkin and non-Hodgkin lymphoma: the Lugano classification. J Clin Oncol. 2014;32(27):3059-3068. 10. Hamadani M, Hari PN, Zhang Y, et al. Early failure of frontline rituximab-containing chemo-immunotherapy in diffuse large B cell lymphoma does not predict futility of autologous hematopoietic cell transplantation. Biol Blood Marrow Transplant. 2014;20(11):17291736. 11. Commenges D, Andersen PK. Score test of homogeneity for survival data. Lifetime Data Anal. 1995;1(2):145-156; discussion 157-149. 12. Shah NN, Ahn KW, Litovich C, et al. Is autologous transplant in relapsed DLBCL patients achieving only a PET+ PR appropriate in the CAR T-cell era? Blood. 2021;137(10):1416-1423. 13. Crump M, Neelapu SS, Farooq U, et al. Outcomes in refractory diffuse large B-cell lymphoma: results from the international SCHOLAR-1 study. Blood. 2017;130(16):1800-1808. 14. Jagadeesh D, Majhail NS, He Y, et al. Outcomes of rituximab-BEAM versus BEAM conditioning regimen in patients with diffuse large B cell lymphoma undergoing autologous transplantation. Cancer. 2020;126(10):2279-2287. 15. Van Den Neste E, Schmitz N, Mounier N, et al. Outcome of patients with relapsed diffuse large B-cell lymphoma who fail second-line salvage regimens in the International CORAL study. Bone Marrow Transplant. 2016;51(1):51-57.

1217

COMMENTS Serological response following anti-SARS-CoV-2 vaccination in hematopoietic stem cell transplantation patients depends upon time from transplant, type of transplant and “booster”dose We read with great interest the systematic review and meta-analysis by Gagelmann et al. on antibody response after vaccination against SARS-CoV-2 in adults with hematological malignancies recently published in Haematologica.1 Among others, the authors analyzed thirteen studies evaluating 1,324 patients who had undergone hematopoietic cell transplantation (HSCT). The pooled response was 83% for autologous and 82% for allogeneic HSCT recipients, respectively. Though limited evidence showed higher responses for patients receiving vaccination at least 6-12 months after HSCT, the analysis was not able to differentiate the results according to specific timing and type of transplantation, an aspect that was not systematically assessable, due to the low number and heterogeneity of studies. We recently published a prospective, cohort study,2 not included in the meta-analysis, of 114 fully vaccinated patients who had received an autologous (52 patients) or allogeneic (62 patients) HSCT at least 3 months before the first dose of vaccination. Overall, serological response rate (>50 AU/mL of anti-spike protein immunoglobulin G [IgG] antibodies detected 4 weeks after the second dose of BNT162b2 mRNA COVID-19 vaccine) was 84%, thus perfectly in line with the results of the meta-analysis. Interestingly, responders after an allogeneic HSCT performed better, in terms of magnitude of serological response, than those treated with an autologous HSCT. However, 6% of autologous and 24% of allogeneic HSCT recipients did not respond at all. In that study, aiming to explore in depth the response according to the time elapsed from transplant, we stratified the patients into three groups: G1=<1 year; G2=1-5 years; G3=>5 years. Among 16% of patients who failed to respond, the large majority was constituted of individuals transplanted within 1 year before vaccination and who had received an allogeneic HSCT. When compared to 107 healthy controls (HC) matched for age and sex, lower antibody titers were observed in both allogeneic and autologous HSCT recipients in G1, while no differences emerged in G2. Interestingly, results in G3 between HC and allogeneic recipients were comparable, whereas patients in the autologous subgroup showed significantly lower titers than HC.2 Thus, we confirmed that most of transplanted patients respond to a complete vaccination cycle and we also observed that failures generally occur within the first year from transplant, mainly in allogeneic HSCT recipients. Furthermore, our analysis revealed that patients who had received an allogeneic transplantation develop higher antibody production than those who had received an autologous transplantation, particularly if vaccinated more than 5 years after HSCT. Time requested for a quantitative and functional recovery of B and T cells after HSCT (up to 1 year or even more), as well as the use of graft-versus-host disease prophylaxis with immunosuppressive agents in the allogeneic setting, might explain the lower antibody titers and the larger number of non-responders in G1. Differences between autologous and allogeneic groups in patients transplanted more than 5 years before vaccination might be instead related to a more frequent persistence of a still active disease and to ongoing salvage treatments in autologous HSCT recipients, which are well known risk factors for a poor response to vaccination.1 For example, our experience confirmed that patients with

1218

myeloma in remission phase after autologous HSCT show significantly higher antibody titers than patients with active disease.2 On the other hand, we hypothesize that the presence of a “healthy” and consolidated immune system provided by the donors in the allogeneic setting, could play a role in producing a more robust response after a longer time period, like that in normal individuals, as we observed in our patients. Administration of a “booster” dose could change such a scenario. Recent data suggest that a third dose of BNT162b2 anti-SARS-CoV-2 mRNA vaccine may improve the humoral response in allogeneic HSCT recipients.3,4 Some preliminary data from our Institution also indicate a significant increase of serological response after a third dose of BNT162b2 antiSARS-CoV-2 mRNA vaccine in transplanted patients, particularly, but not only, in the allogeneic group. Of note, a not negligible proportion of non-responders after the first two doses (mainly patients belonging to the allogeneic HSCT group in G1, more rarely in the autologous setting) was able to mount a serological response 1 month after the third dose (Attolico et al., manuscript in preparation). Most of the available studies have only evaluated the serological response in terms of anti-spike IgG antibodies. Clearcut relationships between these antibodies and protection against the virus has not been unequivocally established. As observed in other contexts, neutralizing antibodies, development of memory B cells and T-cell immune response after vaccination could play an even more important role in protecting against SARS-CoV-2 infection. Notwithstanding, we underline that time from transplant, type of transplant (allogeneic vs. autologous) and a third dose of vaccine significantly affect the serological response in HSCT patients. If further confirmed in larger studies, we think these aspects should be considered in planning anti-SARS-CoV-2 vaccine strategies for these patients. Immacolata Attolico,1 Francesco Tarantini,2 Paola Carluccio1 and Pellegrino Musto1,2 1 Hematology and Stem Cell Transplantation Unit, AOUC Policlinico and 2Department of Emergency and Organ Transplantation, “Aldo Moro”University School of Medicine, Bari, Italy. Correspondence: IMMACOLATA ATTOLICO - immattolico@gmail.com doi:10.3324/haematol.2022.280619 Received: January 5, 2022. Accepted: January 11, 2022. Pre-published: January 20, 2022. Disclosures: no conflicts of interest to disclose. Contributions: IA and PM wrote the manuscript; FT performed data analysis; IA, FT and PC performed research; PM supervised the study.

References 1. Gagelmann N, Passamonti F, Wolschke C, et al. Antibody response after vaccination against SARS-CoV-2 in adults with haematological malignancies: a systematic review and meta-analysis. Haematologica. 2021 Dec 16. [Epub ahead of print] 2. Attolico I, Tarantini F, Carluccio P, et al. Serological response following BNT162b2 anti-SARS-CoV-2 mRNA vaccination in haematopoietic stem cell transplantation patients. Br J Haematol. 2022;196(4):928931. 3. 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. 4. Moreau P, Le Gouill S, Bene MC, Chevallier P. Interest of a third dose of BNT162b2 anti-SARS-CoV-2 messenger RNA vaccine after allotransplant. Br J Haematol. 2022;196(5):e38-e40.

haematologica | 2022; 107(5)

CASE REPORTS Platelet-activating anti-PF4 antibodies mimic VITT antibodies in an unvaccinated patient with monoclonal gammopathy Transient prothrombotic disorders caused by plateletactivating antibodies against platelet factor 4 (PF4) include heparin-induced thrombocytopenia (HIT), spontaneous HIT syndrome,1 and, most recently, vaccine-induced immune thrombotic thrombocytopenia (VITT).2 Here, we identified prothrombotic, platelet-activating anti-PF4 antibodies, not associated with heparin treatment, in a patient with monoclonal gammopathy that resulted in a chronic hypercoagulability state. Mid-2019, a 79-year-old caucasian female with a history of unprovoked right-lower-limb deep-vein thrombosis (DVT) had experienced 1 year earlier thrombocytopenia and recurrent DVT with subsequent pulmonary embolism and stroke despite therapeutic anticoagulation (apixaban, 5 mg twice daily) (Figure 1A). Anticoagulation was switched to a vitamin K antagonist. Her platelet count was low (115×109/L) but had been within the normal range (250–330×109/L) in previous years. Prothrombin-international normalized ratio (PT-INR) values where within the therapeutic range and stable; however platelet count remained persistently low. In July 2020, she was re-admitted with pulmonary embolism (INR, 3.8; platelet count, 105×109/L; D-dimers, 1.5 mg/L). Anticoagulation was switched to low-molecular-weight heparin (enoxaparin, 1 mg/kg twice daily). Two weeks later, pulmonary embolism progressed with signs of rightventricular strain (platelet count, 81×109/L; D-dimers, 10.4 mg/L). Due to suspected HIT, anticoagulation was switched to fondaparinux (7.5 mg daily). Six weeks later she developed a frontal paramedian stroke (platelet count, 100×109/L). Since the end of 2020, the patient has been anticoagulated with apixaban (5 mg twice daily) and lowdose acetylsalicylic acid (100 mg daily), without new thromboembolic events as of September 2021. In 2020, repeated SARS-CoV-2 PCR analyses of nasopharyngeal swabs and antibodies against SARS-CoV-2 nucleocapsid, spike protein (receptor-binding domain), and trimeric spike protein were negative. The patient has not received a COVID-19 vaccine at time of reporting. In August 2020, IgG-specific PF4/heparin (HIT)enzyme-linked immunosorbant assay (ELISA) was positive (optical density >2.0 [reference range, <0.5]), while

functional testing excluded presence of heparin-dependent, platelet-activating antibodies. She also tested negative for antiphospholipid syndrome, JAK2 V617F mutation, and paroxysmal nocturnal hemoglobinuria. There was no evidence for underlying malignancy (negative gastro-/colonoscopy and computerized tomography [CT] imaging of abdomen/pelvis) or rheumatologic diseases (antinuclear/ds-DNA antibodies negative, no complement consumption) and bone marrow aspirate was without pathological findings. In June 2021, serum immunofixation electrophoresis revealed a monoclonal paraprotein of IgG-κ type (M-gradient was 9.6%), with IgG-specific PF4/heparin-ELISA remaining strongly positive (Figure 1A). We re-analyzed the patient serum sample of August 2020 in a washed platelet aggregation assay. As typically seen in VITT, patient serum induced platelet activation that was amplified by addition of PF4. In contrast, addition of heparin did not enhance patient serum-triggered platelet aggregation (Figure 1B). Together the data indicate VITT-like anti-PF4 antibodies. In order to confirm the existence of VITT-like anti-PF4 antibodies in this unvaccinated monoclonal gammopathy patient, we used the deglycosylated monoclonal anti-PF4 antibody (DG-1E12)— which binds the identical epitope on PF4 as VITT antibodies, without activating platelets.3 DG-1E12 interfered with VITT patient serum-driven platelet aggregation in the presence of PF4 (see control in Figure 1B) and also markedly inhibited PF4-dependent platelet aggregation induced by the gammopathy patient serum. We affinity purified anti-PF4 antibodies from the gammopathy patient serum.4 The k-light chain IgG monoclonal band (equal to paraprotein) strongly cross-reacted with immobilized PF4/heparin complexes in an ELISA (optical density >2.0). Similar to the monoclonal gammopathy patient serum, immunopurified antibodies also initiated platelet aggregation strictly in a PF4-dependent manner (Figure 1B). Thus, our patient’s IgG-k paraprotein shares similarities with pathologic VITT antibodies, by (i) binding to PF4 and (ii) activating platelets in a FcgIIa receptor-dependent mechanism, producing hypercoagulability. Our gammopathy patient, with a persisting PF4-reactive monoclonal IgG paraprotein that directly activates platelets leading to persistent thrombocytopenia and recurrent thrombosis, has a chronic hypercoagulability state that strongly correlates both with the degree of thrombocytopenia and D-dimer elevation (Figure 1C). A previous case of spontaneous HIT syndrome associated

Figure 1. Continued on following page.

haematologica | 2022; 107(5)

1219

Letters to the Editor

Figure 1: History of platelet counts, thromboembolic events, and PF4-based diagnostic tests in a patient with monoclonal gammopathy. (A) The shaded area indicates the normal reference range of peripheral platelet counts (150–400×109/L). (B) Patient serum or affinity purified anti-PF4 antibodies were tested with washed platelets from 3 healthy donors in the presence of buffer, low-molecular-weight heparin (reviparin, 0.2 aFXaU/mL), PF4 10 µg/mL, or deglycosylated anti-PF4 antibody DG-1E12 (100 µg/mL) in the functional heparin-induced platelet activation test (HIPA).4 The lag time until platelet aggregation occurred is indicated in minutes (min). Shorter lag time indicates stronger platelet activation. As reactivity between different platelet donors can vary, reactivity of the serum with each platelet preparation is given as one data point. Serum samples from August 2020 and November 2020 as well as affinity purified anti-PF4 antibodies induced platelet activation in the presence of PF4, but were negative following addition of buffer or heparin both at low (LMWH 0.2 aFXaU/mL, or high heparin (100 IU/mL; not shown). Also the monoclonal antibody IV.3 inhibited platelet activation in the presence of patient serum or the affinity purified anti-PF4 antibody fraction (data not shown). The serum of a vaccine-induced immune thrombotic thrombocytopenia (VITT) patient was used as positive control. (C) Correlation between peripheral platelet counts and plasma D-dimers during the course of treatment indicates that D-dimer levels increased when platelet counts decreased, a finding consistent with platelet count reduction due to procoagulant activation and consumption.DVT: deep vein thrombosis; PE: pulmonary embolism; VKA: vitamin K antagonist; LMWH: low-molecular-weight heparin; FPX: fondaparinux; ASA: acetylsalicylic acid; PF4: platelet factor 4; ELISA: enzyme-linked immunosorbent assay; OD: optical density; FEU: fibrinogen equivalent units; aFXaU/mL: anti-factor Xa activity in units/mL; DG-E12: deglycosylated monoclonal antibody E12; Ig: immunoglobulin.

with IgG-k paraprotein has been reported (although in that patient PF4-dependent reactivity profile was not reported).5 In conclusion, PF4-dependent platelet-activating antibodies causing chronic thrombocytopenia and persisting hypercoagulability may underly chronic prothrombotic disorders such as monoclonal gammopathy. The spectrum of anti-PF4 antibody mediated hypercoagulability states should be extended beyond heparin (HIT) and vaccine (VITT) exposure to some paraproteins in neoplastic disease. Andreas Greinacher,1* Florian Langer,2* Linda Schönborn,1 Thomas Thiele,1 Munif Haddad,3 Thomas Renné,3,4,5 Jerome Rollin,6,7 Yves Gruel6,7# and Theodore E. Warkentin8# 1 Institut für Immunologie und Transfusionsmedizin, Abteilung Transfusionsmedizin, Universitätsmedizin Greifswald, Greifswald, Germany; 2Zentrum für Onkologie, II. Medizinische Klinik und Poliklinik, Universitätsklinikum Eppendorf, Hamburg, Germany; 1220

Institute of Clinical Chemistry and Laboratory Medicine, University Medical Center Hamburg-Eppendorf, Hamburg, Germany; 4Center for Thrombosis and Hemostasis (CTH), Johannes Gutenberg University Medical Center, Mainz, Germany; 5Irish Center for Vascular Biology, School of Pharmacy and Biomolecular Sciences, Royal College of Surgeons in Ireland, Dublin, Ireland; 6Regional University Hospital Centre Tours, Department of Hemostasis, Tours, France; 7University of Tours, EA7501 GICC, Tours, France and 8 Deptarment of Pathology and Molecular Medicine, and Department of Medicine, McMaster University, Hamilton, Ontario, Canada *AG and FL contributed equally as co-first authors # YG and TEW contributed equally as co-senior authors Correspondence: ANDREAS GREINACHER- andreas.greinacher@med.uni-greifswald.de doi:10.3324/haematol.2021.280366 Received: November 16, 2021. Accepted: December 17, 2021. haematologica | 2022; 107(5)

Case Reports

Pre-published: December 30, 2021. Disclosures: no conflicts of interest to disclose. Contributions: AG, FL, YG, TEW developed the concept; TEW collected data on medical spontaneous HIT syndrome; MH and TR performed the immuno-electrophoresis; YG and JR developed and provided the 1E12 monoclonal antibody and its deglycosylated form AG, TT, LS performed the laboratory studies; FL took care for the patient; AG, FL, TEW and YG wrote the manuscript; AG and FL verified the underlying data. All authors critically revised and approved the final version of the manuscript. Funding: the study has been supported by Deutsche Forschungsgemeinschaft, grant/award numbers: 374031971-TRR 240 and KFO306.

References 1. Warkentin TE, Greinacher A. Spontaneous HIT syndrome: knee

haematologica | 2022; 107(5)

replacement, infection, and parallels with vaccine-induced immune thrombotic thrombocytopenia. Thromb Res. 2021;204:40-51. 2. Greinacher A, Thiele T, Warkentin TE, Weisser K, Kyrle PA, Eichinger S. Thrombotic thrombocytopenia after ChAdOx1 nCov19 vaccination. N Engl J Med. 2021;384(22):2092-2101. 3. Vayne C, Nguyen TH, Rollin J, et al. Characterization of new monoclonal PF4-specific antibodies as useful tools for studies on typical and autoimmune heparin-induced thrombocytopenia. Thromb Haemost. 2021;121(3):322-331. 4. 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. 5. Faille D, Hurtado-Nedelec M, Ouedrani A, et al. Isolation of a monoclonal IgG kappa with functional autoantibody activity against platelet factor 4/heparin from a patient with a monoclonal gammopathy of undetermined significance and clinically overt heparin thrombocytopenia. Res Pract Thromb Haemost. 2017;1(Suppl 1):S1355.

1221

Case Reports

Kikuchi-Fujimoto disease associated with hemophagocytic lymphohistiocytosis following the BNT162b2 mRNA COVID-19 vaccination Kikuchi-Fujimoto disease (KD) is a self-limiting histiocytic necrotizing lymphadenitis. One case of KD following administration of the BNT162b2 mRNA COVID-19 vaccine was recently reported.1 Hemophagocytic lymphohistiocytosis (HLH) is a life-threatening hyperinflammatory state brought on by uncontrolled histiocytes, macrophages and T-cell activation, which have also been occasionally observed after BNT162b2 vaccination.2 KD and HLH present overlapping pathogenesis and symptoms, and their association has been previously described in children and adult patients.3,4 Here, we report the first case of KD associated with HLH following the BNT162b2 mRNA COVID-19 vaccination. A 38-year-old previously healthy woman was admitted to the hospital with a history of a fever of 40°C for more than ten days, associated with chills and fatigue. She presented with a diffuse cutaneous eruption of erythematous papules, which were subsequently confluent (Figure 1, AB). She had previously been treated with antibiotics (amoxicillin and clavulanic acid, followed by teicoplanin and doxycycline), with no improvement. The first dose of

the BNT162b2 mRNA COVID-19 vaccine was administered two months after giving birth to a healthy baby. Three weeks before the onset of fever, the patient received a second dose of the vaccine, inoculated in the same left arm as the first dose. At hospital admission, her nasopharyngeal swab for SARS CoV-2 PCR was negative; SARS CoV-2 IgG antibodies were positive (>2.080 BAU/mL; cut-off: 33.8 BAU/mL; LIAISON SARS-CoV-2 TrimericS IgG, Diasorin, Saluggia, Italy). Her physical exam was normal except for the cutaneous rash and multiple enlarged tender lymph nodes in the left axillary zone. This was confirmed by a contrastenhanced computed tomography (CT) examination (Figure 1, C-D). Laboratory tests showed bi-cytopenia with leukopenia and anemia (neutrophil count, 0.9 x 109/L; lymphocyte count, 0.3 x 109/L, hemoglobin, 9.8 g/L), increased lactate dehydrogenase and transaminase levels, high serum ferritin levels (500 µg/L), mild hypertriglyceridemia (225 mg/dL) and normal fibrinogen. Her serum soluble interleukin-2 receptor (IL-2R) level was increased to 2.610 U/mL (normal value 223-710) and her natural killer (NK) cell count was low (<35 cells/µL; normal value 200-400). IL-6, IL-8 and IL-10 levels were normal, but the tumor necrosis factor (TNF) level was increased. The patient underwent a bone marrow aspiration and trephine proce-

Figure 1. Patient’s clinical and radiological findings. A and B) Diffuse maculo-papular skin rash; C and D) CT scan showing left axillary lymph node enlargement.

1222

haematologica | 2022; 107(5)

Case Reports

dure, which showed hemophagocytosis (Figure 2, A-C). A diagnosis of HLH was confirmed based on the fulfilment of 6/8 HLH-2004 diagnostic criteria.5 In addition, an excisional lymph node biopsy of the left axillary was performed. Histopathological examination of the lymph node revealed histiocytic necrotizing lymphadenitis, characterized by paracortical, well-circumscribed necrotic areas with karyorrhexis and fibrin deposits. Immunohistochemistry revealed numerous CD68-positive histiocytes as well as several CD3-positive T cells and few CD20-positive B cells. All these features are considered typical of KD (Figure 2, D-F). Further analysis did not show ongoing infections of HCV, HBV hepatitis virus, HIV, toxoplasma, rubeovirus,

brucella, leptospirosis, bartonella, chlamydia, morbillivirus, mycoplasma or yersinia. The tuberculosis Quantiferon test was also negative. Epstein Barr virus (EBV), Parvovirus B19, cytomegalovirus, JC virus and herpes-6 DNA were also absent. Antinuclear antibodies (ANA), antibodies to double-stranded DNA (antidsDNA), and antibodies to extractable nuclear antigens (anti-ENA) were absent. The rheumatoid factor test was negative and both complement C3 and C4 serum levels were normal. The patient was promptly initiated on steroids (methylprednisolone 1 mg/kg i.v.), in adherence with the recommended treatment for HLH. Sudden fever lysis was observed and clinical conditions improved. Methyl-pred-

Figure 2. Lymph node and bone marrow histological findings. A) Increased diffusion of histiocytes (brown color) in the bone marrow trephine biopsy; CD68 (PGM1) stain, 10X magnification; B) An histiocyte showing phagocytosis of hemopoietic elements; CD68 (PGM1) stain, 60X magnification; C) Histiocytes showing phagocytosis of red cells, platelets, erythroid precursor cells; hematoxylin-eosin stain, 60X magnification; D) A low power view of the lymph node showing paracortical necrotizing zones (N) in an otherwise preserved architecture; hematoxylin-eosin stain; E) The paracortical necrotizing areas were characterized by many reactive T lymphocytes, as shown by immunohistochemistry for CD3; 10X magnification; F) The immunostaining for CD68 revealed numerous histiocytes within the necrosis; 10X magnification.

haematologica | 2022; 107(5)

1223

Case Reports

nisone was substituted with prednisone (1 mg/day) at hospital discharge one week later. This schedule was maintained for a further two weeks, reduced to a half dosage for three more weeks, then tapered over the following three weeks. IL-2R levels decreased to 1.170 U/mL and 740 U/mL after two and three weeks of steroid treatment, respectively, and reached normal levels (460 U/mL) after four weeks. NK cells also returned to normal levels. The patient is doing well and follow-up is still ongoing. To the best of our knowledge, this is the first case of KD associated with HLH following COVID-19 vaccination. HLH is a severe hyperinflammatory syndrome, secondary to aberrant cytokine production and uncontrolled histiocyte activation, following infections, in particular EBV, hematological malignancies, autoimmune diseases, hematopoietic stem cells or organ transplantation. It is characterized by cytopenia, unremitting fever, hepatic dysfunction and a fatal multiple organ failure without early recognition and prompt treatment. The first line of treatment is steroids and/or immunoglobulins, with etoposide added in poor responders.5 KD is a self-limiting histiocytic necrotizing lymphadenitis that commonly occurs in Asia, although it is present globally.6 It is characterized by fever, lymphadenopathy and leukopenia, may be associated with a skin rash and is mainly characterized by transient red, millet-sized maculopapules.7 Since no skin biopsy was carried out, it remains unclear whether the patient’s skin rash was associated with KD or was a possible adverse effect of teicoplanin (Figure 1, A-B). KD pathogenesis is unknown, although it is believed to be a consequence of an aberrant T cells and histiocyte immune response to an immunogenic antigen. An association with organisms such as toxoplasma, cytomegalovirus, varicella-zoster virus, EBV, human herpes virus-6, HIV, and yersinia enterocolitica, has been suggested, although convincing evidence has yet to be presented. Moreover, a possible association with autoimmune disorders, including antiphospholipid antibody syndrome and mixed connective tissue disease, has been reported.1 Steroids and immunoglobulins treatment may be beneficial. The association between HLH and KD has been described in both children and adult patients and has a potentially fatal outcome if left untreated.3,4,8 Some authors have reported that that CD8+ T lymphocytes in patients with HLH-KD may be excessively activated, altering the course of the self-limited KD progress and resulting in HLH.8 Both HLH and KD following BNT162b2 mRNA COVID-19 vaccination have recently been described.1,2 Notably, after mass vaccination against COVID-19, a vaccine-associated hypermetabolic lymphadenopathy (VAHL) in the axillary or supraclavicular lymph nodes, ipsilateral to the vaccination site, has been reported.9,10,11 This is frequently observed after BNT162b2 administration, with higher intensities following the booster dose, and lasting until three weeks after vaccination. We hypothesize that our patient first developed a left-axillary VHAL following two ipsilateral vaccine dose inoculations, followed by a systemic inflammatory response syndrome (SIRS) with KD features in the axillary lymph nodes, and HLH symptoms as a systemic inflammatory reaction. Dermal histiocytes and macrophages represent the first resident antigen-presenting cell (APC) transfected by the mRNA vaccine that presents antigenic peptides on major histocompatibility (MHC) class I and MHC class II molecules of CD4+ and CD8+ cells, resulting in immune response expansion. Additionally, intramuscular vaccine injection leads to a local increase in proinflamma1224

tory cytokines, which form an immune-stimulatory environment in draining axillary lymph nodes.12 It remains unclear whether an alternate arm inoculation in our patient might have been less immune-reactive. In conclusion, we have described the first case of KD associated with HLH following COVID-19 vaccination. This is a rare event and does not compromise the safety and efficacy of the BNT162b2 mRNA vaccine in the fight against COVID-19. Physicians should be aware of rare systemic inflammatory reactions that require early diagnosis and treatment. Giovanni Caocci,1 Daniela Fanni,2 Mariagrazia Porru,3 Marianna Greco,1 Sonia Nemolato,2 Davide Firinu,3 Gavino Faa,2 Angelo Scuteri3 and Giorgio La Nasa1 1 Department of Medical Sciences and Public Health, Haematology, University of Cagliari, Businco Hospital; 2 Department of Medical Sciences and Public Health, Pathology, University of Cagliari, S.Giovanni di Dio and Businco Hospital and 3Department of Medical Sciences and Public Health, Internal Medicine, University of Cagliari, Policlinico Hospital, Cagliari, Italy Correspondence: GIOVANNI CAOCCI - giovanni.caocci@unica.it doi:10.3324/haematol.2021.280239 Received: October 25, 2021. Accepted: December 20, 2021. Pre-published: December 30, 2021. Disclosures: no conflicts of interest to disclose. Contributions: GC, GF, AS and GLN conceived and designed the study; GC, MP, DF, AS and GLN managed patients; DF, SN and GF carried out the pathological analysis; MG carried out the immunophenotype and interleukin analysis: GC wrote the manuscript; GC, DF, MP, DF, MG, SN, GF, AS, GLN approved the final draft of the manuscript. Acknowledgments: we thank Dr. Mariano Cabiddu and Dr. Giovanna Manconi for patient management in the early phases; Prof. Fabio Medas for performing lymph node biopsy; Dr. Salvatore Labate for CT images and Dr. Valeria Fresu for interleukin analysis. Compliance with ethical standards: all procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards Data sharing statement: the data that support the findings of this study are available from the corresponding author, upon reasonable request. Informed consent: informed consent was obtained from the patient, including the publication of personal images..

References 1. Soub HA, Ibrahim W, Maslamani MA, A Ali G, Ummer W, AbuDayeh A. Kikuchi-Fujimoto disease following SARS CoV2 vaccination: Case report. IDCases. 2021;25:e01253. 2. Tang LV, Hu Y. Hemophagocytic lymphohistiocytosis after COVID19 vaccination. J Hematol Oncol. 2021;14(1):87. 3. Nishiwaki M, Hagiya H, Kamiya T. Kikuchi-Fujimoto disease complicated with reactive hemophagocytic lymphohistiocytosis. Acta Med Okayama. 2016;70(5):383-388. 4. Duan W, Xiao Z-H, Yang L-G, Luo H-Y. Kikuchi’s disease with hemophagocytic lymphohistiocytosis: a case report and literature review. Medicine (Baltimore). 2020;99(51):e23500. 5. Henter J-I, Horne A, Aricó M, et al. HLH-2004: Diagnostic and therapeutic guidelines for hemophagocytic lymphohistiocytosis. Pediatr Blood Cancer. 2007;48(2):124-131. 6. Perry AM, Choi SM. Kikuchi-Fujimoto Disease: A Review. Arch Pathol Lab Med. 2018;142(11):1341-1346.

haematologica | 2022; 107(5)

Case Reports

7. Resende C, Araújo C, Duarte M da L, Vieira AP, Brito C. Kikuchi’s disease of the xanthomathous type with cutaneous manifestations. An Bras Dermatol. 2015;90(2):245-247. 8. Yang Y, Lian H, Ma H, et al. Hemophagocytic lymphohistiocytosis associated with histiocytic necrotizing lymphadenitis: A clinical study of 13 children and literature review. J Pediatr. 2021;229:267274. 9. Cohen D, Krauthammer SH, Wolf I, Even-Sapir E. Hypermetabolic lymphadenopathy following administration of BNT162b2 mRNA Covid-19 vaccine: incidence assessed by [18F]FDG PET-CT and relevance to study interpretation. Eur J Nucl Med Mol Imaging.

haematologica | 2022; 107(5)

2021;48(6):1854-1863. 10. Steinberg J, Thomas A, Iravani A. 18F-fluorodeoxyglucose PET/CT findings in a systemic inflammatory response syndrome after COVID-19 vaccine. Lancet. 2021;397(10279):e9. 11. Adin ME, Isufi E, Kulon M, Pucar D. Association of COVID-19 mRNA vaccine with ipsilateral axillary lymph node reactivity on imaging. JAMA Oncol. 2021;7(8):1241-1242. 12. Mascellino MT, Di Timoteo F, De Angelis M, Oliva A. Overview of the main anti-SARS-CoV-2 vaccines: mechanism of action, efficacy and safety. Infect Drug Resist. 2021;14:3459-3476.

1225

Case Reports

Zanubrutinib, rituximab and lenalidomide induces deep and durable remission in TP53-mutated B-cell prolymphocytic leukemia: a case report and literature review B-cell prolymphocytic leukemia (B-PLL) is a rare lymphoid neoplasm accounting for approximately 1% of all cases of lymphocytic leukemia. In this disease, TP53 abnormalities are found in more than half of the cases and about 50% of patients have MYC abnormalities (rearrangement and/or increased copy number). Similar to chronic lymphocytic leukemia (CLL), shortened survival in B-PLL is associated with TP53 mutations. Due to the rarity of this disease, most therapeutic approaches have been executed according to CLL guidelines. Specifically, Bruton’s tyrosine kinase (BTK) inhibitors show significant efficacy in CLL patients with 17p deletion/TP53 mutation. However, little is known about the treatment outcome of BTK inhibitors (BTKi) in B-PLL. Here, we report for the first time the efficacy of a nextgeneration BTKi, zanubrutinib, combined with rituximab and lenalidomide (ZR2), in a B-PLL patient with TP53 and MYC abnormalities. A 52-year-old man visited our hospital in October 2020 with a 3-year history of high white blood cell (WBC) count and splenomegaly. Physical examination revealed multiple palpable lymphadenopathies (bilateral neck, bilateral supraclavicular, bilateral axillary, and bilateral inguinal regions) and massive splenomegaly (19 cm below left costal margin). Complete blood count showed an elevated WBC count of 31.4×109/L, hemoglobin concentration of 134 g/L and platelet count of 87×109/L. Peripheral blood (PB) smear revealed 88% of prolymphocytes. Serum lactate dehydrogenase (313 U/L, normal <250 U/L) and β2-microgloubulin (5.42 mg/L, normal <2.8 mg/L) were elevated. The concentration of serum monoclonal immunoglobulin G (IgG) with λ light chain, detected by immunofixation electrophoresis, was 2.8 g/L. Abdominal ultrasound showed splenomegaly (24.6×8.2 cm, 19 cm below left costal margin). 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography/computed tomography (PET/CT) scan showed slightly increased FDG metabolism of lymph nodes in the neck, axilla, mediastinum, retroperitoneum, abdominal cavity, pelvic cavity, and groin (the largest diameter is 1.5 cm, SUVmax 4.3). It also demonstrated increased FDG metabolism of spleen (SUVmax 4.3). A bone marrow (BM) biopsy showed B-PLL representing 75% of the marrow cells. Flow cytometry showed these cells were positive for CD19,

CD79a, FMC7, CD81, CD22, CD20, and restricted monoclonal λ light chain, together with weakly positive for CD23, CD25, CD38, CD200, surface immunoglobulin M (sIgM), and negative for CD5, CD10, CD43, CD71, CD123, CD103, CD11c, surface IgD (sIgD) and κ light chain. These cells were stained negative for cyclin D1. Immunoglobulin heavy chain (IGH) somatic hypermutation analysis showed mutated IGH variable region genes. Cytogenetics revealed t(8;14) and fluorescense in situ hybridization (FISH) showed MYC gene rearrangement without CCND1/IgH rearrangement which exclude the diagnosis of mantle cell lymphoma (MCL). Molecular studies showed mutations in both MYC and TP53 genes. Based on cell morphology, histopathology, immunohistochemistry, genetic analysis, and clinical features, the patient was diagnosed as B-PLL. This patient was treated with ZR2 regimen (zanubrutinib, 160 mg twice daily on day 1-21; lenalidomide, 25 mg once daily on day 1-14; rituximab, 375 mg/m2 on day 1) every 28 days. Following the initiation of ZR2 treatment, the patient experienced resolution of splenomegaly, with the WBC decreasing from 31.4×109/L to 8.43×109/L after one cycle treatment (Table 1). Minimal residual disease (MRD) negative complete remission (CR) (by PB flow cytometry and PET-CT scan) was achieved after four cycles of ZR2 treatment, with monoclonal IgG disappearance. The patient refused allogeneic hematopoietic stem cell transplantation (HSCT). After 6 cycles of ZR2 treatment, MRD negative CR was further verified by PET-CT and flow cytometric analysis of bone marrow aspirates. Subsequently, the patient received two cycles of ZR (zanubrutinib, 160 mg twice daily on day 1-21; rituximab, 375 mg/m2 on day 1) as consolidation therapy. Thereafter, zanubrutinib and lenalidomide (zanubrutinib, 160 mg twice; lenalidomide, 25 mg once daily on day 114; administered every 28 days) were used as maintenance therapy. Moreover, the patient has been well and has remained in sustained MRD-negative CR for 12 months by now. The most hematological adverse events were common grade 1-2 neutrophil count decrease, which can be recovered in a few days, and no significant non-hematological adverse events such as nausea and fatigue were noted. There are an estimated 120 new cases of B-PLL per year in the United States and prospective clinical trials are currently not available. There is neither clear expert consensus nor are there guidelines for the treatment of BPLL, and the treatment according to CLL are frequently recommended as upfront therapy. In patients with TP53 mutation and/or deletion, alemtuzumab was an effective

Table 1. Treatment course.

Time point Baseline Cycle 1 Cycle 2 Cycle 3 Cycle 4 Cycle 5 Cycle 6 Cycle 7 Cycle 8

ZR2 ZR2 ZR2 ZR2 ZR2 ZR2 ZR ZR

WBC count (x109/L)

ALC (x109/L)

31.4 8.43 4.23 3.52 3.06 3.69 3.04 4.62 2.41

27.6 5.10 2.67 1.0 1.0 1.27 1.29 1.59 0.83

Response evaluation LDH (U/L) PET/CT 313 245 362 255 192 184 196 220 206

Spleen, lymph node N/A N/A N/A CR N/A N/A N/A N/A

PB MRD

BM MRD

88% N/A 14.26% NEG NEG NEG N/A N/A NEG

75% N/A N/A N/A N/A N/A NEG N/A N/A

WBC: white blood cell count; ALC: absolute lymphocyte count; LDH: lactate dehydrogenase; PB: peripheral blood; BM: bone marrow; MRD: measurable residual disease; N/A: not available; NEG: negative; CR: complete remission; ZR2: zanubrutinib+lenalidomide+rituximab; ZR: zanubrutinib+rituximab.

1226

haematologica | 2022; 107(5)

Case Reports

Ibrutinib combination

Ibrutinib monotherapy

Table 2. Ibrutinib treatment in patients with B-cell prolymphocytic leukemia.

Reference

Age, Cytogenetics years

Gordon et al., 20172

73, 77

del(17p) del(13q)

Algrin et al., 20173

del(17p)

Coelho et al., 20174

Damlaj et al., 20185

Bindra et al., 20196

Patil et al., 20197

Christoforidou et al., 20208

George et al., 20209

Oka et al.,202010

Moore et al.,202011

Siddiqui et al.,202112

Patient 1: single dose R, Ibrutinib started on day 40 Patient 2: single dose BR, Ibrutinib started on day 44 Ibrutinib monotherapy

Outcome

del(17p), gain MYC

Ibrutinib treatment-related adverse events

Patient 1: CCR, PFS: 1. Mild fatigue and easy bruising, 15 months 2. Atrial fibrillation Patient 2: CCR, PFS: 12 months 3CR/1PR, Median PFS: 9 months After allo-HSCT: CR, PFS: Transient lymphocytosis 10 months

del(17p), 1. Idelalisib-rituximab del(11q), for 5 months del(13q), 2. Ibrutinib for 2 months trisomy 12, 3. Allo-HSCT t(11; 14) (reduced intensity) MYC Ibrutinib monotherapy PFS: 8 months rearrangement del(17p), 1. Ibrutinib for 12 months 1. Response to del(13q) 2. Venetoclax and ibrutinib for 12 months leukapheresis for 2. Patient went to 5 weeks hospice care after Venetoclax TP53 deletion 1. Alemtuzumab 1. 3rd line Ibrutinib for 18 months (hematological PR, 2. Idelalisib-rituximab PFS: 12months ) for 12 months 2. OS: 5 years 3. Ibrutinib for 12 months 4. Venetoclax for 8 months del(17p) 1. BR for 4 months 1. 2nd line Ibrutinib 2. Ibrutinib for 5 months (hematological PR, 3. Venetoclax for 6 months PFS: 5 months) 4. Idelalisib-rituximab 2. OS: 2.5 years for 10 months t(4;14) Ibrutinib monotherapy PR, PFS: 15 months (p16.3;q32) , TP53 mutation del(17p) Low dose CR, PFS: 12 months Ibrutinib monotherapy

median TP53 age: disruption 67.3 years

Treatment

IRA(n=2); IR(n=2); Ibrutinib (n=2)

Response criteria iwCLL

iwCLL Author’s report

Easy cutaneous bruising, transient lymphocytosis -

Author’s report Author’s report

Transient lymphocytosis

Author’s report

Atrial fibrillation, gastric hemorrhage, transient lymphocytosis

Author’s report

Transient lymphocytosis

Author’s report

IRA (2 CCR) Recurrent urinary tract IR (1 CCR,1 PR) infections (1), cutaneous Nocardia Ibrutinib (1 PR,1 SD) infection (1), cytomegalovirus Median PFS: 34.7 months reactivation (1), musculoskeletal pain (3), atrial fibrillation (1), stomatitis requiring dose reduction (1) 1. Ibrutinib for 10 months After IV: CR; 2. IR for 3 months Total PFS: 4.5 years 3. IV for 3 years

iwCLL

R: rituximab; BR: bendamustine + rituximab; IRA: ibrutinib+rituximab+alemtuzumab; IR: ibrutinib+rituximab; I: ibrutinib monotherapy; IV: ibrutinib+venetoclax; iwCLL: international workshop on chronic lymphocytic leukemia; CCR: patients without restaging bone marrow biopsies or imaging, meeting iwCLL clinical and laboratory complete remission criteria were considered to have a clinical complete remission (CCR); PR: partial response; CR: complete response; PFS: progression-free survival; OS: overall survival.

therapeutic option for these patients, despite showing a short reaction time. In a small series of idelalisib plus rituximab in B-PLL, responses were seen in all five patients, which lasts more than 6 months at the time of the report.1 BTKi is known to promote high response rates, leading to durable remissions in all genetic subsets of CLL patients including patients with TP53 abnormalities. As shown in Table 2, ibrutinib has shown efficacy in individual case reports and small case series studies in B-PLL patients.2-12 Zanubrutinib is a new-generation, irreversible BTKi demonstrating high selectivity and low toxicities. haematologica | 2022; 107(5)

Zanubrutinib has demonstrated single agent safety and efficacy in B-cell malignancies including CLL, lymphocytic lymphoma (LPL) and MCL in several clinical trials. Lenalidomide is an immunomodulatory drug which has direct anti-tumor activity and indirect effects by enhancing anti-tumor immune responses. Lenalidomide can downregulate the expression of MYC and its target genes. Chamuleau et al. conducted a prospective R2CHOP study on newly diagnosed MYC rearrangementpositive DLBCL patients which was safe and achieved 67% complete metabolic response.13 Bühler et al. reported that in a study of relapsed/refractory CLL patients treated 1227

Case Reports

with lenalidomide, there was no statistically significant difference in OS between TP53-mutant and wild-type patients.14 Additionally, for CLL patients who have high risk factors including TP53 mutation, the application of lenalidomide maintenance therapy can bring significant clinical benefits.15 Lenalidomide can enhance antibody-dependent cellular cytotoxicity (ADCC), and antibody-dependent cellmediated phagocytosis (ADCP) of rituximab, showing rational combination strategy with rituximab. Several studies have shown efficacy and safety of the combination of BTKi (ibrutinib or zanubrutinib) with R2 (lenalidomide and rituximab) and with or without chemotherapy (ibrutinib+R2+CHOP, clinicaltrials gov. Identifier: NCT02077166; ibrutinib+R2, clinicaltrials gov. Identifier: NCT02460276; ZR2+CHOP, clinicaltrials gov. Identifiers: EHA2021 EP548), to treat DLBCL and MCL. In a phase II trial (clinicaltrials gov. Identifier: NCT04460248) previously untreated elderly patients with DLBCL will be treated with ZR2 regimen. Taken together, lenalidomide showed anti-tumor potential for MYC rearrangement-positive DLBCL and TP53-mutant CLL, in addition, ibrutinib+R2 showed efficacy for MCL, which may have similar disease characteristics as B-PLL. This patient has both TP53 and MYC mutations along with MYC rearrangement, which predicted the poor outcome with short survival period by conventional chemoimmunotherapy. Therefore, we employed ZR2 as a first line therapy for this patient. As far as we know, this is the first case report to document a successful treatment outcome with ZR2 as upfront therapy for a B-PLL patient. Although effective standard treatment strategies have not yet been established for patients with B-PLL, we here demonstrate that ZR2 regimen induces a deep and durable response in one B-PLL patient with TP53 and MYC mutations along with MYC rearrangement. Given the poor prognosis of B-PLL and lack of effective established treatment modalities, this case report could represent a promising indication of ZR2 for B-PLL treatment. Further investigations in large cohort will be needed to characterize the efficacy, safety, and tolerability of this combination treatment. Lijie Xing, Qiang He, Linna Xie, Hui Wang and Zengjun Li Department of Hematology, Shandong Cancer Hospital and Institute, Shandong First Medical University and Shandong Academy of Medical Sciences, Jinan, China Correspondence: ZENGJUN LI - zengjunli@163.com doi:10.3324/haematol.2021.280259 Received: October 26, 2021. Accepted: December 16, 2021. Pre-published: December 23, 2021. Disclosures: no conflicts of interest to disclose. Contributions: LX and ZL designed the study, performed treatments, collected and analyzed data, and wrote the manuscript; QH, LX and HW collected data on clinical follow-up. All authors approved the final version of the manuscript.

1228

Acknowledgements: we are grateful to Teru Hideshima and Kenneth Wen at the Dana-Farber Cancer Institute for expert assistance in the revision and editing of the manuscript. Funding: this investigation was supported by the grant ZR2021MH072 to LX from Shandong Provincial Natural Science Foundation, China. Ethics approval and consent to participate: this study was approved by the Medical Ethical Committee of Shandong Cancer Hospital and Institute. All patients’ samples were obtained with informed consent in accordance with the Declaration of Helsinki. Data sharing statement: the data sets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

References 1. Eyre TA, Fox CP, Boden A, et al. Idelalisib-rituximab induces durable remissions in TP53 disrupted B-PLL but results in significant toxicity: updated results of the UK-wide compassionate use programme. Br J Haematol. 2019;184(4):667-671. 2. Gordon MJ, Raess PW, Young K, et al. Ibrutinib is an effective treatment for B-cell prolymphocytic leukaemia. Br J Haematol. 2017;179(3):501-503. 3. Algrin C, Chapiro E, Saviellis J, et al. B-Cell prolymphocytic leukemia, a rare lymphoproliferative disorder: analysis of 35 cases, a study on behalf of the French Innovative Leukemia Organization (FILO). Blood. 2017;130(Suppl 1):S4310. 4. Coelho H, Badior M, Melo T. Sequential kinase inhibition (idelalisib/ibrutinib) induces clinical remission in B-cell prolymphocytic leukemia harboring a 17p deletion. Case Rep Hematol. 2017;2017:8563218. 5. Damlaj M, Al Balwi M, Al Mugairi AM. Ibrutinib therapy is effective in B-cell prolymphocytic leukemia exhibiting MYC aberrations. Leuk Lymphoma. 2018;59(3):739-742. 6. Bindra BS, Kaur H, Portillo S, et al. B-cell prolymphocytic leukemia: case report and challenges on a diagnostic and therapeutic forefront. Cureus. 2019;11(9):e5629. 7. Patil N, Went RG. Venetoclax is an option in B-cell prolymphocytic leukaemia following progression on B-cell receptor pathway inhibitors. Br J Haematol. 2019;186(4):e80-e82. 8. Christoforidou A, Bezirgiannidou Z, Vrachiolias G, et al. B-cell prolymphocytic leukemia successfully treated with B-cell receptor antagonists, but resistant to venetoclax. Leuk Lymphoma. 2020;61(3):749-752. 9. George P, Brown A, Weinkove R. B-cell prolymphocytic leukaemia with a t(4;14) FGFR3/IGH translocation: response to ibrutinib. Pathology. 2020;52(4):491-492. 10. Oka S, Ono K, Nohgawa M. Effective upfront treatment with low-dose ibrutinib for a patient with B cell prolymphocytic leukemia. Invest New Drugs. 2020;38(5):1598-1600. 11. Moore J, Baran AM, Meacham PJ, et al. Initial treatment of B-cell prolymphocytic leukemia with ibrutinib. Am J Hematol. 2020;95(5):E108-E110. 12. Shindiapina P, Brown JR, Danilov AV. A new hope: novel therapeutic approaches to treatment of chronic lymphocytic leukaemia with defects in TP53. Br J Haematol. 2014;167(2):149-161. 13. Chamuleau MED, Burggraaff CN, Nijland M, et al. Treatment of patients with MYC rearrangement positive large B-cell lymphoma with R-CHOP plus lenalidomide: results of a multicenter HOVON phase II trial. Haematologica. 2020;105(12):2805-2812. 14. Buhler A, Wendtner CM, Kipps TJ, et al. Lenalidomide treatment and prognostic markers in relapsed or refractory chronic lymphocytic leukemia: data from the prospective, multicenter phase-II CLL-009 trial. Blood Cancer J. 2016;6(3):e404. 15. Fink AM, Bahlo J, Robrecht S, et al. Lenalidomide maintenance after first-line therapy for high-risk chronic lymphocytic leukaemia (CLLM1): final results from a randomised, doubleblind, phase 3 study. Lancet Haematol. 2017;4(10):e475-e486.

haematologica | 2022; 107(5)