Haematologica, Volume 109, Issue 1 by Haematologica

VOL.

haematologica

109

Journal of the Ferrata Storti Foundation

JANUARY 2024

haematologica.org

ISSN 0390 - 6078

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

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

Associate Editors Shai Izraeli (Tel Aviv), Steve Lane (Brisbane), Pier Mannuccio Mannucci (Milan), Jessica Okosun (London), Pavan Reddy (Ann Arbor), David C. Rees (London), Paul G. Richardson (Boston), Francesco Rodeghiero (Vicenza), Gilles Salles (New York), Kerry Savage (Vancouver), Aaron Schimmer (Toronto), Richard F. Schlenk (Heidelberg)

Statistical Consultant Catherine Klersy (Pavia)

AI Consultant Jean Louis Raisaro (Lausanne)

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), Ahmet Dogan (New York), Mary Eapen (Milwaukee), Francesca Gay (Torino), Ajay Gopal (Seattle), Alex Herrera (Duarte), Martin Kaiser (London), Marina Konopleva (Houston), Nicolaus Kröger (Hamburg), Austin Kulasekararaj (London), Shaji Kumar (Rochester), Ann LaCasce (Boston), Matthew J. Mauer (Rochester) Neha MehtaShah (St. Louis), Moshe Mittelman (Tel Aviv), Alison Moskowitz (New York), Yishai Ofran (Haifa), Farhad Ravandi (Houston), John W. Semple (Lund), Liran Shlush (Toronto), Sarah K. 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), Debora Moscatelli (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), Anne Freckleton (English Editor), Rosangela Invernizzi (Scientific Consultant), Marianna Rossi (Scientific Consultant), Massimo Senna (Information Technology), Luk Cox (Graphic Artist)

Haematologica | 109 - January 2024

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 | 109 - January 2024

Table of Contents Volume 109, Issue 1: January 2024

About the Cover Image taken from the editorial by Isabelle Plo and Iléana Antony-Debré in this issue.

Landmark Papers in Hematology 1

The first human acute myeloid leukemia genome ever fully sequenced Brunangelo Falini https://doi.org/10.3324/haematol.2022.282118

Power out chronic lymphocytic leukemia: unplugging OXPHOS/mTOR pathways to overcome venetoclax resistance Elodie Viry and Jerome Paggetti https://doi.org/10.3324/haematol.2023.283847

Is it time for age and clinically adjusted minimal residual disease interpretation in acute myeloid leukemia? Yishai Ofran https://doi.org/10.3324/haematol.2023.283693

Moving low-risk myelodysplastic syndromes from humans to mice: is it truly that simple? Daniel Coriu and Maria-Camelia Stancioaica https://doi.org/10.3324/haematol.2023.283652

Patient-reported treatment response in chronic graft-versus-host disease Hildegard T. Greinix https://doi.org/10.3324/haematol.2023.283524

The double-edged sword of adenosine Isabelle Plo and Iléana Antony-Debré https://doi.org/10.3324/haematol.2023.283469

Secondary acute myeloid leukemia: time to turn the page Christian Récher https://doi.org/10.3324/haematol.2023.283468

Exploring the role of viral hepatitis in plasma cell disorders Elizabeth O’Donnell https://doi.org/10.3324/haematol.2023.283461

BRCA1/2 mutations and de novo hematologic malignancies: true, true and not clearly related Payal D. Shah and Katherine L. Nathanson https://doi.org/10.3324/haematol.2023.283348

Editorials

Review Articles 23

Unresolved laboratory issues of the heterozygous state of b-thalassemia: a literature review Shyamali Thilakarathne et al. https://doi.org/10.3324/haematol.2022.282667 Haematologica | 109 - January 2024

Ubiquitin-proteasome pathway-mediated regulation of the Bcl-2 family: effects and therapeutic approaches Galvin Le Qian Tang et al. https://doi.org/10.3324/haematol.2023.283730

Spotlight Review Articles 44

Tagraxofusp for blastic plasmacytoid dendritic cell neoplasm Marlise R. Luskin and Andrew A. Lane https://doi.org/10.3324/haematol.2022.282171

Articles 53

115

129

143

Acute Lymphoblastic Leukemia Genomic analysis of venous thrombosis in children with acute lymphoblastic leukemia from diverse ancestries Yan Zheng et al. https://doi.org/10.3324/haematol.2022.281582 Acute Myeloid Leukemia Effect of age and treatment on predictive value of measurable residual disease: implications for clinical management of adult patients with acute myeloid leukemia Francesco Mannelli et al. https://doi.org/10.3324/haematol.2023.283196 Acute Myeloid Leukemia Fludarabine, cytarabine, and idarubicin with or without venetoclax in patients with relapsed/refractory acute myeloid leukemia Rabia Shahswar et al. https://doi.org/10.3324/haematol.2023.282912 Acute Myeloid Leukemia Twist family BHLH transcription factor 1 is required for the maintenance of leukemia stem cell in MLL-AF9+ acute myeloid leukemia Nan Wang et al. https://doi.org/10.3324/haematol.2023.282748 Acute Myeloid Leukemia SUMOylation inhibitor TAK-981 (subasumstat) synergizes with 5-azacytidine in preclinical models of acute myeloid leukemia Ludovic Gabellier et al. https://doi.org/10.3324/haematol.2023.282704 Acute Myeloid Leukemia Outcomes after intensive chemotherapy for secondary and myeloid-related changes acute myeloid leukemia patients aged 60 to 75 years old: a retrospective analysis from the PETHEMA registry David Martínez-Cuadrón et al. https://doi.org/10.3324/haematol.2022.282506 Cell Therapy & Immunotherapy Treg-targeted IL-2/anti-IL-2 complex controls graft-versus-host disease and supports anti-tumor effect in allogeneic hematopoietic stem cell transplantation Allan Thiolat et al. https://doi.org/10.3324/haematol.2022.282653 Cell Therapy & Immunotherapy Patient-reported treatment response in chronic graft-versus-host disease Annie Im et al. https://doi.org/10.3324/haematol.2023.282734

Haematologica | 109 - January 2024

151

163

175

186

200

209

220

231

245

256

272

Chronic Lymphocytic Leukemia Electron transport chain and mTOR inhibition synergistically decrease CD40 signaling and counteract venetoclax resistance in chronic lymphocytic leukemia Zhenghao Chen et al. https://doi.org/10.3324/haematol.2023.282760 Chronic Lymphocytic Leukemia Pro-inflammatory cells sustain leukemic clonal expansion in T-cell large granular lymphocyte leukemia Cristina Vicenzetto et al. https://doi.org/10.3324/haematol.2022.282306 Hematopoiesis Adenosine signaling inhibits erythropoiesis and promotes myeloid differentiation Mahmoud Mikdar et al. https://doi.org/10.3324/haematol.2022.281823 Non-Hodgkin Lymphoma Hypomethylating agent decitabine sensitizes diffuse large B-cell lymphoma to venetoclax Fen Zhu et al. https://doi.org/10.3324/haematol.2023.283245 Non-Hodgkin Lymphoma Diffuse large B-cell lymphoma involving osseous sites: utility of response assessment by PET/CT and good long-term outcomes Paola Ghione et al. https://doi.org/10.3324/haematol.2022.282643 Non-Hodgkin Lymphoma Safety and efficacy of tenalisib in combination with romidepsin in patients with relapsed/refractory T-cell lymphoma: results from a phase I/II open-label multicenter study Swaminathan P. Iyer et al. https://doi.org/10.3324/haematol.2022.281875 Plasma Cell Disorders Daratumumab in first-line treatment of patients with light chain amyloidosis and Mayo stage IIIb improves treatment response and overall survival Sara Oubari, et al. https://doi.org/10.3324/haematol.2023.283325 Plasma Cell Disorders DIS3 depletion in multiple myeloma causes extensive perturbation in cell cycle progression and centrosome amplification Vanessa K. Favasuli et al. https://doi.org/10.3324/haematol.2023.283274 Plasma Cell Disorders Elotuzumab plus pomalidomide and dexamethasone in relapsed/refractory multiple myeloma: a multicenter, retrospective, real-world experience with 200 cases outside of controlled clinical trials Massimo Gentile et al. https://doi.org/10.3324/haematol.2023.283251 Plasma Cell Disorders S-adenosylmethionine biosynthesis is a targetable metabolic vulnerability in multiple myeloma Yanmeng Wang et al. https://doi.org/10.3324/haematol.2023.282866 Plasma Cell Disorders Impact of viral hepatitis therapy in multiple myeloma and other monoclonal gammopathies linked to hepatitis B or C viruses Alba Rodríguez-García et al. https://doi.org/10.3324/haematol.2023.283096

Haematologica | 109 - January 2024

III

Letters 283

SARS-CoV-2 vaccination in 361 non-transplanted patients with aplastic anemia and/or paroxysmal nocturnal hemoglobinuria Morag Griffin et al. https://doi.org/10.3324/haematol.2023.283863

287

Molecular predictors of response and survival following IDH1/2 inhibitor monotherapy in acute myeloid leukemia Naseema Gangat et al. https://doi.org/10.3324/haematol.2023.283732

293

Myeloid lineage switch in KMT2A-rearranged acute lymphoblastic leukemia treated with lymphoid lineage-directed therapies Alex Bataller et al. https://doi.org/10.3324/haematol.2023.283705

298

Red cell distribution width as a new prognostic biomarker in refractory chronic graft-versus-host disease Eduard Schulz et al. https://doi.org/10.3324/haematol.2023.283646

303

Impaired T-cell response to mRNA vaccination heralds risk of COVID-19 in long-term allogeneic hematopoietic stem cell transplantation survivors Sigrun Einarsdottir et al. https://doi.org/10.3324/haematol.2023.283551

308

Colchicine reduces inflammation in a humanized transgenic murine model of sickle cell disease Raghda T. Fouda et al. https://doi.org/10.3324/haematol.2023.283377

312

318

In-depth cytogenetic and immunohistochemical analysis in a real world cohort - reconsidering the role of primary and secondary aberrations in multiple myeloma Sina Wenger et al. https://doi.org/10.3324/haematol.2023.283142

325

Inflammation is predictive of outcome in Waldenström macroglobulinemia treated by Bruton tyrosine kinase inhibitors: a multicentric real-life study Pierre-Edouard Debureaux et al. https://doi.org/10.3324/haematol.2023.283141

331

Actin-bundling protein L-plastin promotes megakaryocyte rigidity and dampens proplatelet formation Li Guo et al. https://doi.org/10.3324/haematol.2023.283016

337

An accessible patient-derived xenograft model of low-risk myelodysplastic syndromes Patric Teodorescu et al. https://doi.org/10.3324/haematol.2023.282967

343

351

Germline loss-of-function BRCA1 and BRCA2 mutations and risk of de novo hematopoietic malignancies Ryan J. Stubbins et al. https://doi.org/10.3324/haematol.2022.281580

Haematologica | 109 - January 2024

Case Reports & Case Series 357

Immune-related adverse events with bispecific T-cell engager therapy targeting B-cell maturation antigen Bénédicte Piron et al. https://doi.org/10.3324/haematol.2023.282919

362

CD49d in chronic lymphocytic leukemia: a molecule with multiple regulation layers. Comment to “Sialylation regulates migration in chronic lymphocytic leukemia” Federico Pozzo et al. https://doi.org/10.3324/haematol.2023.283237

364

Erratum to: Comparing the two leading erythroid lines BEL-A and HUDEP-2 Deborah E. Daniels et al. https://doi.org/10.3324/haematol.2023.284510

365

Erratum to: DUSP22-rearranged ALK-negative anaplastic large cell lymphoma is a pathogenetically distinct disease but can have variable clinical outcome Kerry J. Savage and Graham W. Slack https://doi.org/10.3324/haematol.2023.284289

Comments

Errata

Haematologica | 109 - January 2024

LANDMARK PAPER IN HEMATOLOGY

B. Falini

The first human acute myeloid leukemia genome ever fully sequenced Brunangelo Falini Institute of Hematology and CREO, University of Perugia, Perugia, Italy. E-mail: brunangelo.falini@unipg.it doi:10.3324/haematol.2022.282118 ©2024 Ferrata Storti Foundation Haematologica material is published under a CC BY-NC license

TITLE

DNA sequencing of a cytogenetically normal acute myeloid leukaemia genome.

AUTHORS Ley TJ, Mardis ER, Ding L, et al. JOURNAL Nature. 2008;456(7218):66-72. DOI: 10.1038/nature07485. PMID: 18987736

>30-fold coverage, the authors finally ended up with the discovery of somatic mutations affecting ten genes. Recurrent mutations of NPM1 and FLT3 had been previously described,2 whereas the other mutations were new and involved members of the protocadherin/cadherin family (CDH24 and PCLKC), G-protein-coupled receptors (GPR123 and EBI2), a protein phosphatase (PTPRT), a potential guanine nucleotide exchange factor (KNDC1), a peptide/drug transporter (SLC15A1) and a glutamate receptor (GRINL1B). With the exception of FLT3, the mutations were detectable in virtually all leukemic cells both at diagnosis and relapse, suggesting that “the patient had a single dominant clone containing all of the mutations”.

In 2008 a paper appeared in Nature that was to change the way we approach the study of the cancer genome. Ley et al.1 succeeded in sequencing the whole genome of a cytogenetically normal acute myeloid leukemia (AML-M1) from a woman in her mid-50s who presented with a high peripheral white blood cell count (105x109/L, 85% myeloblasts), asthenia and bleeding. The patient achieved a complete remission but relapsed 11 months later acquiring a new clonal cytogenetic abnormality, t(10;12) (p12; p13). Whole DNA sequencing was performed on bone marrow cells at diagnosis and relapse, using the patient’s normal skin cells as a control to exclude germline mutations. Using several filtering tools and sequencing to a depth of

Figure 1. Summmary of Roche/454 FLX readcount data obtained for ten somatic mutations and two validated single nucleotide polymorphisms in the primary tumor, relapse tumor and skin specimens. The readcount data for the variant alleles in the primary tumor sample and relapse tumor sample are statistically different from those of the skin samples for all mutations (P<0.000001 for all mutations, Fisher exact test, denoted by an asterisk in all cases). Note that the skin sample was contaminated by leukemic cells containing the somatic mutations. The patient’s white blood cell count was 105x109/L (85% blasts) when the skin punch biopsy was obtained. (Figure 3 from Ley TJ et al. Nature. 2008;456(7218):66-72. Haematologica | 109 January 2024

LANDMARK PAPER IN HEMATOLOGY

B. Falini tification of the mutational landscape of AML, which consists of more than 20 recurrent mutations,4 including NPM1, FLT3, DNMT3A, IDH1 and IDH2. This led to elucidation of the role of DNMT3A mutations (discovered by whole-genome sequencing) in sustaining clonal hematopoiesis and promoting the development of AML in cooperation with other mutations, e.g. NPM1. Moreover, several targeted therapies were developed against IDH1 and IDH2 mutants (also discovered by whole-genome sequencing) in AML. Finally, clinical trials of AML patients stratified according to mutational profiles allowed groups with different prognoses to be defined. Currently, nextgeneration sequencing of a selected panel of key genes (carrying AML driver mutations) is being increasingly used to define the genomic profile of each AML patient before treatment and also to assess the molecular response to therapy. Whole-genome sequencing can be applied in cases in which cytogenetic analysis is unsuccessful since, in addition to mutations, it can detect copy number alterations and chromosomal rearrangements.

When the genome from the same patient was re-sequenced with a greater coverage depth (a technique not available in 2007-2008) an inactivating DNMT3A L723fs mutation causing haploinsufficiency was discovered.3 In retrospect, the eight non-synonymous mutations (other than NPM1 and FLT3), none of which was recurrent, most likely represented “pre-existing” pathogenically irrelevant mutations of the hematopoietic stem cell that were “captured” by the DNMT3Amediated clonal expansion leading, in cooperation with NPM1 and FLT3 mutations, to the development of AML. This landmark study describing the first human AML (and in general the first human cancer) genome ever fully sequenced clearly demonstrated the value of whole-genome sequencing as an unbiased method for unraveling cancer-initiating mutations in previously unidentified genes. Moreover, it highlighted the limits of hypothesisdriven (for example, candidate gene-based) investigation of tumor genomes by polymerase chain reaction-directed or capture-based methods that can miss key mutations. The impact of the study by Ley et al.1 in accelerating the analysis of the genomes of many hematologic and solid malignancies has been dramatic. Thousands of AML genomes have now been fully sequenced, enabling the iden-

Disclosures No conflicts of interest to disclose.

References 1. Ley TJ, Mardis ER, Ding L, et al. DNA sequencing of a cytogenetically normal acute myeloid leukaemia genome. Nature. 2008;456(7218):66-72. 2. Falini B, Mecucci C, Tiacci E, et al. Cytoplasmic nucleophosmin in acute myelogenous leukemia with a normal karyotype. N Engl J Med. 2005;352(3):254-266.

3. Ley TJ, Ding L, Walter MJ, et al. DNMT3A mutations in acute myeloid leukemia. N Engl J Med. 2010;363(25):2424-2433. 4. Cancer Genome Atlas Research Network; Ley TJ, Miller C, Ding L, et al. Genomic and epigenomic landscapes of adult de novo acute myeloid leukemia. N Engl J Med. 2013;368(22):2059-2074.

Haematologica | 109 January 2024

EDITORIAL

E. Viry and J. Paggetti

Power out chronic lymphocytic leukemia: unplugging OXPHOS/mTOR pathways to overcome venetoclax resistance Elodie Viry and Jerome Paggetti Tumor Stroma Interactions, Department of Cancer Research, Luxembourg Institute of Health, Luxembourg, Luxembourg

Correspondence: J. Paggetti jerome.paggetti@lih.lu Received: Accepted: Early view:

July 19, 2023. July 28, 2023. August 3, 2023.

eral blood of patients were activated with CD40L in order to mimic lymph node signaling (Figure 1A). Similar to resistant CLL cells arising from long-term selection with venetoclax during therapy or in the laboratory, in vitro CD40-activated CLL cells exhibit comparable metabolic characteristics. Here, the strategy was to use inhibitors of different metabolic pathways during CD40 stimulation, and subsequently expose the cells to venetoclax. Thereby, the authors identified OXPHOS as the main driver of CD40-mediated resistance to venetoclax, while neither the inhibition of glycolysis nor that of glutaminolysis counteracted venetoclax resistance. To go further into OXPHOS dismantling, the activity of complexes I, II, III and V of the electron transport chain was independently targeted. The authors reported that only the inhibition of electron transport chain complexes involved in proton pumping and ATP production (complexes I, III and V) led to increased susceptibility to venetoclax (Figure 1B). Inhibitors of electron transport chain complexes I, III and V acted by reducing the two critical players in venetoclax resistance, BCL-XL and MCL-1 proteins, and by downmodulating CD40 signaling through a negative feedback loop. Chen et al. also highlighted that induction of OXPHOS and subsequent venetoclax resistance, following CD40-stimulation of CLL cells, is mediated by mammalian target of rapamycin (mTOR) signaling. As for OXPHOS inhibition, specific targeting of mTOR1/2 decreased the basal oxygen consumption and extracellular acidification rates, and limited the spare respiration capacity, therefore sensitizing CLL cells to venetoclax. Furthermore, the authors emphasize the crucial role of aberrant de novo protein translation in CD40-mediated venetoclax resistance in CLL cells. These results strengthen recent findings demonstrating that B-cell receptor- or Tolllike receptor-engagement leads to the activation of translation initiation in CLL cells. It appears that, rather than a general increase in mRNA translation, enhanced translation

In this issue of Haematologica, Chen et al. report their findings on the contribution of metabolic pathways to CD40induced venetoclax resistance in chronic lymphocytic leukemia (CLL).1 The BCL-2 inhibitor venetoclax is a widely used first-line treatment in fit patients with CLL. Venetoclax-based treatment has shown efficacy and a favorable safety profile, but emergence of resistance remains an important challenge to overcome. Development of chemoresistance is profoundly shaped by multifaceted interactions occurring within the tumor microenvironment. Indeed, venetoclax efficiently eliminates quiescent circulating CLL cells in the peripheral blood, while it produces a less satisfactory response in proliferation centers, in which the CLL cells are exposed to a myriad of signals.2 Notably, within the lymph nodes, interactions between CLL and CD4+ T helper cells via CD40-CD40L signaling contribute to enhanced expression of anti-apoptotic proteins (i.e., BCL-XL, MCL-1 and BFL-1) in CLL cells, leading to subsequent resistance to anticancer therapies.3 When entering proliferation centers, B-cell receptor engagement induces metabolic reprograming mainly relying on aberrant mitochondrial oxidative phosphorylation (OXPHOS) and increased glycolytic capacity/reserve in CLL cells, thereby promoting CLL cell survival and proliferation.46 Chen et al. have already provided additional insights into the metabolism of CLL cells in lymph nodes and resistance to therapy.7 They demonstrated that B-cell receptor/CD40engagement boosts energy metabolism in CLL cells, an effect involving enhanced glycolysis, OXPHOS and the tricarboxylic acid cycle. They demonstrated, for the first time, the dependency of B-cell receptor/CD40-induced CLL cells on glutamine, this being the main substrate fueling the tricarboxylic acid cycle. To further elucidate the mechanisms behind CD40-induced resistance to venetoclax in CLL, in the study published in this issue of Haematologica, Chen and colleagues used an in vitro model in which CLL cells isolated from the periph-

Haematologica | 109 January 2024

EDITORIAL

E. Viry and J. Paggetti

Figure 1. CD40-mediated venetoclax resistance in chronic lymphocytic leukemia cells involves mTOR and OXPHOS pathways.

(A) Within the lymph node, stimulation of chronic lymphocytic leukemia (CLL) cells by T helper cells, via the CD40-CD40L interaction, leads to increased oxidative phosphorylation (OXPHOS), which is a crucial driver of resistance to venetoclax. During this process, the mammalian target of rapamycin (mTOR) pathway connects OXPHOS with CD40 signaling. Both mTOR and OXPHOS pathways exhibit similar effects by promoting protein translation. Together, these two pathways likely affect the sensitivity of CLL cells to venetoclax by upregulating the expression of the anti-apoptotic proteins BCL-XL and MCL-1, while also over-activating CD40 signaling. (B) mTOR inhibition by AZD8055 decreases OXPHOS in CLL cells. Both mTOR inhibition and direct OXPHOS inhibition by rotenone, antimycin A, or oligomycin lowers all regulators linked with venetoclax sensitivity, including decreased protein translation, downregulation of the anti-apoptotic proteins BCL-XL and MCL-1, and downmodulation of CD40 signaling. Together, inhibition of both OXPHOS and mTOR synergistically counteracts venetoclax resistance in CD40-activated CLL cells.

kidney cancer, with rituximab, a chimeric anti-CD20 monoclonal antibody, administered after high-dose consolidative therapy to prevent relapse in lymphoma patients. On the other hand, the OXPHOS inhibitor IACS-010759 has recently been investigated in phase I clinical trials involving patients with relapsed or refractory acute myeloid leukemia, as well as those with advanced, metastatic, or unresectable solid tumors. However, no clinical studies have been conducted on the combination of OXPHOS and mTOR inhibitors, particularly in the context of refractory malignancies. Chen and colleagues aim to address this gap by targeting both mTOR and OXPHOS, with the goal of overcoming resistance mechanisms observed in CLL, emphasizing the significance of addressing translation dysregulation and developing innovative therapeutic strategies.

of specific pathways, particularly MYC and MCL-1, is likely to be essential for CLL cell survival and proliferation.8,9 Interestingly, Chen et al. reported synergistic effects on translation when combining OXPHOS and mTOR inhibitors, converging to an almost complete rescue of CD40-mediated venetoclax resistance. In their study, Chen and colleagues shed light on the critical role of the tumor microenvironment in conferring resistance to therapies in CLL. They established that CD40-mediated venetoclax resistance in CLL is mediated by mTOR signaling and aberrant OXPHOS metabolism, both mechanisms sharing an elevated translation rate as a common feature. The authors propose that new therapeutic opportunities for CLL patients who are resistant to venetoclax could come from combining OXPHOS and mTOR inhibitors. While mTOR and OXPHOS inhibitors have been approved for clinical use against different types of pathologies,10 some of them are now being tested in clinical trials for the treatment of resistant malignancies. One notable example is the combination of everolimus, a Food and Drug Administration-approved mTOR inhibitor used in advanced

Disclosures No conflicts of interests to disclose. Contributions Both authors contributed equally.

Haematologica | 109 January 2024

EDITORIAL

E. Viry and J. Paggetti

References 1. Chen Z, Cretenet G, Carnazzo V, et al. Electron transport chain and mTOR inhibition synergistically decrease CD40 signaling and counteract venetoclax resistance in chronic lymphocytic leukemia. Haematologica. 2024;109(1):151-162. 2. Seymour JF, Kipps TJ, Eichhorst B, et al. Venetoclax-rituximab in relapsed or refractory chronic lymphocytic leukemia. N Engl J Med. 2018;378(12):1107-1120. 3. Haselager MV, Kielbassa K, Ter Burg J, et al. Changes in Bcl-2 members after ibrutinib or venetoclax uncover functional hierarchy in determining resistance to venetoclax in CLL. Blood. 2020;136(25):2918-2926. 4. Jitschin R, Hofmann AD, Bruns H, et al. Mitochondrial metabolism contributes to oxidative stress and reveals therapeutic targets in chronic lymphocytic leukemia. Blood. 2014;123(17):2663-2672. 5. Vangapandu HV, Havranek O, Ayres ML, et al. B-cell receptor signaling regulates metabolism in chronic lymphocytic leukemia. Mol Cancer Res. 2017;15(12):1692-1703.

6. Jitschin R, Braun M, Qorraj M, et al. Stromal cell-mediated glycolytic switch in CLL cells involves Notch-c-Myc signaling. Blood. 2015;125(22):3432-3436. 7. Chen Z, Simon-Molas H, Cretenet G, et al. Characterization of metabolic alterations of chronic lymphocytic leukemia in the lymph node microenvironment. Blood. 2022;140(6):630-643. 8. Largeot A, Klapp V, Viry E, et al. Inhibition of MYC translation through targeting of the newly identified PHB-eIF4F complex as a therapeutic strategy in CLL. Blood. 2023;141(26):3166-3183. 9. Wilmore S, Rogers-Broadway KR, Taylor J, et al. Targeted inhibition of eIF4A suppresses B-cell receptor-induced translation and expression of MYC and MCL1 in chronic lymphocytic leukemia cells. Cell Mol Life Sci. 2021;78(17-18):6337-6349. 10. Ali ES, Mitra K, Akter S, et al. Recent advances and limitations of mTOR inhibitors in the treatment of cancer. Cancer Cell Int. 2022;22(1):284.

Haematologica | 109 January 2024

EDITORIAL

Y. Ofran

Is it time for age and clinically adjusted minimal residual disease interpretation in acute myeloid leukemia? Yishai Ofran Department of Hematology, Shaare Zedek Medical Center and Faculty of Medicine, Hebrew University of Jerusalem, Jerusalem, Israel

Correspondence: Y. Ofran yofran@szmc.org.il Received: Accepted: Early view:

July 20, 2023. July 26, 2023. August 3, 2023.

very difficult to execute. However, despite the limitations of the study by Mannelli et al., important principles can be learnt. First and foremost, once again the prognostic power of a MRD2neg status and its role as a desired milestone along AML therapy are being confirmed. Unfortunately, uniformly across all MRD studies in AML patients, rates of false negative (patients who eventually relapse despite being MRD2neg) and false positive (patients who remain in remission despite being MRDpos) MRD2 results are substantial. Relapse rates despite MRD2 negativity vary between 10-25% in younger, standard-risk AML patients2,3 to 4560% in older adults with adverse-risk leukemia.4,5 Therefore, current guidelines strongly encourage the use of MRD in standard-risk AML and are less determined in poorer risk cases.6 The current work by Mannelli and colleagues suggests that adjusting MRD results to patients’ clinical context is feasible and makes interpretation more accurate. Since huge studies that could yield statistically powerful MRD survival data for each subgroup segregated by a combination of age, ELN risk, induction/consolidation regimen, transplantation and MRD laboratory test used are not going to happen, a clinically relevant working interpretation paradigm is required. As a general medical rule, the prevalence of faulty results (relapse despite MRDneg or continuous remission despite MRDpos status) is influenced by a test’s properties but also by pre-test probability of relapse.7 As illustrated in Figure 1, in high-risk leukemia, achievement of a MRD2neg status reduces the risk of relapse, but this risk remains high enough to justify the morbidity and mortality associated with allogeneic stem cell transplantation. In such cases MRD tests are more likely to identify those who are at extreme high risk of relapse (MRDpos high-risk leukemia). In contrast, for patients with standard-risk leukemia, being MRD2neg is reassuring and may justify avoiding allogeneic stem cell transplantation. Mannelli et al. demonstrated that age is probably the strongest discriminator of adverse risk. Dose intensity of induction

The introduction of sensitive molecular methods enabling the detection of minimal residual disease (MRD) revolutionized decision trees in acute myeloid leukemia (AML). As for any other laboratory tests, even sensitive tests assessing MRD at the level of 10-4-10-5 have false negatives and false positives. In the current issue of Haematologica, Mannelli and colleagues retrospectively explored the influence of patients’ clinical context, (age, genetic profile, and intensity of pre-MRD testing therapy) on the relapsefree and overall survival predictive value of MRD results.1 Different methods of MRD evaluation were used: real-time polymerase chain reaction (RT-PCR) MRD testing in patients presenting with core binding factor or NPM1-mutated leukemia, and immunophenotype for all other patients. Only patients who achieved morphological complete remission after one induction cycle and completed a second intensive chemotherapy consolidation cycle, with available MRD results after the first and second chemotherapy cycles were included. In accordance with previously published results, the rate of MRD eradication in the whole group of 194 AML patients was 56.7% and 62.4% following the first and second treatment cycles (MRD2), respectively. A negative result at that point (MRD2neg) was most predictive of relapse-free and overall suvival. Notably, in different subgroups, the achievement of a MRD2neg status was associated with different outcomes. For patients younger than 55 who were MRD2neg with standard-dose cytarabine, the 3-year progressionfree survival was 86.4% while it was only 46.6% for older adults treated with high-dose cytarabine. The rate of 2017 European LeukemiaNet adverse-risk patients were 16.9% and 2% (P=0.014) among those treated with high- or standard-dose cytarabine, respectively. The utilization of allogenic stem cell transplantation was, as expected, higher among high-risk patients and the more sensitive method of MRD evaluation of RT-PCR was only used in standardrisk patients. A study that is powered adequately to evaluate the prognostic value of MRD2 negativity in all potential clinical scenarios would have to be huge and would be

Haematologica | 109 January 2024

EDITORIAL

Y. Ofran

Figure 1. The clinical impact of minimal residual disease in different clinical scenarios. AML: acute myeloid leukemia; MRD: minimal residual disease.

factor to be considered after age. In conclusion, assessing MRD is fundamental during AML therapy, but results should be interpreted completely differently in high- and standard-risk situations.

(by cytarabine dose) also emerged as an important factor in their study, although in another study, when intensive and non-intensive regimens were compared, the pre-MRD testing therapy effect was not statistically significant.8 Due to the retrospective nature of the work and the fact that groups divided by cytarabine doses were imbalanced I would vote for European LeukemiaNet risk as the second

Disclosures No conflicts of interests to disclose.

References 1. Mannelli F, Piccini M, Bencini S, et al. Effect of age and treatment on predictive value of measurable residual disease: implications for clinical management of adult patients with acute myeloid leukemia. Haematologica. 2024;109(1):60-71. 2. Rücker FG, Agrawal M, Corbacioglu A, et al. Measurable residual disease monitoring in acute myeloid leukemia with t (8; 21)(q22; q22. 1): results from the AML Study Group. Blood. 2019;134(19):1608-1618. 3. Hubmann M, Köhnke T, Hoster E, et al. Molecular response assessment by quantitative real-time polymerase chain reaction after induction therapy in NPM1-mutated patients identifies those at high risk of relapse. Haematologica. 2014;99(8):1317-1325. 4. Murdock HM, Kim HT, Denlinger N, et al. Impact of diagnostic genetics on remission MRD and transplantation outcomes in older patients with AML. Blood. 2022;139(24):3546-3557.

5. Dillon LW, Gui G, Page KM, et al. DNA sequencing to detect residual disease in adults with acute myeloid leukemia prior to hematopoietic cell transplant. JAMA. 2023;329(9):745-755. 6. Heuser M, Freeman SD, Ossenkoppele GJ, et al. 2021 Update on MRD in acute myeloid leukemia: a consensus document from the European LeukemiaNet MRD Working Party. Blood. 2021;138(26):2753-2767. 7. Brüggemann M, Ritgen M, Ladetto M, et al. Next-generation sequencing and real-time quantitative PCR for quantification of low-level minimal residual disease in acute lymphoblastic leukemia of adults. Blood. 2013;122(21):351. 8. Bazinet A, Kadia T, Short NJ, et al. Undetectable measurable residual disease is associated with improved outcomes in AML irrespective of treatment intensity. Blood Adv. 2023;7(13):3284-3296.

Haematologica | 109 January 2024

EDITORIAL

D. Coriu and M-C. Stancioaica

Moving low-risk myelodysplastic syndromes from humans to mice: is it truly that simple? Daniel Coriu and Maria-Camelia Stancioaica Center of Hematology and Bone Marrow Transplantation, Fundeni Clinical Institute, “Carol Davila” University of Medicine and Pharmacy, Bucharest, Romania

Correspondence: D. Coriu daniel_coriu@yahoo.com Received: Accepted: Early view:

July 13, 2023. July 27, 2023. August 3, 2023.

understanding of disease evolution, and perhaps the development of new, targeted drugs. A major breakthrough came through the research of Song et al., who found that low-risk MDS cells could reliably engraft MISTRG mice (highly humanized complex immunodeficient mice).7 This model promised to recapitulate the morphological features, immunophenotype and clonal complexity of MDS via transplantation of CD34+ cells from patients with low-risk MDS. Nevertheless, and somewhat surprisingly, this model has not led to the predicted increased in scientific output regarding novel drug testing in low-risk MDS. Although engraftment in MISTRG mice was shown to be superior to that in NSG mice, perhaps the requirement for intrahepatic injections in newborn animals or intrafemoral injections in adults or maybe the lack of unrestricted access to MITRG mice has limited the use of this exciting PDX. In this issue of Haematologica, Teodorescu et al. describe a new xenograft model of low-risk MDS that makes use of readily available reagents (NSG-S mice and clodronate) and involves simple laboratory techniques (intravenous injection)8 (Figure 1). The authors hypothesize that the barrier to xenotransplantation in these heavily immunocompromised animals is represented by residual mouse macrophages that cannot recognize the CD47 present on malignant human cells. Moreover, they show that macrophages are not eliminated by conditioning with radiation and are, therefore, still present during transplantation. So perhaps the use of intrafemoral or intrahepatic routes is less suitable for overcoming this xeno-barrier. Nevertheless, by giving one injection of liposomal clodronate, the authors could completely, albeit transiently, eliminate spleen macrophages.9 The approach is disarmingly simple and amenable to upscaling for complex, clinical trial-like investigations of multiple concomitant treatments. The authors provide proof-of-concept for therapeutic interventions by using azacitidine treatment of a small number of mice; but upscaling this model may need further optimization. Nevertheless, this PDX recapitulates the clonal

Myelodysplastic syndromes (MDS) are a heterogeneous group of clonal bone marrow diseases featuring ineffective hematopoiesis, morphological dysplasia and persistent peripheral cytopenia that can affect one or more cell lines. The spectrum of genetic abnormalities is in ceaseless expansion and includes chromosomal abnormalities, such as del(5q) and del(7q) and recurrent mutations affecting transcription factors (RUNX1, ETV6), splicing factors (SF3B1, SRSF2), and epigenetic modifiers (TET2, ASXL1, DNMT3A).1,2 The therapeutic strategies for and survival of patients with MDS are decided according to the most common predictor, the Revised International Prognostic Scoring System (IPSS-R), defined by number and severity of cytopenia, percentage of blasts in the bone marrow, as well as the type and number of chromosome changes.3 However, with all the data on molecular diagnostics now available, the topography of MDS patients is more precise. This is why new recommendations are in favor of the use of the Molecular-IPSS (IPSS-M), which incorporates data on 31 driver genes and helps to classify people with MDS into six groups with different rates of survival.4 As far as concerns therapy, the approach consists of disease and patient monitoring, supportive care (blood transfusions, antibiotics), luspatercept, lenalidomide and erythropoietin for low-risk MDS, whereas the hypomethylating agents, azacitidine and decitabine, are used for high-risk MDS, which are known to be aggressive diseases with an increased risk of progression to acute myeloid leukemia. The first patient-derived xenograft (PDX) model was described back in 1969 and in 1981 a genetically engineered mouse model was introduced.5 Since then the use of PDX models has improved the predictability of clinical therapeutic response to greater than 80%.6 However, primary MDS cells are difficult to manipulate, which leads to a paucity of cell lines and PDX models. This lack of PDX models hinders the development of novel therapies, particularly in the case of low-risk MDS. Getting to know the biology of low-risk MDS better is the key to a greater

Haematologica | 109 January 2024

EDITORIAL

D. Coriu and M-C. Stancioaica

Figure 1. A simplified diagram of an improved patient-derived xenograft model using CD34 cells from patients with low-risk myelodysplastic syndrome. MDS: myelodysplastic syndromes; BM: bone marrow; hCD45+: human CD45+; qPCR: quantitative polymerase chain reaction.

cytes are rapidly eliminated by murine macrophages and the clodronate-mediated depletion of murine macrophages was performed only once in the current study, during conditioning. It is tempting to hypothesize that subsequent injections of clodronate could improve the persistence of human erythropoiesis in these PDX mice. Clarifying the extent to which MDS dyserythopoiesis is recapitulated is of the utmost importance prior to testing therapeutic interventions aimed at improving anemia in patients with low-risk MDS, particularly MDS with ring sideroblasts. Lastly, while the study by Teodorescu et al. models lowrisk MDS, it would be useful to test whether premalignant

molecular architecture of the original disease and even maintains some of the splicing events characteristic of the respective mutations in spliceosome genes. Although the authors specifically used samples from patients with lowrisk MDS with spliceosome mutations, one may hypothesize that patients with MDS with other molecular features could also be used to create PDX models employing a similar approach. As expected with PDX models of MDS, human engraftment was vastly represented by myeloid cells but, interestingly, some erythroid and megakaryocytic lineage cells were found in the bone marrow at 3 months after transplantation. This is particularly noteworthy as human erythro-

Haematologica | 109 January 2024

EDITORIAL

D. Coriu and M-C. Stancioaica

clonal hematopoietic conditions, such as clonal hematopoiesis of indetermined potential or clonal cytopenias of unknown significance, could also be modeled in mice using this approach. PDX models of these conditions would provide the tools to study clonal competition and uncover the mechanisms responsible for clonal dominance resulting in MDS. We hope that the model described in the current issue of Haematologica will catalyze preclinical research in MDS and

hematopoietic aging and result in an outburst of novel therapeutic approaches in these conditions. Time will tell! Disclosures No conflicts of interests to disclose. Contributions Both CMS and DC contributed to the final version of the manuscript. DC supervised the project.

References 1. Liu W, Teodorescu P, Halene S, Ghiaur G. The coming of age of preclinical models of MDS. Front Oncol. 2022;12:815037. 2. Goto T. Patient-derived tumor xenograft models: toward the establishment of precision cancer medicine. J Pers Med. 2020;10(3):64. 3. Bernard E, Tuechler H, Greenberg PL. Molecular International Prognostic Scoring System for myelodysplastic syndromes. NEJM Evid. 2022;1(7). 4. Baer C, Huber S, Hutter S, et al. Risk prediction in MDS: independent validation of the IPSS-M - ready for routine? Leukemia. 2023;37(4):938-941. 5. Abdolahi S, Ghazvinian Z, Muhammadnejad S, et al. Patientderived xenograft (PDX) models, applications and challenges in cancer research. J Transl Med. 2022;20(1):206.

6. Goto T. Patient-derived tumor xenograft models: toward the establishment of precision cancer medicine. J Pers Med. 2020;10(3):64. 7. Song Y, Rongvaux A, Taylor A, et al. A highly efficient and faithful MDS patient-derived xenotransplantation model for pre-clinical studies. Nat Commun. 2019;10(1):366. 8. Teodorescu P, Pasca S, Choi I, et al. An accessible patient-derived xenograft model of low-risk myelodysplastic syndromes. Haematologica. 2024;109(1):337-342. 9. Li Z, Xu X, Feng X, Murphy PM. The macrophage-depleting agent clodronate promotes durable hematopoietic chimerism and donor-specific skin allograft tolerance in mice. Sci Rep. 2016;6:22143.

Haematologica | 109 January 2024

EDITORIAL

H.T. Greinix

Patient-reported treatment response in chronic graftversus-host disease Hildegard T. Greinix

Correspondence: H.T. Greinix hildegard.greinix@medunigraz.at

Division of Haematology, Medical University Graz, Graz, Austria

Received: Accepted: Early view:

July 4, 2023. July 11, 2023. July 20, 2023.

analysis of PRO results, moderate-severe lung involvement at enrollment was associated with no improvement, whereas NIH-defined responses in eye, mouth and lung were associated with patients' report of improvement at six months. In addition, improvement from enrollment to six months in the LSS eye score was significantly associated with patients' report of improvement. Thus, investigating tools that capture symptom burden allow better insight into improvements in patients' perception of cGvHD and their QoL. Importantly, PRO at six months were significantly associated with subsequent FFS that was primarily driven by changes in immunosuppressive therapies, offering a good demonstration that, in the real world, continuation or change of cGvHD-specific treatments is based on clinicians' response assessment combined with patients' information on QoL, changes in symptom burden, and function. This clinical routine is in contrast to the treatment response assessment requested in clinical trials where clinician-assessed and patient-reported signs and symptoms, LSS, and the clinician-assessed and patientreported global rating scales have been used as cGvHDspecific core measures.5,6 The fact that PRO had limited correlation with NIH cGvHD response criteria, and that per NIH response only 46% compared to 71% of patients in self-reports had cGvHD improvement after six months in the present study is worrisome since an NIH-defined nonresponder would not meet the primary efficacy endpoint in treatment studies. The increasing importance of PRO is also reflected in the US Food and Drug Administration approval of ibrutinib for treatment of patients with refractory cGvHD since it is plausible that clinician-documented measures and results of the LSS were correlated with clinical responses.6 Recently, Lee et al. reported a strong correlation between clinical response measures and clinically meaningful changes in LSS in cGvHD patients treated with belumosudil.7 They observed an excellent correlation of improvements of skin, mouth, eye, upper gastrointestinal (GI) tract, and lung, supporting the use of PRO for response assess-

Treatment response in chronic graft-versus-host disease (cGvHD) is assessed using National Institutes of Health (NIH) Consensus criteria in clinical studies and by clinician assessment in routine practice. The paper by Im et al. in this issue of Haematologica provides new insight into 6month patient-reported treatment response in cGvHD that was associated with subsequent failure-free survival (FFS) but had limited correlation with NIH and clinician-assessed response.1 So far, patient-reported outcomes (PRO) have been mainly investigated to examine the association of cGvHD symptoms with quality-of-life (QoL), revealing poor functional status, inability to return to work, disrupted activities of daily living, and increased symptom burden.2 In a natural history study, cGvHD severity was negatively associated with nearly all functional and QoL outcome measures, including the Lee Chronic GVHD Symptom Scale (LSS), Medical Outcomes Study Short Form-36-item questionnaire (SF-36), Physical Component Scale, 2-minute walk, grip strength, range of motion, and Human Activity Profile (HAP).3 Joint/fascia, skin, and lung involvement affected function and QoL most significantly, and showed the greatest correlation with outcome measures. Of note, physicians generally underestimated changes in patients' QoL.4 In the paper by Im et al. on 382 cGvHD patients of two prospective studies, PRO were investigated to assess whether patient-reported response measures may capture clinical benefit differently from standard NIH or clinicianreported response measures.1 Physicians and patients rated overall changes in cGvHD manifestations with an 8point scale,5 and in addition, patients completed the SF36 and the LSS. At six months, 270 (71%) patients reported cGvHD improvement defined as: completely gone, very much better, moderately better, or a little better. PRO had limited correlation with either clinician-reported or NIH cGvHD response criteria.5 Per NIH response, only 46% of patients had improvement whereas clinicians reported 66% improvement in cGvHD at six months. In multivariable

Haematologica | 109 January 2024

EDITORIAL

H.T. Greinix

ment in cGvHD clinical trials and patient care to capture the patients' perspective on cGvHD disease activity. However, there was a poor correlation between clinical and PRO responses in joints, esophagus, lower GI tract, and pulmonary function tests. The discrepancy in the findings in the paper of Im et al. demonstrate quite well that the selection of PRO tools can possibly have an impact on results and that there remains an urgent need to validate PRO measures besides the LSS, as recently suggested in a systematic review of 64 articles reporting on 27 PRO.8 In this review, only HAP, LSS and the NIH Eleven Point Scale had evidence to support strong reliability, responsiveness, and validity within the cGvHD population. In the study by Im et al., 6-month patient-reported response was associated with subsequent FFS but did not significantly impact overall survival (OS).1 Improved FFS is an important outcome for patients with cGvHD since treatment failure resulting in the need of additional lines of immunosuppressive therapies can substantially increase patients’ risk of severe infectious complications and organ toxicity. There has been little agreement between the results on treatment response and outcome of patients with cGvHD published so far. In a prospective observational trial on 575 cGvHD patients, 6-month clinician-reported response and 2014 NIH-calculated response significantly correlated with subsequent FFS.9 In addition, a change in the 2005 NIH 0 to 3 clinician-reported skin score and 0 to 10 patient-reported itching score predicted subsequent

FFS, whereas change in the LSS predicted subsequent OS and non-relapse-mortality. In a multicenter prospective cohort of 283 cGvHD patients, a correlation between overall and organ-specific responses at six months with the corresponding overall and organ-specific changes in patient-reported symptom burden using multiple PRO was observed.10 However, overall response at six months did not correlate with changes in QoL measures or with OS. PRO measures are a valuable tool for monitoring disease progression and treatment response, and are likely to provide benefit in GvHD clinical practice and research. So far, patient-reported QoL measures have served as secondary endpoints but were not considered in primary study endpoints since these only included quantitative measures of cGvHD activity. Therefore, an area for future investigation would be to determine methods to integrate objective and subjective measures into holistic assessments of cGvHD disease activity and how this changes under immunosuppressive/immunomodulatory therapy. This should take place in cGvHD clinical trials, drug development, and routine clinical practice to objectively document, as well as possible, improvements in cGvHD manifestations, as well as symptom relief and patients’ clinical benefit. Disclosures HTG has received consulting/educational honoraria from Mallinckrodt, Gilead, Novartis, Roche, Sanofi, and Takeda.

References Response Criteria Working Group report. Biol Blood Marrow Transplant. 2015;21(6):984-999. 6. Zeiser R, Lee SJ. Three US Food and Drug Administrationapproved therapies for chronic GVHD. Blood. 2022;139(11):1642-1645. 7. Lee SJ, Cutler C, Blazar BR, Tu A, Yang Z, Pavletic SZ. Correlation of patient-reported outcomes with clinical organ responses: data from the belumosudil chronic graft-versus-host disease studies. Transplant Cell Ther. 2022;28(10):700.e1-700.e6. 8. Kilgour JM, Wali G, Gibbons E, et al. Systematic review of patient-reported outcome measures in graft-versus-host disease. Biol Blood Marrow Transplant. 2020;26(5):e113-e127. 9. Palmer J, Chai X, Pidala J, et al. Predictors of survival, nonrelapse mortality, and failure-free survival in patients treated for chronic graft-versus-host disease. Blood. 2016;127(1):160-166. 10. Inamoto Y, Martin PJ, Chai X, et al. Clinical benefit of response in chronic graft-versus-host disease. Biol Blood Marrow Transplant. 2012;18(10):1517-1524.

1. Im A, Pusic I, Onstad L, et al. Patient-reported treatment response in chronic graft- versus-host disease. Haematologica. 2024;109(1):143-150. 2. Lee SJ, Onstad L, Chow EJ, et al. Patient-reported outcomes and health status associated with chronic graft-versus-host disease. Haematologica. 2018;103(9):1535-1541. 3. Baird K, Steinberg SM, Grkovic L, et al. National Institutes of Health chronic graft-versus-host disease staging in severely affected patients: organ and global scoring correlate with established indicators of disease severity and prognosis. Biol Blood Marrow Transplant. 2013;19(4):632-639. 4. Kurosawa S, Oshima K, Yamaguchi T, et al. Quality of life after allogeneic hematopoietic cell transplantation according to affected organ and severity of chronic graft-versus-host disease. Biol Blood Marrow Transplant. 2017;23(10):1749-1758. 5. Lee SJ, Wolff D, Kitko C, et al. Measuring therapeutic response in chronic graft-versus-host disease. National Institutes of Health consensus development project on criteria for clinical trials in chronic graft-versus-host disease: IV. The 2014

Haematologica | 109 January 2024

EDITORIAL

I. Plo and I. Antony-Debré

The double-edged sword of adenosine Isabelle Plo1,2,3 and Iléana Antony-Debré1,2,3 1

INSERM, UMR 1287, Gustave Roussy, Villejuif; 2Gustave Roussy, Villejuif and 3Université Paris-Saclay, Gif-sur-Yvette, France

Correspondence: I. Plo isabelle.plo@gustaveroussy.fr Received: Accepted: Early view:

June 22, 2023. July 11, 2023. July 20, 2023.

In this issue of Haematologica, Midkar et al. show the dual role of adenosine signaling in erythroid and myeloid differentiation.1 Besides their involvement in DNA and RNA synthesis, nucleosides such as adenosine are also signaling molecules that regulate tissue homeostasis through various biological functions: cell growth, migration, neurotransmission, immune regulation, and vascular tone. Adenosine levels are highly up-regulated during hypoxia, ischemia, physical activity, epilepsy, pain, and cancers, and deal with oxygen demand and inflammation. This regulation implicates a tightly controlled network of catabolic enzymes, nucleoside transporters, and receptors. In particular, the extra-/intracellular levels of adenosine are controlled by the equilibrative nucleoside transporters (ENT) and the concentrative nucleoside transporters (CNT). Extracellular adenosine can be generated from dephosphorylation of extracellular ATP, ADP, and AMP through the action of the cell-surface-associated enzymes CD39 (ectonucleoside triphosphate diphosphohydrolylase 1) and CD73 (ecto-5'-nucleotidase). Adenosine can be rapidly phosphorylated to AMP by adenosine kinase or deaminated to inosine by adenosine deaminase. As a signaling molecule, adenosine can bind to four adenosine G proteincoupled receptor (AR) subtypes: A1AR, A2AAR, A2BAR, and A3AR. Transcriptomic databases of human hematopoiesis showed the predominant expression of ADORA3 and ADORA2B, encoding A3AR and A2BAR, in CD34+ progenitors, and erythroid and myeloid cells. Following adenosine binding, activation of the receptors modulates the AMPc/PKA axis and related transcription factors and kinases, and the Phospholipase C and MAPK pathways (ERK, P38, JNK).2 The involvement of adenosine in the regulation of hematopoiesis is well known. However, very few studies, including those on the role of nucleosides in general, have reported its role in erythropoiesis, the process of differentiating hematopoietic stem cells into red blood cells (RBC).3 Defective erythropoiesis was observed both in ENT1-null patients harboring the extremely rare Augustine blood group system and in Ent1 knock-out (KO) mice. Both models displayed increased extracellular nucleosides, including adenosine.4 In the present study, Midkar et al.

have focused on the impact of extracellular adenosine during early and late erythropoiesis. They used an in vitro differentiation system derived from purified CD34+ progenitors with the addition of stem cell factor (SCF), interleukin-3 (IL-3), and erythropoietin. They showed that, in contrast to other nucleosides, a high concentration of adenosine is able to negatively impact erythropoiesis by inhibiting proliferation and differentiation. The combination of ENT1 potent bidirectional inhibitor (NTBI) and adenosine further potentiated the inhibitory effect on proliferation, leading to altered erythropoiesis, and suggesting the involvement of extracellular adenosine signaling. This hypothesis was elegantly confirmed by the use of the A3AR agonist (CI-IB-MECA), which is in line with a previous study in primary human cells.5 Since A2BAR was also expressed in myeloid and erythroid cells, the authors tested the BAY60-6583 agonist. This delayed erythroid maturation but did not impact proliferation. Taken together, this study provides evidence that extracellular adenosine is a signaling molecule that regulates erythropoiesis at different stages. It would be interesting to investigate the expression of these receptors in order to clarify their role. Interestingly, the authors found that the negative effect of adenosine on erythropoiesis was through P53-mediated apoptosis. This was confirmed by a proteomic analysis of erythroblasts which showed numerous over-expressed proteins in both the P53-related apoptotic and the DNA repair pathways. Of note, P53-induced apoptosis upon extracellular adenosine accumulation has already been described in breast and ovarian cancer cell lines. In particular, it was shown that P53 can promote transcription of ADORA2B, which in turn activates caspase- and Pumadependent apoptotic pathways through P53.6 Alternatively, one could hypothesize that the activation of P53 by AR could be triggered through its phosphorylation by P38. P53 is known to be a critical molecule for effective erythropoiesis. In early stages, P53 should be tightly regulated to maintain a high rate of erythroblast proliferation. In later stages, when decreased ribosome biogenesis occurs in basophilic erythroblasts, P53 is required to re-

Haematologica | 109 January 2024

EDITORIAL

I. Plo and I. Antony-Debré

Figure 1. Extracellular adenosine induces decreased erythropoiesis via adenosine receptors A2BAR and A3AR. While A3AR inhibits erythroid proliferation and differentiation, A2BAR inhibits erythroid maturation with no effect on proliferation. The defective erythropoiesis is due to activation of P53- and caspase 3-dependent apoptosis. Moreover, activation of A3AR decreases erythropoiesis and favors myelopoiesis. spond to ribosome stress and ensure terminal maturation. A premature or abnormal activation of P53 during erythropoiesis, such as that observed upon extracellular adenosine treatment or in ribosomopathies, may lead to differentiation blockade and anemia.7 Midkar et al. report that extracellular adenosine and A3AR agonist decreased erythroid differentiation in favor of myeloid differentiation, while the A2BAR agonist had no effect (Figure 1). Similarly, Ent1 KO mice displayed reduced numbers of erythroid cells while showing increased granulocytic cell counts, at both the mature and progenitor stages.4 A3AR activation has already been shown to promote mouse granulo-monocytic progenitor growth by increasing the effects of growth factors such as IL-3, SCF, and granulocyte macrophage colony-stimulating factor (GM-CSF) in vitro,3 and by promoting granulocyte (G)-CSF production in vivo in mice.8 In contrast, in a sickle cell disease mouse model harboring elevated levels of extracellu-

lar adenosine, treatment with polyethylene glycol-modified adenosine deaminase, which lowers adenosine concentrations, leads to a decrease in leukocyte counts, and increased RBC counts.9 Deletion of A3AR in mice leads to several abnormalities in the peripheral blood, such as decreased neutrophil and monocyte counts, increased erythrocyte and hemoglobin levels, and a decrease in platelet counts.10 However, the erythroid compartment still needs to be studied in more detail. While studies so far agree on the role of adenosine/A3AR signaling in erythroid/myeloid commitment and differentiation, other lineages, such as the megakaryocytic lineage, could also be affected, and there is still no clear indication as to how adenosine signaling can influence erythroid versus myeloid commitment. Could it strengthen the cytokine effect? Does it have an impact on transcription factor expression? Does it generate metabolic changes that will dictate differentiation? The work of Mikdar et al. adds adenosine to the plethora

Haematologica | 109 January 2024

EDITORIAL

I. Plo and I. Antony-Debré

of factors that can modulate normal erythropoiesis. Different types of AR agonists are already in use or are being developed in clinical trials in several diseases: auto-immune diseases, inflammation, neurological diseases, and diabetes. Therefore, one could consider their use in diseases with erythroid hyperplasia (hereditary erythrocytosis or myeloproliferative diseases such as polycythemia vera). In contrast, since adenosine plays a broad immunosuppressive role that regulates both innate and adaptive immune responses, inhibitors of extracellular adenosine have also been widely developed to promote antitumor

activities in cancers. It would be interesting to determine the levels of extracellular adenosine in hematologic malignancies or diseases with defective erythropoiesis, and explore how these levels could contribute to disease development using adenosine inhibitors or agonists. Disclosures No conflicts of interest to disclose. Contributions Both authors contributed equally.

References 1. Mikdar M, Serra M, Colin E, et al. Adenosine signaling inhibits erythropoiesis and promotes myeloid differentiation. Haematologica. 2024;109(1):175-185. 2. Antonioli L, Blandizzi C, Csóka B, Pacher P, Haskó G. Adenosine signalling in diabetes mellitus-pathophysiology and therapeutic considerations. Nat Rev Endocrinol. 2015;11(4):228-241. 3. Hofer M, Pospisil M, Weiterova L, Hoferova Z. The role of adenosine receptor agonists in regulation of hematopoiesis. Molecules. 2011;16(1):675-685. 4. Mikdar M, González-Menéndez P, Cai X, et al. The equilibrative nucleoside transporter ENT1 is critical for nucleotide homeostasis and optimal erythropoiesis. Blood. 2021;137(25):3548-3562. 5. Bhanu NV, Aerbajinai W, Gantt NM, et al. Cl-IB-MECA inhibits human erythropoiesis. Br J Haematol. 2007;137(3):233-236.

6. Long JS, Crighton D, O’Prey J, et al. Extracellular adenosine sensing-a metabolic cell death priming mechanism downstream of p53. Mol Cell. 2013;50(3):394-406. 7. Le Goff S, Boussaid I, Floquet C, et al. p53 activation during ribosome biogenesis regulates normal erythroid differentiation. Blood. 2021;137(1):89-102. 8. Bar-Yehuda S, Madi L, Barak D, et al. Agonists to the A3 adenosine receptor induce G-CSF production via NF-κB activation. Exp Hematol. 2002;30(12):1390-1398. 9. Zhang Y, Dai Y, Wen J, et al. Detrimental effects of adenosine signaling in sickle cell disease. Nat Med. 2011;17(1):79-86. 10. Hofer M, Pospíšil M, Dušek L, Hoferová Z, Komůrková D. Lack of adenosine A3 receptors causes defects in mouse peripheral blood parameters. Purinergic Signal. 2014;10(3):509-514.

Haematologica | 109 January 2024

EDITORIAL

C. Récher

Secondary acute myeloid leukemia: time to turn the page Christian Récher Service d’Hématologie, Centre Hospitalier Universitaire de Toulouse, Institut Universitaire du Cancer de Toulouse Oncopole, Université Toulouse III Paul Sabatier, Toulouse, France

Correspondence: C. RECHER Recher.Christian@iuct-oncopole.fr Received: Accepted: Early view:

June 23, 2023. July 13, 2023. July 20, 2023.

This study raises several points for discussion. Firstly, this picture of real-life practice shows that patients were selected for intensive therapy mainly according to criteria such as age and performance status, rather than on chemosensitivity. Secondly, only a minority of patients who were refractory to a first induction cycle received a second cycle. It is likely that physicians estimated that in the light of the benefit/risk ratio, a second induction was of little value, which, in this specific situation, is probably true. In the European LeukemiaNet 2022 guidelines, however, patients are considered to be refractory if they complete two cycles of induction.3 Since eligibility criteria for clinical trials generally follow this recommendation, this may be of concern for those patients who, having failed a first cycle, could then proceed directly to innovative therapy rather than undergo another futile, and potentially toxic, cycle of chemotherapy.4 Third, although in this patient population bridging to allogeneic HSCT (alloHCT) after induction chemotherapy is a major therapeutic goal, only a minority of patients received transplants, thus leaving room for maintenance therapy for those patients who were able to achieve complete remission. Lastly, even though in secondary AML, NPM1 mutations are a rare event, they retain a major prognostic impact; in fact, recent data showed that therapy-related and de novo mutated NPM1 AML have overlapping features and a similar prognosis.5 Therefore, t-AML and sAML with NPM1 mutation should be managed as de novo AML. This study is likely to be one of the last to evaluate standard chemotherapy in this predefined population, as the definition of secondary AML has now been refined and treatment has been changed. Although in certain respects the recent World Health Organization classification differs from the latest International Consensus Classification, both have significantly modified the subentities of secondary AML: AML-MRC and t-AML have been removed,6,7 and the International Consensus Classification has isolated AML with TP53 mutations as a single entity. Previous exposure to cytotoxic therapy (which now includes immune interventions and PARP inhibitors) and prior MDS or MDS/MPN are now diagnostic

In this issue of Haematologica, Martinez-Cuadron et al. have analyzed data from a multinational registry co-ordinated by the Spanish Programa de Estudio y Tratamiento de las Hemopatías Malignas (PETHEMA) and report the results of intensive treatments that were available before the era of new therapeutic options in older patients with secondary acute myeloid leukemia (AML).1 For physicians used to treating AML, the terms “elderly” or “secondary” have always been synonymous with major problems of patient and disease management. The combination of these two factors results in one of the worst prognoses in oncology, underlying the urgent need for innovative research. For the academics who classify hematologic malignancies, the definition of secondary AML has remained an evolving field in view of recent advances in the molecular characterization of AML with the aim to rationally define patient populations. In everyday language, “secondary” AML encompass therapy-related AML (t-AML) that occur after prior exposure to cytotoxic chemotherapy and/or radiotherapy, and AML that occur after a previous myeloid malignancy such as myelodysplastic syndrome (MDS) or MDS/myeloproliferative neoplasms (MPN). PETHEMA is to be congratulated on having set up such a large registry, thus making it possible to produce one of the largest cohorts of patients with secondary AML who were selected by their physicians in routine practice to receive intensive chemotherapy. The richness of the data is particularly appreciated. The investigators focused on a patient population with criteria similar to that of the pivotal phase III trial comparing standard chemotherapy and the liposomal encapsulation of cytarabine and daunorubicin CPX-351, which led to the registration of CPX351 for t-AML and AML with myelodysplasia-related changes (AML-MRC).2 As expected, the results confirm the poor prognosis of intensive chemotherapy in a realworld setting. Again, not surprisingly, age and poor performance status were adverse prognostic factors for response and overall survival, while favorable/intermediate cytogenetic risk and NPM1 mutations were favorable factors. Post-remission therapy also affected outcome, especially hematopoietic stem cell transplantation (HCT).

Haematologica | 109 January 2024

EDITORIAL

C. Récher intermediate-risk AML who were not included in the pivotal trial have now been reclassified. Whether CPX-351 is also of benefit in AML with AML with MR-gene mutations receptor gene mutations compared to standard chemotherapy is an open question. Venetoclax plus azacitidine (VEN-AZA) has also demonstrated significant activity in secondary AML.8 Although this combination has been approved for patients who are ineligible for intensive chemotherapy, many physicians propose VEN-AZA in high-risk AML as a bridge to transplantation in fit patients. Lastly, targeted agents, such as IDH or FLT3 inhibitors, added to standard chemotherapy may also improve outcome in high-risk patients, and maintenance

qualifiers of disease entities. AML-MRC have been replaced by a new entity, myelodysplasia-related AML (AML-MR), which is biologically defined by specific cytogenetic or molecular abnormalities (secondary AML-like mutations) but no longer by multilineage dysplasia. Meanwhile, novel therapeutic strategies have been developed, including CPX-351, which have proven efficacy over standard 3+7 with lower extra-hematologic toxicity and an impressive outcome in patients who were bridged to alloHCT. However, CPX-351 was specifically approved in the formerly defined-group of t-AML and AML-MRC. By introducing secondary AML-like mutations in the new AMLMR definitions, a sizable proportion of de novo

Figure 1. Management algorithm for secondary acute myeloid leukemia. alloHCT: allogeneic hematopoietic stem cell transplan-

tation; ASAP: as soon as possible; Conso: consolidation chemotherapy; AML: acute myeloid leukemia; t-AML: therapy-related AML; MR-AML: myelodysplasia-related AML; MRC-AML: AML with myelodysplasia-related changes; MDS: myelodysplastic syndrome; MPN: myeloproliferative neoplasms; AZA-VEN: azacitidine venetoclax; CR/CRi: complete remission/CR with incomplete hematologic recovery. Haematologica | 109 January 2024

EDITORIAL

C. Récher

therapy with oral AZA prolongs the duration of both response and survival.9 Things are also moving in the field of transplantation, which is crucial for long-term survival. A recent randomized trial convincingly demonstrated that sequential conditioning and immediate transplantation produce similar outcomes to intensive remission induction chemotherapy followed by alloHCT in refractory or relapsed AML.10 This strategy should undoubtedly facilitate bridging to transplantation in a significant proportion of patients. Therefore, as soon as the diagnosis is made, the treatment strategy for patients with secondary AML should integrate an active search for a suitable donor with a transplant schedule. While awaiting transplantation, achieving disease control and eventually reaching complete remission

without compromising general health is highly desirable, and CPX351 or VEN-AZA appear to fulfill this mission better than 3+7 (Figure 1). Preventing post-transplant relapse is also becoming an important issue. To sum up, the therapeutic landscape for front-line treatment in secondary AML has changed dramatically and brings hope for better outcomes. A registry such as this will surely provide the PETHEMA group with the opportunity to show us the impact of these new treatments in real-life clinical practice. Disclosures CR has received research grants from Abbvie, Astellas, BMS, Jazz Pharmaceuticals and Iqvia; he is on the advisory board of Abbvie, Astellas, BMS, Jazz Pharma, Novartis, Servier, and Takeda.

References 1. Martínez-Cuadrón D, Megías-Vericat JE, Gil C, et al. Outcomes after intensive chemotherapy for secondary and myeloidrelated changes acute myeloid leukemia patients aged 60 to 75 years old: a retrospective analysis from the PETHEMA registry. Haematologica. 2024;109(1):115-128. 2. Lancet JE, Uy GL, Newell LF, et al. CPX-351 versus 7+3 cytarabine and daunorubicin chemotherapy in older adults with newly diagnosed high-risk or secondary acute myeloid leukaemia: 5-year results of a randomised, open-label, multicentre, phase 3 trial. Lancet Haematol. 2021;8(7):e481-e491. 3. Döhner H, Wei AH, Appelbaum FR, et al. Diagnosis and management of AML in adults: 2022 recommendations from an international expert panel on behalf of the ELN. Blood. 2022;140(12):1345-1377. 4. Uy GL, Aldoss I, Foster MC, et al. Flotetuzumab as salvage immunotherapy for refractory acute myeloid leukemia. Blood. 2021;137(6):751-762. 5. Othman J, Meggendorfer M, Tiacci E, et al. Overlapping features of therapy-related and de novo NPM1-mutated AML. Blood. 2023;141(15):1846-1857. 6. Khoury JD, Solary E, Abla O, et al. The 5th edition of the World

Health Organization Classification of Haematolymphoid Tumours: myeloid and histiocytic/dendritic neoplasms. Leukemia. 2022;36(7):1703-1719. 7. Arber DA, Orazi A, Hasserjian RP, et al. International Consensus Classification of Myeloid Neoplasms and Acute Leukemias: integrating morphologic, clinical, and genomic data. Blood. 2022;140(11):1200-1228. 8. DiNardo CD, Jonas BA, Pullarkat V, et al. Azacitidine and venetoclax in previously untreated acute myeloid leukemia. N Engl J Med. 2020;383(7):617-629. 9. Wei AH, Döhner H, Sayar H, et al. Long-term survival with oral azacitidine for patients with acute myeloid leukemia in first remission after chemotherapy: updated results from the randomized, placebo-controlled, phase 3 QUAZAR AML-001 trial. Am J Hematol. 2023;98(4):E84-E87. 10. Stelljes M, Middeke JM, Bug G, et al. In Patients with relapsed/refractory AML sequential conditioning and immediate allogeneic stem cell transplantation (allo-HCT) results in similar overall and leukemia-free survival compared to intensive remission induction chemotherapy followed by allo-HCT: results from the randomized phase III ASAP Trial. Blood. 2022;140(Suppl 1):S9-11.

Haematologica | 109 January 2024

EDITORIAL

E. O’Donnell

Exploring the role of viral hepatitis in plasma cell disorders Elizabeth O’Donnell

Correspondence: E. O’Donnell Elizabeth_odonnell@dfci.harvard.edu

Dana-Farber Cancer Institute, Boston, MA, USA

Received: Accepted: Early view:

July 5, 2023. July 14, 2023. July 20, 2023.

Ig was the HBV, suggesting that the HBV infection initiated the clonal gammopathy. They then went on to evaluate the overall survival of patients with MM with HBV and HCV who had been treated with antiviral therapy versus those who had not. They found statistically significant differences between the treated and untreated cohorts for both HBV and HCV. Their findings strengthen previous findings that anti-HCV therapy may improve outcomes in plasma cell dyscrasias, and extends these to include HBV. The recognition of the association between viral hepatitis and MM and monoclonal gammopathies has important clinical implications. The authors highlight that if the target of a patient’s monoclonal protein can be identified and eliminated, chronic antigen exposure disappears, leading to control of the clonal plasma cells. Not only would antiviral therapy improve outcomes and the morbidity associated with viral hepatitis, but it could also potentially mitigate the monoclonal gammopathy and the risk of progressing to symptomatic disease. As such, there is a need for increased awareness and screening for viral hepatitis in patients with MM and monoclonal gammopathies. Early identification of viral hepatitis in these individuals can lead to appropriate antiviral therapy and subsequent improvement in outcomes. The management of patients with MM with concomitant viral hepatitis requires a multidisciplinary approach. Treatment decisions should consider both the hematologic malignancy and the viral hepatitis, taking into account potential drug-drug interactions and the impact of underlying liver disease on treatment tolerability. Close collaboration between hematologists/oncologists and hepatologists is essential to optimize treatment outcomes and minimize complications. Prevention strategies are also crucial in mitigating the impact of viral hepatitis in patients with MM and other monoclonal gammopathies. Vaccination against HBV should be encouraged in individuals at risk, including patients with MM and those receiving immunosuppressive therapy. In addition, raising awareness about safe injection practices, blood screening,

Multiple myeloma (MM) is a malignant plasma cell disorder characterized by the proliferation of abnormal plasma cells in the bone marrow. Monoclonal gammopathy of undetermined significance (MGUS) and smoldering multiple myeloma (SMM) are precursors to MM. While the exact mechanisms are not yet fully understood, several studies have demonstrated an association between chronic viral hepatitis and the development of MM and other monoclonal gammopathies.1-6 Viral hepatitis, caused by hepatitis B (HBV) and hepatitis C (HCV) viruses, has been a major global health concern for several decades. These viruses can lead to chronic infections that increase the risk of various liver diseases, including cirrhosis and hepatocellular carcinoma. However, recent research has shed light on the link between viral hepatitis and the development of certain hematologic malignancies, particularly MM and other monoclonal gammopathies.1-3 In this issue of Haematologica, Rodriguez-Garcia and colleagues explore the role of HBV in patients infected with HBV and diagnosed with monoclonal gammopathy, and evaluate the impact of antiviral therapy on outcomes for patients with HCV and HBV infections and MM.7 Hepatitis C virus infection has been implicated in the pathogenesis of MM. Chronic HCV infection induces chronic inflammation, immune dysregulation, and clonal B-cell expansion, all of which can promote the development of B-cell malignancies.8 Furthermore, HCV infection has been associated with an unfavorable prognosis and reduced survival in patients with MM.9 However, in patients with HCV whose monoclonal protein reacted against HCV, treating the HCV infection improved MGUS and MM disease.10 Similarly, an increased prevalence of HBV infection has been reported in patients with MM compared to the general population.11 In this issue, Rodriguez-Garcia and colleagues evaluated patients with monoclonal gammopathy who had been infected with HBV and found that, in 36.7% of the HBV-infected patients, the target of the monoclonal

Haematologica | 109 January 2024

EDITORIAL

E. O’Donnell

and the importance of antiviral therapy for viral hepatitis can reduce the incidence and transmission of these infections. The association between viral hepatitis and the development of MM and other monoclonal gammopathies has emerged as an important area of research. Chronic HBV and HCV infections contribute to the pathogenesis of these hematologic malignancies, warranting increased awareness, screening, and management strategies. Early identification of viral hepatitis in patients with MM and

monoclonal gammopathies can lead to improved outcomes through appropriate antiviral therapy and tailored treatment approaches. Moving forward, continued research, collaboration between specialists, and comprehensive prevention efforts will be crucial in reducing the burden of viral hepatitis in this patient population, and in improving their overall prognosis and quality of life. Disclosures No conflicts of interest to disclose.

References myeloma. N Engl J Med. 2016;374(6):555-561. 7. Rodriguez-Garcia A, Mennesson N, Hernandez-Ibarburu G, et al. Impact of viral hepatitis therapy in multiple myeloma and other monoclonal gammopathies linked to hepatitis B or C viruses. Haematologica. 2024;109(1):272-282. 8. Zignego AL, Giannini C, Ferri C. Hepatitis C virus-related lymphoproliferative disorders: an overview. World J Gastroenterol. 2007;13(17):2467-2478. 9. Teng CJ, Liu HT, Liu CY, et al. Chronic hepatitis virus infection in patients with multiple myeloma: clinical characteristics and outcomes. Clinics (Sao Paulo). 2011;66(12):2055-2061. 10. Rodriguez-Garcia A, Linares M, Morales ML, et al. Efficacy of antiviral treatment in hepatitis C virus (HCV)-driven monoclonal gammopathies including myeloma. Front Immunol. 2021;12:797209. 11. Becker N, Schnitzler P, Boffetta P, et al. Hepatitis B virus infection and risk of lymphoma: results of a serological analysis within the European case-control study Epilymph. J Cancer Res Clin Oncol. 2012;138(12):1993-2001.

1. Hermouet S, Corre I, Gassin M, Bigot-Corbel E, Sutton CA, Casey JW. Hepatitis C virus, human herpesvirus 8, and the development of plasma-cell leukemia. N Engl J Med. 2003;348(2):178-179. 2. Bigot-Corbel E, Gassin M, Corre I, Le Carrer D, Delaroche O, Hermouet S. Hepatitis C virus (HCV) infection, monoclonal immunoglobulin specific for HCV core protein, and plasma-cell malignancy. Blood. 2008;112(10):4357-4358. 3. McShane CM, Murray LJ, Engels EA, Landgren O, Anderson LA. Common community-acquired infections and subsequent risk of multiple myeloma: a population-based study. Int J Cancer. 2014;134(7):1734-1740. 4. Li Y, Li Y, Zhang L, Li W. Hepatitis C virus infection and risk of multiple myeloma: evidence from a meta-analysis based on 17 case-control studies. J Viral Hepat. 2017;24(12):1151-1159. 5. Bosseboeuf A, Feron D, Tallet A, et al. Monoclonal IgG in MGUS and multiple myeloma targets infectious pathogens. JCI Insight. 2017;2(19):e95367. 6. Nair S, Branagan AR, Liu J, Boddupalli CS, Mistry PK, Dhodapkar MV. Clonal immunoglobulin against lysolipids in the origin of

Haematologica | 109 January 2024

EDITORIAL

P.D. Shah and K.L. Nathanson

BRCA1/2 mutations and de novo hematologic malignancies: true, true and not clearly related Payal D. Shah1,2 and Katherine L. Nathanson2,3 1

Division of Hematology and Oncology, Department of Medicine, 2Basser Center for BRCA, Abramson Cancer Center and 3Division of Translational Medicine and Human Genetics, Department of Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA

Correspondence: K.L. Nathanson knathans@upenn.edu Received: Accepted: Early view:

July 4, 2023. July 10, 2023. July 20, 2023.

neutral, with tumor pathogenesis occurring independently of, rather than driven by, gBRCA1/2. In an analysis evaluating germline blood and matched tumor tissue from over 17,000 cancer patients among whom 472 harbored a gBRCA1/2 PV, selective pressure for biallelic inactivation, zygosity-dependent phenotype penetrance, and poly-ADP ribose polymerase inhibitor (PARPi) sensitivity were only observed in tumor types classically associated with BRCA1/2, i.e., breast, ovary, prostate or pancreas cancers.3 Arguing against BRCA as a major driver of the observed de novo HM is the presence of LoH in only three of 14 evaluated samples in the study cohort. It is known that classically BRCA1/2-associated solid tumors often (though not always) demonstrate LoH, whereas solid tumors occurring with, but not driven by, BRCA do not.6 The level of LoH observed in this study is consistent with chance, being similar to the level observed with benign gBRCA1/2 variants in a larger dataset.3 As noted by the authors, both determining whether this level is higher than observed in HM with benign gBRCA1/2 variants and evaluating the role of epigenetic silencing should be done. In classically BRCA1/2-associated tumors with and without LoH, additional factors often support BRCA as a driver of tumor pathogenesis, such as vertical transmission, early age of onset, and phenotypic tumor characteristics including homologous recombination deficiency or PARPi sensitivity. The manuscript by Stubbins et al. does not report whether the study cohort or the observed de novo HM display these features. The many HM types reported is inconsistent with BRCA as a major driver of pathogenesis. For solid tumors, BRCA1/2 PV are associated with very specific tumor types - for example, high-grade serous ovarian cancer and pancreatic ductal adenocarcinoma are BRCA1/2-associated neoplasms, whereas low-grade, borderline, and germ-cell ovarian and pancreatic neuroendocrine cancers are not. Furthermore, PV in BRCA1 and BRCA2 have non-identical cancer risk profiles. The lumping of ten different HM di-

Stubbins and colleagues sought to evaluate whether individuals with germline BRCA1 or BRCA2 (gBRCA1/2) pathogenic variants (PV) have an independent risk of developing de novo hematologic malignancies (HM) in addition to therapy-related neoplasms.1 In this single-institution retrospective study, the authors identified 25 patients with gBRCA1 (n=14) or gBRCA2 (n=11) PV concurrent with a HM diagnosis. Eight of 14 (BRCA1) and eight of 11 (BRCA2) patients had de novo HM, rather than therapy-related HM. These patients constituted 1.1% of patients with HM seen over 8 years. Leukemic cells from three of 14 (21%) patients with BRCA1/2 PV had loss of heterozygosity (LoH) of the wildtype allele. In addition to therapy-related HM in BRCA1/2 carriers,2 patients with BRCA1/2 PV developed de novo HM of various types. Most literature examining BRCA1/2 and HM focuses on therapy-related neoplasms, so the characterization of de novo HM is of interest. The study by Stubbins and colleagues ascertained patients based on the presence of a HM providing a valuable perspective on gBRCA1/2-associated cancers. The development of HM in patients with PV in gBRCA1/2 could be either incidental, with a risk similar to that in the general population,3 or causal, based on the gBRCA1/2 PV. Differentiating between these two possibilities is the greatest clinical concern to patients and providers, but the study by Stubbins et al. is neither designed nor powered to address this issue. The authors suggest that the relative frequency of gBRCA1/2 PV is enriched in their HM population compared to a reference (gnomAD) population. However, without ancestry matching, it is impossible to accurately determine whether it is truly higher, as the frequency of gBRCA1/2 PV varies among populations; 1:175 individuals in non-Finnish Europeans (0.6%)4 and 1:40 in Ashkenazi Jews (2.5%).5 The report of a 1.1% rate of gBRCA1/2 PV in HM could be based on representation of individuals from both populations, and enriched due to referral bias. In a significant proportion of tumors occurring in patients with BRCA1/2 PV, the mutant BRCA protein is biologically

Haematologica | 109 January 2024

EDITORIAL

P.D. Shah and K.L. Nathanson

agnoses and looking at gBRCA1/2 together are convenient, however, their consideration in aggregate detracts from the possibility of identifying a specific causal relationship. Previously published studies have rigorously examined qualitative and quantitative cancer risks conferred by gBRCA1/2 PV. A study including 3,184 BRCA1 and 2,157 BRCA2 families from the Consortium of Investigators of Modifiers of BRCA1/2 (CIMBA) estimated absolute risks for 22 first primary cancer types, adjusting for family ascertainment.7 No increased risk of leukemia (BRCA1: relative risk [RR]=0.90, 0.36-2.26, P=0.82; BRCA2: RR=0.91, 0.292.85, P=0.87), lymphoma (BRCA1: RR=1.03, 0.33-3.22, P=0.96; BRCA2: RR=0.97, 0.16-5.87, P=0.97), or multiple myeloma (BRCA1: RR=3.06, 0.83-11.26, P=0.09; BRCA2: RR=0.84, 0.10-7.31, P=0.87) was reported. Stubbins and colleagues note that this study ascertained patients based on known personal or family history of breast or ovarian cancer, with the possibility of pre-selection for a specific disease phenotype. Although bias is possible, it is extraordinarily unlikely that clinically meaningful risks of HM would have been undetected. Furthermore, characterization of cancers in a cohort of nearly 7,000 men with gBRCA1/2 PV showed 51 cases of HM (all subtypes) among 1,634 cancers noted (3.1%).8 By comparison, lymphoma, leukemia and multiple myeloma are estimated to account for 9.4% of new cancers in the USA in 2023.9 Therefore,

even in a BRCA1/2 population without a risk of female breast or ovarian cancer, HM are not overrepresented. We therefore read this exploratory study with interest, but also with concern that its findings, based on 16 patients with de novo HM from a single institution, may be misinterpreted or extrapolated to indicate a causal relationship between gBRCA1/2 and HM in general. The preponderance of currently published data from rigorously conducted studies refutes such causality. The current report, while thought-provoking, does not provide the breadth or depth of evidence necessary to contradict existing data. We agree with the authors’ conclusion that examining study populations specifically ascertained to look at inherited predispositions to HM, and families with BRCA1/2 PV and multiple cases of HM, would be of interest. However, we wish to reassure readers and the BRCA1/2 community that although this paper by Stubbins and colleagues demonstrates that individuals with BRCA1/2 PV are not exempt from HM, it does not substantiate a BRCA1/2associated general predisposition to HM. Disclosures No conflicts of interests to disclose. Contributions Both authors contributed equally.

References BRCA2. Nat Genet. 1996;14(2):185-187. 6. Sokol ES, Pavlick D, Khiabanian H, et al. Pan-cancer analysis of BRCA1 and BRCA2 genomic alterations and their association with genomic Instability as measured by genome-wide loss of heterozygosity. JCO Precis Oncol. 2020;4:442-465. 7. Li S, Silvestri V, Leslie G, et al. Cancer risks associated with BRCA1 and BRCA2 pathogenic variants. J Clin Oncol. 2022;40(14):1529-1541. 8. Silvestri V, Leslie G, Barnes DR, et al. Characterization of the cancer spectrum in men with germline BRCA1 and BRCA2 pathogenic variants: results from the Consortium of Investigators of Modifiers of BRCA1/2 (CIMBA). JAMA Oncol. 2020;6(8):1218-1230. 9. Siegel RL, Miller KD, Wagle NS, Jemal A. Cancer statistics, 2023. CA Cancer J Clin. 2023;73(1):17-48.

1. Stubbins RJ, Asom AS, Wang P, Lager AM, Gary A, Godley LA. Germline loss of function BRCA1 and BRCA2 mutations and risk of de novo hematopoetic malignancies. Haematologica. 2024;109(1):351-356. 2. Schulz E, Valentin A, Ulz P, et al. Germline mutations in the DNA damage response genes BRCA1, BRCA2, BARD1 and TP53 in patients with therapy related myeloid neoplasms. J Med Genet. 2012;49(7):422-428. 3. Jonsson P, Bandlamudi C, Cheng ML, et al. Tumour lineage shapes BRCA-mediated phenotypes. Nature. 2019;571(7766):576-579. 4. Karczewski KJ, Francioli LC, Tiao G, et al. The mutational constraint spectrum quantified from variation in 141,456 humans. Nature. 2020;581(7809):434-443. 5. Roa BB, Boyd AA, Volcik K, Richards CS. Ashkenazi Jewish population frequencies for common mutations in BRCA1 and

Haematologica | 109 January 2024

REVIEW ARTICLE

Unresolved laboratory issues of the heterozygous state of β-thalassemia: a literature review Shyamali Thilakarathne,1,2 Udayanga P. Jayaweera3 and Anuja Premawardhena4 1

Faculty of Graduate Studies, University of Kelaniya, Dalugama; 2Department of Medical Laboratory Science, Faculty of Allied Health Sciences, University of Peradeniya, Peradeniya; 3 Divisional Hospital, Medawala-Haarispatthuwa and 4Department of Medicine, Faculty of Medicine, University of Kelaniya, Ragama, Sri Lanka

Correspondence: S. Thilakarathne shyamali5thilakarathne@gmail.com shyamali.thilakarathne@ahs.pdn.ac.lk Received: Accepted: Early view:

February 15, 2023. May 19, 2023. June 1, 2023.

Abstract Although considered a mild clinical condition, many laboratory issues of the carrier state of β-thalassemia remain unresolved. Accurate laboratory screening of β-thalassemia traits is crucial for preventing the birth of a β-thalassemia major child. Identification of carriers in the laboratory is affected by factors that influence red cell indices and HbA2 quantification. Silent mutations and co-inheriting genetic and non-genetic factors affect red cell indices which decreases the effectiveness of the conventional approach. Similarly, the type of β mutation, co-inheriting genetic and non-genetic factors, and technical aspects, including the analytical method used and variations in the HbA2 cut-off values, affect the HbA2 results, leading to further confusion. However, the combination of mean corpuscular volume, mean corpuscular hemoglobin, and hemoglobin analysis increases the diagnostic accuracy. Diagnostic problems arising from non-genetic factors can be eliminated by carefully screening the patient’s clinical history. However, issues due to certain genetic factors, such as Krüppel-like factor 1 gene mutations and α triplication still remain unresolved. Each laboratory should determine the population-specific reference ranges and be wary of machine-related variations of HbA2 levels, the prevalence of silent mutations in the community.

partner is also a carrier. The World Health Organization (WHO) emphasizes the importance of incorporating carrier screening in basic health services in countries with a high incidence of thalassemia.4 Hematology laboratories play a vital role in carrier identification. Screening thalassemia traits using blood analysis has two methodological approaches: 1) primary screening followed by a secondary screening; 2) complete screening. In the former approach, only subjects with reduced mean corpuscular volume (MCV) and/or mean corpuscular hemoglobin (MCH) will be assessed for hemoglobin patterns and HbA2 levels. In contrast, hemoglobin analysis is carried out for all subjects from the outset in the latter.5 The first approach has notable disadvantages over its limited advantages, which include low cost. In the presence of mild/silent mutations, β-promoter mutations, α thalassemia and HbD and/or β variants, β-thalassemia carriers may show normal/marginal red cell indices.6 In addition, also megaloblastic anemia and hereditary persistence of fetal hemoglobin (HPFH) can mask the microcytosis of β-thalassemia carriers.7,8 Other than these clinical conditions, technical problems such as sample storage can affect

Introduction β-thalassemia is one of the most common hereditary blood disorders.1 It is characterized by the reduction of (β+) or absence of (β0) synthesis of the β globin chains of the hemoglobin (Hb) tetramer. More than 200 different mutations of β globin genes are currently recognized.1 Most of the mutations are single nucleotide substitutions, deletions, or insertions of oligonucleotides leading to frameshifts. More rarely, β-thalassemia results from gene deletions. Clinical severity could be affected by the mutation type.2 Based on clinical and hematologic severity, there are three main forms of β-thalassemia syndromes, i.e., the β-thalassemia carrier (trait) state, thalassemia intermedia (nontransfusion-dependent thalassemia), and thalassemia major (transfusion-dependent thalassemia). The current review will focus on β-thalassemia carrier state (BTT). It has been estimated that about 1.5% of the global population (80-90 million people) are carriers of βthalassemia.3 β-thalassemia traits have a 25% risk of having children affected with thalassemia major if their

Haematologica | 109 January 2024

REVIEW ARTICLE - Laboratory issues of β-thalassemia trait MCV results. Thus the first approach will misdiagnose a proportion of β-thalassemia carriers. The second approach is also not foolproof. The conventional method, high-performance liquid chromatography (HPLC), and capillary electrophoresis (CE) are two techniques used for quantifying HbA2. When the same sample is analyzed using both techniques, machine-related variations in HbA2 have been reported.9,10 Factors such as δ-hemoglobinopathies, δβ0-thalassemia, α-thalassemia, triple-α, hemoglobin variants, iron deficiency anemia (IDA), hyperthyroidism, megaloblastic anemia, HPFH, antiretroviral drugs, and Krüppel-like factor 1 (KLF1) gene mutations can affect the HbA2 levels.8,11,12 Labeling a person as BTT based on HbA2 level without considering the above factors may not be reliable. An HbA2 cut-off of 3.5% is generally used for the diagnosis of BTT,13 though different reference ranges for HbA2 are used in some countries.14 HbA2 values between the upper limit of the normal range and the cut-off value are considered borderline. The diagnosis of borderline samples is usually made only by doing confirmatory genetic tests such as polymerase chain reaction and β-gene sequencing. The current review aims to discuss these unresolved problems in the BTT state with reference to the laboratory aspects.

Methods Search strategy We searched databases of MEDLINE via Pub Med, Google Scholar and Taylor & Francis for research studies published in English over the past 20 years (October 2002 to September 2022) using the following keywords in different combinations: Beta-thalassemia trait, Beta-thalassemia minor, Heterozygous state of beta-thalassemia, Screening carriers, Diagnostic problems, Laboratory issues, HbA2, Borderline, Cutoff, Techniques, Iron deficiency. Inclusion criteria Prospective, descriptive, and retrospective studies, including laboratory issues of the β-thalassemia carrier state in different countries of the world (India, China, Italy, Thailand, Iran, Saudi Arabia, Turkey, Egypt, Malaysia, the US, the UK, Bangladesh, Pakistan, Canada, the Netherlands, the UAE, Portugal, Iraq, Bahrain, Singapore, and Spain) were included in the present review.

S. Thilakarathne et al.

Data extraction Two researchers (ST and UPJ) independently reviewed all abstracts of journal articles gathered by the web search to identify articles that required full-text review. All selected articles were discussed with a third independent reviewer (AP). Data on study design, objectives, methodology, and results of the selected articles were extracted and methodically reviewed.

Results Through the investigation strategy, we identified 810 citations, from which 68 articles were selected for qualitative synthesis (Figure 1). The study designs excluding case studies, abstracts, and reviews, included prospective studies, retrospective studies, and short communications. Out of the 68 articles selected, 14 (20.6%) originated from India, ten (14.7%) from Thailand, nine (13.2%) from Italy, eight (11.8%) from China, six (8.8%) from Iran, two (2.9%) each from Malaysia, Saudi Arabia, Pakistan, the Netherlands, and the UK, and one (1.5%) each from Sri Lanka, Bangladesh, Bahrain, Singapore, Egypt, the UAE, Portugal, Iraq, Canada, the US, Turkey, and Spain. Of these 68 studies, seven (10.3%) analyzed the effectiveness of the conventional approaches of β-thalassemia carrier screening. There were 60 studies (77.9%) focused on factors affecting red cell indices (37 studies) and/or HbA2 value (43 studies). Only one study (1.5%) assessed the conventional cut-off values used in BTT diagnosis. Effectiveness of the conventional approaches Out of seven studies that analyzed the effectiveness of conventional hematologic-analysis-based approaches of β-thalassemia carrier screening, 57.1% (n=4) compared the effectiveness of MCV and MCH values in diagnosing BTT based on the first approach.15-18 Sensitivity, specificity, positive predictive values, negative predictive values, and false negatives in different combinations have been used for comparisons. Of four studies, three concluded that MCH is more appropriate than MCV,15,16,18 while one concluded that MCV is more appropriate than MCH.17 Out of seven studies, 14.3% (n=1) showed that using Hb analysis as the first test is unreliable in detecting β-thalassemia traits.19 About 28.6% (n=2) compared the sensitivity, specificity, positive predictive value, negative predictive value and diagnostic accuracy among the conventional two approaches for thalassemia carrier detection.20,21 According to both studies, the combination of MCV, MCH and Hb analysis resulted in high sensitivity, specificity, and diagnostic accuracy.

Exclusion criteria Studies including β-thalassemia traits that did not describe problems in laboratory diagnosis were excluded. In addition, case reports, abstracts, reviews, unpublished Factors affecting mean corpuscular volume and mean studies, and duplicates of previously included studies corpuscular hemoglobin among β-thalassemia carriers were also excluded from the present review. Of 37 studies regarding factors affecting RBC indices, Haematologica | 109 January 2024

REVIEW ARTICLE - Laboratory issues of β-thalassemia trait

S. Thilakarathne et al.

Figure 1. Flow chart of article selection for the systematic review.

67.6% (n=25) identified silent or mild β-thalassemia carriers with normal or marginal MCV and/or MCH. Of 25 studies, 6 (24%) originated from Italy,6,22-26 5 (20%) from India,14,27-30 4 (16%) from China,31-34 2 (8%) each from Thailand,35,36 Iran37,38 and Pakistan,39,40 one (4%) each from Turkey,41 Bahrain,42 Malaysia,43 and Spain.44 Silent or mild β-thalassemia mutations reported from different countries are listed in Table 1. Out of 37 studies, 13.5% (n=5) analyzed the effect of the coinheritance of α- and β-thalassemia on MCV and MCH.34,35,4547 SEA type was the most common α-thalassemia type that co-existed with β-thalassemia in three studies from Thailand, China and Singapore while 3.7 kb deletion was the most common type in the other two studies from Thailand and Iran. Of five studies, four (80%) reported a significant increase in MCV and MCH values among heterozygotes for α- and β-thalassemia compared to the simple β-thalassemia carriers.34,35,46,47 The same four studies revealed that this increase in MCV and MCH values is higher when one or two α-globin gene deletions co-existed with β-thalassemia than in other α-gene arrangements. However, in all four studies, mean values of MCV and MCH in heterozygotes for α- and β-thalassemia were less than 80 fL and 27 pg. Out of 37 studies, 2.7% (n=1) showed the effect of fetal Hb in HPFH on MCV and MCH of BTT.8 The study showed a significant increase in MCH and MCV values, accounting for 29% and 30% of their respective variances. This effect was highest in individuals who have inherited two copies of the HPFH quantitative trait loci (QTL), least in those without HPFH, and intermediate in subjects with only one copy of the QTL. However, mean values of MCV and MCH were less than 80 fL and 27 pg even in individuals who have inherited two copies of the HPFH QTL.

In the present review, we observed some non-genetic factors affecting MCV and MCH. Out of 37 studies, 2.7% (n=1) focused on normalized red cell parameters of BTT in patients with megaloblastic anemia.7 Treatment with vitamin B12 and folic acid reduced the mean MCV of patients from 85.28±12.49 to 74.45±9.74. According to 13.5% of studies (n=5), antiretroviral therapy (ART) causes an increase in MCV and MCH among HIV-infected patients.48-52 Pornprasert et al.50 revealed that at 6 and 12 months after the ART therapy, the mean MCV of β-thalassemia patients shifted from microcytic level (<80 fL) to normocytic level (80-100 fL). Mean MCH increased to normal levels (27-31 pg). Factors affecting HbA2 Of 43 studies focusing on factors affecting HbA2 levels, 34.9% (n=15) reported the type of β-thalassemia mutation itself as one of the factors associated with reduced HbA2 level among β-thalassemia carriers. Four studies (26.7%) from Thailand,36,53-55 three (20%) from Italy,6,11,56 two (13.3%) from China,33,34 two (13.3%) from India,14,30 and one (6.7%) each from Iran,38 Malaysia,57 Saudi Arabia,58 and Portugal59 reported different types of β-thalassemia mutations resulting in borderline or normal HbA2 levels in β-thalassemia carriers. Out of 43 studies, 39.5% (n=17) investigated the effect of α-thalassemia on HbA2 level. Of 17 studies, two (11.8%) compared pure α-thalassemia traits and normal individuals, and both reported reduced HbA2 levels in α-thalassemia traits.60,61 One study reported that the reduction in HbA2 expression depends on the number of α-globin gene defects.60 Co-inheritance of β-thalassemia with Hb Constant Spring was associated with lower HbA2 expression when compared to other types of α-thalassemia.60 Out of

Haematologica | 109 January 2024

REVIEW ARTICLE - Laboratory issues of β-thalassemia trait

S. Thilakarathne et al.

Table 1. Silent/mild β-thalassemia mutations causing normal or marginal mean corpuscular volume.

Country Italy India Malaysia Pakistan Bahrain Turkey Iran China Spain Thailand

Silent/mild mutations* CAP+1570 (T>C), β -101C →G, β-101, β +45 (G>C), β -101 (C →T), β -92 (C →T), β -54 (G →A), Poly A Capsite +1 (A→C), CD 16 (-C), -88 (C →T), Poly A (T>C), IVS-1-5 (G>C) Codon 17 (A>T), IVS-1-5 (G>C) CAP+1 (A>C) -71 (C>T), -101 (C>T) HBB:c.*+108 A>G and HBB:c.*+132 C>T –101 (C → T) , IVS II 844 (C → G), IVS I 128 (T>G) -31 (A>C), -50 (G>A), -73 (A>T), –90(C>T), -88 (C>T) -27 (A>T), -28 (A>G), -101 (C>T), -29 (A>G), -86 (C>A), -88 (C>T) NT-28 (A>G), NT-31 (A>G), NT-87 (C>A), NT-50 (G>A), CAP+1 (A > C), Codon 126 (T > G)

*Mean corpuscular volume (MCV) / mean MCV >70 fL.

17 studies, 47.1% (n=8) compared HbA2 levels between βthalassemia traits and β-thalassemia traits co-inherited with α-thalassemia.34-36,45-47,55,60 Although four studies (50%) observed a reduction in HbA2 level when α-thalassemia co-exists,34,35,47,55 only two studies (25%) reported a possible effect of this on BTT diagnosis.47,55 Of these two, one showed that HbA2<4.0% can occur when HbH disease (α-thalassemia intermedia) co-exists with the BTT.55 Surprisingly, out of 17 studies, 47.1% (n=8) reported a marginal increase in HbA2 levels among α-thalassemia traits without β-thalassemia heterozygosis.6,11,33,38,54,55,57,58 None of the studies could explain the underlying mechanism. Of the studies that investigated factors affecting HbA2, 13.9% (n=6) from four different countries (Italy,6,11,56 India,62 Thailand,60 and Portugal59) identified that co-inheritance of δ-hemoglobinopathy decreases the HbA2 level in β-thalassemia carriers resulting in misdiagnosis. About 11.6% (n=5) evaluated the effect of α-gene triplication on the HbA2 level. Out of these five studies, four reported the presence of an ααα condition in healthy individuals with borderline increased HbA2 levels but without a β-globin mutation.6,11,33,34 Only one study stated that α-gene triplication does not affect hematologic parameters.63 Similarly, out of 43 studies, only 2.3% (n=1) found that co-inherited HPFH in BTT reduced HbA2 levels.8 The effect increased with the number of copies of the HPFH allele. Similarly, 18.6% (n=8) eligible studies commented on the effect of KLF1 gene mutations. KLF1 or Erythroid KruppelLike Factor is a zinc-finger transcription factor that has different functions in erythropoiesis including modifying chromatin architecture, regulating β-like globin gene switching, and activating or repressing gene transcription.64 Of eight studies on the effect of KLF1 gene mutations, seven (87.5%) observed a high prevalence of KLF1 gene defects among individuals with borderline increased HbA2 levels

without an HBB gene mutation.12,26,33,34,54,65,66 They concluded a possible effect of KLF1 gene mutations on increased HbA2 levels among normal individuals. Only one study (12.5%), which was a cross-sectional analysis in Saudi Arabia, reported that KLF1 gene mutations are not only associated with borderline high HbA2 results but also normal (HbA2<3%) and high (HbA2>4.3%) results.58 Out of seven studies which determined the effect of KLF1 gene mutations on HbA2 level, the studies of Perseu et al.,12 Hariharan et al.,65 and Liu et al.66 revealed 0.8%, 1.1%, and 0.5% increases, respectively, in HbA2 level among individuals with KLF1 gene mutations when compared to the group without KLF1 gene mutations. Interestingly, Liu et al.66 had shown a high prevalence of KLF1 genes in β-thalassemia endemic regions in China and an amelioration of the severity of βthalassemia by KLF1 gene changes.66 We were able to determine the most common KLF1 mutations in different countries; G176Afsx179 was the most common KLF1 mutation in China and Thailand, while p ser. 270 and -148 (G<A) were the most common in Italy and India, respectively. In the present study, we analyzed the effect of non-genetic factors on HbA2 level. Out of 43 studies, 20.9% (n=9) investigated the association between iron deficiency anemia (IDA) and HbA2 levels. Of these nine studies, four (44.4%) determined the effect of IDA on HbA2 results of normal (non-thalassemic) individuals.56,60,67,68 All concluded that IDA causes a significant reduction in HbA2 levels in normal individuals. Eight studies (88.9%) compared the results of HbA2 among β-thalassemia traits with and without IDA. Out of these eight studies, five (62.5%) showed significantly higher HbA2 values for β-thalassemia traits without IDA than in those with IDA;30,61,67,69,70 in the other three, no difference was observed.56,68,71 Of five studies that observed higher values, three (60%) suggested possible effects on βthalassemia diagnosis,30,61,67 while others showed no effect

Haematologica | 109 January 2024

REVIEW ARTICLE - Laboratory issues of β-thalassemia trait due to the steadying of all HbA2 values above the cutoff.69,70 Among those studies that suggested possible effects, one proposed a dose effect of severity of iron deficiency on HbA2 level of β-thalassemia traits.61 Out of 43 studies, 4.6% (n=2) reported the effect of megaloblastic anemia in elevated HbA2 levels among normal individuals without β-thalassemia.7,68 One of the above studies reported a change of mean HbA2 of the study group from 4.56% to 3.81% with vitamin B12 and folic acid treatments, whereas 93.75% of subjects diagnosed as BTT (HbA2>3.5%) before treatment became normal (HbA2<3.5%) after the treatment.7 Among 43 studies that discussed factors affecting HbA2 levels, 2.3% (n=1) indicated the effect of thyroid hormones on the production of HbA2.72 According to the results, HbA2 was significantly higher in hyperthyroid (mean: 2.77%) patients than in the controls (2.39%). However, this increase in HbA2 did not pass the cut-off level of HbA2, meaning there was less chance of misdiagnosing BTT. Five studies analyzed the effect of antiretroviral therapy (ART) on HbA2 results.48-52 All of them showed an increase in HbA2 levels among HIV-infected patients treated with ART leading to a misdiagnosis of BTT. Of five studies, three (60%) confirmed that zidovudine increased HbA2 levels.48,51,52 Bhagat et al.48 evaluated the effect of three different drugs, and concluded that both zidovudine and stavudine increase HbA2 levels, while tenofovir had no effect. In the present review, we observed technical issues due to the variations in analytical methods used to determine HbA2 level in β-thalassemia screening. Out of 43 studies, 6.9% (n=3) compared HbA2 results by different analytical systems in the presence of Hb variants.9,10,73 All three studies determined falsely increased HbA2 in HPLC reports of patients with HbE and HbS variants. However, the CE method clearly separated HbE and HbA2. One study reported the variant Hb Lepore, which co-elutes with HbA2 in the HPLC method causing problems in diagnosis.73 Similarly, when the HPLC technique is used, falsely reduced HbA2 had been observed in the presence of HbD Punjab in two studies.9,73 According to two studies, HbA2 peak was included in HbC peak when the CE method was used.9,10 Interestingly, all three studies reported variations in HbA2 values between HPLC and CE methods when comparing HbA2 measurements in samples within the normal range. Two studies reported higher HbA2 values by the Bio-Rad variant II HPLC method (mean: 2.95% and 2.67%, respectively) than the Sebia CE method (mean: 2.49% and 2.51%, respectively).9,73 On the contrary, one study reported HbA2 levels by the Sebia CE method (mean: 2.8%) was higher than that of the Primus HPLC method (mean: 2.3%).10

S. Thilakarathne et al.

based on borderline HbA2 values) clearly mentioned the HbA2 cut-off value used in diagnosing β-thalassemia carriers. Of 37 articles, the majority (48.6%; n=18) used 3.5% as the cut-off value, while 32.4% (n=12) of studies used 4.0% as the cut-off. Eighteen studies that used 3.5% as the cut-off include five from China,18,21,32,34,70 five from Iran,15,16,37,47,67 two from Thailand,46,53 one each from India,51 Bangladesh,20 Saudi Arabia,19 Portugal,59 the UAE,61 and the UK.52 Similarly, 12 studies that used 4.0% as the cut-off include four from India,7,17,27,48 four from Thailand,35,49,51,55 and one each from Malaysia,43 Bahrain,42 Italy,24 and the Netherlands.73 A 3.9% cut-off was used for three studies from India (n=2)68,71 and Italy (n=1).12 Two studies from Italy26 and Turkey41 used 3.8% as the cut-off while 3.4%, 3.6%, and 3.7% were used as the cut-off values by three separate studies from Italy,56 Thailand,36 and India,28 respectively. Abdel-Messih et al.74 determined the effectiveness of the two most common cut-off values used in BTT screening. The results revealed the sensitivity (100.0% and 97.4%), specificity (70.0% and 72.7%), positive predictive value (75.0% and 92.6%), negative predictive value (100.0% and 88.8%), and accuracy (70.0% and 92.0%) in identifying βthalassemia carriers at 3.5% and 4.0% cut-off values, respectively. There were 16 (23.5%) studies which analyzed borderline HbA2 results in BTT diagnosis; nine different borderline ranges were identified among these studies (Table 2).

Discussion

The diagnosis of β-thalassemia using red cell indices as the primary screening or using HPLC or CE as the first test each seem to have their own deficiencies. In the first approach, MCV and MCH are crucial hematologic parameters to rule out possible β-thalassemia traits. Several studies reported that MCH is more appropriate than MCV.15,16,18 Most of these studies generate sensitivity figures based on their own cut-off values for MCV and MCH. MCH is considered to be more stable in ambient temperature while MCV increases with time after being stored for several hours before testing;75 this stability of MCH may be the reason for its higher sensitivity. Some clinicians use only MCV to exclude β-thalassemia in the primary screening.36 In such cases, falsely high MCV values due to stored samples may give misleading results. Therefore, it is important to consider not one but both MCV and MCH results in screening for β-thalassemia. Reduction in any parameter should be considered as having diagnostic importance. One of the main disadvantages of using red cell indices Problems with HbA2 cut-off values for screening includes the possibility of missing silent βIn the present review, we analyzed the variations in cut- thalassemia carriers, as red cell indices may be normal in off values of HbA2 used in β-thalassemia screening. Out these individuals. Silent β-thalassemia mutations exist not of 68 studies, 54.4% (n=37) articles (excluding studies only in codons, but also in the promoter region and/or the Haematologica | 109 January 2024

REVIEW ARTICLE - Laboratory issues of β-thalassemia trait cap region of the HBB gene.24,28,41,76 Screening for thalassemia based on red cell indices will also result in missing individuals with HbD trait and individuals with triple α genes.77 Although phenotypically not important, and globally rare, the co-inheritance of the above conditions with β-thalassemia mutations leads to severe phenotype in the offspring. In the present review, we analyzed the multitude of factors that may affect HbA2 levels, and thus the diagnosis of β-thalassemia (Table 2). The problems may arise in two circumstances: decreased HbA2 levels in β-thalassemia traits, or deceptively elevated HbA2 levels in non-carriers. The type of β-thalassemia mutation is the main reason for reduced or borderline HbA2 levels among carriers.30 Silent or mild β-thalassemia mutations show reduced/borderline HbA2 levels along with normal or slightly reduced RBC indices. Most of the time, homozygotes for classic and silent β-thalassemia mutations have thalassemia intermedia phenotype.76,78 Unfortunately, silent carriers will be missed until the birth of a child with β-thalassemia intermedia. Co-inheritance of α-thalassemia with β-thalassemia results in decreased HbA2 levels due to the unavailability of a free α-chain pool to partner with δ-globin chains.2 This reduction seems insufficient to cause problems in BTT diagnosis36,60 unless mild or silent β-thalassemia also coexists. Satthakar et al.55 showed HbA2 less than 4.0% only in HbH and HbH-CS diseases; since both these conditions are apparent in HPLC patterns, β-thalassemia traits will not be missed. Surprisingly, borderline increases in HbA2 levels have also been identified among individuals with αthalassemia but without β-thalassemia mutations.33,55,58 The exact reason for that has still not been discovered, while some researchers consider regulation factors or mutations in regulatory genes to be causes.33,38 Co-inheritance of δ-hemoglobinopathy60 and HPFH8 cause reduced HbA2 levels in β-thalassemia carriers. The synthesis of insufficient amounts of δ-globin monomers results in lower HbA2 levels in δ-hemoglobinopathy. On the other hand, co-inheritance of heterocellular HPFH leads to a primary increase in γ-chain synthesis resulting in high HbF levels.8 However, the reason for decreased HbA2 level is not clear. The effect of α-triplication and KLF1 gene mutations on increased HbA2 levels is contradictory. α-triplications are considered uncommon because de novo crossover events generating triplications are rare.63 The KLF1 gene has a regulatory effect on erythropoiesis.79 Perseu et al.12 first identified a delay in the transcriptional switch from the HBD to the HBB gene by KLF1 mutations. The majority of the studies regarding α-triplication and KLF1 gene mutations in the current review reported possible effects on increased HbA2 levels. Most of these studies have specifically selected cohorts with borderline HbA2 results; because of the marginally increased HbA2 levels and

S. Thilakarathne et al. Table 2. Factors affecting HbA2 results of high-performance liquid chromatography and capillary electrophoresis methods.

HPLC

Falsely increased HbA2

Falsely decreased HbA2

KLF1 gene mutation* Megaloblastic anemia* Hyperthyroidism* Antiretroviral therapy* HbE HbS Hb Lepore α-thalassemia* δ-hemoglobinopathy* Hereditary persistence of fetal Hb* Iron deficiency anemia* HbD Punjab HbC

*These factors affect HbA2 results of both high-performance liquid chromatography (HPLC) and capillary electrophoresis (CE) methods. Hb: hemoglobin.

absence of β-thalassemia mutations, possible effects have been hypothesized. Among eligible studies, there were a few of a cross-sectional nature, with some studies including members of the general population, but they revealed no association between increased HbA2 level and α-triplication/KLF1 mutations.58,63 Therefore, larger scale studies that include the general population are needed in order to gather more comprehensive and informative data. IDA is the major cause of anemia in the world.80 Co-existence of IDA may decrease the HbA2 level in β-thalassemia traits. Intracellular lack of iron reduces α-globin chain synthesis more than non-α-globin chains. In β-thalassemia, β-globin chain synthesis is also limited. Therefore, β chains compete more effectively for α-globin chains than δ-globin chains, resulting in reduced levels of HbA2.81 It is widely believed that IDA cannot reduce HbA2 levels of βthalassemia traits below the cut-off values. If this is the case, the effect of IDA on β-thalassemia diagnosis would be minimal, except in cases of severe IDA or silent/mild β-thalassemia mutations. Megaloblastic anemia, ART therapy, and hyperthyroidism are other non-genetic factors which increase HbA2 levels in β-thalassemia traits. Both megaloblastic anemia and long-term ART therapy (in particular, zidovudine and stavudine) affect red cell synthesis. Nuclear maturation of red cell precursors is delayed in megaloblastic anemia, while antiretroviral medications inhibit nucleic acid synthesis.50 So, more Hb synthesis occurs in immature erythroid precursors. Since the synthesis of δ chains is relatively greater in these cells, HbA2 level is increased.81 Besides this, thyroid hormones are considered to specifically increase the expression of δ globin resulting in high HbA2 levels in hyperthyroidism.72 No literature was found which demonstrated an increased HbA2 level above the cut-off in hyperthyroidism. Nevertheless, a careful medical history

Haematologica | 109 January 2024

REVIEW ARTICLE - Laboratory issues of β-thalassemia trait

S. Thilakarathne et al.

Table 3. Nine different borderline ranges for HbA2 in diagnosing β-thalassemia traits by 16 studies included in the present review.

Borderline range (HbA2%)

3.0-3.9

3.0-4.0

3.1-3.9

3.3-3.7 3.3-3.9 3.3-4.0 3.3-4.1 3.5-3.9 3.5-4.0

Study

% β-thalassemia heterozygotes

Bahar et al.57

30.8 (36/117)

Colaco et al.30

72.6 (149/205)

Rangan et al.14 Borgio et al.58

32 (8/25) 83.3 (60/72)

Giambona et al.6

14.8 (61/410)

Chaweephisal et al.36 Moradi et al.38 Mosca et al.11 Hariharan et al.65 Lou et al.33 Perseu et al.12 Paglietti et al.26 Jiang et al.34 Satthakarn et al.55 Rungsee et al.53 Srivorakun et al.54

5.6 (6/106) 36.4 (159/436) 16.2 (38/234) NM 2.4 (4/165) NM 18.9 (7/37) 17.7 (11/62) 5.33 (16/300) 1.9 (3/158) 10.9 (22/202)

Remarks 10 (27.0%) had HbA2 level of 3.0% 20 of 149 [14.0%] of all β-thalassemia traits would be missed if MCV < 80 fL and HbA2≥3.5% were used as cut-off All had HbA2 levels ranging from 3.5% to 3.9% Exclusion criteria: iron deficiency anemia and Hb variants 20 out of 61 β-thalassemia carriers had MCV>80fL Exclusion criteria: MCV>80fL Exclusion criteria: MCV>80fL MCH>27pg Majority with MCV>80fl Tested only for β0 mutations -

MCV: mean corpuscular volume; MCH: mean corpuscular hemoglobin; Hb: hemoglobin; NM: not mentioned.

of the individual is warranted prior to β-thalassemia screening. There are machine-related variations in HbA2 results of normal individuals or β-thalassemia traits. A comparison of the Bio-Rad variant II HPLC and Sebia CE methods showed consistently high results have been obtained from the HPLC method,9,73 although there is good correlation between the two. When using a single cut-off value to diagnose β-thalassemia carriers, this will be a major problem, especially in borderline HbA2 results. Therefore, when issuing a diagnosis of BTT based on borderline values, machine-related variation of the HbA2 level should be borne in mind. The main reason for this unresolved problem is the non-availability of standardized HbA2 results between methods. Non-availability of a reference method and international reference HbA2 calibrators makes this problem more complex.9 In addition, in the present review, we identified marked differences in the HbA2 cut-off values (3.0-4.0%) used for diagnosis of BTT. Although it is recommended to define population or country specific cut-off values, sometimes 2 or 3 different cut-offs have been used within a single country. Similarly, different authors have defined borderline ranges arbitrarily. In the present review, the percentages of β-thalassemia heterozygotes in each borderline group ranged from 1.9% to 83.3%, which is very broad (Table 3);

the reason may be the differences in inclusion/exclusion criteria of the study group and variations in the analytical methods used to identify β-thalassemia mutations. However, it is noteworthy that all these β-thalassemia traits would be missed if HbA2≥4.0% was used as the cut-off. We believe a cut-off with a higher sensitivity than specificity is suitable for screening to prevent missing β-thalassemia traits. If problems arise in the accurate diagnosis of β-thalassemia heterozygosis due to the borderline results, the parental studies may provide prompt solutions. If not, further genetic or molecular tests are necessary. Conclusions and recommendations Many laboratory issues of the heterozygous state of β-thalassemia remain unresolved. During the screening of βthalassemia, both RBC indices and HbA2 level should be taken into consideration. Population-specific reference ranges and machine-related variations in HbA2 level should be determined within the laboratory. Borderline HbA2 levels should be further investigated for factors affecting HbA2 levels. Careful screening of the patient’s clinical history can identify most non-genetic factors that affect HbA2 screening. IDA and α-thalassemia are unlikely to affect the BTT diagnosis unless in individuals with silent β mutations. Based on the prevalence of silent mutations

Haematologica | 109 January 2024

REVIEW ARTICLE - Laboratory issues of β-thalassemia trait and other co-inheriting genetic factors, the diagnostic approach of BTT should be modified for each population. For example, screening the partners of patients with β-thalassemia heterozygosity for common silent β mutations, HbE, HbD, etc. is possible. However, effects of some genetic factors, such as KLF1 gene mutations and α-triplication on HbA2 level are still controversial, and further investigations are needed. Moreover, there may be other as yet unidentified factors, such as the environment, lifestyle, regulation factors or mutations in regulatory genes affecting HbA2 level. Further studies should aim to identify

S. Thilakarathne et al.

these factors and should include large numbers of participants. Disclosures No conflicts of interest to disclose. Contributions ST and UPJ wrote the first version of the manuscript. AP revised the manuscript critically for important intellectual content. All authors read, revised, and approved the final manuscript.

References 1. Weatherall DJ. Phenotype-genotype relationships in monogenic disease: lessons from the thalassaemias. Nat Rev Genet. 2001;2(4):245-255. 2. Thein SL. Genetic insights into the clinical diversity of betathalassaemia. Br J Haematol. 2004;124(3):264-274. 3. Colah R, Gorakshakar A, Nadkarni A. Global burden, distribution and prevention of β-thalassaemias and hemoglobin E disorders. Expert Rev Hematol. 2010;3(1):103-117. 4. Modell B, Darlison M. Global epidemiology of haemoglobin disorders and derived service indicators. Bull World Health Organ. 2008;86(6):480-487. 5. Angastiniotis M, Eleftheriou A, Galanello R, et al. Prevention of thalassaemias and other haemoglobin disorders: Volume 1: principles [Internet]. Old J, editor. 2nd ed. Nicosia (Cyprus): Thalassaemia International Federation; 2013. 6. Giambona A, Passarello C, Renda D, Maggio A. The significance of the hemoglobin A2 value in screening for hemoglobinopathies. Clin Biochem. 2009;42(18):1786-1796. 7. Sahoo S, Sahu N, Das P, Senapati U. Effect of megaloblastic anemia on hemoglobin A2 and diagnosis of β thalassaemia trait. Indian J Pathol Microbiol. 2023;66(2):327-331. 8. Garner C, Dew T, Sherwood R, Rees D, Thein S. Heterocellular hereditary persistence of fetal haemoglobin affects the haematological parameters of beta-thalassaemia trait. Br J Haematol. 2003;123:353-358. 9. Higgins TN, Khajuria A, Mack M. Quantification of HbA(2) in patients with and without beta-thalassaemia and in the presence of HbS, HbC, HbE, and HbD Punjab hemoglobin variants: comparison of two systems. Am J Clin Pathol. 2009;131(3):357-362. 10. Keren DF, Hedstrom D, Gulbranson R, Ou C-N, Bak R. Comparison of Sebia Capillarys capillary electrophoresis with the Primus high-pressure liquid chromatography in the evaluation of hemoglobinopathies. Am J Clin Pathol. 2008;130(5):824-831. 11. Mosca A, Paleari R, Ivaldi G, Galanello R, Giordano PC. The role of haemoglobin A2 testing in the diagnosis of thalassaemias and related haemoglobinopathies. J Clin Pathol. 2009;62(1):13-17. 12. Perseu L, Satta S, Moi P, et al. KLF1 gene mutations cause borderline HbA(2). Blood. 2011;118(16):4454-4458. 13. Ryan K, Bain BJ, Worthington D, et al. Significant haemoglobinopathies: guidelines for screening and diagnosis. Br J Haematol, 2010;149(1):35-49. 14. Rangan A, Sharma P, Dadu T, Saxena R, Verma IC, Bhargava M. βthalassaemia mutations in subjects with borderline HbA2 values: a pilot study in North India. Clin Chem Lab Med.

2011;49(12):2069-2072. 15. Tari K, Alikhani S, Abbaszadehdibavar M, Kianinodeh F, Karami F, Atashi A. Evaluation of the sensitivity and specificity of MCH and MCV for screening of beta-thalassaemia minor. Int J Biomed Pub Health. 2018;1(4):184-186. 16. Karimi M, Rasekhi AR. Efficiency of premarital screening of betathalassaemia trait using MCH rather than MCV in the population of Fars Province, Iran. Haematologia. 2002;32(2):129-133. 17. Rathod DA, Kaur A, Patel V, et al. Usefulness of cell counterbased parameters and formulas in detection of beta-thalassaemia trait in areas of high prevalence. Am J Clin Pathol. 2007;128(4):585-589. 18. Liao C, Xie X-M, Zhong H-Z, Zhou J-Y, Li D-Z. Proposed screening criteria for β-thalassaemia trait during early pregnancy in Southern China. Hemoglobin. 2009;33(6):528-533. 19. Al-Amodi AM, Ghanem NZ, Aldakeel SA, et al. Hemoglobin A2 (HbA2) has a measure of unreliability in diagnosing βthalassaemia trait (β-TT). Curr Med Res Opin. 2018;34(5):945-951. 20. Noor FA, Sultana N, Bhuyan GS, et al. Nationwide carrier detection and molecular characterization of β-thalassaemia and hemoglobin E variants in Bangladeshi population. Orphanet J Rare Dis. 2020;15(1):15. 21. Zhao J, Li J, Lai Q, Yu Y. Combined use of gap-PCR and nextgeneration sequencing improves thalassaemia carrier screening among premarital adults in China. J Clin Pathol. 2020;73(8):488-492. 22. Dell’Edera D, Epifania AA, Malvasi A, et al. Incidence of βthalassaemia carrier on 1495 couples in preconceptional period. J Matern Fetal Neonatal Med. 2013;26(5):445-448. 23. Moi P, Faà V, Marini MG, et al. A novel silent beta-thalassaemia mutation in the distal CACCC box affects the binding and responsiveness to EKLF. Br J Haematol. 2004;126(6):881-884. 24. Vinciguerra M, Passarello C, Cassarà F, et al. Co-heredity of silent CAP + 1570 T>C (HBB:c*96T>C) defect and severe β-thal mutation: a cause of mild β-thalassaemia intermedia. Int J Lab Hematol. 2016;38(1):17-26. 25. De Angioletti M, Lacerra G, Sabato V, Carestia C. Beta+45 G C: a novel silent beta-thalassaemia mutation, the first in the Kozak sequence. Br J Haematol. 2004;124(2):224-231. 26. Paglietti M E, Satta S, Sollaino MC, et al. The problem of borderline hemoglobin A2 levels in the screening for βthalassaemia carriers in Sardinia. Acta Haematol. 2016;135(4):193-199. 27. Colah RB, Surve R, Sawant P, et al. HPLC studies in hemoglobinopathies. Indian J Pediatr. 2007;74(7):657-662.

Haematologica | 109 January 2024

REVIEW ARTICLE - Laboratory issues of β-thalassemia trait 28. Garewal G, Das R, Awasthi A, Ahluwalia J, Marwaha RK. The clinical significance of the spectrum of interactions of CAP+1 (A→C), a silent β-globin gene mutation, with other βthalassaemia mutations and globin gene modifiers in north Indians. Eur J Haematol. 2007;79(5):417-421. 29. Italia K, Sawant P, Surve R, et al. Variable haematological and clinical presentation of β-thalassaemia carriers and homozygotes with the Poly A (T→C) mutation in the Indian population. Eur J Haematol. 2012;89(2):160-164. 30. Colaco S, Colah R, Nadkarni A. Significance of borderline HbA2 levels in β thalassaemia carrier screening. Sci Rep. 2022;12(1):5414. 31. Chen X-W, Mo Q-H, Li Q, Zeng R, Xu X-M. A novel mutation of 73(A-->T) in the CCAAT box of the beta-globin gene identified in a patient with the mild beta-thalassaemia intermedia. Ann Hematol. 2007;86(9):653-657. 32. Zhao Y, Jiang F, Li D-Z. Hematological characteristics of β-globin gene mutation -50 (G>A) (HBB: C.-100G>A) carriers in mainland China. Hemoglobin. 2020;44(4):240-243. 33. Lou J-W, Li D-Z, Zhang Y, et al. Delineation of the molecular basis of borderline hemoglobin A2 in Chinese individuals. Blood Cells Mol Dis. 2014;53(4):261-264. 34. Jiang F, Chen G-L, Li J, Zhou J-Y, Liao C, Li D-Z. Analysis of the genotypes in a Chinese population with increased HbA2 and low hematological indices. Hemoglobin. 2018;42(3):154-158. 35. Yamsri S, Singha K, Prajantasen T, et al. A large cohort of β+thalassaemia in Thailand: molecular, hematological and diagnostic considerations. Blood Cells Mol Dis. 2015;54(2):164-169. 36. Chaweephisal P, Phusua A, Fanhchaksai K, Sirichotiyakul S, Charoenkwan P. Borderline hemoglobin A2 levels in northern Thai population: HBB genotypes and effects of coinherited alpha-thalassaemia. Blood Cells Mol Dis. 2019;74:13-17. 37. Nezhad FH, Nezhad KH, Choghakabodi PM, Keikhaei B. Prevalence and genetic analysis of α- and β-thalassaemia and sickle cell anemia in Southwest Iran. J Epidemiol Glob Health. 2018;8(3-4):189-195. 38. Moradi K, Alibakhshi R, Shafieenia S, Azimi A. Problem of borderline hemoglobin A2 levels in an Iranian population with a high prevalence of α- and β-thalassaemia carriers. Egypt J Med Hum Genet. 2022;23(1):61. 39. Khattak SAK, Ahmed S, Anwar J, Ali N, Shaikh KH. Prevalence of various mutations in beta-thalassaemia and its association with haematological parameters. J Pak Med Assoc. 2012;62(1):40-43. 40. Karim M, Moinuddin M, Babar S. Cap +1 mutation; an unsuspected cause of beta-thalassaemia transmission in Pakistan. Turk J Hematol. 2009;26:167-170. 41. Bilgen T, Clark OA, Ozturk Z, Yesilipek MA, Keser I. Two novel mutations in the 3’ untranslated region of the beta-globin gene that are associated with the mild phenotype of betathalassaemia. Int J Lab Hematol. 2013;35(1):26-30. 42. Al Moamen NJ, Mahdi F, Salman E, et al. Silent β-thalassaemia mutations at -101 (C>T) and -71 (C>T) and their coinheritance with the sickle cell mutation in Bahrain. Hemoglobin. 2013;37(4):369-377. 43. Abdullah UYH, Ibrahim HM, Mahmud NB, et al. Genotypephenotype correlation of β-thalassaemia in Malaysian population: toward effective genetic counseling. Hemoglobin. 2020;44(3):184-189. 44. Ropero P, Erquiaga S, Arrizabalaga B, et al. Phenotype of mutations in the promoter region of the β-globin gene. J Clin Pathol. 2017;70(10):874-878. 45. Law HY, Chee MKL, Tan GP, Ng ISL. The simultaneous presence of alpha- and beta-thalassaemia alleles: a pitfall of

S. Thilakarathne et al. thalassaemia screening. Community Genet. 2003;6(1):14-21. 46. Viprakasit V, Limwongse C, Sukpanichnant S, et al. Problems in determining thalassaemia carrier status in a program for prevention and control of severe thalassaemia syndromes: a lesson from Thailand. Clin Chem Lab Med. 2013;51(8):1605-1614. 47. Saleh-Gohari N, Bami MK, Nikbakht R, Karimi-Maleh H. Effects of α-thalassaemia mutations on the haematological parameters of β-thalassaemia carriers. J Clin Pathol. 2015;68(7):562-566. 48. Bhagat P, Sachdeva RK, Sharma P, et al. Effect of antiretroviral therapy on hemoglobin A2 values can have implications in antenatal beta-thalassaemia screening programs. Infect Dis (Lond). 2016;48(2):122-126. 49. Pornprasert S, Leechanachai P, Klinbuayaem V, et al. Effect of haematological alterations on thalassaemia investigation in HIV1-infected Thai patients receiving antiretroviral therapy. HIV Med. 2008;9(8):660-666. 50. Pornprasert S, Sonboon P, Kiatwattanacharoen S, et al. Evolution of hematological parameters in HIV-1-infected patients with and without thalassaemia carriages during highly active antiretroviral therapy. HIV Clin Trials. 2009;10(2):88-93. 51. Nigam JS, Bharti JN, Kumar D, Sharma A. A comparison of hemoglobin A2 levels in untreated and treated groups of HIV patients on ART including zidovudine. Patholog Res Int. 2013;828214. 52. Wilkinson MJ, Bain BJ, Phelan L, Benzie A. Increased haemoglobin A2 percentage in HIV infection: disease or treatment. AIDS (Lond). 2007;21(9):1207-1208. 53. Rungsee P, Kongthai K, Pornprasert S. Detection of the common South-East Asian β0-thalassaemia mutations in samples with borderline HbA2 levels. Clin Chem Lab Med. 2017;55(1):e17-e20. 54. Srivorakun H, Thawinan W, Fucharoen G, Sanchaisuriya K, Fucharoen S. Thalassaemia and erythroid transcription factor KLF1 mutations associated with borderline hemoglobin A2 in the Thai population. Arch Med Sci. 2020;18:112-120. 55. Satthakarn S, Panyasai S, Pornprasert S. Molecular characterization of β- and α-globin gene mutations in individuals with borderline HbA2 levels. Hemoglobin. 2020;44(5):349-353. 56. Passarello C, Giambona A, Cannata M, Vinciguerra M, Renda D, Maggio A. Iron deficiency does not compromise the diagnosis of high HbA2 β thalassaemia trait. Haematologica. 2012;97(3):472-473. 57. Bahar R, Hassan MN, Marini R, Shafini MY, Abdullah WZ. The diagnosis of beta-thalassaemia with borderline HbA2 level among Kelantan population. J Blood Disord Transfus. 2017;08:396. 58. Borgio JF, AbdulAzeez S, Al-Muslami AM, et al. KLF1 gene and borderline hemoglobin A2 in Saudi population. Arch Med Sci. 2018;14(1):230-236. 59. Morgado A, Picanço I, Gomes S, et al. Mutational spectrum of delta-globin gene in the Portuguese population. Eur J Haematol. 2007;79:422-428. 60. Singha K, Sanchaisuriya K, Fucharoen G, Fucharoen S. Genetic and non-genetic factors affecting hemoglobin A2 expression in a large cohort of Thai individuals: implication for population screening for thalassaemia. Am J Transl Res. 2021;13(10):11632-11642. 61. Denic S, Agarwal MM, Al-Dabbagh B, et al. Hemoglobin A2 lowered by iron deficiency and α-thalassaemia: should screening recommendation for β-thalassaemia change? Int Sch Res Notices. 2013:e858294. 62. Colaco S, Trivedi A, Colah RB, Ghosh K, Nadkarni AH. Masking of a β-thalassaemia determinant by a novel δ-globin gene defect [HbA2-Saurashtra or δ100(G2)Pro→Ser; HBD: C.301C>T] in Cis.

Haematologica | 109 January 2024

REVIEW ARTICLE - Laboratory issues of β-thalassemia trait

S. Thilakarathne et al. Lab Hematol. 2015;37(1):105-111. 72. Jaafar FH, Al-Tameemi W, Ali HH. The influence of thyroid hormones on hemoglobin A2 and F expression. New Iraqi J Med. 2009;5:78-81. 73. Van Delft P, Lenters E, Bakker-Verweij M, et al. Evaluating five dedicated automatic devices for haemoglobinopathy diagnostics in multi-ethnic populations. Int J Lab Hematol. 2009;31(5):484-495. 74. Abdel-Messih IY, Youssef SR, Mokhtar GM, et al. Clinical to molecular screening paradigm for β-thalassaemia carriers. Hemoglobin. 2015;39(4):240-246. 75. Rogers M, Phelan L, Bain B. Screening criteria for betathalassaemia trait in pregnant women. J Clin Pathol. 1995;48(11):1054-1056. 76. Aslan D. “Silent” β-thalassaemia mutation (promoter nt-101 C > T) with increased hemoglobin A2. Turk J Pediatr. 2016;58(3):305-308. 77. Giordano PC. Strategies for basic laboratory diagnostics of the hemoglobinopathies in multi-ethnic societies: interpretation of results and pitfalls. Int J Lab Hematol. 2013;35(5):465-479. 78. Bain BJ. Haemoglobinopathy diagnosis: algorithms, lessons and pitfalls. Blood Rev. 2011;25(5):205-213. 79. Tallack MR, Whitington T, Yuen WS, et al. A global role for KLF1 in erythropoiesis revealed by ChIP-seq in primary erythroid cells. Genome Res. 2010;20(8):1052-1063. 80. Miller JL. Iron deficiency anemia: a common and curable disease. Cold Spring Harb Perspect Med. 2013;3(7):a011866. 81. Steinberg M H, Adams J G. Hemoglobin A2: Origin, evolution, and aftermath. Blood. 1991;78(9):2165-2177.

Hemoglobin. 2014;38(1):24-27. 63. Giordano PC, Bakker-Verwij M, Harteveld CL. Frequency of alpha-globin gene triplications and their interaction with betathalassaemia mutations. Hemoglobin. 2009;33(2):124-131. 64. Gallagher PG. HbA2: at the borderline of the KLF. Blood. 2011;118(16):4301-4302. 65. Hariharan P, Colah R, Ghosh K, Nadkarni A. Differential role of Kruppel like factor 1 (KLF1) gene in red blood cell disorders. Genomics. 2019;111(6):1771-1776. 66. Liu D, Zhang X, Yu L, et al. KLF1 mutations are relatively more common in a thalassaemia endemic region and ameliorate the severity of β-thalassaemia. Blood. 2014;124(5):803-811. 67. Keramati MR, Maybodi N. The effect of iron deficiency anemia (IDA) on the HbA2 level and comparison of hematologic values between IDA and thalassaemia minor. Int J Hematol Oncol. 2006;17:151-156. 68. Rao S, Kar R, Gupta SK, Chopra A, Saxena R. Spectrum of haemoglobinopathies diagnosed by cation exchange-HPLC & modulating effects of nutritional deficiency anaemias from north India. Indian J Med Res. 2010;132(5):513-519. 69. Verma S, Gupta R, Kudesia M, Mathur A, Krishan G, Singh S. Coexisting iron deficiency anemia and beta-thalassaemia trait: effect of iron therapy on red cell parameters and hemoglobin subtypes. ISRN Hematol. 2014;293216. 70. Verhovsek M, So C-C, O’Shea T, et al. Is HbA2 level a reliable diagnostic measurement for β-thalassaemia trait in people with iron deficiency? Am J Hematol. 2012;87(1):114-116. 71. Sharma P, Das R, Trehan A, et al. Impact of iron deficiency on hemoglobin A2% in obligate β-thalassaemia heterozygotes. Int J

Haematologica | 109 January 2024

REVIEW ARTICLE

Ubiquitin-proteasome pathway-mediated regulation of the Bcl-2 family: effects and therapeutic approaches Galvin Le Qian Tang,1 Jolin Xiao Hui Lai1 and Shazib Pervaiz1-5

Correspondence: S. Pervaiz

Department of Physiology, Yong Loo Lin School of Medicine, National University of Singapore; 2Integrative Science and Engineering Programme (ISEP), NUS Graduate School (NUSGS), National University of Singapore; 3NUS Centre for Cancer Research (N2CR), Yong Loo Lin School of Medicine, National University of Singapore; 4NUS Medicine Healthy Longevity Program, National University of Singapore and 5National University Cancer Institute, National University Health System, Singapore

phssp@nus.edu.sg Received: Accepted: Early view:

June 12, 2023. August 10, 2023. August 17, 2023.

Abstract Proteasomal degradation of proteins represents an important regulatory mechanism in maintaining healthy homeostasis in cells. Deregulation of the ubiquitin-proteasome system is associated with various diseases as it controls protein abundance and turnover in cells. Furthermore, proteasomal regulation of protein turnover rate can determine a cell’s response to external stimuli. The Bcl-2 family of proteins is an important group of proteins involved in mediating cell survival or cell death in response to external stimuli. Aberrant overexpression of anti-apoptotic proteins or deletion of pro-apoptotic proteins can lead to the development of cancer. Unsurprisingly, proteasomal degradation of Bcl-2 proteins also serves as an important factor regulating the level of Bcl-2 proteins and thereby affecting the functional outcome of cell death. This review aims to highlight the regulation of the Bcl-2 family of proteins with particular emphasis on proteasomal-mediated degradation pathways and the current literature on the therapeutic approaches targeting the proteasome system.

Apoptosis and the Bcl-2 family of proteins

Introduction Cancer is mostly a disease of aging, with mutations accumulating over time. The first step of cancer development, known as tumor initiation, is thought to be due to mutations leading to abnormal proliferation.1 These mutations confer additional selective advantages to the cells, such as avoiding cell death, and those cells that have such mutations (including their progeny) will consequently become the bulk population. This clonal selection resembles the process of Darwinian evolution, with the caveat that it occurs microscopically instead of in the wild. With each clonal selection, the population will gain advanced neoplastic characteristics, ultimately forming a highly aggressive malignant tumor that threatens the patient.2 In general, mutations that promote oncogenic traits and/or repress tumor-suppressive traits are highly favorable towards tumor progression. Such mutations may occur in the ubiquitination process which affects the stability of onco- and tumor-suppressive proteins. In this review, we discuss how the various elements of the ubiquitination process, more specifically the E3 ubiquitin ligases and deubiquitinating enzymes (DUB) affect Bcl-2 family proteins and their functions in regulating apoptosis.

Physiologically, cell death and cell proliferation must be balanced to ensure normal cellular function. Dysregulated cell death may lead to disorders such as Parkinson disease, while dysregulated cell proliferation may lead to cancer. Therefore, a dynamic balance between cell death and cell proliferation is important and is highly regulated by anti-apoptotic and pro-apoptotic proteins such as Bcl2 family proteins. Bcl-2 family members are categorized into three distinct subsets, but all members contain one Bcl-2 homologous (BH) domain, which is the BH3 domain. The distinct subsets are the anti-apoptotic proteins, pro-apoptotic proteins, and BH3-only proteins. With regards to the anti-apoptotic Bcl-2 family proteins, they suppress apoptosis and are upregulated in cancer cells.3,4 These proteins include Bcl-2, Bcl-xL, and Mcl-1. Besides their canonical anti-apoptotic function, these proteins can also exert noncanonical effects such as redox regulation, thus highlighting their multifunctional role in maintaining cellular homeostasis.5 On the other hand, the pro-apoptotic proteins, namely Bax and Bak, serve to negate the effect of

Haematologica | 109 January 2024

REVIEW ARTICLE - Ubiquitination of Bcl-2 family proteins

G. Le Qian Tang et al.

the anti-apoptotic proteins of triggering apoptosis and are downregulated in cancer cells.4 They are important in the permeabilization of the mitochondrial membrane, resulting in the release of cytochrome c. Lastly, BH3-only proteins, namely BIM, PUMA, and NOXA, are responsible for either (i) directly binding to anti-apoptotic proteins and inactivating them (also known as sensitizers), or (ii) directly activating mitochondrial outer membrane permeabilization (MOMP) by activating the pro-apoptotic proteins (also known as activators). Activators can also bind to anti-apoptotic proteins and trigger their inactivation. Therefore, activators are sensitizers that have an added role, which is to bind and activate pro-apoptotic proteins. A list of binding partners of Bcl-2 family proteins can be found in Table 1. Mechanistically, apoptosis involves many protein interactions and signaling cascades, and is primarily divided into two pathways, the extrinsic and intrinsic pathways. The extrinsic pathway involves receptors and external ligands,6 which are not be covered in this review. The intrinsic pathway promotes downstream pathways such as the release of cytochrome c, Smac/DIABLO, and the serine protease HtrA2/Omi from mitochondria into the cytosol to trigger caspase-dependent apoptosis.7,8 Regardless, these pathways converge into the execution pathway, through which activated caspase 3 triggers downstream signaling cascades, leading to the degradation of chromosomal DNA, protease activation and formation of apoptotic bodies.9,10 Delving deeper into the intrinsic pathway, this pathway involves stimuli that directly target proteins or molecules within the cells. For example, radiation causes DNA dam-

age, which in turn upregulates tumor suppressors such as p53.11 p53 promotes transcription of pro-apoptotic proteins such as NOXA and PUMA.12,13 NOXA can then bind directly to the anti-apoptotic Mcl-1 and displace the pro-apoptotic activator, Bak, allowing Bak to oligomerize with Bax to induce MOMP.12,14 Besides acting on NOXA transcription, p53 has also been shown to activate Bax.15 Subsequently, Bax can be inserted into the mitochondrial membrane to promote MOMP (this process is not p53-dependent), which induces cytochrome c release. As a result, an apoptosome complex can be formed with cytochrome c, APAF-1, and procaspase 9. This apoptosome complex then activates caspase 9 which subsequently activates caspase 3 and the downstream signaling cascade resulting in apoptosis. The dysregulation of these proteins can be detrimental, as observed in tumor samples.16 Therefore, understanding the regulation of these proteins (summarized in Table 2), more specifically ubiquitination and deubiquitination (summarized in Figure 1), is important for future research and identification of potential therapeutic strategies for cancer.

Effect of ubiquitination on Bcl-2 family proteins Mcl-1 (Myeloid cell leukemia sequence-1) Mcl-1 is an anti-apoptotic protein and serves as a protooncogene. Canonically, Mcl-1 localizes to the outer mitochondrial membrane and exerts its function by

Table 1. Ligands of Bcl-2 family proteins.

Bcl-2 family protein

Mcl-1

Bcl-2

Ligand

Summary

Reference

Bak

Mcl-1 displaces Bak from the outer mitochondrial membrane, preventing apoptosis.

NOXA

NOXA enhances the MULE/Mcl-1 interaction which promotes Mcl-1 ubiquitination and degradation.

17, 18, 26, 27

BIM

BIM/Mcl-1 complex stabilizes both Mcl-1 and BIM.

Bak

Binding of Bcl-2 to Bak prevents Bax/Bak activation.

BIM

Binding of BIM to Bcl-2 prevents Bcl-2 activation and promotes apoptosis. ARTS binds to Bcl-2 and XIAP, promoting XIAP-dependent ubiquitination of Bcl-2. BIM targets Bax for oligomerization and activation.

Reciprocal protection when BIM and Mcl-1 form a bi-complex. Dephosphorylation of BIM by PP2A reduces βTrCP binding to BIM for ubiquitination.

19, 53

ARTS Bax Mcl-1 BIM

Bax

PP2A Rsk1/2, Aurora A

Phosphorylation of BIM by Rsk1/2 and Aurora A promotes βTrCP binding and ubiquitination.

Bcl-2

Bax binds to Bcl-2 to inhibit its activity. Responsible for Bax recruitment and activation to the mitochondrial membrane. Bcl-xL suppresses Bax activity through a direct interaction.

VDAC2 Bcl-xL

Haematologica | 109 January 2024

46 36 57 56, 57 39 65 67

REVIEW ARTICLE - Ubiquitination of Bcl-2 family proteins

G. Le Qian Tang et al.

Table 2. Regulatory proteins of Bcl-2 family proteins.

Bcl-2 family protein

Regulatory proteins Fbw7

Mcl-1

MULE USP9x Ku70

Bcl-2

BIM

Bax

PPARα Parkin XIAP USP1, USP8, USP 22 βTrCP APC/Cdc20 USP27x TRAIL USP30 Parkin IBRDC2

Effect Direct ubiquitination and destabilization of Mcl-1 by Fbw7 is inhibited by GSK3β-dependent Mcl-1 phosphorylation. Direct ubiquitination and destabilization of Mcl-1 by MULE. Deubiquitination of Mcl-1 by USP9x prolongs its half-life. Deubiquitination and stabilization of Mcl-1 by Ku70 is dependent on the C-terminus of Ku70. PPARα induces poly-ubiquitination on Bcl-2 and causes its degradation. Parkin induces mono-ubiquitination on Bcl-2 and stabilizes Bcl-2. XIAP expression positively correlates with Bcl-2 expression. No direct interaction concluded. Downregulation of these USP caused downregulation of Bcl-2 expression. Direct ubiquitination of BIM by βTrCP. Direct ubiquitination of BIM by APC/Cdc20. Deubiqutination and stabilization of BIM by USP27x. TRAIL promotes ubiquitination of Bax, leading to its activation. Deubiquitination and stabilization of Bax by USP30. Direct ubiquitination and destabilization of Bax by Parkin. Direct ubiquitination and destabilization of Bax by IBRDC2.

sequestering pro-apoptotic proteins such as Bak from the membrane.14 This prevents the dimerization of Bax and Bak, thereby suppressing MOMP and apoptosis. NOXA has also been shown to bind and reduce Mcl-1 anti-apoptotic activity.17,18 Short interfering (siRNA) knockdown of NOXA reduced camptothecin-induced cell death, possibly due to the increase in Mcl-1 upon campthotecin treatment.17 By forming a complex with Mcl-1, NOXA destabilizes Mcl-1 by promoting Mcl-1 phosphorylation and ubiquitination.18 BIM is another BH3-only protein that negatively regulates Mcl1 activity. However, the BIM/Mcl-1 complex has also been shown to stabilize both BIM and Mcl-1.19 Hence, the net result of such binding, pro- or anti-apoptotic, is dependent on the relative ratio and affinity of these proteins. To date, many E3 ubiquitin ligases and DUB have been documented to regulate Mcl-1 stability, some of which are discussed here. Prominently, Fbw7 ubiquitinates many glycogen synthase kinase (GSK3β)-phosphorylated substrates, including Mcl-1.20 GSK3β-dependent phosphorylation of Mcl-1 at both Ser159 and Thr163 of Mcl-1 is observed to promote its degradation.21 Notably, inhibition of AKT (a negative regulator of GSK3β) through the small molecule API-1 resulted in increased Ser159 and Thr163 phosphorylation and prompted Fbw7-dependent degradation of Mcl1.22 Such findings suggest that phosphorylation at these sites is essential in promoting Mcl-1 stability. However, this might only be true in cells that are dependent on the GSK3β pathway. A study by Nifoussi et al. concluded that phospho-Thr163 resulted in increased Mcl-1 protein expression in cells that are not dependent on the GSK3β pathway, although no ubiquitin ligases were found to be

Reference 22 25 29 32 40 41 44 47-50 56, 57 58 59 60 62 63 66

involved.23 Hence, it is logical to hypothesize that Fbw7 was not involved as these cell lines are not dependent on GSK3β. Instead of GSK3β, ERK was the kinase reported to be responsible for Thr163 phosphorylation of Mcl-1, with such phosphorylation being upregulated upon 12-O-tetradecanoylphorbol 13-acetate-induced ERK activation in cells23 and downregulated in cells treated with the ERK inhibitor, PD98059.24 Taken together, these findings suggest that GSK3β-mediated phosphorylation of Mcl-1 at both Ser159 and Thr163 is a target of Fbw7 for ubiquitination and degradation, while ERK-induced Thr163 phosphorylation of Mcl-1 enhances Mcl-1 stability. Nonetheless, more studies should be carried out to determine whether these events are independent or whether they interact with one another. Another E3 ubiquitin ligase involved in Mcl-1 ubiquitination is Mcl-1 ubiquitin ligase E3 (MULE). MULE contains a wellconserved BH3 domain, which bind specifically to Mcl-1 but not to other Bcl-2 family protein members.25 A synthetic MULE mutant (mutation in the MULE BH3 domain) prevented binding to Mcl-1 and thus downregulated Mcl-1 ubiquitination and its subsequent degradation.25 Additionally, reduction of MULE expression through RNA interference stabilized Mcl-1 protein while impairing tetracycline-induced apoptosis.25 Etoposide-induced degradation of Mcl-1 was abrogated by MULE deficiency, although basal Mcl-1 expression level was not affected by the lack of MULE in these cells.25 These results suggest that there may be other E3 ubiquitin ligases concomitantly involved in the regulation of Mcl-1 stability. A report by Gomez-Bougie et al. described that NOXA enhances the MULE/Mcl-1 interaction, while promoting ubiquitination of

Haematologica | 109 January 2024

REVIEW ARTICLE - Ubiquitination of Bcl-2 family proteins

G. Le Qian Tang et al.

Figure 1. Schematic diagram of Bcl-2 ubiquitin ligases and deubiquitinases. Schematic diagram of the network of ubiquitin ligases and deubiquitinases that are involved in regulating the expression of Bcl-2 proteins (Mcl-1, Bcl-2, Bax and BIM). Solid lines indicate that a direct interaction was reported whereas dotted arrow lines indicate an association with the protein of interest (i.e., no direct binding was observed). Figure created using Biorender.com

Mcl-1, although no mechanism of action was described. Overexpression of NOXA was also observed to displace USP9x, a DUB that targets Mcl-1 for deubiquitination.26,27 This is supported by the observation that siRNA-mediated knockdown of NOXA promotes USP9x availability to Mcl-1, leading to a lower rate of apoptosis.28 Another study showed that phosphorylated Mcl-1 at several sites displaces USP9x.29 Hence, it is possible that MULE- and USP9x-binding to Mcl-1 are phosphorylation-dependent, such that NOXA-mediated phosphorylation of Mcl-1 promotes the interaction with MULE, while suppressing USP9x binding.18 Deubiquitination is the reverse process of ubiquitination and is catalyzed by specific DUB. USP9x is one of the many DUB that deubiquitinate Mcl-1 to prolong its half-life. USP9x expression correlates positively with Mcl-1 in human follicular lymphoma and diffuse large B-cell lymphoma. In multiple myeloma patients, high expression of USP9x is associated with poor prognosis.29 Hence the interaction between USP9x and Mcl-1 has prognostic relevance in these human malignancies. Like many DUB, USP9x stabilizes Mcl1 through cleaving its K48-linked polyubiquitin chain that normally promotes degradation. Knockdown of USP9x increased Mcl-1 turnover but did not alter Mcl-1 mRNA levels, suggesting that the USP9x modifies Mcl-1 post-translationally. The C1566 mutant (lacking DUB activity) of USP9x also failed to stabilize Mcl-1, which further suggests that Mcl-1 stability is dependent on USP9x DUB activity.29 Similarly, the decrease in the Mcl-1/USP9x interaction through knockdown of USP9x29 or WP1130 treatment,30 a USP9x inhibitor, sensitized cells to ABT737-induced cell death. Lupus Ku autoantigen p70 (Ku70) is a protein that has dual

functions as a tumor suppressor. Ku70 is known as a DNA double-strand break repair protein that is important for non-homologous end joining recombination,31 thus preventing apoptosis. At the same time, it serves as a Mcl-1specific DUB, while having no effect on other Bcl-2 anti-apoptotic family proteins.32 Structurally, Ku70 interacts with Mcl-1 through its C terminus, while carrying out its DSB repair via its N terminus.32 With regards to the role of Ku70 in stabilizing Mcl-1, deletion of Ku70 caused an accumulation of K48-linked polyubiquitin chains on Mcl-1, together with a decrease in Mcl-1 expression and half-life. Overexpression of Ku70 in Ku70-/- cells rescued Mcl-1 expression.32 As for USP9x, knockdown of Ku70 was reported to augment the effect of ABT737 on apoptosis,32 and in a separate study, caused accumulated DNA damage and triggered the p53-induced apoptotic pathway.33 Considering the role Ku70 plays in Mcl-1 stability and the DNA damage response, Ku70 inhibition appears to be a promising approach for therapy. However, the feasibility of inhibition of Ku70 as a therapeutic strategy remains to be determined as Ku70 fulfills many other cellular processes. Bcl-2 (B-cell lymphoma-2) Bcl-2, an anti-apoptotic protein, is highly expressed in tumor samples in various cancers, in which its expression correlates positively with poor prognosis.34 Similarly to Mcl1, Bcl-2 is reported to bind to Bak and prevent Bak and Bax allosteric activation.35 Bcl-2 also interacts with BIM, thereby preventing Bcl-2 from exerting its anti-apoptotic function.36 Interestingly, Bcl-2 has a relatively longer halflife as compared to Mcl-1, possibly due to fewer ubiquitin ligases regulating Bcl-2.37,38 Hence, unlike Mcl-1, there is limited evidence of Bcl-2 ubiquitination, with contrasting

Haematologica | 109 January 2024

REVIEW ARTICLE - Ubiquitination of Bcl-2 family proteins

G. Le Qian Tang et al.

findings. It was first thought that stabilization of Bcl-2 was not dependent on the ubiquitin proteasome system, as treatment with the proteasome inhibitor, MG132, did not alter Bcl-2 levels.39 However, some studies showed otherwise that ubiquitination of Bcl-2 has both stabilizing and destabilizing effects. For example, PPARα induces the transfer of polyubiquitin to Bcl-2 and causes its degradation,40 while Parkin promotes mono-ubiquitination on Bcl2 which stabilizes Bcl-2.41 With regards to the former, PPARα was described to play an E3 ubiquitin ligase role by binding directly to Bcl-2 and transferring K48-linked polyubiquitin chains to the K22 site on Bcl-2, governing its stability.40 Short hairpin (shRNA) silencing of PPARα resulted in the inhibition of Bcl-2 ubiquitination, while overexpression of PPARα augmented chemotherapeutic effects through decreased Bcl-2 levels.40 Hence, it is evident that PPARα negatively regulates Bcl-2 protein stability. On the other hand, Parkin, an E3 ubiquitin ligase known to destabilize Mcl-1,42,43 was reported to mono-ubiquitinate Bcl-2. Interestingly, such mono-ubiquitination was associated with Bcl-2 stability and increases in its steady-state levels.41 This increase in Bcl-2 mediated by Parkin inhibits autophagy, as the conversion of LC3-II (a marker for autophagy) was reduced under both normal and starvation conditions.41 Furthermore, it was found that mutation in the RING1 domain, but not the RING2 domain, of Parkin impairs mono-ubiquitination on Bcl-2, suggesting that the RING1 domain is important for the Parkin and Bcl-2 interaction.41 This suggests a potential combinatorial therapy targeting the RING1 domain of Parkin to destabilize Bcl-2, alongside Bcl-2 inhibitors, such as ABT737, to inhibit autophagy and promote apoptosis. However, as mentioned above, Parkin also sensitizes cells to apoptosis by ubiquitinating and destabilizing Mcl-1,42,43 indicating that Parkin can both stabilize and destabilize anti-apoptotic proteins under certain conditions. Hence, targeting Parkin for cancer therapy might not be as straightforward as it seems. Like Parkin, XIAP is another E3 ubiquitin ligase. Its expression was observed to correlate positively with that of Bcl-2. In cells with XIAP knockdown, Bcl-2, and Bcl-xL expression decrease significantly, with an increase in efflux of cytochrome c from mitochondria into cytoplasm to trigger apoptosis.44 Interestingly, Bax was not affected by such knockdown, further implying that XIAP possibly mediates apoptosis through Bcl-2 and Bcl-xL.44 Another study delved deeper by performing a drug screen, followed by in vivo validations. It was concluded that XIAP and Bcl-2 are among the top five targets for combinatorial therapy in acute myeloid leukemia patients, and that inhibition of both Bcl-2 and XIAP eradicated established acute myeloid leukemia in in vivo patient-derived xenografts.45 It would be intriguing to establish the mechanism behind the interaction between these two anti-apoptotic proteins, Bcl-2 and XIAP. Surprisingly, Edison et al. described a mechanism

through which XIAP and Bcl-2 do interact, in a complex formed of ARTS, Bcl-2, and XIAP.46 ARTS is a pro-apoptotic tumor suppressor, contrary to XIAP and Bcl-2. Edison et al. demonstrated that ARTS achieves its tumor suppressive role by binding to Bcl-2 through its BH3 domain, and recruits XIAP to ubiquitinate Bcl-2. This ubiquitination event at K17 of Bcl-2 promoted the destabilization of Bcl-2.46 Consistently, the K17A mutant increased Bcl-2 stability and was able to protect cells from apoptosis.46 This begs the questions of whether XIAP is a bona fide ubiquitin ligase for Bcl-2, or whether the binding of ARTS alters the conformation of either Bcl-2 or XIAP, causing an increase in affinity towards each other for Bcl-2 ubiquitination and destabilization. In either case, answering these questions will improve our understanding and facilitate the development of combinatorial therapy for cancer. Canonically, DUB bind to and promote deubiquitination of their substrate proteins, ultimately promoting their stabilization. Such interactions require direct binding between the proteins. However, reported research on DUB regulating Bcl-2 expression did not investigate specific DUB and/or their interactions, either direct or indirect. The downregulation of USP1, USP8 and USP22 through their respective siRNA decreased Bcl-2 protein, as shown by western blot analysis.47-50 Kuo et al. demonstrated that the non-specific DUB inhibitor, PR619, could downregulate Bcl2 expression to enhance anti-tumor effects and augment cisplatin cytotoxicity.51 Furthermore, Xu et al. observed that transfection of siUSP1 decreased Bcl-2 mRNA levels as well, which might explain the results observed in the western blot analysis.52 Hence, future research should include ubiquitination and/or co-immunoprecipitation assays to investigate the interaction between DUB and Bcl-2. BIM (Bcl-2 interacting mediator of cell death) Besides negating the anti-apoptotic effects of anti-apoptotic Bcl-2 family proteins, BIM also activates Bax by facilitating Bax oligomerization.36 Knockdown of BIM was observed to attenuate the ability of Bax to trigger MOMP, suggesting that BIM plays an ‘activator’ role in triggering apoptosis.36 Interestingly, the BIM and Mcl-1 interaction provides reciprocal protection and potentially balances out their apoptotic effects.19,53 In other words, BIM binding to Mcl-1 prevents the recruitment of ubiquitin ligases to both of these proteins. Disruption of such binding through a mutation in the BH3 domain of BIM was reported to cause an increase in the turnover rates of both proteins.54 Interestingly, ERK1/2-dependent phosphorylation of BIM at Ser65 (for mice)54 and Ser69 (for humans)55,56 also disrupted the BIM-Mcl-1 interaction and promoted the dissociation of BIM. Therefore, BIM becomes a target for ubiquitination and its subsequent destabilization. BIM phosphorylation at Ser69, regulated by ERK1/2, is important for its subsequent phosphorylation at Ser93, Ser94,

Haematologica | 109 January 2024

REVIEW ARTICLE - Ubiquitination of Bcl-2 family proteins

G. Le Qian Tang et al.

and Ser98 by kinases such as Rsk1/2 and Aurora A.56,57 Inhibition of Rsk or Aurora A reduced phosphorylation of BIM on Ser93, Ser94 and Ser98. These phosphorylation sites are negatively regulated by phosphatases such as protein phosphatase 2A (PP2A).57 Such phosphorylation events are important for the binding of BIM to beta-transducin repeat containing E3 ubiquitin protein ligase (βTrCP) for βTrCPmediated degradation of BIM.56,57 An inhibitor of MEK, UO126, strongly reversed phorbol 12-myristate 13-acetate (PMA)-induced ERK activation and reduced binding of BIM to βTrCP.56 Phosphorylation at Ser93, Ser94, and Ser98 was consequently reduced as well. Consistently, an alanine mutation at these sites drastically reduced binding to βTrCP,57 while siβTrCP,56 mutant βTrCP (delta Fbox),56 and MLN8054, an Aurora A inhibitor,57 failed to induce in vitro ubiquitination on BIM, resulting in BIM stabilization. Taken together, the investigations showed that ERK1/2-mediated phosphorylation of BIM at Ser69 creates a binding site for Rsk1/2 and Aurora A, which in turn phosphorylate BIM at Ser93, Ser94, and Ser98 to allow βTrCP binding and ubiquitination of BIM. Besides βTrCP, APC/Cdc20 has also been linked to the ubiquitination of BIM. Wan et al. revealed that the E3 ubiquitin ligase responsible is cell-cycle dependent, with βTrCP regulating BIM during the S phase, and APC/Cdc20 regulating BIM during the M phase.58 Consistently, the level of BIM was reduced in the late S phase, even when APC/Cdc20 activity was low, and such downregulation of BIM expression could be reversed with βTrCP depletion. In contrast, BIM expression was upregulated upon depletion or knockdown of Cdc20. This regulation was specific to BIM and no other BH3-only proteins.58 Since βTrCP and APC/Cdc20 can directly regulate BIM ubiquitination, combinatorial therapies might emerge from these studies to sensitize tumour cells to apoptotic cell death. In opposition to ubiquitin ligases, DUB such as USP27x, stabilize BIM by removing ubiquitin added by ubiquitin ligases, specifically K48- and K63-linked ubiquitin chains.59 Overexpression of USP27x reduced ERK-dependent BIM phosphorylation and ubiquitination in PMA-stimulated cells as well as tumor cells with a constitutively active Raf/ERK pathway. Binding of USP27x to BIM was dependent on the 42-101 amino acid region on BIM, as mutation of these sites on BIM disrupted binding to USP27x.59 Interestingly, binding of USP27x and βTrCP to BIM is independent, as knockdown of βTrCP did not reduce the association of BIM with USP27x. Concomitantly, phosphorylation of BIM at Ser69 is required for USP27x binding,59 which is similar to βTrCP although some further research established the need for Ser93, Ser94 and Ser98 on BIM to be phosphorylated as well.56,57 Nevertheless, it is possible that both USP27x and βTrCP are simultaneously recruited by BIM upon its phosphorylation to regulate its stability.

Bax (BCL2-associated X, apoptosis regulator) Bax, a member of the pro-apoptotic Bcl-2 family of proteins, oligomerizes and promotes cytochrome c release from mitochondria, resulting in apoptosis. Degradation of Bax prevents cytochrome c release and is widely observed in cancer cells.39 Inhibition of proteasome through MG132 treatment increased the accumulation of ubiquitinated Bax without altering its mRNA levels.39 Moreover, inhibition of proteasome in Bcl-2-overexpressing cells led to accumulation of Bax thus allowing Bax to inhibit Bcl-2 activity.39 Additionally, Peng et al. demonstrated that K21 and K123 of Bax are required for ubiquitin binding on Bax, where K21R and K123R mutations have reduced ubiquitinated Bax levels with prolonged half-lives.60 Interestingly, Bax ubiquitination was observed to activate Bax as well.60,61 Previous studies demonstrated that the TNF-related apoptosis-inducing ligand (TRAIL) cytokine can activate Bax while promoting its degradation in malignant B cells.61 Consistently, K21R and K123R mutants decreased ubiquitination of Bax, with decreased pro-apoptotic activity as compared to wild-type Bax upon TRAIL treatment.60 This is concordant with the hypothesis that ubiquitination is required for Bax activation, given that overexpression of USP30, a DUB that binds directly to Bax, was observed to negatively regulate Bax-mediated apoptosis.62 Overexpression of USP30 alleviated Poly ADP-ribose polymerase (PARP) cleavage and apoptosis, whereas siRNA depletion of USP30 increased PARP cleavage and sensitized cells to drug-induced apoptosis.62 Treatment with an USP30 inhibitor, aumdubin, led to higher levels of ubiquitinated Bax, and Bax-mediated apoptosis, possibly due to ubiquitination on Bax, enhances its mitochondrial localization and activation.62 Collectively, these findings suggest that Bax activation is dependent on ubiquitination of the protein, and can be targeted for degradation after activation. Parkin is another E3 ubiquitin ligase reported to interact with Bax. It was shown to destabilize and inhibit Bax activity by binding directly to the BH3 domain on Bax.63 Binding of Parkin to Bax promoted ubiquitination and subsequent degradation on Bax. Parkin-deficient cells had lower ubiquitinated Bax as compared to wild-type cells, consistent with the notion that Parkin is the ubiquitin ligase responsible for Bax ubiquitination.64 Moreover, activated Parkin promoted the dissociation of Bax from mitochondria.63,65 Interestingly, K21 and K64 of Bax are critical for Parkin-dependent regulation of Bax activity, and K21R and K64R mutants increased Bax mitochondrial translocation without affecting the Parkin-Bax physical interaction.63 Double mutant K21R/K63R had the highest rate of mitochondrial translocation as compared to individual K21R and K63R mutants and wild-type Bax, suggesting that these mutations are additive.63 Nevertheless, this suggests that K21 and K64 on Bax are important for Parkin-dependent translocation of Bax away from mitochondria. On the other hand, Parkinson disease-linked Parkin mutations R275W and W453X both re-

Haematologica | 109 January 2024

REVIEW ARTICLE - Ubiquitination of Bcl-2 family proteins

G. Le Qian Tang et al.

sulted in loss of function of Parkin. Both mutants co-immunoprecipitated with Bax and were observed to have higher affinity with Bax than with wild-type Parkin.64 These mutants failed to inhibit Bax translocation to mitochondria, which resulted in apoptosis and cell death, one of the hallmarks of Parkinson disease.64 In addition, Parkin also negatively regulates Bax activity without ubiquitinating Bax, as suppression of Bax activity by Parkin was recorded in the absence of Bax ubiquitination.65 Mechanistically, it was found that Parkin promoted the destabilization of VDAC2, a known Bax activator, potentially altering the conformation of VDAC2 and preventing Bax activation. These reports suggest a dual role of Parkin in regulating Bax activity by both (i) directly ubiquitinating and destabilizing Bax, and (ii) promoting the degradation of Bax activator. IBR-type RING-finger E3 ubiquitin ligase, IBRDC2, is another E3 ubiquitin ligase that was shown to protect cells from unregulated Bax activation and cell death. A direct interaction between IBRDC2 and Bax was documented by Benard and colleagues and was described to destabilize Bax, but not other Bcl-2 family proteins including Bak, through ubiquitination.66 Downregulation of IBRDC2 through shRNA induced accumulation of active Bax, while IBRDC2 overexpression stimulated Bax ubiquitination. Interestingly, IBRDC2, usually localized in the cytosol, accumulated in mitochondria alongside activated Bax upon induction of apoptosis by staurosporine and actinomycin D, suggesting that IBRDC2 localization to Bax on mitochondria is dependent on Bax activation.66 Bcl-xL, which suppresses Bax activation,67 was also shown to inhibit mitochondrial accumulation of IBRDC2, supporting such a hypothesis.66 Taken together, loss of IBRDC2 failed to promote Bax ubiquitination and ultimately sensitizes cells to Bax-mediated MOMP and cell death.

teins is important in regulating their stability and function. As a result, these ubiquitinated proteins are often targeted by the ubiquitin proteasome system for their degradation. Hence, it is not surprising that researchers have investigated inhibiting the ubiquitin proteasome system to prevent the degradation of tumor suppressors as an anti-cancer strategy (summarized in Table 3). Bortezomib, a reversible inhibitor of 26S proteasome, is the first proteasome inhibitor approved by the Food and Drug Administration to treat multiple myeloma patients.68 Mechanistically, bortezomib was shown to inhibit the NFκB signaling pathway by preventing the degradation of IκB, a negative regulator of NFκB.69 Normally, upon activation of the NFκB signaling pathway, IκB is phosphorylated by IKK and such phosphorylation primes IκB for its ubiquitination and subsequent degradation. IκB dissociates from NFκB complexes, releasing its inhibitory effect. Hence, downstream target genes of NFκB, such as the anti-apoptotic genes XIAP, Bcl-2, and Bcl-xL, can be transcribed.70-72 As a result, upon bortezomib treatment, IκB is thought to be protected from the ubiquitin proteasome system, allowing IκB to exert its inhibitory effect on NFκB and attenuate the aforementioned anti-apoptotic gene expression. Another possible mechanism to promote apoptosis employed by bortezomib is through Myc-dependent upregulation of NOXA gene expression.73 Such upregulation mediates the downregulation of the anti-apoptotic activity of Bcl-2 and Bcl-xL proteins through direct binding of NOXA. Consistently, RNA interference of Myc blocks NOXA induction by bortezomib in tumor cells.73 However, because of the effect of bortezomib on general proteasome inhibition, researchers have also investigated targeting the upstream pathways, such as DUB and ubiquitin ligases. In theory, this direct method might produce fewer side effects and toxicities because of it higher specificity for its target protein, such as inhibiting DUB that stabilize oncogenes or ubiquitin ligases that destabilize tumor suppressors. Recruiting oncogenes for ubiquitination through proteolysis-targeting chimeras (PROTAC) is another viable option, and examples are discussed below.

Targeting Bcl-2 family proteins through regulating ubiquitination Ubiquitination of tumour suppressors and proto-oncoproTable 3. Inhibitors targeting the ubiquitin proteasome system.

Drug

Summary

Reference

Proteasome inhibitor targeting the ubiquitin proteasome system.

68, 83 74, 75

WP1130

USP1-specific inhibitors. USP1 inhibitor currently in a phase I study, proposed to eliminate resistance tp PARP inhibitors. Non-selective inhibitor of USP5, USP9x, USP14, USP24 and UCH37.

DT2216

A dual-inhibitor PROTAC targeting Bcl-xL and Bcl-2.

753b

Improved version of DT2216.

XZ424

PROTAC specifically targeting Bcl-xL for ubiquitination and degradation.

Bortezomib Pimozide, GW7647, ML323 KSQ4279

76 77, 78

USP: ubiquitin-specific peptidase; PARP: poly ADP-ribose polymerase; UCH: ubiquitin C-terminal hydrolase; PROTAC: proteolysis-targeting chimera. Haematologica | 109 January 2024

REVIEW ARTICLE - Ubiquitination of Bcl-2 family proteins

G. Le Qian Tang et al.

With regard to DUB inhibitors, pimozide, GW7647 and ML323 are USP1 inhibitors that selectively bind to USP1.74,75 Pimozide and GW7647 were identified by Chen and colleagues to be USP1/UAF1 inhibitors, with both inhibitors binding reversibly to the allosteric site instead of its active site.75 Consistently, treatment of these inhibitors caused the accumulation of ubiquitinated substrates downstream of USP1, such as PCNA and FANCD2, thus regulating the DNA damage response. In addition, the accumulation of ubiquitinated substrates sensitized cells to cisplatin treatment.75 However, these USP1 inhibitors can also bind to and inhibit USP7, albeit with approximately 24-fold higher IC50 for USP7.75 Hence, the use of these inhibitors are limited by offtarget specificity. These limitations led to the development of ML323, which is highly selective towards USP1 and does not have any activity against other ubiquitin-specific peptidases.74 ML323 also augmented cisplatin sensitivity in tumor samples and was shown to increase endogenous mono-ubiquitination of PCNA and FANDC2, similarly to pimozide and GW7647.74 Recently, another highly selective inhibitor of USP1, KSQ4279 was discovered by KSQ Therapeutics.76 KSQ4279 is currently being tested in a phase I clinical trial as a monotherapy or in combination with an oral PARP inhibitor in patients with advanced solid tumors. This small molecule inhibitor (i.e., KSQ4279) serves to eliminate resistance to PARP inhibitors in cases in which resistance is common in tumors with homologous recombination repair deficiencies, such as BRCA1. Unfortunately, not much research has been performed on the effect of USP1 inhibitors on other USP1 target genes such as Bcl-2 family proteins. Since USP1 may potentially regulate Bcl-2, it would be interesting to investigate whether Bcl-2 plays a role in regulating apoptosis induced by the USP1 inhibitors. WP1130, a cell-permeable DUB inhibitor, is known to upregulate pro-apoptotic proteins while suppressing antiapoptotic proteins, such as p53 and Mcl-1. However, WP1130 is not selective towards USP9x but can also inhibit USP5, USP14, USP24 and UCH37.77,78 These DUB are known to regulate survival pathways and 26S proteasome function. Upon WP1130 treatment, p53 and Mcl-1 were observed to be up- and down-regulated respectively.77 Interestingly, it was revealed that the upregulation of p53 was dependent on USP5 inhibition, while the downregulation of Mcl-1 was dependent on USP9x inhibition.77 However, the downregulation of Mcl-1 was also concluded to be dependent on inhibition of USP24 in different cell lines.78 These contradictory results suggest that the effect of WP1130 on Mcl-1 might not be solely dependent on one DUB. At the same time, the exact mechanism by which WP1130 achieves its inhibitory effect on these DUB is still unknown. It is possible that WP1130 targets the electrophilic cysteine residue on the active site and does not recognize specific conformations of the various DUB, which may explain the ability of WP1130 to target multiple DUB.

More progress has been made recently in the development of targeted anti-cancer therapies. PROTAC have emerged as adapter proteins to bring two proteins together. PROTAC have two binding regions, with one end holding/inhibiting the disease-causing protein, also known as the protein of interest, and the other end recruiting the E3 ubiquitin ligase.79 These two binding regions are joined together by a linker peptide. The purpose of PROTAC is to target the protein of interest for degradation instead of solely inhibiting it, unlike many small molecule inhibitors. This allows PROTAC to overcome resistance caused by mutation of the substrates, which reduces the effectiveness of small molecule inhibitors. In the context of Bcl-2 family proteins, one example of an effective PROTAC is DT2216. DT2216 consists of ABT263, a dual inhibitor of Bcl-xL and Bcl-2, a linker peptide, and a von Hippel-Lindau (VHL) ligand.80 The use of ABT263 allows the recruitment of both Bcl-xL and Bcl-2 to DT2216 and their inactivation. At the same time, VHL, an E3 ubiquitin ligase, is recruited to DT2216 through its ligand. The Bcl-xL and Bcl-2 that are recruited to DT2216 are targeted for ubiquitination by VHL and are degraded by the ubiquitin proteasome system.80 DT2216 was documented to bind to both Bcl-xL and Bcl-2 with strong affinities. However, only Bcl-xL was potently targeted for degradation, possibly because of the lack of accessible lysine residues on Bcl-2 upon binding to DT2216.80 Therefore, 753b was developed to ensure that both Bcl-2 and Bcl-xL can be targeted for degradation.81 The only difference between DT2216 and 753b is the linker site on ABT263. This alteration allows both proteins of interest to be recruited for ubiquitination by VHL and subsequently degraded by the proteasome.81 It was concluded that 753b was more effective in inducing apoptosis in acute myeloid leukemia cells than DT2216 while resulting in low toxicity similar to DT2216. This low toxicity of DT2216 and 753b is due to the low level of VHL ligase present in human platelets, which limits the degradation of the protein of interest in these cells.80,81 In addition to DT2216 and 753b, XZ424 is another PROTAC that is intended to target Bcl-xL for ubiquitination. XZ424 utilizes A1155463, another Bcl-xL inhibitor with the Cereblon E3 ubiquitin ligase ligand, pomalidomide.82 The introduction of PROTAC as anti-cancer therapeutics is relatively recent, but these drugs have incredible potential.

Concluding remarks Ubiquitination plays an essential role in maintaining cellular homeostasis and deregulation of this cellular process can lead to cancer progression. Notably, deregulation of the Bcl2 family of proteins is associated with cancer progression. In line with this, while not all Bcl-2 proteins are susceptible to ubiquitin-mediated degradation, this review has highlighted the available data on ubiquitin-mediated degrada-

Haematologica | 109 January 2024

REVIEW ARTICLE - Ubiquitination of Bcl-2 family proteins

G. Le Qian Tang et al.

tion of Bcl-2 proteins such as Mcl-1, Bcl-2, Bim and Bax. Advancements in understanding the regulation of these proteins would certainly lead to improved therapeutic options, but there are still many nuances in the regulation of these proteins which require further investigations.

the manuscript and generated the tables and figures. JXHL and SP reviewed, edited, and revised the manuscript. Acknowledgments The authors would like to apologize to the authors whose contributions could have been inadvertently left out.

Disclosures No conflicts of interest to disclose.

Funding This work was supported by grants from NMRC, Singapore (NMRC/OFIRG/0041/2017) and Yong Loo Lin School of MediContributions SP conceptualized the work. GLQT wrote the first draft of cine, National University of Singapore to SP.

References therapy. Int J Mol Sci. 2018;19(2):448. 17. Mei Y, Xie C, Xie W, Tian X, Li M, Wu M. Noxa/Mcl-1 balance regulates susceptibility of cells to camptothecin-induced apoptosis. Neoplasia. 2007;9(10):871-881. 18. Nakajima W, Hicks MA, Tanaka N, Krystal GW, Harada H. Noxa determines localization and stability of MCL-1 and consequently ABT-737 sensitivity in small cell lung cancer. Cell Death Dis. 2014;5(2):e1052. 19. Wuilleme-Toumi S, Trichet V, Gomez-Bougie P, Gratas C, Bataille R, Amiot M. Reciprocal protection of Mcl-1 and Bim from ubiquitin-proteasome degradation. Biochem Biophys Res Commun. 2007;361(4):865-869. 20. Inuzuka H, Shaik S, Onoyama I, et al. SCF(FBW7) regulates cellular apoptosis by targeting MCL1 for ubiquitylation and destruction. Nature. 2011;471(7336):104-109. 21. Morel C, Carlson SM, White FM, Davis RJ. Mcl-1 integrates the opposing actions of signaling pathways that mediate survival and apoptosis. Mol Cell Biol. 2009;29(14):3845-3852. 22. Ren H, Koo J, Guan B, et al. The E3 ubiquitin ligases beta-TrCP and FBXW7 cooperatively mediates GSK3-dependent Mcl-1 degradation induced by the Akt inhibitor API-1, resulting in apoptosis. Mol Cancer. 2013;12:146. 23. Nifoussi SK, Vrana JA, Domina AM, et al. Thr 163 phosphorylation causes Mcl-1 stabilization when degradation is independent of the adjacent GSK3-targeted phosphodegron, promoting drug resistance in cancer. PLoS One. 2012;7(10):e47060. 24. Elgendy M, Abdel-Aziz AK, Renne SL, et al. Dual modulation of MCL-1 and mTOR determines the response to sunitinib. J Clin Invest. 2017;127(1):153-168. 25. Zhong Q, Gao W, Du F, Wang X. Mule/ARF-BP1, a BH3-only E3 ubiquitin ligase, catalyzes the polyubiquitination of Mcl-1 and regulates apoptosis. Cell. 2005;121(7):1085-1095. 26. Gomez-Bougie P, Menoret E, Juin P, Dousset C, PellatDeceunynck C, Amiot M. Noxa controls Mule-dependent Mcl-1 ubiquitination through the regulation of the Mcl-1/USP9X interaction. Biochem Biophys Res Commun. 2011;413(3):460-464. 27. Czabotar PE, Lee EF, van Delft MF, et al. Structural insights into the degradation of Mcl-1 induced by BH3 domains. Proc Natl Acad Sci U S A. 2007;104(15):6217-6222. 28. Yan J, Zhong N, Liu G, et al. Usp9x- and Noxa-mediated Mcl-1 downregulation contributes to pemetrexed-induced apoptosis in human non-small-cell lung cancer cells. Cell Death Dis. 2014;5(7):e1316. 29. Schwickart M, Huang X, Lill JR, et al. Deubiquitinase USP9X stabilizes MCL1 and promotes tumour cell survival. Nature. 2010;463(7277):103-107.

1. Teimouri H, Kochugaeva MP, Kolomeisky AB. Elucidating the correlations between cancer initiation times and lifetime cancer risks. Sci Rep. 2019;9(1):18940. 2. Ramon YCS, Capdevila C, Hernandez-Losa J, et al. Cancer as an ecomolecular disease and a neoplastic consortium. Biochim Biophys Acta Rev Cancer. 2017;1868(2):484-499. 3. Gores GJ, Kaufmann SH. Selectively targeting Mcl-1 for the treatment of acute myelogenous leukemia and solid tumors. Genes Dev. 2012;26(4):305-311. 4. Adams JM, Cory S. The Bcl-2 apoptotic switch in cancer development and therapy. Oncogene. 2007;26(9):1324-1337. 5. Chong SJF, Marchi S, Petroni G, Kroemer G, Galluzzi L, Pervaiz S. Noncanonical cell fate regulation by Bcl-2 proteins. Trends Cell Biol. 2020;30(7):537-555. 6. Reed JC. Mechanisms of apoptosis. Am J Pathol. 2000;157(5):1415-1430. 7. Du C, Fang M, Li Y, Li L, Wang X. Smac, a mitochondrial protein that promotes cytochrome c-dependent caspase activation by eliminating IAP inhibition. Cell. 2000;102(1):33-42. 8. van Loo G, van Gurp M, Depuydt B, et al. The serine protease Omi/HtrA2 is released from mitochondria during apoptosis. Omi interacts with caspase-inhibitor XIAP and induces enhanced caspase activity. Cell Death Differ. 2002;9(1):20-26. 9. Winter DB, Gearhart PJ, Bohr VA. Homogeneous rate of degradation of nuclear DNA during apoptosis. Nucleic Acids Res. 1998;26(19):4422-4425. 10. Kohler C, Hakansson A, Svanborg C, Orrenius S, Zhivotovsky B. Protease activation in apoptosis induced by MAL. Exp Cell Res. 1999;249(2):260-268. 11. Lee CL, Blum JM, Kirsch DG. Role of p53 in regulating tissue response to radiation by mechanisms independent of apoptosis. Transl Cancer Res. 2013;2(5):412-421. 12. Oda E, Ohki R, Murasawa H, et al. Noxa, a BH3-only member of the Bcl-2 family and candidate mediator of p53-induced apoptosis. Science. 2000;288(5468):1053-1058. 13. Yu J, Wang Z, Kinzler KW, Vogelstein B, Zhang L. PUMA mediates the apoptotic response to p53 in colorectal cancer cells. Proc Natl Acad Sci U S A. 2003;100(4):1931-1936. 14. Willis SN, Chen L, Dewson G, et al. Proapoptotic Bak is sequestered by Mcl-1 and Bcl-xL, but not Bcl-2, until displaced by BH3-only proteins. Genes Dev. 2005;19(11):1294-1305. 15. Chipuk JE, Kuwana T, Bouchier-Hayes L, et al. Direct activation of Bax by p53 mediates mitochondrial membrane permeabilization and apoptosis. Science. 2004;303(5660):1010-1014. 16. Pfeffer CM, Singh ATK. Apoptosis: a target for anticancer

Haematologica | 109 January 2024

REVIEW ARTICLE - Ubiquitination of Bcl-2 family proteins

G. Le Qian Tang et al.

30. Peddaboina C, Jupiter D, Fletcher S, et al. The downregulation of Mcl-1 via USP9X inhibition sensitizes solid tumors to Bcl-xl inhibition. BMC Cancer. 2012;12:541. 31. Lee KJ, Saha J, Sun J, et al. Phosphorylation of Ku dictates DNA double-strand break (DSB) repair pathway choice in S phase. Nucleic Acids Res. 2016;44(4):1732-1745. 32. Wang B, Xie M, Li R, et al. Role of Ku70 in deubiquitination of Mcl-1 and suppression of apoptosis. Cell Death Differ. 2014;21(7):1160-1169. 33. Yu W, Li L, Wang G, Zhang W, Xu J, Liang A. KU70 inhibition impairs both non-homologous end joining and homologous recombination DNA damage repair through SHP-1 induced dephosphorylation of SIRT1 in T-cell acute lymphoblastic leukemia (T-ALL) [corrected]. Cell Physiol Biochem. 2018;49(6):2111-2123. 34. Roh J, Cho H, Pak HK, et al. BCL2 super-expressor diffuse large B-cell lymphoma: a distinct subgroup associated with poor prognosis. Mod Pathol. 2022;35(4):480-488. 35. Cheng EH, Wei MC, Weiler S, et al. BCL-2, BCL-X(L) sequester BH3 domain-only molecules preventing BAX- and BAK-mediated mitochondrial apoptosis. Mol Cell. 2001;8(3):705-711. 36. O'Connor L, Strasser A, O'Reilly LA, et al. Bim: a novel member of the Bcl-2 family that promotes apoptosis. EMBO J. 1998;17(2):384-395. 37. Merino R, Ding L, Veis DJ, Korsmeyer SJ, Nunez G. Developmental regulation of the Bcl-2 protein and susceptibility to cell death in B lymphocytes. EMBO J. 1994;13(3):683-691. 38. Fritsch RM, Schneider G, Saur D, Scheibel M, Schmid RM. Translational repression of MCL-1 couples stress-induced eIF2 alpha phosphorylation to mitochondrial apoptosis initiation. J Biol Chem. 2007;282(31):22551-22562. 39. Li B, Dou QP. Bax degradation by the ubiquitin/proteasomedependent pathway: involvement in tumor survival and progression. Proc Natl Acad Sci U S A. 2000;97(8):3850-3855. 40. Gao J, Liu Q, Xu Y, et al. PPARα induces cell apoptosis by destructing Bcl2. Oncotarget. 2015;6(42):44635-44642. 41. Chen D, Gao F, Li B, et al. Parkin mono-ubiquitinates Bcl-2 and regulates autophagy. J Biol Chem. 2010;285(49):38214-38223. 42. Carroll RG, Hollville E, Martin SJ. Parkin sensitizes toward apoptosis induced by mitochondrial depolarization through promoting degradation of Mcl-1. Cell Rep. 2014;9(4):1538-1553. 43. Zhang C, Lee S, Peng Y, et al. PINK1 triggers autocatalytic activation of Parkin to specify cell fate decisions. Curr Biol. 2014;24(16):1854-1865. 44. Chen C, Liu TS, Zhao SC, Yang WZ, Chen ZP, Yan Y. XIAP impairs mitochondrial function during apoptosis by regulating the Bcl-2 family in renal cell carcinoma. Exp Ther Med. 2018;15(5):4587-4593. 45. Hashimoto M, Saito Y, Nakagawa R, et al. Combined inhibition of XIAP and BCL2 drives maximal therapeutic efficacy in genetically diverse aggressive acute myeloid leukemia. Nat Cancer. 2021;2(3):340-356. 46. Edison N, Curtz Y, Paland N, et al. Degradation of Bcl-2 by XIAP and ARTS promotes apoptosis. Cell Rep. 2017;21(2):442-454. 47. Liu J, Zhu H, Zhong N, et al. Gene silencing of USP1 by lentivirus effectively inhibits proliferation and invasion of human osteosarcoma cells. Int J Oncol. 2016;49(6):2549-2557. 48. Jing X, Chen Y, Chen Y, et al. Down-regulation of USP8 inhibits cholangiocarcinoma cell proliferation and invasion. Cancer Manag Res. 2020;12:2185-2194. 49. Rong Z, Zhu Z, Cai S, Zhang B. Knockdown of USP8 inhibits the growth of lung cancer cells. Cancer Manag Res. 2020;12:12415-12422. 50. Zhang J, Luo N, Tian Y, et al. USP22 knockdown enhanced

chemosensitivity of hepatocellular carcinoma cells to 5-Fu by up-regulation of Smad4 and suppression of Akt. Oncotarget. 2017;8(15):24728-24740. 51. Kuo KL, Liu SH, Lin WC, et al. The deubiquitinating enzyme inhibitor PR-619 enhances the cytotoxicity of cisplatin via the suppression of anti-apoptotic Bcl-2 protein: in vitro and in vivo study. Cells. 2019;8(10):1268. 52. Xu X, Li S, Cui X, et al. Inhibition of ubiquitin specific protease 1 sensitizes colorectal cancer cells to DNA-damaging chemotherapeutics. Front Oncol. 2019;9:1406. 53. Warr MR, Acoca S, Liu Z, et al. BH3-ligand regulates access of MCL-1 to its E3 ligase. FEBS Lett. 2005;579(25):5603-5608. 54. Ewings KE, Hadfield-Moorhouse K, Wiggins CM, et al. ERK1/2dependent phosphorylation of BimEL promotes its rapid dissociation from Mcl-1 and Bcl-xL. EMBO J. 2007;26(12):2856-2867. 55. Chen K, Tu Y, Zhang Y, Blair HC, Zhang L, Wu C. PINCH-1 regulates the ERK-Bim pathway and contributes to apoptosis resistance in cancer cells. J Biol Chem. 2008;283(5):2508-2517. 56. Dehan E, Bassermann F, Guardavaccaro D, et al. betaTrCP- and Rsk1/2-mediated degradation of BimEL inhibits apoptosis. Mol Cell. 2009;33(1):109-116. 57. Moustafa-Kamal M, Gamache I, Lu Y, Li S, Teodoro JG. BimEL is phosphorylated at mitosis by Aurora A and targeted for degradation by betaTrCP1. Cell Death Differ. 2013;20(10):1393-1403. 58. Wan L, Tan M, Yang J, et al. APC(Cdc20) suppresses apoptosis through targeting Bim for ubiquitination and destruction. Dev Cell. 2014;29(4):377-391. 59. Weber A, Heinlein M, Dengjel J, Alber C, Singh PK, Hacker G. The deubiquitinase Usp27x stabilizes the BH3-only protein Bim and enhances apoptosis. EMBO Rep. 2016;17(5):724-738. 60. Peng R, Zhu J, Deng S, et al. Targeting BAX ubiquitin-binding sites reveals that BAX activation is essential for its ubiquitindependent degradation. J Cell Biochem. 2020;121(4):2802-2810. 61. Liu FT, Agrawal SG, Gribben JG, et al. Bortezomib blocks Bax degradation in malignant B cells during treatment with TRAIL. Blood. 2008;111(5):2797-2805. 62. Yan D, Li X, Yang Q, et al. Regulation of Bax-dependent apoptosis by mitochondrial deubiquitinase USP30. Cell Death Discov. 2021;7(1):211. 63. Charan RA, Johnson BN, Zaganelli S, Nardozzi JD, LaVoie MJ. Inhibition of apoptotic Bax translocation to the mitochondria is a central function of parkin. Cell Death Dis. 2014;5(7):e1313. 64. Johnson BN, Berger AK, Cortese GP, Lavoie MJ. The ubiquitin E3 ligase parkin regulates the proapoptotic function of Bax. Proc Natl Acad Sci U S A. 2012;109(16):6283-6288. 65. Bernardini JP, Brouwer JM, Tan IK, et al. Parkin inhibits BAK and BAX apoptotic function by distinct mechanisms during mitophagy. EMBO J. 2019;38(2):e99916. 66. Benard G, Neutzner A, Peng G, et al. IBRDC2, an IBR-type E3 ubiquitin ligase, is a regulatory factor for Bax and apoptosis activation. EMBO J. 2010;29(8):1458-1471. 67. Edlich F, Banerjee S, Suzuki M, et al. Bcl-x(L) retrotranslocates Bax from the mitochondria into the cytosol. Cell. 2011;145(1):104-116. 68. Kane RC, Bross PF, Farrell AT, Pazdur R. Velcade: U.S. FDA approval for the treatment of multiple myeloma progressing on prior therapy. Oncologist. 2003;8(6):508-513. 69. Sung MH, Bagain L, Chen Z, et al. Dynamic effect of bortezomib on nuclear factor-kappaB activity and gene expression in tumor cells. Mol Pharmacol. 2008;74(5):1215-1222. 70. Baud V, Karin M. Is NF-kappaB a good target for cancer therapy? Hopes and pitfalls. Nat Rev Drug Discov. 2009;8(1):33-40.

Haematologica | 109 January 2024

REVIEW ARTICLE - Ubiquitination of Bcl-2 family proteins

G. Le Qian Tang et al.

71. Khoshnan A, Tindell C, Laux I, Bae D, Bennett B, Nel AE. The NFkappa B cascade is important in Bcl-xL expression and for the anti-apoptotic effects of the CD28 receptor in primary human CD4+ lymphocytes. J Immunol. 2000;165(4):1743-1754. 72. Catz SD, Johnson JL. Transcriptional regulation of bcl-2 by nuclear factor kappa B and its significance in prostate cancer. Oncogene. 2001;20(50):7342-7351. 73. Qin JZ, Ziffra J, Stennett L, et al. Proteasome inhibitors trigger NOXA-mediated apoptosis in melanoma and myeloma cells. Cancer Res. 2005;65(14):6282-6293. 74. Dexheimer TS, Rosenthal AS, Liang Q, et al. Discovery of ML323 as a novel inhibitor of the USP1/UAF1 deubiquitinase complex. In: Probe Reports from the NIH Molecular Libraries Program [Internet]. Bethesda (MD): National Center for Biotechnology Information (US); 2010. 2012 Oct 23 [updated 2014 Sep 18]. 75. Chen J, Dexheimer TS, Ai Y, et al. Selective and cell-active inhibitors of the USP1/ UAF1 deubiquitinase complex reverse cisplatin resistance in non-small cell lung cancer cells. Chem Biol. 2011;18(11):1390-1400. 76. Cadzow L, Gokhale PC, Ganapathy S, et al. KSQ-4279, a first-inclass USP1 inhibitor shows strong combination activity in BRCA mutant cancers with intrinsic or acquired resistance to PARP inhibitors. Eur J Cancer. 2022;174(Suppl 1):S37-S38.

77. Kapuria V, Peterson LF, Fang D, Bornmann WG, Talpaz M, Donato NJ. Deubiquitinase inhibition by small-molecule WP1130 triggers aggresome formation and tumor cell apoptosis. Cancer Res. 2010;70(22):9265-9276. 78. Luo H, Jing B, Xia Y, et al. WP1130 reveals USP24 as a novel target in T-cell acute lymphoblastic leukemia. Cancer Cell Int. 2019;19:56. 79. Bekes M, Langley DR, Crews CM. PROTAC targeted protein degraders: the past is prologue. Nat Rev Drug Discov. 2022;21(3):181-200. 80. Khan S, Zhang X, Lv D, et al. A selective BCL-XL PROTAC degrader achieves safe and potent antitumor activity. Nat Med. 2019;25(12):1938-1947. 81. Lv D, Pal P, Liu X, et al. Development of a BCL-xL and BCL-2 dual degrader with improved anti-leukemic activity. Nat Commun. 2021;12(1):6896. 82. Zhang X, Thummuri D, He Y, et al. Utilizing PROTAC technology to address the on-target platelet toxicity associated with inhibition of BCL-XL. Chem Commun (Camb). 2019;55(98):14765-14768. 83. Kane RC, Dagher R, Farrell A, et al. Bortezomib for the treatment of mantle cell lymphoma. Clin Cancer Res. 2007;13(18 Pt 1):5291-5294.

Haematologica | 109 January 2024

SPOTLIGHT REVIEW ARTICLE

Tagraxofusp for blastic plasmacytoid dendritic cell neoplasm Marlise R. Luskin and Andrew A. Lane

Correspondence: M. R. Luskin

BPDCN Center, Department of Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA, USA

marlise_luskin@dfci.harvard.edu Andrew A. Lane andrew_lane@dfci.harvard.edu Received: Accepted: Early view:

February 7, 2023. March 15, 2023. March 23, 2023.

Abstract Blastic plasmacytoid dendritic cell neoplasm (BPDCN) is a rare hematologic malignancy that presents with characteristic dark purple skin papules, plaques, and tumors, but may also involve the bone marrow, blood, lymph nodes, and central nervous system. The disease, which commonly affects older men but can also present in children, is associated with a distinct immunophenotype including universal expression of CD123, the α chain of the interleukin 3 receptor. Recently, tagraxofusp, a CD123-targeting drug consisting of the ligand for CD123, interleukin 3, conjugated to a truncated diphtheria toxin payload was approved for treatment of BPDCN. This was the first agent specifically approved for BPDCN and the first CD123 targeted agent in oncology. Here, we review the development of tagraxofusp, and the key preclinical insights and clinical data that led to approval. Tagraxofusp treatment is associated with a unique toxicity, capillary leak syndrome (CLS), which can be severe but is manageable with proper patient selection and monitoring, early recognition, and directed intervention. We outline our approach to the use of tagraxofusp and discuss open questions in the treatment of BPDCN. Overall, tagraxofusp represents a unique targeted therapy and a step forward in meeting an unmet need for patients with this rare disease.

Introduction Blastic plasmacytoid dendritic cell neoplasm (BPDCN) is an uncommon hematologic malignancy whose rarity (<1% of blood cancers), combined with historical confusion regarding the disease’s defining clinical and pathologic criteria, resulted in a disorienting series of changes in name and classification over decades.1-4 BPDCN is now recognized as a unique disease rather than a subtype of the (relatively) more common acute myeloid leukemia (AML). The diagnosis of BPDCN is made by identifying malignant cells (blasts) that express CD123, with CD4 and/or CD56, and at least one other plasmacytoid dendritic cell (pDC) marker (TCF4, TCL1, CD303, CD304), or that express CD123 with any three pDC markers, without T-cell (CD3), B-cell (CD19), or myeloid/monocytic (lysozyme, myeloperoxidase, or CD14) markers.3 The approval of tagraxofusp, the first agent approved specifically for BPDCN, has made it imperative that hematologists accurately diagnose the disease.

Beyond its rarity and defining immunophenotype (think “123456” for CD123, CD4, CD56), BPDCN is distinguished by an older (median age 65-70 years) and male (male:female ratio approx. 2-5:1) demographic. It has an aggressive clinical course, and involvement of multiple tissue compartments is common. Patients with BPDCN regularly present to a dermatologist with characteristic, dark purple tumors or plaques involving the skin; initially, these are frequently in sun-exposed areas but they can be found throughout the body (Figure 1). Once identified as likely BPDCN, patients should be referred to a hematologist, and further testing often reveals involvement of blood and marrow, lymph nodes, and central nervous system (CNS). Rarely, a patient does not have skin involvement and the disease is recognized after the patient develops symptomatic cytopenias, adenopathy, or constitutional complaints. An additional challenge in the diagnosis of BPDCN is that it may occur concurrent with, prior to, or subsequent to a diagnosis of another myeloid neoplasm such as chronic myelomonocytic leukemia (CMML),

Haematologica | 109 January 2024

SPOTLIGHT REVIEW ARTICLE - Tagraxofusp for BPDCN

M.R. Luskin and A.A. Lane

Figure 1. Characteristic blastic plasmacytoid dendritic cell neoplasm lesions and example of a skin lesion achieving complete response with residual skin abnormality not indicative of active disease. (A) Typical skin lesions at the time of diagnosis showing multiple purple bruise-like nodules, plaques, and tumors. (B) Close-up view of a skin lesion prior to treatment. (C) Skin area from panel (B) after two cycles of tagraxofusp, showing a complete response with residual skin abnormality not indicative of active disease. This area of residual hyperpigmentation was biopsied; pathological review showed no overt active disease.

achieve complete remission (CR) more frequently, and are typically offered allogeneic hematopoietic stem cell transplant (allo-HSCT) in first CR (CR1) with long-term remissions regularly demonstrated.8,9,11,14-19 The role of autologous transplant is not well defined although some series suggest benefit.20 BPDCN that relapses after any approach is typically refractory and almost never cured. Given that few patients with BPDCN can benefit from intensive conventional chemotherapy induction approaches and allo-HSCT, BPDCN has long represented a disease with a significant unmet need.

myelodysplastic syndrome (MDS), or AML.5,6 Co-occurring diagnoses, often confirmed to be clonally related, occur in at least 10% of cases, and may be under-recognized.

Treatment of blastic plasmacytoid dendritic cell neoplasm with chemotherapy Prior to the approval of tagraxofusp, treatment of BPDCN involved applying a multi-agent chemotherapy regimen developed for AML, acute lymphoblastic leukemia (ALL) or lymphoma.7-10 Even patients who present with more indolent “skin-only” disease should be offered systemic therapy. In the absence of systemic treatment, the disease invariably advances to a leukemic or lymphomatous phase, although the time to progression is variable. No chemotherapy regimens have been prospectively compared in patients with BPDCN. However, retrospective studies suggest that acute leukemia regimens, particularly ALL regimens, have superior efficacy.9-11 Despite this, chemotherapy-treated patients fare poorly due to inherently aggressive, chemo-resistant disease biology combined with an older patient demographic precluding many from receiving optimal treatment intensity. As with older adults treated for AML and ALL, older patients with BPDCN experience high early mortality, modest remission rates (approx. 50%), and a median overall survival (OS) of less than two years due to early treatment failure (refractory disease, early mortality), relapses, and deaths in remission from regimen-related toxicity.9,12,13 Younger patients who can tolerate intensive conventional chemotherapy regimens may

Tagraxofusp: pharmacology and initial experience The use of diphtheria toxin (DT) as an anti-leukemia immunotoxin has been under investigation since the late 1970s.21,22 Decades later, in the early 2000s, Frankel et al. developed DT-IL-3, a truncated diphtheria toxin fused to recombinant human IL-3 (Figure 2). DT-IL-3 binds with high affinity to the IL-3 receptor (IL-3R), whose α-subunit is also known as CD123, where it is internalized by receptor-mediated endocytosis. The catalytic domain of the DT is cleaved and translocates into the cytosol where it inactivates elongation factor 2 (EF2), which leads to disruption of protein synthesis and apoptotic cell death. This construct was initially investigated in myeloid neoplasms including AML, with the hope that the drug would be effective against leukemia stem cells (LSC), which strongly express CD123.23,24 In the laboratory, DT-IL-3 efficacy correlated to IL-3R (CD123) density on target cells, but single agent clinical efficacy against MDS

Haematologica | 109 January 2024

SPOTLIGHT REVIEW ARTICLE - Tagraxofusp for BPDCN

M.R. Luskin and A.A. Lane

Figure 2. Tagraxofusp mechanism of action and resistance. (Left) Tagraxofusp is a recombinant fusion protein consisting of interleukin 3 (IL3) fused to a truncated diphtheria toxin payload (DT). Tagraxofusp binds to the IL3 receptor, which consists of an α chain (CD123) and a β chain. Upon binding, the drug-receptor complex is internalized, then DT escapes from an endocytic vesicle into the cytoplasm. DT then catalyzes ADP-ribosylation of a modified histidine residue, called diphthamide, on elongation factor 2 (EF2). ADP-ribosylated EF2 inhibits protein synthesis, which causes cell death. (Right) Cells that acquire resistance to tagraxofusp maintain cell surface CD123/IL3R α and β chain expression, drug internalization, and DT cytoplasmic localization. However, due to DNA methylation-mediated silencing of diphthamide synthesis genes, there is no diphthamide target on EF2 in resistant cells. Thus, DT is unable to catalyze ADPribosylation of EF2, and cells survive.

and AML was limited.25-27 Given that BPDCN cells universally express high levels of CD123, it became an important area of development for DT-IL-3, later re-named SL-401 and then tagraxofusp (TAG).28 Tagraxofusp had potent in vitro and in vivo toxicity against BPDCN cell lines, patient-derived BPDCN cells, and mouse models of BPDCN.23,29 Importantly, in a phase I study of tagraxofusp led by Frankel, 7 of 9 patients with BPDCN (78%) treated with a single course of tagraxofusp at 12.5 μg/kg/day x 5 doses responded (5 achieving CR) with a median duration of response of five months. Although a distinct vascular toxicity, capillary leak syndrome (CLS), occurred in some subjects, the responses and manageable toxicity profile supported further development of tagraxofusp in BPDCN.

refractory (R/R) disease. No maximum-tolerated dose (MTD) was identified, but 12 μg/kg was tested in the stage 2 expansion phase based on overall review of toxicity and efficacy. Stage 3 was the “pivotal” cohort for untreated patients treated at 12 μg/kg, with the goal of demonstrating sufficient efficacy for registration. Stage 4 provided ongoing access to tagraxofusp in the context of a trial. Given the risk of CLS, the trial implemented stringent eligibility criteria. Patients were required to be ≥18 years, with Eastern Cooperative Oncology Group (ECOG) performance status of 0-2 and have normal hepatic (total bilirubin ≤1.5 mg/dL, ALT/AST ≤2.5x ULN), renal (creatinine ≤1.5 mg/dL), and cardiopulmonary function including normal left ventricular ejection fraction. Notably, a baseline albumin of ≥3.2 g/dL was required for study entry. During stage 1 of the trial, a grade 5 CLS event (death) occurred, which prompted the minimum baseline albumin to be raised from 3.0 to 3.2 g/dL. Tagraxofusp as a single agent Response assessments evaluated all major disease comThese pilot data were the basis for the STML-401-0114 partments: skin, marrow, lymph nodes, and extramedulmulti-stage, multi-center phase I / II clinical trial that ad- lary disease. Skin response was evaluated via the modified ministered tagraxofusp in repeating cycles (NCT02113982).30 severity weighted assessment (mSWAT) originally develStage 1 was a 3+3 design where escalating doses of tagra- oped for assessment of cutaneous lymphomas, while xofusp (days 1-5, every 21 days) were administered to pa- marrow and lymph node assessment was according to tients with both newly diagnosed and relapsed and/or AML and lymphoma criteria.31-33 Given the frequency of reHaematologica | 109 January 2024

SPOTLIGHT REVIEW ARTICLE - Tagraxofusp for BPDCN

M.R. Luskin and A.A. Lane

sidual pigment changes with treated skin disease, a new response category called “clinical CR (CRc)” was created to designate patients with persistent skin abnormalities but no clinically significant residual disease (Figure 1B, C). This study did not require assessment of the CNS. The initial publication reported on 47 patients: 32 without previous treatment (median age 68 years, 81% male) and 15 with previous treatment (median age 72 years, 87% male). The vast majority (94%) had skin disease, with frequent involvement of other compartments: marrow (51%), lymph nodes (45%), blood (17%), and viscera (17%). Among patients with untreated BPDCN who received 12 μg/kg of tagraxofusp (n=29), there was a 90% overall response rate (ORR) and 72% CR/CRc rate (54% CR/CRc rate in the “pivotal” stage 3 cohort), which translated into a 24-month OS of 52% (Table 1). Almost half (45%) of patients proceeded to stem cell transplantation. Responses occurred quickly, after one or two cycles. Responses were less frequent in patients who entered the study with relapsed or refractory disease (ORR 67%). Long-term follow-up of STML-401-0114, including the stage 4 continued access phase, was recently published, confirming efficacy of tagraxofusp in a larger group of patients (n=84) with longer follow-up (Table 1).34 Responses were durable (median 24.9 months, 95% CI: 3.9-not reached). The rate of CLS was reported to be 21%, with 7% being grade ≥3. Of note, initial “real-world” clinical practice data from a European expanded access program were recently presented confirming a 67% CR rate (10 out of 15) in previously untreated patients, with a similar safety profile and no deaths related to CLS.35 The STML-401-0114 data resulted in US Food and Drug Administration (FDA) approval for tagraxofusp in 2018, for patients aged ≥2 years with newly diagnosed or R/R BPDCN. Tagraxofusp was the first drug approved specifically for BPDCN, as well as the first anti-CD123 drug approved for any indication.36 The European Medicines Agency followed in 2021 with an approval for first-line treatment in adults. CLS is listed as a “black box warning” on the FDA label.

Key toxicity: capillary leak syndrome Capillary leak syndrome (CLS) is associated with edema, weight gain, hypoalbuminemia, and cardiovascular instability, and can be rapidly fatal if not recognized and managed properly. Fortunately, with proper patient selection, vigilant monitoring for early identification, and appropriate treatment, CLS has become a manageable toxicity (Table 2). Patients offered tagraxofusp should meet STML-401-0114 eligibility criteria and demonstrate adequate performance status, organ function, and albumin ≥3.2 g/dL. Interestingly CLS, when it occurs, is an early toxicity occurring almost universally during cycle 1. CLS does not frequently recur. Thus, patients should be admitted to the hospital for administration of the first cycle of therapy to enable increased monitoring and early intervention, while later cycles may be administered in an outpatient setting. The median time to CLS was six days in STML-401-0114, with a median duration of six days. It is prudent to monitor patients in the hospital for 1-2 days after completion of five days of dosing during cycle 1. Monitoring protocols should include at least daily clinical assessments (including weight), and measurement of albumin and organ function (creatinine, liver function tests) which should be reviewed each day prior to dosing. The expertise of experienced nursing staff is important, and thus safe administration of tagraxofusp requires a multidisciplinary collaboration and training. When there is concern for CLS based on clinical or lab (particularly albumin) parameters, treatment is prompt administration of intravenous (IV) albumin, at least once or twice per day, highdose IV steroids (methylprednisolone 1-2 mg/kg/day or equivalent), and careful fluid management (often diuresis, but sometimes fluid repletion). If there is any clinically significant CLS, subsequent tagraxofusp doses should be withheld, and likely omitted, in the first cycle. The package insert should be carefully reviewed and followed. Other common toxicities that typically occur during cycle 1 include transaminitis and mild thrombocytopenia. These

Table 1. Responses for first-line patients with blastic plasmacytoid dendritic cell neoplasm treated with tagraxofusp at 12 μg/kg/day. Stages 1-3, N=29

Overall, N=65

21 (72%)

37 (57%)

43 days (range, 14-131)

39 days (range, 14-131)

13 (45%)

21 (32%)

Median duration CR + CRc in months (95% CI) Probability at 24 months, %

NR (5.9-NR) 65

24.9 (3.8-NR) 53

Median OS in months (95% CI) Probability at 12 months, % Probability at 24 months, %

25.8 (9.6-53.9) 62 52

15.8 (9.7-25.8) 55 40

CR + CRc Median time to CR + CRc Bridged to transplant

Stage 1-3 is initial dose finding, expansion, and pivotal cohorts.30 Overall is Stage 1-3 plus a continued access expansion cohort.34 CR: complete response; CRc: CR with residual skin abnormality not indicative of active disease; CI: confidence interval; OS: overall survival; NR: not reached.

Haematologica | 109 January 2024

SPOTLIGHT REVIEW ARTICLE - Tagraxofusp for BPDCN

M.R. Luskin and A.A. Lane

typically resolve quickly and do not recur. Notably, there is no significant myelosuppression or infection risk with tagraxofusp. The reason for tagraxofusp-related toxicity being restricted to cycle 1 is not known.

basis for ongoing studies that combine tagraxofusp with azacitidine and/or venetoclax (NCT03113643), or with conventional chemotherapy with or without venetoclax (NCT04216524) (Table 3).

Open questions for the use of tagraxofusp in blastic plasmacytoid Cells that develop acquired resistance to tagraxofusp gendendritic cell neoplasm Resistance to tagraxofusp

erally do not lose CD123.37 Instead, cells develop resistance to diphtheria toxin, which is mediated by DNA methylation-mediated silencing of genes in the diphthamide synthesis pathway (Figure 2). Diphthamide is a modified histidine amino acid on elongation factor 2 (EF2) and is the target for ADP ribosylation by diphtheria toxin. Diphthamide silencing is reversible by azacitidine, and this also reverses tagraxofusp resistance. Furthermore, cells that persist after tagraxofusp treatment have increased apoptotic priming and dependence on the anti-apoptotic protein BCL2. These adaptations promote increased sensitivity to the BCL2 inhibitor venetoclax and to conventional chemotherapy. These data provide a rational

Much is known about tagraxofusp, its mechanism of action, modes of resistance, toxicity profile, and strategies for proper patient selection and supportive management. However, many questions and challenges still remain regarding the use of this novel therapeutic.

Is tagraxofusp the optimal ﬁrst-line therapy for blastic plasmacytoid dendritic cell neoplasm? Although tagraxofusp is the only drug specifically approved for BPDCN, it may not always be the most appropriate firstline therapy. Some patients are not eligible for tagraxofusp due to laboratory derangements (hypoalbuminemia, trans-

Table 2. Tagraxofusp administration: basics. Patient selection (based on STML-401-0114 eligibility criteria)

Normal hepatic function (Tbili ≤1.5 mg/dL, ALT/AST ≤2.5x ULN) Normal renal function (creatinine ≤1.5 mg/dL) Normal cardiopulmonary function, including normal left ventricular ejection fraction Baseline albumin of ≥3.2 g/dL Adults (US FDA and EMA) Children ≥2 years (US FDA only)

Patient age (approval indication)

Newly diagnosed (US FDA and EMA) Relapsed, refractory (US FDA only)

Treatment line Dose and schedule

IV infusion (15 min) 12 μg/kg/day on days 1-5 of a 21-day cycle

Number of cycles

No maximum, continue until progression or transplant

Treatment setting

Cycle 1, inpatient Cycle 2+, outpatient permitted

Pre-medication, for infusion hyper-sensitivity reactions Monitoring, minimum daily pre-dose Major toxicities

Corticosteroids, H1 and H2 blockers, and acetaminophen (see package insert) Vital signs, weight (consistent scale), examination (focus on cardiopulmonary status, volume status) Renal function (creatinine), liver function tests (ALT, AST, bilirubin), albumin Capillary leak syndrome, transaminitis, thrombocytopenia

Timing of major toxicities

Cycle 1, doses may be delayed up to 10 days for management of toxicities. Omission of subsequent doses recommended in setting of clinically significant CLS

Reasons to withhold drug

ALT, AST >5x ULN (resume when <2.5x ULN) CLS (consider discontinuation until next cycle; see below)

Capillary leak syndrome Key features

Key management tools

Edema, weight gain, hypoalbuminemia, cardiovascular instability IV albumin, to achieve ≥3.5 g/dL or pre-treatment value High-dose IV steroids (at least methylprednisolone 1-2 mg/kg/day or equivalent) Fluid management (often diuresis, but in some cases fluid repletion) Withhold drug, likely for remainder of cycle

STML-401-0114: multi-stage, multi-center phase I / II clinical trial (NCT02113982); Tbili: total bilirubin; ALT: alanine aminotransferase; AST: aspartate aminotransferase; ULN: Upper Limit of Normal; US FDA: US Food and Drug Administration; EMA: European Medicines Agency; IV: intravenous; min: minutes; H1/H2: histamine blockers at H1/H2 receptor; CLS: capillary leak syndrome. Haematologica | 109 January 2024

SPOTLIGHT REVIEW ARTICLE - Tagraxofusp for BPDCN

M.R. Luskin and A.A. Lane

Table 3. Select clinical trials in blastic plasmacytoid dendritic cell neoplasm. Regimen

NCT number (sponsor)

Tagraxofusp combinations Tagraxofusp with azacitidine, venetoclax Tagraxofusp with chemotherapy (hyper-CVAD), venetoclax

NCT03113643 (Dana-Farber Cancer Institute) NCT04216524 (MD Anderson Cancer Center)

Other CD123 approaches IMGN632 (pivekimab sunirine), a CD123 directed antibody-drug conjugate MB-102, a CD123 auto-CAR T

NCT03386513 (Immunogen) NCT04109482 (Mustang)

NCT: National Clinical Trial; CVAD: cyclophosphamide, vincristine sulfate, doxorubicin hydrochloride and dexamethasone; auto-CAR T: autologous chimeric antigen receptor T-cell therapy.

aminitis), or tenuous cardiovascular or pulmonary status as a result of high burden, proliferative disease. These patients should receive conventional chemotherapy for urgent disease control. They may be able to receive tagraxofusp later during treatment. Other patients will never be eligible for tagraxofusp due to baseline, unmodifiable comorbidities such as a low cardiac ejection fraction. These patients may receive less intensive conventional chemotherapy approaches including ALL regimens with dose adjustments or AML-style therapy with a hypomethylating agent (HMA) and venetoclax (VIALE-A regimen).38 BPDCN cells are uniformly dependent on BCL2 and this translates to high sensitivity to venetoclax.39 For patients eligible for either tagraxofusp or conventional chemotherapy, no randomized comparative data exist. Each regimen’s toxicities must be considered along with the patient’s comorbidities, disease features, and personal preferences, along with the capabilities, preferences, and experience of the treating physician and center. Older patients who may be vulnerable to complications from intensive chemotherapy, particularly myelosuppression, may benefit most from tagraxofusp over other approaches. On the other hand, for younger, fit patients, recent retrospective data generated by investigators at the University of Texas MD Anderson Cancer Center and Moffitt Cancer Center suggest that tagraxofusp and the intensive HCVAD regimen may result in an equivalent CR rate.8,16 We currently lack clinicopathologic features that predict the likelihood of an excellent response to tagraxofusp versus other approaches. Patients are encouraged to enroll on prospective trials of tagraxofusp, other CD123-directed investigational therapies, chemotherapy, and/or venetoclax (Table 3). The optimal approach may involve combining and/or sequencing multiple active agents to maximize depth and durability of responses. Ideal combination and maintenance therapy might eventually obviate the need for allo-HSCT for cure, although we still recommend prompt transplantation in eligible patients in CR1. It is important to recognize that tagraxofusp is not approved in all countries or may not be available for other reasons such as expense, lack of infrastructure, or inadequately trained staff for urgent administration. Thus, con-

ventional chemotherapy remains an important treatment for many patients diagnosed with BPDCN even when tagraxofusp might be considered otherwise. Cost and resource barriers to accessing tagraxofusp and other novel treatment approaches will be important considerations in the coming years.

How should the central nervous system be staged and managed in tagraxofusp-treated patients? It is now recognized that BPDCN, like ALL, commonly involves the CNS at diagnosis and/or at relapse, ultimately affecting up to one-third of patients.40-42 All patients diagnosed with BPDCN should have CNS assessment and serial intrathecal (IT) chemo-prophylaxis during initial treatment. As tagraxofusp is not known to cross the blood-brain barrier, IT chemotherapy is recommended concurrently with tagraxofusp-based treatment. Although the best approach to management of the CNS in BPDCN has not been established and requires further study, we recommend adopting standards used to prevent and treat CNS disease in patients with high-risk ALL. How should BPDCN with associated hematologic malignancy be managed? At least 15-20% of patients with BPDCN have a prior or concurrent diagnosis of MDS, CMML, or AML. As outlined above, the early studies conducted by Frankel suggest that tagraxofusp alone may not be effective treatment for these diagnoses, although this could depend on CD123 expression level or other, yet undetermined, predictive factors. The best approach to these complex situations requires further study. For those with concurrent myeloid neoplasms, HMA and venetoclax-based therapies are attractive in combination with tagraxofusp. What is the role of measurable residual disease monitoring? As BPDCN commonly involves multiple compartments (marrow, blood, lymph nodes, skin, and CNS), response assessments must involve imaging (positron emission tomography-computed tomography), bone marrow examinations, cerebrospinal fluid assessments, and detailed skin examinations. An area requiring further study is the clinical rel-

Haematologica | 109 January 2024

SPOTLIGHT REVIEW ARTICLE - Tagraxofusp for BPDCN

M.R. Luskin and A.A. Lane

evance and best method to assess measurable residual disease (MRD). Assessing MRD in the skin is particularly challenging as residual pigment changes must be distinguished from active disease (Figure 2B, C). Finally, it is possible that MRD kinetics may be different in patients treated with conventional chemotherapy, tagraxofusp, and combination therapies.

apy,46 most pediatric oncologists recommend high-risk ALL-type chemotherapy rather than tagraxofusp as initial therapy, and do not recommend allo-HSCT in CR1. The optimal integration of tagraxofusp into treatment of children with BPDCN requires further study.

Summary: how I treat

What about CD123-targeted agents under development? Tagraxofusp was the first drug developed and approved for BPDCN, and the first CD123 targeted agent in oncology. Now, the field is rapidly evolving with other CD123-directed agents under development. The antibody-drug conjugate, IMGN632 (pivekimab sunirine), received US FDA breakthrough therapy designation in 2020 for relapsed or refractory BPDCN based on encouraging early data,43 and is currently being tested for both newly diagnosed and R/R disease (Table 3). Notably, IMGN632 activity includes patients who previously received tagraxofusp, supporting sequential CD123 therapy and the observation that CD123 expression is often maintained after tagraxofusp. Of note, IMGN632 does not cause CLS, although peripheral edema is reported. As additional CD123-directed agents become available, the relative characteristics of each regarding side-effect profile, route and schedule of administration, cost, ability to combine with other therapeutics, and approved indication, may shift the therapeutic landscape for this disease.

Blastic plasmacytoid dendritic cell neoplasm is a rare, aggressive leukemia with unique clinical and pathologic features. Most commonly affecting older men, BPDCN can involve the skin, bone marrow, lymph nodes, and CNS. Given the rarity of the diagnosis and unique management considerations, patients should, when possible, be referred to a cancer center with clinical trial availability and with experience in evaluating and managing this disorder. Comprehensive staging and expert hematopathology review are imperative. BPDCN universally expresses CD123, the IL-3 receptor, which is an important therapeutic target. Treatment is systemic, regardless of clinical presentation. In 2023, patients may be treated with tagraxofusp, an IL3-diphtheria toxin conjugate targeting CD123, or a conventional chemotherapy regimen. Novel CD123-directed agents are under development. Consideration of allo-HSCT in CR1 is recommended whenever possible. Treatment selection is dictated by disease features, comorbidities, treatment goals, patient choice, and center expertise. Patients being treated with tagraxofusp require How should tagraxofusp be incorporated into the treatment close monitoring for established toxicities including infuof children with BPDCN? sion reactions, transaminitis, cytopenias, and particularly The US FDA approval of tagraxofusp includes children ≥2 for CLS, which all have the highest risk during the first years, based on activity with no unexpected side-effects cycle of therapy. in a small series of pediatric patients with relapsed BPDCN.44 Pediatric BPDCN has similar clinical character- Disclosures istics at diagnosis as BPDCN in adults, although the gen- AAL has received research support from Abbvie and Stemetics differ. Adult BPDCN is characterized by “secondary line Therapeutics. AAL has received consulting fees from Cimyeloid-type” mutations (TET2, ASXL1, splicing factors), meio Therapeutics, IDRx, Jnana Therapeutics, and Qiagen. whereas pediatric BPDCN lacks those mutations and instead is enriched for translocations that encode fusions Contributions of the transcription factor MYB.45 Given that pediatric pa- MRL and AAL conceived, wrote, and edited the manuscript. tients with BPDCN respond well to ALL-type chemother-

References 1. Vardiman JW, Thiele J, Arber DA, et al. The 2008 revision of the World Health Organization (WHO) classification of myeloid neoplasms and acute leukemia: rationale and important changes. Blood. 2009;114(5):937-951. 2. 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. 3. Khoury JD, Solary E, Abla O, et al. The 5th edition of the World Health Organization Classification of Haematolymphoid Tumours: Myeloid and Histiocytic/Dendritic Neoplasms.

Leukemia. 2022;36(7):1703-1719. 4. Guru Murthy GS, Pemmaraju N, Atallah E. Epidemiology and survival of blastic plasmacytoid dendritic cell neoplasm. Leuk Res. 2018;73:21-23. 5. Batta K, Bossenbroek HM, Pemmaraju N, et al. Divergent clonal evolution of blastic plasmacytoid dendritic cell neoplasm and chronic myelomonocytic leukemia from a shared TET2-mutated origin. Leukemia. 2021;35(11):3299-3303. 6. Luskin MR, Kim AS, Patel SS, Wright K, LeBoeuf NR, Lane AA. Evidence for separate transformation to acute myeloid

Haematologica | 109 January 2024

SPOTLIGHT REVIEW ARTICLE - Tagraxofusp for BPDCN

M.R. Luskin and A.A. Lane

leukemia and blastic plasmacytoid dendritic cell neoplasm from a shared ancestral hematopoietic clone. Leuk Lymphoma. 2020;61(9):2258-2261. 7. Pagano L, Valentini CG, Grammatico S, Pulsoni A. Blastic plasmacytoid dendritic cell neoplasm: diagnostic criteria and therapeutical approaches. Br J Haematol. 2016;174(2):188-202. 8. Pemmaraju N, Wilson NR, Garcia-Manero G, et al. Characteristics and outcomes of patients with blastic plasmacytoid dendritic cell neoplasm treated with frontline HCVAD. Blood Adv. 2022;6(10):3027-3035. 9. Taylor J, Haddadin M, Upadhyay VA, et al. Multicenter analysis of outcomes in blastic plasmacytoid dendritic cell neoplasm offers a pretargeted therapy benchmark. Blood. 2019;134(8):678-687. 10. Haddadin M, Taylor J. Chemotherapy options for blastic plasmacytoid dendritic cell neoplasm. Hematol Oncol Clin North Am. 2020;34(3):539-552. 11. Garnache-Ottou F, Vidal C, Biichle S, et al. How should we diagnose and treat blastic plasmacytoid dendritic cell neoplasm patients? Blood Adv. 2019;3(24):4238-4251. 12. Luskin MR. Acute lymphoblastic leukemia in older adults: curtain call for conventional chemotherapy? Hematology Am Soc Hematol Educ Program. 2021;2021(1):7-14. 13. Kantarjian H, O'Brien S, Cortes J, et al. Results of intensive chemotherapy in 998 patients age 65 years or older with acute myeloid leukemia or high-risk myelodysplastic syndrome: predictive prognostic models for outcome. Cancer. 2006;106(5):1090-1098. 14. Roos-Weil D, Dietrich S, Boumendil A, et al. Stem cell transplantation can provide durable disease control in blastic plasmacytoid dendritic cell neoplasm: a retrospective study from the European Group for Blood and Marrow Transplantation. Blood. 2013;121(3):440-446. 15. Kharfan-Dabaja MA, Cherry M. Hematopoietic cell transplant for blastic plasmacytoid dendritic cell neoplasm. Hematol Oncol Clin North Am. 2020;34(3):621-629. 16. Yun S, Chan O, Kerr D, et al. Survival outcomes in blastic plasmacytoid dendritic cell neoplasm by first-line treatment and stem cell transplant. Blood Adv. 2020;4(14):3435-3442. 17. Kharfan-Dabaja MA, Pemmaraju N, Mohty M. Therapeutic approaches for blastic plasmacytoid dendritic cell neoplasm: allogeneic hematopoietic cell transplantation and novel therapies. Clin Hematol Int. 2019;1(1):2-9. 18. Kharfan-Dabaja MA, Reljic T, Murthy HS, Ayala E, Kumar A. Allogeneic hematopoietic cell transplantation is an effective treatment for blastic plasmacytoid dendritic cell neoplasm in first complete remission: systematic review and meta-analysis. Clin Lymphoma Myeloma Leuk. 2018;18(11):703-709.e1. 19. Kharfan-Dabaja MA, Al Malki MM, Deotare U, et al. Haematopoietic cell transplantation for blastic plasmacytoid dendritic cell neoplasm: a North American multicentre collaborative study. Br J Haematol. 2017;179(5):781-789. 20. Aoki T, Suzuki R, Kuwatsuka Y, et al. Long-term survival following autologous and allogeneic stem cell transplantation for blastic plasmacytoid dendritic cell neoplasm. Blood. 2015;125(23):3559-3562. 21. Thorpe PE, Ross WC, Cumber AJ, Hinson CA, Edwards DC, Davies AJ. Toxicity of diphtheria toxin for lymphoblastoid cells is increased by conjugation to antilymphocytic globulin. Nature. 1978;271(5647):752-755. 22. Ross WC, Thorpe PE, Cumber AJ, Edwards DC, Hinson CA, Davies AJ. Increased toxicity of diphtheria toxin for human lymphoblastoid cells following covalent linkage to anti-(human lymphocyte) globulin or its F(ab')2 fragment. Eur J Biochem. 1980;104(2):381-390.

23. Frankel AE, Ramage J, Kiser M, Alexander R, Kucera G, Miller MS. Characterization of diphtheria fusion proteins targeted to the human interleukin-3 receptor. Protein Eng. 2000;13(8):575-581. 24. Frankel AE, McCubrey JA, Miller MS, et al. Diphtheria toxin fused to human interleukin-3 is toxic to blasts from patients with myeloid leukemias. Leukemia. 2000;14(4):576-585. 25. Pemmaraju N, Konopleva M. Approval of tagraxofusp-erzs for blastic plasmacytoid dendritic cell neoplasm. Blood Adv. 2020;4(16):4020-4027. 26. Alexander RL, Ramage J, Kucera GL, Caligiuri MA, Frankel AE. High affinity interleukin-3 receptor expression on blasts from patients with acute myelogenous leukemia correlates with cytotoxicity of a diphtheria toxin/IL-3 fusion protein. Leuk Res. 2001;25(10):875-881. 27. Frankel A, Liu JS, Rizzieri D, Hogge D. Phase I clinical study of diphtheria toxin-interleukin 3 fusion protein in patients with acute myeloid leukemia and myelodysplasia. Leuk Lymphoma. 2008;49(3):543-553. 28. Frankel AE, Woo JH, Ahn C, et al. Activity of SL-401, a targeted therapy directed to interleukin-3 receptor, in blastic plasmacytoid dendritic cell neoplasm patients. Blood. 2014;124(3):385-392. 29. Angelot-Delettre F, Roggy A, Frankel AE, et al. In vivo and in vitro sensitivity of blastic plasmacytoid dendritic cell neoplasm to SL-401, an interleukin-3 receptor targeted biologic agent. Haematologica. 2015;100(2):223-230. 30. Pemmaraju N, Lane AA, Sweet KL, et al. Tagraxofusp in blastic plasmacytoid dendritic-cell neoplasm. N Engl J Med. 2019;380(17):1628-1637. 31. Olsen EA, Whittaker S, Kim YH, et al. Clinical end points and response criteria in mycosis fungoides and Sezary syndrome: a consensus statement of the International Society for Cutaneous Lymphomas, the United States Cutaneous Lymphoma Consortium, and the Cutaneous Lymphoma Task Force of the European Organisation for Research and Treatment of Cancer. J Clin Oncol. 2011;29(18):2598-2607. 32. 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. 33. Cheson BD, Bennett JM, Kopecky KJ, et al. Revised recommendations of the International Working Group for Diagnosis, Standardization of Response Criteria, Treatment Outcomes, and Reporting Standards for Therapeutic Trials in Acute Myeloid Leukemia. J Clin Oncol. 2003;21(24):4642-4649. 34. Pemmaraju N, Sweet KL, Stein AS, et al. Long-term benefits of tagraxofusp for patients with blastic plasmacytoid dendritic cell neoplasm. J Clin Oncol. 2022;40(26):3032-3036. 35. Deconinck E, Anant M, Manteigas D, et al. Preliminary results from an observational multicenter study of patients with blastic plasmacytoid dendritic cell neoplasm treated with tagraxofusp in the European Expanded Access Program. Blood. 2022;140(Suppl 1):8115-8116. 36. Jen EY, Gao X, Li L, et al. FDA approval summary: tagraxofusperzs for treatment of blastic plasmacytoid dendritic cell neoplasm. Clin Cancer Res. 2020;26(3):532-536. 37. Togami K, Chung SS, Madan V, et al. Sex-biased ZRSR2 mutations in myeloid malignancies impair plasmacytoid dendritic cell activation and apoptosis. Cancer Discov. 2022;12(2):522-541. 38. DiNardo CD, Jonas BA, Pullarkat V, et al. Azacitidine and venetoclax in previously untreated acute myeloid leukemia. N Engl J Med. 2020;383(7):617-629. 39. Montero J, Stephansky J, Cai T, et al. Blastic plasmacytoid

Haematologica | 109 January 2024

SPOTLIGHT REVIEW ARTICLE - Tagraxofusp for BPDCN

M.R. Luskin and A.A. Lane

dendritic cell neoplasm is dependent on BCL2 and sensitive to venetoclax. Cancer Discov. 2017;7(2):156-164. 40. Martin-Martin L, Almeida J, Pomares H, et al. Blastic plasmacytoid dendritic cell neoplasm frequently shows occult central nervous system involvement at diagnosis and benefits from intrathecal therapy. Oncotarget. 2016;7(9):10174-10181. 41. Davis JA, Rizzieri DA, Lane AA, et al. Treatment patterns and outcomes of patients with CNS involvement of blastic plasmacytoid dendritic cell neoplasm (BPDCN). Leuk Lymphoma. 2022;63(11):2757-2759. 42. Pemmaraju N, Wilson NR, Khoury JD, et al. Central nervous system involvement in blastic plasmacytoid dendritic cell neoplasm. Blood. 2021;138(15):1373-1377. 43. Pemmaraju N, Martinelli G, Todisco E, et al. Clinical profile of

IMGN632, a novel CD123-targeting antibody-drug conjugate (ADC), in patients with relapsed/refractory (R/R) blastic plasmacytoid dendritic cell neoplasm (BPDCN). Blood. 2020;136(Suppl 1):11-13. 44. Sun W, Liu H, Kim Y, et al. First pediatric experience of SL-401, a CD123-targeted therapy, in patients with blastic plasmacytoid dendritic cell neoplasm: report of three cases. J Hematol Oncol. 2018;11(1):61. 45. Suzuki K, Suzuki Y, Hama A, et al. Recurrent MYB rearrangement in blastic plasmacytoid dendritic cell neoplasm. Leukemia. 2017;31(7):1629-1633. 46. Kim MJ, Nasr A, Kabir B, et al. Pediatric blastic plasmacytoid dendritic cell neoplasm: a systematic literature review. J Pediatr Hematol Oncol. 2017;39(7):528-537.

Haematologica | 109 January 2024

ARTICLE - Acute Lymphoblastic Leukemia

Genomic analysis of venous thrombosis in children with acute lymphoblastic leukemia from diverse ancestries Yan Zheng,1* Wenjian Yang,2* Jeremie Estepp,3,4 Deqing Pei,5 Cheng Cheng,5 Clifford M. Takemoto,4 Hiroto Inaba,6 Sima Jeha,3,6 Ching-Hon Pui,6 Mary V. Relling2 and Seth E. Karol6

Correspondence: S. E. Karol Seth.Karol@stjude.org

Department of Pathology; Department of Pharmaceutical Sciences; Department of Global 4

Pediatric Medicine; Department of Hematology; Department of Biostatistics and 6

Department of Oncology. St. Jude Children’s Research Hospital, Memphis, TN, USA

Received: Accepted: Early view:

January 22, 2023. June 29, 2023. July 6, 2023.

https://doi.org/10.3324/haematol.2022.281582 *

YZ and WY contributed equally as first authors.

Abstract Venous thrombosis is a common adverse effect of modern therapy for acute lymphoblastic leukemia (ALL). Prior studies to identify risks of thrombosis in pediatric ALL have been limited by genetic screens of pre-identified genetic variants or genome-wide association studies (GWAS) in ancestrally uniform populations. To address this, we performed a retrospective cohort evaluation of thrombosis risk in 1,005 children treated for newly diagnosed ALL. Genetic risk factors were comprehensively evaluated from genome-wide single nucleotide polymorphism (SNP) arrays and were evaluated using Cox regression adjusting for identified clinical risk factors and genetic ancestry. The cumulative incidence of thrombosis was 7.8%. In multivariate analysis, older age, T-lineage ALL, and non-O blood group were associated with increased thrombosis while non-low-risk treatment and higher presenting white blood cell count trended toward increased thrombosis. No SNP reached genome-wide significance. The SNP most strongly associated with thrombosis was rs2874964 near RFXAP (G risk allele; P=4x10-7; hazard ratio [HR] =2.8). In patients of non-European ancestry, rs55689276 near the α globin cluster (P=1.28x10-6; HR=27) was most strongly associated with thrombosis. Among GWAS catalogue SNP reported to be associated with thrombosis, rs2519093 (T risk allele, P=4.8x10-4; HR=2.1), an intronic variant in ABO, was most strongly associated with risk in this cohort. Classic thrombophilia risks were not associated with thrombosis. Our study confirms known clinical risk features associated with thrombosis risk in children with ALL. In this ancestrally diverse cohort, genetic risks linked to thrombosis risk aggregated in erythrocyte-related SNP, suggesting the critical role of this tissue in thrombosis risk.

Introduction

V Leiden have yielded inconsistent results.5-9 However, several of these studies have only ascertained thrombophilia genetics in patients who developed thromboses.7 8 Other studies used targeted screening but lacked genome-wide assessments.10 Prior genome-wide assessments of thrombosis risk in ALL have been limited by homogenous population selection, with analyses limited to patients of European ancestry.4,11,12 As a result, there have been no comprehensive genetic evaluations of the risk of thrombosis during ALL therapy in genetically diverse cohorts. Because studies have demonstrated an increased risk of venous thrombosis among Americans of African ancestry as compared to those with European ancestry in both the general population13,14 and adults with cancer,15 assessment of this risk factor in the context of ALL therapy and identification of any genetic features driving such differences are needed.

Contemporary treatment for children with acute lymphoblastic leukemia (ALL) cures more than 90% of patients.1,2 However, therapy-related toxicities including venous thrombosis can reduce event-free survival and overall survival as well as quality of life of patients and survivors. The risk of thrombosis in children with ALL is dramatically higher than the general pediatric population, likely related to a combination of known prothrombotic risk factors, including the presence of central venous catheters, inflammation and other direct effects from leukemia, and chemotherapy, particularly glucocorticoids and asparaginase.3 As pediatric therapeutic strategies have expanded to treat young adults, thrombosis is increasingly occurring in this population.4 Prior studies of the association between inherited thrombophilia variants such as prothrombin G20210A and factor

Haematologica | 109 January 2024

ARTICLE - Thrombosis in pediatric ALL

Y. Zheng et al.

Methods

with data present on 26 March 2020.18 Identified traits and SNP linked to thrombosis are found in the Online Supplementary Table S1.

Patients and clinical trials All patients were treated at our institution and enrolled on the Total XV (change both to: clinicaltrials gov. Identifier: NCT00137111) or Total XVI (change both to: clinicaltrials gov. Identifier: NCT00549848) studies for newly diagnosed ALL. Details of these cohorts and their treatment characteristics have been previously reported and are included in the Online Supplementary Appendix.1,16 Neither trial included recommendations for thromboprophylaxis. Prospective thrombophilia screening was not routinely performed in patients. All clinically significant thromboses (Common Terminology Criteria for Adverse Events versions 2 [Total XV] or 3 [Total XVI] grade 2 or higher) occurring on therapy were prospectively gathered in research databases and were independently reviewed for accuracy. Thrombosis-associated symptoms were assessed at the time of data collection. Data used in the current analyses was current as of 20 January 2023. All studies were approved by the institutional review boards of participating institutions. Patients and/or legal guardians provided informed written consent/assent consistent with the Declaration of Helsinki.

Results

Genotyping and genomic analyses Germline DNA for patients was collected from peripheral blood following remission induction and genotyping was performed in a research setting without return of results to clinicians or patients. Details of genotyping and assessment of genetic ancestry are available in the Online Supplementary Appendix. Statistics Risk for thrombosis was assessed using time-dependent Cox proportional hazard models until the occurrence of the first thrombosis with death in remission, relapse, and termination of protocol therapy treated as competing events. For each single nucleotide polymorphism (SNP), Cox proportional-hazard regressions were performed separately for patients treated on Total XV and Total XVI adjusting for age, ancestry, and risk arm. The results were then combined in a meta-analysis weighted by square-root of the number of patients included in each protocol. Analyses were performed in R17 4.1.1 and Stata 16.1 (StataCorp LLC, College Station, TX). The threshold of statistical significance was P<0.05 for clinical analyses and P<5x10-8 for GWAS analyses. In multivariate analysis, first all factors associated at P<0.05 in univariate analysis were included to define the initial model. Then, the final model was determined by removing factors showing P<0.2 in the initial model. In order to identify variants previously associated with thrombosis or associated phenotypes, we extracted relevant diseases/traits from the GWAS catalogue version 1.02

Clinical associations with thrombosis Thrombosis was detected in 79 of 1,005 evaluated patients. Thrombosis was detected by imaging without clinical symptoms in seven patients, including three patients with cerebral sinus thrombosis discovered on magnetic resonance imaging (MRI) and four with line-associated thromboses. A second episode of thrombosis occurred in nine of 79 patients (11.4% of thromboses), including two patients with three episodes of thrombosis. Subsequent thromboses were at discreet sites from the initial thrombosis in ten of 11 subsequent episodes; seven occurred after completion of the initial anticoagulation course and four occurred while patients were still receiving anticoagulation. Thromboses included 27 central venous line-associated thromboses, eight pulmonary emboli, 21 cerebral venous thromboses, and 34 other deep vein thromboses. Thrombosis was fatal in two cases, including a patient with a fungal pulmonary thrombosis during high-risk pretransplant re-intensification therapy and a patient with Trisomy 21 who developed a saddle pulmonary embolus during week 11 of standard-risk continuation therapy. Univariate analysis showed that older age, T-ALL, and standard- or high-risk therapy were risk factors for thrombosis development (Table 1). Patients who developed thrombosis were older (mean 11 vs. 6.9 years; P<0.001; hazard ratio [HR] =1.17; 95% confidence interval [CI]: 1.12-1.22%; Online Supplementary Figure S1). Thrombosis occurred in 4.3% of children 1-9.99 years old at diagnosis, 8.3% of infants less than 1 year old at diagnosis (HR=2.3; 95% CI: 0.3-16.5 vs. children; P=0.4), and in 17.3% of adolescents 10 years old and older at diagnosis (HR=4.5; 95% CI: 2.9-7.1 vs. children, P<0.001). There was no difference by patients’ genetic ancestry, with 5.5% of African ancestry, 8.7% of Hispanic/Native American ancestry, 8.8% of European ancestry, and 4.9% of patients with other ancestries developing thrombosis (P=0.4; Online Supplemental Figure S2). There was also no difference by sex, with 7.7% of females and 7.8% of males developing thrombosis (P=0.9). Thrombosis rates were similar across protocol, with 7.4% of Total XV and 8.2% of Total XVI patients developing thrombosis (P=0.7). However, certain disease and therapy characteristics were associated with thrombosis risk. Patients treated with standard- or high-risk therapy had a higher thrombosis rate (11.5%) than those treated with low-risk therapy (3.5%, HR=3.6; 95% CI: 2.1-6.2; P<0.001). Thrombosis risk was higher in patients with T-ALL than B-cell precursor ALL (16.8% vs. 6.1%, HR=3; 95% CI: 1.9-4.8; P<0.001). Since all patients with T-ALL received intensified asparagi-

Haematologica | 109 January 2024

ARTICLE - Thrombosis in pediatric ALL

Y. Zheng et al.

nase and dexamethasone treatment in standard-/high-risk arm, comparison was also made between patients with T-ALL and those with B-ALL treated in standard-/highrisk arm and showed a persistent significant increased thrombosis risk for patients with T-ALL (16.8% vs. 9.2%, HR=1.9; 95% CI: 1.2-3.2; P=0.009). There was a non-significant increase in thrombosis in patients with central

nervous system 3 (CNS3) status (cerebrospinal fluid [CSF] white blood cell count [WBC] ≥5/μL with blasts) compared to other patients (17% vs. 7.6%, HR=2.4; 95% CI: 0.97-5.9; P=0.06). Intensified intrathecal (IT) treatment was associated with increased thrombosis risk (9.7% in those receiving 4 to 7 doses during induction vs. 6.3% in those receiving 2 to 3 doses, HR=1.56; 95% CI: 1-2.44; P=0.048). Patients

Table 1. Analysis of risk factors for thrombosis development in children with acute lymphoblastic leukemia.

No thrombosis N=926 Age in years at diagnosis, mean (SD)

Thrombosis Multivariate Multivariate Multivariate N=79 Univariate P with P without P with (% with P blood type and ABO type ABO type thrombosis) rs2519093

6.9 (4.6)

11 (5.1)

<0.001

Age in years at diagnosis 0-0.99 1-9.99 10-18.99

11 691 224

1 (8.3) 31 (4.3) 47 (17.3)

<0.001

Sex F M

394 532

34 (7.9) 45 (7.8)

0.97

Study Total XV Total XVI

377 549

30 (7.4) 49 (8.2)

0.69

Immunophenotype and risk B low-risk B SH-risk T lineage, SH-risk

442 345 139

16 (3.5) 35 (9.2) 28 (16.8)

Ref <0.001 <0.001

Ref * 0.021

Ref * 0.014

Ref 0.15 0.008

CNS3 status at diagnosis Yes No

24 902

5 (17.2) 74 (7.6)

0.093

Intrathecal therapy in induction, N of doses 2-3 4-7

517 409

35 (6.3) 44 (9.7)

0.047

52 (104.4)

93.7 (194.3)

<0.001

0.045

0.011

0.025

Mediastinal mass on initial chest x-ray Yes No Unavailable

92 830 4

14 (13.2) 63 (7.1) 2 (33.3)

0.033

Genetic ancestry African Hispanic/ Native American Other ancestry European Unavailable

138 84 77 612 15

8 (5.5) 8 (8.7) 4 (4.9) 59 (8.8) 0 (0)

0.21 0.99 0.26 Ref NA

NA NA NA NA NA

Blood type OO No Yes Unavailable

444 442 40

49 (10) 25 (5.4) 5 (11.1%)

0.007

0.001

0.11

rs2519093 TT TC CC Unavailable

35 215 656 20

4 (10.3) 36 (14.3) 38 (5.5) 1 (4.8)

<0.001

0.05

WBC count x103/mm3 at diagnosis, mean (SD)

SD: standard deviation; NA: factor not considered in multivariate analysis; *: not selected in multivariate analysis; CNS: central nervous system; WBC: white blood cell; M: male; F: female.

Haematologica | 109 January 2024

ARTICLE - Thrombosis in pediatric ALL

Y. Zheng et al.

who developed thrombosis also had a higher presenting WBC count (mean 93.7 vs. 52x103/mm3, HR=1.25; 95% CI: 1.1-1.41 per 100x103/mm3; P<0.001) and were more likely to have a mediastinal mass on initial chest x-ray (13.2% vs. 7.1%, HR=1.97; 95% CI: 1.1-3.52; P=0.02). In order to identify factors associated which might mediate the increased risk of thrombosis in T-ALL compared to standard/high-risk B-ALL, we assessed factors reported to be more common in T-ALL than B-ALL including CNS status and CNS-directed therapy, presenting WBC count, and the presence of mediastinal mass for their association with thrombosis development. Among patients treated on the standard- or high-risk arm, there was no increase in CNS3 status (CSF WBC count ≥5/μL with blasts) comparing B-ALL to T-ALL (5% vs. 6%; P=0.63). More patients with T-ALL received intensified CNS-directed therapy with 82% receiving at least four intrathecal therapies during induction compared to 50% among standard/high-risk B-ALL (P<0.001). Patients with T-ALL were also more likely to have a mediastinal mass (53.6% vs. 1.3%; P<0.001) and had a higher mean presenting WBC count (154 vs. 57.2x103/ mm3; P<0.001). In this subset of patients, only WBC count was associated with increased thrombosis risk (HR=1.15 per 100x103/mm3; 95% CI: 1.003-1.32; P=0.045). In multivariate analysis, age (HR=1.15 for each additional 1 year; 95% CI: 1.09-1.2; P<0.001) and T-ALL (HR=2.69; 95% CI: 1.29-5.62; P=0.009) remained associated with increased thrombosis risk while B-ALL receiving standard/high-risk

therapy (HR=1.63; 95% CI: 0.83-3.2; P=0.16) and presenting WBC count (HR=1.15; 95% CI: 0.99-1.34 per 100x103/mm3; P=0.07) were also retained in the model but did not reach the threshold of statistical significance (Table 1). Genetic associations with thrombosis Genotype data were available on 983 patients, including 387 treated on Total XV and 596 treated on Total XVI. No SNP were associated with thrombosis at the genome-wide association threshold of 5x10-8 (Figure 1). The SNP most strongly associated with thrombosis in the meta-analysis was rs2874964 near RFXAP (G risk allele, P=4x10-7; HR=2.8; 95% CI: 1.9-4.2), a variant which decreases SREBP binding.19 The coding variant most strongly associated with thrombosis was rs11540822 in CCHCR1 (T risk allele, P=1.1x10-5; HR=3.4; 95% CI: 2-5.9). This missense variant is predicted to be damaging to CCHCR1 function (Ensembl release 10920), and is an expression quantitative trait locus for HLA-C and HCG27 in whole blood (GTeX accessed May 14, 2023).21 Among GWAS catalogue SNP, 279 previously reported to be associated with thrombosis were typed or imputed in our cohort (Online Supplementary Table S1). The GWAS catalog thrombosis SNP most strongly associated with thrombosis in this cohort was rs2519093 (T risk allele, P=4.8x10-4; HR=2.1; 95% CI: 1.4-3.1), an intronic variant in ABO. Notably, there was no association between rs6025 (Leiden mutation) in factor V (P=0.4), rs1799963 (G2020A) in prothrombin (P=0.3), or rs1801133 (C677T) in MTHFR on the development

Figure 1. Manhattan plot of venous thrombosis risk in children with acute lymphoblastic leukemia. Non-coding variants are shown as blue and gray circles. The top variant is rs2874964 near RFXAP (P=4x10-7). Coding variates are shown with blue and gray crosses and the top variant is rs11540822 in CCHCR1 (P=1.1x10-5). Variants previously associated with thrombosis in the genome-wide association studies catalogue are shown as red triangles, with rs2519093 in ABO the top ranked single nucleotide polymorphism (P=4.8x10-4).

Haematologica | 109 January 2024

ARTICLE - Thrombosis in pediatric ALL

Y. Zheng et al.

fication does not differentiate patients of Middle Eastern ancestry from other patients and this population is not a large portion of the communities contributing patients to these trials. This is notable because patients of Middle Eastern ancestry have previously been reported to have a higher incidence of classic thrombophilia genetic variants than the European population.9,10,22,23 Whether our genomic findings are extensible to this population is unknown and requires further study. No SNP reached genome-wide significance in this analysis, suggesting that therapy factors and physiological changes with age are the dominant drivers of thrombosis risk during ALL therapy. Classic thrombophilia polymorphisms in prothrombin, factor V, and MTHFR were not associated with an increased risk of thrombosis during therapy in this cohort, consistent with some prior findings.11,12 This supports our clinical practice to not routinely test for these variants in patients at diagnosis, although such routine testing may be indicated in populations with high rates of thrombophilia variants as noted above. Non-O blood type was associated with a 2-fold increased incidence of thrombosis in this cohort. Non-O blood has previously been associated with an increased risk of thrombosis in children treated on Dana Farber trials for newly diagnosed ALL.24 This increased risk of thrombosis is consistent with the increased risk thrombotic complications observed in individuals with non-O blood type in the general population.25 The preservation of this risk in multivariable analysis and despite the strong impact of therapy suggests that the mechanism of increased risk associated with non-O blood is additive with the risks imposed by ALL therapy. Non-O blood type is associated with elevated levels of factor VIII and von Willebrand factor26 and appears to increase thrombosis risk through this and other mechanisms.25 Interestingly, the intronic rs2519093 variant in ABO increased thrombosis risk despite imperfect correlation with blood type (Online Supplementary Table S2) and in multivariate analysis including blood type. Prior work has linked this variant to increased soluble E-selectin and soluble intercellular adhesion molecule-1 independent of its association with blood group.27 These two adhesion molecules have been associated with increased intravascular inflammation and thrombosis risk,28 suggesting an alternative biological pathway driving thrombosis risk in this population. Limitations of this study include its moderate size relative to most genomic discovery studies. This may limit our ability to identify new genetic variants associated with thrombosis. We addressed this limitation in part by evaluating variants previously associated with thrombosis risk in prior published GWAS. Findings in this asparaginase intensive regimen may be attenuated in children with ALL treated with less asparaginase-intensive therapy. However, because the rs2519093 variant and non-O blood are known to be risk factors in the general population (who receive no asparaginase), we

of thrombosis. Prior GWAS SNP12 rs1804772 and rs570684 were not associated with thrombosis in our cohort (P=0.41 and P=0.82, respectively; Online Supplementary Table S1). When the analysis was restricted to 317 patients of non-European ancestry (CEU <90%), the top SNP was rs55689276 near the α globin cluster (P=1.28x10-6; HR=27; 95% CI: 7.2105; Online Supplementary Figure S3). This SNP had a weak trend in the same direction in European ancestry patients (P=0.19; HR=1.5; 95% CI: 0.83-2.6) and is an eQTL for Ζ globin in whole blood. Because of the association between rs2519093 in ABO with thrombosis in our cohort, we assessed the impact of blood type on thrombosis risk. Clinical blood types were available for 960 patients, including 493 with type non-O blood and 467 with O blood. Non-type O blood was associated with an increased risk of thrombosis (HR=1.89; 95% CI: 1.17-3.07; P=0.007), an association which was retained in multivariate analysis (HR=2.34; 95% CI: 1.42-3.87; P=0.001; Table 1). Despite the strong correlation between rs2519093 genotype and blood group, both factors retained strong trends toward association in multivariable models. (Table 1; Online Supplementary Table S2).

Discussion Despite excellent cure rates, therapy for children with ALL can result in life-threatening consequences. Within our cohort, symptomatic venous thrombosis occurred in 7.8% of 1,004 patients and was fatal in two patients (0.2%). Older age, T-ALL immunophenotype, higher WBC count at diagnosis, standard-/high-risk therapy, and non-O blood group were all associated with increased with thrombosis and retained in multivariate analysis. No genetic variant reached genome-wide significance; the GWAS catalog SNP most strongly associated with thrombosis was the ABO variant rs2519093. In this cohort with diverse ancestries, no genetic ancestry group experienced an increased risk of thrombosis. This contrasts with increased risks of thrombosis in self-identified Blacks in adults with cancer15 and the general population.13,14 This suggests that risk factors associated with ALL therapy override any genetic contribution to the increased risk of thrombosis experienced by Blacks in those studies. A related explanation is that social determinants of health may play a larger role in thrombosis risk in adults than in children receiving therapy for ALL. Interestingly, the SNP most strongly associated with thrombosis in non-European ancestry patients in our cohort differed from that seen in the population at large, suggesting the drivers of thrombosis may differ in patients of different ancestries being treated with ALL. This provides a potential explanation for our failure to replicate thrombosis SNP from a previously published ALL cohort.12 Notably, although this cohort is more ancestrally diverse than prior analyses of European ancestry patients, our genomic ancestry classi-

Haematologica | 109 January 2024

ARTICLE - Thrombosis in pediatric ALL

Y. Zheng et al.

believe these findings are applicable to all children receiving ALL treatment. Interestingly, a variant in the α globin cluster rs55689276 was the top SNP in analyses limited to the non-European population, suggesting genetic risks of thrombosis converge on erythrocytes. The contribution of erythrocytes to thrombosis is now being documented in multiple contexts. Interactions between erythrocytes and oxidative stress contribute to thrombosis initiation and expansion,29 erythrocyte transfusion increases the risk of thrombosis after orthopedic surgery,30 and dysfunctional interactions between erythrocytes and vascular endothelium appear to contribute to COVID-19 associated thrombosis.31 While the findings of this study suggest that age, disease, and treatment factors predominate in modulating the risk of thrombosis in children with ALL, it is possible that genetic risks may play a larger role in cohorts treated with less asparaginase. The increased risk associated with T-immunophenotype, older age, and non-O blood group are all likely to be maintained across cohorts due to their confirmation in multiple other settings. The convergence of genetic risks from this study around erythrocyte characteristics including ABO and globin SNP suggests a consistent impact from this tissue on thrombosis risk, consistent with identified risks in other settings. Further research is needed to understand the interaction of demographic, phenotypic, therapeutic, and genomic contributors to thrombosis risk in the ALL population, and whether the transfusions frequently given to this population impacts this risk. In this large, ancestrally diverse cohort of children treated for ALL, clinical rather than genetic risk factors dom-

inated the risk of thrombosis. This study both identifies easily assessable factors (older age, T immunophenotype, and non-O blood type) to further refine thrombosis risk while providing areas for future evaluation in other diverse cohorts. Such work may provide opportunities to target prophylactic interventions to reduce the frequency of this common toxicity from ALL therapy. Disclosures SEK has served as a consultant to Servier and on an advisory board for Jazz Pharmaceuticals. All other authors have no conflicts of interest to disclose. Contributions WY, MVR and SEK designed the study. YZ, SJ and CHP provided data and/or supervised the clinical trials. WY, DP, CC and SEK performed analyses. YZ, WY and SEK drafted the manuscript which was finalized with input from all authors. Funding This research was supported by National Institutes of Health CA250418, CA142665, CA21765, and GM115279; and ALSAC. 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 De-identified data used in the preparation of this manuscript are available from the authors upon reasonable request for non-commercial use.

References 8. Rank CU, Toft N, Tuckuviene R, et al. Thromboembolism in acute lymphoblastic leukemia: results of NOPHO ALL2008 protocol treatment in patients aged 1 to 45 years. Blood. 2018;131(22):2475-2484. 9. Barzilai-Birenboim S, Arad-Cohen N, Nirel R, et al. Thrombophilia screening and thromboprophylaxis may benefit specific ethnic subgroups with paediatric acute lymphoblastic leukaemia. Br J Haematol. 2019;184(6):994-998. 10. Barzilai-Birenboim S, Nirel R, Arad-Cohen N, et al. Venous thromboembolism and its risk factors in children with acute lymphoblastic leukemia in Israel: a population-based study. Cancers (Basel). 2020;12(10):2759. 11. Abaji R, Gagne V, Xu CJ, et al. Whole-exome sequencing identified genetic risk factors for asparaginase-related complications in childhood ALL patients. Oncotarget. 2017;8(27):43752-43767. 12. Mateos MK, Tulstrup M, Quinn MC, et al. Genome-wide association meta-analysis of single-nucleotide polymorphisms and symptomatic venous thromboembolism during therapy for acute lymphoblastic leukemia and lymphoma in Caucasian children. Cancers (Basel). 2020;12(5):1285. 13. DeMonaco NA, Dang Q, Kapoor WN, Ragni MV. Pulmonary embolism incidence is increasing with use of spiral computed tomography. Am J Med. 2008;121(7):611-617.

1. Jeha S, Pei D, Choi J, et al. Improved CNS control of childhood acute lymphoblastic leukemia without cranial irradiation: St Jude Total Therapy Study 16. J Clin Oncol. 2019;37(35):3377-3391. 2. Vrooman LM, Blonquist TM, Harris MH, et al. Refining risk classification in childhood B acute lymphoblastic leukemia: results of DFCI ALL Consortium Protocol 05-001. Blood Adv. 2018;2(12):1449-1458. 3. Levy-Mendelovich S, Barg AA, Kenet G. Thrombosis in pediatric patients with leukemia. Thromb Res. 2018;164(Suppl 1):S94-S97. 4. Jarvis KB, LeBlanc M, Tulstrup M, et al. Candidate single nucleotide polymorphisms and thromboembolism in acute lymphoblastic leukemia - a NOPHO ALL2008 study. Thromb Res. 2019;184:92-98. 5. Payne JH, Vora AJ. Thrombosis and acute lymphoblastic leukaemia. Br J Haematol. 2007;138(4):430-445. 6. Nowak-Gottl U, Wermes C, Junker R, et al. Prospective evaluation of the thrombotic risk in children with acute lymphoblastic leukemia carrying the MTHFR TT 677 genotype, the prothrombin G20210A variant, and further prothrombotic risk factors. Blood. 1999;93(5):1595-1599. 7. Wermes C, von Depka Prondzinski M, Lichtinghagen R, Barthels M, Welte K, Sykora KW. Clinical relevance of genetic risk factors for thrombosis in paediatric oncology patients with central venous catheters. Eur J Pediatr. 1999;158(Suppl 3):S143-146.

Haematologica | 109 January 2024

ARTICLE - Thrombosis in pediatric ALL

Y. Zheng et al.

14. Deitelzweig SB, Lin J, Johnson BH, Schulman KL. Venous thromboembolism in the US: does race matter? J Thromb Thrombolysis. 2011;31(2):133-138. 15. Addo-Tabiri NO, Chudasama R, Vasudeva R, et al. Black patients experience highest rates of cancer-associated venous thromboembolism. Am J Clin Oncol. 2020;43(2):94-100. 16. Pui CH, Evans WE. Treatment of acute lymphoblastic leukemia. N Engl J Med. 2006;354(2):166-178. 17. Team RDC. R: a language and environment for statistical computing. R Foundation for Statistical Computing. version 3.5.1 ed. Vienna, Austria; 2019. 18. Buniello A, MacArthur JAL, Cerezo M, et al. The NHGRI-EBI GWAS Catalog of published genome-wide association studies, targeted arrays and summary statistics 2019. Nucleic Acids Res. 2019;47(D1):D1005-D1012. 19. Kheradpour P, Ernst J, Melnikov A, et al. Systematic dissection of regulatory motifs in 2000 predicted human enhancers using a massively parallel reporter assay. Genome Res. 2013;23(5):800-811. 20. Cunningham F, Allen JE, Allen J, et al. Ensembl 2022. Nucleic Acids Res. 2022;50(D1):D988-D995. 21. Consortium GT. The Genotype-Tissue Expression (GTEx) project. Nat Genet. 2013;45(6):580-585. 22. Jadaon MM, Dashti AA. HR2 haplotype in Arab population and patients with venous thrombosis in Kuwait. J Thromb Haemost. 2005;3(7):1467-1471. 23. De Stefano V, Chiusolo P, Paciaroni K, Leone G. Epidemiology of factor V Leiden: clinical implications. Semin Thromb Hemost.

1998;24(4):367-379. 24. Mizrahi T, Leclerc JM, David M, Ducruet T, Robitaille N. ABO group as a thrombotic risk factor in children with acute lymphoblastic leukemia: a retrospective study of 523 Patients. J Pediatr Hematol Oncol. 2015;37(5):e328-332. 25. Ohira T, Cushman M, Tsai MY, et al. ABO blood group, other risk factors and incidence of venous thromboembolism: the Longitudinal Investigation of Thromboembolism Etiology (LITE). J Thromb Haemost. 2007;5(7):1455-1461. 26. Gill JC, Endres-Brooks J, Bauer PJ, Marks WJ, Jr., Montgomery RR. The effect of ABO blood group on the diagnosis of von Willebrand disease. Blood. 1987;69(6):1691-1695. 27. Sliz E, Kalaoja M, Ahola-Olli A, et al. Genome-wide association study identifies seven novel loci associating with circulating cytokines and cell adhesion molecules in Finns. J Med Genet. 2019;56(9):607-616. 28. Budnik I, Brill A. Immune factors in deep vein thrombosis initiation. Trends Immunol. 2018;39(8):610-623. 29. Bettiol A, Galora S, Argento FR, et al. Erythrocyte oxidative stress and thrombosis. Expert Rev Mol Med. 2022;24:e31. 30. Sheth BK, Yakkanti R, Ravipati K, Arif B, Castro G, Hernandez V. Red blood cell transfusions and risk of postoperative venous thromboembolism. J Am Acad Orthop Surg. 2022;30(13):e919-e928. 31. Venter C, Bezuidenhout JA, Laubscher GJ, et al. Erythrocyte, platelet, serum ferritin, and P-selectin pathophysiology implicated in severe hypercoagulation and vascular complications in COVID-19. Int J Mol Sci. 2020;21(21):8234.

Haematologica | 109 January 2024

ARTICLE - Acute Myeloid Leukemia

Effect of age and treatment on predictive value of measurable residual disease: implications for clinical management of adult patients with acute myeloid leukemia Francesco Mannelli,1,2 Matteo Piccini,1 Sara Bencini,3 Giacomo Gianfaldoni,1 Benedetta Peruzzi,3 Roberto Caporale,3 Barbara Scappini,1 Laura Fasano,1 Elisa Quinti,1 Gaia Ciolli,1 Andrea

Correspondence: F. Mannelli

Pasquini, Francesca Crupi, Sofia Pilerci, Fabiana Pancani, Leonardo Signori, Danilo

francesco.mannelli@unifi.it

Tarantino,1,2 Chiara Maccari,1,2 Vivian Paradiso,1 Francesco Annunziato,3 Paola Guglielmelli1,2 and

Received: Accepted: Early view:

1,2

Alessandro M. Vannucchi1,2

March 24, 2023. June 15, 2023. June 22, 2023.

SOD Ematologia, Università di Firenze, AOU Careggi, 2Centro Ricerca e Innovazione Malattie

https://doi.org/10.3324/haematol.2023.283196

Mieloproliferative (CRIMM), AOU Careggi and 3Centro Diagnostico di Citofluorimetria e

Immunoterapia, AOU Careggi, Firenze, Italy

Abstract Measurable residual disease (MRD) is a powerful predictor of outcome in acute myeloid leukemia. In the early phases of treatment, MRD refines initial disease risk stratification and is used for the allocation to allogeneic transplant. Despite its well-established role, a relatively high fraction of patients eventually relapses albeit achieving MRDneg status. The aim of this work was to assess specifically the influence of baseline features and treatment intensity on the predictive value of an MRDneg status, particularly focusing on MRD2, measured after two consecutive chemotherapy cycles. Among baseline features, younger MRD2neg patients (<55 years) had a significantly longer disease-free survival (median not reached) compared to their older counterparts (median 25.0 months, P=0.013, hazard ratio=2.08). Treatment intensity, specifically the delivery of a high dose of cytarabine in induction or first consolidation, apparently had a pejorative effect on the outcome of MRD2neg patients compared to standard dose (P=0.048, hazard ratio=1.80), a finding also confirmed by the analysis of data extracted from the literature. The combination of age and treatment intensity allowed us to identify categories of patients, among those who reached a MRD2neg status, characterized by significantly different disease-free survival rate. Our data showed that variables such as age and intensity of treatment administered can influence the predictive value of MRD in patients with acute myeloid leukemia. In addition to underscoring the need for further improvement of MRD analysis, these findings call for a reasoned application of MRD data, as currently available, to modulate consolidation therapy on adequately estimated relapse rates.

Introduction The determination of measurable residual disease (MRD) relies on the detection of leukemic cells below the morphology-based threshold that defines complete remission.1 The persistence of MRD is a powerful predictor of disease relapse and poor outcome in acute myeloid leukemia (AML),2-6 a role formally recognized in the European LeukemiaNet (ELN) recommendations, which defines a specific response category of complete response (CR) without MRD.1 The prognostic impact of MRD measurement in AML has been demonstrated regardless of the method applied to assess it,6 although the ELN consensus suggests employing a molecular probe for core binding factor (CBF)-related and NPM1-mutated AML, and a multiparameter flow cytometry

(MFC) approach in all other subgroups.7 In the early phases of treatment, MRD assessment refines initial disease risk stratification and is primarily used for the allocation of eligible patients to allogeneic hematopoietic stem cell transplant (HSCT), generally reserved for categories deemed to be at high risk of relapse.8 From this perspective, the most informative time-point is set after two cycles of intensive chemotherapy (i.e., MRD2), as it has been shown to yield the greatest predictive value on outcome.3,5,9 Such a strategy is stated as standard by ELN guidelines7,10 and is increasingly becoming the basis for clinical decision-making in AML management.9,11,12 Despite its indisputable role, the prognostic value of MRD may be influenced by the clinical context, including age range13 and genetic profile,14 as well as by pre-MRD intensity

Haematologica | 109 January 2024

ARTICLE - The effect of age and treatment on MRD

F. Mannelli et al.

board (protocol number: 2013/0021560). Written informed consent was obtained from study patients in accordance with the Declaration of Helsinki. Enrollment criteria required were (i) intensive treatment, (ii) CR achievement after first induction therapy, and (iii) availability of MRD data after induction (MRD1) and first consolidation cycle (MRD2).

of treatment. In fact, the predictive value of MRD2 reflects the degree of chemosensitivity of leukemic cells: its positioning after two chemotherapy cycles provides a reliable estimation of the likelihood that further chemotherapy may lead to disease eradication. Overall, the reliability of MRD2 largely depends on previously delivered treatment, including the dosages of the backbone agents (cytarabine and anthracycline), the use of a third drug alongside (etoposide, fludarabine, gemtuzumab ozogamicin and, more recently for FLT3-mutant cases, an FLT3 inhibitor), as well as the regimen schedule (standard, double induction, sequential). Therefore, MRD2 status somehow captures information on disease sensitivity to the treatment actually delivered. Furthermore, MRD is measured after achievement of CR following induction therapy, and the intensity of chemotherapy in the first cycle is known to influence the rate of CR.15-17 In other words, different treatment intensity in the first cycle contributes to shape the population of patients in which MRD is later assessed starting from CR achievement onward, and this might affect the ability of MRD itself to estimate outcome. Although the influence of such an effect is systematically stated in guidelines,7 its impact has not been fully evaluated and is generally not taken into account in clinical decision-making. The aim of this work was to assess specifically the influence of baseline features (age, white blood cell count, ELN stratification), and treatment intensity on the predictive value of MRD, focusing specifically on MRD2-negative (MRD2neg) status because of the key clinical information it can provide. For this purpose, we retrospectively analyzed our database and then interrogated the available literature to validate our findings.

Measurable residual disease study Multiparameter flow cytometry MFC study files reporting individual leukemia-aberrant immunephenotype (LAIP) profiles were acquired locally according to pre-defined standard operating procedures. LAIP profiles were defined by antigen expression on blasts from fresh diagnostic bone marrow samples (or peripheral blood in the case of punctio sicca). Standard MFC methodologies for LAIP definition are detailed in the Online Supplementary File and reported elsewhere.19 MRD was expressed as the percentage of LAIP+ cells on CD45+ cells. MRD was defined as positive for any detection ≥0.1%, in accordance with ELN recommendations.7,20 Polymerase chain reaction-based assessment Sensitive real-time quantitative-polymerase chain reaction assays (RQ-PCR) were used for detection of MRD in patients with a suitable molecular probe (RUNX1T1 and CBFB-MYH11 transcripts and NPM1 gene mutations). RQ-PCR was performed following the Europe Against Cancer (EAC) program recommendations21 with a sensitivity of 10-5. MRD was defined as positive for any detection ≥0.01%. Standard PCR methodologies are detailed in the Online Supplementary File. Definitions CR, disease-free survival (DFS), and overall survival (OS) were defined according to standard criteria.1 As regards prognostic stratification, the criteria used for the therapeutic decision-making across different time periods, and particularly allocation to HSCT, are specified in the Online Supplementary File. In post hoc analysis, patients were stratified, according to the availability of genetic data, by Medical Research Council (MRC) criteria,22 and ELN version 2010-2017.1,23

Methods Patients Patients entering the study had a diagnosis of non-promyelocytic AML, based on morphological, immunophenotypic, and molecular criteria. Once eligible for intensive chemotherapy, they were consecutively assigned to treatment based on the availability of a clinical trial at the time of diagnosis and on local institutional treatment choices as previously reported18 and detailed in the Online Supplementary Data. For the purpose of the study, the treatment intensity of the first two chemotherapy cycles was categorized according to the dosage of cytarabine (ARA-C): standard dose (SDAC) included “3+7” and ”3+7”-like regimens sharing a single-dose infusion of 100-200 mg/m2 and a cumulative dose of up to 1,400 mg/m2 per cycle; high dose (HDAC) included several schedules sharing a single dose of at least 1,000 mg/m2 and a cumulative dose of at least 6,000 mg/m2 (Online Supplementary Data). CPX-351 was categorized as SDAC. The study was approved by the local institutional review

Analysis of literature We carried out a search in PubMed for articles published between 2000 and 2021 by filtering for keywords (AML, acute myeloid leukemia, or acute myelogenous leukemia, and MRD, minimal residual disease, or measurable residual disease). The analytical method, the criteria for paper selection and the extracted data are detailed in the Online Supplementary Materials and Methods S3, Online Supplementary Table S1. Based on this assessment, a total of 33 articles were selected. We then extracted survival data from Kaplan-Meier curves by using commercial graph digitizer software (Digitizelt, version 2.1; Bormisoft) and applying a previously published algorithm to reconstruct survival data

Haematologica | 109 January 2024

ARTICLE - The effect of age and treatment on MRD

F. Mannelli et al.

for MRDpos and MRDneg cases.24 The main characteristics of treatment in the first two cycles (drugs, ARA-C dosage, schedule) were obtained for each report and tabulated as in Online Supplementary Table S2.

patients are detailed in Table 1. In addition to baseline features, we classified patients according to treatment intensity (SDAC-based and HDAC-containing regimens) and MRD status after induction (MRD1) and first consolidation (MRD2); patients receiving HDAC in at least one of the first two cycles, in induction (cycle 1) or consolidation (cycle 2) (n=124, 63.9%), were compared with those receiving SDAC in both cycles (n=70, 36.1%) (Figure 1).

Statistical methods All statistical analyses were performed using R version 3.5.0 and SPPS version 26. Statistical methods are detailed in the Online Supplementary File (Materials and Methods S5).

Measurable residual disease study For CBF-related and NPM1-mutated AML, the categorization according to MRD was based on PCR data, available for all 33 CBF-related cases, and for 62/80 (77.5%) NPM1-mutated ones. In all other cases, MFC was used. After induction, 110 (56.7%) patients were defined as MRD1neg and 84 (43.3%) as MRD1pos. After first consolidation, a total of 121 (62.4%) cases reached MRD2neg status, whereas 73 (37.6%) patients resulted MRD2pos (Figure 1). MRD1 and MRD2 status showed an impact on DFS (Online Supplementary Figures S2 and 3), with MRD2 being more effective in discriminating outcome

Results Characteristics of patients and treatment flow From April 2004 to January 2022, 194 patients affected by AML met the inclusion criteria. Considering the large enrollment period, we carried out an analysis of OS to check for an influence by the year of diagnosis (Online Supplementary Figure S1): no relevant impact on survival emerged. The clinical and biological characteristics of our

Table 1. Characteristics of patients according to treatment group after induction and the first consolidation cycle. Overall N=194

SDAC N=70 (36.1%)

HDAC N=124 (63.9%)

Age in years, median (range)

55 (16-73)

55 (22-70)

53 (16-73)

0.75

WBC, x109/L, median (range)

14.8 (0.6-435.0)

14.9 (0.6-191.0)

13.6 (0.9-435.0)

0.90

Hb, g/dL, median (range)

9.2 (3.9-14.9)

9.2 (5.4-14.5)

9.0 (3.9-14.9)

0.25

Platelets, x109/L, median (range)

53 (1-281)

47 (10-152)

57 (1-281)

0.053

Peripheral blasts, %, median (range)

53 (0-100)

58 (0-100)

52 (0-98)

0.30

Secondary AML, N (%)

12 (6.2%)

3 (4.3)

9 (7.3)

0.54

Karyotype, N (%) Favorable Normal Intermediate, non-normal Adverse Lack of growth

33 (17.0) 110 (56.7) 26 (13.4) 15 (7.7) 10 (5.2)

13 (18.6) 41 (58.6) 9 (12.9) 2 (2.9) 5 (7.1)

20 (16.1) 69 (55.6) 17 (13.7) 13 (10.5) 5 (4.0)

0.69 0.76 1.0 0.09 0.50

Molecular genetics, N (%) NPM1-mutated FLT3-ITD CEBPA-bZIP

80 (41.2) 41 (21.1) 10 (5.2)

39 (55.7) 16 (22.9) 4 (5.7)

47 (37.9) 25 (20.2) 6 (4.8)

0.23 0.71 0.74

ELN 2010 risk group, N (%) Favorable Intermediate-1 Intermediate-2 Adverse

90 (46.4) 80 (41.2) 8 (4.1) 16 (8.2)

39 (55.7) 27 (38.6) 2 (2.9) 2 (2.9)

51 (41.1) 53 (42.7) 6 (4.8) 14 (11.3)

0.06 0.64 0.71 0.06

ELN 2017 risk group, N (%) Favorable Intermediate Adverse Not assessable

102 (52.5) 62 (32.0) 24 (12.4) 6 (3.1)

41 (58.7) 19 (27.1) 5 (7.1) 5 (7.1)

61 (49.2) 43 (34.7) 19 (15.3) 1 (0.8)

0.23 0.33 0.11

0.38

0.01

0.05

Differences between treatment groups were evaluated using the Mann-Whitney test for continuous variables and Fisher exact test or χ2 for categorical variables. P values <0.05 are statistically significant. SDAC: standard-dose cytarabine (both in induction and consolidation cycle); HDAC: high-dose cytarabine (at least in one cycle, induction or first consolidation); WBC: white blood cells; Hb: hemoglobin; AML: acute myeloid leukemia; ITD: internal tandem duplication; ELN: European LeukemiaNet.

Haematologica | 109 January 2024

ARTICLE - The effect of age and treatment on MRD

F. Mannelli et al.

categories (C-index 0.61 vs. 0.547, P=0.009). MRD2 status also significantly influenced OS (P=0.0014, hazard ratio [HR]=2.08) (Online Supplementary Figure S3). Analysis according to baseline features Due to its greater informativeness for prognosis, further analyses were focused on MRD2. We searched for any interaction between the prognostic impact of MRD2 and baseline features. First, patients were stratified for age: the observed median age was 55 years. By means of a receiving operating characteristic curve analysis using relapse as a binary endpoint, we confirmed an age threshold of 55.5 years as having the best performance (Youden score=0.224) on the basis of the combination of a sensitivity and specificity of 0.564 and 0.66, respectively. Accordingly, we separated patients for their age <55 years and ≥55 years. We also categorized patients for gender, white blood cell count (<30x109/L and ≥30x109/L), and ELN category (favorable and intermediate). ELN adverse category was too limited in size (n=24, of whom 13 MRD2neg) for such an analysis. Significant differences between MRD2neg

and MRD2pos patients were found as regarded DFS and OS in subgroups defined by age, white blood count, and ELN category (Online Supplementary Figures S4-S6) and within the female patient category (Online Supplementary Figure S6). In male patients, we observed a non-significant trend in favor of the MRD2neg group (Online Supplementary Figure S7). Since we observed a still remarkable rate of relapse in the MRD2neg group (DFS rate at 2 years, 60.2% vs. 32.1% in MRD2pos patients, P=0.00013), we searched for variables having an impact on DFS by comparing the outcome of MRD2neg patients in different disease categories. We found that younger (i.e., <55 years) MRD2neg patients showed a significantly longer DFS than their older (i.e., ≥55 years) counterparts (P=0.013, HR=2.08) (Online Supplementary Figure S6), while no significant differences emerged for gender- (P=0.11), white blood count- (P=0.29) and ELN- (P=0.1) related strata (Online Supplementary Figure S8). Analysis according to treatment groups We then analyzed the baseline characteristics and outcomes of patients stratified according to treatment intensity (SDAC

Figure 1. Kinetics of measurable residual disease according to timepoint of assessment. Flow of data for measurable residual disease (MRD) status (negative, MRDneg; positive, MRDpos) at the post-induction (MRD1) and post-consolidation (MRD2) timepoints. The chemotherapy regimens and the number of patients treated with each scheme are listed for induction (pink panel) and consolidation (blue panel) cycle. The cycles were classified according to the dose of cytarabine as standard dose (SDAC) or high dose (HDAC). Treatment details for each chemotherapy scheme are provided in the Online Supplementary File. GO: gemtuzumab ozogamicin; Mido: midostaurin; CPX: CPX-351; ICE: idarubicin, cytarabine, etoposide; HDS: high-dose sequential; Ida: idarubicin; HDAC 10: high-dose cytarabine 10 g/m2; HDAC 1,3,5: high-dose cytarabine days 1, 3, and 5; DIA: intermediate-dose cytarabine; PCR: polymerase chain reaction; LAIP: leukemia-associated immunophenotype.

Haematologica | 109 January 2024

ARTICLE - The effect of age and treatment on MRD

F. Mannelli et al.

vs. HDAC). The only statistically significant difference was younger age in patients treated with HDAC-containing induction (47 vs. 57 years; P<0.0001) (Online Supplementary Table S3). No significant difference in outcome was observed according to cytarabine dose in induction (Online Supplementary Figure S9). We further stratified patients according to the cumulative effect of induction and first consolidation, i.e., those who received at least one HDAC-containing cycle versus those

treated only with SDAC. We only observed a trend towards enrichment in 2010 ELN adverse-risk categories within the HDAC-treated group (Table 1). However, no significant outcome benefit was highlighted for patients receiving HDAC during induction or consolidation (Online Supplementary Figure S9). Effects of treatment intensity on MRD2 predictive capability We searched for interactions between MRD2 status and

Figure 2. Analysis of survival according to measurable residual disease status after first consolidation in different treatment groups. (A, C) Disease-free survival and (B, D) overall survival according to measurable residual disease status after first consolidation (MRD2) in patients treated with standard-dose cytarabine in both induction and consolidation (A, B) or with at least one cycle including high-dose cytarabine (C, D). The curves of patients with MRD2neg status are depicted in blue; the curves of patients with MRD2pos status are depicted in red. Median survival estimates are reported in months. NR: not reached; HR: hazard ratio; 95% CI: 95% confidence interval; NA: not available.

Haematologica | 109 January 2024

ARTICLE - The effect of age and treatment on MRD

F. Mannelli et al.

outcome within different treatment groups. We observed statistically significant longer DFS in MRD2neg versus MRD2pos patients who received SDAC during induction and first consolidation (HR=3.87, P<0.0001, C-index=0.652) (Figure 2A). Contrariwise, MRD2 status failed to significantly resolve DFS in patients treated with at least one HDAC-containing cycle, in induction or consolidation (HR=1.60, P=0.066, C-index=0.585) (Figure 2C). Similar findings concerned the impact of MRD2 on OS, which was associated with a significant value in patients treated with SDAC (P<0.0001, HR=3.85, 95% confidence interval [95% CI]: 1.87-7.93), unlike what was observed for HDAC (P=0.35, HR=1.33, 95% CI: 0.73-2.41) (Figure 2B-D). The above differences were also maintained in the category of intermediate-risk karyotype: the favorable impact of MRD2neg status on DFS was more evident in patients receiving SDAC (HR=3.98, P=0.00078, C-index=0.65) than HDAC (HR=2.13, P=0.01, C-index=0.613) (Online Supplementary Figure S10), al-

though statistical significance was reached also in the latter case. Conversely, among MRD2pos patients, DFS was largely unaffected by treatment arm (median 11.6 and 11.4 months in SDAC- and HDAC-treated patients, respectively) and the worse performance of MRD2 in the HDAC group was mainly the consequence of a relatively high rate of relapse in MRD2neg patients. OS estimates followed the same pattern in the intermediate-risk karyotype tier (Online Supplementary Figure S10). Characteristics of MRD2neg patients according to treatment arm We then analyzed MRD2neg patients according to the intensity of the first two chemotherapy cycles. At baseline, there was a statistically significant enrichment of ELN adverse-risk patients in HDAC-treated (16.9% according to 2017 version) compared to SDAC-treated MRD2neg cases (2.0%, P=0.014) (Table 2). When comparing patients

Table 2. Characteristics of MRD2neg patients according to treatment group after induction and first consolidation cycle. MRD2neg overall N=121

MRD2neg SDAC N=50 (41.3%)

MRD2 HDAC N=71 (58.7%)

Age in years, median (range)

55 (22-73)

55 (22-69)

55 (22-73)

0.23

WBC, x109/L, median (range)

8.9 (0.6-380.0)

11.8 (0.6-191.0)

7.5 (0.9-380.0)

0.30

Hb, g/dL, median (range)

9.3 (3.9-14.1)

9.3 (5.4-13.3)

9.3 (3.9-14.1)

0.55

Platelets, x109/L, median (range)

56 (9-271)

47 (10-152)

71 (9-271)

0.03

Peripheral blasts, %, median (range)

47 (0-100)

57 (0-100)

40 (0-98)

0.06

Secondary AML, N (%)

10 (8.3)

7 (14.0)

3 (4.2)

0.09

Karyotype, N (%) Favorable Normal Intermediate, non-normal Adverse Lack of growth

13 (10.7) 72 (59.5) 19 (15.7) 9 (7.5) 8 (6.6)

8 (16.0) 31 (62.0) 6 (12.0) 1 (2.0) 4 (8.0)

5 (7.0) 41 (57.7) 13 (18.3) 8 (11.3) 4 (5.6)

0.14 0.71 0.45 0.07 0.72

Molecular genetics, N (%) NPM1-mutated FLT3-ITD CEBPA-bZIP

50 (41.3) 20 (16.5) 10 (8.3)

25 (50.0) 9 (18.0) 4 (8.0)

25 (35.2) 11 (15.5) 6 (8.5)

0.13 0.19 1.0

ELN 2010 risk group, N (%) Favorable Intermediate-1 Intermediate-2 Adverse

55 (45.5) 50 (41.3) 6 (5.0) 10 (8.2)

30 (60.0) 17 (34.0) 2 (4.0) 1 (2.0)

25 (35.2) 33 (46.5) 4 (5.6) 9 (12.7)

0.009 0.19 1.0 0.045

ELN 2017 risk group, N (%) Favorable Intermediate Adverse Not assessable

62 (51.3) 42 (34.7) 13 (10.7) 4 (3.3)

32 (64.0) 13 (24.0) 1 (2.0) 4 (10.0)

30 (42.2) 29 (40.8) 12 (16.9) 0 (-)

0.026 0.12 0.014 0.027

0.148

0.003

0.001

Differences between treatment groups were evaluated using the Mann-Whitney test for continuous variables and Fisher exact test or χ2 for categorical variables. P values <0.05 are statistically significant. MRD2: measurable residual disease assessed after induction therapy and first consolidation cycle; SDAC: standard-dose cytarabine (both in induction and consolidation cycle); HDAC: high-dose cytarabine (at least in one cycle, induction or first consolidation); WBC: white blood cells; Hb: hemoglobin; AML: acute myeloid leukemia; ITD: internal tandem duplication; ELN: European LeukemiaNet.

Haematologica | 109 January 2024

ARTICLE - The effect of age and treatment on MRD

F. Mannelli et al.

reaching MRD2neg status after SDAC or HDAC, significantly different DFS (P=0.048, HR=1.80, 95% CI: 1.0-3.23) and OS (P=0.049, HR=1.94, 95% CI: 0.99-3.8) were highlighted (Online Supplementary Figure S11). After being superimposable at 6 months (93.9% and 90.0% for SDAC- and HDAC-treated patients, respectively), DFS rates started to diverge at 12 months (83.4% vs. 71.5%) and were 70.6%

vs. 51.7% at 24 months. Similar trends in DFS were observed in HDAC- vs. SDAC-treated patients with intermediate-risk karyotype (P=0.092) and ELN 2017 (P=0.23) categories (Online Supplementary Figure S12).

Combined model including age and treatment intensity Based on the observed influence of age and treatment on

Figure 3. Analysis of major outcomes for patients with negative measurable residual disease after first consolidation according to the combined model. (A, C) Disease-free survival and (B, D) overall survival according to the combined model including age range and treatment intensity in patients with measurable residual disease negativity after first consolidation (MRD2neg): the curves of patients aged less than 55 years and treated with standard-dose cytarabine in both induction and consolidation (COMB_1) are depicted in blue. In the upper panels, the curves of patients aged more than or equal to 55 years and treated with high-dose cytarabine in at least one of the first two cycles (COMB_3) are depicted in red; the remaining patients (COMB_2) are depicted in gray. In the lower panels, COMB_2 and COMB_3 categories are grouped, and their survival curves depicted in red. Median survival estimates are reported in months. NR: not reached; HR: hazard ratio; 95% CI: 95% confidence interval; NA: not available.

Haematologica | 109 January 2024

ARTICLE - The effect of age and treatment on MRD

F. Mannelli et al.

the prognostic value of MRD2neg status, we combined these two variables and stratified MRD2neg patients accordingly. Twenty-four (out of 121, 19.8%) patients were younger than 55 years and reached MRD2 negativity after two SDAC-based cycles (COMB_1): they showed a DFS rate of 86.4% at 3 years. At the opposite, the group of 38 (31.4%) elderly patients who had received at least one HDAC-based cycle (COMB_3) had a median DFS of 22.2 months and a 3-year DFS rate of 46.6% (Figure 3A, B). The remaining patients (COMB_2) displayed an intermediate behavior, but with DFS largely superimposable to the latter group in the first 2 years after CR achievement (Figure 3A). After combining COMB_2 and _3 patients, survival estimates for DFS (P=0.0034, HR=4.13) and OS (P=0.0014, HR=7.36) were significantly different with respect to that of COMB-1 patients (Figure 3C, D). The significant effect of the model on DFS was maintained in intermediate-risk stratification tiers according to karyotype (P=0.013) (Online Supplementary Figure S13) and ELN 2017 category (P=0.016) (Online Supplementary Figure S13).

according to the intensity of treatment (SDAC- vs. HDACbased regimens). First, we carried out an analysis unselected for AML subset, and we observed a difference in DFS in favor of the SDAC-treated group (P=0.014) (Figure 4A). To adjust for the fact that CBF-related cases were restricted to the group receiving HDAC, we performed further analyses after excluding CBF-related cases (Figure 4B) and focusing on intermediate-risk karyotype (Figure 4C). In both analyses, the SDAC-treated category displayed longer DFS (P<0.0001 and P=0.0018, respectively), an observation consistent with the findings in our dataset.

Discussion MRD assessment is an essential tool for the management of patients with AML. In fact, it integrates the baseline and pre-treatment prognostic variables by conveying information on the sensitivity of the disease to treatment in the early phases of therapy.2-5,9,12,25 In terms of decision-making, the persistence of MRD after intensive treatment correlates with a very high risk of disease recurrence, and patients are immediately allocated to allogeneic HSCT.26 An undetectable MRD is far trickier to interpret and translate into a clinical decision. Although with some variability based on the methodology used, the rate of patients who relapse despite the achievement of an MRDneg status is not negligible, ranging from 20 to 40% in the first 2 years from CR.6,27 The reasons for the failure of MRDneg to predict for maintenance of CR in such a relevant fraction of patients are not fully understood. Relapse is supposed to derive from a minor population of leukemic cells that escape MRD detection, likely due to a combination of factors. These include technical issues in the MRD measurement, as supported by the improved prognostic performance obtained with the use of more sensitive molecular methods28-30 or by increasing the number of parameters and cells analyzed by MFC.8,10 Indeed, it has been shown that introduction of more stringent criteria for the limit of detection and quantitation resulted in a more accurate ability to predict patients’ outcome.31 Furthermore, residual AML cells might escape detection as the result of changes in their phenotypic profile or confinement in more immature cell compartments. To address these specific issue, some investigators have proposed advancements to standard approaches, aimed at extending MRD evaluation to deviations from physiological hematopoiesis (i.e., so-called “different from normal”)7 or by highlighting abnormalities in leukemic stem cells,32,33 but both strategies remain to be validated. Whatever the reasons, failure of MRDneg status to appropriately predict for maintenance of CR has relevant clinical consequences. Particularly when a relapse occurs in the first year after completion of treatment, it can be challenging to obtain a second response, outside the CBF-related subset.34

Allogeneic hematopoietic stem cell transplant Further analysis included censoring for patients who received HSCT. The differential effect of age and treatment intensity on DFS of MRD2neg patients was confirmed in this analysis, either as single variables (Online Supplementary Figure S14) or in the combined model (Online Supplementary Figure S15). A non-significant trend for a lower actual rate of HSCT in first CR was observed for MRD2neg (32/121, 26.4%) versus MRD2pos groups (28/73, 38.3%, P=0.108). We highlighted a similar pattern within MRD2neg patients for SDAC- (10/50, 20.0%) and HDAC-treated cases (22/71, 31.0%, P=0.212). Then, to estimate the benefit from HSCT in MRD2neg cases according to the combined model (age + treatment) category, we used the Mantel-Byar method and Simon-Makuch plots to correct for pre-HSCT events considering HSCT as a time-dependent variable. In this analysis of DFS, the protection from relapse was not significant in the COMB_1 group (P=0.64, HR=0.60) (Online Supplementary Figure S15), whereas it benefitted the COMB_2-3 patients (P=0.00433, HR=0.26) (Online Supplementary Figure S16). Analysis of literature Our data indicated an interaction between the predictive value of MRD2neg status and the intensity of treatment. We thus carried out an analysis of available literature to validate our findings. We selected a set of papers as detailed in the Methods section; studies selected for analysis of DFS in MRD2neg cases based on treatment intensity were processed as in conventional meta-analyses and extracted MRD data are summarized in a forest plot (Online Supplementary Data S5). Then we performed DFS analysis after merging extracted data annotated for MRD status after two chemotherapy cycles, AML subset and treatment schedule. We selected cases that resulted negative after two chemotherapy cycles (i.e., MRD2neg), stratifying patients

Haematologica | 109 January 2024

ARTICLE - The effect of age and treatment on MRD

F. Mannelli et al.

In these patients, a negative MRD status can thus misguide the treatment planning, including the timely decision to exploit allogeneic HSCT when the disease burden is low. Pending further methodological improvements (e.g., next-generation sequencing), the clinical application of current MRDneg data should take into an account those limitations, and efforts should be devoted to shape the settings in which MRDneg predictive value is robust enough or, contrariwise, relatively weak to influence important ther-

apeutic decisions. With such an aim in mind, we searched for interactions between MRD, baseline characteristics (age, gender, white cell count, ELN stratification) and the intensity of treatment in a series of AML patients from our Institution. We focused on the predictive value of MRD2neg status after two chemotherapy cycles (MRD2) in view of its greater capability to discriminate DFS in comparison with earlier assessment (i.e., after induction, MRD1), consistent with previous experiences.4,35

Continued on following page. Haematologica | 109 January 2024

ARTICLE - The effect of age and treatment on MRD

F. Mannelli et al.

Figure 4. Analysis of disease-free survival according to treatment intensity based on data extracted from literature. Disease-free survival in patients with negative measurable residual disease status after first consolidation (MRD2neg) according to treatment intensity, standard-dose cytarabine (SDAC, blue curve) versus high-dose cytarabine (HDAC, red curve) in different acute myeloid leukemia subsets: (A) unselected; (B) after exclusion of CBF-related acute myeloid leukemia; (C) intermediate-risk karyotype. Next to each Kaplan-Meier curve, the papers from which data were extracted for each analysis are listed by first author’s name and year of publication.

Among baseline features, we found that younger MRD2neg patients (i.e., aged less than 55 years) had a significantly longer DFS (median not reached) compared to their elderly counterparts (median 25.0 months, P=0.013, HR=2.08, 95% CI: 1.15-3.77). Treatment intensity, specifically the delivery of HDAC in either induction or first consolidation, apparently had a pejorative effect on the outcome of MRD2neg patients. This finding can likely be explained by the progressive selection of patients with unfavorable characteristics exerted by HDAC, as supported by the enrichment in ELN 2017 adverse-risk category for HDAC-treated in comparison to SDAC-treated patients from after induction (19.0% vs. 11.2, respectively, P=0.46) to after consolidation (16.9% vs. 2.0%, P=0.014). In other words, one might surmise that the early delivery of HDAC could promote CR attainment, eventually making eligible for MRD assessment a category of patients featured by inferior chemosensitivity, who would otherwise have resulted refractory (or MRDpos) when treated with SDAC. HDAC could then transiently mask intrinsically chemorefractory disease, thereby correlating with a higher relapse rate than that with SDAC, despite both treatments resulting in MRD2neg status. Our findings on the effect of treatment on the significance of MRD are consistent with a previous report by Maurillo et al: aiming to use MRD as a biomarker for optimal ARA-C dosing, they also described lower reliability of MRD in discriminating prognosis in patients treated with HDAC compared to SDAC.36

To validate our observations, we selected and interrogated the available literature on the interaction between treatment effect and MRD2neg status. The results of this analysis confirmed the pattern in our patients, with longer DFS for those receiving SDAC in induction and first consolidation (Figure 4). The combination of age and treatment effects allowed us to identify categories of patients with remarkably different rates of relapse, in spite of the achievement of MRD2 negativity. The group of elderly, HDAC-receiving (COMB_2-3) patients displayed a DFS rate of around 50% at 3 years. This figure challenges the usual implications of MRDneg status, as current ELN guidelines recommend consolidation with allogeneic HSCT as the preferred post-CR option for patients with an estimated relapse risk exceeding 35-40%.1,37 Of note, this category of patients showed the greatest benefits from the delivery of HSCT in first CR, as emerging from a specific analysis using the Mantel-Byar method (Online Supplementary Figure S16). This finding is consistent with the established influence of disease burden, usually estimated by MRD, on the efficacy of the antileukemic effect of HSCT.26,38,39 Indeed, it is plausible that, although both characterized by MRD2neg status, COMB_1 patients achieved a more profound level of response than COMB_2-3 patients, likely explaining the difference in outcome in the post-HSCT setting. The longer DFS for SDAC- vs. HDAC-treated MRD2neg patients remained in the intermediate-risk tier, as assessed by karyotype and ELN (Online Supplementary Figure S13), the

Haematologica | 109 January 2024

ARTICLE - The effect of age and treatment on MRD

F. Mannelli et al.

category in which MRD2 assessment is generally employed for the clinical management of AML patients. These data emphasize some of the current limitations of MRD assessment. Although MRD represents the strongest predictive parameter for long-term survival overall, in some settings it actually lacks the robustness needed to support key clinical decisions. In particular, the translation of an MRD2neg status into a consolidation program including or omitting HSCT should take into account other contributing variables. To this regard, we herein provide evidence that age and treatment intensity may help to effectively delineate settings (younger age, SDAC-treated) in which the application of MRD2neg data to support clinical decisions is justified based on its correlation with an excellent outcome, unlike in others instances (older patients who were HDAC-treated) in which it did not add to the baseline stratification and should not be used as a key decision-making parameter. As with the dose of ARA-C, other treatment variables (e.g., gemtuzumab ozogamicin, FLT3 inhibitors) may also have an effect on MRD, an issue that deserves to be evaluated in large series of patients. Our work has some limitations that should be acknowledged. We recognize the limits of a retrospective study with a long enrollment period, with changes in risk assessment and treatment allocation (in particular with regards to HSCT), although we checked for an impact of year of diagnosis on OS (Online Supplementary Figure S1). Another point concerns the limited number of patients in the adverse-risk category according to karyotype and ELN, preventing specific analyses. This finding is clearly due to the selection of patients who had a response to chemotherapy, at least initially, and in any case represents the disease subset in which MRD is generally not crucial in clinical decisions. The lack of validation of our results in an independent cohort is a further weakness that we tried to address with the analysis of data extracted from the available literature, which confirmed our findings on

the impact of treatment. In conclusion, we showed a variable reliability of MRD in different AML settings, as defined by age and intensity of treatment. In addition to underscoring the need for further improvement of MRD approaches, these findings call for a reasoned application of MRD information, as currently available, to modulate consolidation therapy on adequately estimated relapse rates. Disclosures No conflicts of interest to disclose. Contributions FM designed and performed the research, analyzed, and interpreted the data, and wrote the manuscript. MP performed the research, analyzed, and interpreted the data. SB performed flow cytometric analysis and collected the data. BP, RC performed flow cytometric analysis. GG, BS, LS, EQ, GC, FC, SP, and VP collected clinical data. FP, LS, CM, and PG performed molecular analysis. FA and AMV interpreted the data and critically reviewed the manuscript. Acknowledgments This work is dedicated to the memory of Dr. Franco Leoni, for his invaluable mentorship in the care of patients with acute leukemias. Funding The work was supported by AIRC 5x1000 call “Metastatic disease: the key unmet need in oncology” to MYNERVA project, #21267 (MYeloid NEoplasms Research Venture AIRC) and by Legato Zottola. A detailed description of the MYNERVA project is available at http://www.progettomynerva.it Data-sharing statement The data that support the findings of this study are available on request from the corresponding author.

References 1. Döhner H, Wei AH, Appelbaum FR, et al. Diagnosis and management of AML in adults: 2022 ELN recommendations from an international expert panel. Blood. 2022;140(12):1345-1377. 2. Miguel JFS, Martínez A, Macedo A, et al. Immunophenotyping investigation of minimal residual disease is a useful approach for predicting relapse in acute myeloid leukemia patients. Blood. 1997;90(6):2465-2470. 3. Venditti A, Buccisano F, Poeta DG, et al. Level of minimal residual disease after consolidation therapy predicts outcome in acute myeloid leukemia. Blood. 2000;96(12):3948-3952. 4. Terwijn M, Putten WL van, Kelder A, et al. High prognostic impact of flow cytometric minimal residual disease detection in acute myeloid leukemia: data from the HOVON/SAKK AML 42A study. J Clin Oncol. 2013;31(31):3889-3897. 5. Ivey A, Hills RK, Simpson MA, et al. Assessment of minimal

residual disease in standard-risk AML. N Engl J Med. 2016;374(5):422-433. 6. Short NJ, Zhou S, Fu C, et al. Association of measurable residual disease with survival outcomes in patients with acute myeloid leukemia. JAMA Oncol. 2020;6(12):1890-1899. 7. Heuser M, Freeman SD, Ossenkoppele GJ, et al. 2021 update measurable residual disease in acute myeloid leukemia: European LeukemiaNet Working Party Consensus Document. Blood. 2021;138(26):2753-2767. 8. Buccisano F, Maurillo L, Principe MID, et al. Minimal residual disease as a biomarker for outcome prediction and therapy optimization in acute myeloid leukemia. Expert Rev Hematol. 2018;11(4):307-313. 9. Freeman SD, Hills RK, Virgo P, et al. Measurable residual disease at induction redefines partial response in acute myeloid

Haematologica | 109 January 2024

ARTICLE - The effect of age and treatment on MRD

F. Mannelli et al.

leukemia and stratifies outcomes in patients at standard risk without NPM1 mutations. J Clin Oncol. 2018;36(15):1486-1497. 10. Tettero JM, Freeman S, Buecklein V, et al. Technical aspects of flow cytometry-based measurable residual disease quantification in acute myeloid leukemia: experience of the European LeukemiaNet MRD Working Party. Hemasphere. 2021;6(1):e676. 11. Zhu HH, Zhang HX, Qin ZY, et al. MRD-directed risk stratification treatment may improve outcomes of t(8;21) AML in the first complete remission: results from the AML05 multicenter trial. Blood. 2013;121(20):4056-4062. 12. Venditti A, Piciocchi A, Candoni A, et al. GIMEMA AML1310 trial of risk-adapted, MRD-directed therapy for young adults with newly diagnosed acute myeloid leukemia. Blood. 2019;134(12):935-945. 13. Buccisano F, Maurillo L, Piciocchi A, et al. Minimal residual disease negativity in elderly patients with acute myeloid leukemia may indicate different postremission strategies than in younger patients. Ann Hematol. 2015;94(8):1319-1326. 14. Jentzsch M, Grimm J, Bill M, et al. Clinical value of the measurable residual disease status within the ELN2017 risk groups in AML patients undergoing allogeneic stem cell transplantation. Am J Hematol. 2021;96(7):E237-E239. 15. Burnett AK, Goldstone A, Hills RK, et al. Curability of patients with acute myeloid leukemia who did not undergo transplantation in first remission. J Clin Oncol. 2013;31(10):1293-1301. 16. Castaigne S, Chevret S, Archimbaud E, et al. Randomized comparison of double induction and timed-sequential induction to a “3 + 7” induction in adults with AML: long-term analysis of the Acute Leukemia French Association (ALFA) 9000 study. Blood. 2004;104(8):2467-2474. 17. Bassan R, Intermesoli T, Masciulli A, et al. Randomized trial comparing standard vs sequential high-dose chemotherapy for inducing early CR in adult AML. Blood Adv. 2019;3(7):1103-1117. 18. Mannelli F, Bencini S, Piccini M, et al. Multilineage dysplasia as assessed by immunophenotype in acute myeloid leukemia: a prognostic tool in a genetically undefined category. Cancers. 2020;12(11):3196. 19. Mannelli F, Gianfaldoni G, Bencini S, et al. Early peripheral blast cell clearance predicts minimal residual disease status and refines disease prognosis in acute myeloid leukemia. Am J Hematol. 2020;95(11):1304-1313. 20. Schuurhuis GJ, Heuser M, Freeman S, et al. Minimal/measurable residual disease in AML: a consensus document from the European LeukemiaNet MRD Working Party. Blood. 2018;131(12):1275-1291. 21. Gabert J, Beillard E, van der Velden VHJ, et al. Standardization and quality control studies of ‘real-time’ quantitative reverse transcriptase polymerase chain reaction of fusion gene transcripts for residual disease detection in leukemia – a Europe Against Cancer program. Leukemia. 2003;17(12):2318-2357. 22. 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 younger adult patients treated in the United Kingdom Medical Research Council trials. Blood. 2010;116(3):354-365. 23. 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. 24. Guyot P, Ades A, Ouwens MJ, Welton NJ. Enhanced secondary

analysis of survival data: reconstructing the data from published Kaplan-Meier survival curves. BMC Med Res Methodol. 2012;12(1):9. 25. Balsat M, Renneville A, Thomas X, et al. Postinduction minimal residual disease predicts outcome and benefit from allogeneic stem cell transplantation in acute myeloid leukemia with NPM1 mutation: a study by the Acute Leukemia French Association group. J Clin Oncol. 2016;35(2):185-193. 26. Buccisano F, Maurillo L, Piciocchi A, et al. Pre-transplant persistence of minimal residual disease does not contraindicate allogeneic stem cell transplantation for adult patients with acute myeloid leukemia. Bone Marrow Transpl. 2017;52(3):473-475. 27. Paiva B, Vidriales M-B, Sempere A, et al. Impact of measurable residual disease by decentralized flow cytometry: a PETHEMA real-world study in 1076 patients with acute myeloid leukemia. Leukemia. 2021;35(8):2358-2370. 28. Pastore F, Levine RL. Next-generation sequencing and detection of minimal residual disease in acute myeloid leukemia: ready for clinical practice? JAMA. 2015;314(8):778-780. 29. 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. 30. Jongen-Lavrencic M, Grob T, Hanekamp D, et al. Molecular minimal residual disease in acute myeloid leukemia. N Engl J Med 2018;378(13):1189-1199. 31. Buccisano F, Palmieri R, Piciocchi A, et al. Clinical relevance of an objective flow cytometry approach based on limit of detection and limit of quantification for measurable residual disease assessment in acute myeloid leukemia. A post-hoc analysis of the GIMEMA AML1310 trial. Haematologica. 2022;107(12):2823-2833. 32. Li S-Q, Xu L-P, Wang Y, et al. An LSC-based MRD assay to complement the traditional MFC method for prediction of AML relapse: a prospective study. Blood. 2022;140(5):516-520. 33. Zeijlemaker W, Kelder A, Oussoren-Brockhoff YJM, et al. A simple one-tube assay for immunophenotypical quantification of leukemic stem cells in acute myeloid leukemia. Leukemia. 2016;30(2):439-446. 34. Breems DA, Putten WLJV, Huijgens PC, et al. Prognostic index for adult patients with acute myeloid leukemia in first relapse. J Clin Oncol. 2005;23(9):1969-1978. 35. Maurillo L, Buccisano F, Principe M, et al. Toward optimization of postremission therapy for residual disease–positive patients with acute myeloid leukemia. J Clin Oncol. 2008;26(30):4944-4951. 36. Maurillo L, Buccisano F, Piciocchi A, et al. Minimal residual disease as biomarker for optimal biologic dosing of ARA-C in patients with acute myeloid leukemia. Am J Hematol. 2015;90(2):125-131. 37. Cornelissen JJ, Blaise D. Hematopoietic stem cell transplantation for patients with AML in first complete remission. Blood. 2016;127(1):62-70. 38. Walter RB, Gooley TA, Wood BL, et al. Impact of pretransplantation minimal residual disease, as detected by multiparametric flow cytometry, on outcome of myeloablative hematopoietic cell transplantation for acute myeloid leukemia. J Clin Oncol. 2011;29(9):1190-1197. 39. Hourigan CS, Dillon LW, Gui G, et al. Impact of conditioning intensity of allogeneic transplantation for acute myeloid leukemia with genomic evidence of residual disease. J Clin Oncol. 2019;38(12):1273-1283.

Haematologica | 109 January 2024

ARTICLE - Acute Myeloid Leukemia

Fludarabine, cytarabine, and idarubicin with or without venetoclax in patients with relapsed/refractory acute myeloid leukemia Rabia Shahswar,1 Gernot Beutel,1 Razif Gabdoulline,1 Adrian Schwarzer,1 Arnold Kloos,1 Christian Koenecke,1 Michael Stadler,1 Gudrun Gohring,2 Yvonne Lisa Behrens,2 Zhixiong Li,1 Louisa-Kristin Dallmann,1 Piroska Klement,1 Catherin Albert,1 Martin Wichmann,1 Yasmine Alwie,1 Axel Benner,3 Maral Saadati,4 Arnold Ganser,1 Felicitas Thol1 and Michael Heuser1 1

Department of Hematology, Hemostasis, Oncology, and Stem Cell Transplantation, Hannover Medical School, Hannover; 2Department of Human Genetics, Hannover Medical School, Hannover; 3DKFZ German Cancer Research Center, Heidelberg and 4Saadati Solutions, Ladenburg, Germany

Correspondence: M. Heuser heuser.michael@mh-hannover.de Received: Accepted: Early view:

February 8, 2023. July 13, 2023. July 20, 2023.

Abstract Treatment options for relapsed and refractory acute myeloid leukemia patients (R/R AML) are limited. This retrospective cohort study compares safety and efficacy of fludarabine, cytarabine, and idarubicin (FLA-IDA) without or with venetoclax (FLAVIDA) in patients with R/R AML. Thirty-seven and 81 patients received one course FLA-IDA with or without a 7-day course of venetoclax, respectively. The overall response rate (ORR) was significantly higher in FLAVIDA compared to FLAIDA-treated patients (78% vs. 47%; P=0.001), while measurable residual disease was negative at a similar proportion in responding patients (50% vs. 57%), respectively. Eighty-one percent and 79% of patients proceeded to allogeneic hematopoietic cell transplantation or donor lymphocyte infusion after FLAVIDA and FLA-IDA, respectively. Event-free and overall survival were similar in FLAVIDA- and FLA-IDA-treated patients. Refractory patients could be salvaged more successfully after FLA-IDA compared to FLAVIDA pretreatment. Neutrophil and platelet recovery times were similar in the venetoclax and the control group. In conclusion, short-term venetoclax in combination with FLA-IDA represents an effective treatment regimen in R/R AML identifying chemosensitive patients rapidly and inducing measurable residual disease-negative remission in a high proportion of R/R AML patients.

venetoclax combined with HMA is approved for newly diagnosed patients with AML who are 75 years old, or unfit for intensive chemotherapy providing a new robust standard of care option for this frail population.10 Subgroup biomarker analyses identified an association between NPM1-, IDH1- or IDH2-mutated AML and favorable response, whereas outcomes were similar in patients irrespective of FLT3-internal tandem duplication (ITD) mutational status.10,13,14 Further analyses suggested that outcomes of patients with poor-risk cytogenetics in the absence of mutated TP53 were similar to those with intermediate risk when treated with azacitidine plus venetoclax.15 In order to improve response to induction or salvage chemotherapy, venetoclax has been recently combined with intensive chemotherapy. The addition of venetoclax to cytarabine and idarubicin (FLAVIDA) or FLAG-IDA induced remarkable response rates (composite CR [CRc] of 97% and 90%, respectively) and promising OS (median not reached [NR] and 11.2 months, respectively) in newly diagnosed AML patients.16,17

Introduction Primary refractory disease is found in 10-20% of younger and 50% of older acute myeloid leukemia (AML) patients after two courses of intensive induction therapy, and 5070% of patients who obtain complete remission (CR) will relapse.1,2 The primary goal in these patients is to reliably determine eligibility for allogeneic hematopoietic cell transplantation (alloHCT) and donor availability.3 Salvage treatment with intensive chemotherapy regimens like fludarabine, cytarabine, and idarubicin (FLA-IDA), high-dose cytosine arabinoside and mitoxantrone (HAM), or mitoxantrone, etoposide and intermediate-dose Ara-C (MEC) induce CR in 35-58%, and median overall survival (OS) of 6.5-11.9 months in patients undergoing alloHCT and of 1.511.9 months in patients not eligible for alloHCT.4-9 Lower-intensity regimens combining the B-cell lymphoma2 (BCL-2) inhibitor venetoclax and hypomethylating agents (HMA) or low-dose cytarabine (LDAC) demonstrated high efficacy in AML.10-12 Based on the results of the VIALE-A trial

Haematologica | 109 January 2024

ARTICLE - FLAVIDA in R/R AML

R. Shahswar, et al.

In a subsequent propensity score matched analysis the efficacy of venetoclax combined with cytarabine, idarubicin and either fludarabine or cladribine was compared to treatment with cytarabine, idarubicin and either fludarabine, cladribine or clofarabine without venetoclax. The analysis included 279 newly diagnosed AML patients, with 85 patients in the venetoclax and 194 in the control cohort.18 Eighty-six percent of patients in the venetoclax cohort achieved measurable residual disease (MRD)negative CRc compared to 61% in the control cohort (odds ratio [OR] =3.2 ; 95% confidence interval [CI]: 1.5-6.7; P=0.0028). While event-free survival (EFS) was significantly improved (median NR; 95% CI: NR-NR vs. 14.3 months; 95% CI: 10.7-33.5; hazard ratio [HR] =0.57; 95% CI: 0.340.95; P=0.03), OS was not significantly different between both cohorts (median NR; 95% CI: 24-NR vs. 32 months; 95% CI: 19-NR; HR=0.63; 95% CI: 0.35-1.1; P=0.13. In R/R AML, 67% of patients treated with FLAG-IDA with venetoclax (FLAVIDA) achieved a CRc and 46% transitioned to alloHCT. Median OS in R/R AML patients was 13 months (95% CI: 7-NR). Due to high rates of sepsis in the first six patients treated with FLAG-IDA with a 21-day regimen, venetoclax was reduced to 200 mg days 1-14 in the subsequent patients.17 In another retrospective analysis of 25 patients including mainly R/R AML patients, venetoclax combined with FLAG induced CRc in 72%. In this study, venetoclax was administered at a dose of 400 mg once daily for a median of 8 days. Three patients (12%) experienced early death within 30 days of therapy initiation and 1-year OS was 50%.19 Zucenka et al.20 retrospectively compared the outcomes of 20 patients receiving venetoclax plus low-dose cytarabine plus actinomycin D with 29 patients receiving FLAG-IDA as salvage therapy for R/R AML after alloHCT. The CR/CRp rate was significantly higher in the venetoclax (FLAVIDA) compared to the FLAG-IDA groups (70%, n=14/20 vs. 34% n=10/29, respectively; P= 0.02). Yet, it remains unclear, whether the high-response rates of regimens combining purine analogues and cytarabine with venetoclax translate into better survival outcomes. We, therefore, compared patients treated with FLA-IDA with a short 7-day course of venetoclax as salvage regimen in R/R AML patients with a historical cohort of FLA-IDA treated patients from our institution.

Methods Inclusion criteria Patients aged 18 years or older with refractory or relapsed (R/R) AML, who had been treated with FLAVIDA between May 2018 and July 2021 and had been reported to the venetoclax registry (venreg.org) were included in the analysis. Patients with acute promyelocytic leukemia or

significant cardiovascular, renal or hepatic comorbidities were excluded. All patients had given written informed consent to the off-label use of venetoclax, genetic analysis and use of clinical data according to the Declaration of Helsinki and institutional guidelines. The registry was approved by the local Ethics Review Committee (ethical vote No.7972_BO_K_2018). Treatment administration All patients had received venetoclax for salvage chemotherapy in combination with fludarabine, cytarabine, and idarubicin (FLA-IDA).21 Venetoclax was administered without dose ramp-up at a dose of 100 mg instead of 400 mg once daily per orally (days 1-7) due to mandatory co-medication with a CYP3A4 inhibitor for fungal prophylaxis, primarily posaconazole. The control patients were selected from the in-house database of Hannover Medical School and were treated with FLA-IDA between 2000 and 2018 for relapsed or refractory AML. Eighty-one patients were identified that fulfilled the treatment criteria of the FLAVIDA patients and were compared to the FLAVIDA patients included in this analysis (detailed information about the dosing scheme is provided in the Online Supplementary Appendix). Cytogenetic and molecular analysis Molecular and cytogenetic analysis was performed centrally by G- and R-banding analysis and next-generation sequencing (NGS) using peripheral blood or bone marrow as reported previously.22 Molecular analysis was performed before start of chemotherapy, at time of relapse, and at time of refractory disease. MRD was assessed on bone marrow specimens using either mutation-specific realtime quantitative polymerase chain reaction (qRT-PCR) for NPM1, capillary electrophoresis for FLT3-ITD, or NGSbased MRD detection for any other MRD marker as reported previously.23-26 In patients without detectable mutations by NGS, multi-parameter flow cytometry using leukemia-associated immunophenotypes (LAIP) was performed before start of FLAVIDA chemotherapy and for detection of measurable residual disease after the first cycle. MRD assessment was performed at the end of cycle one. Details on MRD assessment are provided in the Online Supplementary Appendix. Safety and efficacy assessment The primary objectives included safety and tolerability of a 7-day venetoclax regimen with FLA-IDA and assessment of the overall response rate (ORR). Secondary objectives included assessment of the MRD-negative response rate and survival outcomes including OS (time from treatment start to death) and EFS (time from treatment start until refractory disease, relapse or death, whichever occurred first). The ORR was defined by European LeukemiaNet

Haematologica | 109 January 2024

ARTICLE - FLAVIDA in R/R AML

R. Shahswar, et al.

2017 (ELN) criteria and included CR, CR with incomplete blood count recovery (CRi) (composite complete remission CRc=CR+CRi), and morphologic leukemia-free state (MLFS, defined as less than 5% blasts in an aspirate sample without hematological recovery).27 Details on efficacy and safety assessment are provided in the Online Supplementary Appendix. Statistical analysis Safety and efficacy analyses were performed for all patients who received at least one cycle of FLAVIDA chemotherapy. Demographics were analyzed by descriptive statistics. The propensity score for each patient was calculated as a probability from a logistic regression model that included all covariates deemed likely to have affected treatment decisions and response including age, ELN 2017 risk, sex, AML type, prior alloHCT, and refractory versus relapsed disease. Estimated propensity scores were incorporated into survival models using Inverse Probability of Treatment Weighting (IPTW) (see the Online Supplementary Appendix). The statistical analyses were performed using statistical software environment R, version 3.5.1 using packages survival, survminer, PSweight and cmprsk (R Foundation for Statistical Computing, Vienna, Austria), and statistical software package SPSS 27.0 (IBM Corporation, Armonk, NY, USA).

Results Patient characteristics Thirty-seven and 81 sequentially treated patients received FLAVIDA and FLA-IDA, respectively, and had safety and efficacy outcomes reported. Baseline characteristics were similarly distributed between FLAVIDA and FLA-IDA patients (Table 1). A similar proportion of patients had refractory AML (49% and 42%) or relapsed AML (51% and 58%) in the FLAVIDA and FLA-IDA groups, respectively. Refractoriness was not uniformly defined due to the retrospective design of the study. However, the number of prior chemotherapy cycles was comparable in FLAVIDA and FLA-IDA treated patients (Table 1). In addition, bone marrow blasts in patients considered refractory after one cycle of chemotherapy were similar between FLAVIDA and FLA-IDA treated patients (Table 1). Thus, characteristics defining refractory disease were met by a similar number of patients in both cohorts. The number of patients with prior alloHCT in relapsed FLAVIDA-treated patients was six of 19 (32%), and was similar in FLA-IDA control patients with 21 of 47 (44%). The median duration of the most recent CR before FLA(V)IDA therapy was 13.3 months in relapsed FLAVIDA patients and 11.9 months in relapsed FLA-IDA control patients (Table 1). The number of prior lines of therapy was comparable between both groups

(Table 1). Mutation analysis was available for 36 FLAVIDA and 63 FLA-IDA patients. Molecular aberrations were similarly distributed between FLAVIDA and FLA-IDA patients except a lower frequency of SRSF2 mutations in the FLAIDA control group (Table 1). Treatment response All patients received either one cycle of FLAVIDA or one cycle of FLA-IDA chemotherapy. The ORR was 78% (n=29) in the FLAVIDA group compared to 47% (n=38) in the FLAIDA group (P=0.001). The CRc rate was 59% and 30% in FLAVIDA and FLA-IDA patients, respectively (P=0.003, Table 2; Figure 1). MRD data after one treatment cycle was available for 26 and only 23 patients with overall response (OR) in FLAVIDA and FLA-IDA cohorts, respectively. MRD-negative response was achieved in a similar proportion of patients treated with the two regimens (FLAVIDA 50% and FLA-IDA 57%; P=0.65, Table 2; Online Supplementary Figure S1). Exploratory analysis of response in clinical and genetic subgroups showed a high consistency of the favorable response to FLAVIDA compared to FLA-IDA across most subgroups, notably in patients with ELN adverse risk, and IDH1/2 mutations, while the benefit was less pronounced in patients with refractory disease, and KRAS/NRAS mutations (Figure 2). In summary, significantly more patients treated with FLAVIDA compared to FLA-IDA achieved OR and CRc, while the MRD rate was similar among responding patients. Allogeneic hematopoietic cell transplantation After FLAVIDA chemotherapy, 30 (81%) patients transitioned to alloHCT (n=27) or received donor lymphocyte infusions (DLI, n=3). Among responding patients, 90% (n=26/29) proceeded to alloHCT or received DLI. In the FLA-IDA cohort, 64 (79%) patients proceeded to alloHCT (n=49) or received DLI (n=15) (Online Supplementary Table S1). Among relapsed and refractory patients, usage of alloHCT as well as treatment outcome was similar (Online Supplementary Appendix; Online Supplementary Figures S2A, B and S3A, B). Dose intensity for conditioning was comparable in the FLAVIDA and FLA-IDA groups (reducedintensity conditioning [RIC] FLAVIDA vs. FLA-IDA 61% vs. 67%; myeloablative conditioning [MAC] FLAVIDA vs. FLAIDA 39% vs. 33%) (Table 1). In summary, both regimens allowed bridging to alloHCT or DLI in a high proportion of patients. Survival OS rates at 1 and 2 years were 52% and 48% in the FLAVIDA cohort after a median follow-up of 22.4 months and 61% and 52% in the FLA-IDA cohort after a median follow-up of 62.9 months, respectively. Despite a higher ORR in FLAVIDA patients, OS was similar between FLAVIDA and FLA-IDA-treated patients after propensity score

Haematologica | 109 January 2024

ARTICLE - FLAVIDA in R/R AML

R. Shahswar, et al.

Table 1. Patient demographics, baseline and treatment characteristics of FLAVIDA- and FLA-IDA-treated patients.

Baseline characteristics Age in years, median (range) Sex, N (%) Male Female Type of AML, N (%) De novo Secondary Therapy-related ELN 2017 risk group, N (%) Favorable Intermediate Adverse Missing Treatment lines before FLA(V)IDA Median (range) Salvage 1, N (%) Salvage 2, N (%) Salvage 3 or greater, N (%) Disease status, N (%) Refractory AML Relapsed AML Refractory after cycles of chemotherapy 1 cycle, N (%) ≥2 cycles, N (%) Bone marrow blasts in patients considered refractory after 1 cycle Median, % (range) Median duration in months of most recent CR before FLA(V)IDA therapy Molecular mutations, N (%) DNMT3A NPM1 SRSF2 FLT3-ITD IDH1 IDH2 RUNX1 NF1 K/NRAS TP53 Prior alloHCT before FLA(V)IDA, N (%) Overall (relapsed and refractory) In relapsed patients Conditioning regimen, N (%) Reduced-intensity conditioning Myeloablative conditioning Use of G-CSF after FLA(V)IDA, N (%) Yes No Missing

FLAVIDA N=37 54 (19-70)

FLA-IDA N=81 52 (22-72)

21 (57) 16 (43)

45 (56) 36 (44)

P 0.95 0.9

0.53 26 (70) 10 (27) 1 (3)

62 (77) 15 (19) 4 (5) 0.2

6 (16) 19 (51) 12 (33) 0 (0)

20 (25) 37 (46) 20 (25) 4 (5)

1 (1-5) 26 (70) 9 (25) 2 (5)

1 (1-4) 57 (70) 16 (20) 8 (10)

0.91

0.5 18 (49) 19 (51) N=18 14 (78) 4 (22)

34 (42) 47 (58) N=34 28 (82) 6 (18)

0.34

0.94 55 (15-90)

44 (5-95)

13.3

11.9

0.81

N=36 13 (36) 11 (31) 9 (25) 8 (22) 2 (6) 8 (22) 7 (19) 6 (17) 4 (11) 2 (6)

N=63 21 (33) 14 (22) 3 (5) 19 (30) 5 (6) 9 (14) 5 (8) 5 (8) 8 (13) 4 (6)

0.78 0.36 <0.01 0.39 0.66 0.31 0.09 0.18 0.82 0.87

8 (22) 6/19 (32) N=27 16 (59) 11 (41)

23 (28) 21/47 (44) N=49 33 (67) 16 (33)

0.44 0.33 0.48

0.63 25 (68) 12 (32) 0 (0)

56 (69) 20 (25) 5 (6)

FLA-IDA: fludarabine, cytarabine, and idarubicin; FLAVIDA: FLA-IDA with venetoclax; AML: acute myeloid leukemia; alloHCT: allogeneic hematopoietic cell transplantation; CR: complete remission; ELN: European LeukemiaNet; G-CSF: granulocyte colony stimulating factor; N: number; P: P value.

Haematologica | 109 January 2024

ARTICLE - FLAVIDA in R/R AML

R. Shahswar, et al.

Table 2. Treatment response after one cycle of FLAVIDA and FLA-IDA treatment.

ORR (CRc+MLFS), N (%) CRc (CR+CRi), N (%) CR CRi MLFS (%) MRD- ORR No Response (PR+SD+RD), N (%) Early death (within 30 days), N (%)

FLAVIDA N=37 29 (78) 22 (59) 20 (54) 2 (5) 7 (19) 13/26 (50) 7 (19) 1 (3)

FLA-IDA N=81 38 (47) 34 (42) 24 (30) 10 (12) 4 (5) 13/23 (57) 42 (52) 1(1)

OR (95%CI)

0.001 0.003

4.1 (1.67-10.1) 2.0 (1.0-4.5)

0.02 0.65 0.001 0.6

4.5 (1.23-16.5) 0.77 (0.25-2.4) 0.24 (0.1-0.6) 2.2 (0.14-36.5)

FLA-IDA: fludarabine, cytarabine, and idarubicin; FLAVIDA: FLA-IDA with venetoclax; CI: confidence interval; CR: complete remission; CRc: composite complete remission; CRi: complete remission with incomplete hematological recovery; MLFS: morphologic leukemia-free state; MRD: measurable residual disease; N: number; OR: odds ratio; ORR: overall response rate; P: P value; PR: partial remission; RD: refractory disease; SD: stable disease.

Figure 1. Frequency of complete remission, complete remission with incomplete hematological recovery, morphologic leukemia-free state, and overall response rate in FLAVIDA- (N=37) and FLA-IDA- (N=81) treated patients. FLA-IDA: fludarabine, cytarabine, and idarubicin; FLAVIDA: FLA-IDA with venetoclax; CR: complete remission; CRi: complete remission with incomplete hematological recovery; MLFS: morphologic leukemia-free state; ORR: overall response rate; CI: confidence interval.

weighting (HR=1.25; P=0.4) (Figure 3A; Online Supplementary Table S2). Median EFS was comparable between FLAVIDA and FLA-IDA patients (11.3 months vs. 6.9 months; HR=0.7; P=0.1) (Figure 3B; Online Supplementary Table S2). Survival outcomes and response rates were also similar when excluding patients treated with FLA-IDA before 2007, when antifungals were not available (Online Supplementary Appendix; Online Supplementary Figure S4A, B). In the 26 responding FLAVIDA patients with available MRD data, survival was significantly longer in MRD-negative patients compared to MRD-positive patients (HR=0.1; 95% CI: 0.01-0.59; P=0.01) (Figure 4A). MRD was not prognostic for OS in responding FLA-IDA patients, while it should be recognized that MRD was available only for 28% of FLA-IDA-

treated patients due to missing samples (HR=0.41; 95% CI: 0.1-1.7; P=0.23) (Figure 4B). However, OS rates at 1 year and 2 years were similar in MRD-negative FLAVIDA and FLAIDA patients (OS at 1 year and 2 years FLAVIDA 89% and 89% vs. FLA-IDA 85% and 77%; P=0.4) (Online Supplementary Figure S5). While RFS appeared superior by trend for MRD-negative compared to MRD-positive CR/CRi patients in the FLAVIDA group, RFS was comparable between MRD-negative and MRD-positive CR/CRi patients in the FLA-IDA group (FLAVIDA: HR for relapse or death, MRD- vs. MRD+ 0.2 ; 95% CI: 0.04-1.0; P=0.11; FLA-IDA: HR for relapse or death, MRDvs. MRD+ 0.76; 95% CI: 0.27-2.1; P=0.59) (Figures 4C, D). In order to better understand why a higher ORR and longer OS of MRD- patients did not result in a survival benefit in

Haematologica | 109 January 2024

ARTICLE - FLAVIDA in R/R AML

R. Shahswar, et al.

Figure 2. Impact of patient and disease characteristics on achieving overall response in FLAVIDA and FLA-IDA patients. TP53 status not shown due to small number of mutated patients (FLAVIDA TP53mut N=2; FLA-IDA TP53mut N=4). FLA-IDA: fludarabine, cytarabine, and idarubicin; FLAVIDA: FlA-IDA with venetoclax; CI: confidence interval; sAML: secondary acute myeloid leukemia; tAML: therapy-related AML; ELN: European LeukemiaNet.

FLAVIDA treated patients, we evaluated survival differences across clinical and genetic subgroups between FLAVIDA and FLA-IDA regimens. EFS in IDH1/2 mutated patients favored FLAVIDA over FLA-IDA, while OS was similar in these patients (Online Supplementary Figures S6 and S7). The only subgroup with different EFS and OS between FLAVIDA and FLA-IDA was the patient group not responding to these regimens, in which FLA-IDA treatment was associated with improved EFS and OS compared to FLAVIDA treatment (Online Supplementary Figures S6 and S7). Median OS from the time of non-response was significantly longer in non-responding FLA-IDA compared to

non-responding FLAVIDA patients (HR=0.3; 95% CI: 0.140.71; P<0.001), suggesting that FLA-IDA treated patients could be salvaged more successfully (Figure 4). In addition, non-responding FLA-IDA patients numerically more often received alloHCT/DLI (FLA-IDA cohort 81%, n=35/43 vs. FLAVIDA cohort 50%, n=4/8; P=0.05), and had less pretreatment lines (FLA-IDA median lines 1; range, 1-4 vs. FLAVIDA median lines 2; range 1-5; P=0.02) (Online Supplementary Tables S3 and S4). Safety Most commonly observed treatment-related toxicities of

Haematologica | 109 January 2024

ARTICLE - FLAVIDA in R/R AML

R. Shahswar, et al.

any grade were hematological adverse events (AE) being reported in all patients. Non-hematological AE of all grades included bacteremia, sepsis, and fungal pneumonia occurring in 27%, 11%, and 11%, respectively. One patient (3%) died within 30 days after start of FLAVIDA treatment from multiorgan failure following pneumonia. The most common grade 3 and 4 all-causality event was neutropenic fever (97%), while thrombocytopenia, anemia, and neutropenia occurred in all patients at nadir (Online Supplementary Table S5). The median time to neutrophil recovery (>500/nL) in responding FLAVIDA-treated patients was 33 days (95% CI: 30-36) and 28 days (95% CI: 23-33) in the FLA-IDA cohort

(P=0.94) from day 1 of chemotherapy (Online Supplementary Figure S8A). Recovery times for ANC >1,000/nL for responding FLAVIDA patients were 35 days (95% CI: 34-36), and were similar to those in the FLA-IDA cohort (34 days; 95% CI: 3038; P=1.0) (Online Supplementary Figure S8B). Median time for platelet recovery (>50/nL) was 35 days (95% CI: 32-38) in the FLAVIDA cohort versus 34 days (95% CI: 27-41) in the FLA-IDA cohort (P=0.85) (Online Supplementary Figure S8C). Recovery time for platelet counts >100/nL for responding FLAVIDA patients was 36 days (95% CI: 33-39) compared to 34 days (95% CI: 31-37) in the FLA-IDA cohort (P=0.86) (Online Supplementary Figure S8D). Thus, absolute neutrophil

Figure 3. Survival outcomes in FLAVIDA- and FLA-IDA-treated patients. (A) Inverse probability of preatment weighted Kaplan Meier curves of overall survival (OS) in FLAVIDA- and FLA-IDA-treated patients. (B) Inverse probability of treatment weighted Kaplan Meier curves of event-free survival (EFS) in FLAVIDA- and FLA-IDA-treated patients. FLA-IDA: fludarabine, cytarabine, and idarubicin; FLAVIDA: FlA-IDA with venetoclax; HR: hazard ratio; CI: confidence interval. Haematologica | 109 January 2024

ARTICLE - FLAVIDA in R/R AML

R. Shahswar, et al.

count (ANC) and platelet recovery times were comparable between FLAVIDA and FLA-IDA regimens (Online Supplementary Table S6). Recovery times were also similar between both regimens when including patients with MLFS (FLAVIDA vs. FLA-IDA: ANC recovery >500/nL 35 days vs. 32 days, P=0.43; ANC recovery >1,000/nL 36 days vs. 38 days, P=0.86; PLT recovery >50/nL 36 days vs. 34 days, P=0.39; PLT recovery >100/nL 38 days vs. 34 days, P=0.22) suggesting similar hematologic toxicity of the two regimens.

IDA, while OS was similar for the two regimens. High response rates were observed in patients with de novo, secondary and therapy-related AML, and across ELN and mutational subgroups. However, due to the small sample size subgroup analyses should be interpreted with caution. The response rate in the FLA-IDA-treated cohort was consistent with the previously reported response rates for this regimen.28,29 Responding patients with available MRD data attained MRD negativity in a similar proportion in the FLAVIDA and FLA-IDA groups, respectively. DiNardo and colleagues17 reported flow cytometry MRD negativity in 69% of responding patients who were treated with FLAGDiscussion IDA and venetoclax. Lachowiecz and colleagues18 reported In this study, short-term venetoclax in combination with higher rates of MRD-negative remission in patients receivFLA-IDA salvage chemotherapy was well tolerated and ing intensive chemotherapy with versus without venetodemonstrated higher response rates compared to FLA- clax during first line treatment (86% vs. 61%). In their study

Continued on following page. Haematologica | 109 January 2024

ARTICLE - FLAVIDA in R/R AML

R. Shahswar, et al.

Figure 4. Survival outcomes in FLAVIDA- and FLA-IDA-treated patients depending on measurable residual disease status. (A) Overall survival (OS) in FLAVIDA-treated patients with measurable residual disease (MRD)-negative or MRD-positive overall response (including CR, CRi, and MLFS). Landmark analysis was performed from time of overall response. (B) OS in FLA-IDA-treated patients with MRD-negative or MRD-positive overall response (including CR, CRi, and MLFS). Landmark analysis was performed from time of overall response. (C) Relapse-free survival in FLAVIDA patients with MRD-negative or MRD-positive CR/CRi. Landmark analysis was performed from time of CR/CRi. (D) Relapse-free survival in FLA-IDA patients with MRD-negative or MRD-positive CR/CRi. Landmark analysis was performed from time of CR/Cri. FLA-IDA: fludarabine, cytarabine, and idarubicin; FLAVIDA: FlAIDA with venetoclax; CR: complete remission; CRc: composite complete remission; CRi: complete remission with incomplete hematological recovery; MLFS: morphologic leukemia free state

a higher rate of MRD negativity was only found in ELN adverse-risk patients. The high response rate enabled a high transplant/DLI rate of 81%, which was similar to the transplant rate in FLAIDA-treated patients (79%; P=0.88). Despite a significantly higher response rate in the FLAVIDA cohort, EFS and OS were similar in FLA-IDA-treated patients. Only IDH1/2-mutated patients showed an improved EFS with FLAVIDA, consistent with previous studies showing

high sensitivity of these patients towards venetoclax, while other clinical and genetic subgroups showed similar EFS and OS with FLAVIDA or FLA-IDA.13,30 The previously reported median OS of R/R AML patients treated with FLAG-IDA and venetoclax of 13 months and 12 months were similar to the median OS of 12 months found in our study.17,19 In first line treated AML patients, the addition of venetoclax to intensive chemotherapy improved EFS but not OS.18 Improved OS was only observed in the subgroup of ELN adverse risk patients.18

Haematologica | 109 January 2024

ARTICLE - FLAVIDA in R/R AML

R. Shahswar, et al.

Figure 5. Overall survival in non-responding FLAVIDA- or FLA-IDA-treated patients. Landmark analysis was performed from time of first response assessment. OS: overall survival; HR: hazard ratio; CI: confidence interval. FLA-IDA: fludarabine, cytarabine, and idarubicin; FLAVIDA: FlA-IDA with venetoclax.

In the subgroup of patients with response to either regimen, RFS and OS was similar between FLAVIDA and FLA-IDA-treated cohorts. MRD-negative patients in the FLAVIDA group had a favorable 2-year-OS of 89%, while the 2-year OS of MRD-positive patients was 40%. Based on the higher response rate and comparable MRD negativity rate in FLAVIDA compared to FLA-IDA-treated patients, more patients with MRD- response can be expected from FLAVIDA treatment (corresponding to 39% vs. 27%), if our results are extrapolated to all treated patients. We further evaluated how non-responding patients to either regimen survived after being refractory. We show that non-responding patients in the FLA-IDA cohort could be salvaged more effectively than non-responding patients in the FLAVIDA cohort. This is partially explained by a higher use of salvage alloHCT and fewer lines of prior treatments in FLA-IDA treated patients and is consistent with the dismal outcome in unfit AML patients who failed treatment with a hypomethylating agent and venetoclax.31 These data suggest that FLAVIDA efficiently selects the most chemotherapy refractory patients, while some nonresponding patients in the FLA-IDA group may still be rescued with additional alloHCT. Our study raises important questions. First, it will be important to improve the outcome of MRD-positive patients in the FLAVIDA group for example with donor-lymphocyte infusions or maintenance treatment after alloHCT. Second, patients not responding to FLAVIDA have a dismal outcome and these patients should be treated in clinical trials. Third, it is of interest, whether MRD-negative patients in the FLAVIDA group benefit from alloHCT or can also achieve long-term remission with additional cycles of FLAVIDA since MRD negativity in the relapse setting has not yet proven its clinical value.

Since toxicity is a major concern when adding novel drugs to existing regimens, safety and toxicity are important outcomes of this study. Common grade 3 and 4 AE occurring in the FLAVIDA cohort were hematological toxicity and febrile neutropenia, which were managed with transfusions and anti-infective therapy. The early death rate was low with both treatment regimens. Comparison of ANC and platelet count recovery were overall comparable between FLAVIDA and FLA-IDA regimens suggesting that a 7-day course of venetoclax does not add significant hematologic toxicity to the FLA-IDA regimen. As the response and survival rates were similar to other published studies using FLAG-IDA and longer duration venetoclax (days 1-14 during induction and days 1-7 during consolidation when combined with FLAG-IDA; days 1-14 when combined with cytarabine+idarubicin 5+2), treatment with 7 days venetoclax simultaneously with chemotherapy appears similarly effective as venetoclax for longer periods of time.17-19 Based on emerging data we now use venetoclax 50 mg combined with posaconazole or voriconazole. Limitations of our study include the moderate sample size in the FLAVIDA group, retrospective analysis of FLAVIDAtreated patients, the non-randomized comparison to the FLA-IDA cohort and the long period in which FLA-IDA patients were treated. Using historical controls bears the risk of historical bias. However, since rates of alloHCT and OS were similar between the two cohorts, the historical control appears a valid comparator. Our data demonstrate that FLAVIDA is an effective intensive salvage treatment option for R/R AML, particularly as a bridge to alloHCT, inducing high ORR including molecular remissions and allowing high rates of alloHSCT in this difficult to treat AML population. From a patient perspective it may seem preferable to identify chemotherapy respon-

Haematologica | 109 January 2024

ARTICLE - FLAVIDA in R/R AML

R. Shahswar, et al.

sivess rapidly with a regimen like FLAVIDA instead of sequencing multiple salvage regimens until best response is achieved. In light of comparable toxicity this currently argues for the continued use of FLAVIDA in relapsed and refractory AML patients who are eligible for allogeneic transplantation. Disclosures RS declares honoraria from Abbvie and Jazz Pharmaceuticals. MH declares honoraria from Abbvie, Eurocept, Jazz Pharmaceuticals, Janssen, Novartis and Takeda; paid consultancy for Abbvie, Agios, BMS, Daiichi Sankyo, Glycostem, Jazz Pharmaceuticals, Kura Oncology, Novartis, Pfizer, PinotBio, Roche and Tolremo; research funding to his institution from Abbvie, Agios, Astellas, Bayer Pharma AG, BergenBio, Daiichi Sankyo, Glycostem, Jazz Pharmaceuticals, Loxo Oncology, Novartis, Pfizer, PinotBio and Roche. The other authors have no conflict of interest to disclose.

the database. PK, LD, CA, MW and YA contributed to the collection of clinical and biological data. GB, RG, AS, AK, CK, MS, GG, BS, ZL, LD, PK, CA MW, YA, AB, MS, AG, and FT contributed to the analysis of clinical and biological data. RS, RG and MH performed the statistical analysis. RS and MH interpreted the data and wrote the manuscript. All authors read and agreed to the final version of the manuscript. Funding This study was supported by grants HE 5240/6-1, HE 5240/6-2 to MH from DFG, an ERC grant to MH from the European Union’s Horizon 2020 research and innovation programme (no. 638035), grant 16 R/2021 to MH from DJCLS, and grant 70114189 to MH from Deutsche Krebshilfe. RS received a grant from PRACTIS.

Data-sharing statement Individual patient data will not be made available in order to maintain health information privacy. De-identified MRD Contributions and mutation information will be shared upon reasonable RS and MH designed the study. RS, MH, and GB controlled request to the corresponding author.

References Engl J Med. 2020;383(7):617-629. 11. DiNardo CD, Maiti A, Rausch CR, et al. 10-day decitabine with venetoclax for newly diagnosed intensive chemotherapy ineligible, and relapsed or refractory acute myeloid leukaemia: a single-centre, phase 2 trial. Lancet Haematol. 2020;7(10):e724-36. 12. Wei AH, Montesinos P, Ivanov V, et al. Venetoclax plus LDAC for newly diagnosed AML ineligible for intensive chemotherapy: a phase 3 randomized placebo-controlled trial. Blood. 2020;135(24):2137-2145. 13. Lachowiez CA, Loghavi S, Kadia TM, et al. Outcomes of older patients with NPM1-mutated AML: current treatments and the promise of venetoclax-based regimens. Blood Adv. 2020;4(7):1311-1320. 14. Konopleva M, Thirman MJ, Pratz KW, et al. Impact of FLT3 mutation on outcomes after venetoclax and azacitidine for patients with treatment-naïve acute myeloid leukemia. Clin Cancer Res. 2022;28(13):2744-2752. 15. Pollyea DA, Pratz KW, Wei AH, et al. Outcomes in patients with poor-risk cytogenetics with or without TP53 mutations treated with venetoclax and azacitidine. Clin Cancer Res. 2022;28(24):5272-5279. 16. Chua CC, Roberts AW, Reynolds J, et al. Chemotherapy and venetoclax in elderly acute myeloid leukemia Trial (CAVEAT): a phase Ib dose-escalation study of venetoclax combined with modified intensive chemotherapy. J Clin Oncol. 2020;38(30):3506-3517. 17. DiNardo CD, Lachowiez CA, Takahashi K, et al. Venetoclax combined with FLAG-IDA induction and consolidation in newly diagnosed and relapsed or refractory acute myeloid leukemia. J Clin Oncol. 2021;39(25):2768-2778. 18. Lachowiez CA, Reville PK, Kantarjian H, et al. Venetoclax combined with induction chemotherapy in patients with newly diagnosed acute myeloid leukaemia: a post-hoc, propensity

1. Döhner H, Wei AH, Appelbaum FR, et al. Diagnosis and management of AML in adults: 2022 recommendations from an international expert panel on behalf of the ELN. Blood. 2022;140(12):1345-1377. 2. Walter RB, Othus M, Burnett AK, et al. Resistance prediction in AML: analysis of 4601 patients from MRC/NCRI, HOVON/SAKK, SWOG and MD Anderson Cancer Center. Leukemia. 2015;29(2):312-320. 3. Heuser M, Ofran Y, Boissel N, et al. Acute myeloid leukaemia in adult patients: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up [published correction appears in Ann Oncol. 2021;32(6):821]. Ann Oncol. 2020;31(6):697-712. 4. Thol F, Heuser M. Treatment for relapsed/refractory cute myeloid leukemia. Hemasphere. 2021;5(6):e572. 5. Nath R, Reddy V, Kapur A, et al. Survival of relapsed/refractory acute myeloid leukemia (R/R AML) patients receiving stem cell transplantation (SCT). Biol Blood Marrow Transplant. 2019;25(3):S125. 6. Westhus J, Noppeney R, Dührsen U, Hanoun M. FLAG salvage therapy combined with idarubicin in relapsed/refractory acute myeloid leukemia. Leuk Lymphoma. 2019;60(4):1014-1022. 7. Marconi G, Talami A, Abbenante MC, et al. MEC (mitoxantrone, etoposide, and cytarabine) induces complete remission and is an effective bridge to transplant in acute myeloid leukemia. Eur J Haematol. 2020;105(1):47-55. 8. Canaani J, Nagar M, Heering G, et al. Reassessing the role of high dose cytarabine and mitoxantrone in relapsed/refractory acute myeloid leukemia. Oncotarget. 2020;11(23):2233-2245. 9. Megías-Vericat JE, Martínez-Cuadrón D, Sanz MÁ, Montesinos P. Salvage regimens using conventional chemotherapy agents for relapsed/refractory adult AML patients: a systematic literature review. Ann Hematol. 2018;97(7):1115-1153. 10. DiNardo CD, Jonas BA, Pullarkat V, et al. Azacitidine and venetoclax in previously untreated acute myeloid leukemia. N

Haematologica | 109 January 2024

ARTICLE - FLAVIDA in R/R AML

R. Shahswar, et al.

score-matched, cohort study. Lancet Haematol. 2022;9(5):e350-e360. 19. Wolach O, Frisch A, Shargian L, et al. Venetoclax in combination with FLAG-IDA-based protocol for patients with acute myeloid leukemia: a real-world analysis. Ann Hematol. 2022;101(8):1719-1726. 20. Zucenka A, Vaitekenaite V, Maneikis K, et al. Venetoclax-based salvage therapy followed by Venetoclax and DLI maintenance vs. FLAG-Ida for relapsed or refractory acute myeloid leukemia after allogeneic stem cell transplantation. Bone Marrow Transplant. 2021;56(11):2804-2812. 21. Shahswar R, Beutel G, Klement P, et al. FLA-IDA salvage chemotherapy combined with a seven-day course of venetoclax (FLAVIDA) in patients with relapsed/refractory acute leukaemia. Br J Haematol. 2020;188(3):e11-e15. 22. Heuser M, Gabdoulline R, Löffeld P, et al. Individual outcome prediction for myelodysplastic syndrome (MDS) and secondary acute myeloid leukemia from MDS after allogeneic hematopoietic cell transplantation. Ann Hematol. 2017;96(8):1361-1372. 23. Gabert J, Beillard E, van der Velden VHJ, et al. Standardization and quality control studies of ‘real-time’ quantitative reverse transcriptase polymerase chain reaction of fusion gene transcripts for residual disease detection in leukemia a Europe Against Cancer Program. Leukemia. 2003;17(12):2318-2357. 24. Thol F, Gabdoulline R, Liebich A, et al. Measurable residual disease monitoring by NGS before allogeneic hematopoietic cell

transplantation in AML. Blood. 2018;132(16):1703-1713. 25. Krönke J, Schlenk RF, Jensen K-O, et al. Monitoring of minimal residual disease in NPM1-mutated acute myeloid leukemia: a study from the German-Austrian acute myeloid leukemia study group. J Clin Oncol. 2011;29(19):2709-2716. 26. Schuurhuis GJ, Heuser M, Freeman S, et al. Minimal/measurable residual disease in AML: a consensus document from the European LeukemiaNet MRD Working Party. Blood. 2018;131(12):1275-1291. 27. 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. 28. Bergua JM, Montesinos P, Martinez-Cuadrón D, et al. A prognostic model for survival after salvage treatment with FLAG-Ida +/- gemtuzumab-ozogamicine in adult patients with refractory/relapsed acute myeloid leukaemia. Br J Haematol. 2016;174(5):700-710. 29. Roboz GJ, Rosenblat T, Arellano M, et al. International randomized phase III study of elacytarabine versus investigator choice in patients with relapsed/refractory acute myeloid leukemia. J Clin Oncol. 2014;32(18):1919-1926. 30. DiNardo CD, Pratz K, Pullarkat V, et al. Venetoclax combined with decitabine or azacitidine in treatment-naive, elderly patients with acute myeloid leukemia. Blood. 2019;133(1):7-17. 31. Maiti A, Rausch CR, Cortes JE, et al. Outcomes of relapsed or refractory acute myeloid leukemia after frontline hypomethylating agent and venetoclax regimens. Haematologica. 2021;106(3):894-898.

Haematologica | 109 January 2024

ARTICLE - Acute Myeloid Leukemia

Twist family BHLH transcription factor 1 is required for the maintenance of leukemia stem cell in MLL-AF9+ acute myeloid leukemia Nan Wang,1,2,3* Jing Yin,1,2,3* Na You,1,2,3* Wenqi Zhu,1,2,3 Nini Guo,1,2,3 Xiaoyan Liu,1,2,3 Peiwen Zhang,1,2,3 Wanling Huang,1,2,3 Yueqiao Xie,1,2,3 Qian Ren1,2,3 and Xiaotong Ma1,2,3

Correspondence: X. Ma maxt@ihcams.ac.cn

State Key Laboratory of Experimental Hematology, National Clinical Research Center for

Blood Diseases, Haihe Laboratory of Cell Ecosystem, Institute of Hematology & Blood

Received: Accepted: Early view:

Janury 15, 2023. September 20, 2023. September 28, 2023.

Diseases Hospital, Chinese Academy of Medical Sciences & Peking Union Medical College; 2

Tianjin Institutes of Health Science and 3Center for Stem Cell Medicine, Chinese Academy

https://doi.org/10.3324/haematol.2023.282748

of Medical Sciences, Tianjin, China

NW, JY and NY contributed equally as first authors.

Abstract Leukemia stem cells (LSC) are a rare population capable of limitless self-renewal and are responsible for the initiation, maintenance, and relapse of leukemia. Elucidation of the mechanisms underlying the regulation of LSC function could provide novel treatment strategies. Here, we show that TWIST1 is extremely highly expressed in the LSC of MLL-AF9+ acute myeloid leukemia (AML), and its upregulation is positively regulated by KDM4C in a H3K9me3 demethylation-dependent manner. We further demonstrate that TWIST1 is essential for the viability, dormancy, and self-renewal capacities of LSC, and that it promotes the initiation and maintenance of MLL-AF9-mediated AML. In addition, TWIST1 directly interacts and collaborates with HOXA9 in inducing AML in mice. Mechanistically, TWIST1 exerts its oncogenic function by activating the WNT5a/RAC1 axis. Collectively, our study uncovers a critical role of TWIST1 in LSC function and provides new mechanistic insights into the pathogenesis of MLL-AF9+ AML.

Introduction Acute myeloid leukemia (AML) is a heterogeneous disease with a multitude of chromosomal aberrations, transcriptional deregulation and epigenetic changes.1 Chromosomal translocations involving the mixed lineage leukemia (MLL) gene cause aggressive leukemia, namely MLL-rearranged (MLL-r) leukemia that accounts for 5-10% of adult AML and 50% of infant AML.2,3 Patients with MLL-r AML have unfavorable outcomes due to refractoriness to current chemotherapy regimens and a shorter period to relapse.4 MLL-r leukemia exhibits unique biological and clinical characteristics that are distinct from those of non-MLL-r AML.5 Multiple chromatin-associated proteins and epigenetic regulators interact with MLL-fusion protein to form large multisubunit complexes, which are involved in chromatin modification and transcriptional elongation.5,6 More than 130 partner genes that fuse with MLL have been identified, and one of the most common driver fusion genes in AML is AF9.7 A better understanding of the molecular mechanisms underlying the pathogenesis of MLL-AF9-mediated AML

will lead to the development of new therapeutic strategies. MLL-AF9 protein transforms hematopoietic stem and progenitor cells (HSPC) through induction of aberrant expression of stem-cell-associated gene programs.8 The most well-known targets of MLL-AF9 protein include HOXA9 and its co-factors MEIS1 and PBX3, all of which are important for MLL-AF9+ AML development.9-12 HOXA9 is overexpressed in more than one half of AML cases and is a strong predictor of poor prognosis.13,14 Overexpression of HOXA9 induces aberrant self-renewal capability and blocks differentiation of HSC, leading to myeloproliferative phenotypes in mice.15 Co-expression of HOXA9 with either MEIS1 or PBX3 induces fatal AML.12,15-17 TWIST1, a key epithelial-to-mesenchymal transition (EMT)-related transcription factor, is shown to play a crucial role in generation and maintenance of cancer stem cells in a number of human solid tumor types.18,19 In the blood system, TWIST1 is required for normal HSC maintenance.20,21 Although TWIST1 expression was overall increased in AML,22-24 among AML subtypes, it has only been reported to be implicated in the progression of acute promyelocytic

Haematologica | 109 January 2024

ARTICLE - TWIST1 is required for leukemia stem cell function

leukemia, which arises from committed promyelocytic progenitors rather than leukemic stem cell (LSC).23 The role of TWIST1 in LSC of AML remains unknown. Recently Luo et al. reported that, in MLL-AF9+ AML, HOX gene loci-associated long non-coding RNA HOTTIP recruits WDR5/MLL complex to promote HOXA gene expression and contributes to disease progression.25,26 HOTTIP aberration in mice leads to an increased LSC pool.25,26 Interestingly, TWIST1 expression positively correlates with HOTTIP, and HOTTIP deletion reduces the recruitment of MLL1 to TWIST1 and leads to a significant decrease in TWIST1 expression in MLL-AF9+ AML cells.25 These data indicate that TWIST1 may be involved in the pathogenesis and LSC regulation of MLL-AF9+ AML. Here we report for the first time that TWIST1 is required for the LSC function of MLL-AF9+ AML through activation of the WNT5a/RAC1 axis, and it directly interacts and collaborates with HOXA9 in malignant transformation.

Methods Mice ER-Cre+/-;Twist1flox/flox mice were generated as previously described.27 Cre expression was induced by daily intraperitoneal injection of tamoxifen (75 mg/kg; Sigma-Aldrich, St. Louis, MO, USA) for 5 days. C57BL/6-Ly5.1 or NOD/SCID mice were purchased from the animal facility of the State Key Laboratory of Experimental Hematology (SKLEH). All animal care and experimental procedures complied with the animal care guidelines and were approved by the Institutional Animal Care and Use Committees of SKLEH. Mouse acute myeloid leukemia model Bone marrow (BM) cells were harvested from C57BL/6Ly5.1, Twist1-knockout, ER-Cre+/-;Twist1flox/flox or ER-Cre mice. c-kit-positive BM cells were enriched by positive selection using CD117 microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany) and incubated overnight in Iscove’s modified Dulbecco medium supplemented with 5% fetal bovine serum, 50 ng/mL murine SCF (Peprotech, Rocky Hill, NJ), 10 ng/mL murine IL-6 (Peprotech), and 10 ng/mL murine IL-3 (Peprotech). After prestimulation, cells were transduced with retroviruses expressing MLL-AF9, Hoxa9, Twist1, or Hoxa9 and Twist1 at a multiplicity of infection (MOI) of 1525 in the presence of 8 μg/mL polybrene (Sigma-Aldrich) on retronectin-coated plates. Seventy-two hours after beginning the transduction, infected cells were sorted for in vitro culture and in vivo transplantation studies. Statistical analysis All data are expressed as the mean ± standard deviation (SD) and are representative of at least two trials with *P<0.05, **P<0.01, and ***P<0.001. Statistical significance was calculated using a paired or unpaired Student’s t test. In the survival experiments, mouse survival rates were analyzed

using the log-rank test. Details of other experimental procedures are described in the Online Supplementary Appendix. Data availability RNA-sequencing data have been deposited in NCBI’s Gene Expression Omnibus (GEO accession: GSE215026). The mass spectrometry proteomics data are available at the ProteomeXchange Consortium via the iProx partner with the dataset identifier PXD037217.

Results TWIST1 is highly expressed in MLL-AF9+ acute myeloid leukemia stem cells and its expression is regulated by KDM4C In order to uncover the role of TWIST1 in MLL-r AML, we first analyzed the expression level of TWIST1 mRNA in patients with MLL-r AML by surveying publicly available datasets. We found that TWIST1 expression was significantly increased in BM mononuclear cells (BMMNC) from patients with MLL-r AML compared with those from healthy volunteers (Figure 1A). In addition, the expression levels of TWIST1 were substantially heterogeneous among MLL-r AML patients, and patients with t (9; 11) MLL-AF9 fusion gene exhibited significantly higher TWIST1 expression than those with other translocations (Figure 1B). We then examined the expression of Twist1 in leukemic granulocyte/macrophage progenitors (L-GMP) and c-kit+Gr1- cells, both of which have been identified as LSC in mice with MLL-AF9-driven AML.3,28 GMP and HSC from normal mice were used as controls. Interestingly, the expression level of Twist1 in L-GMP was extremely high, approximately 35- and 18-fold higher than those in normal GMP and HSC, respectively (Figure 1C), and its level in c-kit+Gr1- cells was 14- and 7-fold higher than those in GMP and HSC, respectively (Figure 1C). These findings point to a potential pathogenetic role of TWIST1 in MLL-AF9-induced AML. We next sought to understand the upstream epigenetic mechanisms controlling TWIST1 expression in MLL-AF9+ AML. In order to clarify whether TWIST1 is directly upregulated by the MLL-AF9 protein, we analyzed published MLL-AF9 chromatin immunoprecipitation sequencing (ChIP-seq) data generated from MLL-AF9 transformed murine hematopoietic cells, and found that MLL-AF9 exhibited no binding to the Twist1 locus (Online Supplementary Figure S1). Aberrant histone modification has been identified as a regulatory mechanism of altered expression of various oncogenes, such as Hox clusters, that are responsible for MLL-AF9+ AML development.29,30 In order to determine whether TWIST1 expression is regulated by histone modification, we analyzed the publicly available ChIP-seq data of H3K9me3, H3K4me3, H3K27me3, H3K36me3, and H3K79me2 in LSC from MLL-

Haematologica | 109 January 2024

N. Wang et al.

ARTICLE - TWIST1 is required for leukemia stem cell function

N. Wang et al.

Continued on following page.

Haematologica | 109 January 2024

ARTICLE - TWIST1 is required for leukemia stem cell function

N. Wang et al.

Figure 1. TWIST1 is expressed at high levels in leukemic stem cells of MLL-AF9+ acute myeloid leukemia and its expression is positively regulated by KDM4C. (A) Comparison of TWIST1 expression between primary bone marrow mononuclear cells (BMMNC) from patients with MLL-rearranged (MLL-r) acute myeloid leukemia (AML, N=38) and those from healthy volunteers (N=73) in acute myeloid leukemia (AML) datasets Microarray Innovations in Leukemia (MILE) (GSE13159). (B) TWIST1 expression in BMMNC from patients with MLL-AF9 AML (N=16) versus all the other subtypes of MLL-r AML (N=36) (GSE61804 and GSE19577). (C) Relative expression levels of Twist1 in BMMNC, granulocyte/macrophage progenitors (GMP), hematopoietic stem cell (HSC) from wild-type mice, and in GFP+c-kit+Gr1-, and leukemic GMP (L-GMP) cells from MLL-AF9 leukemic mice were evaluated by quantitative polymerase chain reaction (qPCR) (N=3). (D) Integrative genomics viewer plots showing signal tag density of H3K9me3 at the Twist1 gene in c-kit+ cells from MLL-AF9+ AML (MA9-1 and MA9-2) and wild-type control mice (GSE132175). (E) H3K9me3 chromatin immunoprecipitation (ChIP)-qPCR of Twist1 in c-kit+ cells from MLL-AF9+ AML and wild-type control mice. Genomic DNA was immunoprecipitated with anti-H3K9me3 antibody or immunglobulin (Ig)G control and ampliﬁed with primers spanning the Twist1 promoter and gene body sequences (N=3). (F) The expression of TWIST1 and KDM4C in KDM4C small hairpin RNA (shRNA)-transduced MOLM-13 and THP-1 cells was examined by quantitative real-time PCR (N=3). (G) Analysis of protein expression of TWIST1 by western blot. (H) ChIP-qPCR analysis of H3K9me3 and KDM4C enrichment on the Twist1 locus in MOLM-13 and THP-1 cells transduced with shCtrl or shKDM4C (N=3). Data are represented as mean ± standard deviation. *P<0.05; **P<0.01; ***P<0.001, by Student’s t test (A-C, E, F and H).

AF9-driven AML and control cells from wild-type mice. We found that the Twist1 locus showed a significantly lower level of repressive histone marks H3K9me3 in LSC (Figure 1D; Online Supplementary Figure S1). Furthermore, ChIP-quantitative polymerase chain reaction (qPCR) assays demonstrated that the enrichment of H3K9me3 at the promoter and gene body regions of the Twist1 locus was greatly decreased (by 2.7-4.7-fold) (Figure 1E). These results indicate that reduced repressive mark H3K9me3 of the Twist1 locus could contribute to the upregulation of Twist1 expression in LSC of MLL-AF9+ leukemia. H3K9me3 is tightly controlled by demethylase such as the KDM4 family which can remove H3K9me3 and play a strong regulatory inﬂuence on oncogenic gene transcription in AML.31 We analyzed a published dataset and found that, among the KDM4 family members, only KDM4C expression was positively correlated with TWIST1 in MLL-r AML samples (Online Supplementary Figure S2A-C). In addition, analysis of the published KDM4C ChIP-seq data on MLL-AF9-transformed LSC revealed that KDM4C bound to the promoter and spread into the gene body of Twist1 (Online Supplementary Figure S3A), whereas the binding was impaired upon loss of KDM4C (Online Supplementary Figure S3A). Published KDM4C ATAC-sequencing datasets from MOLM-13 cells showed a similar binding distribution at the TWIST1 locus (Online Supplementary Figure S3B). In order to verify these data, we knocked down KDM4C in MLL-AF9+ cell lines THP-1 and MOLM-13 or treated these cells with KDM4C inhibitor SD70, and measured TWIST1 expression. Inhibition of KDM4C resulted in significantly reduced expression of TWIST1 at both mRNA and protein levels (Figure 1F, G; Online Supplementary Figure S3C). In addition, ChIP-qPCR analysis demonstrated that KDM4C was significantly enriched at the transcription sites of the TWIST1 gene in MOLM-13 and THP-1 cells (Figure 1H). Knockdown of KDM4C led to markedly reduced recruitment of endogenous KDM4C at the TWIST1 loci (Figure 1H), with a corresponding increase in H3K9me3 level (Figure 1H). Taken together, these results indicate that KDM4C is required for TWIST1 expression in a H3K9me3 demethylation-dependent manner.

TWIST1 promotes the initiation and progression of MLLAF9-mediated acute myeloid leukemia In order to explore the role of TWIST1 in MLL-AF9+ AML, we performed transformation assays, both in vitro and in vivo, using a previously described Twist1-knockout mouse model. We transduced c-kit+ BM cells from Twist1-knockout and wild-type mice with a retrovirus expressing MLL-AF9 (Figure 2A). The resulting cells were referred to as Twist1-/- and control (Ctrl), respectively, in the figures hereafter. Colony formation and serial replating assays revealed that Twist1 deficiency led to a signiﬁcant reduction in colony-forming activity (Online Supplementary Figure S4A). This result was corroborated by BM transplantation (BMT). Equal numbers of MLL-AF9-transduced c-kit+ cells from Twist1 knockout or wild-type mice were transplanted into lethally irradiated recipients. Deletion of Twist1 was verified by quantitative reverse transcription PCR (qRT-PCR) in Twist1-/- BM cells from the transplantation recipients (Online Supplementary Figure S4B). The expression of Twist2 was not altered in Twist1-deficient BM cells (Online Supplementary Figure S4B). Mice transplanted with MLL-AF9-transduced wild-type cells developed AML and died within 41 days after transplantation (Figure 2B). In contrast, recipients transplanted with MLL-AF9-transduced cells from Twist1-/- mice showed a significantly prolonged survival (Figure 2B), suggesting that loss of Twist1 delays the onset of leukemia. Moreover, these recipients showed remarkable reductions in splenomegaly, white blood counts (WBC), and GFP+ AML cells in the BM and spleen compared with the control (Online Supplementary Figure S4C-F). The histological analysis also revealed prominently reduced leukemic infiltration of BM, spleen, and liver (Figure 2C), suggesting a decrease in disease burden upon Twist1 loss. Meanwhile, Twist1 deletion led to significantly enhanced apoptosis (Figure 2D), as well as differentiation of leukemia cells (Figure 2E; Online Supplementary Figure S4G). No obvious difference in proliferation was observed between Ctrl and Twist1-deleted leukemia cells (Online Supplementary Figure S4H). These results suggest that TWIST1 promotes the initiation of MLL-AF9-mediated AML by enhancing differentiation blockage and inhibiting apoptosis of leukemia cells.

Haematologica | 109 January 2024

ARTICLE - TWIST1 is required for leukemia stem cell function

We next assess the effect of Twist1 deficiency on the maintenance of established leukemia. We transplanted primary MLL-AF9 (ER-Cre+/-;Twist1ﬂox/ﬂox or ER-Cre+/-;Twist1wt/ wt ) AML cells into irradiated wild-type mice, followed by injection of tamoxifen (TAM) to ablate Twist1 (Figure 2F). As expected, the deletion of Twist1 significantly prolonged the survival of recipient mice and retarded the progression of leukemia (Figure 2G; Online Supplementary Figure S5A). Significantly decreased disease burden in the peripheral blood (PB), BM, and spleen (Online Supplementary Figure S5B-E), and increased myeloid differentiation of AML cells

N. Wang et al.

were observed upon Twist1 deletion (Online Supplementary Figure S5F, G). In order to further validate the role of TWIST1, we transplanted murine AML cells transduced with retroviral vectors containing Twist1 or control into irradiated recipient mice (Online Supplementary Figure S6A). Enforced expression of Twist1 signiﬁcantly accelerated MLL-AF9-induced AML development (Online Supplementary Figure S6B-D), and arrested myeloid differentiation (Online Supplementary Figure S6E). Together, our findings indicate that TWIST1 plays a critical role in the initiation and progression of MLL-AF9-induced AML.

Figure 2. TWIST1 is important for the initiation and maintenance of MLL-AF9-mediated acute myeloid leukemia. (A) Schematic of the colony formation and leukemic transplant experiments. (B) Kaplan-Meier survival curves of mice injected with control or Twist1-/- cells transformed with MLL-AF9 (N=5-6). (C) One hundred thousand MLL-AF9-transduced c-kit+ cells from wild-type or Twist1-knockout mice were transplanted into lethally irradiated recipients. The leukemia burden was examined 1 month after transplantation. Representative hematoxylin and eosin-stained histology sections of bone marrow (BM), spleen and liver. The length of the bars represents 100 μm. (D-E) Leukemic BM cells from sick primary mice were transplanted into the secondary recipients. The biological characteristics of leukemia cells were examined when the recipient mice became moribund. Apoptosis in GFP+ BM and GFP+ spleen cells (D, N=4-5). Frequency of leukemic myeloid Gr1+ CD11b+ cells in GFP+ BM and GFP+ spleen cells (E, N=3-4). (F) Schematic of the experimental procedure for evaluation of the role of TWIST1 in acute myeloid leukemia (AML) maintenance. (G) Kaplan-Meier survival curves of mice in AML maintenance experiment (N=10). Data are represented as mean ± standard deviation. *P<0.05; **P<0.01; ***P<0.001, by Student’s t test (D and E) or log-rank test (B and G).

Haematologica | 109 January 2024

ARTICLE - TWIST1 is required for leukemia stem cell function

TWIST1 is essential for the maintenance of leukemia stem cell In order to elucidate the role of TWIST1 in LSC, we evaluated the number and apoptotic status of the LSC from Twist1-/- AML recipients. Flow cytometric analysis revealed that the percentages of L-GMP and c-kit+Gr1- cells were markedly reduced in the BM and spleen of recipients of Twist1-/- AML cells (Figure 3A, B), probably due to the increase in apoptotic rate (Figure 3C). In addition, cell cycle analysis showed a significant decrease in the proportion of these cells in the G0 phase (Figure 3D). In order to assess the impact of Twist1 deletion on LSC function, we performed secondary transplantation into irradiated mice using leukemic whole BM cells from sick primary mice. Similar to primary transplantation, secondary recipient mice receiving Twist1-/- leukemia cells exhibited significantly longer leukemia-free survival (Online Supplementary Figure S7A). No significant difference in homing capacity was observed between Twist1-/- and control AML cells (Online Supplementary Figure S7B). In order to quantitate the functional LSC, we performed a limiting dilution assay (LDA). The frequency of LSC in Twist1-/- AML mice was dramatically reduced by approximately 69-fold compared to the control (1:21,472 vs. 1:310 cells) (Figure 3E). In addition, transplantation of Twist1-/- LSC also led to markedly delayed onset of AML compared to that of control LSC (Online Supplementary Figure S7C). Overall, these data suggest that TWIST1 is necessary for LSC survival and function in MLL-AF9+ AML. TWIST1 promotes leukemogenesis of human MLL-AF9+ acute myeloid leukemia In order to verify the function of TWIST1 in human MLL-AF9 leukemia, we performed small haipin RNA (shRNA)-mediated TWIST1 knockdown (KD) in THP-1 and MOLM-13 cell lines (Online Supplementary Figure S8A). TWIST1 insufficiency signiﬁcantly attenuated cell growth and colony formation, and promoted differentiation, while its induction of apoptosis was mild but statistically significant compared with the controls (Figure 4A, B; Online Supplementary Figure S8B, C). In order to ascertain if TWIST1 KD inhibits leukemia development in vivo, we developed a xenograft model of human AML. MOLM-13-luc2 cells transduced with TWIST1 shRNA or control scramble shRNA were injected into NOD/ SCID mice. TWIST1 KD led to remarkably reduced tumor burden (Figure 4C, D), and significantly prolonged survival in comparison with the control (Figure 4E). Collectively, our results demonstrate a critical role of TWIST1 in the propagation of human MLL-AF9+ AML cells. TWIST1 promotes the development of MLL-AF9-induced acute myeloid leukemia via activation of the WNT5a/ RAC1 axis In order to identify the downstream targets of TWIST1, we performed RNA sequencing on LSC from Twist1-/- and

Twist+/+ MLL-AF9 AML mice. A total of 1,033 genes were differentially expressed, of which 329 were downregulated and 704 were upregulated (log2 fold change >1, and P<0.05; Figure 5A). Gene ontology (GO) analysis revealed that regulation of Rho protein signal transduction was among the top enriched terms (Figure 5B). Enrichment of the Rho protein signal pathway was confirmed through gene set enrichment analysis (GSEA) (Figure 5C). RAC1, RHOA, and CDC42 are the three best-characterized members of the Rho GTPase family,32 and RAC1 plays an important role in MLL-AF9-induced leukemogenesis.33,34 In order to determine whether RAC1 is involved in TWIST1-mediated LSC maintenance, we first examined RAC1 activity using a Rac1-GTP pull-down assay in Twist1-deficient LSC and control cells. The results demonstrated that the depletion of Twist1 led to a significant decrease in RAC1 activity (Figure 5D), while RAC2 activity was unchanged (Figure 5D). We next examined whether RAC1 mediated the contribution of TWIST1 to MLLAF9-driven AML development. A constitutively active form of RAC1 (Q61L), a dominant-negative RAC1 mutant (T17N), or an empty vector was overexpressed in Twist1-deleted LSC, and the transduced cells were then injected into recipient mice (Figure 5E). As expected, overexpression of Q61L Rac1 markedly abrogated the impact of Twist1 ablation on murine survival (Figure 5F), spleen size, LSC frequency, and myeloid differentiation (Online Supplementary Figure S9A-C). In contrast, overexpression of T17N Rac1 or control vector showed no similar effects (Figure 5F; Online Supplementary Figure S9A-C). In summary, these data suggest that RAC1 activation mediates, at least partially, the pathogenic role of TWIST1 in MLL-AF9-driven AML. GSEA analysis also revealed significant negative enrichment of the non-canonical WNT signaling pathway in Twist1deficient LSC (Online Supplementary Table S1). Validation of the expression of genes associated with the Wnt pathway revealed that the expression levels of Wnt5a were most dramatically reduced (Figure 6A, B; primer sequences and antibodies are listed in Online Supplementary Tables S2 and S3). In addition, in a cohort of MLL-r AML, WNT5a expression correlates significantly with TWIST1 (Online Supplementary Figure S10). We then reintroduced TWIST1 expression in Twist1-deleted LSC, and found that the reintroduction of TWIST1 completely restored Wnt5a expression (Figure 6C). In order to investigate whether TWIST1 played a direct role in the transcription regulation of Wnt5a, we first searched the JASPAR database and found E-box consensus sequences in the promoter region of Wnt5a. We then generated luciferase constructs to examine the promoter activity of various lengths of Wnt5a 5’ flanking sequences (0.5, 1, 1.5, and 1.8 kb). Only the 1.8 kb fragment responded to TWIST1 overexpression by enhancing the luciferase activity (Figure 6D), indicating that TWIST1 bind to the Wnta5a promoter region from -1.8 kb to -1.5 kb. ChIP assay further confirmed a direct binding of endogenous TWIST1 to the Wnt5a promoter in LSC isolated from MLL-AF9+ AML mice (Figure

Haematologica | 109 January 2024

N. Wang et al.

ARTICLE - TWIST1 is required for leukemia stem cell function

6E). Taken together, these results indicate that TWIST1 is a transcriptional activator of WNT5a. Previous studies have reported that WNT5a promotes the self-renewal of tumor stem cells,35 and it induces RAC1 activation in a β-catenin-independent pathway.36,37 Therefore, we speculated that TWIST1 may activate RAC1 by

upregulating WNT5a. In order to test this, we ectopically expressed Wnt5a in Twist1-wild-type and Twist1-deficient LSC and analyzed RAC1 activity. The results revealed that enforced expression of Wnt5a in Twist1-wild-type LSC led to significantly enhanced activities of RAC1 (Figure 6F). In addition, WNT5a overexpression substantially diminished

N. Wang et al.

Figure 3. Twist1 deletion reduces the frequency and function of leukemic stem cell in MLL-AF9+ acute myeloid leukemia. (A-D) The frequency and biological characteristics of leukemic stem cells (LSC) were examined when the secondary recipient mice became moribund. Frequencies of c-kit+Gr1- subpopulations in the bone marrow (BM) and spleen (right) and representative fluorescence-activated cell sorting plots (left) (A, N=4-5). Frequencies of leukemic granulocyte/macrophage progenitor (L-GMP, Lin-c-kit+Sca1-CD127-CD34+CD16/32+) in the BM and spleen (B, N=4-5). Apoptosis analysis of c-kit+Gr1- LSC in the BM and spleen (C, N=4-5). Cell-cycle analysis of c-kit+Gr1- LSC in the BM and spleen (D, N=3-4). (E) Limiting dilution assay using BM cells from primary Twist1-deleted or control MLL-AF9+ acute myeloid leukemia mice. Logarithmic plot (left) showing the percentage of non-responding recipients transplanted with different cell doses of GFP+ cells. Table (right) showing the number of mice used as BM transplantation recipients in the assay and the number of recipients that developed leukemia. Data are represented as mean ± standard deviation. *P<0.05; **P<0.01; ***P<0.001, by Student’s t test (A-D) or χ2 test (E).

Haematologica | 109 January 2024

ARTICLE - TWIST1 is required for leukemia stem cell function

the inhibitory effect of Twist1 deletion on RAC1 activity (Figure 6F). We further transplanted Twist1-/- AML cells overexpressing Wnt5a or empty vector into recipient mice, and found that enforced expression of Wnt5a completely abrogated the changes in murine survival, leukemia cell infiltration, LSC frequency, and myeloid differentiation caused by Twist1 deletion (Figure 6G; Online Supplementary Figure S11A-D). These data suggest that the oncogenic

N. Wang et al.

role of TWIST1 in MLL-AF9-induced AML is mediated via activation of WNT5a-RAC1 signaling pathway. TWIST1 directly binds to HOXA9 and the co-expression of Twist1 and Hoxa9 induces fatal acute myeloid leukemia in mice Transcription factors often interact with other factors, in complexes, to alter DNA binding specificity or enhance

Figure 4. Human MLL-AF9+ acute myeloid leukemia cells are sensitive to TWIST1 knockdown. (A, B) Effects of TWIST1 knockdown on cell proliferation (A) and colony formation (B) in MOLM-13 and THP-1 cells, as determined by cell counting and colony-forming assay, respectively (N=3-5). (C-E) MOLM-13-luc2 cells were transfected with TWIST1 small hairpin RNA (shRNA) or control scramble shRNA and engrafted into NOD/SCID mice (8×105 cells per mouse). Bioluminescence imaging of representative mice from each group was taken on day 7 and day 14 post-transplantation (C). Quantification of the bioluminescence imaging at the indicated time points (D, N=3-4). Kaplan-Meier survival curves of MOLM13-transplanted mice (E, N=8). Data are represented as mean ± standard deviation. *P<0.05; **P<0.01; ***P<0.001, by Student’s t test (A, B and D) or log-rank test (E).

Haematologica | 109 January 2024

ARTICLE - TWIST1 is required for leukemia stem cell function

transcription.38 in order to investigate whether TWTST1 binds to any factors, we performed a liquid chromatography-tandem mass spectrometry analysis of endogenous TWIST1 immunoprecipitates in the LSC isolated from MLL-AF9driven AML mice, and identified HOXA9, the only protein reported to be essential for MLL-dependent leukemogenesis and LSC maintenance,9,10 as a potential TWIST1-interacting protein (dataset available via ProteomeXchange

N. Wang et al.

with identifier PXD037217). The interaction of TWIST1 with HOXA9 in the LSC was further confirmed by reciprocal co-immunoprecipitation using antibodies against either TWIST1 or HOXA9 (Figure 7A; Online Supplementary Figure S12A). Similar interactions were also observed in MOLM-13 and THP-1 cells (Figure 7A; Online Supplementary Figure S12A). In order to further test whether this interaction is direct, GST pull-down assay was performed using purified

Figure 5. TWIST1 accelerates the development of MLL-AF9-driven acute myeloid leukemia through activation of RAC1. (A) Volcano plots show the differences in gene expression between Twist1-deficient (cKO) leukemic stem cells (LSC) and the control cells (Ctrl). (B) Gene ontology (GO) term enrichment analysis of differentially expressed genes. Only the top ten GO terms are listed. (C) Gene set enrichment analysis showing reduced Rho protein signal pathway in Twist1-deficient LSC. The normalized enrichment score (NES) and P value are shown. (D) Western blot analysis of activated RAC1 (GTP-RAC1), total RAC1, activated RAC2 (GTP-RAC2), and total RAC2 in Twist1-deficient LSC and the control cells. (E) Schematic of the experimental procedure for evaluation of the role of RAC1 in TWIST1-mediated development of acute myeloid leukemia (AML). (F) Kaplan-Meier survival curves of mice transplanted with the indicated transduced cells (N=9-12). *P<0.05; **P<0.01; ***P<0.001, by log-rank test.

Haematologica | 109 January 2024

ARTICLE - TWIST1 is required for leukemia stem cell function

proteins of GST-tagged TWIST1 and MBP-tagged HOXA9. The GST-TWIST1, but not GST alone, could pull down MBPtagged HOXA9 (Figure 7B), suggesting that TWIST1 directly interacts with HOXA9. Our previous study revealed that the enforced expression of Twist1 promotes HSC self-renewal and myeloid skewing, but it is not sufficient to induce hematopoietic malignancies.21

Overexpression of individual Hoxa9 leads to myeloproliferative neoplasm in mice.15 Given that TWIST1 and HOXA9 are both highly expressed in patients with MLL-r AML and TWIST1 interacts with HOXA9, we next sought to determine whether co-expression of Twist1 and Hoxa9 could initiate AML. c-kit+ BM cells from wild-type mice co-transduced with Twist1 and Hoxa9 retroviruses were subjected to trans-

N. Wang et al.

Figure 6. Wnt5a is a direct functional target of TWIST1 and mediates TWIST1-induced RAC1 activation. (A) Relative mRNA expression levels of non-canonical Wnt pathway genes in Twist1-deficient leukemic stem cells (LSC) and control cells (Ctrl) (N=3). (B) Western blot analysis of WNT5a protein in Twist1-deficient LSC and control cells. (C) Western blot analysis of WNT5a protein in Ctrl and Twist1-/- LSC transduced with Twist1-expressing or control vector. (D) Promoter activity assay using Twist1 expression plasmid and luciferase reporter constructs driven by various lengths of the Wnt5a 5’ flanking region in the 293T cell line (N=3). (E) Chromatin immunoprecipitation quantitative polymerase chain reaction assay was performed with TWIST1 and immunoglobulin (Ig)G control in LSC (N=3). (F) Western blot analysis of GTP-RAC1 and total RAC1 in Ctrl and Twist1-/- LSC transduced with Wnt5a-expressing vector. (G) Kaplan-Meier survival curves of mice transplanted with Twist1-deficient LSC or control cells with or without Wnt5a overexpression. Data are represented as mean ± standard deviation. *P<0.05; **P<0.01; ***P<0.001, by Student’s t test (A, D and E) or log-rank test (G).

Haematologica | 109 January 2024

ARTICLE - TWIST1 is required for leukemia stem cell function

plant experiments (Figure 7C). The results revealed that the recipient mice transplanted with cells co-expressing Twist1 and Hoxa9 have significantly shorter survival than those with Hoxa9 or Twist1 alone (Figure 7D). Histological

N. Wang et al.

staining revealed that these mice developed fatal AML, characterized by expansion of myeloid blast cells and leukemic infiltration of the BM, spleen, liver, and kidney (Figure 7E; Online Supplementary Figure S12B), whereas Hoxa9 or Twist1

Figure 7. TWIST1 directly interacts and collaborates with HOXA9 to induce acute myeloid leukemia in transplanted mice. (A) Co-immunoprecipitation (Co-IP) and immunoblot (IB) analyses in murine leukemic stem cells (LSC), MOLM-13, and THP-1 cells using anti-HOXA9 or anti-rabbit immunoglobulin (Ig)G for IP and anti-HOXA9 or anti-TWIST1 for IB. (B) GST pull-down assay was performed by using bacterially purified GST-TWIST1 and MBP-HOXA9, followed by immunoblotting. GST and MBP alone were used as negative controls, respectively. (C) Schematic of the experimental procedure for evaluation of the role of co-expression of Twist1 and Hoxa9 in inducing acute myeloid leukemia. (D) Kaplan-Meier curves of mice transplanted with c-kit+ cells transduced with the indicated genes, and secondary recipient mice transplanted with primary leukemic bone marrow (BM) cells of the Hoxa9/ Twist1 group (N=5-8). (E) Representative images of tissue morphologies. The BM, spleen, liver, and kidney tissues were stained with hematoxylin and eosin. The length of bars represents 100 μm. (F) Flow cytometric analysis of c-kit+CD11b- cells of BM and spleen from the recipient mice. *P<0.05; ***P<0.001, by log-rank test (D).

Haematologica | 109 January 2024

ARTICLE - TWIST1 is required for leukemia stem cell function

overexpression alone did not initiate leukemia, consistent with the previous reports.15,21 Fluorescence-activated cell sorting analysis of BM and spleen cells from the recipient mice showed that co-expression of Twist1 and Hoxa9 led to a higher proportion of c-kit+CD11b- blast cells than those expressing Hoxa9 alone (Figure 7F). In order to determine whether Twist1/Hoxa9-induced AML is transplantable, we performed secondary transplantation with leukemic whole BM cells from diseased primary mice. The secondary recipients transplanted with leukemic cells co-expressing Twist1 and Hoxa9 exhibited signiﬁcantly shortened survival (median 27 days) than the primary ones (median 227 days) (Figure 7D), indicating a more aggressive disease. In sum, TWIST1 cooperates with HOXA9 to induce AML. We next carried out RNA sequencing to characterize transcriptional changes upon Hoxa9 knockdown in LSC from MLL-AF9+ AML mice. A total of 324 genes were downregulated and 367 upregulated (log2 fold change >1, and P<0.05; Online Supplementary Figure S13A). GO and Reactome pathway analyses revealed that, consistent with Twist1 deletion, the non-canonical WNT signaling and Rho GTPase signaling were also significantly negatively enriched (Online Supplementary Figure S13B, C), indicating that the WNT5a-RAC1 signaling pathway may be the common downstream mediator for the pathogenic effects of HOXA9 and TWIST1. These data provide further support for the collaboration between TWIST1 and HOXA9 in leukemogenesis.

Discussion High relapse rate leads to high mortality in patients with MLL-r AML, indicating the importance of targeted therapy for LSC. Here, our data demonstrate that TWIST1 is highly expressed in BMMNC from patients with MLL-AF9 AML, and its level in murine LSC is strikingly higher than that in normal HSC. In addition, TWIST1 plays essential roles in maintaining the survival and function of LSC, and promotes the leukemogenesis of MLL-AF9+ AML. Although our previous study showed that TWIST1 preserves dormancy and self-renewal capacities of HSC,20 Twist1 knockout does not induce apoptosis of HSC, inconsistent with its role in LSC, suggesting that TWIST1 is much more important for LSC than HSC. More and more modulators of EMT, a crucial process for the invasion and metastasis of epithelial tumors, have been identified as tumor promoting factors in AML.39,40 However, because of the lack of relevance in non-EMT contexts, the specific roles of these EMT regulators in AML remain enigmas. Here, we demonstrate that TWIST1, a key EMT-inducing transcription factor, plays crucial roles in LSC maintenance, consistent with what has been observed for SNAI1, ZEB2 and AXL, three other factors regulating EMT.39,41,42 In addition, as a TWIST1 downstream mediator in the present study, Rac1 GTPase has also been implicated in EMT trigger,43 as

well as LSC function.44,45 It is documented that activation of the EMT program in epithelial cells induces the acquisition of stem cell properties.19 Therefore, it appears that the relevance of EMT with stemness in solid tumors extends to hematologic malignancies, in which the EMT inducing factors promote LSC maintenance. Recently, several compounds specifically targeting EMT-associated AXL kinase in AML LSC have advanced into clinical trials.42 Albeit caution needs to be exercised in developing therapeutic drugs targeting TWIST1, as our previous study demonstrates that, opposite to the tumor-promoting role in LSC, TWIST1 expressed in niche cells exerts anti-leukemia activities.27 It appears that the roles of TWIST1 in hematological malignancies are cell-type dependent. Despite the extraordinary diversity of fusion partners of MLL-r AML, the presence of common core regulators implies that targeting such factors will be broadly applicable to MLL-r leukemia. High expression of HOXA9 is an important step toward MLL-fusion-mediated leukemic transformation.10 The ability for HOXA9 to drive leukemia development is dependent on interaction with different co-factors.46 The most important co-factors are MEIS1 and PBX3, which mainly play supportive roles in promoting HOXA9 to activate or repress the transcription of downstream targets.12,15 Co-expression of HOXA9 with either MEIS1 or PBX3 promotes leukemogenesis, while MEIS1 or PBX3 expression alone is insufficient to initiate HSPC transformation.12,15 Our present study demonstrates for the first time that TWIST1 not only directly interacts with HOXA9, but also cooperates with HOXA9 to induce AML. These findings indicate that TWIST1 acts as a novel co-factor of HOXA9. Growing evidence shows that decreased H3K9me3 accelerates AML disease progression,47-50 but the underlying gene loci have not been well defined. Our findings demonstrate that TWIST1 is one of the related loci marked by a low level of H3K9me3 in LSC. Furthermore, we demonstrate that lysine demethylase KDM4C is required for TWIST1 expression in LSC of MLL-AF9+ AML by removing H3K9me3 silencing marks on the TWIST1 locus. KDM4C has been implicated in various malignancies, acting as an oncoprotein.47,51 Recent reports have identified the roles of KDM4C in maintaining AML LSC,48 and in the leukemogenesis of AML.47 We observed that ectopic expression of TWIST1 partially rescued the reduction in colony formation and enhancement in apoptosis of human MLL-AF9+ AML cells caused by KDM4C knockdown (Online Supplementary Figure S14A, B), suggesting that the tumor-promoting activities of KDM4C are partly mediated by TWIST1. In preclinical models of MLL-r AML, KDM4C specific inhibitor SD70 can suppress the transcription and transformation ability of MLL fusions in AML cells.47 Our findings provide new insights into the molecular mechanisms underlying its anti-leukemia activity. In conclusion, our studies demonstrate for the first time that TWIST1 plays a critical role in LSC maintenance, and it directly interacts and collaborates with HOXA9 in induc-

Haematologica | 109 January 2024

N. Wang et al.

ARTICLE - TWIST1 is required for leukemia stem cell function

ing AML in mice. Our results provide new insights into the leukemogenesis and suggest potential therapeutic target for treatment of MLL-AF9+ AML.

N. Wang et al.

provided plasmid constructs. All authors approved the final version of the manuscript.

Disclosures No conflicts of interest to disclose. Contributions XTM designed the project. XTM and NW wrote the manuscript. NW organized the experiments and performed statistical analysis of the data. NW, JY, NY, WQZ, NNG, XYL, PWZ, WLH and YQX performed experiments and analyzed data. QR

Funding This work was supported by grants from CAMS Innovation Fund for Medical Sciences (2021-I2M-1-019), National Natural Science Foundation of China (82270122, 82070113, 81730006). Data-sharing statement All data are available upon request sent to the corresponding author.

References 16. Thorsteinsdottir U, Kroon E, Jerome L, Blasi F, Sauvageau G. Defining roles for HOX and MEIS1 genes in induction of acute myeloid leukemia. Mol Cell Biol. 2001;21(1):224-234. 17. Jin G, Yamazaki Y, Takuwa M, et al. Trib1 and Evi1 cooperate with Hoxa and Meis1 in myeloid leukemogenesis. Blood. 2007;109(9):3998-4005. 18. Vesuna F, Lisok A, Kimble B, Raman V. Twist modulates breast cancer stem cells by transcriptional regulation of CD24 expression. Neoplasia. 2009;11(12):1318-1328. 19. Singh A, Settleman J. EMT, cancer stem cells and drug resistance: an emerging axis of evil in the war on cancer. Oncogene. 2010;29(34):4741-4751. 20. Wang N, Yin J, You N, et al. TWIST1 preserves hematopoietic stem cell function via the CACNA1B/Ca2+/mitochondria axis. Blood. 2021;137(21):2907-2919. 21. Dong CY, Liu XY, Wang N, et al. Twist-1, a novel regulator of hematopoietic stem cell self-renewal and myeloid lineage development. Stem Cells. 2014;32(12):3173-3182. 22. Wang N, Guo D, Zhao YY, et al. TWIST-1 promotes cell growth, drug resistance and progenitor clonogenic capacities in myeloid leukemia and is a novel poor prognostic factor in acute myeloid leukemia. Oncotarget. 2015;6(25):20977-20992. 23. Lin J, Zhang W, Niu LT, et al. TRIB3 stabilizes high TWIST1 expression to promote rapid APL progression and ATRA resistance. Clin Cancer Res. 2019;25(20):6228-6242. 24. Li H, Wang Y, Pang X, et al. Elevated TWIST1 expression in myelodysplastic syndromes/acute myeloid leukemia reduces efficacy of hypomethylating therapy with decitabine. Haematologica. 2020;105(10):e502. 25. Luo H, Zhu G, Xu J, et al. HOTTIP lncRNA promotes hematopoietic stem cell self-renewal leading to AML-like disease in mice. Cancer Cell. 2019;36(6):645-659. 26. Singh AP, Luo H, Matur M, et al. A coordinated function of lncRNA HOTTIP and miRNA-196b underpinning leukemogenesis by targeting FAS signaling. Oncogene. 2022;41(5):718-731. 27. Liu X, Ma Y, Li R, et al. Niche TWIST1 is critical for maintaining normal hematopoiesis and impeding leukemia progression. Haematologica. 2018;103(12):1969-1979. 28. Somervaille TC, Cleary ML. Identification and characterization of leukemia stem cells in murine MLL-AF9 acute myeloid leukemia. Cancer Cell. 2006;10(4):257-268. 29. Milne TA, Briggs SD, Brock HW, et al. MLL targets SET domain methyltransferase activity to Hox gene promoters. Mol Cell. 2002;10(5):1107-1117.

1. Estey EH. Acute myeloid leukemia: 2013 update on riskstratification and management. Am J Hematol. 2013;88(4):318-327. 2. Tsakaneli A, Williams O. Drug repurposing for targeting acute leukemia with KMT2A (MLL)-gene rearrangements. Front Pharmacol. 2021;12:741413. 3. Krivtsov AV, Twomey D, Feng Z, et al. Transformation from committed progenitor to leukaemia stem cell initiated by MLL-AF9. Nature. 2006;442(7104):818-822. 4. Wong NM, So CWE. Novel therapeutic strategies for MLLrearranged leukemias. Biochim Biophys Acta Gene Regul Mech. 2020;1863(9):194584. 5. Takahashi S, Yokoyama A. The molecular functions of common and atypical MLL fusion protein complexes. Biochim Biophys Acta Gene Regul Mech. 2020;1863(7):194548. 6. Muntean AG, Hess JL. The pathogenesis of mixed-lineage leukemia. Annu Rev Pathol. 2012;7:283-301. 7. Meyer C, Burmeister T, Groger D, et al. The MLL recombinome of acute leukemias in 2017. Leukemia. 2018;32(2):273-284. 8. Armstrong SA, Staunton JE, Silverman LB, et al. MLL translocations specify a distinct gene expression profile that distinguishes a unique leukemia. Nat Genet. 2002;30(1):41-47. 9. Faber J, Krivtsov AV, Stubbs MC, et al. HOXA9 is required for survival in human MLL-rearranged acute leukemias. Blood. 2009;113(11):2375-2385. 10. Ayton PM, Cleary ML. Transformation of myeloid progenitors by MLL oncoproteins is dependent on Hoxa7 and Hoxa9. Genes Dev. 2003;17(18):2298-2307. 11. Zeisig BB, Milne T, Garcia-Cuellar MP, et al. Hoxa9 and Meis1 are key targets for MLL-ENL-mediated cellular immortalization. Mol Cell Biol. 2004;24(2):617-628. 12. Li Z, Zhang Z, Li Y, et al. PBX3 is an important cofactor of HOXA9 in leukemogenesis. Blood. 2013;121(8):1422-1431. 13. Collins CT, Hess JL. Role of HOXA9 in leukemia: dysregulation, cofactors and essential targets. Oncogene. 2016;35(9):1090-1098. 14. Li Z, Huang H, Li Y, et al. Up-regulation of a HOXA-PBX3 homeobox-gene signature following down-regulation of miR-181 is associated with adverse prognosis in patients with cytogenetically abnormal AML. Blood. 2012;119(10):2314-2324. 15. Kroon E, Krosl J, Thorsteinsdottir U, Baban S, Buchberg AM, Sauvageau G. Hoxa9 transforms primary bone marrow cells through specific collaboration with Meis1a but not Pbx1b. EMBO J. 1998;17(13):3714-3725.

Haematologica | 109 January 2024

ARTICLE - TWIST1 is required for leukemia stem cell function

30. Krivtsov AV, Armstrong SA. MLL translocations, histone modifications and leukaemia stem-cell development. Nat Rev Cancer. 2007;7(11):823-833. 31. Monaghan L, Massett ME, Bunschoten RP, et al. The emerging role of H3K9me3 as a potential therapeutic target in acute myeloid leukemia. Front Oncol. 2019;9:705. 32. Karlsson R, Pedersen ED, Wang Z, Brakebusch C. Rho GTPase function in tumorigenesis. Biochim Biophys Acta. 2009;1796(2):91-98. 33. Mizukawa B, Wei J, Shrestha M, et al. Inhibition of Rac GTPase signaling and downstream prosurvival Bcl-2 proteins as combination targeted therapy in MLL-AF9 leukemia. Blood. 2011;118(19):5235-5245. 34. Nimmagadda SC, Frey S, Edelmann B, et al. Bruton’s tyrosine kinase and RAC1 promote cell survival in MLL-rearranged acute myeloid leukemia. Leukemia. 2018;32(3):846-849. 35. Zhou Y, Kipps TJ, Zhang S. Wnt5a signaling in normal and cancer stem cells. Stem Cells Int. 2017;2017:5295286. 36. Naskar D, Maiti G, Chakraborty A, Roy A, Chattopadhyay D, Sen M. Wnt5a-Rac1-NF-kappaB homeostatic circuitry sustains innate immune functions in macrophages. J Immunol. 2014;192(9):4386-4397. 37. Akoumianakis I, Sanna F, Margaritis M, et al. Adipose tissuederived WNT5A regulates vascular redox signaling in obesity via USP17/RAC1-mediated activation of NADPH oxidases. Sci Transl Med. 2019;11(510):eaav5055. 38. Lelli KM, Slattery M, Mann RS. Disentangling the many layers of eukaryotic transcriptional regulation. Annu Rev Genet. 2012;46:43-68. 39. Carmichael CL, Wang J, Nguyen T, et al. The EMT modulator SNAI1 contributes to AML pathogenesis via its interaction with LSD1. Blood. 2020;136(8):957-973. 40. Almotiri A, Alzahrani H, Menendez-Gonzalez JB, et al. Zeb1 modulates hematopoietic stem cell fates required for suppressing acute myeloid leukemia. J Clin Invest. 2021;131(1):e129115.

41. Li H, Mar BG, Zhang H, et al. The EMT regulator ZEB2 is a novel dependency of human and murine acute myeloid leukemia. Blood. 2017;129(4):497-508. 42. Niu X, Rothe K, Chen M, et al. Targeting AXL kinase sensitizes leukemic stem and progenitor cells to venetoclax treatment in acute myeloid leukemia. Blood. 2021;137(26):3641-3655. 43. Yang WH, Lan HY, Huang CH, et al. RAC1 activation mediates Twist1-induced cancer cell migration. Nat Cell Biol. 2012;14(4):366-374. 44. Chen S, Li H, Li S, et al. Rac1 GTPase promotes interaction of hematopoietic stem/progenitor cell with niche and participates in leukemia initiation and maintenance in mouse. Stem Cells. 2016;34(7):1730-1741. 45. Grey W, Rio-Machin A, Casado P, et al. CKS1 inhibition depletes leukemic stem cells and protects healthy hematopoietic stem cells in acute myeloid leukemia. Sci Transl Med. 2022;14(650):eabn3248. 46. de Bock CE, Demeyer S, Degryse S, et al. HOXA9 cooperates with activated JAK/STAT signaling to drive leukemia development. Cancer Discov. 2018;8(5):616-631. 47. Cheung N, Fung TK, Zeisig BB, et al. Targeting aberrant epigenetic networks mediated by PRMT1 and KDM4C in acute myeloid leukemia. Cancer Cell. 2016;29(1):32-48. 48. Wang J, Li Y, Wang P, et al. Leukemogenic chromatin alterations promote AML leukemia stem cells via a KDM4C-ALKBH5-AXL signaling axis. Cell Stem Cell. 2020;27(1):81-97. 49. Chu Y, Chen Y, Guo H, et al. SUV39H1 regulates the progression of MLL-AF9-induced acute myeloid leukemia. Oncogene. 2020;39(50):7239-7252. 50. Wang Y, Lu T, Sun G, et al. Targeting of apoptosis gene loci by reprogramming factors leads to selective eradication of leukemia cells. Nat Commun. 2019;10(1):5594. 51. Ernst P, Schnoder TM, Huber N, et al. Histone demethylase KDM4C is a functional dependency in JAK2-mutated neoplasms. Leukemia. 2022;36(7):1843-1849.

Haematologica | 109 January 2024

N. Wang et al.

ARTICLE - Acute Myeloid Leukemia

SUMOylation inhibitor TAK-981 (subasumstat) synergizes with 5-azacytidine in preclinical models of acute myeloid leukemia Ludovic Gabellier,1,2* Marion De Toledo,1* Mehuli Chakraborty,1 Dana Akl,1 Rawan Hallal,1 Mays Aqrouq,1 Giovanni Buonocore,1 Clara Recasens-Zorzo,1 Guillaume Cartron,1,2 Abigaïl Delort,3 Marc Piechaczyk,1 Denis Tempé1 and Guillaume Bossis1

Correspondence: M. De Toledo marion.detoledo@igmm.cnrs.fr

IGMM, University of Montpellier, CNRS; 2Service d’Hématologie Clinique, CHU de Montpellier

D. Tempé denis.tempe@igmm.cnrs.fr

and 3MGX, University of Montpellier, CNRS, INSERM, Montpellier, France

G. Bossis guillaume.bossis@igmm.cnrs.fr

Received: Accepted: Early view:

LG and MDT contributed equally as first authors.

January 6, 2023. August 16, 2023. August 24, 2023.

Abstract Acute myeloid leukemias (AML) are severe hematomalignancies with dismal prognosis. The post-translational modification SUMOylation plays key roles in leukemogenesis and AML response to therapies. Here, we show that TAK-981 (subasumstat), a first-in-class SUMOylation inhibitor, is endowed with potent anti-leukemic activity in various preclinical models of AML. TAK-981 targets AML cell lines and patient blast cells in vitro and in vivo in xenografted mice with minimal toxicity on normal hematopoietic cells. Moreover, it synergizes with 5-azacytidine (AZA), a DNA-hypomethylating agent now used in combination with the BCL-2 inhibitor venetoclax to treat AML patients unfit for standard chemotherapies. Interestingly, TAK-981+AZA combination shows higher anti-leukemic activity than AZA+venetoclax combination both in vitro and in vivo, at least in the models tested. Mechanistically, TAK-981 potentiates the transcriptional reprogramming induced by AZA, promoting apoptosis, alteration of the cell cycle and differentiation of the leukemic cells. In addition, TAK-981+AZA treatment induces many genes linked to inflammation and immune response pathways. In particular, this leads to the secretion of type-I interferon by AML cells. Finally, TAK-981+AZA induces the expression of natural killer-activating ligands (MICA/B) and adhesion proteins (ICAM-1) at the surface of AML cells. Consistently, TAK-981+AZA-treated AML cells activate natural killer cells and increase their cytotoxic activity. Targeting SUMOylation with TAK-981 may thus be a promising strategy to both sensitize AML cells to AZA and reduce their immune-escape capacities.

Introduction Acute myeloid leukemias (AML) are severe hematologic malignancies resulting from the acquisition of oncogenic mutations by hematopoietic stem or progenitor cells. AML cells, which are blocked at intermediate stages of differentiation, proliferate and infiltrate the bone marrow, thereby disrupting normal hematopoiesis.1 Fit patients are usually treated with intensive chemotherapy based on the combination of an anthracycline (daunorubicin [DNR] or idarubicin) and the nucleoside analogue, cytarabine (ARA-C). Relapses are, however, frequent and overall survival (OS) remains

very poor. In the past few years, various new molecules improving AML prognosis have been approved. In most cases, they target mutated oncogenes such as IDH1/2 or FLT3, which restricts their use to patients carrying these mutations.2 Unfit patients who cannot receive chemotherapy because of age or comorbidities are generally treated with hypomethylating agents, in particular 5-azacytidine (AZA).3 In cells, this cytidine analog is metabolized into 5-aza-dCTP and incorporated into DNA during replication, where it can form covalent adducts with DNA-methyl transferases (DNMT). This triggers ubiquitin-proteasome-dependent depletion of DNMT4 and results in a progressive

Haematologica | 109 January 2024

ARTICLE - TAK-981 synergizes with 5-azacytidine in AML

L. Gabellier et al.

loss of DNA methylation at CpG dinucleotide motifs. The prevalent model to explain the therapeutic effect of AZA is that reduced methylation of CpG leads to the reactivation of silenced tumor suppressor genes as well as genes involved in differentiation, which are generally hypermethylated at cis-regulatory regions in cancer cells.5 The clinical benefit of AZA treatment is, however, limited with a 4- to 5-month increased OS compared to other AML therapies.6 Its combination with the BCL2 inhibitor venetoclax (VEN) significantly improves patient response and OS.7,8 This combo is now used as first-line therapy for patients unfit for standard chemotherapies. Nevertheless, a proportion of AML patients respond poorly to this regimen or acquire resistance.9,10 SUMO proteins are peptide post-translational modifiers with structural homology to ubiquitin. Whilst human genome encodes for five SUMO genes (SUMO-1 to -5), the main conjugated isoforms are SUMO-1, -2 and -3, the latter two being almost identical.11 SUMOylation involves a SUMO-activating E1 enzyme (UBA2/SAE1), a SUMO-conjugating E2 enzyme (UBC9) and several E3 factors. SUMOylation is highly dynamic thanks to various isopeptidases, which can release SUMO from conjugated targets. SUMO are conjugated to lysines of thousands of proteins (>6,000 identified in cancer cells,12 around 1,000 in healthy mouse tissues13) to modify their function and fate.14 As such, SUMOylation has been implicated in the regulation of most cellular functions.11 One of its best-characterized roles concerns the regulation of gene expression.15,16 We have previously shown that SUMOylation limits the anti-leukemic activity of both chemotherapies (DNR and Ara-C)17 and differentiation therapies18 in AML. This suggested that targeting the SUMO pathway could improve AML responses to therapies. A recent breakthrough in the field of SUMOylation is the discovery of TAK-981 (subasumstat), a first-in-class SUMO E1 inhibitor with very high potency and specificity.19 TAK981 has potent anti-tumoral activity in syngenic mouse models grafted with murine lymphoma or pancreatic tumor cells through the induction of a strong type-1 interferon (IFN-I)-dependent anti-tumor immune response.20,21 Indeed, TAK-981 activates dendritic cells, cytotoxic CD8+ T cells, memory B cells, natural killer (NK) cells and macrophages.20-24 Moreover, TAK-981 increases antigen presentation by cancer cells, further enhancing anti-tumor immune response.22 In addition to these effects on the immune micro-environment, TAK-981 can directly induce cancer cells death.21,25-27 However, the relative contribution of the direct and indirect anti-tumoral activities of TAK-981 remains to be clarified. Here, we have addressed the therapeutic potential of TAK981 in AML. We found that TAK-981 has potent anti-leukemic activity, in particular when combined with AZA. TAK-981 exacerbates the transcriptional reprogramming induced by AZA. In addition to genes involved in differentiation and apoptosis, TAK-981+AZA induces inflammation- and immune

response-related transcriptional programs. In particular, AML cells exposed to TAK-981+AZA show increased secretion of IFN-I. Finally, they express, at their membrane, NK adhesion molecules (ICAM-1) and ligands of NK activating receptors (MICA/B), leading to an enhanced NK cell-mediated cytotoxicity towards AML cells. Altogether, our data suggest that combining the inhibition of SUMOylation by TAK-981 and DNA methylation by AZA could be a promising strategy for AML treatment.

Methods Bioluminescent cell line generation, patient cell cultures, flow cytometry and RNA-sequencing (RNA-seq) analysis are described in the Online Supplemental Appendix. Pharmacological inhibitors and reagents TAK-981 was obtained from Takeda Development Center Americas, Inc. and 5-azacytidine (AZA) was from StemCell and resuspended in RPMI prior to each experiment. DNR and aracytine (aracytosine-β-D-arabinofuranoside-ARA-C) were from Sigma-Aldrich, VEN from MedChemExpress. All antibodies used are described in Online Supplementary Table S1. Cell lines and culture conditions Human AML cell lines (U937, THP-1, HL-60, MOLM14, MV4.11) were obtained by the American Type Culture Collection and regularly tested for Mycoplasma contamination. They were cultured as previously described28 at 37°C in the presence of 5% CO2 in RPMI 1640 medium supplemented with 10% heat-decomplemented fetal bovine serum, penicillin and streptomycin. Cells were seeded at 0.3x106/mL 1 day before being drug-treated. Patient and healthy donor samples Bone marrow aspirates and blood samples were collected after obtaining written informed consent from patients or donors under the frame of the Declaration of Helsinki and after approval by the Institutional Review Board (Ethical Committee “Sud Méditerranée 1,” ref 2013-A00260-45, HemoDiag collection). Healthy donor leukocytes were collected from blood donors of the Montpellier Etablissement Français du Sang. Fresh leukocytes were purified by density based centrifugation using Histopaque 1077 (Sigma Aldrich). NK cells were purified using EasySep Human NK Cell Isolation kit (StemCell Technologies). Acute myeloid leukemia mouse xenograft model All experiments on animals were approved by the Ethics Committee of the Languedoc-Roussillon (2018043021198029 #14905 v3). For cell line-derived xenograft (CLDX) and patient-derived xenograft (PDX) experiments, female NOD/ LtSz-SCID/IL-2Rγchain null (NSG) mice (Charles River) (6-10

Haematologica | 109 January 2024

ARTICLE - TAK-981 synergizes with 5-azacytidine in AML

L. Gabellier et al.

weeks old) were treated using respectively 20 mg/kg and 30 mg/kg busulfan intravenous injections (SIGMA B2635, France) 48 hours before cell engraftment. 1x106 cells (cell lines) or 1.5x106 cells (patient’s cells) were injected in the tail vein.

As several rounds of replication are necessary to obtain maximal AZA-induced hypomethylation, we performed viability assays after 72 hours of treatment. Under these conditions, all cell lines showed sensitivity to TAK-981 alone, HL-60 and U937 being however less sensitive than THP-1 cells (Online Supplementary Figure S1B). A synergistic cytotoxicity was seen between TAK-981 and AZA for these three cell lines (Figures 1E; Online Supplementary Figure S1D). TAK-981 and AZA combination was also more efficient than the single treatments on MOLM14 and MV4.11 cell lines (Online Supplementary Figure S1E). For all cell lines, AZA+TAK-981 combination was more efficient than AZA+VEN combination at decreasing AML cell line viability in vitro (note the differences in VEN and TAK-981 doses used; Online Supplementary Figure S1E). We then treated primary AML cells from 17 different patients. Both AZA and TAK-981 treatment led to a significant reduction in the number of leukemic cells, but the most important effect was obtained with the combination of the two drugs, with however some variability between patient samples (Figure 1F; Online Supplementary Table S2). This variability might be related to the cytogenetic characteristics of the patients, as those with abnormal karyotypes were more sensitive to TAK-981+AZA than those with a normal karyotype (Online Supplementary Figure S2A). In addition, patients from the more differentiated M4/M5 subgroups of French American British (FAB) classification were more sensitive to TAK-981 than those from the more immature M1/M2 subgroups (Online Supplementary Figure S2B). Finally, treatments had no significant effects on bone marrow-derived mononuclear cells (BMMC) from healthy donors cultured under the same conditions, although TAK-981, when used alone, tend to slightly increase their numbers (Figure 1G). Altogether, these data suggest that inhibition of SUMOylation with TAK-981 affects the viability of AML cells and synergizes in vitro with the hypomethylating agent AZA.

Assessment of natural killer cell cytotoxicity towards acute myeloid leukemia cells Target cells (THP-1-LucZsGreen) were co-cultured at a 1:1 ratio in 96-well plates with primary NK cells purified from fresh peripheral blood mononucelar cells (PBMC) of healthy donors. Real-time fluorescence was assessed for 15 hours using the IncuCyte S3 Live Cell imaging system (Sartorius) in the non-adherent cell-by-cell mode, using a 20X objective, four images/well and one image/hour. Analyses were performed by measuring the relative integrated green fluorescence intensity using IncuCyte 2021C software. Statistical analyses Statistical analyses of the differences between data sets were performed using one-way ANOVA for normal distribution data and Kruskal-Wallis or Friedman tests for non-Gaussian distribution data (GraphPad Prism, GraphPad v9.4.0). Overall mouse survivals were estimated for each treatment group using the Kaplan-Meier method and compared with the log-rank test. P values of less than 0.05 were considered significant (*P< 0.05; **P< 0.01 and ***P< 0.001; NS: not significant).

Results TAK-981 synergizes with 5-azacytidine to induce acute myeloid leukemia cell death in vitro In order to characterize the effect of the SUMOylation inhibitor TAK-981 on AML cells, we first monitored its impact on SUMOylation in three model AML cell lines, HL-60, U937 and THP-1. After 24 hours, TAK-981 induced a strong decrease in global SUMOylation by SUMO-1 and SUMO-2/3 at 10 nM and an almost complete loss at 100 nM (Figure 1A; Online Supplementary Figure S1A). Although the extent of deSUMOylation was similar between the three cell lines, U937 and THP-1 cells were highly sensitive to TAK-981, whilst HL-60 were unaffected after 24 hours of treatment (Figure 1B). Importantly, TAK-981 did not significantly affect the viability of normal mononuclear cells derived from either bone marrow (BMDMC) or peripheral blood (PBMC) under the same conditions (Figure 1C). Then, in order to determine whether SUMOylation inhibition could synergize with the drugs most commonly used for AML treatment, we combined TAK-981 with either DNR, ARA-C or AZA to treat HL-60, the least sensitive cell line to TAK-981, for 24 hours (Figure 1D; Online Supplementary Figure S1C). TAK981 synergized with all three drugs, the strongest synergy being with AZA (Figures 1D).

TAK-981 +5-azacytidine combination has anti-leukemic activity in preclinical acute myeloid leukemia models In order to further delineate the therapeutical potential of the TAK-981+AZA combination, we performed in vivo experiments using NOD-SCID-gammaIL2Rnull (NSG) mice. First, to ensure the efficiency of TAK-981 in vivo, we used a microbead-based assay29,30 to monitor its ability to inhibit SUMOylation activity. TAK-981 treatment led to a 70% decrease in SUMOylation activity in the bone marrow of NSG mice after 5 hours and a progressive recovery to basal levels after 24 hours (Online Supplementary Figure S3A). Second, NSG mice were grafted intravenously with bioluminescent THP-1 (Online Supplementary Figure S3B) or U937 cells (Online Supplementary Figure S4A). TAK-981 and AZA monotherapies limited tumor progression (Figure 2A; Online Supplementary Figures 3C and S4B, C) and significantly extended mice survival (Figure 2B; Online Supplementary Figure S4D) for both THP-1 and U937. TAK-981+AZA com-

Haematologica | 109 January 2024

100

ARTICLE - TAK-981 synergizes with 5-azacytidine in AML

L. Gabellier et al.

Continued on following page.

Haematologica | 109 January 2024

101

ARTICLE - TAK-981 synergizes with 5-azacytidine in AML

L. Gabellier et al.

Figure 1. TAK981 synergizes with 5-azacytidine to induce acute myeloid leukemia cell death in vitro. (A) HL-60 cells were treated with increasing doses of TAK-981 for 24 hours and immunoblotted for SUMO-1, SUMO-2/3 and GAPDH. SUMO conjugates appear as smears. (B) TAK-981 half-maximal inhibitory concentrations (IC50) were determined using HL-60, U937 and THP-1 cells treated with varying drug concentrations. Cell viability was determined using MTS assays after 24 hours of treatment. Concentration response curves were generated by comparing the viability of TAK-981-treated cells with mock-treated controls. Data are shown as mean +/- standard error of the mean (SEM) of replicate samples (N=5). Absolute IC50 are indicated in the figure. (C) Bone marrow-derived mononuclear cells (BMMC) (N=2) and peripheral blood mononulcear cells (PBMC) (N=3) collected from healthy donors were treated with varying TAK-981 concentrations for 24 hours. Cell viability was determined by flow cytometry after 24 hours of treatment. Concentration response curves were generated by comparing the viability of TAK-981-treated cells with mock-treated controls. Data are shown as mean +/- SEM of replicate samples. (D) HL-60 cells were treated for 24 hours with TAK-981 and either daunorubicin (DNR), cytarabine (ARA-C) or 5-azacytidine (AZA) and viability was assessed by MTS assay (median of 3 independent experiments for each drug). Left panels: heat maps showing the synergy ZIP-score between the 2 drugs, and the “most synergistic areas” (black squares) estimated using the SynergyFinder software v2.0. Right panel: mean of synergy ZIP-scores in most synergistic areas for HL-60 cells. Data are shown as mean +/- standard deviation (SD) of replicate samples (N=3 for each drug). Dotted line at 10% represents the threshold for significant synergy. (E) HL-60, U937 and THP-1 cells were treated with TAK-981 and AZA every day for 3 consecutive days. Viability was analyzed at day 4 and compared to that in mock-treated conditions (median of 3 independent experiments for each cell line). Left panels: heat maps for the corresponding synergy ZIP-score between the TAK-981 and AZA, and the “most synergistic areas” (black squares) estimated using the SynergyFinder software v2.0. Right panel: mean of synergy ZIP-scores in most synergistic areas for the 3 cell lines. Data are shown as mean +/- SD of replicate samples (N=3 for each cell line). Dotted line at 10% represents the threshold for significant synergy. (F, G) Patient (N=17) (F) or healthy donor (N=6) (G) BMMC were treated for 3 consecutive days (day 1, 2, 3) with TAK-981 (10 nM) and/ or AZA (100 nM) and kept in culture. After 8 days, cells were collected and the number of CD45+ cells was analyzed by flow cytometry in each condition and compared to the mock-treated condition. For each group, plain lines represent the median value, and dotted lines are the quartiles. Groups were compared using RM one-way ANOVA test.

bination had a much higher anti-leukemic effect than the monotherapies (Figure 2A, B; Online Supplementary Figure 4B, D). Finally, it was also more efficient than VEN+AZA combination at both limiting THP-1 cells proliferation in vivo and extending mice survival (Figure 2C, D; Online Supplementary Figure S3D). We then turned to PDX (Online Supplementary Table S2). Once engrafted patient cells became detectable in blood, mice were treated with one cycle of TAK-981 and/or AZA and the number of leukemic cells was analyzed in spleen and bone marrow. For the first patient tested (PDX #1), TAK-981 decreased tumor burden on its own in the spleen. TAK-981+AZA treatment was more efficient than TAK-981 alone at decreasing tumor burden in both spleen and bone marrow (Figure 2E). Similar results were obtained with two other PDX (PDX #2 and PDX #3) (Figure 2F; Online Supplementary Figure S4E). AZA was highly efficient on its own on one of them (PDX #3), thus limiting the benefit of its combination with TAK-981 in this case (Online Supplementary Figure S4E). Altogether, these data show in preclinical models that targeting SUMOylation with TAK-981 can exert an anti-leukemic effect in vivo, which is increased when combined with AZA. TAK-981 amplifies 5-azacytidine-induced transcriptional reprogramming and favors apoptosis In order to study the molecular mechanisms underlying the synergy between TAK-981 and AZA in AML, we performed RNA-seq experiments in U937 cells treated for 72 hours. TAK-981 showed limited effects on gene expression with 112 genes upregulated and three genes downregulated more than 2-fold (Figure 3A). AZA induced a much broader transcriptional reprogramming with 1,684 genes up- and 225 genes downregulated (Figure 3B). The highest impact

on transcription occurred with the TAK-981+AZA combo, with 2947 genes upregulated and 850 genes downregulated more than 2-fold (Figure 3C; Online Supplementary Table S3). Most genes up- or downregulated upon AZA (orange and purple dots, respectively) had higher fold changes upon TAK-981+AZA combination (Figure 3D). This suggests that inhibition of SUMOylation with TAK-981 amplifies AZA-induced modulation of gene expression. Accordingly, gene set enrichment analysis (GSEA) revealed that most gene signatures enriched in AZA-treated cells have higher normalized enrichment scores (NES) upon TAK-981+AZA treatment (Figure 3E). The most enriched pathways in the TAK-981+AZA versus mock- (Figure 3E), AZA- (Online Supplementary Figure S5A) and TAK-981- (Online Supplementary Figure S5B) treated cells, are linked to cell death as well as inflammation and immune system (see below). In order to avoid measuring transcriptional effects indirectly linked to the induction of cell death, we performed the RNA-seq analysis in U937 treated with doses of AZA and TAK-981 (10 nM each) suboptimal to induce apoptosis (Figure 4A). However, in line with the activation of transcriptional programs related to cell death, combinations of AZA and TAK-981 at higher doses led to a massive apoptosis both in U937 and THP-1 cells (Figure 4A). For both cell lines, TAK-981+AZA was more efficient than VEN+AZA at inducing apoptosis (Figure 4A). Finally, the most downregulated gene signatures in TAK-981+AZA treated cells are related to cell cycle progression, in particular MYC and E2F target genes (Figure 3E; Online Supplementary Figure S3A, B). We confirmed that c-MYC itself is downregulated upon TAK-981+AZA treatment in U937 cells (Online Supplementary Figure S5C). In the THP-1 and HL-60 cell line, we did not observe a significant modulation of c-MYC but TAK-981+AZA induced a strong up-regulation of CDKN1a, which encodes the p21(WAF1/CIP1) cell

Haematologica | 109 January 2024

102

ARTICLE - TAK-981 synergizes with 5-azacytidine in AML

L. Gabellier et al.

Figure 2. TAK-981 and 5-azacytidine combination has a higher anti-leukemic activity than monotherapies in vivo. (A, B) NSG mice were injected with THP-1 cells and treated with TAK-981 (15 mg/kg, intravenously [IV]), 5-azacytidine (AZA) (2 mg/kg, intraperitoneally [IP]) or the combination according to the schedule presented in Online Supplementary Figure S2B (N=5/group). (A) Quantification, as relative luminescence units, of tumor burden evolution monitored by luminescence intensity in mice injected with bioluminescent THP-1 cells. (B) Overall survival after treatment start of mice injected with bioluminescent THP-1 cells was estimated in each group and compared using Kaplan-Meier method and log-rank test. (C, D) NSG mice were injected with THP-1 cells and treated with TAK-981 (15 mg/kg, IV) and AZA (2 mg/kg, IP) or venetoclax (VEN) (50 mg/kg, oral gavage [OG]) according to the schedule presented in Online Supplementary Figure S2D (N=7/group). (C) Quantification (as relative luminescence units) of tumor burden evolution monitored by luminescence intensity in mice injected with bioluminescent THP-1 cells. (D) Overall survival after treatment start of mice injected with bioluminescent THP-1 cells was estimated in each group of treatment and compared using Kaplan-Meier method and log-rank test. (E, F) NSG mice were injected with primary cells from 2 different AML patients. After engraftment, mice were treated with AZA and/or TAK-981 and euthanized at day 9. The total number of human CD45+ cells (hCD45) was estimated by flow cytometry in bone marrow (PDX #1, N=7) and spleen (PDX #1, N=6; PDX #2, N=4-6), and compared to the mean number of cells collected in the mock-treated group of mice. For each group, plain lines represent the median value, and dotted lines are the quartiles. Groups were compared using ordinary one-way ANOVA test.

Haematologica | 109 January 2024

103

ARTICLE - TAK-981 synergizes with 5-azacytidine in AML

L. Gabellier et al.

Continued on following page.

Haematologica | 109 January 2024

104

ARTICLE - TAK-981 synergizes with 5-azacytidine in AML

L. Gabellier et al.

Figure 3. TAK-981 enhances 5-azacytidine-induced transcriptional reprogramming in U937 acute myeloid leukemia cell line. (A, B, C) Volcano plots showing differentially expressed genes (DEG) in U937 cell line treated each day for 3 days with 5-azacytidine (AZA) (10 nM) (A), TAK-981 (10 nM) (B) or the combination TAK-981+AZA (10 nM each) (C), analyzed at day 4 by RNA-sequencing and compared to mock-treated cells (N=3). Red dotted lines indicate the 2-fold change cutoff (absolute log2 fold change [abs log2FC]=1) and a P value of 0.05 (log10=1.3). The total numbers of up- and downregulated DEG are indicated. Orange and violet dots indicate genes whose expression is respectively up- or downregulated more than 2-fold by AZA-alone treatment. (D) Scatter plot displaying the DESeq2 fold change of DEG in TAK-981+AZA treated U937 cells as a function of their fold change in AZA-treated cells. The red line represents the linear regression of the fold changes for the comparison of AZA+TAK-981 versus AZA replicates. As a control, the grey line represents the linear regression from the comparison of AZA versus AZA replicates. P value was calculated using student t test between the two linear regressions. (E) Gene set enrichment analysis was performed using Hallmark datasets on the RNA-sequencing data obtained from U937 cells. All pathways significantly enriched in the AZA+TAK-981 compared to mock-treated cells are shown (absolute normalized enrichment score [abs NES]>1, P<0.05 and false discovery rate [FDR]<0.05). NES and FDR are also indicated for the AZA and TAK-981 versus mock-treated cells.

cycle inhibitor and is known to be repressed by c-MYC,31 suggesting a downregulation of the MYC pathway in these cell lines as well (Online Supplementary Figure 5D). TAK981, in particular when combined with AZA, altered cell cycle progression with decreased percentage of cells in G1 (Figure 4B). Altogether, this suggests that the amplification of AZA-induced transcriptional reprogramming by TAK-981 leads to decreased proliferation and increased apoptosis of the leukemic cells, providing an explanation for the synergy between the two drugs. TAK-981+5-azacytidine favors acute myeloid leukemia cell differentiation Restoration of differentiation participates to the anti-leukemic action of various drugs, including hypomethylating agents.32 Gene signatures related to myeloid differentiation were enriched in TAK-981+AZA compared to mock, AZA and TAK-981 treatments (Figure 5A; Online Supplementary Table S4). We confirmed by quantitative reverse transcription polymerase chain reaction (qRT-PCR) that TAK-981+AZA combo leads to a stronger increase in the expression of the myeloid marker CD14 compared to the single treatments in U937 cells (Figure 5B). TAK-981+AZA also induced CD14 expression at the surface of U937 xenografted in mice (Figure 5C). Similarly, CD14 expression was found induced in THP-1 cells, with again a maximal effect obtained for the TAK-981+AZA combination (Figure 5D). As a comparison, VEN+AZA combination also induced CD14 expression in THP1 cells at however lower levels (Online Supplementary Figure S6). Finally, the prodifferentiation effect of TAK-981+AZA treatment was confirmed on patient cells in vivo, both in the blood (Figure 5E) and bone marrow (Figure 5F) of PDX mice (PDX #1), with increased expression of CD14 and CD15 at the surface of the leukemic cells. Altogether, our data suggest that the anti-leukemic action of the combination of TAK-981 and AZA is associated to the reactivation of leukemic cells differentiation. TAK-981+5-azacytidine induces the secretion of type-1 interferon by acute myeloid leukemia cells As mentioned above, inhibition of SUMOylation increases AZA-induced expression of genes linked to inflammatory response and immunity (Figure 3E). In particular, this

concerns IFN-I response pathway (Figure 6A), with genes such as interferon regulatory factors (IRF) being maximally upregulated upon TAK-981+AZA in both U937 (Figure 6B) and THP-1 cells (Figures 6C, D). Accordingly, TAK-981+AZA induced the production of IFN-α by THP1-cells (Figure 6E). We then analyzed IFN-α production in vivo by intracellular labeling of IFN-I on AML patient cells recovered from the bone marrow of PDX mice (PDX #1). Only TAK-981+AZA induced an increase in IFN-α production by the AML cells (Figure 6F). Thus, in addition to a direct effect on AML cells differentiation, proliferation and viability, TAK-981+AZA combination enhances the secretion of IFN-I by AML cells, which may stimulate innate and/or adaptive anti-tumor immune response. TAK-981 induces the expression of natural killer cell ligands on acute myeloid leukemia cells and activates natural killer cell cytotoxicity NK cells play critical roles in cancer immune surveillance, including in AML.33 Within the gene signatures linked to immune response enriched in TAK-981+AZA versus mock-, AZA- or TAK-981-treated U937 cells, we identified several related to the activation of NK cells (Figure 7A; Online Supplementary Table S4). Among the genes of this signature, we focused on the adhesion molecule ICAM-1, which is required for target cells to bind to NK through its interaction with LFA-1,34 as well as on the MICA/B ligands of the NK-activating receptor NKG2D present on NK cells.35 TAK-981 increased the expression of ICAM-1 and MICA/B at the surface of THP-1 cells, which was further increased by addition of AZA (Figures 7B, C). In vivo, ICAM-1 expression at the surface of xenografted patient cells was also increased by both TAK-981 and TAK-981+AZA treatments (Figure 7D) whereas, only TAK-981+AZA led to an increased MICA/B expression (Figure 7E). In order to assess whether these treatments could enhance the activation of NK cells, we co-cultured PBMC purified from the blood of six different healthy donors as a source of NK cells together with THP1 cells, previously treated with TAK-981, AZA or the drugs combination. The expression of the activation marker CD69 was increased at the surface of NK cells when they were co-cultured with TAK-981- or TAK-981+AZA-treated THP-1

Haematologica | 109 January 2024

105

ARTICLE - TAK-981 synergizes with 5-azacytidine in AML

L. Gabellier et al.

Figure 4. TAK-981+5-azacytidine induces apoptosis and cell cycle defects in acute myeloid leukemia cells. (A) U937 or THP1 cells were treated every day for 3 consecutive days with TAK-981, 5-azacytidine (AZA) or venetoclax (VEN) at the indicated doses and stained on day 4 with Annexin-V/7AAD to quantify apoptotic and dead cells (N=3, mean +/- standard deviation [SD], conditions were compared to mock-treated condition using ordinary one-way ANOVA). (B) Cells treated as in (A) were stained with propidium iodide to analyze cell cycle distribution (N=3, mean +/- SD).

compared to mock- or AZA-treated cells (Figure 7F). Finally, we monitored the cytotoxicity of purified NK cells isolated from five healthy donors towards THP-1 cells using live cell imaging. NK cell cytotoxicity was higher on THP-1 cells that had been previously treated with TAK-981+/-AZA compared to mock- or AZA-treated cells (Figures 7G, H). Altogether, our data suggest that TAK-981+AZA favors the recognition and lysis of AML cells by NK cells, which could contribute to the anti-leukemic activity of this treatment.

Discussion Here, we report that the SUMOylation inhibitor TAK-981 has anti-leukemic activity in various AML preclinical models. Moreover, it synergizes with AZA, a DNA hypomethylating

agent widely used for AML treatment. TAK-981 and AZA combo induces a broad transcriptional reprogramming of AML cells underlying pleiotropic effects. These include increased apoptosis, alteration of the cell cycle, differentiation of the leukemic cells, induction of IFN-I secretion and enhanced expression of NK cell ligands at the surface of AML cells, which stimulates NK cytotoxicity towards them (Figure 8). Accumulating evidence suggest that alteration in SUMOylation can both contribute to tumorigenesis and affect response to therapies in various cancers.36 This is notably the case for AML.37 Different inhibitors of SUMOylation such as anacardic acid,17 2D-0818,38 and McM02504439 showed in vitro toxicity for leukemic cell. However, their low activity (μM range) and poor pharmacological properties prevented further preclinical studies. The discovery of TAK-981 now

Haematologica | 109 January 2024

106

ARTICLE - TAK-981 synergizes with 5-azacytidine in AML

L. Gabellier et al.

Figure 5. TAK-981+5-azacytidine induces differentiation of acute myeloid leukemia cells. (A) Gene set enrichment analysis enrichment plot for the gene signature “GOBP Myeloid Cell Differentiation” in TAK-981+5-azacytidine [AZA]-treated U937 cells compared to mock- (upper panel) or AZA- (lower panel) treated cells. (B) mRNA expression of CD14 was analyzed by quantitative reverse transcription polymerase chain reaction (qRT-PCR) in U937 cells treated for 72 hours with 10 nM AZA, 10 nM TAK-981 or the drug combination. Results were normalized to GAPDH mRNA levels and expressed as ratio to mock-treated cells (N=3, mean +/- standard deviation [SD], conditions were compared using RM one-way ANOVA test). (C) NSG mice were injected with U937 cells (N=4 or 5/group) and, after engraftment, treated according to treatment schedule presented in Figure 2A. Bone marrow were collected at day 9 after treatment start and the level of CD14 was assessed by flow cytometry at cell surface of hCD45+ cells. Data were normalized to the mean of CD14 expression (mean fluorescence intensity [MFI]) in mock-treated group of mice. For each group, plain lines represent the median value and dotted lines are the quartiles. Groups were compared using Ordinary one-way ANOVA test. (D) Expression of CD14 was measured by flow cytometry on THP-1 treated with 10 nM AZA, 10 nM TAK-981 or the drug combination for 72 hours. MFI were normalized to that of mock-treated cells (N=6, mean +/- SD, RM one-way ANOVA test). (E, F) NSG mice were injected with primary patient cells (PDX #1), and treated according to treatment schedule presented in Figure 2A. Peripheral blood samples (N=3-6/group) were collected at day 30 after treatment start (E) or, in an independent experiment, spleens (N=6 or 7/group) were collected at day 9 after treatment start (F). The level of CD14 and CD15 protein expression at cell surface of hCD45+ cells was evaluated by flow cytometry. Data were normalized on the mean of CD14 and CD15 expression (MFI) in cells from the mock-treated group. For each group, plain lines represent the median values, and dotted lines are the quartiles. Groups were compared using Kruskall-Wallis test, due to lack of normality of data.

Haematologica | 109 January 2024

107

ARTICLE - TAK-981 synergizes with 5-azacytidine in AML

L. Gabellier et al.

Figure 6. TAK-981 and 5-azacytidine combination activates type-1 interferon secretion by acute myeloid leukemia cells. (A) Gene set enrichment analysis plot for genes involved in interferon α (IFN-α) response in TAK-981+AZA treated U937 cells compared to mock- (upper panel) and 5-azacytidine (AZA)- (lower panel) treated cells. (B) Heatmap representing the RNA-sequencing results for the expression of interferon regulatory factor (IRF) in U937 cells. (C, D) mRNA expression of IRF7 (C) and IRF9 (D) was analyzed by quantitative reverse transcription polymerase chain reaction (qRT-PCR) in THP-1 cells treated for 72 hours with 10 nM TAK-981 and the indicated concentrations of AZA. Results were normalized to GAPDH mRNA levels and expressed as ratio to mock-treated cells (N=3-5, mean +/- standard deviation [SD], all conditions were compared using RM one-way ANOVA test, only those with significant P values are shown). (E) THP-1 cells were treated for 3 consecutive days with TAK-981 (10 nM) and/or AZA (10 nM or 100 nM). After 8 days, cells were collected and the production of intracellular IFN-α was analyzed by flow cytometry. Background was subtracted and data normalized to the mock-treated condition (N=3, mean +/- SD, all conditions were compared using RM one-way ANOVA test, only those with significant P values are shown). (F) NSG mice were injected with primary patient cells (PDX #1) and treated according to treatment schedule presented in Figure 2A. Spleens (N=6/group) were collected at day 9 after treatment start and the level of intracellular IFN-α in hCD45+ cells was assessed by flow cytometry. Data were normalized to the mean of IFN-α expression (mean fluorescence intensity [MFI]) in mock-treated group of mice. For each group, plain lines represent the median value and dotted lines are the quartiles. Groups were compared using ordinary one-way ANOVA test.

Haematologica | 109 January 2024

108

ARTICLE - TAK-981 synergizes with 5-azacytidine in AML

L. Gabellier et al.

Continued on following page.

Haematologica | 109 January 2024

109

ARTICLE - TAK-981 synergizes with 5-azacytidine in AML

L. Gabellier et al.

Figure 7. TAK-981 treatment of acute myeloid leukemia cells leads to natural killer cell activation and increased natural killer cell cytotoxicity. (A) Gene set enrichment analysis enrichment plot for genes involved in natural killer (NK) cell-mediated cytotoxicity in TAK-981+5-azacytidine (AZA)-treated U937 cells compared to mock- (upper panel) or AZA- (lower panel) treated cells. (B, C) Expression of ICAM-1 (B) and MICA/B (C) was measured by flow cytometry on THP-1 treated with 10 nM TAK-981 and AZA at the indicated concentrations for 72 hours. Background was subtracted and results were normalized to mock-treated condition (N=10 for ICAM-1, N=5 for MICA/B, mean +/- standard deviation [SD], RM one-way ANOVA test). (D, E) NSG mice (N=7/group) were injected with primary patient cells (PDX #1), and treated according to treatment schedule presented in Figure 2A. Bone marrows were collected at day 9 after treatment start and ICAM-1 and MICA/B expression on the membrane of human CD45+ cells was assessed by flow cytometry. Data were normalized to the mean fluorescence intensity (MFI) of mock-treated mice group. Plain lines represent the median values, and dotted lines are the quartiles. Groups were compared using Kruskall-Wallis test. (F) THP1 cells were treated each day for 3 consecutive days with 10 nM AZA, 10 nM TAK-981 or the drug combination, and co-cultured at day 8 with healthy donor peripheral blood mononuclear cells (PBMC) at a 1:10 acute myeloid leukemia (AML):PBMC ratio. After 24 hours of co-culture, expression of activation marker CD69 was assessed by flow cytometry on NK cells (CD3-/CD56+ cells). Data were normalized to the MFI of CD69 expression on NK cells co-cultured with untreated THP-1, (mean +/- SD, N=6, RM oneway ANOVA test). (G, H) Real-time immune cell killing assay. To evaluate NK cells cytotoxicity against AML cells treated with TAK981-/+AZA, co-culture experiments were performed during 15 hours using an Incucyte device between, on one hand, THP-1-LucZsGreen cells previously treated for 72 hours with 10 nM AZA, 10 nM TAK-981 or the drug combination, and on the other hand, NK cells, purified from healthy donor PBMC. Cells were used at a 1:1 AML:NK ratio. (G) Relative green fluorescence intensities for THP-1-Luc-ZsGreen cells mock-treated without (grey curve) or with NK cells (light blue curve) and treated with TAK-981+AZA without (red curve) or with NK cells (dark blue curve) (N=5). Cytotoxicity of NK cells was calculated by comparing grey- (mock-treated THP-1-/+NK cells) and red areas (AZA+TAK-981 treated THP-1-/+NK cells). (H) NK cells cytotoxicity was determined by calculating areas between the curves for each treatment condition (10 nM AZA, 10 nM TAK-981 or the drug combination) with or without NK. Data were normalized to mock-treated THP-1 cells (N=5, mean +/- SD, RM one-way ANOVA test).

allows to envision SUMOylation inhibition in cancer patients. Five phase I/II clinical trials are ongoing in solid tumors (clinicaltrials gov. Identifier: NCT03648372, NCT04381650), multiple myelomas (clinicaltrials gov. Identifier: NCT04776018) and lymphomas (clinicaltrials gov. Identifier: NCT03648372, NCT04074330). Our data obtained in preclinical models of AML provide a rationale for evaluating TAK-981 in AML treatment. Importantly, we observed minimal toxicity of TAK-981 on normal blood and BMMC (Figure 1C, G) or when administered to mice (Figure 2). Although TAK-981 can kill leukemic cells in vitro and in vivo, its anti-tumor activity is relatively limited when used as monotherapy. However, its combination with AZA has a largely superior anti-leukemic activity. Combination therapies are increasingly considered to achieve stronger responses to cancer treatments and to limit relapses, including in AML.40 This notably concerns AZA, whose combination with various drugs, in particular VEN, has improved clinical responses.3 Nevertheless, many patients are refractory to VEN+AZA regimen or relapse after treatment.9,10 This is for example the case of patients suffering from monocytic AML (FAB M4 and M5).41 It is, therefore, interesting that TAK-981+AZA shows more efficiency than VEN+AZA for all cell lines tested. In particular, TAK-981+AZA was more efficient than VEN+AZA at inducing apoptosis, cell cycle defects and differentiation of the monocytic cell lines U937 and THP-1. In addition, in our in vitro experiments on patient cells (Figure 1F; Online Supplementary Figure S2B), AML cells from the more differentiated M4 and M5 subgroups were as sensitive (even slightly more sensitive) to TAK-981+AZA treatment than less differentiated AML cells from the M1 and M2 subgroups. Patients who are refractory to VEN+AZA regimen might therefore be sensitive to TAK-981+AZA treatment. The synergy between AZA and TAK-981 likely resides in the

ability of TAK-981 to enhance the action of AZA on transcription. A large number of transcription factors, co-activators and co-repressor complexes, the basal transcription machinery and histones are SUMOylated.15,16,42 In general, SUMOylation of protein complexes rather than individual proteins within these complexes mediates the biological effects of SUMOylation, which include the stabilization of these complexes43 or the recruitment of SUMO interacting motif (SIM)-containing proteins.44,45 For example, SUMOylation of chromatin bound proteins can favor the recruitment of co-repressor complexes such as those containing histone deacetylases (HDAC)46-48 or the histone methyl transferase SETDB149 via SUMO-SIM interactions. We show that inhibition of SUMOylation per se has limited effects on gene expression. Although surprising considering the high number of SUMOylated proteins present on gene regulatory regions, SUMOylation is thus dispensable for gene expression in basal conditions. However, inhibition of SUMOylation largely increases the expression of most AZA-induced genes. Hence, it is likely that the effect of TAK-981 on AZA-induced transcriptional reprogramming is due to the global deSUMOylation of proteins bound to gene regulatory regions rather than the consequence of the deSUMOylation of specific proteins present in these regions. This deSUMOylation would create a permissive environment for transcription, likely by affecting the recruitment and activity of transcription regulating- and/or chromatin remodeling-complexes. This would, therefore, amplify the transcriptional regulation of genes, whose cis-regulatory regions have been hypo-methylated by AZA. Of note, we have recently shown that inhibition of SUMOylation limits the transcriptional reprogramming induced by DNR in AML cells after few hours of treatment.50 This suggests that inhibitors of SUMOylation could have different global impacts on gene expression depending on the duration of

Haematologica | 109 January 2024

110

ARTICLE - TAK-981 synergizes with 5-azacytidine in AML

L. Gabellier et al.

Figure 8. Model for the anti-leukemic activity of TAK-981+5-azacytidine in acute myeloid leukemia. Combined inhibition of SUMOylation with TAK-981 and DNA methylation with 5-azacytidine (AZA) induces a transcriptional reprogramming in acute myeloid leukemia (AML) cells. This includes the activation of genes involved in the induction of apoptosis and differentiation and a repression of genes linked to cell cycle progression. In addition, TAK-981+AZA induces interferon 1 (IFN-1) secretion as well as the expression of natural killer (NK) cell ligands at the surface of acute myeloid leukemia (AML) cells. This activates NK and increases their cytotoxicity towards AML cells.

the treatment (hours vs. days) and/or the drugs they are associated with. We also do not exclude that other mechanisms than transcriptional reprogramming might be at play to explain the synergy between AZA and TAK-981. Poly-SUMOylation of DNMT1 was shown to be triggered by decitabine, another hypomethylating agent, which induces crosslink between DNMT1 and DNA. This leads to its RNF4-mediated ubiquitylation and promotes the resolution of the DNA-protein crosslinks (DPC).4,51 Accordingly, inhibition of SUMOylation increases decitabine-induced DPC, ultimately resulting in DNA damage and cell death in models of lymphoma.52 However, these studies were performed with high doses (110 μM) of decitabine, which is much more prone to induce DPC and DNA damage than AZA at the same doses.53 In the condition used in our transcriptomic study (10 nM of AZA for 72 hours), it is unlikely that AZA induces massive DPC and subsequent DNA damage. Moreover, the transcriptional reprogramming we characterized in AML was observed at sublethal doses of the drugs, further supporting the idea that it did not involve DPC-induced DNA damages. Our data point to pleiotropic effects of the TAK-981+AZA combo to eliminate AML cells through increased apoptosis, decreased proliferation and induction of myeloid differentiation. The contribution of these different pathways to the anti-leukemic activity of the TAK-981+AZA combo differs depending on the concentrations of the drugs. At the high-

est doses, apoptosis is likely central to the anti-leukemic action of TAK-981+AZA. At lower doses, differentiation might be the main contributor to the decreased cell proliferation upon TAK-981+AZA treatment. Indeed, induction of differentiation is critical for the action of various AML therapies, including all trans retinoic acid (ATRA) and arsenic trioxide in acute promyelocytic leukemia (APL) subtype of AML54 as well as IDH1 and FLT3 inhibitors.55 In addition to the direct effect on AML cells proliferation and survival, TAK-981+AZA combination induces many genes linked to inflammation and immunity. This is notably the case for the IFN-I pathway, which has been reported to be activated by both AZA in cancer cells56,57 and by TAK981 in immune cells.20,21 IFN-I induction can then induce an anti-tumor immune response in these models.20,21,23 Although TAK-981 on its own weakly induced IFN-I pathway, its combination with AZA largely increased IFN-I secretion by AML cells themselves. Although systemic IFN-I based therapies have proven disappointing in terms of clinical efficacy with important toxicities,58 controlled and localized secretion of IFN-I by AML cells might be sufficient to elicit an anti-leukemic immune response devoid of toxicity. Finally, our work suggests an important role for NK cells in the elimination of TAK-981-treated AML cells. NK cells can eliminate tumor cells directly by inducing their lysis or indirectly through the secretion of cytokines such as

Haematologica | 109 January 2024

111

ARTICLE - TAK-981 synergizes with 5-azacytidine in AML

L. Gabellier et al.

IFN-γ or TNF-α. AML patients survival is highly correlated with the number and activity of NK cells.35 However, NK cells are often poorly functional in these patients. In particular, NK cells from AML patients are often defective in the expression of activating receptors (DNAM1, NKp30, NKp46) or overexpress inhibitory KIR receptors. Finally, AML cells develop immunosuppressive strategies to escape NK cell-mediated cell lysis, such as downregulation or shedding from their cell surface of ligands of the NK cell-activating receptor NKG2D (MICA/MICB and ULBP1-6).35 It was, therefore, interesting to observe that TAK-981+AZA treatment leads to upregulation of MICA/MICB at the surface of AML cells. In addition, we also detected an upregulation of adhesion molecule ICAM-1, whose binding to the LFA-1 receptor on NK cells is required for efficient lysis of AML cells.34 The combined secretion of cytokines and expression of NK cell-activating markers by AML cells could explain the increased activation and enhanced cytotoxicity of NK cells towards TAK-981+AZA-treated AML cells. In conclusion, our work suggests that targeting SUMOylation with TAK-981 may be a promising strategy to enhance the clinical efficacy of AZA in AML patients. This combination treatment is expected to exert cell-autonomous effects on AML cells by inducing their differentiation and apoptosis and cell-extrinsic effects by triggering an anti-leukemic immune response.

formed in vivo experiments. MC prepared samples for RNA-sequencing and performed qPCR. DA, MA and GiB analyzed viability and differentiation of AML cell lines. RH participated in PBMC and NK purifications. MA, RH and DA performed validation qPCR. AD prepared RNA-sequencing libraries and sequenced them. CRZ performed microbead-based assay. GC provided patient samples. DT conceived and analyzed RNA-sequencing experiments and performed GSEA analysis. MP provided funding for the study. GB supervised the study and provided funding. LG, MDT, DT, MP and GB wrote the manuscript.

Disclosures TAK-981 was obtained from Takeda Development Center Americas, Inc. GC is a consultant for Roche and BMS-Celgene; is a member of the advisory boards of MabQi, Ownards Therapeutics, MedXcell and receives honoraria from Abbvie, Sanofi, Gilead, Janssen, Roche, BMS-Celgene, Takeda Pharmaceuticals. The remaining authors have no conflicts of interest to disclose. Contributions LG designed and performed the experiments on cell lines, patient cells and mouse models. MDT conceived and per-

Acknowledgments We thank the members of the “Ubiquitin Family in Hematological Malignancies” group, Takeda Development Center Americas (Lexington, MA), for providing TAK-981 and Dr Allison Berger for critically reading the manuscript. Funding Funding was provided by the CNRS, the Ligue Nationale contre le Cancer (to DA, MC, MP), the Fondation pour la Recherche Médicale (contract FDM201906008566, LG), INCa_16072, the Fédération Leucémie Espoir, the Association Laurette Fugain, the Fondation ARC pour la recherche sur le cancer and the Agence Nationale pour la Recherche (“Investissements d’avenir” program; ANR-16-IDEX-0006; project SUMOLAM). The HEMODIAG_2020 collection was funded by the Montpellier University Hospital, the Montpellier SIRIC and the Languedoc Roussillon Region. MGX acknowledges financial support from the France Génomique National infrastructure, funded as part of “Investissements d’Avenir” program managed by the Agence Nationale pour la Recherche (contract ANR 10 INBS 09). Data-sharing statement The RNA-sequencing data are available on Gene Expression Omnibus with accession number GSE212330. All other data are presented in the main text or the Online Supplementary Appendix.

References 1. Saultz JN, Garzon R. Acute myeloid leukemia: a concise review. J Clin Med. 2016;5(3):33. 2. Ahmadmehrabi K, Haque AR, Aleem A, Griffiths EA, Roloff GW. Targeted therapies for the evolving molecular landscape of acute myeloid leukemia. Cancers (Basel). 2021;13(18):4646. 3. Ciotti G, Marconi G, Martinelli G. Hypomethylating agent-based Ccmbination therapies to treat post-hematopoietic stem cell transplant relapse of acute myeloid leukemia. Front Oncol. 2022;11:810387. 4. Liu JCY, Kühbacher U, Larsen NB, et al. Mechanism and function of DNA replication-independent DNA-protein crosslink repair via the SUMO-RNF4 pathway. EMBO J. 2021;40(18):e107413. 5. Lund K, Cole JJ, VanderKraats ND, et al. DNMT inhibitors

reverse a specific signature of aberrant promoter DNA methylation and associated gene silencing in AML. Genome Biol. 2014;15(8):406. 6. Dombret H, Seymour JF, Butrym A, et al. International phase 3 study of azacitidine vs conventional care regimens in older patients with newly diagnosed AML with >30% blasts. Blood. 2015;126(3):291-299. 7. DiNardo CD, Pratz K, Pullarkat V, et al. Venetoclax combined with decitabine or azacitidine in treatment-naive, elderly patients with acute myeloid leukemia. Blood. 2019;133(1):7-17. 8. DiNardo CD, Jonas BA, Pullarkat V, et al. Azacitidine and venetoclax in previously untreated acute myeloid leukemia. N Engl J Med. 2020;383(7):617-629.

Haematologica | 109 January 2024

112

ARTICLE - TAK-981 synergizes with 5-azacytidine in AML

L. Gabellier et al.

9. Garciaz S, Saillard C, Hicheri Y, Hospital M-A, Vey N. Venetoclax in acute myeloid leukemia: molecular basis, evidences for preclinical and clinical efficacy and strategies to target resistance. Cancers. 2021;13(22):5608. 10. Ong F, Kim K, Konopleva MY. Venetoclax resistance: mechanistic insights and future strategies. Cancer Drug Resist. 2022;5(2):380-400. 11. Vertegaal ACO. Signalling mechanisms and cellular functions of SUMO. Nat Rev Mol Cell Biol. 2022;23(11):715-731. 12. Hendriks IA, Lyon D, Young C, Jensen LJ, Vertegaal ACO, Nielsen ML. Site-specific mapping of the human SUMO proteome reveals co-modification with phosphorylation. Nat Struct Mol Biol. 2017;24(3):325-336. 13. Hendriks IA, Lyon D, Su D, et al. Site-specific characterization of endogenous SUMOylation across species and organs. Nat Commun. 2018;9(1):2456. 14. Pichler A, Fatouros C, Lee H, Eisenhardt N. SUMO conjugation - a mechanistic view. Biomol Concepts. 2017;8(1):13-36. 15. Boulanger M, Chakraborty M, Tempé D, Piechaczyk M, Bossis G. SUMO and transcriptional regulation: the lessons of large-scale proteomic, modifomic and genomic studies. Molecules. 2021;26(4):828. 16. Rosonina E. A conserved role for transcription factor sumoylation in binding-site selection. Curr Genet. 2019;65(6):1307-1312. 17. Bossis G, Sarry J-E, Kifagi C, et al. The ROS/SUMO axis contributes to the response of acute myeloid leukemia cells to chemotherapeutic drugs. Cell Rep. 2014;7(6):1815-1823. 18. Baik H, Boulanger M, Hosseini M, et al. Targeting the SUMO pathway primes all-trans retinoic acid–induced differentiation of nonpromyelocytic acute myeloid leukemias. Cancer Res. 2018;78(10):2601-2613. 19. Langston SP, Grossman S, England D, et al. Discovery of TAK981, a first-in-class inhibitor of SUMO-activating enzyme for the treatment of cancer. J Med Chem. 2021;64(5):2501-2520. 20. Lightcap ES, Yu P, Grossman S, et al. A small-molecule SUMOylation inhibitor activates antitumor immune responses and potentiates immune therapies in preclinical models. Sci Transl Med. 2021;13(611):eaba7791. 21. Kumar S, Schoonderwoerd MJA, Kroonen JS, et al. Targeting pancreatic cancer by TAK-981: a SUMOylation inhibitor that activates the immune system and blocks cancer cell cycle progression in a preclinical model. Gut. 2022;71(11):2266-2283. 22. Demel UM, Böger M, Yousefian S, et al. Activated SUMOylation restricts MHC class I antigen presentation to confer immune evasion in cancer. J Clin Invest. 2022;132(9):e152383. 23. Nakamura A, Grossman S, Song K, et al. SUMOylation inhibitor subasumstat potentiates rituximab activity by IFN1-dependent macrophage and NK cell stimulation. Blood. 2022;139(18):2770-2781. 24. Demel UM, Wirth M, Yousefian S, et al. Small molecule SUMO inhibition for biomarker-informed B-cell lymphoma therapy. Haematologica. 2023;108(2):555-567. 25. Du L, Liu W, Pichiorri F, Rosen ST. SUMOylation inhibition enhances multiple myeloma sensitivity to lenalidomide. Cancer Gene Ther. 2022;30(4):567-574. 26. Du L, Liu W, Aldana-Masangkay G, et al. SUMOylation inhibition enhances dexamethasone sensitivity in multiple myeloma. J Exp Clin Cancer Res. 2022;41(1):1-16. 27. Kim HS, Kim B-R, Dao TTP, et al. TAK-981, a SUMOylation inhibitor, suppresses AML growth immune-independently. Blood Adv. 2023;7(13):3155-3168. 28. Paolillo R, Boulanger M, Gâtel P, et al. The NADPH oxidase NOX2

is a marker of adverse prognosis involved in chemoresistance of acute myeloid leukemias. Haematologica. 2022;107(11):2562-2575. 29. Gâtel P, Brockly F, Reynes C, et al. Ubiquitin and SUMO conjugation as biomarkers of acute myeloid leukemias response to chemotherapies. Life Sci Alliance. 2020;3(6):e201900577. 30. Recasens-Zorzo C, Gâtel P, Brockly F, Bossis G. A microbeadbased flow cytometry assay to assess the activity of ubiquitin and ubiquitin-like conjugating enzymes. Methods Mol Biol. 2023;2602:65-79. 31. Gartel AL, Ye X, Goufman E, et al. Myc represses the p21(WAF1/ CIP1) promoter and interacts with Sp1/Sp3. Proc Natl Acad Sci U S A. 2001;98(8):4510-4515. 32. Stubbins RJ, Karsan A. Differentiation therapy for myeloid malignancies: beyond cytotoxicity. Blood Cancer J. 2021;11(12):1-9. 33. Rahmani S, Yazdanpanah N, Rezaei N. Natural killer cells and acute myeloid leukemia: promises and challenges. Cancer Immunol Immunother. 2022;71(12):2849-2867. 34. Allende-Vega N, Marco Brualla J, Falvo P, et al. Metformin sensitizes leukemic cells to cytotoxic lymphocytes by increasing expression of intercellular adhesion molecule-1 (ICAM-1). Sci Rep. 2022;12(1):1341. 35. Carlsten M, Järås M. Natural killer cells in myeloid malignancies: immune surveillance, NK cell dysfunction, and pharmacological opportunities to bolster the endogenous NK cells. Front Immunol. 2019;10:2357. 36. Seeler J-S, Dejean A. SUMO and the robustness of cancer. Nat Rev Cancer. 2017;17(3):184-197. 37. Boulanger M, Paolillo R, Piechaczyk M, Bossis G. The SUMO pathway in hematomalignancies and their response to therapies. Int J Mol Sci. 2019;20(16):3895. 38. Zhou P, Chen X, Li M, et al. 2-D08 as a SUMOylation inhibitor induced ROS accumulation mediates apoptosis of acute myeloid leukemia cells possibly through the deSUMOylation of NOX2. Biochem Biophys Res Commun. 2019;513(4):1063-1069. 39. Benoit YD, Mitchell RR, Wang W, et al. Targeting SUMOylation dependency in human cancer stem cells through a unique SAE2 motif revealed by chemical genomics. Cell Chem Biol. 2021;28(10):1394-1406. 40. Ishii H, Yano S. New therapeutic strategies for adult acute myeloid leukemia. Cancers (Basel). 2022;14(11):2806. 41. Pei S, Pollyea DA, Gustafson A, et al. Monocytic subclones confer resistance to venetoclax-based therapy in patients with acute myeloid leukemia. Cancer Discov. 2020;10(4):536-551. 42. Ryu H-Y, Hochstrasser M. Histone sumoylation and chromatin dynamics. Nucleic Acids Res. 2021;49(11):6043-6052. 43. Seifert A, Schofield P, Barton GJ, Hay RT. Proteotoxic stress reprograms the chromatin landscape of SUMO modification. Sci Signal. 2015;8(384):rs7. 44. González-Prieto R, Eifler-Olivi K, Claessens LA, et al. Global non-covalent SUMO interaction networks reveal SUMOdependent stabilization of the non-homologous end joining complex. Cell Rep. 2021;34(4):108691. 45. Lascorz J, Codina-Fabra J, Reverter D, Torres-Rosell J. SUMOSIM interactions: from structure to biological functions. Semin Cell Dev Biol. 2022;132:193-202. 46. Ouyang J, Shi Y, Valin A, Xuan Y, Gill G. Direct binding of CoREST1 to SUMO-2/3 contributes to gene-specific repression by the LSD1/CoREST1/HDAC complex. Mol Cell. 2009;34(2):145-154. 47. Hua G, Ganti KP, Chambon P. Glucocorticoid-induced tethered transrepression requires SUMOylation of GR and formation of a

Haematologica | 109 January 2024

113

ARTICLE - TAK-981 synergizes with 5-azacytidine in AML

L. Gabellier et al.

SUMO-SMRT/NCoR1-HDAC3 repressing complex. Proc Natl Acad Sci U S A. 2016;113(5):E635-E643. 48. Pascual G, Fong AL, Ogawa S, et al. A SUMOylation-dependent pathway mediates transrepression of inflammatory response genes by PPAR-γ. Nature. 2005;437(7059):759-763. 49. Ivanov AV, Peng H, Yurchenko V, et al. PHD domain-mediated E3 ligase activity directs intramolecular sumoylation of an adjacent bromodomain required for gene silencing. Mol Cell. 2007;28(5):823-837. 50. Boulanger M, Aqrouq M, Tempé D, et al. DeSUMOylation of chromatin-bound proteins limits the rapid transcriptional reprogramming induced by daunorubicin in acute myeloid leukemias. Nucleic Acids Res. 2023;51(16):8413-8433. 51. Borgermann N, Ackermann L, Schwertman P, et al. SUMOylation promotes protective responses to DNA-protein crosslinks. EMBO J. 2019;38(8):e101496. 52. Kroonen JS, de Graaf IJ, Kumar S, et al. Inhibition of SUMOylation enhances DNA hypomethylating drug efficacy to reduce outgrowth of hematopoietic malignancies. Leukemia.

2023;37(4):864-876. 53. Hollenbach PW, Nguyen AN, Brady H, et al. A comparison of azacitidine and decitabine activities in acute myeloid leukemia cell lines. PLoS One. 2010;5(2):e9001. 54. Yilmaz M, Kantarjian H, Ravandi F. Acute promyelocytic leukemia current treatment algorithms. Blood Cancer J. 2021;11(6):1-9. 55. Madan V, Koeffler HP. Differentiation therapy of myeloid leukemia: four decades of development. Haematologica. 2021;106(1):26-38. 56. Roulois D, Loo Yau H, Singhania R, et al. DNA-demethylating agents target colorectal cancer cells by inducing viral mimicry by endogenous transcripts. Cell. 2015;162(5):961-973. 57. Chiappinelli KB, Strissel PL, Desrichard A, et al. Inhibiting DNA Methylation causes an interferon response in cancer via dsRNA including endogenous retroviruses. Cell. 2015;162(5):974-986. 58. Aricò E, Castiello L, Capone I, Gabriele L, Belardelli F. Type I interferons and cancer: an evolving story demanding novel clinical applications. Cancers (Basel). 2019;11(12):1943.

Haematologica | 109 January 2024

114

ARTICLE - Acute Myeloid Leukemia

Outcomes after intensive chemotherapy for secondary and myeloid-related changes acute myeloid leukemia patients aged 60 to 75 years old: a retrospective analysis from the PETHEMA registry David Martínez-Cuadrón,1,2,3 Juan E. Megías-Vericat,4 Cristina Gil,5 Teresa Bernal,6 Mar Tormo,7 Pilar Martínez-Sánchez,8 Carlos Rodríguez-Medina,9 Josefina Serrano,10 Pilar Herrera,11 José A. Pérez Simón,12 María J. Sayas,13 Juan Bergua,14 Esperanza Lavilla-Rubira,15 María Luz Amigo,16 Celina Benavente,17 Jose L. López Lorenzo,18 Manuel M. Pérez-Encinas,19 María B. Vidriales,20 Mercedes Colorado,21 Beatriz de Rueda,22 Raimundo García-Boyero,23 Sandra Marini,24 Julio García-Suárez,25 María López-Pavía,26 Maria I. Gómez-Roncero,27 Víctor Noriega,28 Aurelio López,29 Jorge Labrador,30 Ana Cabello,31 Claudia Sossa,32 Lorenzo Algarra,33 Mariana Stevenazzi,34 Antonio Solana-Altabella,4 Blanca Boluda1,2 and Pau Montesinos1,3

Correspondence: D. Martínez-Cuadrón martinez_davcua@gva.es Received: Accepted: Early view:

January 16, 2023. May 8, 2023. May 18, 2023.

Hospital Universitari i Politècnic La Fe, Valencia, Spain; 2Instituto de Investigación Sanitaria La Fe (IISLAFE), Valencia, Spain; 3CIBERONC, Instituto Carlos III, Madrid, Spain; 4 Servicio de Farmacia, Área del Medicamento, Hospital Universitari i Politècnic La Fe, Valencia, Spain; 5Hospital General Universitario de Alicante, Alicante, Spain; 6Hospital Universitario Central de Asturias, Instituto Universitario de Oncología del Principado de Asturias (IUOPA), Instituto de Investigación Sanitaria del Principado de Asturias (ISPA), Asturias, Spain; 7Hospital Clínico Universitario, INCLIVA Biomedical Research Institute, Valencia, Spain; 8Hospital Universitario 12 de Octubre, Madrid, Spain; 9Hospital Universitario de Gran Canaria Doctor Negrín, Las Palmas, Spain; 10Hospital Universitario Reina Sofía-IMIBIC, Córdoba, Spain; 11Hospital Universitario Ramón y Cajal, Madrid, Spain; 12 Hospital Virgen del Rocío, Instituto de Biomedicina de Sevilla (IBIS/CISC), Universidad de Sevilla, Sevilla, Spain; 13Hospital Universitario Doctor Peset, Valencia, Spain; 14Hospital San Pedro Alcántara, Cáceres, Spain; 15Hospital Universitario Lucus Augusti, Lugo, Spain; 16 Hospital General Universitario Morales Meseguer, Murcia, Spain; 17Hospital Clínico San Carlos, Madrid, Spain; 18Hospital Universitario Fundación Jiménez Díaz, Madrid, Spain; 19 Hospital Clínico Universitario de Santiago, Santiago de Compostela, Spain; 20Hospital Clínico Universitario de Salamanca, Salamanca, Spain; 21Hospital Universitario Marqués de Valdecilla, Santander, Spain; 22Hospital Universitario Miguel Servet, Zaragoza, Spain; 23 Hospital General Universitari de Castelló, Castellón, Spain; 24Centro Hospitalar e Universitário de Coimbra, EPE, Coimbra, Portugal; 25Hospital Universitario Príncipe de Asturias, Madrid, Spain; 26Hospital General Universitario de Valencia, Valencia, Spain; 27 Hospital Universitario de Toledo, Toledo, Spain; 28Complejo Hospitalario, Universitario A Coruña, A Coruña, Spain; 29Hospital Arnau de Vilanova, Valencia, Spain; 30Hospital Universitario de Burgos, Burgos, Spain; 31Hospital Universitario Nuestra Señora de Candelaria, Tenerife, Spain; 32FOSCAL, Facultad de Ciencias de la Salud, Universidad Autónoma de Bucaramanga, Bucaramanga, Colombia; 33Hospital General, Universitario de Albacete, Albacete, Spain and 34Hospital de Clínicas Dr. Manuel Quintela, Montevideo, Uruguay 1

Abstract Treatment options for patients with secondary acute myeloid leukemia (sAML) and AML with myeloid-related changes (AMLMRC) aged 60 to 75 years are scarce and unsuitable. A pivotal trial showed that CPX-351 improved complete remission with/without incomplete recovery (CR/CRi) and overall survival (OS) as compared with standard "3+7" regimens. We retrospectively analyze outcomes of 765 patients with sAML and AML-MRC aged 60 to 75 years treated with intensive chemotherapy, reported to the PETHEMA registry before CPX-351 became available. The CR/CRi rate was 48%, median OS was 7.6 months (95% confidence interval [CI]: 6.7-8.5) and event-free survival (EFS) 2.7 months (95% CI: 2-3.3), without differences between intensive chemotherapy regimens and AML type. Multivariate analyses identified age ≥70 years, Eastern Cooperative Oncology Group performance status ≥1 as independent adverse prognostic factors for CR/CRi and OS, while Haematologica | 109 January 2024

115

ARTICLE - Intensive chemotherapy in older sAML and AML-MRC

D. Martínez-Cuadrón et al.

favorable/intermediate cytogenetic risk and NPM1 were favorable prognostic factors. Patients receiving allogeneic stem cell transplant (HSCT), autologous HSCT, and those who completed more consolidation cycles showed improved OS. This large study suggests that classical intensive chemotherapy could lead to similar CR/CRi rates with slightly shorter median OS than CPX-351.

Introduction Treatment approaches in patients aged over 60 years diagnosed with acute myeloid leukemia (AML) are still showing unsatisfactory results.1-3 Many studies have shown poor prognosis features in this setting, as compared to younger patients, such as more adverse cytogenetics, worse performance status or higher rate of secondary AML (sAML).1-2 Frequency of sAML varies between 20-30% of all cases, showing dismal prognosis across different study cohorts.4-18 Several treatment options are now available for older sAML patients, such as hypomethylating agents (HMA) with or without venetoclax for those considered unfit for intensive therapy, and intensive chemotherapy (3+7 regimens or CPX-351) followed by an allogeneic hematopoietic stem cell transplant (allo-HSCT) for those considered fit enough. CPX-351 is a liposomal encapsulation of cytarabine and daunorubicin indicated for the treatment of patients diagnosed with AML after receiving cytotoxic, radiation or immunosuppressive treatments (t-AML) and those patients diagnosed with AML with myelodysplasia-related changes (AML-MRC),19 including this therapeutic indication a slightly different population than that enrolled in the pivotal phase III trial leading to its regulatory approval.20 In that study, CPX-351 showed improved complete remission with or without incomplete recovery (CR/CRi) rate and median overall survival (OS) compared with standard “3+7” regimens (47% vs. 33%, and 9.6 vs. 5.9 months, respectively). However, there is limited evidence regarding efficacy of classical intensive chemotherapy regimens in a similar real-life cohort of sAML and AML-MRC patients aged 60 to 75 years. This study aims to analyze the main characteristics and therapeutic results of sAML and AML-MRC patients receiving classical intensive chemotherapy (before CPX-351 became available) in a large epidemiological registry of the “Programa Español para el Tratamiento de las Hemopatías Malignas” (PETHEMA) (clinicaltrials gov. Identifier: NCT02607059).

Methods Eligibility All patients with diagnosis of sAML or AML-MRC aged between 60 and 75 years who were treated with intensive chemotherapy (excluding CPX-351) and reported to the

multinational PETHEMA AML registry until May, 31st 2020 were included in the study.17 Informed consent was a requisite for patients and the corresponding Research Ethics Board approved the study according to the Declaration of Helsinki. Induction treatment Patients receiving standard-dose cytarabine (100-200 mg/m2/day, days 1 to 7) with idarubicin (10-12 mg/m2/day, days 1 to 3) or daunorubicin (60 mg/m2/day, days 1 to 3) were classified in the “3+7” cohort. When the 3+7 regimen was dose-reduced or another drug was added to the induction schedule, therapy was considered not standard, and patients were considered as “other intensive therapy” group (Online Supplementary Table S1). Study definitions and endpoints According to the 2016 World Health Organization (WHO) classification,19 AML-MRC diagnosis includes AML with MDS-related cytogenetic abnormalities, those previously diagnosed with MDS (MDS-AML) or chronic myelomonocytic leukemia (CMML-AML) and AML with morphological multilineage dysplasia. Patients without information related to previous neoplastic antecedents but with MDSrelated cytogenetic abnormalities were assessed in a different type of AML group (unknown). Similarly, those patients with dysplasia but without information regarding previous therapy, neoplastic antecedents or lack of cytogenetic diagnosis were also analyzed in the unknown group.19 MDS-AML, CMML-AML and t-AML were grouped as sAML. Cytogenetic results were stratified as per the Medical Research Council (MRC) classification.21 Complete remission (CR) and CR with incomplete recovery (CRi) required the absence of extramedullary disease, no blast cells in the peripheral blood (PB) and <5% of them in the bone marrow (BM).22 Neutrophil and platelet counts in PB should be >1x109/L and >100x109/L, respectively to achieve CR. Patients not achieving these neutrophil or platelet counts in PB after chemotherapy were considered CRi.22 Reduction of BM blasts >50% compared to the basal value and a total count between 5-25% were necessary to reach partial remission (PR). Induction death was defined as death before the patient was assessed for response after starting the first cycle of induction. Resistance was the result if patients did not meet the aforementioned criteria. Relapse required an increase of ≥5% in BM blast cells or their resurgence in PB or as extramedullary AML after having achieved CR/CRi.

Haematologica | 109 January 2024

116

ARTICLE - Intensive chemotherapy in older sAML and AML-MRC The main objective was to analyze the baseline characteristics in a real-life population and the response to intensive therapy in terms of CR/CRi rate, OS and event-free survival (EFS) in all registered patients. Other secondary endpoints were to compare different therapeutic approaches (“3+7” vs. “other intensive therapies”) and type of AML (sAML subgroups, de novo AML with MDS-related cytogenetic abnormalities or solely multilineage dysplasia). Causes to stop the intensive treatment cycles were also recorded, during induction and after remission. Postremission therapy was considered as completed when hemtopoietic stem cell transplantation (HSCT) was performed or when treatment was discontinued because of relapse before transplantation. Statistical analysis Patients’ baseline characteristics and response to induction therapy were expressed with the median and the interquartile range (IQR) and were analyzed with Χ2 with Yates’ correction, ANOVA, Wilcoxon, Kruskal-Wallis, ANOVA or Student’s t test, determined by the type of variable. OS and EFS prediction were computed from the date of AML diagnosis until the date of death for the first time-toevent element, whereas EFS was assessed until the date of treatment failure, relapse or death, depending on which occured first. These lifetime variables were estimated by the Kaplan-Meier estimator23 and they were compared with the log-rank test.24 In order to calculate medians and confidence intervals (CI), the corrected method of Kalbfleisch and Prentice was seleted.25 Multivariate analysis for OS was performed through a Cox proportional hazards model. Characteristics with clinically and statistically significant association in the univariate analysis (P<0.1) and those available variables with a possible relationship in previous studies were included in the analysis. Allogeneic HSCT (allo-HSCT) and autologous HSCT (auto-HSCT) were included as time-dependent variables, in a Cox proportional hazards regression with timedependent covariates. Afterwards, the Mantel-Byar test was performed and a Simon-Makuch plot was the method to represent them.26,27 Living patients were censored the last day they were alive before October, 31st 2021, when patients’ follow-up was updated. All mentioned P values are two-sided. Statistical computing and graphics were conducted using the R software.

D. Martínez-Cuadrón et al.

defined. A Consolidated Standards of Reporting Trials (CONSORT) diagram is detailed in Figure 1. Among the 765 included patients, median age was 66 years (IQR, 63-69), median Eastern Cooperative Oncology Group performance status (ECOG PS) was 1 (IQR, 0-1) and 61% were male. White blood cell (WBC) count was ≥20×109/L in 25% of patients and median blast cells in BM at diagnosis was 48% (IQR, 31-70). Karyotype was not available in 102 (13%) patients, and it was normal in 174 (23%). When available, cytogenetic risk was adverse in 358 (55%) patients and favorable in ten (2%), NPM1 mutation was present in 30 (8%) patients, and 32 (8%) were positive for FLT3-internal tandem duplication (ITD) mutation. Characteristics of “3+7” versus other intensive-therapy cohorts Overall, 389 (51%) received 3+7 and 376 (49%) other intensive therapy (Online Supplementary Table S1), being the “3+7” cohort older (P<0.001) and with more t-AML (P<0.001) (Table 1; Online Supplementary Table S2).

Results

Front-line therapy BM assessments were performed after every cycle of treatment. From 753 (98%) patients with available response assessment, 95 (13%) died during induction, CR/CRi was achieved after cycle 1 in 337 (45%), partial response (PR) in 78 (10%) and resistance in 243 (38%) (Figure 2). A second identical induction cycle was administered in 64 (8%) subjects, increasing the CR/CRi rate up to 49%. No differences in CR/CRi were observed between “3+7” and other intensive regimens (P=0.65) (Table 2). Factors associated with induction response and P<0.05 included age (P=0.02), ECOG PS (P<0.001), white blood cells (WBC) count (P=0.04), creatinine (P=0.04), uric acid (P=0.003), which were negative predictors, and albumin levels (P<0.001), MRC cytogenetic risk (P=0.002), and NPM1 mutation status (P<0.001), which were positively associated with response. No differences were found according to type of AML (therapy-related AML [t-AML] vs. secondary to MDS/CMML vs. de novo AML with MDS-cytogenetic vs de novo AML with dysplasia) or FLT3-ITD mutation status (P=0.37 and P=0.68, respectively; Table 3, Online Supplementary Table S3). When patients were included in a multinomial logistic regression, older age (P=0.004), and higher ECOG PS (P<0.001) were independent adverse prognostic factors for CR/CRi; while presence of NPM1 mutation (P<0.001), and favorable (P=0.03) and intermediate cytogenetic risk (P<0.001) were favorable prognostic factors.

Accrual and patient characteristics Until May 2020, 12,426 patients were registered and 5,090 (41%) were aged between 60 and 75 years. Of them, 2,376 (47%) had received intensive schedules as front-line therapy and 765 (15%) were sAML or AML-MRC as previously

Post-remission treatment Data on HSCT was available in 350 of 367 patients who achieved CR/CRi after induction. Of them, 28 (8%) received autologous HSCT (auto-HSCT) and 73 (21%) allo-

Haematologica | 109 January 2024

117

ARTICLE - Intensive chemotherapy in older sAML and AML-MRC HSCT (Table 2). More patients in the “3+7” cohort received subsequent HSCT in first CR/CRi (statistical trend P=0.08). The total number of cycles of intensive chemotherapy during front-line therapy was available in 719 patients, with a mean number of 1.69 cycles (median: 1 cycle; IQR, 1-2). One cycle was administered in 399 (55%) patients, two cycles in 166 (23%), three cycles in 130 (18%) patients, four cycles in 20 (3%), five cycles in two (0.3%) and six cycles in one (0.1%) (Figure 2). Detailed information regarding second induction and consolidation regimens are shown in Online Supplementary Tables S4 and S5. Causes of front-line treatment withdrawal Overall, 143 patients completed post-remission therapy, 224 had incomplete post-remission schedule, and information on schedule completeness was not available in 30 patients. Causes of withdrawal were: 37 patients (11%) died in CR/CRi because of treatment complications, five (1%) rejected further intensive treatment, five (1%) failed to perform a planned auto-HSCT due to mobilization failure, 147 (44%) were considered unfit to receive further

D. Martínez-Cuadrón et al.

cycles (20 [6%] because of serious events during prior cycles, and 126 [38%] due to concomitant comorbidities, age or any other reason according to physicians’ criteria). In addition, three patients achieving PR after one cycle of induction therapy had no further information and 32 did not receive a second identical cycle: three died before starting it, 23 switched to alternative schedules, two suffered serious complications during the first cycle, and four were considered unfit to receive more therapies. Overall survival The median OS of the 765 patients included in the study was 7.6 months (95% CI: 6.7-8.5), 8.3 months (95% CI: 6.910.1) in the “3+7” cohort versus 6.8 months (95% CI: 5.6-8.1) in other intensive therapies group (P=0.05; Figure 3A; Table 2). In univariate analyses, prognostic factors for prolonged OS were: age <70 years (Figure 3B; P=0.003), female (Figure 3C; P=0.009), ECOG PS ≤2 (Figure 3D; P<0.001), WBC count <20×109/L (Figure 3E; P=0.048), platelet count >20×109/L (P=0.01), creatinine levels ≤1.3 mg/dL (P=0.001), uric acid ≤7 mg/dL (P=0.001), albumin >3.5 g/dL (P=0.003), lower per-

Figure 1. Consolidated Standards of Reporting Trials (CONSORT) diagram for adult patients aged between 60 and 75 years with secondary acute myeloid leukemia and acute myeloid leukemia with myeloid-related changes. AML: acute myeloid leukemia; sAML: secondary AML; MRC-AML: AML with myeloid-related changes; MDS: myelodisplastic syndrome; CMML: chronic myelomonocytic leukemia; t-AML: therapy-related AML; BSC: best supportive care. Haematologica | 109 January 2024

118

ARTICLE - Intensive chemotherapy in older sAML and AML-MRC centage of blast cells in BM (P=0.04), intermediate and favorable cytogenetic risk (Figure 3F; P=0.009) and mutated NPM1 (Figure 3G; P<0.001). No statistical differences were found as per FLT3-ITD mutational status, type of AML (Figure 3H), CR/CRi after one or two cycles of induction (Figure 3I; P=0.75), or prior treatment with anthracyclines or HMA (Table 3; Online Supplementary Table S3). Multivariate Cox regression showed that age ≥70 years (P=0.007), ECOG PS ≥1 (P=0.046) and higher WBC counts

D. Martínez-Cuadrón et al.

(P=0.04) were independent adverse prognostic factors for OS. Intermediate cytogenetic risk (P=0.016), and NPM1 mutation (P=0.007) were favorable prognostic factors (Online Supplementary Table S6). The Cox proportional hazard regression with time-dependent covariates showed ECOG PS 3 (hazard ratio [HR] =2.27; 95% CI: 1.50-3.42; P<0.001), ECOG PS 4 (HR=3.83; 95% CI: 1.42-10.38; P=0.008) and no allo- or auto-HSCT (HR=17.26; 95% CI: 7.74-38.47; P<0.001) as independent adverse prognostic factors.

Table 1. Demographic and baseline characteristics of the study population (“3+7” vs. other intensive therapy). Overall Characteristic

Median (IQR)

Total Age in years <70 ≥70

“3+7” N (%)

Median (IQR)

765 (100) 66 (63-69)

Sex Male Female

765 (100) 592 (77) 173 (23)

Other intensive therapy N (%)

Median (IQR)

389 (100) 64 (62-67)

765 (100) 466 (61) 299 (39)

389 (100) 335 (86) 54 (14)

N (%)

376 (100) 67(64-70)

389 (100) 232 (60) 157 (40)

376 (100) 257 (68) 119 (32) 376 (100) 234 (62) 142 (38)

<0.001* <0.001

0.51

ECOG PS 0 1 2 3 4

1 (0-1)

676 (100) 214 (32) 317 (47) 112 (17) 28 (4) 5 (1)

1 (0-1)

342 (100) 115 (34) 149 (44) 59 (17) 15 (4) 4 (1)

1 (0-1)

334 (100) 29 (30) 168 (50) 53 (16) 13 (4) 1 (0)

0.80* 0.34

WBC, ×109/L <20 ≥20

4.3 (2-20)

728 (100) 545 (75) 183 (25)

4.7 (2-20.8)

359 (100) 263 (73) 96 (27)

4.1 (1.8-17.3)

369 (100) 282 (76) 87 (24)

0.60* 0.37 0.83

MRC cytogenetic risk Favorable Intermediate Adverse

654 (100) 10 (2) 286 (44) 358 (55)

332 (100) 6 (2) 144 (43) 182 (55)

322 (100) 4 (1) 142 (44) 176 (55)

FLT3-ITD Positive Negative

412 (100) 32 (8) 380 (92)

240 (100) 21 (9) 219 (91)

172 (100) 11 (6) 161 (94)

0.49

NPM1 Positive Negative

375 (100) 30 (8) 345 (92)

227 (100) 20 (9) 207 (91)

148 (100) 10 (7) 138 (93)

0.60

Type of AML t-AML Previous anthracycline No previous anthracycline NA Previous MDS/CMML Previous HMA No previous HMA NA MDS cytogentics Dysplastic changes only Unknown#

765 (100) 155 (20) 40 (26) 82 (53) 33 (21) 235 (31) 39 (17) 119 (51) 77 (33) 242 (32) 82 (11) 51 (7)

389 (100) 84 (22) 20 (24) 50 (60) 14 (17) 89 (23) 15 (17) 45 (51) 29 (33) 145 (37) 47 (12) 24 (6)

376 (100) 71 (19) 20 (28) 32 (45) 19 (27) 146 (39) 24 (16) 74 (51) 48 (33) 97 (26) 35 (9) 27 (7)

<0.001 0.16†

0.99†

IQR: interquartile range; ECOG PS: Eastern Cooperative Oncology Group performance status; sAML: secondary acute myeloid leukemia; WBC: white blood cells; BM: bone marrow; MRC: Medical Research Council; FLT3: FMS-like tyrosine kinase 3; ITD: internal tandem duplication; NPM1: Nucleophosmin1; HMA: hypomethylating agents; MDS: myelodisplastic syndrome; t-AML: therpay-related AML; CMML: chronic myelomonocytic leukemia; NA: not applicable; *P compares continuous variables; †P compares previous treatment; #this group includes patients with dysplasia but no information regarding previous therapy, neoplastic antecedents or cytogenetics and patients with MDS/CMML-related cytogenetics but no information regarding previous therapy or neoplastic antecedents.

Haematologica | 109 January 2024

119

ARTICLE - Intensive chemotherapy in older sAML and AML-MRC Impact of post-remission therapy on overall survival We analyzed the impact of post-remission schedule on OS (Table 4). Among 176 patients withdrawing post-remission schedule as per physicians’/patient decision or mobilization failure, statistical differences were found after stratifying between one cycle and >1 cycle, with longer OS in the later group in comparison to the former one i.e., 1 cycle: median and 5 years OS of 13.6 months; 95% CI: 8.9-23 and 5%; 95% CI: 0.8-30.4 versus >1 cycle: 18.3 months; 95% CI: 15.2-23.7 and 14%; 95% CI: 8.9-23; P=0.01, respectively (Figure 4A). Median and 5 years OS of patients undergoing allo-HSCT after achieving first CR/CRi were 27 months (95% CI: 20.5not applicable [NA]) and 34.2% (95% CI: 21.8-53.5) versus 37 months (95% CI: 27.3-NA) and 31.4% (95% CI: 18.0-54.7) in those receiving auto-HSCT, and 12.1 months (95% CI: 10.1-14.1) and 8.8% (95% CI: 5.5-14.1) in those not consolidated with HSCT (P<0.001), respectively. According to the

D. Martínez-Cuadrón et al.

Mantel-Byar test, patients receiving HSCT in first CR/CRi had better OS in comparison with those who did not (P<0.001), with a significant difference both allo-HSCT and auto-HSCT versus no HSCT (P=0.003 and P=0.008, respectively) (Figure 4B). No differences were observed in OS between allo-HSCT and auto-HSCT i.e., the median OS from the date of HSCT was 18.9 months (95% CI: 9.9-81.7) and 25.2 (95% CI: 13-54.1), respectively (P=0.97). Event-free survival The median EFS of the 715 AML patients with available data for this analysis was 2.7 months (95% CI: 2-3.3), with 1- and 5-year EFS of 22.8% (95% CI: 19.9-26.2) and 6.8% (95% CI: 5-9.3), respectively. No significant differences were observed between “3+7” and other intensive schedules cohorts, with 1- and 5-year EFS of 23.5% (95% CI: 19.4-28.4) and 7.9 (95% CI: 5.3-11.9) versus 22.2% (95% CI: 18.1-27.1) and 5.7 (95% CI: 3.5-9.3), respectively (P=0.31)

Figure 2. Induction response and post-remission therapy. Auto-HSCT: autologous stem cell transplant; Allo-HSCT: allogeneic HSCT; HMA: hypomethylating agent; CR: complete remission; CRi: complete remission with incomplete recovery. Haematologica | 109 January 2024

120

ARTICLE - Intensive chemotherapy in older sAML and AML-MRC (Online Supplementary Figure S1; Table 2). The median EFS according to HSCT in first CR/CRi from the 698 patients with available data (allo-HSCT vs. auto-HSCT vs. no HSCT) were 25.1 months (95% CI: 16.0-NA) versus 28.3 (95% CI: 15.2-94.1) versus 1.6 (95% CI: 1.5-1.9), respectively. From the date of diagnosis, the 5-year EFS were 35.3% (95% CI: 23.1-54.1) in the allo-HSCT group, 28.6% (95% CI: 15.9-51.3) in the auto-HSCT, and 2.9% (95% CI: 1.7-4.8) in the noHSCT group (P<0.001). According to the Mantel-Byar test, patients receiving HSCT in first CR/CRi had better EFS in comparison with those who did not (P<0.001), with significant difference both allo-HSCT and auto-HSCT versus no HSCT (P<0.001 and P=0.02, respectively) (Figure 4C). No differences were observed in EFS between allo-HSCT and auto-HSCT (P=0.81). CPX-351-like cohort The so called “CPX-351-like” cohort included 673 patients, after removing 92 patients who did not fulfill disease characteristics for inclusion in the pivotal phase III trial leading to CPX-351 approval (i.e., those patients with multilineage dysplasia-AML as sole criterion for AML-MRC classification). Median age of this cohort was 66 years (IQR, 63-69), 264 (39%) patients were female, 28 (5%) had an ECOG PS >2, 466 (73%) had a WBC count <20x109/L,

D. Martínez-Cuadrón et al.

and 356 (62%) had an adverse MRC cytogenetic risk. From 661 with available data, 293 (44%) achieved CR/CRi after one cycle of induction cycle and 27 (4%) after two cycles with the same schedule (CR/CRi rate of 48%). Eighty-two patients (12%) died before assessment, 44 (7%) achieved partial response (PR) and 215 (33%) were considered resistant. Median OS was 7.6 months (95% CI: 6.5-8.4), without differences in CR/CRi rate and OS between “3+7” and other intensive therapy cohorts (P=0.6 and P=0.2, respectively). Median EFS was 2.0 months (95% CI: 1.9-3.3) and no differences were observed regarding induction therapy i.e., the median EFS in “3+7” cohort was 2.7 months (95% CI: 2.0-3.8) versus 2.0 (95% CI: 1.6-3.6) in the other group (Online Supplementary Figure S2A; P=0.66) or other type of AML i.e., 3.3 months in t-AML group (95% CI: 1.9-5.1) versus 2.5 in MDS/CMML-AML (95% CI: 1.9-4.5) versus 1.8 (95% CI: 1.5-3.2) in the MDS cytogenetics (Online Supplementary Figure S2B; P=0.38). Among those patients achieving CR/CRi, there were data related to HSCT in front-line therapy in 305 (64 underwent an allo-HSCT, 21 auto-HSCT, and 220 did not receive HSCT). Median OS was higher in patients receiving an auto-HSCT or an alloHSCT versus no HSCT i.e., 35.8 months (95% CI: 14-NA) versus 27.0 (95% CI: 20.5-NA) versus 12.7 (95% CI: 10.314.6), with 1-year OS of 81% (95% CI: 65.8-99.6), 74.6%

Table 2. Induction response, hematopoietic stem cell transplantation rates in first complete remission/complete remission with incomplete recovery, overall survival and event-free survival according to therapeutic group. Variable

Overall N (%)

"3+7" N (%)

Other N (%)

All patients ORR (CR + CRi) CR + CRi after first induction CR + CRi after second induction PR Resistance Death

753 (100) 367 (49) 337 (45) 30 (4) 53 (7) 238 (32) 95 (13)

385 (100) 193 (50) 176 (46) 17 (4) 30 (8) 113 (29) 49 (13)

303 (100) 174 (47) 161 (44) 13 (4) 23 (16) 125 (34) 46 (12)

HSCT in first CR/CRi Allogeneic HSCT rate Autologous HSCT rate No HSCT

350 (100)* 73 (21)* 28 (8) 249 (71)

182 (100) 44 (24) 18 (10) 120 (66)

168 (100) 29 (17) 10 (6) 129 (77)

OS Median, months (95% CI) 1 year, % (95% CI) 3 years, % (95% CI) 5 years, % (95% CI)

765 (100) 7.6 (6.7-8.6) 35.1 (31.7-38.8) 13.8 (11.0-16.5) 7.9 (5.9-10.6)

389 (100) 8.3 (6.9-10.1) 37.4 (32.6-42.9) 15.8 (12.1-20.5) 10.8 (7.6-15.4)

376 (100) 6.8 (5.6-8.1) 32.8 (28.3-38.0) 11.8 (8.7-15.9) 5.6 (3.4-9.0)

EFS Median, months (95% CI) 1 year, % (95% CI) 3 years, % (95% CI) 5 years, % (95% CI)

715 (100)# 2.7 (2-3.3) 22.8 (19.9-26.2) 10.2 (7.8-12.6) 6.8 (5-9.3)

364 (100)# 2.9 (2.2-3.8) 23.5 (19.4-28.4) 11.6 (8.5-15.8) 7.9 (5.3-11.9)

351 (100)# 2.2 (1.7-3.6) 22.2 (18.1-27.1) 8.8 (6.1-12.6) 5.7 (3.5-9.3)

0.65

0.08

0.05

0.31

PR: partial response; CR: complete remission; CRi: CR with incomplete recovery; HSCT: hematopoietic stem cell transplantation; OS: overall survival; CI: confidence interval; EFS: event-free survival; *1 patient who received allogeneic HSCT as first cycle and 2 patients without information regarding consolidation cycles, but underwent allogeneic HSCT, were included in this analysis; #patients with available data for EFS analysis.

Haematologica | 109 January 2024

121

ARTICLE - Intensive chemotherapy in older sAML and AML-MRC (95% CI: 64.2-86.6) and 52.1 (95% CI: 45.7-59.3), respectively (P<0.001). According to the Mantel-Byar test, patients receiving HSCT in first CR/CRi had better OS in comparison with those who did not (P=0.018), with a significant difference between allo-HSCT and no HSCT (P=0.029), and a trend between auto-HSCT and no HSCT (P=0.08) (Figure 4D). No differences were observed in OS between alloHSCT and auto-HSCT i.e., the median OS from the date of HSCT were 18.9 months (95% CI: 9.9-81.7) and 29.4 (95% CI: 7.3-149.5), respectively (P=0.83). After selecting patients with ECOG PS 0-2, serum creatinine <2.0 mg/dL and serum total bilirubin <2.0 mg/dL, 368 were analyzed. Median age was 66 years (IQR, 63-69) and baseline characteristics were similar to patients included in the CPX-

D. Martínez-Cuadrón et al.

351-like cohort. The CR/CRi rate in this group was 51%, with a median OS and EFS of 7.7 months (95% CI: 6.6-9.1) and 2.8 months (95% CI: 2-3.9), respectively (Online Supplementary Table S7).

Discussion This study shows the baseline characteristics and the clinical outcomes in a real-life cohort of patients aged between 60 and 75 years, who were diagnosed with t-AML or MRCAML and treated with intensive schedules. This cohort included 765 patients (roughly 7% of all AML registered patients in the study period), achieving a CR/CRi rate of 48%

Table 3. Induction response and overall survival according to baseline characteristics. Induction response, N(%) Variable

CR/CRi PR/RES

Age in years <70 ≥70

300 (51) 216 (37) 67 (40) 75 (44)

Sex Male Female

212 (46) 190 (42) 155 (52) 101 (34)

ECOG PS 0-2 3-4

313 (49) 248 (39) 7 (22) 13 (41)

Induction death

OS P

N (%)

Median (95% CI)

At 1 year (95% CI)

765 (100) 0.02 592 (77) 8.1 (6.9-9.2) 37 (33.1-41.3) 173 (23) 6.3 (4.6-7.7) 28.8 (22.7-36.6) 55 (12) 40 (14)

0.12

At 3 years (95% CI)

At 5 years (95% CI)

16 (13-19.7) 6.5 (3.5-11.9)

9.4 (6.9-12.8) 3.6 (1.6-8.4)

0.003

765 (100) 466 (61) 6.8 (6.0-8.1) 32.4 (28.3-37.1) 10.5 (7.8-14) 5.8 (3.8-9.1) 0.009 299 (39) 8.7 (7.1-10.3) 39.3 (33.9-45.5) 19.1 (14.7-24.9) 11.3 (7.6-16.7)

74 (12) 12 (37)

676 (100) <0.001 643 (95) 7.8 (6.8-9.1) 36.5 (32.9-40.6) 14.2 (11.5-17.5) 33 (5) 1.5 (1.2-4.2) 6.2 (1.6-23.9) 3.1 (0.5-21.5) 728 (100) 0.04 545 (75) 8.1 (6.8-9.4) 35.8 (31.8-40.3) 183 (25) 5.4 (4.2-7.9) 32.1 (25.9-39.8)

WBC, ×109/L <20 ≥ 20

277 (51) 196 (36) 75 (41) 83 (45)

65 (12) 25 (14)

MRC cytogenetic risk Favorable Intermediate Adverse

8 (80) 1 (10) 161 (57) 91 (32) 153 (43) 156 (44)

1 (10) 30 (11) 47 (13)

0.002

FLT3-ITD Positive Negative

15 (47) 12 (37) 202 (54) 131 (35)

5 (16) 43 (11)

0.68

NPM1 Positive Negative

27 (90) 3 (10) 170 (50) 129 (34)

0 (0) 42 (12)

Type of AML t-AML 83 (55) 47 (31) Previous anthracycline 20 (50) 15 (37) No previous anthracycline 46 (57) 23 (28) Previous MDS/CMML 105 (46) 94 (41) Previous HMA 15 (41) 18 (49) No previous HMA 55 (47) 44 (38) MDS cytogentics 112 (46) 100 (41) Dysplastic changes only 44 (54) 27 (33)

22 (14) 5 (12) 12 (15) 28 (12) 4 (11) 17 (15) 29 (12) 11 (13)

13.9 (11-17.6) 11.8 (7.8-18)

0.5†

<0.001

8.5 (6.1-11.9) 5.5 (2.8-10.9)

0.048

654 (100) 10 (2) 11.5 (3.9-NA) 45.7 (22.4-93.2) 22.9 (6.8-76.8) 286 (44) 9.2 (7.1-11.9) 42.6 (37-49) 17.7 (13.4-23.3) 8.6 (5.4-13.9) 358 (55) 6.9 (5.6-8.2) 29.8 (25.3-35.1) 10.2 (7.3-14.4) 6.9 (4.4-10.7) 412 (100) 32 8) 6.9 (3.8-15.5) 380 (92) 8.1 (6.9-9.8)

714 (100) 155 (22) 6.8 (5.2-8.7) 40 (33) 6.1 (4.2-9.8) 82 (67) 7.1 (4.3-10.3) 235 (33) 8.1 (6-10.3) 39 (25) 7.8 (5.1-13.6) 119 (75) 9.1 (5.4-12) 242 (34) 7.2 (6-9.3) 82 (12) 9.2 (6-13.6)

0.009

34.1 (21-55.4) 23.4 (12.2-44.8) 23.4 (12.2-44.8) 0.71 37.6 (32.8-43) 13.9 (10.5-18.4) 6.5 (3.8-11.2)

375 (100) <0.001 30 88) 16.4 (8.1-NA) 59.1 (43.7-79.9) 41.1 (26.4-64) 345 (92) 7.7 (6.7-9.2) 34.5 (29.6-40.2) 11.8 (8.5-16.4) 0.37 0.6*

8 (5.9-11) -

27.4 (11-68.4) <0.001 5 (2.6-9.4)

33.1 (26.2-41.8) 15.7 (10.6-23.3) 9.6 (5-18.3) 0.59 24.3 (13.4-43.9) 15.2 (6.8-33.8) 10.1 (3.3-31.4) 0.59# 35.2 (25.9-47.8) 19.1 (11.9-30.9) 10.6 (3.9-28.9) 38.6 (32.7-45.6) 15.1 (10.9-21.1) 6.8 (3.7-12.2) 38.7 (25.8-58) 15 (6.5-35) 10 (3.1-32.1) 0.93¥ 39.3 (31.3-49.5) 19.2 (12.9-2837) 9.2 (4.5-18.7) 32.7 (27.1-39.4) 11.5 (7.8-17.1) 7.7 (4.6-12.8) 37.7 (28-50.8) 12.4 (6.4-23.8) 6.2 (2.2-17.4)

CI: confidence interval; CMML: chronic myelomonocytic leukemia; sAML: secondary acute myeloid leukemia; CR: complete remission; CRi: CR with incomplete recovery; HMA: hypomethylating agents; MDS: myelodysplastic syndrome; OS: overall survival; PR: partial response; RES: resistance; t-AML: therapy-related acute myeloid leukemia; ECOG PS: Eastern Cooperative Oncology Group performance status; WBC: white blood cells; MRC: Medical Research Council. *P value from 121 patients with available data regarding previous treatment with anthracyclines; #P value from 122 patients with available data regarding previous treatment with anthracyclines; †P value from 158 patients with available data regarding previous treatment with HMA; ¥ P value from 153 patients with available data regarding previous treatment with HMA.

Haematologica | 109 January 2024

122

ARTICLE - Intensive chemotherapy in older sAML and AML-MRC

D. Martínez-Cuadrón et al.

Continued on following page. Haematologica | 109 January 2024

123

ARTICLE - Intensive chemotherapy in older sAML and AML-MRC

D. Martínez-Cuadrón et al.

Figure 3. Impact of patient’s characteristics at diagnosis and number of cycles to achieve complete remission/complete remission with incomplete recovery. (A) Overall survival (OS) in secondary acute myeloid leukemia (sAML) and AML patients with myeloid-related changes (AML-MRC) according to therapeutic approach (“3+7” vs. other intensive therapy). (B) OS in sAML and AML-MRC according to age (60-69 vs. ≥70 years). (C) OS in sAML and AML-MRC according to sex. (D) OS in sAML and AML-MRC according to Eastern Cooperative Oncology Group performance status (ECOG PS). (E) OS in sAML and AML-MRC according to value of white blood cell (WBC) count. (F) OS in sAML and AML-MRC according to cytogenetic risk. (G) OS in sAML and AML-MRC according to NPM1 mutation. (H) OS according to the type of AML (t-AML, sAML myelodysplastic syndrome/chronic myelomonocytic leukemia [MDS/CMML], MDS cytogenetics and multilineage dysplasia). (I) OS sAML and AML-MRC between patients who achieved complete remission/complete remission with incomplete recovery (CR/CRi) after 1 or 2 cycles of induction; yo: years old.

after one or two induction cycles, and median OS of 7.6 months (95% CI: 6.7-8.5), roughly 2 months better than observed in the 3+7 control arm of the pivotal phase III trial.20 In this study, we have analyzed outcomes of patients with AML and multilineage dysplasia at diagnosis according to the CPX-351 regulatory approval, although isolated morphological multilineage dysplasia was not among inclusion criteria for the pivotal phase III trial.20 When we take into account the similar study population to the pivotal phase III trial (CPX-351-like cohort), median OS remained unchanged at 7.6 months. We must underline that the CPX351 data comes from a randomized trial, which is the gold standard to have new therapies for our patients. Thus, our comparison should be taken with a great deal of caution. We can speculate on reasons for non-comparable results between our real-life series and the control arm of the pivotal phase III trial: i) our cohort included patients with worse clinical features (i.e., deteriorated ECOG PS or organ

function, more patients with adverse MRC cytogenetic risk or WBC count ≥20x109/L) as compared to those enrolled in clinical trial; ii) given the open label design of the phase III trial, control arm patients could withdraw earlier frontline therapy as they were not assigned to experimental arm; iii) although our study cohort comes from an epidemiologic registry, a positive or negative reporting bias could occur; and iv) a marked variability in cytarabine plus anthracycline induction regimens was noted in our series, where half of the patients received upfront “3+7”, and half received “other intensive therapies” with addition of a third drug, dose reductions or intensive clinical trials. On the other hand, we can also hypothesize that the adverse effect of FLT3-ITD mutations was not observed in our study due to: i) its association with higher WBC counts, which was shown as an independent factor for survival; and ii) our cohort shows lower prevalence of FLT3-ITD mutation (8%) in comparison with published series with

Haematologica | 109 January 2024

124

ARTICLE - Intensive chemotherapy in older sAML and AML-MRC

D. Martínez-Cuadrón et al.

Table 4. Overall survival according to cycles of intensive chemotherapy and post-remission approach. Characteristic

N (%)

Median OS months (95% CI)

1-year OS % (95% CI)

3-year OS % (95% CI)

5-year OS % (95% CI)

Overall* 1 cycle 2 cycles 3 cycles >3 cycles

718 (100) 399 (56) 166 (23) 130 (18) 23 (3)

3.5 (3-4.5) 9.7 (8.1-13.4) 16 (14-23.7) 22.3 (14-NA)

18.6 (15.1-23.1) 43.2 (36.2-51.7) 64.6 (56.7-73.6) 78.3 (63.1-97.1)

3.8 (2.1-6.8) 18.7 (13.2-26.6) 30.7 (23.4-40.2) 41.9 (25.6-68.5)

2.9 (1.5-5.8) 11.6 (6.9-19.7) 14.8 (9.3-23.4) 28.7 (13.9-59.3)

<0.001

Withdrawn patients# After 1 cycle After 2 cycles After 3 cycles After >3 cycles

176 (100) 43 (24) 40 (23) 77 (44) 16 (9)

13.6 (9.6-27.8) 14.6 (10.1-30) 18.6 (15.3-25) 19.4 (14-NA)

53.8 (40.5-71.4) 57.1 (43.1-75.4) 75.2 (65.8-85.8) 81.2 (64.2-100)

14.4 (5.7-36.3) 25.4 (14.3-45) 30.1 (21-43) 37.5 (19.9-70.6)

4.8 (0.8-30.4) 17.4 (8-38) 12.2 (6.3-23.8) 20 (6.6-60.7)

0.07

Withdrawn patients# After 1 cycle After >1 cycle

176 (100) 43 (24) 133 (76)

13.6 (9.6-27.8) 18.3 (15.2-23.7)

53.8 (40.5-71.4) 70.6 (63.1-79)

14.4 (5.7-36.3) 29.6 (22.5-39.1)

4.8 (0.8-30.4) 14.2 (8.9-23)

0.01

Withdrawn patients# Switch to HMA/Non-IC EoT

176 (100) 40 (23) 136 (77)

12 (9.8-27.8) 17.2 (15.2-23.5)

48.9 (35.4-67.4) 71.9 (64.5-80.1)

13.9 (5.3-36.8) 29.3 (22.3-38.7)

13.7 (8.6-21.9)

0.1

HSCT performed in first CR/CRi Allogeneic HSCT rate Autologous HSCT rate No HSCT

350 (100)* 73 (21) 28 (8) 249 (71)

27.0 (20.5-NA) 37.0 (27.3-NA) 12.1 (10.1-14.1)

75.5 (65.7-86.7) 82.1 (69.1-97.6) 50.2 (44.2-57.0)

46.3 (34.9-61.4) 50 (34.5-72.4) 19.5 (14.8-25.6)

34.2 (21.8-53.5) 31.4 (18.0-54.7) 8.8 (5.5-14.1)

<0.001

CI: confidence interval; EoT: end of treatment; HMA: hypomethylating agents; NA: not applicable; Non-IC: non-intensive chemotherapy; OS: overall survival; HSCT: hematopoietic stem cell trasnplantation; CR: complete remission; CRi: CR with incomplete recovery. *Overall cohort with available data; #withdrawn patients according to physicians’ criteria, patient decision or failure of mobilization (N=176).

de novo and younger patients. Interestingly, no significant differences in CR/CRi, OS or EFS were observed between both treatment groups, with the exception of a trend to better OS in the “3+7” group, probably related to older age observed in the “other intensive therapies” group. The main clinical outcomes in our CPX-351-like cohort were better as compared with the “3+7” control arm of the pivotal phase III trial of CPX-351.19 Although similar baseline characteristics were observed in both studies regarding age (66 vs. 67.8 years) and sex (39% vs. 38.6% female), some differences were observed in WBC count <20x109/L (73% vs. 85.6%) and ECOG PS >2 (5% vs. 0%).19 Our CPX-351-like cohort achieved similar CR/CRi (48% vs. 47%) and EFS (2.4 vs. 2.5 months), and worse OS than experimental CPX-351 arm (7.6 vs. 9.6 months), but our cohort showed better outcomes than the “3+7” comparator of phase III trial (CR/CRi 33%, EFS 1.3 months, OS 5.9 months).19 Early mortality rates with CPX351 were 13.7% through day 60, while our cohort showed 13% mortality during the first induction cycle. Regarding HSCT, 29.4% received allo-HSCT in phase III trial (34% CPX-351 cohort vs. 25% 3+7 cohort) compared to 21% of patients in our series. As in the Lancet et al. study19 and a recent population-based study,26 we show that patients undergoing allo-HSCT after induction had significantly better OS and EFS. Moreover, we were able to show that performing auto-HSCT or administering more consolidation cycles increased OS, highlighting the inherent favorable selection bias of AML patients completing their

planned consolidation and intensification schedules. Recent real-world studies analyzed relatively small cohorts of t-AML and AML-MRC older patients treated with CPX-351.29-32 A German31 series showed 47% CR/CRi rate, but Italian29 and French30 cohorts reported higher responses (70% and 59%). Interestingly, several real-life studies with CPX-351 have reported better median OS as compared to the phase III trial (e.g., 16.130 2131 1532 months and median not reached29). Several factors could explain better results observed in real world evidence studies with CPX-351: i) they included adult patients without limit of age, and ii) they analyzed more recent periods were allo-HSCT could be improved and more available as compared with the phase III and our historical cohort. We found that older age, higher ECOG PS and WBC count were independent adverse factors for CR/CRi and OS, whereas favorable/intermediate cytogenetic risk and NPM1 mutation were favorable factors. Similarly, Lancet et al. described ECOG PS, cytogenetic risk classification, and WBC count as prognostic factors after multivariate analysis.20 Recent series of patients treated with CPX-351 showed that non-spliceosome,30 TP53, and PTPN11 mutations30 were adverse risk factors, but we could not analyze their impact due to the limited information on these mutations in our series. We revealed the favorable impact of NPM1 mutations, while others failed probably due to the scarce number of NPM1-mutated patients among the targeted population. As expected, there were differences in age and

Haematologica | 109 January 2024

125

ARTICLE - Intensive chemotherapy in older sAML and AML-MRC

D. Martínez-Cuadrón et al.

Figure 4. Impact of post-remission therapy. (A) Overall survival (OS) according to cycles of intensive treatment (1 cycle vs. >1 cycle) in those patients withdrawn according to physicians’ criteria, patient decision or failure of mobilization (n=176). (B) SimonMakuch plot of OS according to hematopoietic stem cell transplantation (HSCT) in first complete remission/complete remission with incomplete (CR/CRi) (allogeneic HSCT [allo-HSCT] vs. autologous HSCT [auto-HSCT] vs. no HSCT). (C) Simon-Makuch plot of event-free survival (EFS) according to HSCT in first CR/CRi (allo-HSCT vs. auto-HSCT vs. no HSCT). (D) Simon-Makuch plot of OS in the CPX-351-like cohort according to HSCT in first CR/CRi (allo-HSCT vs. auto-HSCT vs. no HSCT).

type of AML among patient subgroups according to HSCT (Online Supplementary Table S8). After performing the Mantel-Byar test to analyze the impact of HSCT on survival, only ECOG PS ≥3 and not performing a HSCT in first CR/CRi remained as independent adverse factors for OS. Some limitations should be addressed: i) this is a real-life analysis with great heterogeneity and not a truly population-based registry; ii) missing data and retrospective analyses; and iii) secondary-AML type mutations (e.g., TP53, ASXL1, RUNX1, IDH) were not available. However, the

Lancet study did not show these genetic characteristics as well, so the mutational landscape of both series cannot justify similar results in OS between our historical cohort and the CPX-351 phase III. In conclusion, this large study in older patients with t-AML or MRC-AML suggests that classical intensive chemotherapy could lead to similar CR/CRi rates with slightly shorter median OS than CPX-351. We confirm improved OS after allo-HSCT in this setting, and we suggest the role of intensification cycles to improve long-term outcomes. The

Haematologica | 109 January 2024

126

ARTICLE - Intensive chemotherapy in older sAML and AML-MRC advantage of CPX-351 has been established in a phase III pivotal trial, but this should be confirmed in larger registry studies focusing on baseline mutational status, measurable residual disease, and toxicities. Disclosures DM-C is on the speaker´s bureau and advisory board and to have received travel and accommodation expenses from Jazz Pharmaceuticals. PM has received research grants from Jazz Pharmaceuticals, he is on the speaker´s bureau and advisory board of Jazz Pharmaceuticals, he has received research grants from Abbvie, is on the speaker´s bureau, advisory board from Abbvie. All other authors have no conflicts of interest to disclose. Contributions DMC, JEMV and PM conceived the study, analyzed, interpreted the data and wrote the paper; DMC, JEMV, CG, TB,

D. Martínez-Cuadrón et al.

MT, PMS, CRMJS, PH, JAPS, MJS, JB, ELR, MLA, CB, JLLL, MMPE, MBV, MC, BdR, RGB, SM, JGS, MLP, MIGR, VN, AL, JL, AC, CS, LA, MS,ASA, BB and PM included data of patients treated in their institutions, reviewed the manuscript and contributed to the final draft. Acknowledgments The authors would like to thank María D. García, Carlos Pastorini, Asier Laria, Yolanda Mendizábal, and Teresa Martinez Sena for data collection and management. Funding This study was partially supported by the Jazz Pharmaceuticals and Cooperative Research Thematic Network (RTICC) grant RD12/0036/014 (ISCIII & ERDF). Data-sharing statement Requests for data sharing should be sent to P. Montesinos (montesinos_pau@gva.es).

References 1. 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. 2. Tallman MS, Wang ES, Altman JK, et al. Acute myeloid leukemia, version 3.2019, NCCN Clinical Practice Guidelines in Oncology. J Natl Compr Canc Netw. 2019;17(6):721-749. 3. Daver N, Wei AH, Pollyea DA, Fathi AT, Vyas P, DiNardo CD. New directions for emerging therapies in acute myeloid leukemia: the next chapter. Blood Cancer J. 2020;10(10):107. 4. Leone G, Mele L, Pulsoni A, Equitani F, Pagano L. The incidence of secondary leukemias. Haematologica. 1999;84(10):937-945. 5. Granfeldt Østgård LS, Medeiros BC, Sengeløv H, et al. Epidemiology and clinical significance of secondary and therapy-related acute myeloid leukemia: a national populationbased cohort study. J Clin Oncol. 2015;33(31):3641-3649. 6. Hulegårdh E, Nilsson C, Lazarevic V, et al. Characterization and prognostic features of secondary acute myeloid leukemia in a population-based setting: a report from the Swedish Acute Leukemia Registry population-based study of secondary AML. Am J Hematol. 2015;90(3):208-214. 7. Szotkowski T, Rohon P, Zapletalova L, Sicova K, Hubacek J, Indrak K. Secondary acute myeloid leukemia - a single center experience. Neoplasma. 2010;57(2):170-178. 8. Xu X-Q, Wang J-M, Gao L, et al. Characteristics of acute myeloid leukemia with myelodysplasia-related changes: A retrospective analysis in a cohort of Chinese patients: characteristics of AMLMRC in Chinese patients. Am J Hematol. 2014;89(9):874-881. 9. Aström M, Bodin L, Nilsson I, Tidefelt U. Treatment, long-term outcome and prognostic variables in 214 unselected AML patients in Sweden. Br J Cancer. 2000;82(8):1387-1392. 10. Schoch C, Schnittger S, Klaus M, Kern W, Hiddemann W, Haferlach T. AML with 11q23/MLL abnormalities as defined by the WHO classification: incidence, partner chromosomes, FAB subtype, age distribution, and prognostic impact in an unselected series of 1897 cytogenetically analyzed AML cases. Blood. 2003;102(7):2395-2402.

11. Wheatley K, Brookes CL, Howman AJ, et al. Prognostic factor analysis of the survival of elderly patients with AML in the MRC AML11 and LRF AML14 trials. Br J Haematol. 2009;145(5):598-605. 12. Ostgård LSG, Kjeldsen E, Holm MS, et al. Reasons for treating secondary AML as de novo AML. Eur J Haematol. 2010;85(3):217-226. 13. Kayser S, Döhner K, Krauter J, et al. The impact of therapyrelated acute myeloid leukemia (AML) on outcome in 2853 adult patients with newly diagnosed AML. Blood. 2011;117(7):2137-2145. 14. Oran B, Weisdorf DJ. Survival for older patients with acute myeloid leukemia: a population-based study. Haematologica. 2012;97(12):1916-1924. 15. Gangatharan SA, Grove CS, P’ng S, et al. Acute myeloid leukaemia in Western Australia 1991-2005: a retrospective population-based study of 898 patients regarding epidemiology, cytogenetics, treatment and outcome: AML in Western Australia. Intern Med J. 2013;43(8):903-911. 16. Medeiros BC, Satram-Hoang S, Hurst D, Hoang KQ, Momin F, Reyes C. Big data analysis of treatment patterns and outcomes among elderly acute myeloid leukemia patients in the United States. Ann Hematol. 2015;94(7):1127-1138. 17. Martínez-Cuadrón D, Megías-Vericat JE, Serrano J, et al. Treatment patterns and outcomes of 2310 patients with secondary acute myeloid leukemia: a PETHEMA registry study. Blood Adv. 2022;6(4):1278-1295. 18. Bertoli S, Tavitian S, BoriesP, et al. Outcome of patients aged 60-75 years with newly diagnosed secondary acute myeloid leukemia: a single-institution experience. Cancer Med. 2019;8(8):3846-3854 19. 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. 20. Lancet JE, Uy GL, Cortes JE, et al. CPX-351 (cytarabine and daunorubicin) liposome for injection versus conventional cytarabine plus daunorubicin in older patients with newly

Haematologica | 109 January 2024

127

ARTICLE - Intensive chemotherapy in older sAML and AML-MRC diagnosed secondary acute myeloid leukemia. J Clin Oncol. 2018;36(26):2684-2692. 21. Grimwade D, Walker H, Harrison G, et al. The predictive value of hierarchical cytogenetic classification in older adults with acute myeloid leukemia (AML): analysis of 1065 patients entered into the United Kingdom Medical Research Council AML11 trial. Blood. 2001;98(5):1312-1320. 22. Cheson BD, Bennett JM, Kopecky KJ, et al. Revised recommendations of the International Working Group for diagnosis, standardization of response criteria, treatment outcomes, and reporting standards for therapeutic trials in acute myeloid leukemia. J Clin Oncol. 2003;21(24):4642-4649. 23. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53(282):457. 24. Mantel N. Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep 1966;50(3):163-170. 25. Kalbfleisch, JD, Prentice, RL. The statistical analysis of failure time data. New York: John Wiley & Sons. 1980. 26. Mantel N, Byar D: Evaluation of response-time data involving transient states: an illustration using heart transplant data. J Am Stat Assoc. 1974;69(345):81-86.

27. Simon R, Makuch RW. A non-parametric graphical representation of the relationship of an event: application to responser versus non-responder bias. Stat Med. 1984;3(1):35-44. 28. Nilsson C, Hulegårdh E, Garelius H, et al. Secondary acute myeloid leukemia and the role of allogeneic stem cell transplantation in a population-based setting. Biol Blood Marrow Transplant. 2019;25(9):1770-1778. 29. Guolo F, Fianchi L, Minetto P, et al. CPX-351 treatment in secondary acute myeloblastic leukemia is effective and improves the feasibility of allogeneic stem cell transplantation: results of the Italian compassionate use program. Blood Cancer. 2020;10(10):96. 30. Chiche E, Rahmé R, Bertoli S, et al. Real-life experience with CPX-351 and impact on the outcome of high-risk AML patients: a multicentric French cohort. Blood Adv. 2021;5(1):176-184. 31. Rautenberg C, Stölzel F, Röllig C, et al. Real-world experience of CPX-351 as first-line treatment for patients with acute myeloid leukemia. Blood Cancer. 2021;11(10):164. 32. Lee D, Jain AG, Deutsch Y, et al. CPX-351 yields similar response and survival outcome in younger and older patients with secondary acute myeloid leukemia. Clin Lymphoma Myeloma Leuk. 2022;22(10):774-779.

Haematologica | 109 January 2024

128

D. Martínez-Cuadrón et al.

ARTICLE - Cell Therapy & Immunotherapy

Treg-targeted IL-2/anti-IL-2 complex controls graft-versushost disease and supports anti-tumor effect in allogeneic hematopoietic stem cell transplantation Allan Thiolat,1* Caroline Pilon,1,2* Pamela Caudana,3,4 Audrey Moatti,1 Nhu Hanh To,1 Christine Sedlik,3,4 Mathieu Leclerc,1,5 Sébastien Maury,1,2,5 Eliane Piaggio3,4# and José L Cohen1,2# 1

University Paris Est Créteil, INSERM U955, IMRB, Créteil; 2AP-HP, Groupe Hospitalo-

Universitaire Chenevier Mondor, Centre d’Investigation Clinique Biothérapie, Fédération Hospitalo-Universitaire TRUE, Créteil; 3INSERM U932, PSL Research University, Institute Curie Research Center, Paris; 4Department of Translational Research, PSL Research University, Institute Curie Research Center, Paris and 5AP-HP, Groupe Hospitalo-Universitaire

Correspondence: J.L. Cohen jose.cohen@inserm.fr Received: Accepted: Early view:

January 6, 2023. September 7, 2023. September 14, 2023.

Chenevier Mondor, Service d’Hématologie Clinique, Créteil, France *

AT and CP contributed equally as first authors.

EP and JLC contributed equally as senior authors.

Abstract Modulating an immune response in opposite directions represents the holy grail in allogeneic hematopoietic stem cell transplantation (allo-HSCT) to avoid insufficient reactivity of donor T cells and hematologic malignancy relapse while controlling the potential development of graft-versus-host disease (GVHD), in which donor T cells attack the recipient’s tissues. IL-2/anti-IL-2 complexes (IL-2Cx) represent a therapeutic option to selectively accentuate or dampen the immune response. In dedicated experimental models of allo-HSCT, including also human cells injected in immunodeficient NSG mice, we evaluated side-by-side the therapeutic effect of two IL-2Cx designed either to boost regulatory T cells (Treg) or alternatively to activate effector T cells (Teff), on GVHD occurrence and tumor relapse. We also evaluated the effect of the complexes on the phenotype and function of immune cells in vivo. Unexpectedly, both pro-Treg and pro-Teff IL-2Cx prevented GVHD development. They both induced Treg expansion and reduced CD8+ T-cell numbers, compared to untreated mice. However, only mice treated with the pro-Treg IL-2Cx, showed a dramatic reduction of exhausted CD8+ T cells, consistent with a potent anti-tumor effect. When evaluated on human cells, pro-Treg IL-2Cx also preferentially induced Treg expansion in vitro and in vivo, while allowing the development of a potent anti-tumor effect in NSG mice. Our results demonstrate the clinical relevance of using a pro-Treg, but not a pro-Teff IL2Cx to modulate alloreactivity after HSCT, while promoting a graft-versus-leukemia effect.

Introduction Initially identified for its capacity to stimulate T cells in vitro, interleukin-2 (IL-2) has been used in the clinic for boosting effector immune responses for the treatment of metastatic melanoma and renal carcinoma,1 or for inducing remission maintenance in patients with acute myeloid leukemia,2 although with limited (5-15% of responder patients) or non-significant benefit, respectively. Additionally, highdose IL-2 administration is limited by its high associated toxicity. Low efficacy of IL-2 in cancer can be explained by (i) an insufficient activation of CD8+ T and natural killer (NK) cells,

which respond to IL-2 through the intermediate affinity IL-2 receptor, composed of IL-2Rβ and IL-2Rγc,3 and/or (ii) by the undesired activation of regulatory T cells (Treg). Once activated by IL-2, Treg block the anti-tumor immune response.4 Treg constitutively express the high affinity IL-2 receptor composed of three subunits: IL-2-Rα, IL-2Rβ and IL-2Rγc,4 giving them an advantage over other T cells in the consumption of IL-2, especially in environments with poor IL-2 content. Thus, although IL-2 has a pleiotropic activity, its major role is to favor Treg survival and suppressive functions.5 Consequently, low efficacy of IL-2 treatment in cancer is mainly due to low response of CD8+ T and NK cells expressing the intermediate affinity IL-2 receptor, as

Haematologica | 109 January 2024

129

ARTICLE - Treg-targeted IL-2/anti-IL-2 complex sustains GVHD prevention and GVL effect

well as the unwanted effect of IL-2 on Treg due to their constitutive high affinity IL-2 receptor expression. Understanding of the effect of IL-2 on Treg has led to the design of IL-2-based immunosuppressive strategies targeting Treg. Thus, contrary to what was initially proposed, IL-2 administered at low dose represents a novel immunosuppressive drug for the treatment of autoimmune and inflammatory diseases, acting by Treg activation. Allogeneic hematopoietic stem cell transplantation (alloHSCT) is a particularly interesting situation in which to test the therapeutic efficacy of IL-2. Indeed, the risk associated with insufficient reactivity between donor and recipient is the hematologic malignancy relapse for which allo-HSCT has been performed initially. In this case, high-dose IL-2 therapy could accentuate the graft-versus-leukemia/tumor (GVL/GVT) response through expansion and activation of donor effector T cells (Teff). On the other hand, the major complication of allo-HSCT is the development of life-threatening graft-versus-host disease (GVHD), in which donor T cells attack the recipient’s tissues.6 In this situation, lowdose IL-2 therapy could lead to Treg expansion, and thus to GVHD inhibition. Before the paradoxical effect of IL-2 on the immune response was understood, the group of David Sachs7-9 initially observed that 3 days of low-dose IL-2 administration controlled experimental GVHD, while preserving the GVL/GVT effects. At odds, Shin et al.10 reported a lack of therapeutic effect of low-dose IL-2 in experimental GVHD. We also studied different preventive or curative approaches based on low-dose IL-2 administration +/- rapamycin, but did not observe any effect on GVHD, including in a model of xenogeneic GVHD using human cells administered to immunodeficient mice.11 In humans, low-dose IL-2 therapy has been administered to dampen inflammation in GVHD,12-14 but with either no, limited, or not yet evidenced clinical recuperation. In our models, as well as in healthy volunteers or in allo-HSCT patients, low-dose IL-2 administration induced an approximately 1.5-2-fold expansion of Treg without activating conventional T cells (Tconv). This increase in Treg numbers was far away from the desired 20-fold needed to reach a 1:1 Treg:Tconv ratio and consequently, a protective effect of Treg in experimental GVHD, as we and other previously observed.15-19 Thus, it is actually difficult to conclude on the potential of sole IL-2 administration to prevent GVHD or to induce GVL/GVT effect. Overall, improvements on IL-2-based therapy are needed to increase therapeutic efficacy in both situations. IL-2/anti-IL-2 complexes (IL-2Cx) represent an alternative to enhance IL-2 therapeutic usefulness. Depending on the IL-2 monoclonal antibody (mAb) used to generate the IL2Cx, they can direct IL-2 action to either Teff or to Treg.20 Consequently, some complexes induce dramatic anti-tumor responses by vigorously activating effector immune cells.21 Other complexes show a selective effect on Treg expansion, resulting in immune response attenuation and autoimmune control.22 In allo-HSCT, the possibility of modulating

A. Thiolat et al.

donor T-cell alloreactivity in both directions depending on the used complexes remains unsolved.23-26 Moreover, no study has systematically compared the effect of different complexes side-by-side. Here, we tested two IL-2Cx. One is made with a pro-Treg mAb (i.e., preferentially activating Treg bearing the IL-2Rαβγ, now referred to as Cx25 as CD25 is the IL-2Rα chain). The mAb used in Cx25 binds to the part of IL-2 that normally binds to the IL-2Rβγ. As a result, IL-2 effect will be mainly mobilized for Treg that constitutively express CD25. The second one is made with a pro-Teff mAb (i.e., preferentially activating effector IL-2Rβγ- bearing cells, such as CD4+ conventional T cell, CD8+ T cells and NK cells, now referred to as Cx122 as CD122 is the IL-2Rβ chain). The mAb used in Cx122 binds to the part of IL-2 that normally binds to the high affinity IL-2Rα. As a consequence, IL-2 loses its preferential affinity for constitutively expressing CD25 Treg, leading to an improved availability of IL-2 for Tconv. We show that even if both complexes are effective in preventing GVHD, only the pro-Treg complex allows the emergence of the GVL/GVT effect.

Methods Mice and graft-host disease models Female C57BL/6J (B6, H-2b), BALB/cJ (H-2d), (B6xDBA2) F1 (B6D2F1, H-2bxd), (B6xC3H) F1 (B6C3F1, H-2bxk) from Charles River Laboratories (France) were used at 8-10 weeks of age. NSG (NOD.Cg-PrkdcscidIL2rgtm1Wjl/SzJ) mice were obtained from our own breeding and used at 10-12 weeks of age. All experiment protocols were approved by the local ethics committee (authorization no. APAFIS#11511-2017092610086943) and are in compliance with European Union Guidelines. GVHD was induced in B6C3F1 female injection of CD3+ T cells from B6 mice as previously described and evaluated three times a week.27,28 Treatment Recombinant human IL-2 (rhIL-2, Proleukine, Novartis), anti-CTLA4 (clone 9H10; Bio X Cell), anti-human IL-2 (clone MAB602; Bio-Techne), and anti-human IL-2 (clone 5344.111; BD Biosciences) were purchased. IL-2Cx were prepared by mixing 15,000 UI of rhIL-2 with 4.5 μg of mAb (molar ratio 2:1) and incubated for 30 minutes at 37°C. IL-2Cx is called Cx122 when made using MAB602 clone; and Cx25 when made using 5344.111 clone. IL-2Cx were administered in five or ten intraperitoneal injections from day 0 (d0) to d4 and d7 to d11 or only from d0 to d4 after bone marrow transplant (BMT) according to experiments. Anti-CTLA-4 was administered in three intraperitoneal mAb injections of 200 μg each on d0, d3, and d6 after BMT. Phenotypic analysis of immune cell populations Spleens were harvested on d12 and splenocytes stained with the antibodies listed in Online Supplementary Table

Haematologica | 109 January 2024

130

ARTICLE - Treg-targeted IL-2/anti-IL-2 complex sustains GVHD prevention and GVL effect

S1. Non-specific binding was blocked using anti-CD16/ CD32 (Miltenyi Biotec). For intracellular cytokine staining, cells were stimulated for 5 hours with phorbol 12-myristate 13-acetate (PMA) 50 ng/mL and ionomycin 1 μg/mL (Sigma– Aldrich, Saint Louis, CA, USA,). Brefeldin A (BD Pharmingen, San Diego, CA, USA) was added for the last 4 hours. Then, cells were fixed and permeabilized with fixation/permeabilization solution (ThermoFischer), following the manufacturer’s instructions. Data acquired with a Canto 2 flow cytometer were compensated and exported into FlowJo (version 10.0.8, TreeStar Inc). Tumor relapse models Irradiated recipient B6D2F1 female mice received retro-orbital infusion of P815-GFP mastocytoma cells (2x104 per mouse) +/- bone marrow (BM) and T cells. GVHD symptoms were evaluated three times per week. Blood samples were systematically collected at d14 to identify the presence of tumor cells before tumor apparition at the eye, the site of injection. GFP+ cells were detected by flow cytometry analysis. In addition to blood sampling, presence of an eye tumor was assessed daily. Mice found to have tumors were sacrificed as previously described.27-29 RS4;11-GFP+ acute lymphoblastic cells were used to test the possibility for human cells to support a GVL effect in NSG mice. For this, 3x106 CD3+ cells contained in peripheral blood mononulcear cells (PBMC) from two healthy donor blood samples (Etablissement Francais du Sang, Créteil, France) and 0.5x106 RS4;11-GFP+ were intravenously injected at d0. Mice were treated with Cx25 for 5 days from d2 to d6. At d20 and d30, mice were randomly sacrificed, and BM, spleen, blood and liver were harvested for the exploration of the presence of tumor cells by GFP detection by flow cytometry. Short-term in vivo xenogeneic human lymphocyte activation NSG mice were used to compare the effect of IL-2 and Cx25 on human healthy donor immune cells in vivo; 12x106 CD3+ cells among PBMC from two healthy donor blood samples were injected at d0 in NSG mice. IL-2 and Cx25 were injected at d0, d1 and d2. After 5 days, their spleens were collected, and analyzed by flow cytometry.

Results In allogeneic hematopoietic stem cell transplantation, clinical effects are determined by the timing and duration of Cx25 and Cx122 administration, rather than by the monoclonal antibody used to generate the complexes We performed a series of experiments to evaluate the effects of Cx25 and Cx122 on GVHD development. When administered at the time of BMT, both IL-2Cx reduced clinical manifestations of GVHD, and prolonged survival of grafted mice, compared to untreated mice or mice receiving low

A. Thiolat et al.

doses of IL-2 (Figure 1A). Thus, unexpectedly, Cx25 and Cx122 induced a similar preventive effect on GVHD development when administered at the time of BMT and for a short period. Using the same anti-IL-2 mAb than in our study to generate IL-2Cx, the distinct effector T-cell activation of the Cx122 complex30 or, the pro-Treg effect of Cx2531 were previously documented in other clinical settings. We thus questioned whether the indissociable effects observed in our hands with the two IL-2Cx were due to the particular setting of allo-HSCT. We grafted B6D2F1 skins onto B6 recipient mice and treated them for 4 days, initiating the treatments at time of skin grafting. In untreated mice, skin grafts were rapidly rejected at d10 as well as in IL-2 or in Cx122 treated mice. As expected, skin rejection was delayed only in mice treated with Cx25 (Online Supplementary Figure S1). These results are in accordance with the dissociated clinical effects previously described with Cx25 or Cx122 and suggest that the similar protective effects observed in our experiments (Figure 1A) are exclusively due to the particular setting of allo-HSCT. We then reproduced experiments of BMT but extending the duration of IL-2Cx administration from d0 to d11. It led not only to a loss of their protective effect observed with a short-time treatment, but even to an accentuation of GVHD, and accelerated death (Figure 1B). Prolonged low-dose IL2 administration did not induce any significant effect on GVHD control. Thus, during allo-HSCT, the timing and/or the duration of IL-2Cx administration seem more important in determining the clinical effect than the different nature of the complexes. These results indicate that short-time administration of IL-2Cx at the time of BMT is more effective than low-dose IL-2 and similarly alleviates GVHD independently of the mAb used to generate the complexes. Cx25 and CX122 are supposed to act on different target cells. However, these two complexes induced comparable effects on GVHD when administered preventively. A possibility to assess whether or not their respective effects depend on Treg is to eliminate them in vivo at the same time as the complexes are administered. For this, we took advantage of our recently published data29 showing that the in vivo administration of an anti-CTLA4 mAb at time of BMT efficiently eliminated Treg. We, therefore, tested the effect of the IL-2Cx when Treg are eliminated by an anti-CTLA4 treatment. In this experiment, both IL-2Cx still fully prevented GVHD, whereas anti-CTLA4 worsened it. Interestingly, the protective effect of both IL2Cx was reduced when anti-CTLA4 was co-administered, suggesting at least a partial link between IL-2Cx effects on Treg and GVHD attenuation (Online Supplementary Figure S2). In allogeneic stem cell transplantation, both Cx25 and Cx122 treatments increase the proportion of regulatory T cells Next, we evaluated the biological effects of both IL-2Cx on immune reconstitution. First, we assessed IL-2Cx effects on immune cell engraftment at d12 (Figure 2A), as by this

Haematologica | 109 January 2024

131

ARTICLE - Treg-targeted IL-2/anti-IL-2 complex sustains GVHD prevention and GVL effect

A. Thiolat et al.

Figure 1. Short treatments with Cx25 or CX122 prevent graft-versus-host disease after allogeneic hematopoietic stem cell transplantation. Lethally irradiated B6C3F1 received semi-allogeneic hematopoietic stem cell transplantation (HSCT) (10x106 bone marrow cells plus 2x106 T cells). Weight and graft-versus-host disease (GVHD) signs were evaluated 3 times a week. (A) Mice were untreated (U.T) (N=21) or treated from day (d)0 to d4 after HSCT with 15,000 IU of human interleukin 2 (hIL-2) alone (N=16) or with Cx25 (N=16) or with Cx122 (N=16). Survival curves and clinical grade histograms are the cumulative data of 3 independent experiments. (B) Mice were untreated (N=15) or treated daily from d0 to d4 then from d7 to d11 after HSCT with 15,000 IU of hIL-2 alone (N=15) or with Cx25 (N=16) or with Cx122 (N=16). Survival curves and clinical grade histograms are the cumulative data of 3 independent experiments. Histograms show area under the curve (AUC) of GVHD manifestations evaluated for each mouse for all the duration of the experiment. The AUC were calculated for the GVHD clinical-grade curve for each mouse and are presented as the mean ± standard error of the mean. Kruskal-Wallis tests were performed. *P<0.05, **P<0.01, ***P<0.001. Kaplan-Meier survival curves were compared using the log-rank test. *P<0.05, **P<0.01, ***P<0.001.

time four of 21 (19%) mice had died in the untreated control group, 2 of 16 (12.5%) in the IL2-treated group, 3 of 16 (18%) in the Cx122 group, and none in the Cx25 group. By the time the mice were sacrificed, weight curves and clinical scores already differed according to treatment (Figure 2B). Thus, d12 appears to be a tipping point for GVHD. The frequencies of Treg and Teff of recipient mouse origin were only residual, with only few CD4+Foxp3+, CD4+Foxp3and CD8+Foxp3- cells expressing the H-2Kk recipient-type molecules at their surface being detected in untreated mice (Figure 2C). Administration of Cx122, and to a lesser extent of Cx25, increased the percentage of recipient CD4+ T cells (Foxp3+ or Foxp3-) which, however, remained in very small proportions in the spleen of treated mice. Thus, both complexes induced a slight decrease in the engraftment as evidenced by the persistence of a larger, although minor, proportion of cells of recipient’s origin. Following this observation, we considered that cells collected at d12 were mainly of donor origin (between 95% and 100%) and

consequently attested for the biologic and clinical effects observed with both IL-2Cx. Whereas the absolute numbers of total spleen cells or of T cells were not modified in Cx25 treated mice or slightly decreased for CD8 T cells in Cx122-treated mice compared to untreated mice, we observed that the clinical immunosuppressive effects of both IL-2Cx were compatible with the important decrease in CD8+ T cells percentage in Cx25- and Cx122-treated animals, compared to untreated mice; as well as with the increase of the proportion of CD4+Foxp3+ Treg (Figure 3A, B). Comparable observations were made in the liver of grafted animals, a target organ of GVHD (Online Supplementary Figure S3). It has been previously observed that a small immunosuppressive CD8+Foxp3+ T-cell population can emerge early during GVHD both in experimental models32,33 and in humans.32,34 Here, we observed that the frequency of spleen CD8+Foxp3+ T cells increased under IL-2Cx administration. Finally, we observed that the IL-2Cx induced a dramatic

Haematologica | 109 January 2024

132

ARTICLE - Treg-targeted IL-2/anti-IL-2 complex sustains GVHD prevention and GVL effect

A. Thiolat et al.

Figure 2. Impact of IL-2/anti-IL-2 complexes on early graft-versus-host disease and chimerism at day 12 after hematopoietic stem cell transplantation. (A) Lethally irradiated B6C3F1 received 10x106 semi-allogeneic B6 bone marrow (BM) cells and 2x106 B6 T cells. (B) Weight loss and clinical grade were evaluated at day (d)12 in untreated mice (UT) (N=9) or mice treated from d0 to d4 after hematopoietic stem cell transplantation (HSCT) with Cx25 (N=21) or with Cx122 (N=22). Data cumulative from 3 independent experiments are presented as the mean ± standard error of the mean. Kruskal-Wallis tests were performed. *P<0.05, **P<0.01, ***P<0.001. (C) Mice were sacrificed and splenocytes were collected for chimerism analysis in untreated mice (N=5) or mice treated from d0 to d4 after HSCT with Cx25 (N=8) or with Cx122 (N=8). Kruskal-Wallis tests were performed. *P<0.05, **P<0.01, ***P<0.001.

reduction of the CD8/Treg ratios both in spleen and liver of grafted animals, likely underlying GVHD attenuation (Figure 3; Online Supplementary Figure 3B). Then, we analyzed the expression of CTLA-4, CD25 and TNFR2 markers of Treg function and fitness (Online Supplementary Figure S4) among the remaining CD4+ FOXP3+ Treg. Although similar percentages of Treg expressing CTLA-4 were observed, the mean fluorescence intensity (MFI) of CTLA-4 expression was statistically significantly increased after treatment with both IL-2Cx. For CD25, only mice receiving Cx122 showed statistically significantly increased percentages of CD25+ Treg, with no major changes in its MFI values. Thus, the remaining Treg in IL-2Cx-treated mice displayed a more activated phenotype, compatible with the observed reduced GVHD. However, this observation must be counterbalanced with the observed decrease in

the percentage of Treg expressing TNFR2 in mice treated with the Cx25 (Online Supplementary Figure S4). The combined use of CD44 and CD62L membrane markers allows distinguishing naïve (CD44-CD62L+), memory effector (CD44+CD62L-) and central memory (CD44+CD62L+) T cells35,36 (Online Supplementary Figure S5A). Compared to untreated mice, Cx25 and Cx122 treatments dramatically increased the proportion of CD4+ Foxp3- naïve T cells, whereas CD4+ effector memory T cells were decreased. The proportion of CD8+ effector memory T cells also decreased with Cx122, but not Cx25 administration. No major modification was observed in the central memory T- cell compartment (Figure 4A). In the liver of grafted animals, CD8+ effector memory T cells decreased and CD8+ central memory T cells increased under Cx122 treatment (Online Supplementary Figure S6A).

Haematologica | 109 January 2024

133

ARTICLE - Treg-targeted IL-2/anti-IL-2 complex sustains GVHD prevention and GVL effect

A. Thiolat et al.

Figure 3. IL-2/anti-IL-2 complexes increase donor regulatory T cells after hematopoietic stem cell transplantation. (A) Lethally irradiated B6C3F1 received 10x106 semi-allogeneic B6 bone marrow (BM) cells and 2x106 B6 T cells. On day (d)12, mice were sacrificed and splenocytes were collected for T-cell analysis. Mice were untreated (U.T) (N=14) or treated from d0 to d4 after hematopoietic stem cell transplantation (HSCT) with Cx25 (N=16) or with Cx122 (N=17). Numbers of spleen cells and CD4+ and CD8+ T cells (A) and percentages (B) of CD4+, CD4+Foxp3+, CD8+, CD8+Foxp3+ cells among live splenocytes and CD8+Foxp3-/ CD4+Foxp3+ ratio are shown for each group. Data cumulative from 3 independent experiments are presented as the mean ± standard error of the mean. Kruskal-Wallis tests were performed. *P<0.05, **P<0.01, ***P<0.001.

Haematologica | 109 January 2024

134

ARTICLE - Treg-targeted IL-2/anti-IL-2 complex sustains GVHD prevention and GVL effect

Additionally, we evaluated the effects of complexes on the expression levels of the anergy/exhaustion T-cell markers PD1, T-bet and Eomes (Online Supplementary Figure S5B, C).37,38 Both IL-2Cx induced a decrease in the percentage of splenic CD4+Foxp3- and CD8+Foxp3- PD-1-expressing conventional T cells and of liver CD8+Foxp3- PD-1-expressing conventional T cells, suggesting a global attenuation of T-cell activation, compared to untreated animals (Figure 4B; Online Supplementary Figure S6B). Among the 80% of CD8+Foxp3- T cells that expressed PD-1, more than 70% co-expressed Eomes but not T-bet in untreated mice, attesting a very high fraction of potentially exhausted CD8+ T cells. Interestingly, this fraction of cells dramatically dropped in mice treated with Cx25 (Figure 4C). The percentage of cells with the Eomes+T-bet- phenotype represented 20% of the CD4+PD-1+

A. Thiolat et al.

cells and Cx122 treatment negatively impacted this cell population (Figure 4C). Thus, the effect of the complexes seems associated with a reduction in the percentage of activated Teff and a more marked and statistically significant decrease in the fraction of CD8+ T cells with an exhausted phenotype when administering Cx25. Similar observations were made in the liver of grafted animals (Online Supplementary Figure S5C). IFNγ, TNFα and IL17 are considered as essential players during the cytokine storm associated with GVHD (Online Supplementary Figure S7).6 We did not observe any major effects of both IL-2Cx treatments on cytokine production by Teff, except for IFNγ production by CD8+ cells, with a statistically significant decrease only when Cx122 was used. The production of the three tested cytokines by Treg statistically

Figure 4. IL-2/anti-IL-2 complexes modify the distribution of naïve/memory populations and the activation state of T cells after hematopoietic stem cell transplantation. B6C3F1 mice were lethally irradiated and grafted as for Figures 2 and 3 and were either untreated (U.T) (N=8) or treated from day (d)0 to d4 after hematopoietic stem cell transplantation (HSCT) with Cx25 (N=7) or with Cx122 (N=8). On d12, mice were sacrificed and memory, effector and naïve T cells were analyzed from splenocytes of grafted animals by flow cytometry. (A) Percentage of effector memory (defined as CD44+ CD62L- cells), central memory (defined as CD44+ CD62L+ cells), and naïve (defined as CD44- CD62L+ cells) cells, among CD4+Foxp3- cells, CD4+Foxp3+ cells and CD8+Foxp3cells are shown for each group. (B) PD-1 among CD4+ and CD8+ T cells according to Foxp3 expression and (C) Eomes and T-bet among CD4+Foxp3-PD-1+ and CD8+Foxp3-PD1+ T cells were also analyzed by flow cytometry. Shown are the cumulative data of 2 independent experiments. Untreated mice (N=5), mice treated from d0 to d4 after HSCT with Cx25 (N=8) or with Cx122 (N=8). Data are presented as the mean ± standard error of the mean. Kruskal-Wallis tests were performed. *P<0.05, **P<0.01, ***P<0.001.

Haematologica | 109 January 2024

135

ARTICLE - Treg-targeted IL-2/anti-IL-2 complex sustains GVHD prevention and GVL effect

significantly decreased in Cx25 and Cx122 treated animals, compared to control group (Figure 5). Thus, as soon as d12, the inflammatory status of Treg was dramatically modified when complexes were administered in grafted mice. Cx25 but not Cx122 prevents graft-versus-host disease while inducing a graft-versus-leukemia/tumor effect After observing that both complexes inhibit GVHD through different mechanisms of action, we evaluated their ability to induce a GVL effect. We used our recently published

A. Thiolat et al.

model in which BM cells (5x106 per mouse) are engrafted together with a defined number of tumor cells (2x104 P815-GFP cells per mouse) and T cells (1x106 CD3+ T cells per mouse) into recipient mice, allowing tumor development without inducing GVHD. In this model, P815 tumors develop in 100% of mice, inexorably leading to 100% of tumor-related mortality by d22, without the development of measurable clinical signs of GVHD (Figure 6A, B). We did not observe any differences in survival, tumor incidence or GVHD clinical grade in mice treated with Cx122, when

Figure 5. IL-2/anti-IL-2 complexes inhibit pro-inflammatory cytokines production by regulatory T cells after hematopoietic stem cell transplantation. B6C3F1 mice were lethally irradiated and grafted as described above and were either untreated (U.T) (N=14) or treated from day (d)0 to d4 after hematopoietic stem cell transplantation (HSCT) with Cx25 (N=16) or with Cx122 (N=17). On d12, mice were sacrificed and splenocytes were collected and then stimulated with PMA/ionomycin and golgi plug for 5 hours before analysis by flow cytometry. TNF-α, IFN-γ, and IL-17 expression in CD4+Foxp3-, CD4⁺Foxp3+ and CD8+Foxp3- T cells are shown for each group of mice. Data are presented as the mean ± standard error of the mean. Kruskal-Wallis tests were performed. *P<0.05, **P<0.01, ***P<0.001.

Haematologica | 109 January 2024

136

ARTICLE - Treg-targeted IL-2/anti-IL-2 complex sustains GVHD prevention and GVL effect

compared to untreated mice. In marked contrast, Cx25 administration led to the protection of half of the grafted mice from leukemia-induced death. No tumor cells were detected before d20 and, over the duration of the experiment (d60), P815-GFP cells were detected only in 55% of the Cx25-treated mice. Interestingly, the anti-tumor effect in Cx25-treated mice was not associated to increased GVHD clinical score (Figure 6B). Thus, only Cx25, but not Cx122, induces a protective GVL effect in mice without inducing GVHD. We recently showed in this same experimental model that

A. Thiolat et al.

anti-CTLA4, but not anti-PD-1 administration, induces a strong GVL effect, but is associated to lethal GVHD, rendering this therapeutic approach not compatible with a clinical use.29 Thus, given that both IL-2Cx efficiently reduce GVHD, we hypothesized that their combination with anti-CTLA4 would represent an ideal situation associating the IL-2Cx GVHD-protective effect with the potent anti-tumor effect of anti-CTLA4 mAb. Combination of anti-CTLA4 with Cx122, did not improve mice survival, tumor incidence or GVHD clinical signs, compared to either IL-2Cx alone. However, the GVL/GVT effect as well as the improved sur-

Figure 6. Only Cx25 but not Cx122 treatment triggers a graft-versus-leukemia/tumor effect after allogeneic hematopoietic stem cell transplantation but that is reverted when combined with an anti-CTLA4 treatment. (A) Lethally irradiated B6D2F1 recipient mice received 5x106 bone marrow cells, 1x106 T cells and 2x104 P815-GFP tumor cells. Recipient mice were divided into 5 groups: untreated mice (U.T) (N=10), mice treated with 5 injections from day (d)0 to d4 either with Cx25 (N=9) or Cx122 (N=10) alone or combined with 3 injections at d0, d3, and d6 with an anti-CTLA-4 monoclonal antibody (mAb) (N=8 with Cx25) or (N=12 with Cx122). (B) Survival curves, tumor incidence curves and areas under the curve (AUC) are presented for all experimental groups of mice. Shown are the cumulative data of 2 independent experiments. Kaplan-Meier survival curves were compared using the logrank test. *P<0.05, **P<0.01, ***P<0.001. The AUC were calculated for the graft-versus-host disease (GVHD) clinical-grade curve for each mouse and are presented as the mean ± standard error of the mean. To compare AUC, Mann-Whitney tests were performed. *P<0.05, **P<0.01, **P<0.001.

Haematologica | 109 January 2024

137

ARTICLE - Treg-targeted IL-2/anti-IL-2 complex sustains GVHD prevention and GVL effect

vival obtained with Cx25 were lost when combined with anti-CTLA4 despite an increased alloreactivity, as attested by the statistically significant augmentation of the clinical grade compared to Cx25 alone (Figure 6B). This suggested that the GVL/GVT effect observed with Cx25 was not due to the sole alloreactivity, but rather to a not yet identified mechanism that could partially depend on CTLA4. Preclinical considerations before Cx25 therapeutic utilization Given that Cx25 was the sole IL-2Cx enabling GVHD control concomitant with GVL triggering, we assessed its effects on human immune cells in vitro and in a newly designed humanized mice model of leukemia. Of note, and for translational rational, this Cx25 is made of human IL-2 and anti-human IL-2 mAb. First, we co-cultured in vitro CD2+-sorted human cells containing T and NK cells with allogeneic APC and evaluated the effects of IL-2 and Cx25 on cell proliferation (Online Supplementary Figure S8). Addition of IL-2 induced a significant proliferation of CD8+ and NK cells, but not of Treg. At odds, Cx25 selectively increased the percentage of Treg. Next, we evaluated the effect of Cx25 on the initial steps of human T-cell activation in vivo, 5 days after the adoptive transfer of human PBMC containing 12x106 T cells into immunodeficient NSG mice (Figure 7A). At this time point, IL-2 or Cx25 administration did not modify the percentage of human colonizing immune cells (evaluated as percentage of human CD45+ cells), in accordance with the similar levels of Ki67+ proliferating CD4+ and CD8+ T cells (Figure 7B). Of note, the percentage of Treg strongly increased in mice treated with Cx25 compared to untreated or IL-2-treated mice. These results suggest that the protective effect of Cx25 evidenced by a preferential expansion of Treg is triggered very early after the adoptive transfer of human cells. Even if the xenograft GVHD model, due to the high xenograft reactivity and the paucity of human cytokines and immune cell growth factors, only partially reflects the physiopathology of human GVHD, it is the only model allowing the in vivo evaluation of therapeutic molecules on human cells before clinical evaluation. This model is based on the administration in immunodeficient NSG mice of a number of human tumor cells mimicking a leukemia relapse and a suboptimal number of human T cells not allowing the development of an anti-leukemic effect. In order to evaluate the GVL effect of the Cx25, we infused NSG mice with the 0.5x106 RS4-GFP lymphoblast T cells, together with human PBMC containing 3x106 T cells, and treated the mice with PBS or Cx25 for 5 consecutive days, starting on d2 after cells injection (Figure 7C). When untreated mice were screened between d20 and d30 (a time line compatible with a leukemia development), RS4-GFP cells were detected in six of eight untreated mice. In marked contrast, no GFP-cells were detected in Cx25-treated mice, attesting of a potent GVL effect.

A. Thiolat et al.

Discussion In this work, we comparatively analyzed the effects of IL2Cx made with two different anti-IL-2 mAb with IL-2, with the aim to modulate the alloreactivity after allo-HSCT. Unexpectedly, both complexes, when administered early and for a short period, attenuated GVHD, resulting in prolonged mice survival in the absence of clinical signs of GVHD (Figure 1A). At odds with our data, the group of Bruce Blazar observed an acceleration of acute GVHD in Cx25-treated mice, and even an inhibition of the protective effect of transferred therapeutic Treg.24 These discordant results could be explained by the nature of the complexes (murine in their article, human in our work), but more probably by the level of histo-incompatibility between donor and recipient mice (fully allogeneic in Blazar et al., semi-allogeneic in our work). In order to confirm this hypothesis, we evaluated the effect of preventive treatment of Cx25 and Cx122 in a second B6 in BALB/c fully allogeneic model. No protective effect was observed (Online Supplementary Figure S9A). Along these lines, they did observe the efficacy of the Cx25 complex, when used in the B6 into B10. BR chronic GVHD model. In our model, administration of IL-2Cx over a prolonged period of time, not only leads to a loss of their protective effects, but even accentuates GVHD development (Figure 1B). Because GVHD is a highly inflammatory pathology, it is likely that both complexes indiscriminately, activate Teff more than Treg when administered for a prolonged period, thus reflecting their loss of specificity of action. Indeed, we observed a strong increase in CD25 expression on a very high proportion of conventional T cells at d4 after transplantation, compared to those of donor T cells before their injection (Online Supplementary Figure S9B). In accordance, IL-2Cx treatments initiated at d5, when an elevated ratio of conventional T cells already strongly express CD25, has no protective effect on the occurrence of GVHD (Figure 9C). We, therefore, focused the rest of this study on short treatments in order to understand the mechanisms of action of the two IL-2Cx in HSCT and their potential clinical relevance. The protective properties of both complexes were mainly associated with comparable biological effects on donor T cells (Figure 2B), inducing an efficient control of GVHD as the result of IL-2Cx effect at early time of BMT. In order to evaluate early events that complexes could have on Treg, we performed KI67 staining of donor T cells collected in grafted mice at d4. First, we did not observe major difference on KI67 staining on Treg depending on whether IL-2Cx were administered or not. This suggested that at d4, treatments did not significantly influence the percentage of cycling Treg. However, whereas each mouse was grafted with 2x105 Treg at d0, the number of Treg collected at d4 in IL-2Cx-treated mice tended to increase compared to untreated mice. As the Treg viability were

Haematologica | 109 January 2024

138

ARTICLE - Treg-targeted IL-2/anti-IL-2 complex sustains GVHD prevention and GVL effect

A. Thiolat et al.

Figure 7. Cx25 induces human regulatory T-cell expansion in vivo and triggers a potent graft-versus-leukemia/tumor effect in a model of xenograft-transplantation. (A) NSG mice were receiving human peripheral blood mononuclear cells (PBMC) containing 12x106 CD3+ T cells from 2 healthy volunteer donors, circle or triangle distinguish the 2 donors. IL-2 and Cx25 were injected at day (d)0, d1 and d2 and spleen were harvested at d5 for flow cytometry analysis. (B) The percentage of human CD45+ (huCD45+) cells, of CD4+ and CD8+ among huCD45+, in the spleen of grafted animals, the percentage of CD4 and CD8 proliferating T cells attested by Ki67 staining as well as the proportion of regulatory T cells (Treg) (CD25+Foxp3+) among CD4+ T cells are shown. Dot blot sum the data with each dot representing 1 mouse. Kruskal-Wallis tests were performed. *P<0.05, ***P<0.001. (C) NSG mice were receiving human PBMC containing 3x106 CD3+ T cells from 2 healthy volunteer donors, circle or triangle distinguish the 2 donors; 0,5x106 RS4-GFP+ tumor cells were injected in mice at time of human PBMC infusion and were treated or not (U.T) with 5 injection of Cx25 from d2 to d6. Mice were randomly sacrificed at d20 or d30 and spleen, liver, blood and bone marrow were harvested to determine the presence or not of tumor cells by GFP detection. Pie chart represent the presence (grey) or not (white) of RS4-GFP tumor cells.

Haematologica | 109 January 2024

139

ARTICLE - Treg-targeted IL-2/anti-IL-2 complex sustains GVHD prevention and GVL effect

comparable (data not shown), this suggested that IL-2Cx likely induced Treg proliferation (Online Supplementary Figure S9D). It should be noted that CD8+Foxp3+ T cells appear during GVHD development, and increase in mice treated with both IL-2Cx (Figure 3B), suggesting that CD8+ Treg arise with increased alloreactivity, as we and others have recently demonstrated.29,34,39 Furthermore, CD8+ Treg represent an attractive therapeutic target since their frequencies are increased when GVHD is controlled. Along these lines, administration of a murine pro-Treg complex in combination with rapamycin was reported to expand CD8+Foxp3+ Treg cells with protective properties against GVHD.33 Additionally, the controlled alloreactivity achieved by both IL-2Cx could be attributed to a decrease of CD4+ EM T-cell frequencies, an increase of naive T cells, and a global decrease of T-cell activation attested by reduced PD-1 expression by CD4+ and CD8+ T-cell activation. Concerning Treg functions, both IL-2Cx led to higher CTLA4 expression levels, without increasing the already high percentage of CTLA4-expressing Treg (Online Supplementary Figure S2). In mice developing GVHD, Treg atypically produced TNF-α, IFN-γ and IL-17. Along these lines, it has been described that during autoimmune diseases Treg can convert to pathological, dysregulated Treg with reduced or no immunosuppressive properties.40,41 Our results prone for IL-2Cx likely halting this process of Treg dysregulation, thus ensuring the maintenance of their suppressive capacity (Figure 5). IL-2Cx also induced distinct effects on donor T cells. The statistically significant increase in the percentage of CD25+ Treg was only observed with the Cx122. Alternatively, the statistically significant decrease in TNFR2 expression was only observed with the Cx25 (Online Supplementary Figure S4). We have recently described that TNFR2 is essential for the suppressive functions of Treg during allo-HSCT.27 Thus, all these observations taken together, it could be concluded that GVHD is more efficiently controlled by Cx25, compared to Cx122. In Cx25-treated mice at d12, Treg seem less mobilized to exert their suppressive effect which is reflected by their less activated/suppressive phenotype. This is consistent with discrete but noticeable differences observed between the two complexes, with the better probability of survival and the weaker clinical score (Figure 1A) obtained with Cx25 when compared to untreated mice (P<0.001), compared to those obtained with Cx122 (P<0.005). We also studied the CD8+PD-1+EOMES+T-bet- T-cell compartment that identifies an exhausted population of T cells. In GVHD-experiencing mice, it represents nearly 80% of CD8+ T cells. Although Cx122 administration tended to reduce this cell population, only Cx25 induced a statistically significant decrease in the proportion of exhausted CD8+ T cells. This observation may have important implications regarding the ability of donor cells to mediate a GVL/GVT

A. Thiolat et al.

effect. Indeed, when tumor cells were injected with T cells at the time of BMT, a GVT effect was only observed in mice that were concomitantly protected from GVHD by Cx25 administration. Importantly, this GVT effect was not associated with an increased clinical score of GVHD, suggesting that it could develop partly independently of an increased alloreactivity. At odds, tumor cells were consistently detected in all untreated or Cx122 treated mice. Overall, our study is the first, to our knowledge, to assess side by side the therapeutic effects of IL-2Cx on the occurrence of acute GVHD and the compatibility of such treatment with the development of a GVL/GVT effect. We demonstrate that only Cx25 effectively prevents GVHD while ensuring a GVL/GVT effect through a CTLA4-dependent mechanism. Indeed, co-administration of Cx25 and anti-CTLA4 was incompatible with the maintenance of a GVT effect, despite increased in alloreactivity. Our results are in accordance with the recent description of a method using an orthogonal IL-2/IL-2Rβ system targeting Treg that allows to prevent GVHD while maintaining a GVL effect.42 In this work, we made the choice to use human-compatible complexes in order to generate information that could be directly transferred to humans. We first demonstrated that these complexes can be evaluated in murine experimental GVHD. More importantly, this allowed us to directly test the Cx25 candidate complex on human cells, in vitro and in vivo. As observed in the whole murine mice model, Cx25 induced human Treg expansion in vitro at d2 and in vivo at d5 in NSG immunodeficient mice grafted with human PBMC. In the same xenograft-GVHD model, the co-injection of RS4 cells induced leukemia in six of eight mice between d20 and d30, while no leukemic cells were detected when mice were treated with Cx25. Taken together, our results demonstrate the clinical relevance of using complexes to modulate alloreactivity after HSCT. Due to its protective effects against GVHD and its capacity in inducing the GVL effect, our data sustains the use of Cx25 in a first clinical trial in patient with high-risk to develop GVHD. A secondary objective in these patients would be to assess the rate of leukemic relapse. It would be also interesting in this context to compare the effect of such complex with the direct in vivo injection of an anti-IL-2 mAb, especially since two anti-IL-2 pro-Treg human mAb have been described.43,44 In this context, IL-2Cx have been described to be directly formed in vivo, and the efficacy of administration of a murine anti-IL-2 mAb in the prevention of GVHD has recently been demonstrated.25 Disclosures EP is a co-founder and consultant for Egle-TX. All other authors have no conflicts of interest to disclose. Contributions AT, CP, EP and JLC designed the study. AT, CP, PC and CS

Haematologica | 109 January 2024

140

ARTICLE - Treg-targeted IL-2/anti-IL-2 complex sustains GVHD prevention and GVL effect

performed experiments. AT, CP, PC, AM, ML, SM EP and JLC analyzed the data. AT, CP, EP and JLC wrote the first draft of the manuscript and all authors contributed in the final version. Acknowledgments We are grateful to the IMRB for providing access to their animal facility team and the flow cytometry platform team for their help.

A. Thiolat et al.

Funding AM received a PhD grant from the Université Paris-EstCréteil (UPEC). This work was fully supported and funded by a common grant from the French Ministry of Health and The French National Cancer Institute (PRTK-2015) (to JLC). Data-sharing statement Data and protocols are available on request to corresponding author.

References 1. Rosenberg SA. Raising the bar: the curative potential of human cancer immunotherapy. Sci Transl Med. 2012;4(127):127ps8. 2. Buyse M, Squifflet P, Lange BJ, et al. Individual patient data meta-analysis of randomized trials evaluating IL-2 monotherapy as remission maintenance therapy in acute myeloid leukemia. Blood. 2011;117(26):7007-7013. 3. Malek TR. The biology of interleukin-2. Annu Rev Immunol. 2008;26:453-479. 4. Sakaguchi S, Sakaguchi N, Shimizu J, et al. Immunologic tolerance maintained by CD25+ CD4+ regulatory T cells: their common role in controlling autoimmunity, tumor immunity, and transplantation tolerance. Immunol Rev. 2001;182:18-32. 5. Papiernik M, de Moraes ML, Pontoux C, Vasseur F, Pénit C. Regulatory CD4 T cells: expression of IL-2R alpha chain, resistance to clonal deletion and IL-2 dependency. Int Immunol. 1998;10(4):371-378. 6. Ferrara JL, Levine JE, Reddy P, Holler E. Graft-versus-host disease. Lancet. 2009;373(9674):1550-1561. 7. Sykes M, Abraham VS, Harty MW, Pearson DA. IL-2 reduces graft-versus-host disease and preserves a graft-versusleukemia effect by selectively inhibiting CD4+ T cell activity. J Immunol. 1993;150(1):197-205. 8. Sykes M, Romick ML, Sachs DH. Interleukin 2 prevents graftversus-host disease while preserving the graft-versus-leukemia effect of allogeneic T cells. Proc Natl Acad Sci U S A. 1990;87(15):5633-5637. 9. Sykes M, Romick ML, Hoyles KA, Sachs DH. In vivo administration of interleukin 2 plus T cell-depleted syngeneic marrow prevents graft-versus-host disease mortality and permits alloengraftment. J Exp Med. 1990;171(3):645-658. 10. Shin HJ, Baker J, Leveson-Gower DB, Smith AT, Sega EI, Negrin RS. Rapamycin and IL-2 reduce lethal acute graft-versus-host disease associated with increased expansion of donor type CD4+CD25+Foxp3+ regulatory T cells. Blood. 2011;118(8):2342-2350. 11. Pérol L, Martin GH, Maury S, Cohen JL, Piaggio E. Potential limitations of IL-2 administration for the treatment of experimental acute graft-versus-host disease. Immunol Lett. 2014;162(2 Pt B):173-184. 12. Koreth J, Matsuoka K, Kim HT, et al. Interleukin-2 and regulatory T cells in graft-versus-host disease. N Engl J Med. 2011;365(22):2055-2066. 13. Kennedy-Nasser AA, Ku S, Castillo-Caro P, et al. Ultra low-dose IL-2 for GVHD prophylaxis after allogeneic hematopoietic stem cell transplantation mediates expansion of regulatory T cells without diminishing antiviral and antileukemic activity. Clin Cancer Res. 2014;20(8):2215-2225. 14. Ito S, Bollard CM, Carlsten M, et al. Ultra-low dose interleukin-2

promotes immune-modulating function of regulatory T cells and natural killer cells in healthy volunteers. Mol Ther. 2014;22(7):1388-1395. 15. Cohen JL, Trenado A, Vasey D, Klatzmann D, Salomon BL. CD4(+)CD25(+) immunoregulatory T Cells: new therapeutics for graft-versus-host disease. J Exp Med. 2002;196(3):401-406. 16. Trenado A, Charlotte F, Fisson S, et al. Recipient-type specific CD4+CD25+ regulatory T cells favor immune reconstitution and control graft-versus-host disease while maintaining graftversus-leukemia. J Clin Invest. 2003;112(11):1688-1696. 17. Trenado A, Sudres M, Tang Q, et al. Ex vivo-expanded CD4+CD25+ immunoregulatory T cells prevent graft-versushost-disease by inhibiting activation/differentiation of pathogenic T cells. J Immunol. 2006;176(2):1266-1273. 18. Kellner JN, Delemarre EM, Yvon E, et al. Third party, umbilical cord blood derived regulatory T-cells for prevention of graft versus host disease in allogeneic hematopoietic stem cell transplantation: feasibility, safety and immune reconstitution. Oncotarget. 2018;9(86):35611-35622. 19. Brunstein CG, Miller JS, Cao Q, et al. Infusion of ex vivo expanded T regulatory cells in adults transplanted with umbilical cord blood: safety profile and detection kinetics. Blood. 2011;117(3):1061-1070. 20. Boyman O, Kovar M, Rubinstein MP, Surh CD, Sprent J. Selective stimulation of T cell subsets with antibody-cytokine immune complexes. Science. 2006;311(5769):1924-1927. 21. Létourneau S, van Leeuwen EM, Krieg C, et al. IL-2/anti-IL-2 antibody complexes show strong biological activity by avoiding interaction with IL-2 receptor alpha subunit CD25. Proc Natl Acad Sci U S A. 2010;107(5):2171-2176. 22. Webster KE, Walters S, Kohler RE, et al. In vivo expansion of T reg cells with IL-2-mAb complexes: induction of resistance to EAE and long-term acceptance of islet allografts without immunosuppression. J Exp Med. 2009;206(4):751-760. 23. Mahr B, Unger L, Hock K, et al. IL-2/α-IL-2 complex treatment cannot be substituted for the adoptive transfer of regulatory T cells to promote bone marrow engraftment. PLoS One. 2016;11(1):e0146245. 24. McDonald-Hyman C, Flynn R, Panoskaltsis-Mortari A, et al. Therapeutic regulatory T-cell adoptive transfer ameliorates established murine chronic GVHD in a CXCR5-dependent manner. Blood. 2016;128(7):1013-1017. 25. Song Q, Wang X, Wu X, et al. Tolerogenic anti-IL-2 mAb prevents graft-versus-host disease while preserving strong graft-versusleukemia activity. Blood. 2021;137(16):2243-2255. 26. Zhang P, Tey SK, Koyama M, et al. Induced regulatory T cells promote tolerance when stabilized by rapamycin and IL-2 in vivo. J Immunol. 2013;191(10):5291-5303.

Haematologica | 109 January 2024

141

ARTICLE - Treg-targeted IL-2/anti-IL-2 complex sustains GVHD prevention and GVL effect

27. Leclerc M, Naserian S, Pilon C, et al. Control of GVHD by regulatory T cells depends on TNF produced by T cells and TNFR2 expressed by regulatory T cells. Blood. 2016;128(12):1651-1659. 28. Naserian S, Leclerc M, Thiolat A, et al. Simple, Reproducible, and Efficient Clinical Grading System for Murine Models of Acute Graft-versus-Host Disease. Front Immunol. 2018;9:10. 29. Moatti A, Debesset A, Pilon C, et al. TNFR2 blockade of regulatory T cells unleashes an antitumor immune response after hematopoietic stem-cell transplantation. J Immunother Cancer. 2022;10(4):e003508. 30. Caudana P, Núñez NG, De La Rochere P, et al. IL2/Anti-IL2 complex combined with CTLA-4, but not PD-1, blockade rescues antitumor NK cell function by regulatory T-cell modulation. Cancer Immunol Res. 2019;7(3):443-457. 31. Lee E, Kim M, Lee YJ. Selective expansion of Tregs using the IL-2 cytokine antibody complex does not reverse established Alopecia Areata in C3H/HeJ mice. Front Immunol. 2022;13:874778. 32. Beres AJ, Haribhai D, Chadwick AC, Gonyo PJ, Williams CB, Drobyski WR. CD8+ Foxp3+ regulatory T cells are induced during graft-versus-host disease and mitigate disease severity. J Immunol. 2012;189(1):464-474. 33. Robb RJ, Lineburg KE, Kuns RD, et al. Identification and expansion of highly suppressive CD8(+)FoxP3(+) regulatory T cells after experimental allogeneic bone marrow transplantation. Blood. 2012;119(24):5898-5908. 34. Zheng J, Liu Y, Liu M, et al. Human CD8+ regulatory T cells inhibit GVHD and preserve general immunity in humanized mice. Sci Transl Med. 2013;5(168):168ra9. 35. Lee WT, Pelletier WJ. Visualizing memory phenotype development after in vitro stimulation of CD4(+) T cells. Cell

Immunol. 1998;188(1):1-11. 36. Kagamu H, Touhalisky JE, Plautz GE, Krauss JC, Shu S. Isolation based on L-selectin expression of immune effector T cells derived from tumor-draining lymph nodes. Cancer Res. 1996;56(19):4338-4342. 37. Larbi A, Fulop T. From “truly naïve” to “exhausted senescent” T cells: when markers predict functionality. Cytometry A. 2014;85(1):25-35. 38. Blank CU, Haining WN, Held W, et al. Defining ‘T cell exhaustion’. Nat Rev Immunol. 2019;19(11):665-674. 39. Vieyra-Lobato MR, Vela-Ojeda J, Montiel-Cervantes L, LópezSantiago R, Moreno-Lafont MC. Description of CD8. J Immunol Res. 2018;2018:3758713. 40. Xu K, Yang WY, Nanayakkara GK, et al. GATA3, HDAC6, and BCL6 regulate FOXP3+ Treg plasticity and determine Treg conversion into either novel antigen-presenting cell-like Treg or Th1-Treg. Front Immunol. 2018;9:45. 41. Korn T, Reddy J, Gao W, et al. Myelin-specific regulatory T cells accumulate in the CNS but fail to control autoimmune inflammation. Nat Med. 2007;13(4):423-431. 42. Lopes Ramos T, Bolivar-Wagers S, Jin S, et al. Prevention of acute GVHD disease using an orthogonal IL-2/IL-2Rβ system to selectively expand regulatory T-cells in vivo. Blood. 2023;141(11):1337-1352. 43. Trotta E, Bessette PH, Silveria SL, et al. A human anti-IL-2 antibody that potentiates regulatory T cells by a structurebased mechanism. Nat Med. 2018;24(7):1005-1014. 44. Raeber ME, Rosalia RA, Schmid D, Karakus U, Boyman O. Interleukin-2 signals converge in a lymphoid-dendritic cell pathway that promotes anticancer immunity. Sci Transl Med. 2020;12(561):eaba5664.

Haematologica | 109 January 2024

142

A. Thiolat et al.

ARTICLE - Cell Therapy & Immunotherapy

Patient-reported treatment response in chronic graftversus-host disease Annie Im,1 Iskra Pusic,2 Lynn Onstad,3 Carrie L. Kitko,4 Betty K. Hamilton,5 Amin M. Alousi,6 Mary E. Flowers,3 Stefanie Sarantopoulos,7 Paul Carpenter,3 Jennifer White,8 Sally Arai,9 Najla El Jurdi,10 George Chen,11 Corey Cutler,12 Stephanie Lee3 and Joseph Pidala13 1

University of Pittsburgh/UPMC Hillman Cancer Center, Pittsburgh, PA, USA; 2Division of Medicine and Oncology, Washington University, Saint Louis, MO, USA; 3Clinical Research Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA; 4Pediatric Blood and Marrow Transplantation Program, Vanderbilt University Medical Center, Nashville, TN, USA; 5 Blood and Marrow Transplantation, Department of Hematology and Medical Oncology, Taussig Cancer Institute, Cleveland Clinic, Cleveland, OH, USA; 6Department of Stem Cell Transplantation and Cellular Therapy, The University of Texas MD Anderson Cancer Center, Houston, TX, USA; 7Division of Hematological Malignancies and Cellular Therapy, Duke University Department of Medicine, Duke Cancer Institute, Durham, NC, USA; 8Leukemia/Bone Marrow Transplant Program of British Columbia, British Columbia Cancer Agency, Vancouver, British Columbia, Canada; 9Department of Medicine, Division of Blood and Marrow Transplantation, Stanford University School of Medicine, Stanford, CA, USA; 10Division of Hematology, Oncology and Transplantation, University of Minnesota Medical Center, Minneapolis, MN, USA; 11Department of Medicine, Roswell Park Comprehensive Cancer Center, Buffalo, NY, USA; 12Division of Stem Cell Transplantation and Cellular Therapy, Dana-Farber Cancer Institute, Boston, MA, USA and 13Blood and Marrow Transplantation and Cellular Immunotherapy, H. Lee Moffitt Cancer Center and Research Institute, Tampa, FL, USA

Correspondence: J. Pidala joseph.pidala@moffitt.org Received: Accepted: Early view:

January 18, 2023. May 17, 2023. May 25, 2023.

Abstract Chronic graft-versus-host disease (GvHD) treatment response is assessed using National Institutes of Health (NIH) Consensus Criteria in clinical trials, and by clinician assessment in routine practice. Patient-reported treatment response is central to the experience of chronic GvHD manifestations as well as treatment benefit and toxicity, but how they correlate with clinician- or NIH-responses has not been well-studied. We aimed to characterize 6-month patientreported response, determine associated chronic GvHD baseline organ features and changes, and evaluate which patientreported quality of life and chronic GvHD symptom burden measures correlated with patient-reported response. From two nationally representative Chronic GVHD Consortium prospective observational studies, 382 subjects were included in this analysis. Patient and clinician responses were categorized as improved (completely gone, very much better, moderately better, a little better) versus not improved (about the same, a little worse, moderately worse, very much worse). At six months, 270 (71%) patients perceived chronic GvHD improvement, while 112 (29%) perceived no improvement. Patient-reported response had limited correlation with either clinician-reported (kappa 0.37) or NIH chronic GvHD response criteria (kappa 0.18). Notably, patient-reported response at six months was significantly associated with subsequent failure-free survival. In multivariate analysis, NIH responses in eye, mouth, and lung had significant association with 6-month patient-reported response, as well as a change in Short Form 36 general health and role physical domains and Lee Symptom Score skin and eye changes. Based on these findings, patient-reported responses should be considered as an important complementary endpoint in chronic GvHD clinical trials and drug development.

Introduction Chronic graft-versus-host disease (GvHD) is the most common cause of late morbidity and mortality after allogeneic hematopoietic cell transplantation (HCT).1,2 Prior to the National Institutes of Health (NIH) Consensus Conferences, there was limited standardization of chronic GvHD severity scoring and treatment response assess-

ment. In 2005, the NIH Consensus Conference on Clinical Trials for Treatment of Chronic GVHD provided standardized criteria to assess organs involved in chronic GvHD and to determine response to therapies for use in clinical trials;3-5 consensus criteria were updated in 2014 based on data that had accumulated since the original guidelines.6,7 These criteria have elevated the scientific rigor of current clinical trials in chronic GvHD therapy. However,

Haematologica | 109 January 2024

143

ARTICLE - Patient-reported response

A. Im et al.

substantial improvements in affected organs may be required to achieve response by these criteria, and clinician- or patient-reported response metrics may capture lesser but still important degrees of clinical benefit. For example, while NIH response criteria have been correlated with outcomes,8 other studies reported limited correlation between NIH response and clinician-reported response,9,10 and certain quality of life measures.11 In chronic GvHD management outside of clinical trials, clinicians consider patient-reported symptom burden, findings from physical examination, and routine laboratory tests to determine clinical response, rather than strictly applying the NIH Response Criteria. In total, this clinician assessment of improvement, stability, or worsening is reduced to a perceived presence or absence of clinical benefit, and informs immunosuppressive (IS) therapy management. Clinician-reported response has been correlated with survival,8 and subsequent analyses found that changes in serum bilirubin, NIH 0 to 3-point scores of lower gastrointestinal tract, mouth, joint/fascia, lung, and skin were factors that correlated most with clinician-reported responses.9 Thus, while clinician-assessed response is less standardized and rigorous compared to NIH Response Criteria, it appears to be of value. Patients affected by chronic GvHD experience the physical manifestations, symptoms, functional limitations and impairment in quality of life known to be due to chronic GvHD, and can identify their own benefit and toxicity from chronic GvHD therapies. Thus, a patient’s self-report of treatment response may be of critical importance. While patient-reported responses are captured in clinical trials according to the NIH criteria, they are not used to determine overall NIH response. In fact, clinical trials have yet to incorporate patient-reported outcomes into determination of success; however, in chronic GvHD, improvement in quality of life is an important treatment goal.12 It is unknown whether patient-reported responses correlate with clinician or NIH responses, nor do we know the factors that contribute to patient-reported responses. The aims of this study were to characterize 6-month patient-reported treatment responses, including associated chronic GvHD baseline organ features and changes, and to evaluate which patient-reported quality of life and chronic GvHD symptom measures correlate with patient-reported response. Ultimately, we sought to determine whether patient-reported response measures may capture clinical benefit in a different way to standard NIH or clinician-reported response measures in chronic GvHD.

Chronic GVHD Consortium observational studies. The “Chronic GVHD Consortium Improvement Outcomes Assessment in Chronic GVHD” study enrolled 601 patients between 2007 and 2012.13 Patients in this study enrolled at any time after starting systemic treatment for chronic GvHD. At enrollment and every six months thereafter, providers and patients recorded standardized information regarding current chronic GvHD organ involvement and symptoms using forms developed according to the 2005 NIH Chronic GVHD Consensus Criteria. For incident cases, providers and patients also recorded the same information at three months after enrollment. The “Chronic GVHD Consortium Response Measures Validation Study” enrolled 383 patients with chronic GvHD between 2013 and 2017.14 Patients in this study enrolled within four weeks before or after starting a new systemic treatment for chronic GvHD. At enrollment and 3, 6, and 18 months thereafter, providers and patients recorded standardized information according to the 2014 NIH Consensus Conference Criteria. Exclusion criteria in both studies included primary disease relapse and inability to comply with study procedures. At each assessment, providers and patients rated overall changes in GvHD manifestations from enrollment according to an 8-point scale with categories of “completely gone”, “very much better”, “moderately better”, “a little better”, “about the same”, “a little worse”, “moderately worse”, or “a lot worse”. This is in the form of a single question that patients and providers complete. In addition, patients completed the Short Form 36 (SF-36) survey, and rated the severity of GvHD symptoms according to the Lee Chronic GVHD Symptom Scale.15 The protocols were approved by the Institutional Review Board at each site, and all patients provided informed consent in accordance with the Declaration of Helsinki. For the purposes of this analysis, patients from the two parent cohort studies were included based on the availability of patient-reported response data at cohort enrollment and at six months. Responses were categorized as “improved” (completely gone, very much better, moderately better, a little better) versus “not improved” (about the same, a little worse, moderately worse, a lot worse).

Statistical analysis Comparisons of characteristics by patient perception of chronic GvHD improvement versus no improvement were performed using the χ2 test and Fisher’s Exact Test for categorical variables and the Wilcoxon rank sum test for continuous variables. Multivariable logistic regression was used to examine the relationships between patient perception of chronic GvHD improvement, and transplant and chronic GvHD characteristics, as well as 6-month organ Methods responses, using a stepwise procedure with entry and rePatients tention criteria of P≤0.1. The initial set of variables included Patients were enrolled in two prospective, multicenter all those found to be univariately related at the P≤0.1 level. Haematologica | 109 January 2024

144

ARTICLE - Patient-reported response

A. Im et al.

This same procedure was used to examine the relationships between patient perception of chronic GvHD improvement with transplant and patient-reported outcomes (PRO) of the Lee Symptom Scale and the SF36. The log-rank test was used to compare overall survival (OS) and failure-free survival (FFS) (composite outcome including death, malignancy relapse, and start of new line of systemic IS therapy) by patient perception of chronic GvHD improvement. OS was calculated from the 6-month visit until death; FFS was calculated from the 6-month visit until malignancy relapse, death, or addition of a new systemic IS medication for chronic GvHD among those who had not started a new IS medication for chronic GvHD before six months. Analyses were performed in SAS 9.4 (SAS Institute, Cary NC, USA).

Results From the overall parent cohort (N=605), this study population was limited to 382 patients with baseline and 6month patient-reported response data. The 223 excluded for missing 6-month patient response were predominantly due to missed visit or missed patient survey (N=166, 74.4%), with a lesser contribution from other reasons (withdrew from study before 6-month visit N=9, 4%; relapsed before six months N=23, 10.3%; or died before six months N=25, 11.2%). A comparison of those with versus without 6-month response data is presented in the Online Supplementary Appendix. Table 1 summarizes baseline patient and chronic GvHD characteristics (see also the Online Supplementary Appendix). The median time from chronic GvHD to enrollment was 0.5 months (Interquartile Range [IQR] 0-8.6 months). Most patients had moderate (50%) or severe (38%) chronic GvHD, and the most involved organ sites were skin (73%), mouth (60%), and eyes (54%). Thirty-four percent of the cases were prevalent, while 66% were incident cases. Chronic GvHD features are presented in Table 1. At six months, 71% of patients reported improvement in their chronic GvHD, while 29% reported no improvement. Among those categorized as improved (N=270), patient response included: completely gone (N=31, 11.5%), very much better (N=119, 44.1%), moderately better (N=68, 25.2%), and a little better (N=52, 19.3%). For the not improved group (N=112), patient responses were: about the same (N=46, 41.1%), a little worse (N=37, 33%), moderately worse (N=18, 16.1%), and very much worse (N=11, 9.8%). Notably, there was limited correlation with clinician-reported response (kappa 0.37) and NIH response (kappa 0.18). Among clinicians, 66% reported improvement in chronic GvHD, while 34% reported no improvement. Per NIH response, 46% of patients had improvement in chronic GvHD while 54% had

no improvement. Clinician-reported and NIH responses grouped per patient-reported response are presented in the Online Supplementary Appendix. In univariate analysis, enrollment lung involvement was associated with patients’ perception of response at six months (P<0.001). NIH response in skin (P=0.003), eye (P<0.001), mouth (P=0.004), and lung (P<0.001) at six months was also associated with patient-reported response at six months (Online Supplementary Appendix). In multivariable analysis of patient-reported response, moderate-severe lung involvement at enrollment (P=0.007) was associated with report of no improvement, and NIH responses in eye (P=0.009), mouth (P=0.01), and lung (P=0.03) were associated with patient report of improvement at six months (Table 2). In multivariable analysis of patient perception of chronic GvHD improvement with quality of life and symptom burden measures, a higher SF-36 general health score at enrollment (P<0.001) (indicating better quality of life) and improvement from enrollment to six months in the SF-36 general health score (P<0.001) were associated with patient report of improvement, while worsening in SF-36 role physical domain score was associated with patient report of no improvement (P=0.04) (Table 3). A lower LSS skin score at enrollment (P=0.1), indicating lower burden of symptoms, and improvement from enrollment to six months in the LSS eye score (P=0.007) were associated with patient report of improvement, while worsening of LSS skin score (P<0.001) was associated with patient report of no improvement (Table 3). Failure-free survival after six months was higher among patients who reported improvement compared to those who reported no improvement (P=0.005) (Figure 1). The most common cause of failure was starting new treatment for chronic GvHD (Table 4). The cumulative incidence of new systemic IS therapy stratified by 6-month patient-reported response is included in the Online Supplementary Appendix. In a multivariable analysis of the association of 6-month response with subsequent time to new systemic IS therapy, we found no significant difference in the Hazard Ratios for patient, clinician, or NIH response (data not shown). There was no difference in overall survival at six months from enrollment between the two groups (Figure 2).

Discussion This study provides new insight into patient-reported treatment response in chronic GvHD, a previously understudied area. Six-month patient-reported response was associated with subsequent FFS, and may represent a clinically meaningful measure. Moreover, given the limited correlation with NIH and clinician-assessed response, it

Haematologica | 109 January 2024

145

ARTICLE - Patient-reported response

A. Im et al.

Table 1. Baseline chronic graft-versus-host disease characteristics by patient perception.

Variables

Skin*

Fascia*

Liver

Lung

Eye

Mouth

Joint

Genital

Categories

Total N=382

GvHD not improved N=112

GvHD improved N=270

0 1 2 3 0 1 2 3 0 1 2 3 0 1 2 3 0 1 2 3 0 1 2 3 0 1 2 3 0 1 2 3 0 1 2 3

103 (27.0%) 66 (17.3%) 109 (28.5%) 104 (27.2%) 266 (69.6%) 92 (24.1%) 20 (5.2%) 4 (1.0%) 259 (68.0%) 84 (22.0%) 35 (9.2%) 3 (0.8%) 294 (78.8%) 27 (7.2%) 43 (11.5%) 9 (2.4%) 274 (71.7%) 82 (21.5%) 21 (5.5%) 5 (1.3%) 176 (46.2%) 130 (34.1%) 69 (18.1%) 6 (1.6%) 152 (39.8%) 173 (45.3%) 52 (13.6%) 5 (1.3%) 236 (61.9%) 91 (23.9%) 47 (12.3%) 7 (1.8%) 228 (83.8%) 31 (11.4%) 8 (2.9%) 5 (1.8%)

31 (27.7%) 18 (16.1%) 29 (25.9%) 34 (30.4%) 75 (67.0%) 30 (26.8%) 7 (6.3%) 0 (0.0%) 77 (68.8%) 24 (21.4%) 11 (9.8%) 0 (0.0%) 88 (81.5%) 11 (10.2%) 9 (8.3%) 0 (0.0%) 69 (61.6%) 26 (23.2%) 14 (12.5%) 3 (2.7%) 45 (40.5%) 37 (33.3%) 26 (23.4%) 3 (2.7%) 50 (44.6%) 51 (45.5%) 11 (9.8%) 0 (0.0%) 68 (60.7%) 24 (21.4%) 18 (16.1%) 2 (1.8%) 59 (76.6%) 13 (16.9%) 3 (3.9%) 2 (2.6%)

72 (26.7%) 48 (17.8%) 80 (29.6%) 70 (25.9%) 191 (70.7%) 62 (23.0%) 13 (4.8%) 4 (1.5%) 182 (67.7%) 60 (22.3%) 24 (8.9%) 3 (1.1%) 206 (77.7%) 16 (6.0%) 34 (12.8%) 9 (3.4%) 205 (75.9%) 56 (20.7%) 7 (2.6%) 2 (0.7%) 131 (48.5%) 93 (34.4%) 43 (15.9%) 3 (1.1%) 102 (37.8%) 122 (45.2%) 41 (15.2%) 5 (1.9%) 168 (62.5%) 67 (24.9%) 29 (10.8%) 5 (1.9%) 169 (86.7%) 18 (9.2%) 5 (2.6%) 3 (1.5%)

0.77

0.52

0.86

0.07

<0.001

0.16

0.22

0.53

0.19

GvHD: graft-versus-host disease; GI: gastrointestinal. *Skin sclerosis and fascia involvement presented in the Online Supplementary Appendix.

appears patient-reported response may capture unique aspects of clinical benefit. We also identified both baseline chronic GvHD organ involvement and organ responses that had association with patient-reported response, and determined which PRO measure changes had greatest association with patient-reported response. Taken together, the findings from this study advance a potential additional approach for assessing clinical benefit in chronic GvHD therapy, and lay a foundation for future research in this area.

We determined that baseline lung and liver involvement, and NIH responses in eyes, mouth, and lungs had the greatest association with patient-reported responses in this cohort. Patient-reported response appeared to be most associated with organs where responses can be profound in their impact on symptom burden or functionality. These results may have been influenced by the baseline frequency and reversibility of specific chronic GvHD organ manifestations, scale definitions in calculated NIH response, as well as patient bias focused on more recent

Haematologica | 109 January 2024

146

ARTICLE - Patient-reported response

A. Im et al.

Table 2. Multivariate analysis of patient characteristics, chronic graft-versus-host disease characteristics, National Institutes of Health organ responses by patient perception.

Variable* Liver (at enrollment)

Lung (at enrollment)

NIH eye response

NIH mouth response

NIH lung response

Class Not involved Mild Moderate/severe Not involved Mild Moderate/severe Not involved/stable CR PR Prog Not involved/stable CR PR Prog Not involved/stable CR PR Prog

OR (95% CI) Reference 0.3 (0.1-0.9) 1.5 (0.6-3.7) Reference 0.7 (0.3-1.6) 0.1 (0.0-0.6) Reference 1.7 (0.6-4.7) 0.7 (0.2-2.2) 0.3 (0.1-0.7) Reference 2.3 (1.1-4.8) 3.8 (1.6-9.1) 1.4 (0.5-4.0) Reference 2.7 (0.8-9.1) 2.2 (0.3-15.9) 0.4 (0.1-1.0)

P 0.04 0.41 0.46 0.007 0.32 0.5 0.003 0.03 0.003 0.51 0.1 0.43 0.05

Global P-value 0.06 0.02 0.009 0.01 0.03 -

OR: Odd's Ratio; CI. Confidence Interval; NIH: National Institutes of Health; CR: complete response; PR: partial response; Prog: progression. *Model adjusted for disease diagnosis and parent cohort study.

Table 3. Multivariate analysis of patient-reported outcome by patient perception.

Variable* SF-36 general health (enrollment) LSS skin (enrollment) SF-36 general health (change score) SF-36 role-physical (change score) LSS skin (change score) LSS eye (change score)

OR (95% CI)** 2.7 (1.9-3.8) 0.9 (0.7-1.0) 3.3 (2.0-5.3) 1.4 (1.0-1.9) 0.7 (0.6-0.8) 0.8 (0.7-1.0)

P <0.001 0.1 <0.001 0.04 <0.001 0.007

OR: Odd's Ratio; CI: Confidence Interval; SF-36: Short-Form 36; LSS: Lee Symptom Scale. *Model adjusted for parent cohort study. **OR presented for each variable represents OR per 10-point change.

changes versus change from baseline, as previously reviewed.9 The organ site-specific LSS change measures most associated with patient-reported response demonstrated some, but not complete, agreement with these same organ-specific NIH response findings. Finally, we acknowledge that the association of 6-month patient-reported response with subsequent FFS was primarily driven by changes in systemic IS treatment, and that this did not impact OS. The dominance of treatment change in the composite FFS outcome was well-established in prior studies.16,17 In chronic GvHD therapy, it represents a significant failure where additional lines of therapy lead to additional risk of infectious complications, treatment toxicity, medication costs, and associated healthcare costs. Additional investigation is needed in larger patient populations, both for overall validation of this work, and to

further refine conclusions regarding specific organ manifestations. Following validation, patient-reported response could be analyzed in clinical trials, potentially as a secondary outcome measure to complement NIH responses. This would allow for a formal determination of benefit that could lead to approvals for treatments in the situations where NIH responses do not capture the full clinical benefit in the setting of a clinical trial. Patient-reported response could also be formally captured in routine clinical care through use of a simple patient-reported ordinal response scale to provide structured assessment above and beyond information currently shared by patients regarding their perceived treatment benefit. While inviting patients’ impressions of benefit of treatments is routine in clinical practice, this structured assessment may provide insight into clinical benefit not captured through usual response measures.

Haematologica | 109 January 2024

147

ARTICLE - Patient-reported response

A. Im et al.

This analysis has some limitations. First, this is an unplanned retrospective analysis of existing data from two parent national cohort studies. While these parent studies were rigorously designed, this analysis was facilitated using only those subjects with available baseline and 6-

month patient-reported response data. Second, we note heterogeneity in the included subjects, both regarding differing parent cohort enrollment criteria, and other factors, most notably variation in the number of lines of systemic therapy and the actual agents used to treat the chronic

Figure 1. Failure-free survival and patient-reported response. 5-year failure-free survival by patient perception of chronic graftversus-host disease at six months among those failure free at six months.

Figure 2. Overall survival and patient-reported response. Haematologica | 109 January 2024

148

ARTICLE - Patient-reported response

A. Im et al.

Table 4. Causes of failure in failure-free survival.

At 1 year GvHD

At 2 years

At 3 years

At 5 years

Not improved

Improved

Not improved

Improved

Not improved

Improved

Not improved

Improved

New medication

42.4%

26.5%

54.2%

39.4%

65.1%

43.6%

65.1%

51.2%

Relapse

4.3%

5.6%

6.0%

6.7%

6.0%

7.8%

6.0%

9.8%

Non-relapse mortality

0.0%

1.6%

3.5%

2.8%

3.5%

4.6%

3.5%

7.8%

GvHD: graft-versus-host disease.

GvHD after cohort. Third, in the study of association between NIH responses and patient-reported response, we acknowledge certain organ sites were better represented, and that these may differ in treatment responsiveness overall. Given this, we were unable to conduct more detailed analyses according to each type of organ involvement per affected organ site. Patient-reported responses are an important response measure that is associated with FFS in chronic GvHD and should be considered as a complementary response outcome for clinical trials and drug development in chronic GvHD. Disclosures IM reports research funding from Incyte, and consultancy for Abbvie and CTI Biopharma. IP sits on the advisory board for Incyte. CLK sits on the advisory board for Horizon Therapeutics. BKH sits on the advisory board for NKARTA, Equilium, Incyte, Kadmon, and Syndax, and has received speaker fees from Mallinkrodt. AMA is a consultant for Prolacta and Genentech, has received research support from Incyte, and has been a speaker for Johnson and Johnson. MEF has received research support from Pharmacyclics Inc. and Novartis/Incyte Corp., has been a Speaker Honorarium for Janssen, Pharmaceutical Companies of Johnson & Johnson, Astellas, Mallinckrodt Pharmaceutics, and is a consultant for Pharmacyclics Inc., CSL Behring, and Fresenius Kabi. SS has received drug support, sits on the advisory board, and offers consultancy services to Rigel. PC has been involved

in research finding/consulting for AbbVie, for Incyte, and Seres. SA is a Member on the Board of Directors or advisory committees for Kadmon. CC has received consultancy/advisory honararia from Sanofi, Equillium, CTI Biopharma, Mallinckrodt, Omeros, CSL Behring, Incyte, InhibRx, Cellarity Scientific Advisory Board with Equity: Cimeio, Oxford Algorithmics. SL has offered consultancy services, and has received honoraria, and research funding from Kadmon, research funding from Pfizer, Syndax, Amgen, Incyte, and Novartis, and consultancy services and honoraria from Equillium, Consultancy from Mallinckrodt, and honoraria from AstraZeneca, sits on the Steering Committee of Novartis, and is a member of the Board of Directors or advisory committees of the National Marrow Donor Program. JP reports consulting and advisory board membership for Syndax, CTI Biopharma, Amgen, Regeneron, and Incyte, and clinical trial support from Novartis, Amgen, Takeda, Janssen, Johnson & Johnson, Pharmacyclics, Abbvie, CTI Biopharma, and BMS. LO, JW, NEJ and GC have no conflicts of interest to disclose. Contributions AI, IP, LO, SJL and JP designed and performed the research, analyzed data, and wrote the paper. All other authors contributed to the data analysis and writing of the paper. Data-sharing statement For original data or further information, contact the corresponding author.

References 1. Lee SJ, Vogelsang G, Flowers ME. Chronic graft-versus-host disease. Biol Blood Marrow Transplant. 2003;9(4):215-233. 2. Arai S, Arora M, Wang T, et al. Increasing incidence of chronic graft-versus-host disease in allogeneic transplantation: a report from the Center for International Blood and Marrow Transplant Research. Biol Blood Marrow Transplant. 2015;21(2):266-274. 3. Martin PJ, Weisdorf D, Przepiorka D, et al. National Institutes of Health Consensus Development Project on Criteria for Clinical Trials in Chronic Graft-versus-Host Disease: VI. Design of Clinical Trials Working Group report. Biol Blood Marrow Transplant. 2006;12(5):491-505.

4. Filipovich AH, Weisdorf D, Pavletic S, et al. National Institutes of Health consensus development project on criteria for clinical trials in chronic graft-versus-host disease: I. Diagnosis and staging working group report. Biol Blood Marrow Transplant. 2005;11(12):945-956. 5. Pavletic SZ, Martin P, Lee SJ, et al. Measuring therapeutic response in chronic graft-versus-host disease: National Institutes of Health Consensus Development Project on Criteria for Clinical Trials in Chronic Graft-versus-Host Disease: IV. Response Criteria Working Group report. Biol Blood Marrow Transplant. 2006;12(3):252-266.

Haematologica | 109 January 2024

149

ARTICLE - Patient-reported response

A. Im et al.

6. Lee SJ, Wolff D, Kitko C, et al. Measuring therapeutic response in chronic graft-versus-host disease. National Institutes of Health consensus development project on criteria for clinical trials in chronic graft-versus-host disease: IV. The 2014 Response Criteria Working Group report. Biol Blood Marrow Transplant. 2015;21(6):984-999. 7. Jagasia MH, Greinix HT, Arora M, et al. National Institutes of Health Consensus Development Project on Criteria for Clinical Trials in Chronic Graft-versus-Host Disease: I. The 2014 Diagnosis and Staging Working Group report. Biol Blood Marrow Transplant. 2015;21(3):389-401. 8. Palmer J, Chai X, Pidala J, et al. Predictors of survival, nonrelapse mortality, and failure-free survival in patients treated for chronic graft-versus-host disease. Blood. 2016;127(1):160-166. 9. Martin PJ, Storer BE, Palmer J, et al. Organ changes associated with provider-assessed responses in patients with chronic graft-versus-host disease. Biol Blood Marrow Transplant. 2019;25(9):1869-1874. 10. Palmer JM, Lee SJ, Chai X, et al. Poor agreement between clinician response ratings and calculated response measures in patients with chronic graft-versus-host disease. Biol Blood Marrow Transplant. 2012;18(11):1649-1655.

11. Inamoto Y, Martin PJ, Chai X, et al. Clinical benefit of response in chronic graft-versus-host disease. Biol Blood Marrow Transplant. 2012;18(10):1517-1524. 12. Pidala J, Kurland B, Chai X, et al. Patient-reported quality of life is associated with severity of chronic graft-versus-host disease as measured by NIH criteria: report on baseline data from the Chronic GVHD Consortium. Blood. 2011;117(17):4651-4657. 13. Chronic GC. Rationale and design of the chronic GVHD cohort study: improving outcomes assessment in chronic GVHD. Biol Blood Marrow Transplant. 2011;17(8):1114-1120. 14. Lee SJ, Hamilton BK, Pidala J, et al; The Chronic GVHD Consortium. Design and patient characteristics of the Chronic Graft-versus-Host Disease Response Measures Validation Study. Biol Blood Marrow Transplant. 2018;24(8):1727-1732. 15. Lee S, Cook EF, Soiffer R, Antin JH. Development and validation of a scale to measure symptoms of chronic graft-versus-host disease. Biol Blood Marrow Transplant. 2002;8(8):444-452. 16. Inamoto Y, Flowers ME, Sandmaier BM, et al. Failure-free survival after initial systemic treatment of chronic graft-versushost disease. Blood. 2014;124(8):1363-1371. 17. Inamoto Y, Storer BE, Lee SJ, et al. Failure-free survival after second-line systemic treatment of chronic graft-versus-host disease. Blood. 2013;121(12):2340-2346.

Haematologica | 109 January 2024

150

ARTICLE - Chronic Lymphocytic Leukemia

Electron transport chain and mTOR inhibition synergistically decrease CD40 signaling and counteract venetoclax resistance in chronic lymphocytic leukemia Zhenghao Chen,1-5 Gaspard Cretenet,1-5 Valeria Carnazzo,1,6 Helga Simon-Molas,1-5 Arnon P. Kater,2-5 Gerritje J. W. van der Windt7 and Eric Eldering1,3-5 1

Experimental Immunology, Amsterdam UMC location University of Amsterdam, Amsterdam, the Netherlands; 2Hematology, Amsterdam UMC location University of Amsterdam, Amsterdam, the Netherlands; 3Amsterdam Institute for Infection and Immunity, Cancer Immunology, Amsterdam, the Netherlands; 4Cancer Center Amsterdam, Cancer Immunology, Amsterdam, the Netherlands; 5Lymphoma and Myeloma Center, Amsterdam, the Netherlands; 6Department of Clinical Pathology, S.M. Goretti Hospital, Latina, Italy and 7Genmab, Utrecht, the Netherlands.

Correspondence: E. Eldering e.eldering@amsterdamumc.nl Received: Accepted: Early view:

January 16, 2023. July 5, 2023. July 13, 2023.

Abstract CD40 signaling upregulates BCL-XL and MCL-1 expression in the chronic lymphocytic leukemia (CLL) lymph node microenvironment, affording resistance to the BCL-2 inhibitor, venetoclax. Venetoclax resistance in the therapeutic setting and after long-term laboratory selection has been linked to metabolic alterations, but the underlying mechanism(s) are unknown. We aimed here to discover how CD40 stimulation as a model for tumor microenvironment-mediated metabolic changes, affects venetoclax sensitivity/resistance. CD40 stimulation increased oxidative phosphorylation and glycolysis, but only inhibition of oxidative phosphorylation countered venetoclax resistance. Furthermore, blocking mitochondrial import of pyruvate, glutamine or fatty acids affected CLL metabolism, but did not prevent CD40-mediated resistance to venetoclax. In contrast, inhibition of the electron transport chain (ETC) at complex I, III or V attenuated CLL activation and ATP production, and downregulated MCL-1 and BCL-XL, correlating with reduced CD40 surface expression. Moreover, ETC inhibition equaled mTOR1/2 but not mTOR1 inhibition alone for venetoclax resistance, and all three pathways were linked to control of general protein translation. In line with this, ETC plus mTOR inhibition synergistically counteracted venetoclax resistance. These findings link oxidative CLL metabolism to CD40 expression and cellular signaling, and may hold clinical potential.

Introduction Chronic lymphocytic leukemia (CLL) has become a paradigm for cancer in which the microenvironment plays a key role in the physiopathology of the disease.1-3 CLL cells circulate between lymph node (LN) niches and peripheral blood, which confer proliferating and quiescent states, respectively. CLL cells receive various signals in the LN microenvironment, which contains macrophages, stromal cells, monocyte-derived nurse-like cells and T cells. These signals promote cell survival, growth, proliferation and trafficking of cells between the peripheral blood and LN. The microenvironment in the LN activates and protects CLL cells through several mechanisms, such as chemokines, CLL surface molecules, adhesion molecules and tumor necrosis factor receptor members, which protect CLL cells from apoptosis.4-6 Among these mechanisms, the interactions between CLL and CD4+ T helper

cells via CD40-CD40L play a significant role in contributing to resistance to apoptosis.7-11 The expression of antiapoptotic proteins BCL-XL and MCL-1 in CLL cells from LN samples is higher than that in CLL cells from peripheral blood.12 We and others also found that in vitro CD40 stimulation increases the expression of BCL-XL, MCL-1 and Bfl-1, thereby mimicking LN signaling,13,14 which is important for the resistance of CLL to the widely used BCL-2 inhibitor venetoclax.11,13 In line with this, recent clinical trials show that in contrast to responses in the blood, LN responses are less complete.15 Hence, remaining LN sites are a likely source of emerging resistance, which is a growing clinical problem.16,17 Importantly, changes in cellular energy metabolism and mitochondrial reprogramming have also been linked to resistance to venetoclax,18-20 but the underlying mechanism(s) are still unknown. The well-known Warburg effect holds that aerobic glycolysis, i.e., limiting energy metabolism largely to glycoly-

Haematologica | 109 January 2024

151

ARTICLE - Venetoclax resistance in CLL

Z. Chen et al.

sis even in the presence of oxygen, is a hallmark of cancer.21,22 However, CLL does not behave metabolically like other “Warburg malignancies”, since enhanced mitochondrial oxidative phosphorylation (OXPHOS) but not glycolysis was seen in circulating CLL cells as compared to healthy B cells.23 In addition, we recently reported that CLL cells in LN have higher mitochondrial mass and glucose uptake than CLL cells in the blood, and that in vitro B-cell receptor and CD40 stimulation recapitulates various metabolic alterations observed in LN cells.24 The enhanced glycolysis and OXPHOS pathways were confirmed on gene expression and cellular metabolic levels. An important finding was that amino acids, in particular glutamine, fuel the mitochondrial tricarboxylic acid (TCA) cycle and thereby drive OXPHOS, while the contribution of glucose to OXPHOS is much lower.24 Even in cancers driven by the Warburg effect, it has been demonstrated that mitochondrial OXPHOS is crucial for survival.25-27 OXPHOS promotes the generation of metabolites for nucleotide, lipid and amino acid synthesis, which are essential for cell proliferation, yet it also generates onco-metabolites that contribute to tumorigenesis.28,29 As mentioned above, induced venetoclax resistance in cell lines has been associated with increased OXPHOS, and metabolic modulators can cooperate with venetoclax to overcome resistance.18 However, the link between OXPHOS and sensitivity to venetoclax in relationship to primary CLL in the microenvironment remains unclear. Therefore, using CD40 signaling as a common denominator, we investigated if and how energy metabolism is associated with venetoclax resistance in primary human CLL cells.

Methods Patients’ material and reagents Material was obtained from CLL patients, after having obtained their written informed consent, during routine follow-up or diagnostic procedures at the Academic Medical Center, Amsterdam, the Netherlands. The studies were approved by our Ethical Review Board and conducted in agreement with the Helsinki Declaration of 1975, revised in 1983. Peripheral blood mononuclear cells from patients with CLL, obtained after Ficoll density gradient centrifugation (Pharmacia Biotech, Roosendaal, the Netherlands) were cryopreserved and stored in liquid nitrogen. All samples contained more than 85% CD5+CD19+ cells, as assessed by flow cytometry. More details on the patients are provided in Online Supplementary Table S1. All reagents and products used are listed in Online Supplementary Table S2. Cell culture, flow cytometry and the venetoclax sensitivity assay, microarrays, microarray data normalization and differential expression, metabolic assays, western blot analy-

sis, real-time polymerase chain reaction, and the protein translation assay are described in detail in the Online Supplementary Data. Statistics The Student t test was used to analyze paired observations. One-way analysis of variance with multiple testing corrections was used to analyze differences between groups. The specific statistic test applied is indicated in the figure legends. For the figures to display relative values normalized to the condition of growth on 3T3 fibroblasts, statistical analysis was performed on data that were normalized. Statistically significant data are indicated as *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001.

Results Oxidative phosphorylation contributes to the resistance of primary chronic lymphocytic leukemia cells to venetoclax We previously showed that CD40 activation mimics the metabolic conditioning of CLL in the tumor microenvironment (TME).24 We corroborated this further by analyzing the oxygen consumption rate and extracellular acidification rate, as indicators of OXPHOS and glycolytic activity, respectively, using a Seahorse flux analyzer. After 48 h of CD40 stimulation in vitro, CLL cells showed higher oxygen consumption rate at both the basal and maximal levels (Online Supplementary Figure S1A, B), indicating that OXPHOS was increased. The same occurred with glycolysis, as CD40 activation increased extracellular acidification rate (Online Supplementary Figure S1C). Secondly, increased OXPHOS and glycolysis were further validated by analyzing our previously published microarray dataset.30 Accordingly, the majority of genes involved in OXPHOS and glycolysis were upregulated by CD40 stimulation (Figure 1A, B). In addition, pathway analysis confirmed that both OXPHOS and glycolysis were significantly upregulated by CD40 signaling, with both of them ranking among the top eight pathways (Figure 1C). To investigate whether these major changes in metabolic pathways are linked to venetoclax resistance, CLL cells were treated with OXPHOS or glycolysis inhibitors during CD40 stimulation and subsequently exposed to venetoclax. OXPHOS inhibition by oligomycin substantially counteracted venetoclax resistance (1,000-fold shift in IC50), while glycolysis inhibition by 2-deoxy-D-glucose had a very modest, non-significant effect (Figure 1D). The addition of venetoclax did not have a direct effect on oxygen consumption rate (Online Supplementary Figure S1D). Of note, neither inhibition of OXPHOS nor of glycolysis directly induced apoptosis of CLL cells, although 2-deoxyD-glucose reduced the pro-survival effect of CD40

Haematologica | 109 January 2024

152

ARTICLE - Venetoclax resistance in CLL

Z. Chen et al.

stimulation back to the unstimulated level (Figure 1E). In conclusion, inhibition of OXPHOS/ATPase by oligomycin during CD40 stimulation did not affect CLL viability, but rapidly and strongly attenuated induction of venetoclax resistance.

investigate the cross-talk between CD40 signaling, resistance to venetoclax and mitochondrial metabolism, we studied the role of the TCA cycle. To do this, we inhibited import of the three main fuels into the mitochondria and TCA cycle was inhibited alone or in combination during CD40 stimulation: pyruvate was blocked by UK5099, glutaInhibition of glutaminolysis impairs chronic lymphocytic mine by 6-diazo-5-oxo-L-norleucine (DON), and fatty acids leukemia cell metabolism but not sensitivity to venetoclax by etomoxir. The increase in mitochondrial mass observed ATP is produced by OXPHOS via the electron transport chain upon CD40 stimulation was strongly reduced by oligomycin (ETC). The TCA cycle supplies FADH2 and NADH, which sub- and at best marginally affected (e.g., DON) by these three sequently transfer their electrons to the ETC.31 To further fuel inhibitors (Figure 2A). Increased glucose uptake follow-

E Figure 1. CD40 stimulation enhances the metabolic activity of chronic lymphocytic leukemia cells, and blocking oxidative phosphorylation but not glycolysis attenuates resistance to venetoclax. Gene expression data were retrieved from a previously published microarray dataset of chronic lymphocytic leukemia (CLL) cells cultured on 3T3 or 3T40 fibroblasts for 16 h.30 (A) Heatmap of the expression of the top 30 significantly changed oxidative phosphorylation-related genes by 3T40 (N=3). (B) Heatmap of the expression of glycolysis-related genes by 3T40 (N=3). (C) Top eight pathways upregulated by 3T40. Pathway analysis was performed using CAMERA.60 Gene sets were retrieved from MSigDB v5.1 (hallmark gene sets, Entrez Gene ID): http://www.broadinstitute.org/gsea/msigdb/index.jsp. (D) CLL cells (N=6) were cultured on 3T3 or 3T40 fibroblasts and simultaneously treated with 10 nM oligomycin or 4 μM 2-deoxy-D-glucose for 24 h. Subsequently, CLL cells were harvested and incubated with a titration of 1-10,000 nM venetoclax for another 24 h after which viability was measured using DiOC6 and TO-PRO-3. Specific apoptosis was calculated. (E) Viability of CLL cells was also measured by flow cytometry directly after harvesting from fibroblasts. Data are shown as mean ± standard error of mean. One-way analysis of variance was performed to compare culture on 3T3 fibroblasts to the other conditions. Each patient is labeled with a distinct symbol across all the conditions. 2DG: 2-deoxy-D-glucose.

Haematologica | 109 January 2024

153

ARTICLE - Venetoclax resistance in CLL

Z. Chen et al.

ing CD40 stimulation was attenuated by oligomycin and DON, but not by either UK-5099 or etomoxir (Figure 2A). In addition, DON clearly suppressed both basal oxygen consumption rate and extracellular acidification rate, whereas the other two fuel inhibitors, UK5099 and etomoxir, had much weaker effects (Figure 2B). The combination of UK5099 and etomoxir did not have a significant impact on metabolic parameters (Online Supplementary Figure S2A, B), and only the combinations including DON resulted in significant decreases of glycolytic and mitochondrial parameters (Online Supplementary Figure S2A, B). Importantly, none of the combinations decreased CLL cell viability (Online Supplementary Figure S2C). These results indicate that inhibition of glutamine conversion to glutamate by DON is key for the maintenance of CLL metabolism, rather than pyruvate or long-chain fatty acids. This is in accordance with our previous findings that CLL cells predominantly use

glutamine/glutamate to fuel the TCA cycle.24 Although the fuel inhibitors had an impact on CLL metabolic activity, none of them affected CD40 stimulation itself, represented by CD95 induction. The induction of CD95 remained stable after single or combinations of inhibition of the three main fuels of the TCA cycle, while it was significantly decreased by the OXPHOS inhibitor oligomycin (Figure 2C, Online Supplementary Figure S2D). Most importantly, venetoclax resistance was strongly affected by oligomycin as before (Figure 1D), yet hardly affected, not even by DON, by any of the fuel inhibitors (Figure 2D, Online Supplementary Figure S2E). As previously indicated, CLL cells showed high metabolic flexibility when OXPHOS was inhibited; after CLL cells were treated with oligomycin, oxygen consumption rate was largely decreased while extracellular acidification rate was elevated (Figure 2B). This suggested that CLL cells could easily switch to glycolysis as their main energy path-

Figure 2. Inhibition of glutaminolysis impairs chronic lymphocytic leukemia cell metabolism but not sensitivity to venetoclax. Chronic lymphocytic leukemia (CLL) cells were cultured on 3T3 or 3T40 fibroblasts and simultaneously treated with 10 nM oligomycin, 20 μM 6-diazo-5-oxo-L-norleucine, 2 μM UK-5099 or 4 μM etomoxir for 24 h. After resuspension, (A) mitochondrial mass and glucose uptake were measured by flow cytometry (N=5). (B) A Seahorse MitoStress test was performed (N=4), basal oxygen consumption rate and basal extracellular acidification rate were measured on a Seahorse XF96 analyzer. (C) CD95 expression was measured by flow cytometry (N=5). (D) Cells were subsequently incubated with a titration of 1-10,000 nM venetoclax for 24 h after which viability was measured using DiOC6 and TO-PRO-3 staining. Then specific apoptosis was calculated (N=5). Data are shown as mean ± standard error of mean. One-way analysis of variance was performed to compare culture on 3T40 fibroblasts to the other conditions. Each patient is labeled with a distinct symbol across all the conditions. *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001; n.s.: not significant. MFI: mean fluorescence intensity; OCR: oxygen consumption rate; FCCP: carbonyl cyanide-p-trifluoromethoxyphenylhydrazone; ECAR: extracellular acidification rate; DON: 6-diazo-5-oxo-L-norleucine.

Haematologica | 109 January 2024

154

ARTICLE - Venetoclax resistance in CLL

Z. Chen et al.

way when OXPHOS was inhibited. In contrast, DON decreased OXPHOS of CLL cells, but to a lesser extent than oligomycin, and apparently did not enforce CLL cells to enhance lactate production (Figure 2B). Thus, inhibition of import of these three fuels was not capable of affecting venetoclax sensitivity, despite attenuation of OXPHOS. These data suggested that either other (not tested) substrates also contribute to fueling the TCA cycle in CD40stimulated CLL cells or that oligomycin has additional effects besides ETC inhibition. We continued our search by investigating the latter option. Electron transport chain inhibitors interfere with CD40 activation and anti-apoptotic protein expression As oligomycin inhibits the mitochondrial complex V, the last stage of the ETC, we studied whether resistance to venetoclax could also be overcome by inhibiting the activity of the other ETC complexes, using inhibitors of complex I (rotenone), II (dimethyl malonate) and III (antimycin A) of OXPHOS. These mitochondrial complexes were blocked by inhibitors during CD40 activation, which had no direct effect on CLL viability (Online Supplementary Figure S3A). Complex I and III inhibitors (rotenone, antimycin A) also increased sensitivity to venetoclax, but to a lesser extent than oligomycin (Figure 3A). ATP levels were decreased upon inhibition of complexes I, III and V (Online Supplementary Figure S3B). In contrast, inhibiting complex II of the ETC with dimethyl malonate had no effect on venetoclax sensitivity and also did not decrease ATP levels (Online Supplementary Figure S3B). These data indicate that the specific inhibition of complexes involved in proton pumping and ATP production leads to venetoclax sensitivity. Except for the complex II inhibitor, all other ETC inhibitors decreased CD95 expression to similar levels (Figure 3B). Venetoclax resistance is directly correlated with the expression levels of anti-apoptotic proteins MCL-1 and BCLXL.13,14 Indeed, the CD40-induced proteins BCL-XL and MCL-1 were strongly downregulated by all ETC inhibitors except for dimethyl malonate (Figure 3C). These analyses were done by flow cytometry, as described before;14,32 details are provided in the Online Supplementary Methods and Online Supplementary Figure S3C. BCL-2 expression was altered in a opposite way (Figure 3C), indicating that not all pro-survival proteins were similarly affected. OXPHOS inhibition can lead to increased production of reactive oxygen species, which can affect cell survival pathways,33,34 but levels of total or mitochondrial reactive oxygen species levels upon ETC inhibition were not affected by oligomycin (Online Supplementary Figure S3D). Additionally, N-acetylL-cysteine, a reactive oxygen species scavenger, was utilized in conjunction with rotenone, antimycin A and oligomycin. The results demonstrated that the decrease in CD95 expression induced by these ETC inhibitors was not recovered but even further reduced by N-acetyl-L-cysteine,

while viability was preserved (Figure 3D). Since CD40 signaling was affected by ETC inhibition, we investigated overall expression of CD40 protein and RNA expression (Figure 3E, F). Interestingly, CD40 stimulation had a huge impact on CD40 expression itself: RNA expression increased more than 300-fold (Figure 3F). Protein expression also increased, but not as dramatically as that of the RNA. Importantly, these changes were blocked by ETC inhibition, except for inhibition of complex II (Figure 3E). These findings indicate that ETC inhibition at complex I, III or V reduces CD40 signaling and venetoclax resistance in CLL, while oxidation of succinate to fumarate by complex II is not involved in this process. Furthermore, these effects are not due to overproduction of reactive oxygen species. mTOR connects oxidative phosphorylation with CD40 signaling and venetoclax resistance We next investigated how CD40 downstream kinases link CD40 signaling, metabolism and venetoclax resistance. We therefore inhibited phosphatidyl inositol 3-kinase (PI3K) by idelalisib, Bruton tyrosine kinase (BTK) by ibrutinib, mTOR1 by rapamycin, mTOR1/2 by AZD8055 and protein kinase B (AKT) by MK2206 (Online Supplementary Figure S4A). The results showed that besides oligomycin, only AZD8055 decreased mitochondrial mass (Figure 4A, Online Supplementary Figure S4B), and both rapamycin and AZD8055, but not the others, modestly suppressed glucose uptake (Figure 4A, Online Supplementary Figure S4B). mTOR inhibition, especially by AZD8055, decreased basal oxygen consumption rate, extracellular acidification rate and spare respiration capacity, while inhibition of BTK or PI3K did not (Figure 4B). Importantly, AZD8055, but not rapamycin or other kinase inhibitors, suppressed the overexpression of the anti-apoptotic proteins BCL-XL and MCL-1 by CD40 signaling (Figure 4C, Online Supplementary Figure S4C), which consequently strongly alleviated resistance to venetoclax (Figure 5A, Online Supplementary Figure S4D). These results suggested that the effects of ETC inhibition on CD40 signaling might be analogous to those of inhibition of mTOR and, more prominently, both mTOR1 and mTOR2 by AZD8055. Since AZD8055 showed similar effects as ETC inhibitors, we checked CD40 expression and activation upon mTOR inhibition. Figure 5B shows that only AZD8055, like oligomycin, clearly suppressed the expression of CD40 itself and the activation marker CD95 (Figure 5C). In addition, the quantification data indicated that the effects of rapamycin on CD40 expression were quite variable across patients. Importantly, rapamycin as a single agent did not affect CD40-induced resistance to venetoclax (Figure 5A). To further probe the relationship between mTOR and OXPHOS, several downstream targets of mTOR were evaluated in the presence of oligomycin in comparison with AZD8055 or rapamycin. CD40 activation induced p-AKTS473, p-GSK3β,S9 p-eIF2αS51 and p-AKTT450, while oligomycin,

Haematologica | 109 January 2024

155

ARTICLE - Venetoclax resistance in CLL

Z. Chen et al.

AZD805 and rapamycin suppressed these important downstream effectors (Online Supplementary Figure S5). There were differences in terms of quantification of these phosphorylation events, which can be expected when analyzing primary patients’ samples, but the overall pattern was consistent. Of note, ibrutinib and idelalisib as upstream inhibitors of BTK and PI3K did not have effects (Online Supplementary Figure S5). In conclusion, the compiled data indicate that inhibition of mTOR1/2 had equiv-

alent effects on metabolism and venetoclax sensitivity as ETC inhibition by oligomycin. Electron transport chain and mTOR inhibition suppress general protein translation and are synergistic in reversing resistance to venetoclax AKT and mTOR1/signaling have an important effect on translation, including the regulation of expression of specific pro-survival proteins in the context of CLL.30,35 There-

Figure 3. Inhibition of complex I, III or V of the electron transport chain during CD40 stimulation affects expression of BCL-2 members and resistance to venetoclax. Chronic lymphocytic leukemia (CLL) cells were cultured on 3T3 or 3T40 fibroblasts and simultaneously treated with various electron transport chain inhibitors for 24 h. (A) After resuspension, CLL cells were incubated with a titration of 1-10,000 nM venetoclax for 24 h after which viability was measured using MitoTracker Orange CMTMROS and TOPRO-3. Specific apoptosis was calculated and is presented (N=4). (B) Right after the resuspension of CLL cells, expression of CD95 (N=6), and (C) BCL-2 family members BCL-XL (N=4), MCL-1 (N=4) and BCL2 (N=4) were measured by flow cytometry. (D) CLL cells were cultured on 3T3 or 3T40 fibroblasts and simultaneously treated with rotenone (N=3), antimycin A (N=2) or oligomycin (N=3) with or without Nacetyl-L-cysteine for 24 h. After resuspension, CD95 and viability were measured by flow cytometry. (E) CLL cells were cultured on 3T3 or 3T40 fibroblasts and simultaneously treated with various electron transport chain inhibitors for 24 h. After resuspension, CD40 protein expression was measured by western blot; a representative scanning of three independent experiments is shown. The bar chart shows the relative quantification of CD40 expression to actin (N=5). (F) CD40 expression at the RNA level was measured by quantitative polymerase chain reaction (N=4). Data are shown as mean ± standard error of mean. One-way analysis of variance was performed to compare growth on 3T40 fibroblasts to the other conditions. Each patient is labeled with a distinct symbol across all the conditions. *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001; n.s.: not significant. DMM: dimethyl malonate; MFI: mean fluorescence intensity; NAC: N-acetyl-L-cysteine. Haematologica | 109 January 2024

156

ARTICLE - Venetoclax resistance in CLL

Z. Chen et al.

fore, we measured newly translated proteins after treatment with oligomycin, AZD8055 or rapamycin during CD40 activation to determine this effect. After CD40 activation, CLL cells generated much more newly synthesized proteins and, indeed, rapamycin and AZD8055 but also oligomycin counteracted this increase (Figure 6A). Broadly, this shows that these agents do not seem to affect the CD40-mediated activation of the CLL cells, but more specifically the consequences on protein synthesis. As ETC and mTOR inhibition showed overlapping effects, we determined whether these can act complementarily by performing synergy tests. Combinations of various concentrations of oligomycin and AZD8055 or rapamycin were added to CLL cells during CD40 activation. The results showed that inhibitors by themselves had no effect on cell viability, and that only at high doses did the combinations decrease cell viability (Online Supplementary Figure S6A). Oligomycin and AZD8055 exhibited dosagedependent suppression of CD95 expression, and effects

of combinations on CD95 were comparable to those of oligomycin alone (Online Supplementary Figure S6B). Strikingly, a synergistic effect was clear in venetoclax sensitivity tests; oligomycin with AZD8055 or rapamycin sensitized CLL cells to venetoclax much more than the inhibitors alone. This was tested with various combinations of mTOR inhibitors, oligomycin and venetoclax (Figure 6B, Online Supplementary Figure S6C), which were able to almost fully revert the treated cells to the venetoclax sensitivity of control cells. In line with the stronger effect of AZD8055 as a single agent, synergy of AZD8055 with oligomycin was stronger than that of oligomycin with rapamycin, as shown by the synergy score tables (Figure 6C). In conclusion, collectively these data indicate that suppression of protein translation is a common effect of both ETC and mTOR inhibition, and dual inhibition acts synergistically to converge on an almost complete reversal of CD40-mediated resistance to venetoclax.

Figure 4. mTOR1/2 inhibition suppresses metabolic activities and expression of BCL-XL and MCL-1. Chronic lymphocytic leukemia (CLL) cells were cultured on 3T3 or 3T40 fibroblasts and simultaneously treated with various inhibitors of downstream regulators of CD40 signaling for 24 h. After resuspension, (A) mitochondrial mass (N=3) and glucose uptake (N=3) were measured by flow cytometry. (B) A Seahorse MitoStress test was performed after CLL cells were collected (N=4), basal oxygen consumption rate, basal extracellular acidification rate and spare respiration capacity were measured and calculated. (C) Intracellular staining of BCL-xL (N=4) and MCL-1 (N=4) was done after resuspension of CLL cells and their expressions were measured by flow cytometry. Data are shown as mean ± standard error of mean. One-way analysis of variance was performed to compare culture on 3T40 fibroblasts to the other conditions. Each patient is labeled with a distinct symbol across all the conditions. *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001; n.s.: not significant. MFI: mean fluorescence intensity; OCR: oxygen consumption rate; FCCP: carbonyl cyanide-p-trifluoromethoxyphenylhydrazone; ECAR: extracellular acidification rate; SCR: spare respiration capacity. Haematologica | 109 January 2024

157

ARTICLE - Venetoclax resistance in CLL

Z. Chen et al.

Discussion The mitochondrion is crucial for both energy production and survival/apoptosis sensing. In this study, the contribution of metabolic pathways to CD40-induced venetoclax resistance was examined. By attenuating the expression of the anti-apoptotic proteins BCL-XL and MCL-1, which are localized on mitochondria where OXPHOS is also taking place, ETC inhibition clearly exhibits the capacity to counteract venetoclax resistance. In contrast, inhibition of the utilization of glucose, glutamine or long-chain fatty acids in the TCA cycle did not attenuate venetoclax resistance. This might indicate that other fuels can be used by activated CLL cells and that the mechanisms of venetoclax resistance are linked to maintenance of the ETC.

Recently, we showed that upon CD40 stimulation, CLL cells upregulate a wide range of metabolic pathways, many of them related to mitochondrial metabolism, and that glutamine contributes more than glucose to the TCA cycle in CD40-stimulated cells.24 In the present study we have extended this by showing that only inhibition of the conversion of glutamine to glutamate, but not inhibition of pyruvate utilization in the mitochondria, decreased oxygen consumption rate. Interestingly, when using an inhibitor of the plasma membrane glutamine transporter ASCT2, CLL cells became more sensitive to venetoclax,24 while only inhibiting glutaminolysis did not have this effect. This can be explained by the essential contribution of glutamine to other biosynthetic pathways such as nucleotide synthesis or tRNA loading for protein synthesis. In addition, ASCT2 also transports other amino acids; for

Figure 5. mTOR1/2 inhibition affects venetoclax sensitivity and CD40 expression. Chronic lymphocytic leukemia (CLL) cells were cultured on 3T3 or 3T40 fibroblasts and simultaneously treated with various inhibitors of downstream regulators of CD40 signaling for 24 h. (A) After resuspension, CLL cells were incubated with a titration of 1-10,000 nM venetoclax for 24 h after which viability was measured by flow cytometry using MitoTracker Orange CMTMROS and TO-PRO-3. Then specific apoptosis was calculated (N=5) (B) After resuspension, protein of CLL cells was directly isolated and CD40 protein expression was measured by western blot. A representative scanning of three independent experiments is shown. The bar chart shows the relative quantification of CD40 expression to actin (N=5). (C) CD95 was measured by flow cytometry (N=7). Data are shown as mean ± standard error of mean. One-way analysis of variance was performed to compare growth on 3T40 fibroblasts to the other conditions. Each patient is labeled with a distinct symbol across all the conditions. *P<0.05; **P<0.01; ***P<0.001; n.s.: not significant. Haematologica | 109 January 2024

158

ARTICLE - Venetoclax resistance in CLL

Z. Chen et al.

example, alanine metabolism was also a hit in our previous metabolomics and pathway analyses.24 We have identified key aspects of how the ETC interacts with CD40 signaling to affect venetoclax resistance, most

prominently by attenuating the requisite protein synthesis of CD40 itself, and pro-survival BCL-2 proteins MCL-1 and BCL-XL. Various searches into clinical venetoclax resistance did not reveal a unifying underlying mechanism, although

Figure 6. Cellular translation is suppressed by electron transport chain and mTOR inhibition. (A) Chronic lymphocytic leukemia (CLL) cells were cultured on 3T3 or 3T40 fibroblasts and simultaneously treated with oligomycin, AZD8055 or rapamycin for 24 h. After resuspension, newly translated proteins were measured with a Click-iT Plus OPP Kit (N=4). (B) CLL cells were cultured on 3T3 or 3T40 fibroblasts and simultaneously treated with 5 nM oligomycin, 250 nM AZD8055, 500 nM rapamycin or combinations for 24 h. After resuspension, cells were incubated with 1-10,000 nM venetoclax for 24 h after which viability was measured by flow cytometry using MitoTracker Orange CMTMROS and TO-PRO-3. Then specific apoptosis was calculated (N=3). (C) CLL cells were cultured on 3T3 or 3T40 fibroblasts and simultaneously treated with various concentrations of oligomycin, AZD8055, rapamycin or combinations for 24 h. After resuspension, CLL cells were incubated with 10 or 100 nM venetoclax for 24 h after which viability was measured by flow cytometry using MitoTracker Orange CMTMROS and TO-PRO-3. Synergy scores for electron transport chain/mTOR inhibitors and venetoclax were calculated (from Online Supplementary Figure S6C) (N=3). The scores in blue represent no synergy whereas those in red indicate strong synergy. Data are shown as mean ± standard error of mean. One-way analysis of variance was performed to compare culture on 3T40 fibroblasts to the other conditions. Each patient is labeled with a distinct symbol across all the conditions. *P<0.05; **P<0.01; ***P<0.001; n.s.: not significant. MFI: mean fluorescence intensity; AZD: AZD8055; VEN: venetoclax; Oligo: oligomycin; Rapa: rapamycin. Haematologica | 109 January 2024

159

ARTICLE - Venetoclax resistance in CLL

Z. Chen et al.

in many instances changes in BCL-2 family members were noted, including BCL-2 mutations, induction of BCL-XL, MCL-1, and BFL-1, and deletion of Bax.18,20,36,37 Importantly, several reports have also linked venetoclax resistance to mitochondrial metabolism, mostly focusing on the selection of resistant cell lines obtained after extended culture in the presence of venetoclax.18,19 Furthermore, changes in metabolism, specifically higher OXPHOS, were found in samples from patients who developed venetoclax resistance.18,20,38 Our study applied a distinct approach in which primary CLL cells were activated in vitro with CD40 signaling which mimics the LN-TME and rapidly induces venetoclax resistance. Intriguingly, TME-inducible venetoclax resistance exhibited a similar higher metabolic status as long-term selection under therapy or in the laboratory. This indicates that CLL cells that become resistant to venetoclax in patients due to either LN-TME signals or long-term venetoclax treatment might have comparable metabolic characteristics. This notion is further supported by the fact that ETC inhibitors attenuate venetoclax resistance in both settings, as we found in this study and as previously noted by others.18,20 Our data also demonstrate a unique function of mTOR1/2 signaling in sustaining venetoclax resistance. While both rapamycin and AZD8055 decreased the metabolic activity of CLL cells, only AZD8055 as a single agent sensitized the cells to venetoclax. ETC inhibitors and AZD8055 lowered all regulators linked with venetoclax sensitivity, including BCL2 family members and CD40 activation, whereas rapamycin did not. Yet importantly, despite variations in their effects on CD40 signaling and venetoclax sensitivity, AZD8055 and rapamycin exhibited a synergy combined with ETC inhibition on venetoclax sensitivity, leading to almost complete reversal of CD40-induced venetoclax resistance. As a unifying feature general protein translation was inhibited by ETC inhibition, and rapamycin and AZD8055 had a similar effect. Of note, p-eIF2aS51, a key translational regulator,39 was upregulated by CD40 (Online Supplementary Figure S5). P-eIF2aS51 is generally known to inhibit protein synthesis, but it can also promote the selective translation of some specific mRNA that are essential for cellular adaptation to stress.40 As we and others found that protein translation is upregulated in CLL TME,41,42 these apparent counteracting responses may be an interesting subject for future study. In addition to CD40, other TME signals may promote venetoclax resistance in CLL cells. Previously, we determined that B-cell receptor-induced venetoclax resistance can be counteracted by inhibiting glutamine import.24 More recently, we reported that TLR9 can also contribute to venetoclax resistance in patients under treatment with ibrutinib.32 The effect of TLR9 in that setting was to boost protein translation, which is the opposite of the effects of mTOR and ETC inhibition reported here. As both mTOR and ETC inhibitors have been authorized for

use in clinical trials, the synergistic impact of combining ETC and mTOR inhibition may provide an opportunity for enhancing the efficacy of venetoclax. Currently, a large number of clinical trials apply mTOR inhibitors in various cancers,43-46 and also some clinical trials regarding ETC inhibition, are ongoing. Metformin, a putative complex I inhibitor, has been tried in combination with several therapies to treat cancers, e.g., lung adenocarcinoma, ovarian cancer and breast cancer, with limited success.47-50 The complex I inhibitor IACS-010759 had antitumor activity in brain tumor and acute myeloid leukemia.51,52 IACS-010759 was also studied in vitro on CLL cells; it inhibited OXPHOS but exhibited minimal cytotoxicity.53 Other drugs that target mitochondrial metabolism have been approved for clinical trials. CB-839, which inhibits glutaminolysis, and CPI-613, which inhibits the TCA cycle, are in clinical trials for various types of cancer.54-57 In addition, preclinical research indicates that targeting OXPHOS is a promising strategy for enhancing the efficacy of venetoclax in acute myeloid leukemia.52,58,59 There is a paucity of research relating to metabolism and venetoclax, especially for activated CLL in a TME setting. Our research demonstrated the synergy of OXPHOS and mTOR inhibition in relation to venetoclax sensitization, which may hold promise for clinical applications. Disclosures No conflicts of interest to disclose. Contributions EE, GJWvdW, and ZC conceptualized the study. ZC, EE, GJWvdW, and APK were responsible for the methodology. ZC, GC, and VC performed investigations. APK was responsible for patients’ samples. ZC, GC, and EE were responsible for visualization. EE and APK acquired funding. EE and ZC were project administrators. EE, APK, and GJWvdW supervised the study. ZC. EE, and GC wrote the original manuscript draft. ZC, GC, HS-M, APK, GJWvdW, and EE reviewed and edited the manuscript. Acknowledgments The authors thank the patients for their blood donations and cooperation in the studies. Funding This work was supported by the Netherlands Organization for Scientific Research/Netherlands Organization for Health Research and Development Vidi grant 91715337, ERC Consolidator: BOOTCAMP (864815), Lymph and Co: 2018-LYCo008 and Cancer Center Amsterdam grant 2022. Data-sharing statement The original data are not shared online. The details of the experimental protocols are described in the Methods section and figure legends.

Haematologica | 109 January 2024

160

ARTICLE - Venetoclax resistance in CLL

Z. Chen et al.

References 1. Chiorazzi N, Chen SS, Rai KR. Chronic lymphocytic leukemia. Cold Spring Harb Perspect Med. 2021;11(2):1-35. 2. ten Hacken E, Burger JA. Microenvironment dependency in chronic lymphocytic leukemia: the basis for new targeted therapies. Pharmacol Ther. 2014;144(3):338-348. 3. Burger JA. Nurture versus nature: the microenvironment in chronic lymphocytic leukemia. Hematology Am Soc Hematol Educ Program. 2011;2011(1):96-103. 4. Haselager M, Thijssen R, West C, et al. Regulation of Bcl-XL by non-canonical NF-κB in the context of CD40-induced drug resistance in CLL. Cell Death Differ. 2021;28(5):1658-1668. 5. ten Hacken E, Burger JA. Molecular pathways: targeting the microenvironment in chronic lymphocytic leukemia-focus on the B-cell receptor. Clin Cancer Res. 2014;20(3):548-556. 6. Kennedy E, Coulter E, Halliwell E, et al. TLR9 expression in chronic lymphocytic leukemia identifies a promigratory subpopulation and novel therapeutic target. Blood. 2021;137(22):3064-3078. 7. Panayiotidis P, Jones D, Ganeshaguru K, Foroni L, Hoffbrand AV. Human bone marrow stromal cells prevent apoptosis and support the survival of chronic lymphocytic leukaemia cells in vitro. Br J Haematol. 1996;92(1):97-103. 8. Burger JA, Tsukada N, Burger M, Zvaifler NJ, Dell’Aquila M, Kipps TJ. Blood-derived nurse-like cells protect chronic lymphocytic leukemia B cells from spontaneous apoptosis through stromal cell-derived factor-1. Blood. 2000;96(8):2655-2663. 9. Ghia P, Strola G, Granziero L, et al. Chronic lymphocytic leukemia B cells are endowed with the capacity to attract CD4+,CD40L+ T cells by producing CCL22. Eur J Immunol. 2002;32(5):1403-1413. 10. Burger JA, Gandhi V. The lymphatic tissue microenvironments in chronic lymphocytic leukemia: in vitro models and the significance of CD40-CD154 interactions. Blood. 2009;114(12):2560-2561. 11. Kurtova AV, Balakrishnan K, Chen R, et al. Diverse marrow stromal cells protect CLL cells from spontaneous and druginduced apoptosis: development of a reliable and reproducible system to assess stromal cell adhesion-mediated drug resistance. Blood. 2009;114(20):4441-4450. 12. Smit LA, Hallaert DYH, Spijker R, et al. Differential Noxa/Mcl-1 balance in peripheral versus lymph node chronic lymphocytic leukemia cells correlates with survival capacity. Blood. 2007;109(4):1660-1668. 13. Thijssen R, Slinger E, Weller K, et al. Resistance to ABT-199 induced by microenvironmental signals in chronic lymphocytic leukemia can be counteracted by CD20 antibodies or kinase inhibitors. Haematologica. 2015;100(8):e302-306. 14. Haselager M V., Kielbassa K, ter Burg J, et al. Changes in Bcl-2 members after ibrutinib or venetoclax uncover functional hierarchy in determining resistance to venetoclax in CLL. Blood. 2020;136(25):2918-2926. 15. Seymour JF, Kipps TJ, Eichhorst B, et al. Venetoclax-rituximab in relapsed or refractory chronic lymphocytic leukemia. N Engl J Med. 2018;378(12):1107-1120. 16. Roberts AW, Davids MS, Pagel JM, et al. Targeting BCL2 with venetoclax in relapsed chronic lymphocytic leukemia. N Engl J Med. 2016;374(4):311-322. 17. Hallek M. Chronic lymphocytic leukemia: 2020 update on diagnosis, risk stratification and treatment. Am J Hematol. 2019;94(11):1266-1287. 18. Guièze R, Liu VM, Rosebrock D, et al. Mitochondrial

reprogramming underlies resistance to BCL-2 inhibition in lymphoid malignancies. Cancer Cell. 2019;36(4):369-384.e13. 19. Roca-Portoles A, Rodriguez-Blanco G, Sumpton D, et al. Venetoclax causes metabolic reprogramming independent of BCL-2 inhibition. Cell Death Dis. 2020;11(8):616. 20. Thomalla D, Beckmann L, Grimm C, et al. Deregulation and epigenetic modification of BCL2-family genes cause resistance to venetoclax in hematologic malignancies. Blood. 2022;140(20):2113-2126. 21. Warburg O. On the origin of cancer cells. Science. 1956;123(3191):309-314. 22. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011;144(5):646-674. 23. Jitschin R, Hofmann AD, Bruns H, et al. Mitochondrial metabolism contributes to oxidative stress and reveals therapeutic targets in chronic lymphocytic leukemia. Blood. 2014;123(17):2663-2672. 24. Chen Z, Simon-Molas H, Cretenet G, et al. Characterization of metabolic alterations of chronic lymphocytic leukemia in the lymph node microenvironment. Blood. 2022;140(6):630-643. 25. Zheng J. Energy metabolism of cancer: glycolysis versus oxidative phosphorylation (review). Oncol Lett. 2012;4(6):1151-1157. 26. Ashton TM, Gillies McKenna W, Kunz-Schughart LA, Higgins GS. Oxidative phosphorylation as an emerging target in cancer therapy. Clin Cancer Res. 2018;24(11):2482-2490. 27. Farge T, Saland E, de Toni F, et al. Chemotherapy-resistant human acute myeloid leukemia cells are not enriched for leukemic stem cells but require oxidative metabolism. Cancer Discov. 2017;7(7):716-735. 28. Vyas S, Zaganjor E, Haigis MC. Mitochondria and cancer. Cell. 2016;166(3):555-566. 29. Schiliro C, Firestein BL. Mechanisms of metabolic reprogramming in cancer cells supporting enhanced growth and proliferation. Cells. 2021;10(5):1056. 30. Van Attekum MHA, Terpstra S, Slinger E, et al. Macrophages confer survival signals via CCR1-dependent translational MCL-1 induction in chronic lymphocytic leukemia. Oncogene. 2017;36(26):3651-3660. 31. Martínez-Reyes I, Diebold LP, Kong H, et al. TCA cycle and mitochondrial membrane potential are necessary for diverse biological functions. Mol Cell. 2016;61(2):199-209. 32. Kielbassa K, Haselager MV, Bax DJC, et al. Ibrutinib sensitizes CLL cells to venetoclax by interrupting TLR9-induced CD40 upregulation and protein translation. Leukemia. 2023;37(6):1268-1276. 33. van Noorden CJF, Hira VVV, van Dijck AJ, Novak M, Breznik B, Molenaar RJ. Energy metabolism in IDH1 wild‐type and IDH1‐mutated glioblastoma stem cells: a novel target for therapy? Cells. 2021;10(3):1-16. 34. Xiao D, Powolny AA, Moura MB, et al. Phenethyl isothiocyanate inhibits oxidative phosphorylation to trigger reactive oxygen species-mediated death of human prostate cancer cells. J Biol Chem. 2010;285(34):26558-26569. 35. Fabbri L, Chakraborty A, Robert C, Vagner S. The plasticity of mRNA translation during cancer progression and therapy resistance. Nat Rev Cancer. 2021;21(9):558-577. 36. Blombery P, Thompson ER, Nguyen T, et al. Multiple BCL2 mutations cooccurring with Gly101Val emerge in chronic lymphocytic leukemia progression on venetoclax. Blood. 2020;135(10):773-777.

Haematologica | 109 January 2024

161

ARTICLE - Venetoclax resistance in CLL

Z. Chen et al.

37. Thijssen R, Tian L, Anderson MA, et al. Single-cell multiomics reveal the scale of multi-layered adaptations enabling CLL relapse during venetoclax therapy. Blood. 2022;140(20):2127-2141. 38. Bajpai R, Sharma A, Achreja A, et al. Electron transport chain activity is a predictor and target for venetoclax sensitivity in multiple myeloma. Nat Commun. 2020;11(1):1228. 39. Jackson RJ, Hellen CUT, Pestova TV. The mechanism of eukaryotic translation initiation and principles of its regulation. Nat Rev Mol Cell Biol. 2010;11(2):113-127. 40. Pavitt GD, Ron D. New insights into translational regulation in the endoplasmic reticulum unfolded protein response. Cold Spring Harb Perspect Biol. 2012;4(6):1-13. 41. Yeomans A, Lemm E, Wilmore S, et al. PEITC-mediated inhibition of mRNA translation is associated with both inhibition of mTORC1 and increased eIF2a phosphorylation in established cell lines and primary human leukemia cells. Oncotarget. 2016;7(46):74807-74819. 42. Wilmore S, Rogers-Broadway KR, Taylor J, et al. Targeted inhibition of eIF4A suppresses B-cell receptor-induced translation and expression of MYC and MCL1 in chronic lymphocytic leukemia cells. Cell Mol Life Sci. 2021;78(17-18):6337-6349. 43. Guertin DA, Sabatini DM. The pharmacology of mTOR inhibition. Sci Signal. 2009;2(67):pe24. 44. Chiang GG, Abraham RT. Targeting the mTOR signaling network in cancer. Trends Mol Med. 2007;13(10):433-442. 45. Kim JY, Jeon JY, Ko YM, Kang MS, Park SK, Roh K. Characteristics of lymphedema in patients treated with mammalian target of rapamycin inhibitors. Lymphat Res Biol. 2021;19(4):365-371. 46. Bhaoighill MN, Dunlop EA. Mechanistic target of rapamycin inhibitors: successes and challenges as cancer therapeutics. Cancer Drug Resist. 2019;2(4):1069-1085. 47. Brown JR, Chan DK, Shank JJ, et al. Phase II clinical trial of metformin as a cancer stem cell-targeting agent in ovarian cancer. JCI Insight. 2020;5(11):e133247. 48. Goodwin PJ, Parulekar WR, Gelmon KA, et al. Effect of metformin vs placebo on and metabolic factors in NCIC CTG MA.32. J Natl Cancer Inst. 2015;107(3):djv006. 49. Arrieta O, Barrón F, Padilla M-ÁS, et al. Effect of metformin plus tyrosine kinase inhibitors compared with tyrosine kinase

inhibitors alone in patients with epidermal growth factor receptor-mutated lung adenocarcinoma. JAMA Oncol. 2019;5(11):e192553. 50. Pollak M. Repurposing biguanides to target energy metabolism for cancer treatment. Nat Med. 2014;20(6):591-593. 51. Molina JR, Sun Y, Protopopova M, et al. An inhibitor of oxidative phosphorylation exploits cancer vulnerability. Nat Med. 2018;24(7):1036-1046. 52. Liu F, Kalpage HA, Wang D, et al. Cotargeting of mitochondrial complex I and Bcl-2 shows antileukemic activity against acute myeloid leukemia cells reliant on oxidative phosphorylation. Cancers (Basel). 2020;12(9):2400. 53. Vangapandu HV, Alston B, Morse J, et al. Biological and metabolic effects of IACS-010759, an OxPhos inhibitor, on chronic lymphocytic leukemia cells. Oncotarget. 2018;9(38):24980-24991. 54. Alistar AT, Desnoyers R, D’Agostino RJ, Pasche B. CPI-613 enhances FOLFIRINOX response rate in stage IV pancreatic cancer. Ann Oncol. 2016;27(suppl 6):vi228. 55. Pardee TS, Lee K, Luddy J, et al. A phase I study of the first-inclass antimitochondrial metabolism agent, CPI-613, in patients with advanced hematologic malignancies. Clin Cancer Res. 2014;20(20):5255-5264. 56. Raczka AM, Reynolds PA. Glutaminase inhibition in renal cell carcinoma therapy. Cancer Drug Resist. 2019;2(2):356-364. 57. Dos Reis LM, Adamoski D, Souza ROO, et al. Dual inhibition of glutaminase and carnitine palmitoyltransferase decreases growth and migration of glutaminase inhibition-resistant triplenegative breast cancer cells. J Biol Chem. 2019;294(24):9342-9357. 58. Stubbins RJ, Maksakova IA, Sanford DS, Rouhi A, Kuchenbauer F. Mitochondrial metabolism: powering new directions in acute myeloid leukemia. Leuk Lymphoma. 2021;62(10):2331-2341. 59. Hege Hurrish K, Su Y, Wiley S, et al. A novel isoflavone, ME344, enhances venetoclax antileukemic activity against AML via suppression of oxidative phosphorylation and purine biosynthesis. Blood. 2021;138(Suppl 1):2238. 60. Wu D, Smyth GK. CAMERA: a competitive gene set test accounting for inter-gene correlation. Nucleic Acids Res. 2012;40(17):e133.

Haematologica | 109 January 2024

162

ARTICLE - Chronic Lymphocytic Leukemia

Pro-inflammatory cells sustain leukemic clonal expansion in T-cell large granular lymphocyte leukemia Cristina Vicenzetto,1,2° Vanessa Rebecca Gasparini,1,2 Gregorio Barilà,1,2° Antonella Teramo,1,2 Giulia Calabretto,1,2 Elisa Rampazzo,1,2 Samuela Carraro,1° Valentina Trimarco,1 Livio Trentin,1 Monica Facco,1,2 Gianpietro Semenzato1,2 and Renato Zambello1,2

Correspondence: R. Zambello

Department of Medicine, Hematology and Clinical Immunology Branch, University of Padova and 2Veneto Institute of Molecular Medicine (VIMM), Padova, Italy

Received: Accepted: Early view:

Current address of CV: Department of Cardiac, Thoracic, Vascular Sciences and Public Health; University of Padova, Padova, Italy. ° Current address of GB: Hematology Unit, Ospedale San Bortolo, Vicenza, Italy. ° Current address of SC: Department of Medicine, University of Padova, Padova, Italy. °

r.zambello@unipd.it November 18, 2022. July 5, 2023. July 13, 2023.

Abstract T-cell large granular lymphocyte leukemia (T-LGLL) is a chronic lymphoproliferative disorder characterized by the clonal expansion of T-cell large granular lymphocytes (T-LGL). Immunophenotypic and genotypic features contribute to discriminate symptomatic (CD8+ STAT3-mutated T-LGLL) from clinically indolent patients, this latter group including CD8+ wildtype (wt), CD4+ STAT5B-mutated and wt cases. T-LGL lymphoproliferation is sustained both by somatic gain-offunction mutations (i.e., STAT3 and STAT5B) and by pro-inflammatory cytokines, but little information is available on the activity of T-LGLL non-leukemic cells. In this study, we characterized pro-inflammatory cells in the peripheral blood of T-LGLL patients and analyzed their role in supporting the leukemic growth. In symptomatic patients we found that cell populations not belonging to the leukemic component showed a discrete pro-inflammatory pattern. In particular, CD8+ STAT3-mutated cases showed a skewed Th17/Treg ratio and an abnormal distribution of monocyte populations characterized by increased intermediate and non-classical monocytes. We also demonstrated that monocytes released high levels of interleukin-6 after CCL5 stimulation, a chemokine specifically expressed only by leukemic LGL. Conversely, in asymptomatic cases an altered distribution of monocyte populations was not detected. Moreover, T-LGLL patients’ monocytes showed abnormal activation of signaling pathways, further supporting the different pathogenic role of monocytes in patients in discrete clinical settings. Altogether, our data contribute to deepening the knowledge on the different cell subtypes in T-LGLL, focusing particularly on non-leukemic cell populations and thus offering the rationale for new therapeutic strategies.

Introduction According to the 2022 World Health Organization (WHO) classification, T-cell large granular lymphocyte leukemia (TLGLL) is the most frequent disorder (up to 85%) among chronic lymphoproliferative diseases of large granular lymphocytes (LGL). Affected patients show remarkable phenotypic and genotypic heterogeneity as well as a variety of clinical features, ranging from indolent to symptomatic patients requiring therapy and showing a dismal outcome.1 Recently, phenotypic and molecular analyses allowed a better categorization of T-LGLL patients. In particular, most symptomatic cases are included in CD8+ T-LGLL and characterized by a CD3+/CD8+/CD16+/CD56- clone, whereas the remaining patients presenting with either CD8+ or CD4+ T-LGLL, the latter characterized by CD4+ or CD4+/CD8dim

clones, are usually asymptomatic.2,3 Many genes related to epigenetic modification have been found to be mutated frequently in T-LGLL patients, whose mutational status contributes to patients’ stratification.3,4 In particular, signal transducer and activator of transcription 3 (STAT3) mutations correlate with symptomatic disease in nearly 60% of CD8+ T-LGLL cases,5 while STAT5B mutations are found in 66%6 of the asymptomatic CD4+ T-LGLL group and in 100% of the rare aggressive variant of T-LGLL.5-7 Symptomatic patients (38%) are characterized by cytopenias, with neutropenia and related infections being the most frequent clinical features.5 Severe neutropenia (absolute neutrophil count <500/mm3) and/or recurrent infections require treatment which, to date, includes immunosuppressive agents (methotrexate, cyclophosphamide and cyclosporine A).8 T-LGLL may occur in association with auto-

Haematologica | 109 January 2024

163

ARTICLE - Role of pro-inflammatory cells in T-LGLL

C. Vicenzetto et al.

immune diseases and, in particular, rheumatoid arthritis is diagnosed in roughly 20% of T-LGLL patients.9 Interestingly, Felty syndrome, a rare variant of rheumatoid arthritis associated with splenomegaly and neutropenia, has been proposed as a diagnostic continuum due to the biological features consistent with T-LGLL, i.e., the identification of STAT3 mutations and a similar pro-inflammatory cytokine pattern in the peripheral blood.10 Moreover, the concurrence with autoimmune diseases is more frequently described in T-LGLL patients characterized by STAT3 mutations.3 STAT3 is constitutively active in T-LGLL patients, also in cases devoid of the specific mutations,2,11 this finding being considered as an essential factor for the survival of the leukemic clone.12 The state of activation of T-LGL clones is sustained either by the gain-of-function mutations frequently found in the Src homologous 2 (SH2) domain of the gene, with Y640F and D661Y being the most recurrent, or by cytokine stimulation.11,13 In particular, high levels of plasma interleukin (IL)-6 have been reported to elicit STAT3 phosphorylation and, in turn, to contribute to leukemic clone survival. In this setting, monocytes have been regarded as the main cellular source of IL-6.13,14 Many other pro-inflammatory cytokines have been associated with the pathogenesis of T-LGLL, including Fas-L, which is linked to the development of neutropenia,15 and CCL5, whose pathogenic mechanism has not been defined yet.2,16,17 In addition, CCR5, one of the CCL5 receptors, has been reported to be less expressed on the surface of leukemic cells, particularly in neutropenic patients.2,18 Among the different players of the microenvironment, including cellular components (stromal, accessory and regulatory cells, among others) as well as soluble factors (cytokines and chemokines), in the present study we evaluated the role of the peripheral blood pro-inflammatory environment in the pathogenesis of T-LGLL also in relationship to different T-LGLL subgroups. In particular, we analyzed the non-leukemic cell subsets, namely Th17, Treg and monocytes, focusing on their involvement in sustaining the LGL clonal expansion. It is worth mentioning that these cell populations are also abnormally represented in patients with autoimmune diseases.19,20 We herein provide information on the cytokine interplay demonstrating a relationship between monocytes and the leukemic clone in T-LGLL, particularly in symptomatic STAT3-mutated T-LGLL.

Methods Patients Fifty-eight patients and 27 healthy controls (HC, with matched age and sex) were enrolled in the study. The Padova Institutional Review Board approved this study and informed consent was collected from each subject in ac-

cordance with the Declaration of Helsinki. All patients met the 2022 WHO criteria for the diagnosis of T-LGLL.1 No patients had received treatment at the time of the study. Their detailed clinical and diagnostic features are reported in Online Supplementary Tables S1 and S2. Recruited patients were divided into four subgroups according to TLGLL immunophenotype and the STAT3 or STAT5B mutational status as previously described,21 namely CD8+ wildtype (wt) or STAT3-mutated and CD4+ wt or STAT5Bmutated patients. Cell isolation, purification and culture Peripheral blood mononuclear cells (PBMC) were obtained by Ficoll-Hypaque (Sigma-Aldrich) gradient centrifugation. T-LGL and/or monocytes were immunomagnetically purified using magnetic microbeads coated with monoclonal antibodies against CD57 or CD56 for T-LGL or CD14 for monocytes (Miltenyi Biotec). The purity of the populations was checked by flow cytometry (Becton Dickinson, BD) and resulted >97%. In selected cases, in vitro cultures were established. Cells derived from T-LGLL patients and HC were cultured at 2×106 cells/mL in complete RPMI-1640 (EuroClone) supplemented with 10% fetal calf serum (Sigma-Aldrich), 2 mM glutamine, 25 mM Hepes, 100 U/mL penicillin and 100 g/mL streptomycin (EuroClone) and grown in 5% CO2 at 37°C. PBMC and PBMC deprived of monocytes, derived from the same patient, were cultured for 1, 2, 3 and 5 days for the evaluation of activation of T-LGL apoptosis. For analysis of apoptosis and expression evaluations, purified monocytes and PBMC deprived of both T-LGL and monocytes were stimulated with 100 ng/mL CCL5 (R&D) for 12 h. Purified LGL were stimulated with 100 ng/mL IL-17A (CELLgs) for 24 h. Flow cytometric analyses The percentage of Th17 and Treg cells was assessed from PBMC using the commercial protocols Fix&Perm (ThermoFisher Scientific) and Foxp3 Buffer Set (BD), respectively. Th17 cells were stained with anti-human CD161 (FITC), IL17 (PE), IL-23R (APC) and CD4 (APC-Cy7) and Treg with anti-human Foxp3 (PerCP-Cy5.5), CD25 (PE-Cy7) and CD4 (APC-Cy7, BD). For these analyses 500,000 events were acquired and the populations were reported as the percentage of CD4+ cells. The distribution of peripheral blood monocyte populations (namely classical [CD14high/CD16neg], intermediate [CD14high/CD16dim] and non-classical [CD14dim/neg/CD16high]) was determined by staining whole blood with anti-human CD66b (FITC), CD14 (PE), CD16 (PerCP-Cy5.5), CD19 (APC) and CD3 (APC-Cy7, BD). IL-6 intracellular staining was performed following the commercial protocol Fix&Perm on whole blood treated for 6 h at 37°C with the exocytosis inhibitor Golgi Stop (BD)

Haematologica | 109 January 2024

164

ARTICLE - Role of pro-inflammatory cells in T-LGLL

C. Vicenzetto et al.

and 100 ng/mL lipopolysaccharide (Sigma-Aldrich). For this analysis anti-CD14 (APC), CD16 (PerCP-Cy5.5) and IL6 (PE, BD) were used. Apoptosis was evaluated by staining LGL with anti-human CD57 (FITC) and CD8 (PerCP-Cy5.5) and the apoptotic cells with PE-conjugated annexin V (BD). All the cytometric evaluations were performed with a BD

FACS Canto II and the data processed by Diva Software (BD). The gating strategies are reported in Online Supplementary Figures S1, S2 and S3. Expression analysis and statistical evaluation Details of the expression analysis and statistical evaluation are provided in the Online Supplementary Methods.

Figure 1. Distribution of Th17 and Treg cells in patients with T-cell large granular lymphocyte leukemia. (A, C) The percentages of Th17 and Treg were evaluated by flow cytometry. (B, D) The related Th17/Treg ratios were established mathematically. The KruskalWallis test corrected for multiple comparisons was used for the analysis. Data are reported as histograms showing the median with interquartile range. (A, B) Patients with CD8+ T-cell large granular lymphocyte leukemia (N=13) were characterized by an increase of Th17 cells, whereas Treg cells were reduced in CD4+ cases (N=10). The Th17/Treg ratio was imbalanced for both groups in comparison to that of healthy controls (N=4). (C, D) An increased Th17 percentage characterizes CD8+ wildtype (N=6) or STAT3-mutated (N=7) patients. Treg are reduced in CD4+ wildtype (N=7) and STAT5B-mutated (N=3) patients, leading to a significant skewing of the Th17/Treg ratio in this group. T-LGLL: T-cell large granular lymphocyte leukemia; wt: wildtype; mut: mutated; pts: patients. Haematologica | 109 January 2024

165

ARTICLE - Role of pro-inflammatory cells in T-LGLL

C. Vicenzetto et al.

Results Patients with CD8+ and CD4+ T-cell large granular lymphocyte leukemia are characterized by an unbalanced Th17/Treg ratio A significant increase of the percentage of Th17 lymphocytes was found in CD8+ T-LGLL patients as compared with both HC and CD4+ cases (P<0.05 and P<0.01, respectively) (Figure 1A), while Treg percentage was significantly reduced in the CD4+ group compared to HC and CD8+ cases (P<0.001 and P<0.05, respectively) (Figure 1A). These imbalances led to an altered Th17/Treg ratio for both CD4 and CD8 T-LGLL patients related to HC (P<0.05) (Figure 1B). In terms of absolute count, the number of Th17 cells was significantly increased in CD4+ patients as compared to HC (P<0.01) (Online Supplementary Figure S4A), while the number of Treg was unchanged. This resulted in an altered Th17/Treg ratio for CD4+ T-LGLL (P<0.05) (Online Supplementary Figure S4B). Clustering patients according to STAT3 and STAT5B mutations, increased Th17 cell percentages were documented in both CD8+ wt and STAT3-mutated patients (both P<0.05 vs. HC) while CD4+ cases (both STAT5B-mutated and wt) were characterized by a reduced percentage of Treg cells (Figure 1C). The evaluation of the absolute Th17 and Treg counts confirmed the relevant increase of Th17 cells (Online Supplementary Figure S4C, D). To note, the increased Th17 percentage and the skewed Th17/Treg ratio turned out to be a peculiar feature of CD8+ neutropenic patients (Online Supplementary Figure S5A-C). Since an abnormal distribution of Th17 and Treg is reported in autoimmune diseases,22 which are often associated with CD8+ STAT3-mutated T-LGL expansions, we clustered our patients according to the incidence of autoimmune diseases. No differences in the distribution of Th17 and Treg cells were observed between patients with CD8+ STAT3-mutated T-LGLL with or without concomitant autoimmune diseases, also in terms of their related Th17/Treg ratio (Online Supplementary Figure S5C-F). In further support of the results obtained, follow-up analysis showed that neutropenic CD8+ patients also maintained a higher percentage of Th17 cells compared to CD4+ cases at follow-up (53 months [interquartile range, 54.25-59.75]), whereas the CD4+ cases were characterized by a reduced percentage of Treg cells (Online Supplementary Figure S8A-C). The distribution and counts of monocyte populations are altered in T-cell large granular lymphocyte leukemia The distribution of peripheral blood monocyte populations of CD4+ T-LGLL patients displayed a pattern consistent with that of HC (Figure 2A, left panel; Online Supplementary Figure S6A, left panel). Conversely, in CD8+ T-LGLL the percentage of classical monocytes was significantly reduced (P<0.01 vs. HC and P<0.001 vs. CD4+ cases), with a slight increase of in-

termediate and non-classical subsets (Figure 2A, right panel). These alterations resulted specific to CD8+ STAT3-mutated patients. In particular, CD8+ STAT3-mutated cases displayed a significant reduction of classical monocytes (P<0.0001 vs. HC and CD8+ wt) and an increase of intermediate (P<0.05 vs. HC) and non-classical monocytes (P<0.01 vs. HC and P<0.05 vs. CD8+ wt) (Figure 2A; Online Supplementary Figure S6A). The alteration of monocyte percentages and their absolute counts strictly corresponded with the incidence of neutropenia (Online Supplementary Figure S7A, B). The percentage of classical monocytes was not influenced by the incidence of autoimmune diseases, while a significant reduction of their absolute count was observed (P<0.01) (Online Supplementary Figure S7C, D). The percentages of monocytes at follow-up were consistent with those observed at baseline (Online Supplementary Figure S8D). Evaluation of monocytes using a cell blood count test showed that CD8+ STAT3-mutated patients were characterized by a reduced number of peripheral monocytes compared to CD8+ wt cases (P<0.05) (Figure 2B). Cell blood counts also demonstrated that monocyte percentage was inversely correlated with the leukemic clone count when considering the entire cohort of patients (r=-0.5, P<0.001) (Figure 2C). Consistent results were found when dividing patients into CD8+ and CD4+ cases (Figure 2D, E), specifically for wt cases (r=-0.55, P<0.05 for CD8+ wt and r=-0.84, P<0.01 for CD4+ wt) (Online Supplementary Figure S6B-D). In CD8+ STAT3 and CD4+ STAT5B-mutated cases a significant correlation was not demonstrated (Online Supplementary Figure S6E, F). Monocytes activate different cellular pathways among the T-cell large granular lymphocyte leukemia subgroups Several pathways are involved in the activation of inflammatory signals in monocytes, including STAT3 and NFκB signaling.23 Interestingly, a peculiar activation status of these two pathways was found among subgroups of TLGLL patients (Figure 3). More in detail, STAT3 Tyr-705 phosphorylation was significantly increased in monocytes from CD8+ STAT3-mutated patients as compared to HC (P<0.05). This feature was also confirmed at the level of protein expression (P<0.001), with discrete statistical significance among the different subgroups (P<0.001 vs. CD8+ wt and vs. CD4+ wt cases; P<0.05 vs. CD4+ STAT5B-mutated cases) (Figure 3A). On the other hand, increased levels of phosphorylation of the NFκB subunit p65 were detected in CD8+ wt (P<0.05), CD4+ wt (P<0.001) and CD4+ STAT5B-mutated (P<0.001) patients compared to HC (Figure 3B). Monocytes from CD4+ STAT5B-mutated patients were characterized by a higher level of ERK phosphorylation as compared to those from HC (P<0.05) (Figure 3C). Representative western blot images are shown in Online Supplementary Figure S9.

Haematologica | 109 January 2024

166

ARTICLE - Role of pro-inflammatory cells in T-LGLL

C. Vicenzetto et al.

Pattern of IL-6 and CCL5 expression in T-cell large granular lymphocyte leukemia Since monocytes have been regarded as the principal source of IL-6,13 in both T-LGLL patients and HC, we evaluated IL-6 expression levels within the different monocyte

subsets by flow cytometry. Significantly higher IL-6 expression was observed in intermediate and non-classical monocytes compared with the classical ones in terms of percentage of IL-6 positive cells (P<0.0001 for both TLGLL patients and HC) (Figure 4A panel on the left) and of

Continued on following page. Haematologica | 109 January 2024

167

ARTICLE - Role of pro-inflammatory cells in T-LGLL

C. Vicenzetto et al.

Figure 2. Monocyte distribution in the peripheral blood of patients with T-cell large granular lymphocyte leukemia. (A) The percentages of classical, intermediate and non-classical monocytes were assessed by flow cytometry. Data are reported as histograms showing the mean with standard error and were analyzed by two-way analysis of variance. A reduction in classical monocytes was found in CD8+ patients (N=17) in comparison to CD4+ cases (N=12, •••P<0.001) and to healthy controls (N=5, ** P<0.01), as shown in the left panel. The imbalanced monocyte distribution resulted specific to STAT3-mutated cases (N=7) in comparison to CD8+ wildtype (wt) ones (N=10, •) and healthy controls (*) (•••• and ****P<0.0001; **P<0.01 • and *P<0.05), right panel. (B) Monocyte counts in peripheral blood (PB), obtained from complete blood count values, are reported as histograms showing the median with interquartile range. Data were analyzed by the Kruskal-Wallis test and showed that the count was reduced in CD8+ STAT3-mutated cases (N=16) in comparison to CD8+ wt cases (N=17). Eleven CD4+ wt and six STAT5B-mutated patients were also evaluated. (C-E) The correlation between the percentage of monocytes in peripheral blood and the leukemic clone count was evaluated in 50 patients with T-cell large granular lymphocyte leukemia (C), subsequently divided into 33 CD8+ (D) and 17 CD4+ (E) cases; Spearman analysis identified a significant inverse relation. T-LGLL: T-cell large granular lymphocyte leukemia; PB: peripheral blood; HC: healthy controls; mut: mutated; pts: patients; LGL: large granular lymphocytes.

levels of IL-6 expression (for T-LGLL patients, P<0.05 vs. intermediate and P<0.001 vs. non-classical monocytes), as previously described.24 Intermediate monocytes were the cells with the highest IL-6 production (for T-LGLL patients, P<0.001 vs. non-classical monocytes) (Figure 4A panel on the right). Although its precise cell source is still elusive, CCL5 expression was observed to be high in plasma and in patients’ PBMC.16 In our cohort of patients, leukemic LGL was pinpointed as the principal source of CCL5, as shown by higher mRNA level in comparison to its relative negative fraction (P<0.0001) (Figure 4B). Our results also indicated that CCL5 was quantitatively more expressed by CD8+ T-LGLL as compared to CD4+ T-LGLL and HC (P<0.05 and P<0.001, respectively) ) (Figure 4B). In an extended cohort of patients, clustering patients according to STAT3 and STAT5B mutations showed consistently higher CCL5 mRNA expression levels for both CD8+ wt (P<0.05 vs. HC) and CD8+ STAT3-mutated T-LGL (P<0.01 vs. HC) patients (Figure 4C). Monocytes mediate survival of T-cell large granular lymphocyte leukemia cells through a cytokine interchange To assess the impact of monocytes on the survival of the leukemic clone, in vitro cultures of PBMC with or without monocytes derived from the same blood sample from HC and T-LGLL patients were set up. When cultured with monocytes, leukemic LGL had a longer lifespan, at the different timepoints tested, in comparison to the LGL from HC (P<0.05) (Figure 5A). On the other hand, the survival of clonal LGL without monocytes was comparable to that of LGL from HC and significantly lower at 3 and 5 days in comparison to leukemic LGL cultured with monocytes (P<0.01) (Figure 5A). An extended cohort was used to get further insights into the role of monocytes in leukemic LGL survival among the different disease subgroups. When assessing the difference in percentage of apoptosis between PBMC without monocytes and PBMC with monocytes, an increased leukemic LGL mortality upon monocyte depletion, with respect to HC, was observed in all subsets with the only exception of the CD4+

wt subgroup (P<0.001 for CD8+ wt, P<0.01 for CD8+ STAT3mutated and P<0.001 for CD4+ STAT5B-mutated cases) (Figure 5B). Based on the above reported CCL5 overexpression in leukemic LGL and the high IL-6 expression on intermediate and non-classical monocytes (Figure 4), we hypothesized a putative monocyte-LGL communication loop. This suggestion is supported by the reduced expression of CCR5 on leukemic cells, previously described for neutropenic LGLL cases18 and herein reported mainly for CD8+ STAT3mutated patients, and its variable expression on monocytes, with a trend to a higher expression on CD8+ STAT3-mutated patients’ monocytes (Online Supplementary Figure S10A, B). Patients’ monocytes stimulated in vitro for 12 h with CCL5 did not express IL-6 mRNA (Figure 5C). However, a significantly higher IL-6 secretion was observed in monocytes obtained from T-LGLL patients in comparison to CCL5-treated monocytes from HC (P<0.05) (Figure 5D). Clustering for disease subsets did not reveal a subgroup specificity (Online Supplementary Figure S10C).

Discussion In this study we provide evidence that a pro-inflammatory background plays a defined role in sustaining T-cell proliferation in T-LGL disorders. In particular, in CD8+ STAT3-mutated LGLL patients we demonstrated a skewed Th17/Treg ratio and an abnormal distribution of monocyte populations characterized by reduced classical and increased intermediate and non-classical monocytes. These latter cells release high levels of IL-6 following CCL5 stimulation. We also demonstrated that T-LGLL patients’ monocytes take advantage from abnormal intracellular signaling pathways, as compared to monocytes from HC. These data on the ability of the leukemic clone to shape the environment in favor of its own proliferation open a new paradigm in T-LGLL. Accumulating evidence indicates that the intrinsic features of the LGL clone characterize the kinetics of the expanding population and contribute to define the disease subgroups.1,3-5,11 Clustering T-LGLL patients according to immunophenotype and STAT mutational status allowed us

Haematologica | 109 January 2024

168

ARTICLE - Role of pro-inflammatory cells in T-LGLL

C. Vicenzetto et al.

to demonstrate different types of signatures in relationship to the characteristics of non-leukemic cell populations, thus offering additional tools to separate disease

subsets and suggesting peculiar pro-inflammatory mechanisms sustaining the pathogenesis of CD8+ and CD4+ TLGLL. In fact, a skewed Th17/Treg ratio was a distinctive

Figure 3. Activation of pathways in monocytes from patients with T-cell large granular lymphocyte leukemia. (A-C) Protein activation status and basal expression were assessed by western blot analysis in eight healthy controls and 30 patients with T-cell large granular lymphocyte leukemia (7 CD8+ wildtype, 7 CD8+ STAT3-mutated, 9 CD4+ wildtype and 7 CD4+ STAT5B-mutated). Densitometric values were subjected to analysis of variance or Kruskal-Wallis tests, according to the distribution of the data. Data are reported as histograms showing the mean with standard error (A) or median with interquartile range (B, C). (A) STAT3 (densitometric data were obtained by merging the optical density of the isoforms α and β) was more active in monocytes of CD8+ STAT3-mutated cases, whereas (B) p65 was more phosphorylated in CD8+ wildtype and CD4+ patients. (C) CD4+ STAT5B-mutated cases were characterized by a higher activation of ERK (densitometric data were obtained by merging the optical density of the isoforms 1 and 2). wt: wildtype; pts: patients; mut: mutated.

Haematologica | 109 January 2024

169

ARTICLE - Role of pro-inflammatory cells in T-LGLL

C. Vicenzetto et al.

feature in CD8+ and CD4+ T-LGLL patients, CD8+ T-LGLL being characterized by a higher percentage of Th17 while a reduced percentage of Treg cells was detectable in CD4+ T-LGLL; these features were maintained during the patients’ follow-up. The presence of a high percentage of Th17 cells was reported to stimulate CD8+ proliferation and increased cytotoxicity, suggesting a pro-proliferative role for these cells in CD8+ T-LGLL.25,26 On the other hand,

the reduction of Treg might be due to the marked plasticity of these cells, which not only downmodulate FoxP3 to become pro-inflammatory cells, but can also express Th1, Th17 or Th2 master regulators, turning out as helper T celllike Treg.27 All these findings might contribute to the proinflammatory environment needed for LGL proliferation. Looking at each T-LGLL subgroup in more detail, we identified that CD8+ STAT3-mutated and neutropenic pa-

Figure 4. IL-6 and CCL5 expression in T-cell large granular lymphocyte leukemia. (A-C) Data are reported as histograms showing the mean with standard error. (A) The pattern of IL-6 expression, evaluated by flow cytometry in the monocyte subtypes in seven healthy controls and eight patients with T-cell large granular lymphocyte leukemia (T-LGLL), resulted significantly lower in classical monocytes for both groups after two-way analysis of variance; the decrease refers both to the percentage of IL-6-positive cells (left panel) and to the mean fluorescence intensity (right panel). (B) CCL5 expression was assessed by quantitative polymerase chain reaction and results are reported as arbitrary units (AU). Leukemic cells (LGL+) were immunomagnetically purified from CD8+ T-LGLL patients (N=6); higher levels of CCL5 were detected in LGL derived from CD8+ T-LGLL, by paired t test, as compared to the non-leukemic fraction (LGL–), than in LGL from healthy controls (N=3) and CD4+ cases (N=7). (C) The overexpression was confirmed in an enlarged cohort and was consistent in CD8+ wildtype (N=10) and STAT3-mutated (N=9) cases. Twelve CD4+ wildtype T-LGLL and nine CD4+ STAT5B-mutated patients were also evaluated. MFI: mean fluorescence intensity; AU: arbitrary units; pts: patients; LGL: large granular lymphocytes; HC: healthy controls; wt: wildtype; mut: mutated. Haematologica | 109 January 2024

170

ARTICLE - Role of pro-inflammatory cells in T-LGLL

C. Vicenzetto et al.

Figure 5. Evaluation of putative communications between monocytes and the leukemic clone in patients with T-cell large granular lymphocyte leukemia. (A) Monocytes from patients with T-cell large granular lymphocyte leukemia (T-LGLL) mediate the survival of the large granular lymphocytes (LGL) (evaluated by flow cytometry) in vitro. Patients’ LGL (total N=10; 3 CD8+ wildtype [wt], 4 CD8+ STAT3-mutated and 3 CD4+ STAT5Bmutated) cultured in peripheral blood mononuclear cells (PBMC) had a longer lifespan in comparison to healthy controls' LGL cultured in PBMC (HC PBMC, N=3; *P<0.05), HC LGL cultured with PBMC immunomagnetically depleted of monocytes (HC CD14–; •P<0.05) and leukemic LGL cultured without monocytes (T-LGLL CD14–; ♦♦P<0.01). Data are reported as a scatter plot showing the mean with standard error. (B) In an extended cohort of patients, the difference of percentage of LGL apoptosis between in vitro cultures of PBMC without monocytes and PBMC with monocytes was determined by two-way analysis of variance and data are reported as histograms showing mean with standard error. A significant difference with respect to HC (N=6) was observed for CD8+ T-LGLL patients, either STAT3 wt (N=3) or mutated (N=7), and for CD4+ STAT5B-mutated (N=4) cases; conversely, apoptosis levels for CD4+ STAT5B wt patients (N=3) were consistent with those for HC. (C) Immunomagnetically purified monocytes and PBMC deprived of monocytes and LGL, obtained from ten patients (4 CD8+ wt, 2 CD8+ STAT3-mutated, 1 CD4+ STAT5 wt and 3 CD4+ STAT5B-mutated) and five HC were cultured for 12 h under CCL5 stimulation (100 ng/mL). IL-6 induction of expression was evaluated by quantitative polymerase chain reaction and analyzed by t tests. Data are reported as histograms showing the mean ratio between treated and not treated conditions (T/NT) with the standard error. (D) Data obtained by enzyme-linked immunosorbent assay were analyzed by the Mann-Whitney test and are reported as histograms showing the median with interquartile range of T/NT values. A significantly higher IL-6 secretion was observed for TLGLL patients (N=16 in total: 4 CD8+ wt, 4 CD8+ STAT3-mutated, 4 CD4+ wt and 4 CD4+ STAT5B-mutated) in comparison to HC (N=7). mut: mutated; h: hours; d: days.

Haematologica | 109 January 2024

171

ARTICLE - Role of pro-inflammatory cells in T-LGLL

C. Vicenzetto et al.

tients are characterized by the highest number of “autoimmune-like” alterations, including an increased Th17/Treg ratio and the skewed monocyte distribution and count that are typically detected in autoimmune conditions.19,20 These findings would suggest that concomitant autoimmune diseases are involved in shaping the peculiar environment detected in CD8+ STAT3-mutated cases. The evidence that the same changes in cellular distribution were also detected in CD8+ STAT3-mutated patients but without association with autoimmune diseases supports the concept that immune system alterations, identified in these patients, are actually an intrinsic trait of the disease and a potential concomitant autoimmune disease might amplify this feature. Monocytes also contribute to the differences in the TLGLL environment, since their number and population distribution differ across T-LGLL subtypes. In particular, the inverse relationship among leukemic clone burden and monocyte percentage resulted stronger for CD8+ wt and CD4+ wt cases; going into detail, CD8+ STAT3-mutated patients were characterized by an increased count of proinflammatory monocytes notwithstanding the general reduction in monocyte count. Despite being reduced during the expansion of the leukemic clone, monocytes resulted relevant for T-LGLL pathogenesis mostly because of their altered activation status. In fact, monocytes from CD8+ STAT3-mutated patients were found to be characterized by increased STAT3 phosphorylation, a well-known marker of environmental and tumor cell hyperactivation which has been reported to accelerate the development of cancerous cells.28,29 This is consistent with our previous data demonstrating that both intrinsic and extrinsic mechanisms contribute to maintain the JAK/STAT pathway aberrantly activated in T-LGLL patients.13 On the other hand, these findings further highlight the central role of STAT3 activation in T-LGLL pathogenesis. In fact, monocytes were proven to contribute to leukemic cell survival in the majority of T-LGLL subsets, i.e., CD8+ (both STAT3mutated and wt) and CD4+ STAT5B-mutated cases. In fact, monocytes derived from patients not STAT3-mutated (i.e., CD8+ wt and CD4+), were found to be equipped with strong p65 activation. The NFκB and ERK hyperactivation we detected in monocytes from CD4+ STAT5B-mutated patients suggests that NFκB signaling mediates inflammation through cellular senescence in non-classical monocytes.24 Together with the identification of a higher cell-to-cell contact between CD8+ leukemic T-LGL and monocytes, these findings indicate that pro-inflammatory cells could generate discrete settings specific for each TLGLL subgroup, even within the indolent form of T-LGLL.30 The high IL-6 release by intermediate and non-classical monocytes and the overexpression of CCL5 by CD8+ leukemic cells lead to the suggestion that monocytes mediate T-LGL survival through the interchange of cytokines,

especially IL-6 and CCL5. This mechanism is likely to be particularly active in symptomatic patients, since they have high IL-6-secreting monocytes. This adds importance to the role of IL-6 in the pathogenesis of T-LGLL, which was previously described by our group13 and recently linked to the efficacy of alemtuzumab in T-LGLL.31 Moreover, we show a putative pathogenic mechanism of CCL5 in T-LGLL, since its expression was recently associated with CD8+ STAT3-mutated LGL with high NKG2D32 and consistently herein demonstrated to be more expressed in CD8+ clones than in controls and CD4+ cases. Nonetheless, Th17 cells, being increased in CD8+ STAT3-mutated cases, are likely to be involved in the monocyte-LGL loop; it has been reported that this T-cell subset could stimulate CCL5 expression in cytotoxic T lymphocytes.33,34 Accordingly, we noticed a tendency of CCL5 expression induction after stimulation of purified LGL with IL-17A for 24 h (Online Supplementary Figure S11B), suggesting the rationale for the use of IL-17 inhibitors in the treatment of T-LGLL patients, as recently proposed.35 We believe that the exchange of IL-6 and CCL5 between monocytes and leukemic cells is one of several ways of communication between leukemic cells and environmental cells. Many different cytokines have already been described as relevant in T-LGLL, such as PDGF, IL-18 and IL-15, this last recently involved in the accumulation of CD8+ LGL cells in models of STAT3 germinal gain-of-function mutation.16,17,32,36 For these reasons, a better definition of the pro-inflammatory mechanisms, possibly also considering the bone marrow microenvironment, might clarify why, in the presence of different immunophenotypes and STAT somatic mutations, LGL leukemic clones can shape a peculiar environment. Bone marrow analysis is nowadays not currently required in clinical practice in T-LGLL patients.1 This constraint causes a limitation to our results and their interpretation, the evaluation of cell populations being restricted to the peripheral blood. It is, however, plausible that the changes we observed in peripheral blood reflect what is occurring in tissue microenvironment(s). In the bone marrow microenvironment, dendritic cells are in close contact with LGL and they have been claimed to represent the setting in which the putative antigen presentation takes place.37 Moreover, it has also been suggested that dendritic cells may be the relevant presenting cells that trigger the clonal proliferation and maintain LGL expansion by cytokine production, in particular IL-15,38 a cytokine that has been found to induce T-LGLL in transgenic mice.39 This work unveils the role of non-leukemic cells in T-LGLL and demonstrates that T-LGLL subgroups can be further differentiated not only according to biological hallmarks of leukemic cells and/or clinical aspects but also considering environmental features. The potential interactive pathway between leukemic cells, monocytes and Th17

Haematologica | 109 January 2024

172

ARTICLE - Role of pro-inflammatory cells in T-LGLL

C. Vicenzetto et al.

lymphocytes we described in the symptomatic and treatment-requiring patients (i.e., mainly CD8+ STAT3-mutated) might pave the way for the design of new therapeutic strategies, in view of the evidence that several anticancer drugs have been shown to reprogram monocytes in a tumor microenvironment.40 Furthermore, since the communication loop between STAT3-mutated leukemic cells and monocytes might be mediated by the activation of CCR5, with the involvement of Th17 cells, the use of CCR5 or Th17 antagonists might represent an interesting option for the treatment of T-LGLL patients. The environmental alterations identified in non-symptomatic patients indicate that also in these cases leukemic T-LGL rely on proinflammatory stimuli, thus giving rise to new questions on the role of the microenvironment in the pathogensis of TLGLL and its targetability for therapeutic purposes.

Contributions CV, GS, and RZ conceived the study. CV, VRG, ER, AT, and GC purified the cells from patients. CV performed the experiments for the study. CV, SC, and VT performed the flow cytometry analysis. GB, MF, LT, and RZ provided the patients’ clinical features. CV wrote the paper. RZ and GS critically reviewed the manuscript. All authors revised the manuscript.

Disclosures No conflicts of interest to disclose.

Data-sharing statement Data are available upon request.

Funding This work was supported by a fellowship related to a project for rare diseases (DIMAR, from the Department of Medicine, Università degli Studi di Padova) and by a grant from the Associazione Italiana per la Ricerca contro il Cancro (AIRC, Milan) (IG #20216 to GS). VRG was supported by a fellowship from AIRC (IT #26815).

References 12. Epling-Burnette PK, Liu JH, Catlett-Falcone R, et al. Inhibition of STAT3 signaling leads to apoptosis of leukemic large granular lymphocytes and decreased Mcl-1 expression. J Clin Invest. 2001;107(3):351-362. 13. Teramo A, Gattazzo C, Passeri F, et al. Intrinsic and extrinsic mechanisms contribute to maintain the JAK/STAT pathway aberrantly activated in T-type large granular lymphocyte leukemia. Blood. 2013;121(19):3843-3854. 14. Kim D, Park G, Huuhtanen J, et al. STAT3 activation in large granular lymphocyte leukemia is associated with cytokine signaling and DNA hypermethylation. Leukemia. 2021;35(12):3430-3443. 15. Mariotti B, Calabretto G, Rossato M, et al. Identification of a miR-146b-Fas ligand axis in the development of neutropenia in T large granular lymphocyte leukemia. Haematologica. 2020;105(5):1351-1360. 16. Kothapalli R, Nyland S, Kusmartseva I, Bailey R, McKeown T, Loughran T. Constitutive production of proinflammatory cytokines RANTES, MIP-1β and IL-18 characterizes LGL leukemia. Int J Oncol. 2005;26(2)529-535. 17. Isabelle C, Boles A, Chakravarti N, Porcu P, Brammer J, Mishra A. Cytokines in the pathogenesis of large granular lymphocytic leukemia. Front Oncol. 2022;12:849917. 18. Moura J, Rodrigues J, Santos AH, et al. Chemokine receptor repertoire reflects mature T-cell lymphoproliferative disorder clinical presentation. Blood Cells Mol Dis. 2009;42(1):57-63. 19. Yasuda K, Takeuchi Y, Hirota K. The pathogenicity of Th17 cells in autoimmune diseases. Semin Immunopathol. 2019;41(3):283-297. 20. Ma W-T, Gao F, Gu K, Chen D-K. The role of monocytes and macrophages in autoimmune diseases: a comprehensive review. Front Immunol. 2019;10:1140. 21. Gasparini VR, Binatti A, Coppe A, et al. A high definition picture of somatic mutations in chronic lymphoproliferative disorder of natural killer cells. Blood Cancer J 2020;10(4):42. 22. Fasching P, Stradner M, Graninger W, Dejaco C, Fessler J. Therapeutic potential of targeting the Th17/Treg axis in autoimmune disorders. Molecules. 2017;22(1):134.

1. Alaggio R, Amador C, Anagnostopoulos I, et al. The 5th edition of the World Health Organization classification of haematolymphoid tumours: lymphoid neoplasms. Leukemia. 2022;36(7):1720-1748. 2. Teramo A, Barilà G, Calabretto G, et al. STAT3 mutation impacts biological and clinical features of T-LGL leukemia. Oncotarget. 2017;8(37):61876-61889. 3. Cheon H, Xing JC, Moosic KB, et al. Genomic landscape of TCRαβ and TCRγδ T-large granular lymphocyte leukemia. Blood. 2022;139(20):3058-3072. 4. Semenzato G, Zambello R. Interrogating molecular genetics to refine LGLL classification. Blood. 2022;139(20):3002-3004. 5. Barilà G, Teramo A, Calabretto G, et al. Stat3 mutations impact on overall survival in large granular lymphocyte leukemia: a single-center experience of 205 patients. Leukemia. 2020;34(4):1116-1124. 6. Bhattacharya D, Teramo A, Gasparini VR, et al. Identification of novel STAT5B mutations and characterization of TCRβ signatures in CD4+ T-cell large granular lymphocyte leukemia. Blood Cancer J. 2022;1(2)2:31. 7. Teramo A, Barilà G, Calabretto G, et al. Insights into genetic landscape of large granular lymphocyte leukemia. Front Oncol. 2020;10:152. 8. Lamy T, Moignet A, Loughran TP Jr. LGL leukemia: from pathogenesis to treatment. Blood. 2017;129(9):1082-1094. 9. Schwaneck EC, Renner R, Junker L, et al. Prevalence and characteristics of persistent clonal T cell large granular lymphocyte expansions in rheumatoid arthritis: a comprehensive analysis of 529 patients. Arthritis Rheumatol. 2018;70(12):1914-1922. 10. Savola P, Brück O, Olson T, et al. Somatic mutations in Felty syndrome: an implication for a common pathogenesis with large granular lymphocyte leukemia. Haematologica. 2018;103(2):304-312. 11. Koskela HLM, Eldfors S, Ellonen P, et al. Somatic STAT3 mutations in large granular lymphocytic leukemia. N Engl J Med. 2012;366(20):1905-1913.

Haematologica | 109 January 2024

173

ARTICLE - Role of pro-inflammatory cells in T-LGLL

C. Vicenzetto et al.

23. Hildebrand D, Uhle F, Sahin D, Krauser U, Weigand MA, Heeg K. The interplay of Notch signaling and STAT3 in TLR-activated human primary monocytes. Front Cell Infect Microbiol. 2018;8:241. 24. Ong S-M, Hadadi E, Dang T-M, et al. The pro-inflammatory phenotype of the human non-classical monocyte subset is attributed to senescence. Cell Death Dis. 2018;9(3):226. 25. Falivene J, Ghiglione Y, Laufer N, et al. Th17 and Th17/Treg ratio at early HIV infection associate with protective HIV-specific CD8+ T-cell responses and disease progression. Sci Rep. 2015;5:11511. 26. Ankathatti Munegowda M, Deng Y, Mulligan SJ, et al. Th17 and Th17-stimulated CD8+ T cells play a distinct role in Th17induced preventive and therapeutic antitumor immunity. Cancer Immunol Immunother. 2011;60(10):1473-1484. 27. Dominguez-Villar M, Hafler DA. Regulatory T cells in autoimmune disease. Nat Immunol. 2018;19(7):665-673. 28. Yu H, Pardoll D, Jove R. STATs in cancer inflammation and immunity: a leading role for STAT3. Nat Rev Cancer. 2009;9(11):798-809. 29. Wu W-Y, Li J, Wu Z-S, Zhang C-L, Meng X-L. STAT3 activation in monocytes accelerates liver cancer progression. BMC Cancer. 2011;11:506. 30. Huuhtanen J, Bhattacharya D, Lönnberg T, et al. Single-cell characterization of leukemic and non-leukemic immune repertoires in CD8 T-cell large granular lymphocytic leukemia. Nat Commun. 2022;13(1):1981. 31. Gao S, Wu Z, Arnold B, et al. Single-cell RNA sequencing coupled to TCR profiling of large granular lymphocyte leukemia T cells. Nat Commun. 2022;13(1):1982. 32. Masle-Farquhar E, Jackson KJL, Peters TJ, et al. STAT3 gain-offunction mutations connect leukemia with autoimmune disease

by pathological NKG2Dhi CD8+ T cell dysregulation and accumulation. Immunity. 2022;55(12):2386-2404. 33. Lee Y, Awasthi A, Yosef N, et al. Induction and molecular signature of pathogenic TH17 cells. Nat Immunol. 2012;13(10):991-999. 34. Peng X, Xiao Z, Zhang J, Li Y, Dong Y, Du J. IL-17A produced by both γδT and Th17 cells promotes renal fibrosis via RANTESmediated leukocyte infiltration after renal obstruction. J Pathol. 2015;235(1):79-89. 35. Zawit M, Bahaj W, Gurnari C, Maciejewski J. Large granular lymphocytic leukemia: from immunopathogenesis to treatment of refractory disease. Cancers. 2021;13(17):4418. 36. Yang J, Liu X, Nyland SB, et al. Platelet-derived growth factor mediates survival of leukemic large granular lymphocytes via an autocrine regulatory pathway. Blood. 2010;115(1):51-60. 37. Zambello R, Berno T, Cannas G, et al. Phenotypic and functional analyses of dendritic cells in patients with lymphoproliferative disease of granular lymphocytes (LDGL). Blood. 2005;106(12):3926-3931. 38. Zambello R, Facco M, Trentin L, et al. Interleukin-15 triggers the proliferation and cytotoxicity of granular lymphocytes in patients with lymphoproliferative disease of granular lymphocytes. Blood. 1997;89(1):201-211. 39. Mishra A, Liu S, Sams GH, et al. Aberrant overexpression of IL-15 initiates large granular lymphocyte leukemia through chromosomal instability and DNA hypermethylation. Cancer Cell. 2012;22(5):645-655. 40. Casagrande N, Borghese C, Favero A, Vicenzetto C, Aldinucci D. Trabectedin overcomes doxorubicin-resistance, counteracts tumor-immunosuppressive reprogramming of monocytes and decreases xenograft growth in Hodgkin lymphoma. Cancer Lett. 2021;500:182-193.

Haematologica | 109 January 2024

174

ARTICLE - Hematopoiesis

Adenosine signaling inhibits erythropoiesis and promotes myeloid differentiation Mahmoud Mikdar,1* Marion Serra,1* Elia Colin,2 Romain Duval,1,3 Emilie-Fleur Gautier,4 Yann Lamarre,1 Yves Colin,1 Caroline Le Van Kim,1 Thierry Peyrard,1,3 Bérengère Koehl1,5# and Slim Azouzi1,3# Université Paris Cité and Université des Antilles, INSERM, BIGR, F-75014 Paris; 2Laboratory of Molecular Mechanisms of Hematologic Disorders and Therapeutic Implications, Imagine Institute, UMR_ S1163, Inserm, Université Paris Cité, Paris; 3Centre National de Référence pour les Groupes Sanguins, Établissement Français du Sang (EFS), Ile-de-France, F-75011 Paris; 4Université Paris Cité, UMR_S1016, UMR 8104, Plateforme de Protéomique (3P5), Institut Cochin, Inserm, CNRS, Paris and 5Sickle Cell Disease Center, Hematology Unit, Hôpital Robert Debré, Assistance Publique - Hôpitaux de Paris, F-75019 Paris, France 1

Correspondence: S. Azouzi slim.azouzi@inserm.fr Received: Accepted: Early view:

September 20, 2022. May 10, 2023. May 18, 2023.

MM and MS contributed equally as first authors. BK and SA contributed equally as senior authors.

Abstract Intracellular uptake of adenosine is essential for optimal erythroid commitment and differentiation of hematopoietic progenitor cells. The role of adenosine signaling is well documented in the regulation of blood flow, cell proliferation, apoptosis, and stem cell regeneration. However, the role of adenosine signaling in hematopoiesis remains unclear. In this study, we show that adenosine signaling inhibits the proliferation of erythroid precursors by activating the p53 pathway and hampers the terminal erythroid maturation. Furthermore, we demonstrate that the activation of specific adenosine receptors promotes myelopoiesis. Overall, our findings indicate that extracellular adenosine could be a new player in the regulation of hematopoiesis.

Introduction Hematopoiesis is the process by which all blood cells are produced from hematopoietic stem cells (HSC) in the bone marrow. Erythropoiesis and granulopoiesis are two main branches of hematopoiesis leading, respectively, to the production of red blood cells (RBC) and neutrophils. While both erythroid and myeloid differentiation derive from the same common myeloid progenitor (CMP), the interplay between the two processes is complex and tightly orchestrated by different intrinsic and extrinsic factors regulating the commitment of progenitor cells towards one cell lineage or the other.1 Terminal erythropoiesis and granulopoiesis occur in erythroblastic islands, which are specialized microenvironments in the bone marrow consisting of a central macrophage surrounded by erythroid and neutrophil precursors.2 These structures constitute a unique cellular microenvironment, and play a crucial role in supporting cell proliferation and differentiation by providing essential nutrients, removing cellular debris, and secreting cytokines and growth factors.3 A growing body of evidence showed that balanced microenvironmental cues taking place in these niches, as well as metabolite transport and signaling, are additional

key regulators of HSC commitment and differentiation.1,4-6 Under stress or pathological conditions such as hypoxia, infection, inflammation, and anemia, the balance of these factors can be perturbed leading to the overproduction of one cell lineage at the expense of the other. Glucose, lipid, amino acid and purinergic metabolisms, as well as nucleotide biosynthesis, have all been shown to be important for erythropoiesis.1,4-6 Furthermore, we have recently shown that the intracellular uptake of adenosine, a pleiotropic and ubiquitous purine nucleoside, is essential for optimal erythroid commitment and differentiation of progenitor cells.7 However, adenosine is also a major signaling molecule that activates cellular signaling pathways through a family of four different G protein-coupled adenosine receptors (AR): A1, A2A, A2B, and A3.8 Adenosine signaling pathways play critical roles in tissue homeostasis by regulating cellular activation, differentiation, and cell cycling.9 Extracellular levels of adenosine are very low at steadystate condition, ranging from 10 to 200 nM, due to the tightly regulated balance between its production from ATP/AMP by ectonucleotidases (CD39, CD73), its rapid cellular uptake by nucleoside transporters, and its degradation by adenosine deaminase (ADA).10,11 However, the extracellular

Haematologica | 109 January 2024

175

ARTICLE - Adenosine signaling impacts human hematopoiesis levels of adenosine can be rapidly elevated up to 100 μM by release of intracellular adenosine from damaged cells or in response to hypoxia and inflammation, as well as tissue injury.10,12 In sickle cell disease (SCD), plasma adenosine levels are increased, and adenosine signaling is enhanced and associated with an increase in RBC sickling through A2B activation13 and a reduction in inflammation via the activation of iNKT cells.14 The erythrocyte equilibrative nucleoside transporter 1 (ENT1) plays a key role in the regulation of plasma adenosine concentrations and thereby controls adenosine signaling in RBC.15 Under hypoxia, increased extracellular adenosine levels activate the adenosine A2B membrane receptor in RBC, which triggers a downregulation of erythrocyte ENT1 through its proteasomal degradation. These reduced ENT1 levels prevent adenosine cellular uptake and allow its rapid accumulation in the plasma leading to a further activation of adenosine receptors.15 Our previous work demonstrates that ENT1-mediated adenosine uptake is important for erythropoiesis.7 However, the role of extracellular adenosine signaling in erythropoiesis remains to be elucidated. Here, we show that adenosine signaling via its plasma membrane receptors, A2B and A3, influences hematopoiesis both by inhibiting erythropoiesis through the activation of p53 and cell cycle arrest, and by promoting myeloid cell differentiation.

Methods Detailed protocols for all methods are provided in the Online Supplementary Methods. Ethics statement Blood bags from healthy donors were obtained from the Établissement Français du Sang (EFS). The study was performed in accordance with the protocols of the Declaration of Helsinki and was approved by the Ethical Committee of the National Institute of Blood Transfusion (INTS), Paris, France. All the participants gave written informed consent. Ex vivo erythropoiesis Ex vivo erythropoiesis of CD34+ cells isolated from the peripheral blood of healthy donors was performed following the protocol described in Online Supplementary Figure S1, in the presence of increasing extracellular adenosine concentrations of up to 200 μM, which corresponds to adenosine levels in stress conditions. Flow cytometry Viability, cell cycle and expression of surface markers were monitored every two or three days using fluorochrome-conjugated antibodies (Online Supplementary Methods).

M. Mikdar et al.

Western blots Proteins were separated by SDS-PAGE, transferred to nitrocellulose membrane, and incubated with the indicated antibodies, followed by HRP anti-rabbit or antimouse immunoglobulin. Proteomic At day 4 of the differentiation phase, adenosine-treated and untreated cells were washed and conserved as dry pellets at -80°C until protein extraction. Proteomic analysis was performed using liquid chromatographycoupled mass spectrometry analysis (nLC-MS/MS), and DIA-NN version 1.8.1 for identification.

Results High dose of exogeneous adenosine inhibits erythroid proliferation and activates the p53 apoptotic pathway In order to investigate the role of adenosine signaling in regulating erythroid commitment and terminal erythroid maturation, we performed ex vivo erythropoiesis of CD34+ cells purified from peripheral blood of healthy donors, and supplemented with different concentrations of adenosine. Our results showed that adenosine decreases erythroid proliferation in a dose-dependent manner (Online Supplementary Figure S2). When treated with 50 μM of adenosine, erythroid precursors exhibited over a 16-fold lower cell growth than untreated controls by day 12 of terminal differentiation (P<0.0001) (Figure 1A). To explore the potential mechanisms by which adenosine inhibited erythroblast proliferation, we assessed whether adenosine induced apoptosis in erythroid cells. Annexin V and Sytox blue staining revealed that 14% and 31% of cells treated with adenosine are in early and late stage of apoptosis, respectively, at day 5 of differentiation (Figure 1B). The apoptotic effect of adenosine is significantly enhanced in immature erythroblasts and extenuated during terminal erythroid maturation (Figure 1C). In line with the increased apoptosis, adenosine treatment was found to induce the activation of the p53 pathway and its downstream target p21, as well as the increase in the cleaved form of caspase 3 (Figure 1D). Moreover, cell cycle assays revealed that adenosine treatment led to cell accumulation in the G1 phase and a delayed progression towards the S and G2/M phases (P<0.05) (Figure 1E). To further understand how adenosine signaling activates the p53 pathway in immature erythroblasts, we performed a whole proteomic analysis of day 4 erythroblasts cultured in the presence of adenosine (n=2). Interestingly, adenosine treatment led to the overexpression of several proteins involved in the p53 apoptotic signaling pathway (APAF1, PRKCD, HTRA2,

Haematologica | 109 January 2024

176

ARTICLE - Adenosine signaling impacts human hematopoiesis TADA3, and DAPK1) and in repair of reactive oxygen species (ROS)-related DNA damage (TP53, DDB2, and TP53I3) (Online Supplementary Figure S3). These data suggest that adenosine signaling may increase DNA damage which leads to the activation of the p53 pathway. The impact of adenosine on DNA has been pre-

M. Mikdar et al.

viously reported in HL-60 cells.16 On the other hand, numerous myeloid lineage-specific proteins were overexpressed in adenosine-treated cells as compared to untreated control cells, which suggests that adenosine modulates erythroid and myeloid differentiation (Online Supplementary Figure S3).

Figure 1. High dose of adenosine inhibits erythroid proliferation and induces cell death through activation of p53 pathway. (A) CD34+ progenitors from peripheral blood of healthy donors were stimulated with erythropoietin in the absence or presence of adenosine (50 μM) at day 0 of the differentiation phase. Growth curves of erythroid precursors from cells treated with adenosine (red line) and control (black line) cells are shown in absolute number of cells during the differentiation phase. Mann-Whitney test: ***P=0.0002, ****P<0.0001; N=9 independent experiments. (B) Percentages of apoptotic, necrotic, and living cells as monitored by Sytox blue and Annexin V co-staining of cells treated or not with adenosine at day 5 of differentiation. (C) Dead cell rate was determined by Sytox blue staining in flow cytometry and percentages of cells were plotted in bar histograms of mean cell % ± Standard Error of Mean. Unpaired t test: *P<0.05, **P<0.005, ns: not significant; N=3 independent experiments. (D) Representative western blot of p53, p21, caspase3, cleaved caspase3 expression and corresponding expression of β-actin in control (CT) and adenosine-treated (ADO) cells at day 5 of differentiation phase. (E) Cell cycle of GPA+-sorted cells was determined by CytoPhase Violet staining at day 4 of differentiation phase. G1, S, and G2/M phase are represented in blue, yellow, and green, respectively. Percentages of cells in each cell cycle phases are plotted in bar histograms. Two-way Anova test: *P=0.04; N=3 independent experiments.

Haematologica | 109 January 2024

177

ARTICLE - Adenosine signaling impacts human hematopoiesis

M. Mikdar et al.

E D

Continued on following page. Haematologica | 109 January 2024

178

ARTICLE - Adenosine signaling impacts human hematopoiesis

M. Mikdar et al.

Figure 2. High level of adenosine induces myeloid differentiation even in presence of erythropoietin. (A) Flow cytometric analyses of GPA were performed to follow the erythroid-lineage specification and erythroid differentiation. GPA expression was analyzed inside 7-AADneg viable cells. Data are shown as peak histograms at indicated days (erythropoietin [EPO] control, black; adenosinetreated cells, red). Erythroblast maturation was monitored via α4-integrin and Band3 levels of GPApos cells. α4-integrinhigh/Band3low cells represent less mature erythroblasts, whereas α4-integrinlow/Band3high cells are further differentiated. Frames represent gating hierarchy. (B) Percentage of GPAneg cells at day 12 of differentiation in cells cultured with or without adenosine. Data shown as mean cell % ± Standard Error of Mean. Unpaired t test: ****P<0.0001; N=6 independent experiments. (C) At day 12 of EPO-induced erythroid differentiation, adenosine-treated cells show a GPAneg subpopulation (outlined by the purple dotted square below). These GPAneg cells were FACS-sorted and analyzed by flow cytometry for expression levels of myeloid surface markers. Corresponding histograms are shown for unstained controls (shaded gray) and specific stainings (purple line) as indicated. GPAneg subpopulation was also MGG-stained; representative pictures are shown. Bar represents 10μm. (D) Characterization and count of MGG-stained adenosine-induced GPAneg cells and control cells plotted in histograms. Percentages of the respective identified-cell types are shown. (E) CD34+ cells were FACS-sorted on the basis of CD36+ at day 7 of the expansion phase, then replaced in culture and stimulated by EPO. At day 12 of EPO-induced erythroid differentiation, expression levels of GPA were compared between the previously CD36+-sorted and non-sorted cells in both control and adenosine-treated cells. Adenosine-induced GPAneg subpopulation is indicated by the bigger purple arrow in non-sorted cells, and the smaller purple arrow in CD36+-sorted cells.

Extracellular adenosine impacts the erythroid differentiation of hematopoietic stem and progenitor cells and promotes myelopoiesis in the presence of erythropoietin To further characterize the effect of high dose of adenosine on erythropoiesis, we monitored the erythroid surface markers of viable cells during erythroblast maturation. Interestingly, adenosine treatment led to the persistence of a non-erythroid subpopulation of GPAneg cells, despite the presence of erythropoietin (EPO) in the culture medium (Figure 2A). Notably, at day 12 of the culture, whereas all control cells expressed GPA, 24±6% of adenosine-treated cells remained GPAneg (P<0.0001) (Figure 2B). Band3 and α4integrin markers showed normal erythroid differentiation of GPApos cells in both conditions (Figure 2A). To further characterize the non-erythroid population found upon adenosine treatment, GPAneg population was FACS-sorted at day 12 of differentiation, and cells were analyzed for the expression of several surface hematopoietic markers. Surprisingly, these cells up-regulated a wide range of myeloid surface markers, such as CD18, CD11a, CD13, and CD33 (Figure 2C). Moreover, May-Grünwald-Giemsa (MGG) staining of this sorted population revealed granular cells at different stages of differentiation (Figure 2D). These findings indicate that extracellular adenosine allows myeloid differentiation even in an erythroid-favoring environment. Next, to determine the nature of the progenitors resulting in the adenosine-induced myeloid cells in the presence of EPO, erythroid progenitors were FACS-sorted on the basis of CD36 at day 7 of the expansion phase. CD36+ cells were then differentiated in the presence of EPO and adenosine. Interestingly, the percentage of GPAneg cells decreased from 30% to 8%, indicating that adenosine induces the myeloid differentiation of CD36neg progenitors (Figure 2E). This finding was further confirmed by the absence of myeloid cells when adenosine was added in the culture three days after the EPO-induced terminal differentiation (data not shown). Altogether, we demonstrate that, at a higher concentration, adenosine inhibits erythroid proliferation and promotes myeloid differentiation of CD36neg progenitors.

Activation of A3 adenosine receptor by Cl-IB-MECA perturbs the balance of erythroid versus myeloid differentiation of hematopoietic stem and progenitor cells In order to understand how extracellular adenosine disturbs erythropoiesis, we first performed similar experiments with other nucleosides, including guanosine, cytidine or uridine. Neither cytidine nor uridine had any effect on cell proliferation, while guanosine slightly attenuated cell growth (Figure 3A). However, differentiation of erythroid precursors was not impacted by any of these nucleosides (data not shown). Exogenously added adenosine can either enter cells through specific nucleoside transporters (primarily via ENT1) or bind to cell membrane adenosinergic receptors.17 To address the pathway by which adenosine perturbs hematopoiesis, we cotreated cells with adenosine and nitrobenzylthioinosine (NTBI, also called NBMPR), a potent bidirectional inhibitor of ENT1. Accordingly, cell proliferation was further decreased in co-treated cells as compared to untreated or adenosine-only-treated cells (Figure 3B), suggesting that the observed effects are mediated through extracellular adenosine signaling. To confirm this finding, we investigated the pharmacological activation of AR among CD34+ human erythroid progenitor cells cultured in the presence of EPO, using selective agonists. Recently published transcriptomic data showed that only A2B and A3 receptors are expressed in CD34+ progenitors and erythroid precursors.18 We first tested the Cl-IB-MECA (2-chloroN6-[3-iodobenzyl]-adenosine-5’-N-methyluronamide), a highly selective agonist of the A3AR.12,19 In accordance with its known anti-proliferative action, CI-IB-MECA decreased the number of erythroid precursors, as shown by a significant decrease in expansion (P<0.01) (Figure 3C) with an increase in cell death (Figure 3D). Notably, at day 12, only 32±2% of CI-IB-MECA-treated cells were viable. The erythroid commitment potential of hematopoietic progenitors was also significantly decreased, as monitored by the analysis of GPA expression (Figure 4A, B). Furthermore, CI-IB-MECA significantly delayed erythroid maturation of the GPApos-erythroid-committed

Haematologica | 109 January 2024

179

ARTICLE - Adenosine signaling impacts human hematopoiesis cells, as monitored by Band3 and α4-integrin surface markers. Notably, at day 7, only 15±6% of CI-IB-MECAtreated erythroblasts acquired Band3 expression, while 51±2% of untreated-cells expressed Band3 (Figure 4B). Also, at day 12 of differentiation, CI-IB-MECA-treated cells failed to acquire Band3 marker or lose α4-integrin

M. Mikdar et al.

marker and up-regulated c-kit surface markers, which is consistent with the block in erythroid maturation (Figure 4C, D). Moreover, the activation of A3 receptor resulted in the differentiation of myeloid cells, even in the presence of EPO, as indicated by the expression of CD33 marker on CI-IB-MECA-treated cells (Figure 4D). Repre-

Figure 3. Cl-IB-MECA decreases erythroid proliferation and induces cell death. (A) CD34+ progenitors were stimulated with erythroproietin (EPO) and treated with either adenosine, guanosine, uridine or cytidine (50 μM) at day 0 of the differentiation phase. Growth curves are shown. A significant difference was found at day 9 between control and guanosine conditions. Mann Whitney test: P=0.0455. (B) The proliferation curve of EPO-treated control cells was compared to adenosine-only treated cells, and adenosine+NTBI (ENT1 inhibitor) co-treated cells. (C) Growth curve of EPO-stimulated control cells (black) versus EPO+25 μM CI-IB-MECA-treated cells (purple) shown in absolute number of cells during the differentiation phase. Mann-Whitney test: **P<0.01; N=3 independent experiments. (D) Percentages of apoptotic, necrotic and viable cells were monitored by flow cytometry using Sytox blue and Annexin V co-staining of cells treated or not with CI-IB-MECA, at days 5 and 12 of differentiation. Dead cell rate was determined by Sytox blue staining at different times of differentiation, and percentages of cells were plotted in bar histograms of mean cell % ± Standard Error of Mean. Unpaired t test; *P<0.05, **P<0.01; N=3 independent experiments.

Haematologica | 109 January 2024

180

ARTICLE - Adenosine signaling impacts human hematopoiesis sentative images of MGG staining indicate the presence of a mix of immature erythroblasts and myeloblasts among CI-IB-MECA-treated cells, whereas untreated cells present exclusively erythroblasts at mature stages

M. Mikdar et al.

of differentiation and reticulocytes (Figure 4E). Altogether, these findings indicate that A3AR activation on hematopoietic progenitors leads to myeloid differentiation and erythropoiesis inhibition.

Figure 4. Activation of adenosine receptors by Cl-IB-MECA perturbs the erythroid and myeloid differentiation of hematopoietic stem and progenitor cells. (A) GPA expression was analyzed inside viable cells. Data shown as peak histograms at day 7 of differentiation (erythropoietin [EPO] control, black; CI-IB-MECA-treated cells, purple; unstained control, shaded gray), α4-integrin and Band3 expression levels of GPApos cells were analyzed to monitor erythroblast maturation. Frames represent gating hierarchy. (B) Percentages of GPAhigh and GPAlow cells, GPA intensity of fluorescence and percentages of Band3+ cells in control and CI-IBMECA-treated cells plotted in bar histograms and presented at indicated days with corresponding colors. Unpaired t test: *P<0.05, **P<0.01, ns: not significant; N=3 independent experiments. (C) GPA, and α4-integrin and Band3 expression levels of GPApos cells, at day 12 of differentiation for EPO-stimulated control cells (black) and CI-IB-MECA treated cells (purple). Gray-shaded peaks represent unstained controls; N=2 independent experiments. (D) c-kit and CD33 were monitored at day 12 of differentiation for EPO-stimulated control cells (black) and CI-IB-MECA-treated cells (purple); N=2 independent experiments. (E) May-Grünwald Giemsa-stained cytospins of control and CI-IB-MECA-treated cells at day 12 of differentiation. Black arrow corresponds to a large eosinophil cell. Bar represents 10 μm. Haematologica | 109 January 2024

181

ARTICLE - Adenosine signaling impacts human hematopoiesis Activation of A2B adenosine receptor impairs erythroid differentiation Next, we performed similar experiments using the selective A2B agonist BAY 60-658320 to investigate the effect of this receptor activation on erythropoiesis. BAY 60-6583 did not affect the erythroblast proliferation (Figure 5A). Although A2B activation did not impact ery-

M. Mikdar et al.

throid commitment (since all cultured cells were GPA+), BAY 60-6583 treatment strongly abrogated erythroblast maturation (Figure 5B). Notably, the expression of erythroid markers including GPA and Band3 were delayed in the presence of the BAY 60-6583 (Figure 5C-E). In addition, erythroid differentiation blockage of treated cells was further confirmed by attenuated erythroblast

Continued on following page. Haematologica | 109 January 2024

182

ARTICLE - Adenosine signaling impacts human hematopoiesis

M. Mikdar et al.

Figure 5. Activation of A2B adenosine receptor impairs erythroid differentiation. (A) CD34+ progenitors from peripheral blood of healthy donors were stimulated with erythropoietin (EPO) in the absence or presence of BAY60-6583 (25 μM) at day 0 of the differentiation phase. Growth curves of erythroid precursors from control cells (black) and BAY60-6583-treated cells (blue) are shown in absolute number of cells during the differentiation phase; N=4 independent experiments. (B) Expression levels of GPA are shown for EPO-stimulated control cells (black) and EPO + BAY60-6583-treated cells (blue), as well as α4-integrin/Band3 profile at days 5 and 12 of erythroid terminal differentiation. Gray-shaded histograms correspond to unstained controls. (C) Percentages of GPAhigh and GPAlow cells are plotted in bar histograms and are presented at indicated days with corresponding colors. Unpaired t test: **P<0.01, ***P<0.001; N=4 independent experiments. (D) GPA intensity of fluorescence during erythroid differentiation is shown. Unpaired t test: *P<0.05, **P<0.01; N=4 independent experiments. (E) Percentages of Band3+ cells are shown at indicated days. Data are presented as mean ± Standard Error of Mean (SEM). Unpaired t test: *P≤0.05, **P≤0.007, ***P≤0.0009; N=4 independent experiments. (F) May-Grünwald Giemsa-stained cytospins of control and BAY60-6583-treated cells at day 12 of differentiation. Bar represents 10 μm. (G) Dead cell rate was determined by Sytox blue staining in flow cytometry and percentages of cells were plotted in bar histograms of mean cell %±SEM; N=4 independent experiments. Unpaired t test: *P<0.05, ns: not significant.

maturation and enucleation (Figure 5F), and increased apoptosis at days 9 and 12 (Figure 5G).

Discussion Here, we show that adenosine signaling plays an important role in modulating erythroid proliferation and differentiation via the activation of AR. At high concentration, adenosinemediated decrease in proliferation is most likely due to the activation of A3, while the attenuation of erythroid maturation is both mediated by A3 and by A2B receptors. However, A3 seems to be the only adenosine receptor promoting myeloid differentiation. Our findings indicate that high dose of exogenous adenosine attenuates erythroblast proliferation via cell cycle arrest at G1 and increases apoptosis upon the activation of the p53 pathway. Interestingly, it has

been shown that adenosine signaling contributes to p53induced cell death,21 as well as G1 cell cycle arrest in breast cancer stem cells22 and ovarian cancer cell line.23 Furthermore, adenosine-mediated decreased cell proliferation is partly due to A3 receptor since the pharmacological activation of this receptor with its highly selective agonist ClIB-MECA attenuated erythroid cell proliferation and increased cell death. In addition, we showed that extracellular adenosine induces myeloid differentiation even in the presence of erythropoietin through A3 receptor. Indeed, myeloid cells at different stages of granular cell differentiation were found upon adenosine treatment and A3-specific activation led to the presence of CD33+ and ckit+ cells in an erythroid-favoring environment. These findings suggest that adenosine signaling through A3 influences the survival and differentiation of myeloid progenitors. It is important to note that A3 activation via its selective agonist IB-MECA has pre-

Haematologica | 109 January 2024

183

ARTICLE - Adenosine signaling impacts human hematopoiesis viously been related to the stimulation of granulopoiesis,24,25 and A3 knockout mice display a reduction in mature granulocytes and monocytes compared to wild-type mice.26 To conclude, our findings place adenosine signaling as a new player in hematopoiesis regulation by promoting granular cell differentiation, and inhibiting erythroid proliferation and differentiation. Interestingly, adenosine signaling was reported to play a detrimental role in the pathophysiology of SCD, which is characterized by chronic anemia and unexplained high level of neutrophils.27 Interestingly, an SCD mouse model treated with the polyethylene glycol-modified adenosine deaminase (PEG-ADA) (an FDA-approved drug used for the treatment of human ADA genetic deficiency) exhibited significantly reduced levels of plasma adenosine, decreased leukocyte counts, and increased RBC counts, suggesting that adenosine signaling could contribute to dysregulated hematopoiesis in SCDlike mice.13 Future studies using cohorts of SCD patients will be necessary to address the question as to whether adenosine signaling is involved in an abnormal hematopoiesis in this disease.

M. Mikdar et al.

Disclosures No conflicts of interest to disclose. Contributions MM, BK and SA conceived the project and obtained the grant. MM and SA designed the research study. MM, MS, EC, RD, EFG and BK performed the experiments. MM, MS, BK, EC and SA analyzed the data. YL, YC, CLK and TP edited the manuscript. MM, MS and SA wrote the paper. Acknowledgments This work was supported by the Institut National de la Santé et de la Recherche Médicale (Inserm), the Établissement Français du Sang, and the Laboratory of Excellence GR-Ex, reference ANR-11-LABX-0051. The labex GR-Ex is funded by the IdEx program “Investissements d’avenir” of the French National Research Agency, reference ANR-18-IDEX-0001. Data-sharing statement The data supporting our findings are available upon request to the corresponding author.

References 1. Oburoglu L, Romano M, Taylor N, Kinet S. Metabolic regulation of hematopoietic stem cell commitment and erythroid differentiation. Curr Opin Hematol. 2016;23(3):198-205. 2. Romano L, Seu KG, Papoin J, et al. Erythroblastic islands foster granulopoiesis in parallel to terminal erythropoiesis. Blood. 2022;140(14):1621-1634. 3. Chasis JA, Mohandas N. Erythroblastic islands: niches for erythropoiesis. Blood. 2008;112(3):470-478. 4. Oburoglu L, Tardito S, Fritz V, et al. Glucose and glutamine metabolism regulate human hematopoietic stem cell lineage specification. Cell Stem Cell. 2014;15(2):169-184. 5. Huang NJ, Lin YC, Lin CY, et al. Enhanced phosphocholine metabolism is essential for terminal erythropoiesis. Blood. 2018;131(26):2955-2966. 6. Gonzalez-Menendez P, Phadke I, Olive ME, et al. Arginine metabolism regulates human erythroid differentiation through hypusination of eIF5A. Blood. 2023;141(20):2520-2536. 7. Mikdar M, González-Menéndez P, Cai X, et al. The equilibrative nucleoside transporter ENT1 is critical for nucleotide homeostasis and optimal erythropoiesis. Blood. 2021;137(25):3548-3562. 8. Borea PA, Varani K, Vincenzi F, et al. The A 3 adenosine receptor: history and perspectives. Pharmacol Rev. 2015;67(1):74-102. 9. Borea PA, Gessi S, Merighi S, Varani K. Adenosine as a multisignalling guardian angel in human diseases: when, where and how does it exert its protective effects? Trends Pharmacol Sci. 2016;37(6):419-434. 10. Fredholm BB. Adenosine, an endogenous distress signal, modulates tissue damage and repair. Cell Death Differ. 2007;14(7):1315-1323. 11. Schrier SM, Van Tilburg EW, Van der Meulen H, et al. Extracellular adenosine-induced apoptosis in mouse neuroblastoma cells studies on involvement of adenosine receptors and adenosine uptake. Biochem Pharmacol. 2001;61(4):417-425.

12. Burnstock G, Verkhratsky A. Receptors for purines and pyrimidines. Purinergic Signal Nerv Syst. 2012;50(3):119-244. 13. Zhang Y, Dai Y, Wen J, et al. Detrimental effects of adenosine signaling in sickle cell disease. Nat Med. 2011;17(1):79-86. 14. Field JJ, Lin G, Okam MM, et al. Sickle cell vaso-occlusion causes activation of iNKT cells that is decreased by the adenosine A2A receptor agonist regadenoson. Blood. 2013;121(17):3329-3334. 15. Song A, Zhang Y, Han L, et al. Erythrocytes retain hypoxic adenosine response for faster acclimatization upon re-ascent. Nat Commun. 2017;8:14108. 16. Tanaka Y, Yoshihara K, Tsuyuki M, Kamiya T. Apoptosis induced by adenosine in human leukemia HL-60 cells. Exp Cell Res. 1994;213(1):242-252. 17. Di Virgilio F, Adinolfi E. Extracellular purines, purinergic receptors and tumor growth. Oncogene. 2017;36(3):293-303. 18. Ludwig LS, Lareau CA, Bao EL, et al. Transcriptional states and chromatin accessibility underlying human erythropoiesis. Cell Rep. 2019;27(11):3228-3240. 19. Bhanu NV, Aerbajinai W, Gantt NM, et al. Cl-IB-MECA inhibits human erythropoiesis. Br J Haematol. 2007;137(3):233-236. 20. Alnouri MW, Jepards S, Casari A, et al. Selectivity is speciesdependent: characterization of standard agonists and antagonists at human, rat, and mouse adenosine receptors. Purinergic Signal. 2015;11(3):389-407. 21. Long JS, Crighton D, Prey JO, et al. Article extracellular adenosine sensing-a metabolic cell death priming mechanism downstream of p53. Mol Cell. 2013;50(3):394-406. 22. Jafari SM, Joshaghani HR, Panjehpour M, Aghaei M, Zargar Balajam N. Apoptosis and cell cycle regulatory effects of adenosine by modulation of GLI-1 and ERK1/2 pathways in CD44+ and CD24− breast cancer stem cells. Cell Prolif. 2017;50(4):e12345. 23. Shirali S, Aghaei M, Shabani M, et al. Adenosine induces cell

Haematologica | 109 January 2024

184

ARTICLE - Adenosine signaling impacts human hematopoiesis cycle arrest and apoptosis via cyclinD1/Cdk4 and Bcl-2/Bax pathways in human ovarian cancer cell line OVCAR-3. Tumor Biol. 2013;34(2):1085-1095. 24. Hofer M, Pospísil M, Sefc L, et al. Activation of adenosine A(3) receptors supports hematopoiesis-stimulating effects of granulocyte colony-stimulating factor in sublethally irradiated mice. Int J Radiat Biol. 2010;86(8):649-656. 25. Hofer M, Vacek A, Pospíšil M, et al. Activation of adenosine A3 receptors potentiates stimulatory effects of IL-3, SCF, and GM-

M. Mikdar et al.

CSF on mouse granulocyte-macrophage hematopoietic progenitor cells. Physiol Res. 2009;58(2):247-252. 26. Hofer M, Pospíšil M, Dušek L, Hoferová Z, Komůrková D. Lack of adenosine A3 receptors causes defects in mouse peripheral blood parameters. Purinergic Signal. 2014;10(3):509-514. 27. Conran N, Saad STO, Costa FF, Ikuta T. Leukocyte numbers correlate with plasma levels of granulocyte-macrophage colony-stimulating factor in sickle cell disease. Ann Hematol. 2007;86(4):255-261.

Haematologica | 109 January 2024

185

ARTICLE - Non-Hodgkin Lymphoma

Hypomethylating agent decitabine sensitizes diffuse large B-cell lymphoma to venetoclax Fen Zhu,1 Jennifer L. Crombie,1 Wei Ni,1 Nguyet-Minh Hoang,2 Swati Garg,1 Liam Hackett,1 Stephen J. F. Chong,1 Mary C. Collins,1 Lixin Rui,2 James Griffin1 and Matthew S. Davids1

Correspondence: M.S. Davids Matthew_Davids@DFCI.HARVARD.EDU

Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA and Department of Medicine and Carbone Cancer Center, University of Wisconsin School of Medicine and Public Health, Madison, WI, USA 1

Received: Accepted: Early view:

March 29, 2023. July 25, 2023. August 3, 2023.

Abstract Despite recent advances in the therapy of diffuse large B-cell lymphoma (DLBCL), many patients are still not cured. Therefore, new therapeutic strategies are needed. The anti-apoptotic B-cell lymphoma 2 (BCL2) gene is commonly dysregulated in DLBCL due to various mechanisms such as chromosomal translocation t(14;18)(q32;q21) and copy number alterations; however, targeting BCL-2 with the selective inhibitor, venetoclax, led to response in only a minority of patients. Thus, we sought to identify a rational combination partner of venetoclax to improve its activity against DLBCL cells. Utilizing a functional assay, dynamic BH3 profiling, we found that the DNA hypomethylating agent decitabine increased mitochondrial apoptotic priming and BCL-2 dependence in DLBCL cells. RNA-sequencing analysis revealed that decitabine suppressed the pro-survival PI3K-AKT pathway and altered the mitochondria membrane composition in DLBCL cell lines. Additionally, it induced a DNA damage response and increased BAX and BAK activities. The combination of decitabine and venetoclax synergistically suppressed proliferation of DLBCL cells both in vitro and in vivo in a DLBCL cell line-derived xenograft mouse model. Our study suggests that decitabine plus venetoclax is a promising combination to explore clinically in DLBCL.

Introduction With recent advances in treatment, approximately two thirds of patients with diffuse large B-cell lymphoma (DLBCL) can be cured with currently available therapies; however, new treatment strategies are needed to improve outcomes for those who are not cured with existing regimens. Recent comprehensive genomic analyses1–3 have identified genomic heterogeneity in DLBCL beyond the traditional activated B-cell-like DLBCL (ABC DLBCL) and germinal center B-cell-like DLBCL (GCB DLBCL) genotypes.4,5 Notably, some of these newly identified genomic subtypes of DLBCL involve apoptotic signaling, suggesting that it may be promising to target this pathway therapeutically.1–3,6 The intrinsic (mitochondrial) pathway of apoptosis is regulated by the interaction between the proteins in the Bcell lymphoma/leukemia 2 (BCL-2) family including a group of pro-apoptotic proteins (e.g., multidomain effectors [BAK, BAX], BH3 only activators [BIM, BID], BH3 only

sensitizers [BAD, BIK, BMF, HRK, PUMA, NOXA]) and antiapoptotic proteins such as BCL-2, BCL-xL, MCL-1, BCL-w and BFL-1.7,8 The ratio and interactions among these BCL2 family proteins determine whether cancer cells survive or proceed to apoptosis. In GCB DLBCL, the BCL2 gene is often subjected to chromosomal translocation t(14;18)(q32;q21) which places the BCL2 gene under the control of immunoglobulin heavy chain (IgH) enhancer,6 leading to high BCL-2 expression. Increased expression of BCL-2 can also be seen in ABC DLBCL, likely due to copy number alterations.1–3 Thus, BCL-2 is a potential therapeutic target in several genetic subtypes of DLBCL. Preclinical studies have investigated anti-growth effects of the BCL-2-selective inhibitor, venetoclax, in DLBCL cell lines and cell line-derived xenograft mouse models.9–11 Notably, DLBCL cell lines required a higher dosage of venetoclax to kill the cells than other types of lymphoid malignancies.9,10 In a phase I trial evaluating venetoclax in patients with relapsed or refractory non-Hodgkin lymphoma, although single-agent activity of the drug in

Haematologica | 109 January 2024

186

ARTICLE - Decitabine increases BCL-2 dependence

F. Zhu et al.

DLBCL was observed, the complete response rate was only 18%.12 Given the modest single-agent activity of venetoclax in DLBCL, we sought to identify a rational combination partner that could increase the activity of venetoclax in this disease by specifically shifting the balance of BCL2 family proteins toward a more BCL-2-dependent state. In order to identify such a combination partner, we utilized BH3 profiling, an assay that can assess the mitochondrial apoptotic priming of cells, a measure of how close tumor cells are to the threshold of apoptosis.13 The profile also helps to assess the functional dependence of cells on the various anti-apoptotic BCL-2 family proteins. A variation of this assay, known as dynamic BH3 profiling (DBP), is capable of assessing the change in mitochondrial apoptotic priming and dependence induced by ex vivo drug treatments of interest. With DBP, we found that the hypomethylating agents azacitidine and decitabine could increase mitochondrial apoptotic priming and BCL-2 dependence in DLBCL, suggesting the combination of hypomethylating agents with venetoclax could be a potential therapeutic strategy worth exploring in DLBCL. Of note, the combination of venetoclax and a hypomethylating agent has been approved for use in acute myeloid leukemia (AML), where mechanistic studies demonstrated that azacitidine can downregulate MCL-1 and induce NOXA expression.14,15 Decitabine functions differently from azacitidine, as decitabine is a more specific DNA hypomethylating agent. In addition, decitabine was shown to be more potent than azacitidine in suppressing certain cancer cells at equimolar concentrations16 and outperformed azacitidine in anti-leukemic action in a rat model of myeloid leukemia.16 However, mechanistic studies evaluating the combination of hypomethylating agents and venetoclax have been mostly focused on azacitidine. Thus in this study, we focused on decitabine and investigated whether decitabine could sensitize DLBCL cells to venetoclax, and we explored the mechanisms underlying the activity of this combination in DLBCL.

cated using STR profiling at the Molecular Diagnostics Laboratory in January, 2023. Cell lines were routinely tested for mycoplasma contamination. Double-hit lymphoma (DHL) patient-derived xenograft (PDX) cell lines DW19 and DW20 were obtained from Dr. David Weinstock’s laboratory and cultured in X-VIVO 15 media (Lonza, BE02-060Q, Portsmouth, NH, USA) supplemented with 20 ng/mL Fungizone (Invitrogen 15290018, Waltham, MA, USA). They were isolated from PDX mice available from the Public Repository of Xenografts (PRoX). All cell lines were cultured at 37°C in a 5% CO2 atmosphere. BH3 profiling BH3 profiling was performed by flow cytometry as described previously.13 Cells were incubated with antibodies for at least 24 hours before being analyzed by flow cytometry. Baseline BH3 profiling data were presented as percent cytochrome C loss. DBP data were presented as δ percent priming (cytochrome C loss in response to drug treatment minus cytochrome C loss in response to dimethyl sulfoxide [DMSO]). Xenografts All murine studies were performed according to DFCI Institutional Animal Care and Use Committee approved protocol. NOD/SCID/Il2Rγ-/- (NSG) mice were purchased from the Jackson Laboratory; 5x106 of luciferase-labeled OCILy1 cells were resuspended in 100 μL of phosphate-buffered saline (PBS) and injected via the lateral tail veins of 7-week-old male mice. Five days following tumor inoculation, animals with established disease determined by bioluminescent imaging were divided into four cohorts and treated with: (i) vehicle control; (ii) 0.25 mg/kg decitabine in PBS, intraperitoneal, 5 consecutive days per week; (iii) 50 mg/kg venetoclax in 60% Phosal 50PG, 30% PEG400, 10% ethanol, orally, daily; (iv) both drugs at the indicated doses. Imaging was performed once every week. Treatments lasted for 21 days. Afterwards, the mice were observed for changes in total-body bioluminescence and survival. Detailed information can be found in the Online Supplemental Appendix.

Methods Cell lines and culture The DLBCL cell lines (TMD8, OCI-Ly3, OCI-Ly1, OCI-Ly7, SUDHL4) were kindly provided by the laboratory of Dr. Anthony Letai (Dana-Farber Cancer Institute) and grown in RPMI 1640 media (Gibco 11875119, Billings, MT, USA) supplemented with 10% fetal bovine serum (Gibco 26140079), 100 U/mL penicillin-streptomycin (Gibco 15140163), 2 mM L-glutamine (Gibco 25030081). The venetoclax-resistant OCI-Ly1 cell line (OCI-Ly1R) was obtained from Dr. Catherine Wu’s laboratory17 and maintained in the above medium with 0.5 μM of venetoclax. All cell lines were authenti-

Results Diffuse large B-cell lymphoma cells have heterogenous dependence on anti-apoptotic BCL-2 family proteins As DLBCL is a genetically heterogeneous disease, we used a panel of genetically diverse DLBCL cell lines to explore their dependence on various anti-apoptotic BCL-2 family proteins. We first examined the protein expression level of the anti-apoptotic proteins BCL-2, MCL-1, BFL-1 and BCL-xL in a group of DLBCL cell lines including both ABC DLBCL and GCB DLBCL. We found that more than one

Haematologica | 109 January 2024

187

ARTICLE - Decitabine increases BCL-2 dependence

F. Zhu et al.

anti-apoptotic protein was expressed in each cell line (Figure 1A). Of note, one GCB DLBCL cell line, OCI-Ly7, was BCL-2-deficient while the other cell lines had comparable

BCL-2 expression level (Figure 1A; Online Supplementary Figure S1). In order to investigate the dependence of DLBCL on the

D C

Continued on following page. Haematologica | 109 January 2024

188

ARTICLE - Decitabine increases BCL-2 dependence

F. Zhu et al.

Figure 1. Diffuse large B-cell lymphoma displays dependence on multiple BCL-2 family members. (A) Expression of the indicated BCL-2 family proteins in diffuse large B-cell lymphoma (DLBCL) cell lines by western blotting. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as loading control. (B) Interaction map between anti-apoptotic proteins and the BH3 peptides and mimetic used in the BH3 profiling assay. Orange color indicates high affinity interaction. (C) Baseline BH3 profiling of DLBCL cell lines. (D) Dynamic BH3 profiling (DBP) of DLBCL cell lines treated with single BH3 mimetic for 24 hours, data were presented as δ percent priming. (E) Relative cell proliferation of DLBCL cell lines treated with indicated BH3 mimetics alone or in combination for 24 hours. Cell proliferation was measured by CellTiter-Glo assay. Dashed line indicates 50% inhibition of cell proliferation. Data were normalized to dimethyl sulfoxide-treated cells. Data were presented as mean ± standard deviation, N=4. ABC: activated B-cell-like; GCB: germinal center B cell.

anti-apoptotic BCL-2 family members, we next performed baseline BH3 profiling on DLBCL cell lines utilizing a panel of BH3 mimetic peptides or inhibitors that have affinity for either one or multiple anti-apoptotic proteins (Figure 1B). Baseline BH3 profiling revealed that DLBCL cell lines were responsive to more than one BH3 peptide or BH3 mimetic (Figure 1C), suggesting dependence on multiple anti-apoptotic proteins. Cytochrome C loss in response to venetoclax indicates cellular survival dependence on BCL-2, and the variable responses we observed indicate heterogenous dependence on BCL-2 among DLBCL cell lines. Cell lines that showed strong dependence on BCL-2 (OCI-Ly1, SUDHL4) were also dependent on other anti-apoptotic proteins, such as BFL-1 and MCL-1. Utilizing DBP, we treated DLBCL cells with venetoclax and other BH3 mimetics including S63845 (an MCL-1 inhibitor) and A1331852 (a BCL-xL inhibitor), and found that single BH3 mimetic treatment increased overall mitochondrial apoptotic priming and dependence on other anti-apoptotic proteins (Figure 1D). This finding suggests that some DLBCL cells may not be sensitive to venetoclax alone, as they can switch dependence and rely on other anti-apoptotic proteins for survival. Since S63845 or A1331852 treatment can increase BCL2 dependence in DLBCL cells, we hypothesized that combining venetoclax with S63845 or A1331852 may enhance the cytotoxicity of venetoclax. In order to test that hypothesis, we treated multiple DLBCL cell lines with the indicated BH3 mimetics alone or in combination (Figure 1E), and measured cell proliferation by CellTiter-Glo assay. The response to venetoclax, S63845, and A1331852 was heterogenous, with OCI-Ly1 being most sensitive to venetoclax, and the remaining cell lines (TMD8, OCI-Ly3, SUDHL4 and OCI-Ly7) were naturally resistant to venetoclax. In BCL-2-expressing cell lines, combining venetoclax with S63845 or A1331852 had the most significant inhibitory effects on cell proliferation. Together, these data demonstrate heterogenous dependence of DLBCL cells on the anti-apoptotic BCL-2 family members and suggest that these cells utilize multiple anti-apoptotic proteins for survival. These data are consistent with clinical trials data suggesting that venetoclax monotherapy is not effective for most patients with relapsed/refractory DLBCL.12

Decitabine primes diffuse large B-cell lymphoma cells for apoptosis and synergizes with venetoclax in inducing apoptosis We used DBP to identify therapeutic agents that can enhance response to venetoclax in DLBCL. Both hypomethylating agents azacitidine and decitabine increased mitochondrial apoptotic priming and BCL-2 dependence in DLBCL cells (Figure 2A; Online Supplementary Figure S2). Given that azacitidine and decitabine function via different mechanisms and decitabine was shown to be more potent than azacitidine in preclinical studies,16,18 we decided to focus our study on decitabine. In order to mimic acquired venetoclax resistance that could occur in patients receiving venetoclax treatment, we included in our study a venetoclax acquired-resistant cell line OCI-Ly1R which was generated from the venetoclax-sensitive DLBCL cell line OCI-Ly1.17 We found that decitabine could increase BCL-2 dependence in both venetoclax naturally resistant and acquired-resistant DLBCL cells (Figure 2A). Next, we investigated cell proliferation in response to the combination of decitabine and venetoclax in various DLBCL cell lines. Across all the cell lines tested, decitabine consistently sensitized DLBCL cells to venetoclax, including the venetoclax acquired-resistant cell line OCI-Ly1R that had minimal decrease in proliferation in response to venetoclax treatment alone (Figure 2B-D). The effects of this decitabine plus venetoclax combination were synergistic in reducing cell proliferation (Figure 2E). This combination also significantly induced apoptosis (Online Supplementary Figure S3A) and enhanced caspase 3 cleavage (Online Supplementary Figure S3B). Since our DBP data had also revealed increased BCL-xL dependence in some of the cell lines and increased MCL1 dependence in all the cell lines tested (Figure 2A), we also evaluated whether combining decitabine with A1331852 or S63845 could also achieve synergistic cell killing in certain cell lines. For these experiments, we tested one ABC DLBCL (TMD8), two GCB DLBCL (OCI-Ly1, SUDHL4) and one DHL-PDX cell line DW20. When we treated these cell lines with the combination of decitabine and A1331852, we found synergistic cell killing only in OCILy1 and DW20 (Online Supplementary Figure S4A). The combination of decitabine and S63845 induced synergistic cell killing in OCI-Ly1, TMD8 and DW20 but not SUDHL4 (Online Supplementary Figure S4B).

Haematologica | 109 January 2024

189

ARTICLE - Decitabine increases BCL-2 dependence

F. Zhu et al.

Continued on following page. Haematologica | 109 January 2024

190

ARTICLE - Decitabine increases BCL-2 dependence

F. Zhu et al.

Figure 2. Decitabine increases mitochondrial apoptotic priming and synergizes with venetoclax in suppressing the growth of diffuse large B-cell lymphoma cells. (A) Dynamic BH3 profiling (DBP) of diffuse large B-cell lymphoma (DLBCL) cells treated with dimethyl sulfoxide (DMSO) control or 2.5 μM of decitabine for 2 days. Data were presented as δ percent priming. (B-D) Germinal center B cell (GCB) DLBCL cell lines including venetoclax-sensitive parental OCI-Ly1 cell line, experimentally created venetoclaxresistant cell line OCI-Ly1R and venetoclax naturally resistant SUDHL4 cell line (B), venetoclax naturally resistant activated Bcell-like (ABC) DLBCL cell lines (C) and double-hit lymphoma-patient-derived xenograft (DHL-PDX) cell lines (D) were treated with indicated dosage of decitabine for 3 days, and venetoclax was added on the last day. Cell proliferation was determined by CellTiter-Glo assay. Data were normalized to DMSO-treated cells. Data were presented as mean ± standard deviation, N=4. (E) Combination index (CI) was calculated by Compusyn.

Taken together, the combination of venetoclax and decitabine strongly induced apoptotic cell death and suppressed cell proliferation in DLBCL cells, leading us to further explore the mechanisms underlying this synergy. Decitabine induces DNA damage and regulates the expression and activity of BCL-2 family proteins An established role of decitabine is to induce DNA damage.19 When we treated DLBCL cell lines with decitabine, we observed a depletion of DNMT1 and an increase in phospho-H2AX (s139) level, indicating increased DNA damage formation (Online Supplementary Figure S5A). A previous study suggested a link between DNA damage induction and apoptotic priming in hematopoietic stem cells.20 Therefore, we asked whether the DNA damage response is associated with apoptotic priming in DLBCL cells. We tested a chemotherapy drug doxorubicin, which is known to induce DNA damage but lacks the epigenetic functions of decitabine, in three DLBCL cell lines. By DBP, we observed increases both in mitochondrial apoptotic priming and BCL-2 dependence in response to doxorubicin treatment (Figure 3A), suggesting that induction of DNA damage in DLBCL is sufficient to increase apoptotic priming. With cell cycle analysis, we found cell cycle arrest at the G2/M phase upon decitabine treatment, suggesting the formation of DNA damage during S phase (Online Supplementary Figure S5B). Nevertheless, apoptosis was not effectively induced in these cells. Next, we examined the expression level of BCL-2 family proteins in response to decitabine treatment in OCI-Ly1, SUDHL4 and TMD8 cells. We found downregulation of BFL-1 in all three cell lines and downregulation of BCL-xL in OCI-Ly1 and TMD8 cells. The expression level of BCL2, MCL1, BIM and PUMA remained almost the same (Figure 3B). The activity of BAX and/or BAK markedly increased upon decitabine treatment even though their expression level didn’t change (Figure 3B, C). With a co-immunoprecipitation assay, we observed decreased interaction between BCL-2 and BAX upon decitabine treatment, which may account for the increased activity of BAX (Figure 3D). As DLBCL cells are not sensitive to apoptosis induced by decitabine alone, a threshold of BAK/BAX activity may exist that controls the cells’ commitment to apoptosis. Thus, decitabine may increase apoptotic priming in DLBCL cells through DNA

damage induction and regulating the expression and activity of BCL-2 family proteins. Decitabine differentially alters gene expression in venetoclax-sensitive and -resistant diffuse large B-cell lymphoma cells Decitabine regulates gene transcription through DNA hypomethylation.19 In order to investigate the epigenetic role of decitabine in venetoclax-sensitive and -resistant contexts, we performed RNA-sequencing analysis with venetoclax-sensitive OCI-Ly1 parental and -resistant OCI-Ly1R cells. We treated both cell lines with decitabine (1 μM) for 3 days to ensure at least two rounds of replication, and analyzed their transcriptomes by RNA sequencing. Interestingly, we found that there was only about 10% overlap between OCI-Ly1 parental and OCI-Ly1R cells in terms of the significantly upregulated or downregulated genes (Figure 4A). In order to identify significantly altered gene sets, we performed gene set enrichment analysis (GSEA). The significantly altered gene signatures by decitabine were also quite different between the two cell lines (Online Supplementary Figure S6A). Nonetheless, there were still a few commonly deregulated gene signatures between parental and resistant cells. Notably, decitabine led to significant downregulation of the PI3K-AKT-mTOR1 pathway (Figure 4B), which generates a prosurvival signal in multiple genetic subgroups of DLBCL.11 Glycogen synthase kinase 3β (GSK3β) is among one of the first identified substrates of AKT,21 and its phosphorylation by AKT leads to its inactivation. A recent study found that active mTOR1 prevented nuclear localization of GSK3β,22 leading to the stabilization of the nuclear substrates of GSK3β. Thus, the downregulation of PI3K-AKT-mTOR1 pathway by decitabine could potentially activate GSK3β. Indeed, we found decreased phosphorylation of GSK3β serine 9 in both cell lines following decitabine treatment, indicating increased GSK3β activity (Figure 4C). Additionally, we noted decreased expression of c-Myc, one of the substrates of GSK3β (Figure 4C). c-Myc downregulation in response to decitabine treatment has been reported in other types of cancer as well.23,24 However, the underlying mechanism is not well understood. Our data suggest that decitabine could downregulate c-Myc expression via inhibition of the PI3K-AKT-mTOR1 pathway. Decitabine also led to upregulation of the intrinsic component of mitochondrial mem-

Haematologica | 109 January 2024

191

ARTICLE - Decitabine increases BCL-2 dependence

F. Zhu et al.

Continued on following page. Haematologica | 109 January 2024

192

ARTICLE - Decitabine increases BCL-2 dependence

F. Zhu et al.

Figure 3. Decitabine induces DNA damage and regulates the expression and activity of BCL-2 family proteins. (A) Dynamic BH3 profiling (DBP) of diffuse large B-cell lymphoma (DLBCL) cells treated with dimethyl sulfoxide (DMSO) control or 1 μM of doxorubicin for 2 days. Data were presented as δ percent priming. (B) DLBCL cells were treated with DMSO or 2.5 μM of decitabine for 3 days, cells were lysed for western blotting analysis of indicated proteins. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as loading control. For protein bands with dividing lines, images were from different parts of the same gel. (C) DLBCL cells were treated with DMSO or 2.5 μM of decitabine for 3 days, BAX (top) and BAK (bottom) activation were measured with conformation-specific antibody by flow cytometry. Median fluorescence intensity (MFI) was used for quantification, N=3. (D) DLBCL cell lines treated with DMSO or 2.5 μM of decitabine for 3 days. Anti-BAX antibody was used for immunoprecipitation (IP). The bound proteins were analyzed by western blotting with indicated antibodies. Protein band intensity was quantified by Image J. The ratio of BCL2 and BAX in the anti-BAX IP product was calculated and indicated in the figure.

brane-related gene signature in both the venetoclax-sensitive and -resistant cell lines (Figure 4D). The electron transport chain (ETC) complexes are important components of the inner mitochondria membrane and essential for energy metabolism. We assessed their formation using an antibody cocktail which contains five antibodies, detecting complex I subunit NDUFB8, complex II subunit SDHB, complex III subunit UQCRC2, complex IV subunit MTCO1 and complex V subunit ATP5A. These subunits are unstable when its complex is not assembled. We found upregulation of NDUFB8 (Figure 4E), suggesting increased formation of complex I. It has been reported that the combination of azacitidine and venetoclax reduced oxidative phosphorylation in leukemia stem cells (LSC) in AML, leading to selective targeting of LSC and durable disease remission.25 We asked if decitabine and venetoclax would have similar effect on DLBCL cells. In order to evaluate this, we measured the oxygen consumption rate (OCR) in four DLBCL cell lines in response to decitabine. Decitabine alone did not change the OCR in OCI-Ly1 parental or OCI-Ly1R cells (Figure 4F), suggesting that protein expression level does not necessarily correlate with enzymatic activity of the ETC complex, and thus should not be used to predict functional change.26 Decitabine significantly changed basal OCR in only one cell line (TMD8) (Figure 4F). Interestingly, the combination of venetoclax and decitabine significantly reduced basal OCR (Figure 4F) in all these cell lines, suggesting inhibition of OxPhos by decitabine and venetoclax. In addition, when comparing the gene expression signatures between OCI-Ly1 parental and OCI-Ly1R cells, we found that the resistant cells had downregulation in genes involved in several death pathways, including apoptosis, and senescence and autophagy (Online Supplementary Figure S6B). Taken together, our data suggest that decitabine suppresses PI3K-AKT pathway and the combination of decitabine and venetoclax disrupts energy metabolism in DLBCL. Restoration of TGF-β signaling contributes to the cytotoxicity of decitabine in diffuse large B-cell lymphoma It has been previously reported that the tumor suppressive TGF-β/TGF-βR2/SMAD1 signaling axis is frequently inactivated in DLBCL due to promoter hypermethylation of

SMAD1 gene.27 Decitabine has been found to decrease tumor burden in DLBCL cell lines and primary cell-derived xenograft mouse models in part due to restoration of SMAD1 expression.27 We confirmed upregulation of SMAD1 mRNA levels following decitabine treatment in three DLBCL cell lines (Figure 5A). Even though SMAD1 protein level remained largely unaffected, the level of phosphoSMAD1 (s463, s465) increased (Figure 5B; Online Supplementary Figure S7A) in response to decitabine treatment, indicating increased SMAD1 activity. Next, we asked whether SMAD1 activation underlies the synergism observed between decitabine and venetoclax. In order to investigate this, we used a TGF-β receptor type I/II (TβRI/II) dual inhibitor, LY2109761, that can suppress decitabine induced SMAD1 phosphorylation (Figure 5B; Online Supplementary Figure S7A). However, when looking at cell proliferation, adding LY2109761 to decitabine only partially antagonized the toxicity of decitabine in OCI-Ly1 cells (Figure 5C). Adding LY2109761 to the combination of decitabine and venetoclax also partially rescued proliferation and apoptotic cell death of OCI-Ly1 cells (Figure 5C, D). Knockdown of SMAD1 similarly made the cells less sensitive to the combination of decitabine and venetoclax (Figure 5E, F). In another two cell lines, TMD8 and SUDHL4, inhibition of TGF-β signaling had partial or no rescue on cell proliferation and apoptosis (Online Supplementary Figure S7B, C). Thus, the DNA damage response, regulation of BCL-2 family proteins and the other epigenetic targets are likely to also contribute to decitabine-induced cytotoxicity. Together, these data suggest that restoration of SMAD1 expression by decitabine only partially accounts for the synergism between decitabine and venetoclax. Combination of decitabine and venetoclax is synergistic in vivo in a diffuse large B-cell lymphoma xenograft mouse model In order to examine the in vivo activity of decitabine and venetoclax, we established a xenograft mouse model using a luciferase-labeled OCI-Ly1 cell line. Five days after tail vein injection of OCI-Ly1 cells, mice with successful engraftment of cancer cells as documented by bioluminescent imaging were divided into four groups of six to seven mice each. Each group then received one of the following treatments: i) vehicle control; ii) decitabine 0.25

Haematologica | 109 January 2024

193

ARTICLE - Decitabine increases BCL-2 dependence

F. Zhu et al.

Figure 4. Decitabine regulates gene expression in diffuse large B-cell lymphoma. (A) Venn diagram of upregulated and downregulated genes in venetoclax-sensitive (OCI-Ly1 parental) and -resistant (OCI-Ly1R) cells treated with 1 μM of decitabine for 3 days. Cutoff threshold: adjusted P value ≤0.05, absolute (log2 fold change [FC]) ≥1, N=3. (B) Decitabine significantly downregulates PI3K-AKT-MTOR pathway as revealed by gene set enrichment analysis (GSEA). (C) Expression of indicated proteins in cells treated with dimethyl sulfoxide (DMSO) or 0.5 μM of decitabine for 4 days. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as loading control. (D) Decitabine significantly upregulates mitochondrial membrane component-related gene signature as revealed by GSEA. Leading edge genes in each subset were shown. (E) Expression of indicated proteins in cells treated with DMSO or 0.5 μM of decitabine for 4 days. (F) Basal oxygen consumption rate (OCR) was measured by Seahorse XFe96 analyzer in 3 diffuse large B-cell lymphoma (DLBCL) cell lines (OCI-Ly1, OCI-Ly1R, TMD8) and 1 double hit lymphoma-patient-derived xenograft (DHL-PDX) cell line (DW20) treated as indicated. Decitabine 0.5 µM 3 days; venetoclax 1 nM for OCI-Ly1, 5 nM for DW20, 50 nM for OCI-Ly1R and TMD8, 1 day. Data were presented as mean ± standard deviation, N=5. P value was calculated by one-way ANOVA with multiple comparisons test. GO: gene ontology; NES: normalized enrichment score; FDR: false discover rate. Haematologica | 109 January 2024

194

ARTICLE - Decitabine increases BCL-2 dependence

F. Zhu et al.

Continued on following page. Haematologica | 109 January 2024

195

ARTICLE - Decitabine increases BCL-2 dependence

F. Zhu et al.

Figure 5. Inhibition of TGF-β signaling partially blocks the synergism between decitabine and venetoclax. (A) Diffuse large Bcell lymphoma (DLBCL) cell lines were treated with dimethyl sulfoxide (DMSO) or 2.5 μM of decitabine for 3 days. SMAD1 transcript level was measured by quantitative polymerase chain reaction. Data were presented as mean ± standard deviation (SD), N=3. P value was calculated by student’s t test. (B) OCI-Ly1 cells were treated with 2.5 μM of decitabine, 4 μM of TGF-β receptor type I/II (TβRI/II) dual inhibitor, LY2109761 (ly210), alone or in combination for 3 days, phospho-SMDA1 (s463/s465) and total SMAD1 were measured by western blotting. β-actin was used as loading control. (C, D) OCI-Ly1 cells were treated with indicated drugs alone or in combination. Decitabine and ly210 were added to cells for 3 days, venetoclax was added on the last day. Cell proliferation was determined by CellTiter-Glo assay. Data were normalized to DMSO-treated cells. Data were presented as mean ± SD, N=4 (C). Cell viability was measured by propidium iodide/Annexin V assay. Data were presented as mean ± SD, N=3 (D). P value was calculated by one-way ANOVA with multiple comparisons test. (E) Small hairpin RNA (shRNA)-mediated knockdown of SMAD1 in OCI-Ly1 cells. Knockdown efficiency was confirmed by western blotting with anti-SMAD1 antibody, glyceraldehyde3-phosphate dehydrogenase (GAPDH) was used as loading control. (F) OCI-Ly1 cells expressing control shRNA (shCTRL) or shSMDA1 were treated with 2.5 μM of decitabine and 0.01 μM of venetoclax alone or in combination. Decitabine was added to cells for 3 days, venetoclax was added on the last day. Left, cell proliferation was determined by CellTiter-Glo assay. Right, cell viability was measured by propidium iodide/Annexin V assay. Data were presented as mean ± SD, N=3. P value was calculated by one-way ANOVA with multiple comparisons test.

mg/kg, intraperitoneal, 5 days on/2 days off; iii) venetoclax 50 mg/kg, orally, daily; iv) the combination of decitabine and venetoclax. After 21 days, all treatments were stopped, and the status of the remaining mice was monitored. In this model, single-agent treatment with decitabine or venetoclax significantly delayed tumor growth (Figure 6A, B) and extended survival (Figure 6C) compared to vehicle control treatment. The combination of decitabine and venetoclax was able to maintain the tumor burden at a very low level during the treatment period. None of the mice receiving the combination treatment were euthanized due to poor body condition. After treatment was withdrawn, the disease progressed rapidly and all the mice receiving single-drug treatment were found paralyzed and had to be euthanized within 1 week. Mice receiving combination therapy survived longer than single-agent-treated mice did. These data suggest that combination of decitabine and venetoclax effectively inhibited DLBCL growth in vivo.

Discussion Using DBP, we have demonstrated a DNA hypomethylating drug, decitabine, as an agent that can increase the mitochondrial apoptotic priming in DLBCL cells. Although decitabine alone induced DNA damage and cell cycle arrest in DLBCL, the cells were resistant to apoptotic cell death (Online Supplementary Figures S3A and S5) likely due to the expression of multiple anti-apoptotic proteins. In contrast, the combination of decitabine and venetoclax synergistically inhibited the proliferation of DLBCL cells including DHL-PDX cell lines (Figure 2), consistent with our finding using DBP that decitabine increased BCL-2 dependence, together with various other dependences that drive overall priming (Figure 2A). In a DLBCL cell line-derived xenograft mouse model, this combination also significantly suppressed tumor growth and extended survival in vivo (Figure 6). The combined use of a hypomethylating

agent with venetoclax has been approved in AML. Our study provides a rationale for exploring a similar combination therapy in the clinic for patients with DLBCL. In this study, we did not observe a correlation between BCL-2 expression level and venetoclax sensitivity in these cell lines, highlighting the limitations of using expression of a single anti-apoptotic protein alone as a way to predict response to BH3-mimetic drugs. As a cell line usually concurrently expresses more than one anti-apoptotic BCL-2 family protein, it is likely that upon inhibition of BCL-2 by venetoclax, the anti-apoptotic function of BCL-2 can be performed by other proteins such as MCL-1, BFL-1 or BCLxL. Consistent with this hypothesis, targeting BCL-2 concomitantly with another anti-apoptotic protein dramatically inhibited proliferation of BCL-2-expressing DLBCL cells (Figure 1E). Thus, functional measurement of apoptotic proximity through DBP may be able to better guide rational selection of BH3 mimetic-based combination therapy. For example, if DBP of primary DLBCL cells following ex vivo decitabine treatment indicates increased BCL-2 dependence, this patient is likely to benefit from the combined use of decitabine and venetoclax. While a lack of BCL-2 dependence would suggest a different treatment plan. In DLBCL, an increased dependence on MCL-1 or BCL-xL following decitabine treatment was not always accompanied by synergistic cell killing by combined use of decitabine and BH3 mimetics targeting MCL-1 or BCL-xL. The inconsistent degree of inhibition by these BH3-mimetic drugs might be explained by i) non-apoptotic activity of the BCL-2 family members. Non-apoptotic roles have been reported for BCL-2, BCL-xL and MCL-1.28–32 When the BH3 domain is required for non-apoptotic functions, BH3 mimetics could induce non-apoptotic changes in the cell. ii) proliferation is measured by CellTiter-Glo assay which reflects total ATP amount inside the cells. Besides cell apoptosis, other cellular statuses such as cell cycle arrest and senescence can slow down cell proliferation and thereby influence the ATP level. Taken together, our data suggest that combining decitabine with venetoclax more

Haematologica | 109 January 2024

196

ARTICLE - Decitabine increases BCL-2 dependence

F. Zhu et al.

Figure 6. Combination of decitabine and venetoclax is synergistic in diffuse large B-cell lymphoma cell line-derived xenograft model. NSG mice transplanted with luciferized mCherry+ OCI-Ly1 cells were treated with vehicle control (N=6), decitabine (0.25 mg/kg intraperitoneal, 5 days on/2 days off, N=6), venetoclax (50 mg/kg, orally, daily, N=7) or decitabine + venetoclax (N=7). Day 21 was the last day of drug treatment. (A) Representative images from each group. Day 0 is 5 days after injection of tumor cells and the day before treatment starts. (B) Quantification of bioluminescence signal over time. Graph shows mean ± standard errorof the mean. P value was calculated using one-way ANOVA with multiple comparisons test. (C) Kaplan-Meier survival curve showing the percentage of mice without symptoms of disease (weight loss, paralysis). Control versus venetoclax, P=0.0216; control versus decitabine, P=0.0015; control versus combination, P=0.0006; venetoclax versus combination, P=0.0007, decitabine versus combination, P=0.0007. P value was calculated by log-rank test. Min: minimum; Max: maximum.

consistently kills a wide range of DLBCL cells compared to the more selective killing we observed with A1331852 or S63845. A well-established role of hypomethylating agents, including decitabine, is to induce the DNA damage response, which may lead to cell cycle arrest and/or cellular apop-

tosis. In hematopoietic stem cells, a diminished apoptotic priming in response to DNA damage induction was associated with deficits in ATM activity which can be read by the level of γH2AX.20 In DLBCL, we found that decitabine induced DNA double strand breaks as revealed by increased γH2AX, suggesting activation of ATM. However, the mech-

Haematologica | 109 January 2024

197

ARTICLE - Decitabine increases BCL-2 dependence

F. Zhu et al.

anism by which ATM activity contributes to apoptotic priming is not fully understood, especially in cells with mutant p53. A recent study has uncovered a role of ATM as a bridge for PP2A-dependent AKT dephosphorylation and subsequent activation of GSK3β, leading to mitochondria-dependent apoptosis.33 Notably, PP2A plays a major role in the phosphorylation status of anti-apoptotic proteins and affects apoptotic priming in DLBCL cells.34,35 Thus, ATM may regulate apoptotic priming through indirectly regulating the post-translational modification of BCL-2 family proteins. While the decitabine-induced DNA damage response appears to be consistent across different cell types, its function as an epigenetic regulator could be more dependent on the cellular context. By comparing the transcriptional changes in paired venetoclax-sensitive and -resistant OCI-Ly1 cell lines, we found downregulation of PI3K-AKT-mTOR1 pathway and significant changes in mitochondrial membrane-associated gene signatures in both cell lines. Other than that, the genes deregulated by decitabine are different between the two cell lines. In conclusion, our study found that the DNA hypomethylating agent decitabine sensitizes DLBCL cells to venetoclax by increasing mitochondrial apoptotic priming and BCL-2 dependence. This combination significantly inhibited DLBCL growth both in vitro and in vivo and is worthy of clinical evaluation in patients with DLBCL. Disclosures MSD has received consulting fees from AbbVie, Adaptive Biosciences, Ascentage Pharma, AstraZeneca, BeiGene, BMS, Celgene, Eli Lilly, Genentech, Janssen, Merck, Ono Pharmaceuticals, Secura Bio, TG Therapeutics and Takeda;

and research support from AbbVie, AstraZeneca, Ascentage Pharma, Genentech, MEI Pharma, Novartis, Surface Oncology and TG Therapeutics. JLC has received consulting fees from Incite, MorphoSys, Kite, ADC Therapeutics and Genmab; and research funding from Bayer, Abbvie, Genentech, and Merck. Contributions FZ, JLC and MSD developed the study concept. FZ, WN and SG acquired data. FZ, NMH and LR analyzed and interpreted data. SG, LH, SJFC, MCC, and JG provided technical assistance and discussion. MSD supervised the study. FZ wrote the original draft. All authors reviewed and edited the manuscript. Funding This work was funded in part through a PMC FLAMES FLAIR award (to JC). MSD is supported by the National Institutes of Health, National Cancer Institute grant R01CA26629801A1 and is a clinical scholar of the Leukemia and Lymphoma Society. SJFC is a scholar of the Simeon J. Fortin Charitable Foundation postdoctoral fellowship and acknowledges funding support from the Simeon J. Fortin Charitable Foundation, Bank of America Private Bank cotrustee. Data-sharing statement All the data is available from the corresponding author upon request. RNA-sequencing data has been deposited under GEO accession number GSE223598. The remaining data are included in this published article and its Online Supplemental Appendix. The Online Supplementary Appendix is available at Haematologica’s website.

References 1. 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. 2. Schmitz R, Wright GW, Huang DW, et al. Genetics and pathogenesis of diffuse large B-cell lymphoma. N Engl J Med. 2018;378(15):1396-1407. 3. 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. 4. Alizadeh AA, Eisen MB, Davis RE, et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature. 2000;403(6769):503-511. 5. Rosenwald A, Wright G, Chan WC, et al. The use of molecular profiling to predict survival after chemotherapy for diffuse large-B-cell lymphoma. N Engl J Med. 2002;346(25):1937-1947. 6. Huang JZ, Sanger WG, Greiner TC, et al. The t(14;18) defines a unique subset of diffuse large B-cell lymphoma with a germinal center B-cell gene expression profile. Blood.

2002;99(7):2285-2290. 7. Davids MS, Letai A. Targeting the B-cell lymphoma/leukemia 2 family in cancer. J Clin Oncol. 2012;30(25):3127-3135. 8. Valentin R, Grabow S, Davids MS. The rise of apoptosis: targeting apoptosis in hematologic malignancies. Blood. 2018;132(12):1248-1264. 9. Souers AJ, Leverson JD, Boghaert ER, et al. ABT-199, a potent and selective BCL-2 inhibitor, achieves antitumor activity while sparing platelets. Nat Med. 2013;19(2):202-208. 10. Smith VM, Dietz A, Henz K, et al. Specific interactions of BCL-2 family proteins mediate sensitivity to BH3-mimetics in diffuse large B-cell lymphoma. Haematologica. 2020;105(8):2150-2163. 11. Bojarczuk K, Wienand K, Ryan JA, et al. Targeted inhibition of PI3Kα/δ is synergistic with BCL-2 blockade in genetically defined subtypes of DLBCL. Blood. 2019;133(1):70-80. 12. Davids MS, Roberts AW, Seymour JF, et al. Phase I first-inhuman study of venetoclax in patients with relapsed or refractory non-Hodgkin lymphoma. J Clin Oncol. 2017;35(8):826-833. 13. Ryan JA, Brunelle JK, Letai A. Heightened mitochondrial priming

Haematologica | 109 January 2024

198

ARTICLE - Decitabine increases BCL-2 dependence

F. Zhu et al.

is the basis for apoptotic hypersensitivity of CD4+ CD8+ thymocytes. Proc Natl Acad Sci U S A. 2010;107(29):12895-12900. 14. Tsao T, Shi Y, Kornblau S, et al. Concomitant inhibition of DNA methyltransferase and BCL-2 protein function synergistically induce mitochondrial apoptosis in acute myelogenous leukemia cells. Ann Hematol. 2012;91(12):1861-1870. 15. Jin S, Cojocari D, Purkal JJ, et al. 5-azacitidine induces NOXA to prime AML cells for venetoclax-mediated apoptosis. Clin Cancer Res. 2020;26(13):3371-3383. 16. Momparler RL, Momparler LF, Samson J. Comparison of the antileukemic activity of 5-AZA-2’-deoxycytidine, 1-beta-Darabinofuranosylcytosine and 5-azacytidine against L1210 leukemia. Leuk Res. 1984;8(6):1043-1049. 17. Guièze R, Liu VM, Rosebrock D, et al. Mitochondrial reprogramming underlies resistance to BCL-2 inhibition in lymphoid malignancies. Cancer Cell. 2019;36(4):369-384. 18. Karahoca M, Momparler RL. Pharmacokinetic and pharmacodynamic analysis of 5-aza-2’-deoxycytidine (decitabine) in the design of its dose-schedule for cancer therapy. Clin Epigenetics. 2013;5(1):3. 19. Stresemann C, Lyko F. Modes of action of the DNA methyltransferase inhibitors azacytidine and decitabine. Int J Cancer. 2008;123(1):8-13. 20. Gutierrez-Martinez P, Hogdal L, Nagai M, et al. Diminished apoptotic priming and ATM signalling confer a survival advantage onto aged haematopoietic stem cells in response to DNA damage. Nat Cell Biol. 2018;20(4):413-421. 21. Cross DA, Alessi DR, Cohen P, Andjelkovich M, Hemmings BA. Inhibition of glycogen synthase kinase-3 by insulin mediated by protein kinase B. Nature. 1995;378(6559):785-789. 22. Bautista SJ, Boras I, Vissa A, et al. mTOR complex 1 controls the nuclear localization and function of glycogen synthase kinase 3β. J Biol Chem. 2018;293(38):14723-14739. 23. Urakami S, Shiina H, Enokida H, et al. Epigenetic inactivation of Wnt inhibitory factor-1 plays an important role in bladder cancer through aberrant canonical Wnt/β-catenin signaling pathway. Clin Cancer Res. 2006;12(2):383-391. 24. Jiang X, Wang Z, Ding B, et al. The hypomethylating agent decitabine prior to chemotherapy improves the therapy efficacy in refractory/relapsed acute myeloid leukemia patients.

Oncotarget. 2015;6(32):33612-33622. 25. Pollyea DA, Stevens BM, Jones CL, et al. Venetoclax with azacitidine disrupts energy metabolism and targets leukemia stem cells in patients with acute myeloid leukemia. Nat Med. 2018;24(12):1859-1866. 26. Bajpai R, Sharma A, Achreja A, et al. Electron transport chain activity is a predictor and target for venetoclax sensitivity in multiple myeloma. Nat Commun. 2020;11(1):1228. 27. Stelling A, Wu CT, Bertram K, et al. Pharmacological DNA demethylation restores SMAD1 expression and tumor suppressive signaling in diffuse large B-cell lymphoma. Blood Adv. 2019;3(20):3020-3032. 28. Chen ZX, Pervaiz S. Bcl-2 induces pro-oxidant state by engaging mitochondrial respiration in tumor cells. Cell Death Differ. 2007;14(9):1617-1627. 29. Perciavalle RM, Stewart DP, Koss B, et al. Anti-apoptotic MCL-1 localizes to the mitochondrial matrix and couples mitochondrial fusion to respiration. Nat Cell Biol. 2012;14(6):575-583. 30. Williams A, Hayashi T, Wolozny D, et al. The non-apoptotic action of Bcl-xL: regulating Ca(2+) signaling and bioenergetics at the ER-mitochondrion interface. J Bioenerg Biomembr. 2016;48(3):211-225. 31. Escudero S, Zaganjor E, Lee S, et al. Dynamic regulation of long-chain fatty acid oxidation by a noncanonical interaction between the MCL-1 BH3 helix and VLCAD. Mol Cell. 2018;69(5):729-743. 32. Roca-Portoles A, Rodriguez-Blanco G, Sumpton D, et al. Venetoclax causes metabolic reprogramming independent of BCL-2 inhibition. Cell Death Dis. 2020;11(8):1-13. 33. Hotokezaka Y, Katayama I, Nakamura T. ATM-associated signalling triggers the unfolded protein response and cell death in response to stress. Commun Biol. 2020;3(1):1-11. 34. Chong SJF, Lai JXH, Qu J, et al. A feedforward relationship between active Rac1 and phosphorylated Bcl-2 is critical for sustaining Bcl-2 phosphorylation and promoting cancer progression. Cancer Lett. 2019;457:151-167. 35. Chong SJF, Iskandar K, Lai JXH, et al. Serine-70 phosphorylated Bcl-2 prevents oxidative stress-induced DNA damage by modulating the mitochondrial redox metabolism. Nucleic Acids Res. 2020;48(22):12727-12745.

Haematologica | 109 January 2024

199

ARTICLE - Non-Hodgkin Lymphoma

Diffuse large B-cell lymphoma involving osseous sites: utility of response assessment by PET/CT and good longterm outcomes Paola Ghione,1,2 Salma Ahsanuddin,1 Efrat Luttwak,1 Sabela Bobillo Varela,1 Reiko Nakajima,1 Laure Michaud,1 Kanika Gupta,1 Anastasia Navitski,1 David Straus,1,2 M. Lia Palomba,1,2 Alison Moskowitz,1,2 Ariela Noy,1,2 Paul Hamlin,1,2 Matthew Matasar,1,2 Anita Kumar,1,2 Lorenzo Falchi,1,2 Joachim Yahalom,1,3 Steven Horwitz,1,2 Andrew Zelenetz,1,2 Anas Younes,1 Gilles Salles,1,2 Heiko Schöder2,4 and Erel Joffe1,5 1

Lymphoma Service, Memorial Sloan Kettering Cancer Center, New York, NY, USA; 2Weill

Correspondence: P. Ghione ghionep@mskcc.org E. Joffe erelj@tlvmc.gov.il Received: Accepted: Early view:

December 27, 2022. August 22, 2023. August 31, 2023.

Cornell College of Medicine, New York, NY, USA; 3Molecular Imaging and Therapy Service, Memorial Sloan Kettering Cancer Center, New York, NY, USA; 4Radiation Therapy, Memorial Sloan Kettering Cancer Center, New York, NY, USA; 5Sackler School of Medicine, Tel Aviv University, Tel Aviv, Israel

Abstract Osseous involvement by diffuse large B-cell lymphoma (DLBCL-bone) is a heterogeneous disease. There is limited data regarding response assessment by positron emission tomography with fluorodeoxyglucose, which may demonstrate residual avidity despite a complete response. We analyzed clinical data of patients with newly diagnosed DLBCL and identified all cases with DLBCL-bone. End of treatment scans were reviewed by two independent experts classifying osseous lesions into Deauville (DV) ≤3; DV ≥4, or reactive uptake in the bone marrow (M), site of fracture (F) or surgery (S). We compared outcomes of DLBCL-bone to other extranodal sites (EN) matched on International Prognotic Index features and regimen. Of 1,860 patients with DLBCL (bone 16%; EN 45%; nodal 39%), 41% had localized disease and 59% advanced. Only 9% (n=27) of patients with initial bone involvement had residual fluorodeoxyglucose avidity at the osseous site. In half of these cases, the uptake was attributed to F/S/M, and of the remaining 13, only two were truly refractory (both with persistent disease at other sites). Overall survival and progression-free survival (PFS) were found to be similar for early-stage nodal DLBCL and DLBCL-bone, but inferior in EN-DLBCL. Advanced-stage disease involving the bone had a similar 5-year PFS to nodal disease and EN-DLBCL. After matching for International Prognotic Index and treatment regiments, PFS between bone and other EN sites was similar. Osseous involvement in DLBCL does not portend a worse prognosis. End of treatment DV ≥4 can be expected in 5-10% of cases, but in the absence of other signs of refractory disease, may be followed expectantly.

Introduction Diffuse large B-cell lymphoma (DLBCL) is the most common subtype of non-Hodgkin lymphoma,1 involving extranodal sites (EN) in 30% to 60% of the patients.2 Approximately 7% to 21% of DLBCL present with osseous lesions.3 Standard of care treatment for DLBCL is chemoimmunotherapy, followed by response evaluation with functional imaging and consideration of consolidation radiation therapy (RT), particularly for localized disease.4-7 Several retrospective studies evaluated the clinical course of advanced DLBCL with osseous involvement (i.e., stage IV), of early-stage osseous disease (stage I/II-E) and disease confined to bony sites (primary bone lymphoma).8-13 Notably,

in nearly all studies, data was limited to bone involvement as identified by computed tomography (CT) with minimal information about the rate of cortical bone involvement by positron emission tomography (PET) in the absence of CT findings. When evaluated by CT, osseous involvement can be seen in 7.6% in advanced-stage disease and has been associated with a reduced event-free survival.8 In localized DLBCL, osseous involvement can be seen in 3% when evaluated by CT, yet we have recently demonstrated up to 21% of patients may have evidence of involvement by PET.14 In that study, any EN involvement (i.e., stage I/II-E) was associated with a poorer prognosis, however, the analysis did not include a dedicated evaluation for lymphoma involving osseous sites. More recently, analysis of localized

Haematologica | 109 January 2024

200

ARTICLE - DLBCL with bone lesions, utility of PET/CT assessment

DLBCL treated on three consecutive Southwestern Oncology Group studies (S0014, S0313, S1001; clinicaltrials gov. Identifier: NCT00005089, NCT00070018, NCT01359592) has been published.15 Their results contrast with those from Bobillo et al.14 and do not support EN disease as an adverse prognostic factor for patients with localized DLBCL. In that regard, primary bone lymphoma, historically has been shown to carry an excellent prognosis.10-13 Interestingly, better prognosis has also been demonstrated for multi-focal (stage IV) primary bone lymphoma compared to disease involving both nodal and osseous sites.11 16 Possibly contributing to the concerns associated with osseous presentation, is the limited data regarding response assessment by PET with fluorodeoxyglucose (FDG), which may demonstrate residual avidity despite a complete response. Of note that the Deauville (DV) criteria were devised based on data from nodal disease, and it is unclear whether they can be applied to EN sites.17-19 Osseous sites, for example, may be associated with residual uptake at the end of treatment including false-positive uptake due to fractures, reactive bone marrow uptake, or bone reconstruction.20,21 In the present work we aimed to evaluate clinical and PET features of DLBCL involving the bone and to assess their association with treatment outcomes and disease course.

Methods Following institutional review board approval, we reviewed all adult patients (age ≥18 years) with newly diagnosed DLBCL treated with the combination of rituximab, cyclophosphamide, adriamycin, vincristine and prednisone (R-CHOP) and R-CHOP-like chemotherapy at Memorial Sloan Kettering Cancer Center (MSKCC) between 2000 and 2015. All patients underwent a PET-CT scan at diagnosis. We evaluated response assessment by PET-CT in patients with bone involvement and compared treatment outcomes to patients with DLBCL involving other sites of EN disease and to patients with disease limited to nodal sites. Patient data were collected from our institutional lymphoma database. Medical and pathology records were evaluated for clinical characteristics, pathologic and radiologic data, treatment history and outcomes. Treatment was considered to have occurred if at least one cycle of the chemo-immunotherapy was administered. Cortical bone involvement was identified by review of all imaging reports pretreatment and at end of treatment (EOT) by two independent reviewers. Patients with isolated involvement of the bone marrow without involvement of the cortical bone were not considered DLBCL-bone for this analysis. Whenever a high focal bone uptake was noted without CT abnormalities, it was considered as positive for bone lesions.22 EOT scans of patients with any residual FDG avidity above the liver mean standardized uptake value (SUV) at any initially involved osseous sites were referred for

P. Ghione et al.

further review by two independent and blinded radiologists specialized in nuclear medicine. EOT response was classified according to the DV criteria as DV ≤3 (uptake similar or lower than liver); DV ≥4 (uptake greater than liver) with a separate designation for superimposed uptake due to reactive bone marrow (M), site of fracture (F) or site of surgery (S).19 Cases with disagreement between the two radiologists were evaluated by a third reviewer and resolved by consensus. Overall response to treatment was based on the Lugano criteria.23 We grouped patients into osseous, other EN and strictly nodal disease groups. We compared baseline characteristics, response to treatment and survival between the groups separately for localized disease (i.e., stage I/II nodal; I/II-E with osseous involvement; I/II-E with non-osseous EN involvement) and advanced stage disease (i.e., stage III; stage IV osseous; stage IV non-osseous). Finally, we compared survival between the patients with osseous involvement and a cohort of patients with other EN sites matched 1:2 on International Prognostic Index (IPI) features (stage, number of EN sites, age, lactate dehydrogenase [LDH], performance status, cell of origin was determined by the Hans algorithm)24 and on treatment regimen. Baseline characteristics between the groups were compared using the Fisher exact test for categorical variables and the Kruskal-Wallis test for numeric variables. Categorical data are reported as percent (number) and numeric data as median (interquartile range [IQR]). Overall survival (OS) was defined as the time from initiation of treatment to death of any cause censoring at date of last follow-up. Progression-free survival (PFS) was defined as time from initiation of treatment until progression of disease or death of any cause, censoring at date of last follow-up. Time-to-event statistics were estimated using the Kaplan-Meier method, and compared between the groups using the log-rank test. All analyses were performed using R (R version 3.6.3, Austria).

Results Patients We analyzed data of 1,860 patients with DLBCL receiving R-CHOP or R-CHOP-like treatment between 2000 and 2015. Within this population, 39% (n=732) had purely nodal disease, and 61% (n=1,128) had EN involvement. Bone was the most commonly involved EN site with 300 patients having at least one osseous lesion (27% of all EN and 16% of the entire cohort). The most commonly involved other EN sites were lungs 16% (n=180), stomach 13% (n=143), gastro-instestinal tract 12% (n=130), muscle and non-nodal soft tissue 11% (n=126), liver 10% (n=116), bone marrow 9% (n=103), kidney/adrenal 7% (n=83), skin/subcutaneous 6% (n=63), breast 4% (n=46), pancreas 4% (n=45), testis 4% (n=45), with other sites being less frequent. Concurrent systemic and central nervous system involvement was found in 2% (n=21). In the entire cohort, 41% (n=766) had localized dis-

Haematologica | 109 January 2024

201

ARTICLE - DLBCL with bone lesions, utility of PET/CT assessment

ease and 59% (n=1,094) advanced stage. Most patients received frontline chemotherapy with R-CHOP (70%, n=1306), R-EPOCH (13%, n=232) or a regimen of four cycles of R-CHOP followed by three cycles of R-ICE (ifosfamide, carboplatin, etoposide) (17%, n=322).25,26 Progression/ relapse (POD) after frontline chemotherapy were recorded in 20% (n=374) and all-cause deaths in 25% (n=460). Baseline characteristics for limited and advanced stage are presented in Tables 1 and 2, respectively. Limited stage Of the 766 patients with limited stage disease 5.5% (n=42) had DLBCL-bone and 35% (n=271) had EN-DLBCL (Table 1). Patients with DLBCL-bone were significantly younger (age 46 stage I/II-bone versus 62 I/II-EN versus 55 I/II-nodal; P<0.001) with 38% of them (n=16) below the age of 40 at diagnosis. DLBCL-bone was further characterized by a higher rate of germinal center B-cell (GCB) cell of origin (64% stage I/II-bone vs. 36% I/II-EN vs. 45% I/II-nodal; P=0.008). Of note, none of the stageI-II DLBCL localized to the bone showed a transformed histology. Advanced stage Of the 1,094 patients with advanced stage DLBCL 26% (n=279) had a purely nodal disease (stage III), 24% (n=258) had DLBCL-bone and 51% (n=557) EN-DLBCL (Table 2).

P. Ghione et al.

Most DLBCL-bone patients (64%) had involvement of an additional EN site, and these patients were managed more aggressively with 48% (n=125) treated with R-EPOCH or R-CHOP/RICE rather than RCHOP as compared to 33% (n=183) and 31% (n=85) of patients with EN and nodal disease, respectively (P<0.001). Positron emission tomography response assessment Most patient demonstrated a complete metabolic response (CMR) by PET at the site of osseous involvement (DV ≤3 in 91%, n=273). Positive EOT PET (DV ≥4) had low predictive value for residual disease. In half of the PET-positive cases (14/27), after re-review by our three blinded nuclear medicine physicians, the uptake was attributed to local fracture, surgery, or background marrow uptake (F/S/M). The remaining 13 cases (4%) were truly suspicious for residual osseous disease, but only two patients had refractory disease, both with additional extra-osseous sites. One patient was consolidated with radiation as part of the preplanned treatment. Ten patients did not receive further treatment, and four of them had repeat biopsies (3 osseous site; 1 adjacent LN) which were negative. Two of these ten patients subsequently relapsed, but only one had recurrent disease at the initial site of osseous involvement (35 months after initial treatment). With a median follow-up of over 5 years, 25% (n=76) of patients with osseous in-

Table 1. Limited stage (I/II) diffuse large B-cell lymphoma - baseline characteristics. I-II Nodal N=453

I-II EN-bone N=42

I-II EN-other N=271

P overall

P bone vs. EN

55.0 (41.0-68.0)

46.0 (31.0-65.5)

62.0 (51.0-72.0)

<0.001

Sex F, N (%)

228 (50.3)

19 (45.2)

142 (52.4)

0.658

0.411

PS ECOG ≥2, N (%)

35 (7.73)

2 (4.76)

16 (5.90)

0.639

1.000

IPI 3-5, N eval.=740, N (%)

12 (2.75)

0 (0.00)

5 (1.90)

0.389

0.172

Cell of origin (Hans alg.), N (%) GCB non-GCB (available for 563 pts)

196 (43.3) 144 (31.8)

27 (64.3) 9 (21.4)

98 (36.2) 89 (32.8)

0.008

0.001

Transformed, N (%) Bulky ≥10 cm, N eval.=706, N (%) LDH >ULN, N eval.=715, N (%)

53 (11.7) 117 (26.5) 193 (45.8)

0 (0.00) 5 (14.7) 14 (35.0)

19 (7.01) 16 (6.96) 70 (27.6)

0.006 <0.001 <0.001

0.088 0.177 0.469

Treatment, N (%) R-CHOP R-EPOCH/R-CHOP-R-ICE

325 (71.7) 128 (28.2)

38 (90.5) 4 (9.52)

242 (89.3) 29 (10.7)

<0.001

1.000

N of cycles (%) ≥5 3-4 Incomplete treatment

311 (68.7) 132 (29.1) 10 (2.21)

23 (54.8) 19 (45.2) 0 (0.00)

139 (51.3) 127 (46.9) 5 (1.85)

<0.001

1.000

Radiation therapy, N (%)

143 (31.6)

27 (64.3)

127 (46.8)*

<0.001

0.316

Variable, total N=766 Age in years, median (IQR)

EN: extranodal; IQR: interquartile range; F: female; LDH: lactate dehydrogenase; ULN: upper limt of normal; eval: evaluated; R-CHOP: rituximab, cyclophosphamide, adriamycin, vincristine and prednisone; R-EPOCH: rituximab, etoposide, cyclophosphamide, adriamycin, vincristine and prednisone; R-ICE: rituximab ifosfamide, carboplatin, etoposide; PS ECOG: performance status by Eastern Cooperative Oncology Group; IPI: International Prognostic Index; Hans alg.: Hans algorithm; GCB: germinal center B-cell derived; pts: patients. *Excluding 21 cases with radiation therapy to contralateral testis.

Haematologica | 109 January 2024

202

ARTICLE - DLBCL with bone lesions, utility of PET/CT assessment

volvement experienced treatment failure or relapse, but only in 6% did the relapse involve the initial osseous site (Figure 1). PFS among patients with DV ≤3 and ≥4, or with F/S/M avidity was similar (Figure 2). Survival of limited and advanced stage disease OS at 5 years (5y-OS) was similar for early-stage nodal DLBCL and DLBCL-bone and inferior in EN-DLBCL (5y-OS 93% and 95% vs. 88%, respectively; P=0.02). PFS at 5 years (5y-PFS) in patients with nodal DLBCL was superior to that of early-stage bone DLBCL and EN-DLBCL (5y-PFS 92% for nodal DLBCL, 84% for bone DLBCL and 84% for EN-DLBCL; P=0.03). Of note, none of the 34 patients with stage I-bone ever relapsed, translating to a 5y-PFS of 97% (1 late death; compared to 93% for nodal DLBCL and 84% for other EN-DLBCL; P=0.02) (Online Supplementary Figure S1). Most of the patients with stage I-bone (68%, n=23) underwent consolidative radiotherapy (Table 3; Figures 2, 3). Advanced stage disease involving the bone had a similar 5y-PFS to

P. Ghione et al.

nodal disease, and slightly superior to EN-DLBCL (5y-PFS 66% stage IV-bone; 70% stage III and 62% stage IV-EN; P=0.01). OS was significantly better in nodal and bone disease compared with EN-DLBCL (5y-OS 80% stage IV-bone; 85% stage III and 71% stage IV-EN; P<0.0001). In order to allow for a non-biased comparison between DLBCL-bone and DLBCL-EN, we compared the 300 cases with osseous DLBCL with 600 controls from the EN-DLBCL, matching them based on IPI features (stage, number of EN sites, age, LDH, performance status), and treatment regimen (Figure 4). This case-control analysis showed no statistically significant difference in PFS between bone and other EN sites (P=0.1). Slight advantage in OS was still present in osseous cases (P=0.02) (Online Supplementary Figure S1).

Discussion This study aimed to evaluate the disease course of DLBCL

Table 2. Advanced stage diffuse large B-cell lymphoma - baseline characteristics.

Total N=1,094

Stage III-nodal N=279

Stage IV-bone N=258

Stage IV-EN N=557

P overall

P bone versus other EN

Age in years, median (IQR)

61.0 (51.0-71.5)

62.0 (46.0-70.0)

65.0 (53.0-74.0)

<0.001

Sex F, N (%)

125 (44.8)

105 (40.7)

261 (46.9)

0.258

0.116

PS ECOG ≥2, N (%)

47 (16.8)

82 (31.8)

181 (32.5)

<0.001

0.920

IPI 3-5, N eval.=1,058, N (%)

103 (38.4)

194 (76.1)

342 (63.9)

<0.001

0.001

Cell of origin (Hans alg), N (%) GCB non-GCB (available for 848 pts)

130 (46.6) 80 (28.7)

124 (48.1) 87 (33.7)

245 (44.0) 182 (32.7)

0.298

0.806

Double-hit, N (%) (available for 155 pts)

5 (18.5)

6 (13.6)

10 (11.9)

0.695

Bulk ≥10cm, N (%) (available in 991 pts)

36 (13.4)

44 (18.7)

118 (24.2)

0.001

0.027

≥2 EN sites, N (%)

166 (64.3)

183 (32.9)

<0.001

LDH>ULN, N (%)

166 (63.8)

178 (72.1)

338 (66.0)

0.119

Regimen, N (%) R-CHOP R-EPOCH/R-CHOP-R-ICE

194 (69.5) 85 (30.5)

133 (51.6) 125 (48.4)

374 (67.1) 183 (32.8)

<0.001

Cycles, N (%) ≥5 3-4 Incomplete/POD/TRM

263 (94.3) 5 (1.79) 11 (3.94)

249 (96.5) 2 (0.78) 7 (2.71)

507 (91.0) 16 (2.87) 34 (6.10)

0.055

0.088

Radiation therapy, N (%)

6 (2.15)

24 (9.30)

40 (7.18)

0.002

0.365

ORR, N (%)

261 (93.5)

245 (94.9)

492 (88.3)

0.004

0.179

91.0

93.4

85.8

CR %

EN: extranodal; IQR: interquartile range; F: female; PS ECOG: performance status by Eastern Cooperative Oncology Group; IPI: International Prognostic Index; eval.: evaluated; Hans alg: Hands algorithm; GCB: germinal center B-cell derived; pts: patients; LDH: lactate dehydrogenase; ULN: upper limit of normal; R-CHOP: rituximab, cyclophosphamide, adriamycin, vincristine and prednisone; R-EPOCH: rituximab, etoposide, cyclophosphamide, adriamycin, vincristine and prednisone; R-ICE: rituximab ifosfamide, carboplatin, etoposide; POD: progression of disease; TRM: treatment-related mortality; ORR: overall response rate; CR: complete remission.

Haematologica | 109 January 2024

203

ARTICLE - DLBCL with bone lesions, utility of PET/CT assessment

P. Ghione et al.

Figure 1. Consort diagram of diffuse large B-cell lymphoma bone cases. Consort diagram of diffuse large B-cell lymphoma bone (DLBCL-bone) cases comparing resolved versus persistent fluorodeoxyglucose (FDG) avidity at initial osseous site. Of 16 cases with refractory disease, 14 had a refractory disease but resolution of uptake at the initial osseous site, 2 had a refractory disease with uptake in both osseous and other sites. Of 82 cases with relapsed or refractory (R/R) disease only 20 cases involved the initial osseous sites. w/EOT PET: with end-of-treatment positron emission tomography; CR: complete response: POD: progression of disease: BM: bone marrow involvement; DV: Deauville; Cont.: continuous: XRT: radiation therapy.

Figure 2. Progression-free survival of diffuse large B-cell lymphoma bone cases by endof-treatment positron emission tomography response. Comparing patients with Deauville (DV) ≤3, DV ≥4, or with fractures/surgery/ marrow (F/S/M) avidity. PFS: progression-free survival; DLBCL-bone: diffuse large B-cell lymphoma bone; EOT end of treatment.

Haematologica | 109 January 2024

204

ARTICLE - DLBCL with bone lesions, utility of PET/CT assessment

involving osseous site as compared to disease involving other EN and nodal site, with a particular focus on response evaluation by PET/CT. We reviewed data of 1,860 patients with DLBCL treated with RCHOP/RCHOP-like regimens of whom 16% had cortical bones involvement. We demonstrate that osseous involvement in and of itself does not portend a worse prognosis compared to other sites of extra-nodal disease. Further, though residual FDG avidity at osseous sites may be seen in approximately 10% of patients, residual disease seems rare and is usually associated with presence of extra-osseous disease. The rate of osseous involvement in our cohort is considerably higher than that described in prior studies using CT, but similar to that described with the use of PET for baseline staging.8,27,28 For example, in a meta-analysis of patients

P. Ghione et al.

with advanced stage DLBCL treated on nine prospective trials (n=3,840) of the German High-Grade Non-Hodgkin Lymphoma Study Group (DSHNHL), the rate of osseous involvement by CT was approximately 8%, as compared to a sub study of CALGB 50303 in which the reported rate of osseous involvement by PET-CT was approximately 20%.27 In keeping with prior studies, DLBCL-bone, in and of itself, was not associated with a worse prognosis. A similar observation was made in a retrospective study of 60 patients with DLBCL-bone compared with 181 historical controls with EN-DLBCL matched by IPI score and by presence of bone marrow involvement. This study demonstrated no differences in outcome, with a 5-year PFS of 54% and OS 59%.9 Similarly, in the FLYER study, the outcome of patients with localized disease of the bone was virtually identical

Figure 3. Progression-free survival of limited stage diffuse large B-cell lymphoma. PFS: progression-free survival; EN: extranodal.

Table 3. Overall and progression-free survival by stage.

Patients N

5y-PFS median (IQR)

HR median (IQR)

5y-OS median (IQR)

HR median (IQR)

Limited I/II-nodal I/II-bone I/II-other EN

453 42 271

0.93 (0.9-0.95) 0.95 (0.88-1) 0.71 (0.22-2.27) 0.88 (0.84-0.93) 1.55 (1.07-2.24)

1.0 0.56 0.02

0.87 (0.84-0.9) 0.92 (0.83-1) 0.48 (0.15-1.53) 0.84 (0.79-0.89) 1.42 (1.03-1.97)

0.21 0.03

Advanced III-nodal IV-bone IV-other EN

279 258 557

0.70 (0.65-0.76) 0.66 (0.61-0.73) 1.05 (0.8-1.39) 0.62 (0.58-0.66 1.38 (1.1-1.72)

0.71 0.01

0.85 (0.81-0.9) 0.8 (0.74-0.85) 1.13 (0.8-1.58) 0.71 (0.67-0.75)] 1.84 (1.4-2.41)

0.49 <0.0001

5y-PFS: 5-year progression-free survival; 5y-OS: 5-year overall survival; HR: hazard ratio; IQR: interquartile range; EN: extranodal.

Haematologica | 109 January 2024

205

ARTICLE - DLBCL with bone lesions, utility of PET/CT assessment

P. Ghione et al.

Figure 4. Progression-free survival of matched diffuse large B-cell lymphoma bone and extranodal diffuse large B-cell lymphoma. PFS: progression-free survival.

to those with other localized sites of DLBCL.28 In contrast, in the meta-analysis by the DSHNHL, skeletal involvement by CT was associated with a worse outcome after RCHOP (hazard ratio event-free survival =1.5) with a benefit from radiotherapy consolidation to bony sites.8 These differences may be explained, in part, by the limited ability of CT to identify bone involvement (6%, n=60 among patients treated with RCHOP) and absence of data whether in cases considered R/R the site of residual or relapsing the disease was the initial osseous site or another site of disease. A possible explanation for the widespread notion that DLBCL-bone has a worse outcome is the relatively high rate of FDG uptake at the end of therapy at osseous sites, which in the absence of a confirmatory biopsy, may be falsely regarded as sites of refractory disease. In our study, nearly 10% (n=27) of the patients demonstrated residual bone uptake with intensity greater than liver reference region. However, the positive predictive value (PPV) of this finding was very low. After excluding causes of false-positive FDG uptake (i.e., fracture, surgery, marrow reactivity), the PPV was only 15% (2/13), only two cases were confirmed to truly have residual disease. This PPV is much lower than that reported for residual FDG uptake in nodal sites (ranging from 50% to 100%).29-32 Concordant with the Lugano criteria, we, therefore, recommend to verify suspicion for residual disease by either a biopsy, alternative imaging method such as magnetic resonance imaging or a repeated PET/CT before proceeding with further treatment.23 Importantly, our data demonstrate that physicians at a large academic center treat patients with advanced stage osseous involvement more aggressively than those with EN non-osseous disease. Yet, 91% of DLBCL-bone patients demonstrated a cardiovascular magnetic resonance at osseous sites irrespective of their overall systemic response, and less than a third of all relapses involved the initial

osseous sites. These data do in fact suggest an excellent penetration and local control for chemotherapy at osseous sites.33 In patients with limited stage disease we confirmed the prior observation of excellent outcomes of standard treatment with no relapses observed in 38 patients with stage IE disease.11,13,20,34,35 Unlike patients with advanced-stage disease, they were not treated more aggressively than other limited stage patients with nodal or EN disease, though most (68%) received RT consolidation after a short course of R-CHOP. In conclusion, bone involvement in DLBCL, in and of itself, does not portend a worse prognosis. Residual bone FDG uptake with intensity greater than liver reference (DV ≥4) can be seen in 5-10% of cases. However, unlike nodal disease, this finding is only a weak predictor of refractory disease. In the absence of clear signs of refractory disease at other sites or a confirmatory biopsy showing DLBCL, these sites of residual bone uptake can often be monitored expectantly. Disclosures EJ discloses membership on advisory boards of Epizyme and AstraZeneca. MLP discloses honoraria from Pharmacyclics, Celgene, Merck, Novartis, Regeneron and Juno Therapeutics, a Bristol-Meyers Squibb Company; research funding from Genentech and Juno Therapeutics. AN discloses research funding from Rafael Pharma and Pharmacyclics; consultancy for Pharmacyclics, Medscape, Targeted Oncology, Morphosys, Pharmacyclics and Janssen; research funding from NIH. MM discloses consultancy, honoraria and research funding from Genentech, Inc, Bayer and Immunovaccine; consultancy for Merck, Bayer, Juno Therapeutics, F. Hoffmann-La Roche Ltd. Technologies, Daiichi Sankyo, Seattle Genetics, IGM Biosciences, Janssen, Pharmacyclics, Rocket Medical, Takeda, GlaxoSmithKline and Teva. PH discloses research funding

Haematologica | 109 January 2024

206

ARTICLE - DLBCL with bone lesions, utility of PET/CT assessment

from J&J Pharmaceuticals, Portola, Molecular Templates and Incyte; consultancy for Celgene, Karyopharm and Juno Therapeutics. AK discloses honoraria from Kite Pharmaceuticals, Astra Zeneca, Celgene and Seattle Genetics; research funding from Pharmacyclics, Celgene, Adaptive Biotechnologies and AbbVie. AM discloses consultancy for Seattle Genetics, Miragen Therapeutics, Imbrium Therapeutics, L.P. and Merck, research funding from Bristol-Myers Squibb, Incyte, Seattle Genetics and Merck. LF discloses research funding from Genmab and Roche. GvK discloses consultancy for and honoraria from Merck. DS discloses consultancy for and speakers bureau of Targeted Oncology, Seattle Genetics, Elsevier and OncLive; research funding from Takeda and Pharmaceuticals; consultancy for Imedex, Inc. and Karyopharm Therapeutics. SH discloses consultancy for and research funding from Aileron, ADCT Therapeutics, Celgene, Forty Seven, Infinity/Verastem, Kyowa Hakka Kirin, Millenium/Takeda, Seattle Genetics, Trillium, Corvus, Innate Pharma, Mundipharma, Portola, Beigene, C4 Therapeutics, Daiichi Sankyo, GlaxoSmithKline, Janssen, Kura Oncology, Miragen, Myeloid Therapeutics, Verastem, Vividion Therapeutics, Affirmed and ASTEX. AD discloses consultancy for Takeda and Roche; research funding from

P. Ghione et al.

Physicians Education Resource, Corvus Pharmaceuticals, Seattle Genetics, EUSA Pharma, AbbVie and the National Cancer Institute. AZ is on the board of directors or advisory committees of BeiGene; discloses cnsultancy for Adaptive Biotechnology, Novartis, Amgen, Janssen, Celgene, Gilead, Genentech/Roche and Gilead; research funding from Roche, Celgene, Sandoz, MorphoSys and MEI Pharma. Contribution: EJ, PG and SA designed the research. SA developed the script in R and performed the analyses. SA generated the vocabularies. PG assisted with script development. SA and SZ reviewed the notes for gold-standard generation. HS, LM and RN re-reviewed the PET/CT images. EJ supervised note review and resolved questionable cases. EJ, EL and PG wrote the manuscript. All authors reviewed and approved the manuscript. Data-sharing statement All the individual participant data collected during the trial are available after de-identification. Data are available upon reasonable request to the corresponding authors.

References 1. Teras LR, DeSantis CE, Cerhan JR, Morton LM, Jemal A, Flowers CR. 2016 US lymphoid malignancy statistics by World Health Organization subtypes. CA Cancer J Clin. 2016;66(6):443-459. 2. Møller MB, Pedersen NT, Christensen BE. Diffuse large B-cell lymphoma: clinical implications of extranodal versus nodal presentation–a population-based study of 1575 cases. Br J Haematol. 2004;124(2):151-159. 3. Messina C, Christie D, Zucca E, Gospodarowicz M, Ferreri AJ. Primary and secondary bone lymphomas. Cancer Treat Rev. 2015;41(3):235-246. 4. Tao R, Allen PK, Rodriguez A, et al. Benefit of consolidative radiation therapy for primary bone diffuse large B-cell lymphoma. Int J Radiat Oncol Biol Phys. 2015;92(1):122-129. 5. NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines®) B-Cell Lymphomas Version 3.2023. Accessed May 11, 2023. https://www.nccn.org/professionals/physician_gls/ pdf/b-cell.pdf” 6. Yahalom J, Illidge T, Specht L, et al. Modern radiation therapy for extranodal lymphomas: field and dose guidelines from the International Lymphoma Radiation Oncology Group. Int J Radiat Oncol Biol Phys. 2015;92(1):11-31. 7. Vitolo U, Seymour JF, Martelli M, et al. Extranodal diffuse large B-cell lymphoma (DLBCL) and primary mediastinal B-cell lymphoma: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2016;27(Suppl 5):v91-v102. 8. Held G, Zeynalova S, Murawski N, et al. Impact of rituximab and radiotherapy on outcome of patients with aggressive B-cell lymphoma and skeletal involvement. J Clin Oncol. 2013;31(32):4115-4122. 9. Lee HY, Kim SJ, Kim K, Ko YH, Kim WS. Bone involvement in patients with stage IV diffuse large B-cell lymphoma: does it

have a prognostic value? Leuk Lymphoma. 2012;53(1):173-175. 10. Bhagavathi S, Micale MA, Les K, Wilson JD, Wiggins ML, Fu K. Primary bone diffuse large B-cell lymphoma: clinicopathologic study of 21 cases and review of literature. Am J Surg Pathol. 2009;33(10):1463-1469. 11. Messina C, Ferreri AJM, Govi S, et al. Clinical features, management and prognosis of multifocal primary bone lymphoma: a retrospective study of the international extranodal lymphoma study group (the IELSG 14 study). Br J Haematol. 2014;164(6):834-840. 12. Müller A, Dreyling M, Roeder F, et al. Primary bone lymphoma: clinical presentation and therapeutic considerations. J Bone Oncol. 2020;25:100326. 13. Jawad MU, Schneiderbauer MM, Min ES, Cheung MC, Koniaris LG, Scully SP. Primary lymphoma of bone in adult patients. Cancer. 2010;116(4):871-879. 14. Bobillo S, Joffe E, Lavery JA, et al. Clinical characteristics and outcomes of extranodal stage I diffuse large B-cell lymphoma in the rituximab era. Blood. 2021;137(1):39-48. 15. Stephens DM, Li H, Constine LS, et al. Extranodal presentation in limited-stage diffuse large B-cell lymphoma as a prognostic marker in three SWOG trials S0014, S0313 and S1001. Haematologica. 2022;107(11):2732-2736. 16. Coiffier B. State-of-the-art therapeutics: diffuse large B-cell lymphoma. J Clin Oncol. 2005;23(26):6387-6393. 17. Juweid ME, Stroobants S, Hoekstra OS, et al. Use of positron emission tomography for response assessment of lymphoma: consensus of the Imaging Subcommittee of International Harmonization Project in Lymphoma. J Clin Oncol. 2007;25(5):571-578. 18. Barrington SF, Qian W, Somer EJ, et al. Concordance between

Haematologica | 109 January 2024

207

ARTICLE - DLBCL with bone lesions, utility of PET/CT assessment

four European centres of PET reporting criteria designed for use in multicentre trials in Hodgkin lymphoma. Eur J Nucl Med Mol Imaging. 2010;37(10):1824-1833. 19. Meignan M, Gallamini A, Meignan M, Gallamini A, Haioun C. Report on the First International Workshop on Interim-PETScan in Lymphoma. Leuk Lymphoma. 2009;50(8):1257-1260. 20. Govi S, Christie D, Messina C, et al. The clinical features, management and prognostic effects of pathological fractures in a multicenter series of 373 patients with diffuse large B-cell lymphoma of the bone. Ann Oncol. 2014;25(1):176-181. 21. Ng AP, Wirth A, Seymour JF, et al. Early therapeutic response assessment by 18FDG-positron emission tomography during chemotherapy in patients with diffuse large B-cell lymphoma: Isolated residual positivity involving bone is not usually a predictor of subsequent treatment failure. Leuk Lymphoma. 2007;48(3):596-600. 22. Schaefer NG, Strobel K, Taverna C, Hany TF. Bone involvement in patients with lymphoma: the role of FDG-PET/CT. Eur J Nucl Med Mol Imaging. 2007;34(1):60-67. 23. 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. 24. Hans CP, Weisenburger DD, Greiner TC, et al. Confirmation of the molecular classification of diffuse large B-cell lymphoma by immunohistochemistry using a tissue microarray. Blood. 2004;103(1):275-282. 25. Moskowitz C, Hamlin PA, Jr., Maragulia J, Meikle J, Zelenetz AD. Sequential dose-dense RCHOP followed by ICE consolidation (MSKCC protocol 01–142) without radiotherapy for patients with primary mediastinal large B cell lymphoma. Blood. 2010;116(21):420. 26. Moskowitz CH, Schöder H, Teruya-Feldstein J, et al. Riskadapted dose-dense immunochemotherapy determined by interim FDG-PET in advanced-stage diffuse large B-cell Lymphoma. J Clin Oncol. 2010;28(11):1896-1903. 27. Schöder H, Polley M-YC, Knopp MV, et al. Prognostic value of

interim FDG-PET in diffuse large cell lymphoma: results from the CALGB 50303 Clinical Trial. Blood. 2020;135(25):2224-2234. 28. Poeschel V, Held G, Ziepert M, et al. Four versus six cycles of CHOP chemotherapy in combination with six applications of rituximab in patients with aggressive B-cell lymphoma with favourable prognosis (FLYER): a randomised, phase 3, noninferiority trial. Lancet. 2019;394(10216):2271-2281. 29. Cashen AF, Dehdashti F, Luo J, Homb A, Siegel BA, Bartlett NL. 18F-FDG PET/CT for early response assessment in diffuse large B-cell lymphoma: poor predictive value of international harmonization project interpretation. J Nucl Med. 2011;52(3):386-392. 30. Micallef IN, Maurer MJ, Wiseman GA, et al. Epratuzumab with rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone chemotherapy in patients with previously untreated diffuse large B-cell lymphoma. Blood. 2011;118(15):4053-4061. 31. Pregno P, Chiappella A, Bellò M, et al. Interim 18-FDG-PET/CT failed to predict the outcome in diffuse large B-cell lymphoma patients treated at the diagnosis with rituximab-CHOP. Blood. 2012;119(9):2066-2073. 32. Mikhaeel NG, Hutchings M, Fields PA, O’Doherty MJ, Timothy AR. FDG-PET after two to three cycles of chemotherapy predicts progression-free and overall survival in high-grade non-Hodgkin lymphoma. Ann Oncol. 2005;16(9):1514-1523. 33. Landersdorfer CB, Bulitta JB, Kinzig M, Holzgrabe U, Sörgel F. Penetration of antibacterials into bone: pharmacokinetic, pharmacodynamic and bioanalytical considerations. Clin Pharmacokinet. 2009;48(2):89-124. 34. Dubey P, Ha CS, Besa PC, et al. Localized primary malignant lymphoma of bone. Int J Radiat Oncol Biol Phys. 1997;37(5):1087-1093. 35. Heyning FH, Hogendoorn PC, Kramer MH, Holland CT, Dreef E, Jansen PM. Primary lymphoma of bone: extranodal lymphoma with favourable survival independent of germinal centre, postgerminal centre or indeterminate phenotype. J Clin Pathol. 2009;62(9):820-824.

Haematologica | 109 January 2024

208

P. Ghione et al.

ARTICLE - Non-Hodgkin Lymphoma

Safety and efficacy of tenalisib in combination with romidepsin in patients with relapsed/refractory T-cell lymphoma: results from a phase I/II open-label multicenter study Swaminathan P. Iyer,1 Auris Huen,1 Weiyun Z. Ai,2 Deepa Jagadeesh,3 Mary J. Lechowicz,4 Craig Okada,5 Tatyana A. Feldman,6 Paola Ghione,7 Juan P. Alderuccio,8 Rebecca Champion,9 Seo-Hyun Kim,10 Ann Mohrbacher,11 Kasi V. Routhu,12 Prajak Barde,12 Ajit M. Nair12 and Bradley M. Haverkos13 1

The University of Texas MD Anderson Cancer Center, Houston, TX, USA; 2Helen Diller Family Comprehensive Cancer Center, San Francisco, CA, USA; 3Taussig Cancer Institute, Cleveland, OH, USA; 4Winship Cancer Institute, Atlanta, GA, USA; 5Oregon Health and Science University, Portland, OR, USA; 6John Theurer Cancer Center at Hackensack University Medical Center, Hackensack, NJ, USA; 7Roswell Park Comprehensive Cancer Center, Buffalo, NY, USA; 8Sylvester Comprehensive Cancer Center, Miami, FL, USA; 9Norton Cancer Institute, Louisville, KY, USA; 10 Rush University Medical Center, Chicago, IL, USA; 11University of Southern California, Los Angeles, CA, USA; 12Rhizen Pharmaceuticals AG., Basel, Switzerland and 13University of Colorado Cancer Center, Aurora, CO, USA

Correspondence: S.P. Iyer SPIyer@mdanderson.org Received: Accepted: Early view:

August 9, 2022 July 5, 2023. July 13, 2023.

Abstract Tenalisib, a selective phosphoinositide-3-kinase δ/γ, and salt-inducible-kinase-3 inhibitor has shown efficacy and was well-tolerated in patients with T-cell lymphoma (TCL). In vitro studies suggest a synergistic anti-tumor potential for the combination of tenalisib with the histone-deacetylase inhibitor, romidepsin. This multicenter, open-label, phase I/II study was designed to characterize the safety, efficacy and pharmacokinetics of oral tenalisib twice-daily and intravenous romidepsin administered on days 1, 8 and 15 in 28-day cycles in adults with relapsed/refractory TCL. Phase I/dose escalation determined the maximum tolerated dose (MTD)/optimal doses of tenalisib and romidepsin. The phase II/dose expansion assessed the safety and anti-tumor activity of the combination at MTD/optimal dose. Overall, 33 patients were enrolled. In dose escalation, no dose-limiting toxicity was identified. Hence, the recommended doses for dose expansion were tenalisib 800 mg twice daily orally, and romidepsin 14 mg/m2 intravenous. Overall treatment-emergent adverse events of any grade reported in >15% of patients were nausea, thrombocytopenia, increased aspartate aminotransferase, increased alanine aminotransferase, decreased appetite, neutropenia, vomiting, fatigue, anemia, dysgeusia, weight loss, diarrhea, and hypokalemia. Twenty-three patients (69.7%) had related grade ≥3 treatment-emergent adverse events. The overall objective response rate in evaluable patients was 63.0% (peripheral TCL: 75% and cutaneous TCL: 53.3%), with a complete response and partial response of 25.9% and 37.0% respectively. The median duration of response was 5.03 months. Co-administration of tenalisib and romidepsin did not significantly alter the pharmacokinetics of romidepsin. Overall, tenalisib and romidepsin combination demonstrated a favorable safety and efficacy profile supporting its further development for relapsed/refractory TCL (clinicaltrials gov. Identifier: NCT03770000).

Introduction T-cell lymphomas (TCL) are a clinically and biologically heterogeneous group of lymphoid malignancies derived from mature T cells and comprise 15% of non-Hodgkin lymphomas (NHL).1,2 TCL are comprised of aggressive (peripheral TCL [PTCL]) and indolent (cutaneous TCL [CTCL]) subtypes.1,2 Most PTCL and advanced CTCL are characterized by poor outcomes. The 5-year overall survival in relapsed/refractory (r/r) cases of PTCL is reportedly

<35%, which is further compounded by the lack of effective treatment options and treatment consensus.3 Although seven drugs are currently approved for r/r TCL treatment,4-10 their activity remains modest. With the exception of brentuximab in anaplastic large cell lymphoma, objective response rate (ORR) varies between 25% to 30%,7-14 posing a challenge in effectively managing these patients. Hence, novel therapeutics are required to treat such patients. Phosphatidylinositol 3-kinase (PI3K) pathway plays an

Haematologica | 109 January 2024

209

ARTICLE - Tenalisib and romidepsin therapy for TCL

S.P. Iyer et al.

important role in cancer pathophysiology.13,14 In recent years, PI3K inhibitors like idelalisib, duvelisib and copanlisib, have shown promising results as monotherapies in the treatment of r/r TCL.15-18 Tenalisib, a highly specific dual equipotent PI3K δ/γ inhibitor, has shown an acceptable safety profile and showed consistent clinical responses in patients with r/r TCL.19-21 The metabolite of tenalisib, IN0385, is an inhibitor of salt-inducible kinase 3 (SIK3).22-24 SIK are known to play a key role in tumorigenesis in solid tumors by modulating several signaling pathways of tumor cells.25 However, the role of SIK3 in lymphomas has not been evaluated. Histone deacetylases (HDAC) pathways are involved in tumorigenesis. HDAC inhibitors like romidepsin, belinostat and vorinostat modulate epigenetic or non-epigenetic regulation, inducing death, apoptosis and cell cycle arrest in cancer cells.26 Romidepsin, is used to treat PTCL and CTCL in patients who have received at least one prior therapy.4,27 HDAC inhibitors combined with other agents have demonstrated enhanced activity in treatment-resistant tumors.28-30 Combining PI3K and HDAC inhibitors in TCL may result in a synergistic or additive response due to their different mechanisms of action. In vitro studies in TCL cell lines suggest that combining PI3K δ/γ and HDAC inhibitors is synergistic.31 In a phase Ib/II clinical study, the response of the combination of romidepsin with duvelisib, a PI3Kδ/γ inhibitor was higher than observed responses as single agents.32 These findings indicate that the combination of PI3K inhibitors with HDAC is promising. We hypothesized that combining tenalisib and romidepsin would improve responses in patients with r/r TCL with a better safety profile. We designed a study to investigate the safety, efficacy, and potential synergistic effects of tenalisib and romidepsin in these patients. Given the overlapping metabolic pathways of romidepsin and tenalisib,33 extensive pharmacokinetic (PK) assessments were also planned to rule out any drug-drug interaction.

Dose escalation phase The phase I, 3+3 dose escalation study assessed the maximum tolerated dose (MTD)/optimal dose of the combination. Three dose escalation cohorts were planned. For cohort 1 to cohort 3, tenalisib doses were 400, 600, and 800 mg orally twice daily (BID). The corresponding doses of romidepsin were 12, 12, and 14 mg/m2. The study allowed adding or reducing the number of cohorts based on emerging safety and PK data (details are provided in the Online Supplementary Figure S1). Dose expansion phase The phase II study assessed the safety and anti-tumor activity of tenalisib and romidepsin combination at the MTD/optimal dose. Twelve each of PTCL and CTCL patients were to be enrolled in the two patient groups (Online Supplementary Figure S1). Details of the study participants are provided in the Online Supplementary Table S1 and the Online Supplementary Method S1. Treatment and intervention Eligible patients received tenalisib orally, BID, at the same time each day, 1 hour before their meals over a 28-day cycle (day 1–28). Romidepsin was administered intravenously (IV) over 4 hours on days 1, 8, and 15 during the 28-day cycle. Both tenalisib and romidepsin were given until disease progression or discontinuation from the study. Additional details on treatment and compliance are provided in the Online Supplementary Method S2.

Objectives and endpoints Primary objective: to characterize the safety and tolerability and determine the MTD of the combination. Safety endpoints included assessments of adverse events (AE), treatment-emergent AE (TEAE), serious AE (SAE), and DLT (the definition and details of DLT are provided in the Online Supplementary Table S2). The key secondary objectives: (i) ORR: sum of complete response [CR] and partial response [PR] rates), evaluated in PTCL patients according to the Lugano Classification,34 and in CTCL patients as per the global response score,35 Methods (ii) the duration of response (DoR) calculated as the time Study design and participants from the initial response to documented disease progresThis was a multicenter, open-label, non-randomized, sion; and (iii) the PK of tenalisib and romidepsin. Additional two-stage phase I/II study of tenalisib combined with ro- PK assessment details are provided in Online Supplemenmidepsin in adult patients with r/r TCL (clinicaltrails gov. tary Method S3. The details of the study procedures are Identifier: NCT03770000). This TCL cohort represents pa- provided in the Online Supplementary Method S4. tients with either PTCL or CTCL. The study, conducted from April 2019 to May 2021, included dose escalation Statistical analysis and dose expansion phases. Each site’s Institutional Re- Dose escalation phase: three patients per cohort were apview Board approved the study protocol. The study was propriate for assessing MTD/optimal dose. conducted following the International Council for Har- Dose expansion phase: 12 patients per group were conmonization Guideline for Good Clinical Practice and the sidered appropriate for assessing the preliminary antiDeclaration of Helsinki. tumor activity of tenalisib and romidepsin combination. Haematologica | 109 January 2024

210

ARTICLE - Tenalisib and romidepsin therapy for TCL

S.P. Iyer et al.

dose of tenalisib (800 mg BID) and romidepsin (14 mg/m2) combination was considered the MTD/RP2D dose for the study. The demographic and baseline characteristics for all patients are shown in Table 2. Among the 33 patients enrolled, 51.5% were male, the majority (81.8%) being Caucasian. The median age of patients was 66.2 years (range, 42.9-83.4 years). The majority of patients had an Eastern Cooperative Oncology Group (ECOG) performance status of 0 (42.4%) or 1 (54.6%). Twenty-two (66.6%) patients (87.6% in PTCL, 47.1% in CTCL) had stage 3/stage 4 diseases. Overall, 36.4% of patients had relapsed following the last prior therapy (31.3% in PTCL, 41.2% in CTCL). Twenty-one (63.6%) patients (68.8% in PTCL, 58.8% in Results CTCL) were refractory to their last prior therapy. The Patient demographics, baseline characteristics, and median duration from the date of the last prior therapy to disposition treatment was 46 days (range, 8-702 days). The median Overall, 56 patients were screened, of which 33 patients compliance in the study was 100% (range, 99.1-105.3) and were enrolled and received the combination treatment. Of was within the acceptable range of 80-120%. these, 16 PTCL and 17 CTCL patients received tenalisib and romidepsin. Twenty-three patients were screen failures. Safety Of the PTCL patients, 93.7% and all CTCL patients discon- Overall, at least one related TEAE of grade 1 or 2 severity tinued the study. The most common reasons for study was reported in all enrolled patients. In 69.7% of patients, discontinuation were disease progression (57.57%) and AE grade ≥3 were reported (68.8% of PTCL patients, and (18.18%). Three PTCL patients (18.7%) were bridged to 70.6% of CTCL patients). transplant and hence moved out from the study (Table 1; The related TEAE reported in >15% of patients were Online Supplementary Figure S2). No DLT were reported nausea (72.7%); thrombocytopenia (57.6%); fatigue for the combination doses ranging between 400 mg to 800 (54.5%); increased aspartate aminotransferase (AST) mg BID and romidepsin IV 12–14 mg/m2. Hence, the highest (30.3%); increased alanine aminotransferase (ALT) Data were summarized using descriptive statistics for continuous variables and frequencies and percentages for categorical variables. Results from the study are presented by indication. All analyses were performed using SAS® software version 9.4 or higher. All safety analyses were performed on the safety dataset that comprised all patients who received at least one dose of study medication. The efficacy analysis was performed on the modified intent-to-treat (mITT) population (evaluable patients). Further details are provided in the Online Supplementary Method S5.

Table 1. Study disposition by indication.

Patients dosed Discontinued study Reason for discontinuation** Adverse event Disease progression Investigator’s decision Intolerance of study drug Patient bridged to yransplant Withdrawal of consent Drug interruptions due to AE Non-related AE Related AE Dose reduction due to AE Non-related AE Related AE Drug withdrawn permanently due to AE Non-related AE Related AE

PTCL, N (%) 16 15 (93.7)

CTCL, N (%) 17 17 (100)

Overall, N (%) 33 32 (96.9)

1 (6.25) 10 (62.5) 0 0 3 (18.7) 1 (6.25) 12 (75.0) 3 (18.8) 10 (62.5) 7 (43.8) 0 7 (43.8) 1 (6.3) 1 (6.3) 0

5 (29.4) 9 (52.9) 1 (5.8) 1 (5.8) 0 1 (5.8) 12 (70.6) 3 (17.7) 12 (70.6) 8 (47.1) 1 (5.9) 8 (47.1) 5 (29.4) 0 5 (29.4)

6 (18.18) 19 (57.57) 1 (3.03) 1 (3.03) 3 (9.09) 2 (6.06) 24 (72.7) 6 (18.2) 22 (66.7) 15 (45.5) 1 (3.0) 15 (45.5) 6 (18.2) 1 (3.0) 5 (15.2)

Cutaneous T-cell lymphoma (CTCL) percentages are based on the total number of patients dosed in the study. **Percentages for reason of discontinuation are based on the total number of patients discontinued. AE: adverse events; N (%): number (percentage) of patients; PTCL: peripheral T-cell lymphoma.

Haematologica | 109 January 2024

211

ARTICLE - Tenalisib and romidepsin therapy for TCL

S.P. Iyer et al.

(24.2%); decreased appetite, neutropenia, and vomiting (27.3% each); anemia and dysgeusia (21.2% each); weight decrease (18.2%); and diarrhea and hypokalemia (15.2% each) (Table 3). No DLT were reported. Twelve patients (36.4%) reported 19 serious TEAE. Six TEAE led to permanent study treatment discontinuation. Two patients (1 PTCL and 1 CTCL) died due to sepsis; of these, an event of sepsis in the CTCL patient was considered possibly related to the combination (at tenalisib 800 mg BID and ro-

midepsin 14 mg/m2). This was an 81-year-old male with a past medical history of asthma, chronic obstructive pulmonary disease, lung cancer, prostate cancer, diabetes mellitus, hypertension, and stage 4 CTCL. The patient had multiple underlying comorbidities and risk factors. Disease progression could not be completely ruled out based on CT findings. Given this patient’s underlying comorbidities and risk factors, there seemed to be no conclusive evidence that the combination ac-

Table 2. Demographics and baseline characteristics – by indication (safety analyses set).

Age in years Median Min-Max Sex, N (%) Male Female Race, N (%) White Asian Black/African/American Time in days from initial diagnosis to treatment* Median Min-Max PTCL subtype, N (%) PTCL, NOS ALCL T-cell follicular lymphoma (including AITL) CTCL subtype, N (%) MF Sézary syndrome Staging at Screening, N (%) III IV Outcome of the last prior therapy, N (%) Relapse Refractory Time from date of last prior therapy to study treatment in days* Median Min-Max Prior therapies, N Median Min-Max Prior therapies, N (%) ≥3 ≥5 ECOG Performance Status**, N (%) 0 1 2

PTCL, N=16

CTCL, N=17

Overall, N=33

61.85 42.89-83.44

66.53 48.76-80.55

66.21 42.89-83.44

9 (56.3) 7 (43.8)

8 (47.1) 9 (52.9)

17 (51.5) 16 (48.5)

13 (81.3) 2 (12.5) 1 (6.3)

14 (82.4) 0 3 (17.7)

27 (81.8) 2 (6.1) 4 (12.1)

583.5 53-2,643

1,446 253-5,189

754 53-5,189

8 (50) 1 (6.3) 7 (43.8)

NA NA NA

8 (24.2) 1 (3.0) 7 (21.2)

NA NA

12 (70.6) 5 (29.4)

12 (36.4) 5 (15.2)

5 (31.3) 9 (56.3)

1 (5.9) 7 (41.2)

6 (18.2) 16 (48.5)

5 (31.3) 11 (68.8)

7 (41.2) 10 (58.8)

12 (36.4) 21 (63.6)

44 13-702

46 8-435

46 8-702

3 1-5

6 1-17

3 1-17

10 (62.5) 2 (12.5)

15 (88.2) 10 (58.8)

25 (75.8) 12 (36.4)

8 (50.0) 8 (50.0) 0

6 (35.3) 10 (58.8) 1 (5.9)

14 (42.4) 18 (54.6) 1 (3.0)

*Partial dates are imputed using the missing data conventions as mentioned in the statistical analysis plan. **The baseline measurement is the last pretreatment measurement taken on or before cycle 1 day 1. AITL: angioimmunoblastic T-cell lymphoma; ALCL: anaplastic large-cell lymphoma; CTCL: cutaneous T-cell lymphoma; ECOG: Eastern Cooperative Oncology Group; MF: mycosis fungoides; N (%): number (percentage) of patients; NA: not applicable; N: number of evaluable patients; NOS: not otherwise specified; PTCL: peripheral T-cell lymphoma; SD: standard deviation. Haematologica | 109 January 2024

212

ARTICLE - Tenalisib and romidepsin therapy for TCL

S.P. Iyer et al.

centuated or worsened the known adverse effect of individual agents. Overall, in both PTCL and CTCL groups, drug interruptions were observed in 60.6%, 63.6%, and 45.5% of patients for tenalisib, romidepsin, and the combination, respectively. Three or more drug interruptions occurred in 12.1% of patients for tenalisib and 24.2% of patients for romidepsin during the combination treatment. Dose reductions occurred in 27.3%, 33.3%, and 15.2% of patients for tenalisib, romidepsin, and the combination respectively.

olite, IN0385, the exposure was higher for IN0385 (Online Supplementary Table S3). Romidepsin plasma levels were not significantly reduced on co-administration with tenalisib for Cmax and AUC0-t (Figure 1). Efficacy A total of 27 patients (12 PTCL and 15 CTCL) were considered evaluable for efficacy. Six patients (4 PTCL and 2 CTCL) discontinued early in the first cycle, due to disease progression, drug toxicity, and consent withdrawal (2 patients each) and were considered non-evaluable.

Pharmacokinetics Tenalisib was rapidly absorbed in all three cohorts, with the Cmax and AUC0-t increasing with dose. Although the elimination kinetics of tenalisib were similar to its metab-

Objective response rate In evaluable patients, ORR was 75% in PTCL group and 53.3% in CTCL patients. The disease control rate (DCR) was 91.7% in the PTCL patients and 86.7% in the CTCL pa-

Table 3. Overall related treatment-emergent adverse events of any grade reported in >15% patients and corresponding grade ≥3 treatment-emergent adverse events reported - by indication (safety analysis set).

System organ class/ preferred term

PTCL, N=16 N (%), E

CTCL, N=17 N (%), E

Overall, N=33 N (%), E

Any grade

Grade ≥3

Any grade

Grade ≥3

Any grade

Grade ≥3

16 (100.0), 212

11 (68.8), 34

17 (100.0), 191

12 (70.6), 32

33 (100.0), 403

23 (69.7), 66

14 (87.5), 81

8 (50.0), 25

9 (52.9), 39

3 (17.6), 11

23 (69.7), 120

11 (33.3), 36

12 (75.0), 48

6 (37.5), 13

7 (41.2), 14

1 (5.9), 3

19 (57.6), 62

7 (21.2), 16

Neutropenia

6 (37.5), 29

3 (18.8), 11

3 (17.6), 19

2 (11.8), 7

9 (27.3), 48

5 (15.2), 18

Anemia

2 (12.5), 4

1 (5.9), 1

5 (29.4), 6

1 (5.9), 1

7 (21.2), 10

2 (6.1), 2

12 (75.0), 29

13 (76.5), 34

25 (75.8), 63

Nausea

11 (68.8), 14

13 (76.5), 21

24 (72.7), 35

Vomiting

3 (18.8), 5

6 (35.3), 7

9 (27.3), 12

4 (25.0), 6

1 (5.9), 1

5 (15.2), 7

8 (50.0), 13

1 (6.3), 1

12 (70.6), 19

2 (11.8), 2

20 (60.6), 32

3 (9.1), 3

8 (50.0), 9

1 (6.3), 1

10 (58.8), 12

1 (5.9), 1

18 (54.5), 21

2 (6.1), 2

9 (56.3), 34

2 (12.5), 4

9 (52.9), 65

6 (35.3), 12

18 (54.5), 99

8 (24.2), 16

3 (18.8), 5

7 (41.2), 21

1 (5.9), 1

10 (30.3), 26

1 (3.0), 1

2 (12.5), 7

2 (12.5), 2

6 (35.3), 19

4 (23.5), 6

8 (24.2), 26

6 (18.2), 8

3 (18.8), 4

3 (17.6), 4

6 (18.2), 8

8 (50.0), 18

1 (6.3), 1

7 (41.2), 14

1 (5.9), 1

15 (45.5), 32

2 (6.1), 2

5 (31.3), 5

4 (23.5), 8

9 (27.3), 13

4 (25.0), 4

1 (5.9), 1

5 (15.2), 5

Nervous system disorders

5 (31.3), 9

1 (6.3), 1

5 (29.4), 8

10 (30.3), 17

1 (3.0), 1

Dysgeusia Skin and subcutaneous tissue disorders

4 (25.0), 4

3 (17.6), 3

7 (21.2), 7

4 (25.0), 8

2 (11.8), 5

2 (11.8), 4

6 (18.2), 13

2 (6.1), 4

At least one TEAE Blood and lymphatic system disorders Thrombocytopenia

Gastrointestinal disorders

Diarrhea General disorders and administration site conditions Fatigue Investigations Aspartate aminotransferase increased Alanine aminotransferase increased Weight decreased Metabolism and nutrition disorders Decreased appetite Hypokalemia

CTCL: cutaneous T-cell lymphoma; E: events; N (%): number (percentage) of patients; N: number of evaluable patients; PTCL: peripheral Tcell lymphoma; TEAE: treatment-emergent adverse events. Haematologica | 109 January 2024

213

ARTICLE - Tenalisib and romidepsin therapy for TCL

S.P. Iyer et al.

tients. CR was observed in 50% of PTCL patients (2 patients with PTCL not otherwise specified [NOS] and 4 of angioimmunoblastic T-cell lymphoma [AITL]) and 6.7% of CTCL patients (1 with Sezary syndrome [SS]). PR was observed in 25% of PTCL patients (2 patients with PTCL NOS and 1 with AITL) and 46.7% of CTCL patients (4 patients with mycosis fungoides [MF] and 3 with SS) (Table 4). Nine PTCL patients (CR: 6 patients, PR: 3 patients) and eight CTCL patients (CR: 1 patient, PR: 7 patients) had at least 50% improvement in the nodal lesions and modified severity weighted assessment tool (mSWAT) score respectively (Figure 2A, B).

nine patients were confirmed as a CR or PR, with a median DoR of 5.03 months (range, 1.87-25.23+ months). Of the 15 patients with CTCL, eight patients were confirmed as a CR or PR, with a median DoR of 3.80 months (range, 0.8730.83 months) (Table 4; Figure 2C).

Discussion

In previous studies, single agent tenalisib and romidepsin have been well-tolerated with reasonable response rates in r/r TCL patients.19,21,36 Thus, combining these two agents were expected to improve responses in these patients, Duration of response considering the synergistic anti-tumor activity of the comOverall, 17 of 27 patients were confirmed as a CR or PR, bination seen in preclinical studies and with similar class with a median DoR of 5.03 months (range, 0.87-30.83 combinations in the clinic. The safety profile was unlikely months) (Table 4; Figure 2C). Of the 12 patients with PTCL, to be affected due to the non-overlapping toxicities of the

Figure 1. Mean plasma concentrations of romidepsin. (A) Cohort 1. (B) Cohort 2. (C) Cohort 3. C1D1: cycle 1 day 1; C1D8: cycle 1 day 8; N: number of patients. Haematologica | 109 January 2024

214

ARTICLE - Tenalisib and romidepsin therapy for TCL

S.P. Iyer et al.

single agents. In line with these assumptions, the study revealed that the highest dose of tenalisib 800 mg BID and romidepsin IV 14 mg/m2 was found to be tolerable with expected AE. In this study, no DLT were reported with the combination even at the highest dose. This finding is in contrast to studies evaluating other drug combinations with romidepsin (e.g., with pralatrexate or oral 5-azacytidine combination)37,38 where DLT such as thrombocytopenia, neutropenia, and sepsis were reported. Discontinuation rates due to adverse events were around 18% in our study, which is in line with other single agent studies of romidepsin (19-28%).39,40 Despite the high number of dose interruptions/dose reductions reported with the combination, it was in line with other single agent studies of romidepsin. Thus, discontinuation rates or dose modifications did not increase with the combination as compared to single agent romidepsin.41 Romidepsin is primarily metabolized by cytochrome P450 3A4 (CYP3A4),41 while tenalisib and its metabolite IN0385 are moderate inhibitors and substrates for CYP3A4. Due to the CYP3A4 inhibitory potential of tenalisib, there could have been an increase in plasma concentrations of romidepsin leading to increased severity and frequency of romidepsin-induced toxicities. However, analyses in our study revealed that the co-administration of romidepsin and tenalisib did not significantly alter the pharmacokinetic profiles of romidepsin. The study revealed that there were no unexpected AE, or increased incidence of existing AE observed for individual regimens. This was validated by the pharmacokinetic data which showed no drug-drug interaction. Thus, the PK data findings suggest that in a clinical setting, romidepsin and tenalisib do not interact and can be administered as a combination devoid of drug-drug interactions. Although in the study, the efficacy evaluable analysis population considered only patients who had had at least one post baseline assessment (C3D1) when compared with pivotal studies which considers all patients who had one dose of the drug, the combination showed encouraging anti-tumor activity in patients with TCL in our study. For PTCL, 75% of evaluable patients achieved ORR with a CR of 50%. The CR rates observed using this combination were higher than that reported individually with tenalisib and romidepsin monotherapy in PTCL patients (tenalisib CR: 20.0%, romidepsin CR: 14%),21 suggesting synergism. The ORR for the combination seemed to be additive. The ORR in CTCL patients was numerically higher compared to individual single agents tenalisib and romidepsin but not as high as observed in the PTCL population. In the PTCL patients, the median duration of treatment was 3.6 months (range, 0.1-28.96+ months) (Figure 2C). Nine of 16 patients completed seven cycles of treatment

and one patient continues to be on combination therapy for more than 1.5 years. However, the DoR was impacted by three patients who were bridged to transplant and thus taken off the study. In the CTCL patients, the median duration of treatment was 3.46 months (range, 0.76-34.53 months) (Figure 2C). Three patients were treated for seven cycles. Out of these three patients, one patient with Sezary syndrome (SS) was in remission for more than 30 months. Long-term treatment in these patients indicates that tenalisib has been well tolerated without any long-term immune-mediated toxicities associated with PI3K inhibitors, such as colitis or pneumonitis. When the study was conceptualized and began enrolling, romidepsin was approved for both PTCL and CTCL in patients who had received one prior therapy. In a pivotal phase III study of romidepsin plus CHOP (cyclophosphamide, doxorubicin, vincristine and prednisone) in previously untreated PTCL, the addition of romidepsin to CHOP did not improve the responses (progression-free survival, response rates, or overall survival) over CHOP alone but led to increased frequency of grade ≥3 AE.42 As a result, the label for the use of romidepsin in PTCL was withdrawn. However, romidepsin is still approved for CTCL. Given the results seen with tenalisib being combined with romidepsin, the authors believe that the development of the combination should still be explored in both patients with PTCL and CTCL in a larger clinical Table 4. Objective response rate and duration of response – by indication (mITT analysis set).

PTCL, N=12 N (%) 95% CI CR 6 (50.0) 21.09-78.91 PR 3 (25.0) 5.49-57.19 SD 2 (16.7) 2.09-48.41 PD 1 (8.3) 0.21-38.48 ORR (CR+PR) 9 (75.0) 42.81-94.51 DCR (CR+PR+SD) 11 (91.7) 61.52-99.79 Duration of response Median duration of 5.03 response in months (1.87-25.3+) (range)

CTCL, N=15 N (%) 95% CI 1 (6.7) 0.17-31.95 7 (46.7) 21.27-73.41 5 (33.3) 11.82-61.62 2 (13.3) 1.66-40.46 8 (53.3) 26.59-78.73 13 (86.7) 59.54-98.34 3.80 (087-30.83)

CR: complete response; CTCL: cutaneous T-cell lymphoma; DCR: disease control rate; mITT: modified intent-to-treat; N: number of evaluable patients for efficacy; N (%): number (percentage) of patients; ORR: overall response rate; PD: progression of disease; PR: partial response; PTCL: peripheral T-cell lymphoma; SD: stable disease.

Haematologica | 109 January 2024

215

ARTICLE - Tenalisib and romidepsin therapy for TCL

S.P. Iyer et al.

study. In addition, given the favorable results seen in PTCL, the combinations of tenalisib with other approved HDAC inhibitors in PTCL can also be explored. Our study had several strengths. We evaluated drug-drug interaction to establish the safety of the combination. The study classified the patient population separately into

PTCL and CTCL groups in the dose expansion part, allowing for the differential investigation of the safety and efficacy of the drug combination in these populations. Our study was limited by sample size as is seen in the early phase of the drug development. Overall, the combination of tenalisib and romidepsin dem-

Figure 2. Treatment duration and response in evaluable peripheral T-cell lymphoma and cutaneous T-cell lymphoma patients. (A) Percentage change in nodal size from baseline. *Disease progression due to new lesions. (B) Percentage change in modified severity-weighted assessment tool (SWAT) score from baseline. (C) Treatment duration in peripheral T-cell lymphoma (PTLC) and cutaneous T-cell lymphoma (CTCL) patients. AITL: angioimmunoblastic T-cell lymphoma; ALCL: anaplastic large cell lymphoma; SS: Sezary syndrome; CR: complete response; MF: mycosis fungoides; NOS: not otherwise specified; PD: progression of disease; PR: partial response; SCT: stem cell transplant; SD: stable disease. Haematologica | 109 January 2024

216

ARTICLE - Tenalisib and romidepsin therapy for TCL

S.P. Iyer et al.

onstrates potential in patients with hematological malignancies (PTCL/CTCL). This supports further development of this combination for treating TCL. Disclosures SPI receives research support from CRISPR Therapeutics, Myeloid, Merck, Seagen, Ono, Rhizen, Acrotech, Legend, Innate Pharma, Astra Zeneca, Dren Bio, Yingli, and Secura Bio; participates in scientific advisory boards for Seagen, Yingli, and Secura Bio; and participates as speaker for American Society of Hematology, American Society of Cellular Therapy and Transplantation, Biocure and Targeted Oncology’s speaker bureaus. PG reports research funding from Kite Pharma, consulting from AstraZeneca, Kyowa Hakko Kirin, Secura Bio. JPA has served on the advisory board and received research support of ADC Therapeutics. An immediate family member has served on the advisory boards of Agios Pharmaceuticals, and Forma Therapeutics. BH serves as an advisor/consultant for Viracta Therapeutics. AM, S-HK and CO have no conflicts of interest to disclose. WZA has served on the advisory board and is a consultant for Secura Bio, ADC Therapeutics, and More Health. AMN, KVR and PB are full time employees of Rhizen Pharmaceuticals AG., Basel, Switzerland. TAF was on the speaker's bureau of ADC therapeutics, AstraZeneca, Daiichi Sankyo, Genmab, Karyopharm, Kite, MorphoSys and Secura Bio and served on the advisory board or as a consultant for Bristol Myers Squibb, Celgene Corporation, Janssen Biotech, Inc., Pharmacyclics LLC, Seattle Genetics, Inc., and Takeda. DJ reports research funding from AstraZeneca, Atara Biotherapeutics, Debio Pharma, Loxo Pharmaceuticals, MEI Pharma, Regeneron Pharmaceuticals, Seagen and Trillium Pharmaceuticals and served on the advisory board of Daiichi Sankyo. RC was part of the speaker’s bureau of Bristol Myers Squibb and Abbvie. MJL has either served on the advisory board or has received speaker fee from Kyowa Kirin, University of Pennsylvania, Secura Bio Inc. and EUSA Pharma. AH received research support from Rhizen, Kyowa Kirin, Elorac, Trillium, Innate and is a consultant for Kyowa Kirin.

Contributions SPI contributed to design, recruitment, data collection, result interpretation, critical review and approval of the manuscript. AH and WZA contributed to patient recruitment, data collection, review, and final approval of the manuscript. DJ contributed to recruitment, data collection, critical review and approval of the manuscript. MJL contributed to patient recruitment, data collection, critical review and final approval of the manuscript. CO, TAF, RC and SHK contributed to patient recruitment, data collection, review, and final approval of the manuscript. PG and JPA contributed to recruitment, data collection, review and approval of the manuscript. AM contributed to patient recruitment, data collection, review, editing and final approval of the manuscript. KVR contributed to conception, design, analysis, result interpretation, critical review and approval of the manuscript. PB and AMN contributed to conception, design, result interpretation, critical review and approval of the manuscript. BMH contributed to recruitment, data collection, result interpretation, critical review, editing and final approval of the manuscript. Acknowledgments The authors thank the patients and the study site staff for their participation and cooperation during the entire study. The authors acknowledge Dr Timothy M. Kuzel and Dr Don A. Stevens for their contributions in terms of patient recruitment and guidance for the study conduct at their respective study sites. The authors would like to acknowledge Lakshmi Hariharan and Dr Priya Singh from SciVoc Healthcare Pvt. Ltd. for providing manuscript writing and editing services. Funding The sponsorship for this study and publication fee were funded by Rhizen Pharmaceuticals AG., Basel, Switzerland. Data-sharing statement All data related to the study findings are described in the manuscript and the related Online Supplementary Appendix.

References 1. Rodd AL, Ververis K, Karagiannis TC. Current and emerging therapeutics for cutaneous T-cell lymphoma: histone deacetylase inhibitors. Lymphoma. 2012;2012:1-10. 2. Lauenborg B, Christensen L, Ralfkiaer U, et al. Malignant T cells express lymphotoxin α and drive endothelial activation in cutaneous T cell lymphoma. Oncotarget. 2015;6(17):15235-15249. 3. Pellegrini C, Dodero A, Chiappella A, et al. A phase II study on the role of gemcitabine plus romidepsin (GEMRO regimen) in the treatment of relapsed/refractory peripheral T-cell lymphoma patients. J Hematol Oncol. 2016;9:38. 4. Foss F, Pro B, Miles Prince H, et al. Responses to romidepsin by

line of therapy in patients with relapsed or refractory peripheral T-cell lymphoma. Cancer Med. 2017;6(1):36-44. 5. Duvic M, Martin AG, Kim Y, et al. Phase 2 and 3 clinical trial of oral bexarotene (Targretin capsules) for the treatment of refractory or persistent early-stage cutaneous T-cell lymphoma. Arch Dermatol. 2001;137(5):581-593. 6. Mann BS, Johnson JR, Cohen MH, Justice R, Pazdur R. FDA approval summary: vorinostat for treatment of advanced primary cutaneous T-cell lymphoma. Oncologist. 2007;12(10):1247-1252. 7. O'Connor OA, Horwitz S, Masszi T, et al. Belinostat in patients with relapsed or refractory peripheral T-cell lymphoma: results

Haematologica | 109 January 2024

217

ARTICLE - Tenalisib and romidepsin therapy for TCL

S.P. Iyer et al.

of the pivotal phase II BELIEF (CLN-19) study. J Clin Oncol. 2015;33(23):2492-2499. 8. O'Connor OA, Pro B, Pinter-Brown L, et al. Pralatrexate in patients with relapsed or refractory peripheral T-cell lymphoma: results from the pivotal PROPEL study. J Clin Oncol. 2011;29(9):1182-1189. 9. Ogura M, Ishida T, Hatake K, et al. Multicenter phase II study of mogamulizumab (KW-0761), a defucosylated anti-cc chemokine receptor 4 antibody, in patients with relapsed peripheral T-cell lymphoma and cutaneous T-cell lymphoma. J Clin Oncol. 2014;32(11):1157-1163. 10. Flinn IW, O'Brien S, Kahl B, et al. Duvelisib, a novel oral dual inhibitor of PI3K-δ,γ, is clinically active in advanced hematologic malignancies. Blood. 2018;131(8):877-887. 11. Duvic M, Tetzlaff MT, Gangar P, Clos AL, Sui D, Talpur R. Results of a phase II trial of brentuximab vedotin for CD30+ cutaneous T-cell lymphoma and lymphomatoid papulosis. J Clin Oncol. 2015;33(32):3759-3765. 12. Prince HM, Kim YH, Horwitz SM, et al. Brentuximab vedotin or physician's choice in CD30-positive cutaneous T-cell lymphoma (ALCANZA): an international, open-label, randomised, phase 3, multicentre trial. Lancet. 2017;390(10094):555-566. 13. Janku F, Yap TA, Meric-Bernstam F. Targeting the PI3K pathway in cancer: are we making headway? Nat Rev Clin Oncol. 2018;15(5):273-291. 14. Yang J, Nie J, Ma X, Wei Y, Peng Y, Wei X. Targeting PI3K in cancer: mechanisms and advances in clinical trials. Mol Cancer. 2019;18(1):26. 15. Katsuya H, Cook LBM, Rowan AG, Satou Y, Taylor GP, Bangham CRM. Phosphatidylinositol 3-kinase-δ (PI3K-δ) is a potential therapeutic target in adult T-cell leukemia-lymphoma. Biomark Res. 2018;6:24. 16. Flinn IW, Kahl BS, Leonard JP, et al. Idelalisib, a selective inhibitor of phosphatidylinositol 3-kinase-δ, as therapy for previously treated indolent non-Hodgkin lymphoma. Blood. 2014;123(22):3406-3413. 17. Flinn IW, O'Brien S, Kahl B, et al. Duvelisib, a novel oral dual inhibitor of PI3K-δ,γ, is clinically active in advanced hematologic malignancies. Blood. 2018;131(8):877-887. 18. Dreyling M, Panayiotidis P, Egyed M, et al. Efficacy of copanlisib monotherapy in patients with relapsed or refractory marginal zone lymphoma: subset analysis from the CHRONOS-1 trial. Blood. 2017;130(Suppl 1):S4053. 19. Carlo-Stella C, Delarue R, Scarfo L, et al. A First-in-human study of tenalisib (RP6530), a dual PI3K δ/γ inhibitor, in patients with relapsed/refractory hematologic malignancies: results from the European Study. Clin Lymphoma Myeloma Leuk. 2020;20(2):78-86. 20. Grosicki S, Iosava G, Zodelava M, et al. An open label, phase 2 study to assess the efficacy and safety of tenalisib (RP6530), a PI3K δ/γ and SIK3 inhibitor, in patients with relapsed/refractory chronic lymphocytic leukemia (CLL). Blood. 2020;136(Suppl 1):S25. 21. Huen A, Haverkos BM, Zain J, et al. Phase I/Ib study of tenalisib (RP6530), a dual PI3K δ/γ inhibitor in patients with relapsed/refractory T-cell lymphoma. Cancers (Basel). 2020;12(8):2293. 22. Chen F, Chen L, Qin Q, Sun X. Salt-inducible kinase 2: an oncogenic signal transmitter and potential target for cancer therapy. Front Oncol. 2019;9:18. 23. Bon H, Wadhwa K, Schreiner A, et al. Salt-inducible kinase 2 regulates mitotic progression and transcription in prostate cancer. Mol Cancer Res. 2015;13(4):620-635. 24. Sorrentino A, Menevse AN, Michels T, et al. Salt-inducible kinase

3 protects tumor cells from cytotoxic T-cell attack by promoting TNF-induced NF-κB activation. J Immunother Cancer. 2022;10(5):e004258. 25. Charoenfuprasert S, Yang YY, Lee YC, et al. Identification of saltinducible kinase 3 as a novel tumor antigen associated with tumorigenesis of ovarian cancer. Oncogene. 2011;30(33):3570-3584. 26. Eckschlager T, Plch J, Stiborova M, Hrabeta J. Histone deacetylase inhibitors as anticancer drugs. Int J Mol Sci. 2017;18(7):1414. 27. VanderMolen K, McCulloch W, Pearce C, et al. Romidepsin (Istodax, NSC 630176, FR901228, FK228, depsipeptide): a natural product recently approved for cutaneous T-cell lymphoma. J Antibiot. 2011;64:525-531. 28. Sermer D, Pasqualucci L, Wendel HG, Melnick A, Younes A. Emerging epigenetic-modulating therapies in lymphoma. Nat Rev Clin Oncol. 2019;16(8):494-507. 29. Hontecillas-Prieto L, Flores-Campos R, Silver A, de Álava E, Hajji N, García-Domínguez DJ. Synergistic enhancement of cancer therapy using HDAC inhibitors: opportunity for clinical trials. Front Genet. 2020;11:578011. 30. Ranganna K, Selvam C, Shivachar A, Yousefipour Z. Histone deacetylase inhibitors as multitarget-directed Epi-drugs in blocking PI3K oncogenic signaling: a polypharmacology approach. Int J Mol Sci. 2020;21(21):8198. 31. Moskowitz AJ, Koch R, Mehta-Shah N, et al. In vitro, in vivo, and parallel phase I evidence support the safety and activity of duvelisib, a PI3K-δ,γ inhibitor, in combination with romidepsin or bortezomib in relapsed/refractory T-cell lymphoma. Blood. 2017;130(Suppl 1):S819. 32. Horwitz SM, Moskowitz AJ, Jacobsen ED, et al. The Combination of duvelisib, a PI3K-δ,γ inhibitor, and romidepsin is highly active in relapsed/refractory peripheral T-cell lymphoma with low rates of transaminitis: results of parallel multicenter, phase 1 combination studies with expansion cohorts. Blood. 2018;132(Suppl 1):S683. 33. Yan Z, Zhang K, Ji M, Xu H, Chen X. A Dual PI3K/HDAC inhibitor downregulates oncogenic pathways in hematologic tumors in vitro and in vivo. Front Pharmacol. 2021;12:741697. 34. 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. 35. Olsen EA, Whittaker S, Kim YH, et al; International Society for Cutaneous Lymphomas; United States Cutaneous Lymphoma Consortium; Cutaneous Lymphoma Task Force of the European Organisation for Research and Treatment of Cancer. Clinical end points and response criteria in mycosis fungoides and Sézary syndrome: a consensus statement of the International Society for Cutaneous Lymphomas, the United States Cutaneous Lymphoma Consortium, and the Cutaneous Lymphoma Task Force of the European Organisation for Research and Treatment of Cancer. J Clin Oncol. 2011;29(18):2598-2607. 36. Duvic M, Bates SE, Piekarz R, et al. Responses to romidepsin in patients with cutaneous T-cell lymphoma and prior treatment with systemic chemotherapy. Leuk Lymphoma. 2018;59(4):880-887. 37. Amengual JE, Lichtenstein R, Lue J, et al. A phase 1 study of romidepsin and pralatrexate reveals marked activity in relapsed and refractory T-cell lymphoma. Blood. 2018;131(4):397-407. 38. O'Connor OA, Falchi L, Lue JK, et al. Oral 5-azacytidine and romidepsin exhibit marked activity in patients with PTCL: a multicenter phase 1 study. Blood. 2019;134(17):1395-1405. 39. Coiffier B, Pro B, Prince HM, et al. Results from a pivotal, open-

Haematologica | 109 January 2024

218

ARTICLE - Tenalisib and romidepsin therapy for TCL

S.P. Iyer et al.

label, phase II study of romidepsin in relapsed or refractory peripheral T-cell lymphoma after prior systemic therapy. J Clin Oncol. 2012;30(6):631-636. 40. Iyer SP, Foss FF. Romidepsin for the treatment of peripheral Tcell lymphoma. Oncologist. 2015;20(9):1084-1091. 41. ISTODAX® (romidepsin) [prescribing information]. ISTODAX

(romidepsin) Label. Accessdata.fda.gov Accessed 16 February 2023. 42. Bachy E, Camus V, Thieblemont C, et al. Romidepsin plus CHOP versus CHOP in patients with previously untreated peripheral Tcell lymphoma: results of the Ro-CHOP phase III study (Conducted by LYSA). J Clin Oncol. 2022;40(3):242-251.

Haematologica | 109 January 2024

219

ARTICLE - Plasma Cell Disorders

Daratumumab in first-line treatment of patients with light chain amyloidosis and Mayo stage IIIb improves treatment response and overall survival Sara Oubari,1,2 Ute Hegenbart,3 Renate Schoder,4 Maximilian Steinhardt,5 Maria Papathanasiou,2,6 Tienush Rassaf,6 Andreas Thimm,2,7 Tim Hagenacker,2,7 Eyad Naser,8 Ulrich Dührsen,1 Hans C. Reinhardt,1,9 Martin Kortüm,5 Hermine Agis,4 Stefan Schönland3 and Alexander Carpinteiro1,2,8 Department of Hematology and Stem Cell Transplantation, West German Cancer Center, University Hospital Essen, University Duisburg-Essen, Essen, Germany; 2Interdisciplinary Amyloidosis Network, University Hospital Essen, University Duisburg-Essen, Essen, Germany; 3Department of Internal Medicine V, Amyloidosis Center Heidelberg, University Hospital Heidelberg, Heidelberg, Germany; 4Department of Internal Medicine I, Division of Hematology and Hemostaseology, Medical University Vienna, Vienna, Austria; 5Department of Hematology, University Hospital Würzburg, Würzburg, Germany; 6Department of Cardiology and Vascular Medicine, West German Heart and Vascular Center, University Hospital Essen, University Duisburg-Essen, Essen, Germany; 7Department of Neurology and Center for Translational Neuro- and Behavioral Science, University Hospital Essen, Essen, Germany; 8Institute of Molecular Biology, University of Duisburg-Essen, Essen, Germany and 9 German Cancer Consortium (DKTK), Partner Site University Hospital Essen, Essen, Germany 1

Correspondence: A. Carpinteiro alexander.carpinteiro@uk-essen.de Received: Accepted: Early view:

April 10, 2023. July 3, 2023 July 13, 2023.

Abstract Treatment of patients with Mayo stage IIIb light chain (AL) amyloidosis is still challenging, and the prognosis remains very poor. Mayo stage IIIb patients were excluded from the pivotal trial leading to the approval of daratumumab in combination with bortezomib-cyclophosphamide-dexamethasone. This retrospective, multicenter study evaluates the addition of daratumumab to first-line therapy in patients with newly diagnosed stage IIIb AL amyloidosis. In total, data from 119 consecutive patients were analyzed, 27 patients received an upfront treatment including daratumumab, 63 a bortezomibbased regimen without daratumumab, eight received therapies other than daratumumab or bortezomib and 21 pretreated patients or deceased prior to treatment were excluded. In the daratumumab group, median overall survival was not reached after a median follow-up time of 14.5 months, while it was significantly worse in the bortezomib- and the otherwise treated group (6.6 and 2.2 months, respectively) (P=0.002). Overall hematologic response rate at 2 and 6 months was better in the daratumumab group compared to the bortezomib group (59% vs. 37%, P=0.12, 67% vs. 41%, P=0.04, respectively). Landmark survival analyses revealed a significantly improved overall survival in patients with partial hematologic response or better, compared to non-responders. Cardiac response at 6 months was 46%, 21%, 0% in the daratumumab-, bortezomib- and otherwise treated groups, respectively (P=0.04). A landmark survival analysis revealed markedly improved overall survival in patients with cardiac very good partial response vs. cardiac non-responders (P=0.002). This study demonstrates for the first time the superiority of an upfront treatment with daratumumab over standard-of-care in stage IIIb AL amyloidosis.

Introduction Systemic light chain (AL) amyloidosis is a rare acquired protein misfolding disorder with an incidence of about 5-13 cases per million person-years.1 It is characterized by extracellular deposition of misfolded amyloidogenic immunoglobulin light chain fibrils secreted most often from clonal plasma cells in the bone marrow.2 With the exception of the central nervous system, all organs can be affected, most

commonly the heart and kidneys.2,3 The outcome of patients with AL amyloidosis is strongly associated with the severity of organ involvement, especially the heart. Cardiac Mayo classifications from 2004 and 2012 were established to assess the severity of the disease.4-6 In particular, patients with Mayo stage IIIb, who are often excluded from clinical trials, have the worst prognosis and thus represent the group of patients for whom new, efficient treatment strategies need to be developed.6-8

Haematologica | 109 January 2024

220

ARTICLE - First-line daratumumab in AL with Mayo stage IIIb Current treatment protocols for patients with Mayo stage IIIb amyloidosis are based on anti-plasma cell drugs.9,10 Bortezomib, a proteasome inhibitor, has been the most frequently administered drug in first-line treatment since 2010.8 In the case of existing contraindications, other treatments, such as melphalan and dexamethasone, immunomodulatory drugs or bendamustine, are used.8,11 In 2021, daratumumab, an IgGκ monoclonal anti-CD38antibody, was approved for the treatment of patients with AL amyloidosis due to the results from a randomized phase III clinical trial (ANDROMEDA).12 However, in this trial, patients with Mayo stage IIIb were excluded. In contrast to other established therapies, the mechanism of action of daratumumab is rather specific, as the antibody is directed against CD38, which is not present on cardiomyocytes. It is, therefore, unlikely to cause additive cardiotoxicity, which may be a factor favoring the use of daratumumab-containing therapies in advanced cardiac disease. In fact, a post-hoc analysis from the ANDROMEDA study showed that patients with cardiac involvement (up to Mayo stage IIIa) were the ones who benefited most from the addition of daratumumab to standard chemotherapy.12,13 In another retrospective study,14 which included patients with relapsed AL amyloidosis, the median overall survival (OS) in patients with N-terminal prohormone of brain natriuretic peptide (NTproBNP) levels above 8,500 pg/mL was 6 months when treated with daratumumab and dexamethasone (34 patients) and 11 months when treated with daratumumab, bortezomib and dexamethasone (25 patients).14 NTproBNP levels above 8,500 pg/mL are characteristic of Mayo stage IIIb. Similar data were recently published in a report on the use of daratumumab in 19 Mayo stage IIIb patients, which showed a promising efficacy and safety profile with a 12-month OS of 67.5%.15 In another retrospective study including 22 patients treated with daratumumabbased regimens, promising hematologic response rates (64% very good partial response [VGPR] or better at 3 months) but no survival data were reported.16 Overall, these findings suggest that daratumumab-containing regimens may also be an effective and tolerable treatment option for patients with Mayo stage IIIb in the first-line setting, potentially representing a step forward in the treatment of this hitherto underserved group of patients. The present multicenter, retrospective study is the first to investigate the efficacy of the addition of daratumumab to first-line therapy in patients with AL amyloidosis and Mayo stage IIIb compared to bortezomib-containing regimens or other less commonly used therapies.

Methods This retrospective study enrolled 119 consecutive patients with histologically diagnosed AL amyloidosis and Mayo

S. Oubari et al.

stage IIIb, who presented between 2016 and 2022 at four different sites in Germany and Austria (Amyloidosis Center at the University Hospital of Heidelberg, University Hospital of Essen, University Hospital of Vienna, and University Hospital of Würzburg). The study was approved by the institutional ethics committees at each site (Heidelberg S-123/2006, Essen 21-10467-BO, Vienna 1184/2022, Würzburg 20220913 01) and performed according to the Helsinki Declaration of 1975, as revised in 2013. To investigate the role of upfront daratumumab treatment of patients with Mayo stage IIIb, the following treatment groups were compared: (i) patients who received daratumumab with or without bortezomib from the beginning as part of the induction therapy (n=27). Depending on therapy response and toxicity, additional therapy components could be added sequentially during the course. This group was referred to as the daratumumab group; (ii) patients who received bortezomib without daratumumab from the beginning (n=63). Again, in the case of lack of response to therapy, daratumumab could be added during the course. This group was referred to as the bortezomib group; and (iii) patients for whom therapies other than bortezomib or daratumumab were selected (n=8) (Figure 1). Treatment decisions in this retrospective study were left to local physicians' judgement, without any cross-center coordination. Organ involvement was defined according to the consensus criteria of the International Society of Amyloidosis.17 Mayo stage IIIb was defined as NTproBNP level >8,500 pg/mL and cardiac troponin T ≥0.035 ng/mL or high-sensitivity troponin T ≥50 ng/L.4,18 Renal stages were determined according to the 2014 renal classification.19 OS was calculated from the date of treatment initiation to death or loss of follow-up or to the date of heart transplantation. Hematologic responses at 2 and 6 months were assessed as recommended by the International Society of Amyloidosis.20,21 Patients with a small difference between involved and non-involved light chains (dFLC), i.e., <50 mg/L, were excluded. Cardiac response at 6 months was assessed as proposed by Palladini et al. and graded by Muchtar et al.20,22,23 Kaplan-Meier landmark survival analyses of hematologic and cardiac responses were performed. In the case of absence of data, e.g., treatment or organ responses at certain time points, these data were not assessed in the statistical analyses. All treated patients, including those who died, were included in statistical analyses, as indicated. Descriptive statistics were used to summarize the patients’ characteristics. Surrogate continuous variables in the three treatment groups were compared using the Kruskal-Wallis test. Hematologic and cardiac response rates in both the daratumumab group and the bortezomib group were compared using the Fisher exact test. Corresponding odds ratios (OR), 95% confidence intervals (95% CI), and two-sided P values were reported. Time-to-event data variables and the

Haematologica | 109 January 2024

221

ARTICLE - First-line daratumumab in AL with Mayo stage IIIb

S. Oubari et al.

Figure 1. Schematic representation of all treatment groups and regimens. Bor: bortezomib; Cy: cyclophosphamide; Dara: daratumumab; D: dexamethasone; M: melphalan; VMP: bortezomib, melphalan, prednisone; PAD: bortezomib, doxorubicin, dexamethasone; V: bortezomib.

median OS were assessed using the Kaplan-Meier method. The median follow-up time was calculated based on the median time to censoring (reversed Kaplan-Meier). Univariate Cox regression analysis predicting possible prognostic factors in the whole Mayo stage IIIb cohort was performed. A forest plot was constructed and the hazard ratios of various variables with their 95% CI were calculated to determine differences between the daratumumab and the bortezomib treatment groups. The statistical software applied to calculate the statistical significance was GraphPad PRISM.

Results Table 1 shows the baseline characteristics and treatment groups of the included patients broken down by treatment center. While the majority of patients in Heidelberg and Würzburg were treated in the bortezomib group, the larger part of patients in Essen and Vienna were treated in the daratumumab group (Table 1). Table 2 shows the baseline characteristics of treated patients broken down by treatment group. There were no significant differences in baseline characteristics between the treatment groups except for the age of patients treated with other therapies; these patients were significantly older than the ones treated in the bortezomib group (75.7 vs. 68 years; P=0.02). Seventeen patients who did not receive therapy because they had previously died or refused therapy are not included in this table, in which therapies are compared. The

median NTproBNP in this group was 28,912 pg/mL and significantly (P=0.01) increased compared to the patients who received therapy. In the group of patients treated from the beginning with daratumumab and dexamethasone (n=16, 59.2% of the whole daratumumab group), five patients (31.25%) received bortezomib additionally later during the course of treatment, while in the group of patients who were treated with bortezomib and dexamethasone from the beginning (n=55, 87.3% of the whole bortezomib group), seven patients (12.7%) later received additional daratumumab (Figure 1). The response rates to first-line regimens and the secondline therapies are listed in Online Supplementary Table S1. Comparison of OS rates between different treatment groups after a median follow-up time of 31.3 months revealed significant differences favoring daratumumab as first-line treatment in Mayo stage IIIb patients (Figure 2). The 6month OS in the daratumumab group, bortezomib group, and the group treated with other regimens was 70%, 51% and 25%, respectively (log-rank test, P=0.002). The median OS was 6.6 months in the bortezomib group, 2.2 months in the group treated with other regimens, but was not reached in the daratumumab group. Analyzing the OS between single groups also showed significant differences (Figure 2). We also compared hematologic treatment responses at defined timepoints in all evaluable patients. For this purpose, we selected an early timepoint after 2 months and a later one after 6 months. In the daratumumab group, at 2 months the overall response rate (ORR) was 59%, the com-

Haematologica | 109 January 2024

222

ARTICLE - First-line daratumumab in AL with Mayo stage IIIb plete hematologic response (CR) rate was 36.3%, the VGPR rate was 4.5%, and 18.1% of the patients had achieved a partial response (PR). In contrast to that, the ORR was reduced to 36.7% in the bortezomib group. In this group, 4%

S. Oubari et al.

achieved a CR, 18.3% a VGPR and 14.2% a PR. In the group of patients treated with other regimens, the ORR was 28.5% with none having achieved a CR, 14.2% a VGPR and 14.2% a PR (Figure 3A). Comparing ORR at 2 months revealed non-

Table 1. Basic clinical characteristics of the patients in each cohort at the four sites. Heidelberg

Essen

Vienna

Würzburg

N of patients

Age in years median range

70.5 42-83

67 42-84

73 49-87

74 60-83

Time of inclusion

2018-2021

2016-2022

2016-2021

Gender, N (%) male female

33 (63.5) 19 (36.5)

22 (53.7) 19 (46.3)

7 (46.7) 8 (53.3)

6 (54.5) 5 (45.5)

Light chain isotype, N (%) lambda kappa

45 (86.5) 7 (13.5)

32 (78.1) 9 (21.9)

11 (73.3) 4 (26.7)

9 (81.8) 2 (18.2)

PCD, N (%) MG Smoldering MM MM NA

19 (36.5) 24 (46.1) 6 (11.5) 3 (5.8)

15 (36.6) 15 (36.6) 11 (26.8) 0

5 (33.3) 7 (46.7) 2 (13.3) 1 (6.7)

5 (45.4) 3 (27.3) 2 (18.2) 1 (9.1)

dFLC, mg/L median range

287.5 3-16,120

328.7 3.4-2,694.5

233.8 47-9,565.47

406.8 4-11,266

16,960 9,324-132,484

18,156 8,698-187,267

19,529* 8,500-35,000*

15,255* 10,386-34,761*

0.124 0.046-0.72

hsTnT, ng/L median range

132.0 61-790

157.5 54- 408

220.4 57-566

GFR, mL/min/1.73 m2 median range

47 17-93

43 6-106

32 9-64

43 23-78

Proteinuria, g/24 h median range

2.4 0.5-10.5

3.6 0.2-13.4

1.5 0.1-17.5

7.2 3.6-11.3

Renal stages, N (%) stage I stage II stage III dialysis

7 (13.5) 12 (23.1) 6 (12.5) 2 (3.8)

7 (17.1) 12 (29.3) 7 (17.1) 6 (15.6)

1 (6.7) 6 (40.0) 5 (33.3) 1 (6.7)

0 (0) 2 (18.2) 4 (36.3) 0 (0)

Treatment group, N daratumumab group bortezomib group others excluded

4 39 5 4

15 15 3 8

7 2 0 6

1 7 0 3

PCBM infiltration, % median

NTproBNP, pg/mL* median range cTnT, ng/mL median range

The maximal measurable NTproBNP level at the sites in Vienna and Würzburg was 35,000 pg/mL. PCBM: plasma cell bone marrow; PCD: plasma cell disease; MG: monoclonal gammopathy; MM: multiple myeloma; NA: not available data; dFLC: difference between involved and non-involved light chains; NTproBNP: N-terminal prohormone of brain natriuretic peptide; cTnT: cardiac troponin T; hsTnT: high-sensitivity troponin T; GFR: glomerular filtration rate. *

Haematologica | 109 January 2024

223

ARTICLE - First-line daratumumab in AL with Mayo stage IIIb significant differences between the daratumumab- and bortezomib-treated groups (OR=2.48; 95% CI: 0.848-7.036; P=0.12), whereas comparing both groups with respect to CR did show a statistically significant difference (OR=13.43; 95% CI: 2.553-65.49; P=0.0009). The proportion of patients who died within the first 2 months after initiation of therapy was 22% in the daratumumab group, 27% in the bortezomib

S. Oubari et al.

group and 25% in the group treated with other regimens. At 6 months after initiating treatment, in the daratumumab group, the ORR was 66.7%; 41.7% had achieved a CR, 16.7% a VGPR and 8.3% a PR. In contrast to that, the ORR was only 40.7% in the bortezomib group. In this group, 14.8% had achieved a CR, 16.7% a VGPR and 9.3% a PR. In the group of patients treated with other regimens, the ORR was 25%

Table 2. Basic clinical characteristics in each treatment group. Daratumumab group

Bortezomib group

Other regimens

All

N of patients

Age in years median range

67 42-84

68 42-83

75.5 70-83

68.5

0.028

Gender, N (%) male female

15 (55.6) 12 (44.4)

36 (57.1) 27 (42.9)

5 (62.5) 3 (37.5)

56 (57.1) 42 (42.9)

Light chain isotype, N (%) lambda kappa

21 (77.8) 6 (22.2)

51 (80.8) 12 (19.2)

8 (100) 0

80 (81.6) 18 (18.4)

PCD, N (%) MG Smoldering MM MM NA

8 (29.6) 11 (40.7) 8 (29.6) 0

29 (46) 24 (38.1) 9 (14.3) 1 (1.6)

3 (37.5) 4 (50) 0 1 (12.5)

40 (40.8) 39 (40.8) 17 (17.4) 2 (2)

PCBM infiltration, % median

dFLC, mg/L median range

397.5 9.04-2,694.5

303 3.4-16,120

271 52-581

308

0.58

NTproBNP, pg/mL* median rangea

13,668 8,930-87,243

15,047 8,698-132,484

20,659 11,092-40,685

15,943

0.54

GFR, mL/min/1.73 m2 median range

42.5 6.2-92.9

46.5 9-106

28.7 15.3-59.8

44.9

0.24

Proteinuria, g/24 h median range

1.5 0.23-17.5

4.2 0.29-12.1

2.38 1.16-13.4

2.5

0.43

Renal stages stage I stage II stage III dialysis

7 (25.9) 6 (30) 5 (25) 2 (10)

8 (21) 15 (39.5) 12 (31.6) 3 (7.9)

0 (0) 3 (50) 2 (33.3) 1 (16.7)

15 (24.4) 24 (37.5) 19 (29.7) 6 (9.4)

Number of involved organs 1 2 3 4 5 >5

3 (11.1) 8 (22.2) 10 (37.0) 3 (11.1) 1 (3.7) 2 (7.4)

10 (15.9) 19 (30.2) 15 (23.8) 12 (19.0) 7 (11.1) 0

1 (12.5) 1 (12.5) 0 (0) 4 (50) 2 (25) 0

14 (14.3) 28 (28.6) 25 (25.5) 19 (19.4) 10 (10.2) 2 (2)

Follow-up time, months median range

14.5 2.4-39.3

33.8 1.63-44

Undefined

31.3 1.63-44

The maximal measurable NTproBNP level at the sites in Vienna and Würzburg was 35,000 pg/mL. PCD: plasma cell disease; MG: monoclonal gammopathy; MM: multiple myeloma; NA: not available data; PCBM: plasma cell bone marrow; dFLC: difference between involved and noninvolved light chains; NTproBNP: N-terminal prohormone of brain natriuretic peptide; GFR: glomerular filtration rate.

Haematologica | 109 January 2024

224

ARTICLE - First-line daratumumab in AL with Mayo stage IIIb with 12.5% achieving a CR, 12.5% a VGPR and none a PR (Figure 3B). Comparing the ORR at 6 months revealed a significant difference between the daratumumab- and bortezomib-treated groups (OR=2.9; 95% CI: 1.076-8.322; P=0.04). Likewise, the difference in CR rate between the two groups was also statistically significant (OR=4.107; 95% CI: 1.36-11.17; P=0.01). The proportion of patients who died within the first 6 months after initiation of therapy was 29.6% in the daratumumab group, 47.6% in the bortezomib group and 75% in the group treated with other regimens. To evaluate the effect of early hematologic responses on OS we performed landmark survival analyses for survivors starting from 2 and 6 months after treatment initiation (Figure 3C, D). The 12-month OS rates in patients who achieved a CR, VGPR or PR and survived 2 months after treatment initiation were 90%, 82% and 80%, respectively, while it was reduced to 17.4% in hematologic non-responders. Patients who achieved a CR, VGPR or PR and survived 6 months after treatment initiation had a 12-month OS rate of 94%, 79% and 86%, respectively, while all the non-responders who were alive at 6 months after treatment initiation died before

S. Oubari et al.

completing the first year of treatment. Log-rank tests between the groups with a CR, VGPR or PR did not show statistically significant survival differences, whereas nonresponders had a significantly worse survival rate compared to patients with CR, VGPR or PR in both landmark analyses (Figure 3C, 3D). Cardiac response at 6 months after treatment initiation was

Figure 2. Kaplan-Meier survival analysis showing survival rates according to first-line treatment. OS: overall survival.

Figure 3. Hematologic response rates at 2 and 6 months and Kaplan-Meier landmark analyses of overall survival regarding hematologic response at 2 and 6 months. (A) Hematologic response rates at 2 months. In 20 patients, response data were not available or dFLC was lower than 50 mg/L (5 in the daratumumab group, 14 in the bortezomib group and 1 in the group treated with other regimens). (B) Hematologic response rates at 6 months. In 12 patients, response data were not available or dFLC was lower than 50 mg/L (3 in the daratumumab group and 9 in the bortezomib group). (C) Overall survival according to hematologic response at 2 months. (D) Overall survival according to hematologic response at 6 months. Dara: group treated with daratumumab; Bor: group treated with bortezomib; Other: group treated with other regimens; NR: no response; PR: partial response; VGPR: very good partial response; CR: complete response; OS: overall survival. Haematologica | 109 January 2024

225

ARTICLE - First-line daratumumab in AL with Mayo stage IIIb assessed according to treatment group in all patients with evaluable data. None of the patients achieved a cardiac CR. The cardiac ORR was 45.8% in the daratumumab group, compared to 21.4% in the bortezomib group (OR=3.103; 95% CI: 1.171-8.299; P=0.03) (Figure 4A). Patients in the daratumumab group were more likely to have a cardiac VGPR than those in the bortezomib group (25% vs. 14.3%), respectively (OR=2.0; 95% CI: 0.558-5.984; P=0.3). Patients treated with other regimens did not reach a cardiac response after 6 months (Figure 4A). The 12-month OS was 100% in patients who achieved a cardiac VPGR or cardiac PR, compared to 73% in cardiac non-responders (Figure 4B). The log-rank test between patients with a cardiac VGPR and cardiac PR did not show any significant survival differences, whereas patients with who did not have a cardiac response had a

S. Oubari et al.

significantly worse survival compared to patients with a cardiac VGPR (P=0.002) (Figure 4B). In an univariate Cox regression analysis, no risk factors that could predict OS in the whole Mayo IIIb cohort were identified (Table 3). Of note, surrogate parameters such as dFLC ≥500 mg/L or NTproBNP ≥16,000 pg/mL (median 15,944 pg/mL) were not associated with an increased risk of death. Other variables with their hazard ratios are shown in Table 3. Notably, higher hazard ratios were demonstrated in the bortezomib-treated group than in the daratumumab-treated group. To investigate whether any baseline, clonal or organ variables might have affected OS in a given treatment group, we performed another Cox regression analysis (Figure 5). In particular, patients with elevated clonal or organ par-

Figure 4. Cardiac response at 6 months with landmark survival analysis. (A) Cardiac response in patients with evaluable response data. In 11 patients, response data were not available (3 in the daratumumab group, 7 in the bortezomib group and 1 in the group treated with other regimens). (B) Kaplan-Meier landmark analysis of overall survival regarding cardiac response at 6 months. Dara: group treated with daratumumab; Bor: group treated with bortezomib; Other: group treated with other regimens; carNR: no cardiac response; carPR: partial cardiac response; carVGPR: very good partial cardiac response; OS: overall survival.

Table 3. Univariate Cox-regression analysis predicting prognostic factors in the whole cohort. Variables

95% CI of HR

Age ≥65 years (reference age <65 years)

1.225

0.709-2.230

0.484

Male (reference female)

1.016

0.617-1.697

0.950

Light-chain isotype kappa (reference lambda)

1.312

0.679-2.351

0.387

dFLC ≥180 mg/L

0.925

0.552-1.605

0.774

dFLC ≥500 mg/L

0.812

0.459-1.378

0.457

dFLC ≥800 mg/L

0.908

0.461-1.645

0.763

NTproBNP ≥16,000 pg/mL (median 15,944 pg/mL)

1.313

0.801-2.162

0.279

GFR ≥50 mL/min/1.73 m

0.765

0.447-1.276

0.314

GFR ≥30 mL/min/1.73 m2

0.784

0.468-1.355

0.368

Proteinuria ≥5 g/24 h

1.156

0.598-2.118

0.650

Renal stage II (reference stage I)

0.581

0.200-1.792

0.322

Renal stage III (reference stage I)

0.275

0.06-1.027

0.059

Dialysis (reference stage I)

3.389

0.159-30.07

0.311

Treatment group (reference: daratumumab group) bortezomib group other regimens

2.018 4.632

1.061-4.246 1.739-11.96

0.044 0.001

HR: hazard ratio; 95% CI: 95% confidence interval; dFLC: difference between involved and non-involved light chains; NTproBNP: N-terminal prohormone of brain natriuretic peptide; GFR: glomerular filtration rate. Haematologica | 109 January 2024

226

ARTICLE - First-line daratumumab in AL with Mayo stage IIIb ameters (dFLC ≥180 mg/L, NTproBNP ≥16,000 pg/mL and glomerular filtration rate ≥30 mL/min/1.73 m2) benefited markedly from daratumumab in the first-line treatment. The benefit of adding daratumumab to first-line therapy appeared to increase relatively with increasing NTproBNP up to 25,000 pg/mL. However, this advantage appeared to decrease with further increases in NTproBNP (Figure 5).

Discussion To our knowledge, this is the first comparative, multicenter, retrospective study analyzing treatment response and OS in Mayo stage IIIb patients after first-line treatment including daratumumab, compared to bortezomib or other regimens. Patients in the daratumumab group had a significantly

S. Oubari et al.

better outcome (median OS was not reached after the 14.5 months median follow-up time in the daratumumab group) than patients treated in the bortezomib group (median OS was 6.6 months) or with other regimens (median OS was 2.2 months) (Figure 2). While the 6- and 12-month OS rates remained the same at 70% in the daratumumab group, OS decreased from 51% at 6 months to 41% at 12 months in the bortezomib group (Figure 2). None of the patients who were treated with other regimens survived 1 year after treatment initiation. In comparison to previous studies, the median OS rate for Mayo stage IIIb patients enrolled in the EMN23 study between 2004 and 2010 was 5 months.8 In patients enrolled between 2011 and 2018, when bortezomib-containing regimens were being increasingly used, the median OS was 4.5 months and did not improve in contrast to that of patients with Mayo stages II and IIIa.8 In the ALchemy

Figure 5. Forest plot for subgroup analysis of overall survival in patients with AL amyloidosis and Mayo stage IIIb. Data are derived from Cox regression analysis without covariates. HR: hazard ratio; CI: confidence interval; dFLC: difference between involved and non-involved light chains; NTproBNP: N-terminal prohormone of brain natriuretic peptide; GFR: glomerular filtration rate; PU: proteinuria.

Haematologica | 109 January 2024

227

ARTICLE - First-line daratumumab in AL with Mayo stage IIIb study, a single-center, prospective, observational study, the median OS for 179 Mayo IIIb patients enrolled between 2009 and 2015 was 6 months.24 Of note, in these studies daratumumab was not included in first-line treatments. In the randomized ANDROMEDA trial, which included patients with Mayo stages I, II and IIIa, no significant survival differences were found between patients treated with CyBorD (cyclophosphamide, bortezomib and dexamethasone) and CyBorD plus daratumumab after a median follow-up time of 11.4 months (see supplementary data of the ANDROMEDA clinical trial).12 To the best of our knowledge, our comparative study shows for the first time an OS benefit from the use of daratumumab as an upfront treatment over bortezomib in the therapy of AL amyloidosis patients with Mayo stage IIIb. In our patients treated in the daratumumab group, bortezomib was added from the beginning, later in the course, or not at all (Figure 1). The addition of bortezomib did not follow a fixed protocol but was based on the assessment/decision of the treating physician. The experience of the physician, the clinical condition of the patient and the therapy response achieved each played a role. Likewise, the dosage of bortezomib, both at the time of addition or as part of the primary therapy, was not fixed. Generally, the initial dosage of bortezomib was 0.7 mg/m2 or 1 mg/m2 and was eventually increased during the course of therapy, depending on the response to therapy and tolerability. Preliminary evaluation of the ongoing EMN22-trial (daratumumab monotherapy, addition of dexamethasone and bortezomib from the fourth cycle in the case of insufficient response) revealed a median OS of 9.4 months.25 There is a recent short report of a median OS of more than 12 months in Mayo IIIb patients who were treated with upfront Dara-VCD (daratumumab, bortezomib, cyclophosphamide and dexamethasone).15 Preliminary hematologic response data from the ongoing EMN22 trial documented a 6-month ORR of 72%.25 Our patients in the daratumumab group showed comparable results with a 6-month ORR of 67%. Upfront treatment consisting of daratumumab in combination with bortezomib, cyclophosphamide and dexamethasone in 19 patients with Mayo stage IIIb showed good efficacy (VGPR or better, 86.7%) and safety, which is in line with the results of our study.15 In the present study, the 6-month cardiac ORR in 24 evaluable (including 8 deceased) patients in the daratumumab group was 46% Preliminary evaluation of cardiac response at 6 months in the ongoing EMN22 trial revealed a cardiac response rate of 27.6%.25 In a recently published study, in which Mayo IIIb patients were treated with upfront Dara-VCD, a cardiac response was found in 80% of ten evaluable patients; however, deceased patients and patients without evaluable data were excluded from this analysis. Furthermore, differences in the baseline char-

S. Oubari et al.

acteristics may have led to a relevant bias. Remarkably, median NTproBNP in the upfront Dara-VCD study was 10,702 pg/mL, compared to 15,468 pg/mL in the EMN22 trial and 13,668 pg/mL in our study. Therefore, also due to the small number of cases in any of the studies, a clear statement on the efficacy of different treatment-schemes (upfront Dara-VCD vs. upfront daratumumab [+dexamethasone] and sequential addition of bortezomib) cannot be made at the moment. We also determined the hematologic response after 2 and 6 months of therapy and performed landmark survival analyses. We found a significant OS advantage for patients who survived 2 and 6 months and achieved at least a PR compared to patients who had no response (Figure 3C, D). Surprisingly, in these landmark analyses, no significant differences in survival proportion rates were found between patients who achieved CR, VGPR or PR at 2 or 6 months. Our 2-month hematologic response data are in line with the analyses performed in the ALchemy study.24 Here, early hematologic response after 30 days of therapy was shown to predict 12-month OS in patients who were still alive after 6 months. In concordance with our data, no significant differences in the depth of hematologic response (CR, VGPR or PR) were found.24 The impact of early treatment response at 1 and 3 months on survival was investigated in 249 consecutive patients newly diagnosed between 2004 and 2018 at the Amyloidosis Research and Treatment Center in Pavia.26 In this series, patients with a CR or VGPR at 1 month and 3 months showed a significant survival advantage over patients with a PR or no response.26 Furthermore, a landmark analysis of 663 patients with Mayo stage IIIb who survived the first 3 months (60%) showed that the median OS clearly depended on the depth of the response achieved; the median OS in patients with a CR was 47 months, whereas it was 36 months in those who achieved a VGPR, 19 months in those who had a PR and 3 months in non-responders.8 Both, the EMN23 study8 and the study at Pavia26 are from the pre-daratumumab era and early mortality was much higher than in the present study. An explanation of why the difference in depth of hematologic response (CR/VGPR/PR) in our dataset did not allow such a clear prediction of survival may be the sequential strategy in our therapeutic approach, in many cases consisting only of daratumumab and dexamethasone at the beginning with a gradual increase in intensity by adding bortezomib, in order to lessen toxicity, so that the 2-month response may not necessarily have served as a biological hematologic response parameter for survival. To shed some light, in a landmark subanalysis of survival dependent on hematologic response at 2 months, we examined only those patients who were not treated with daratumumab at baseline or later (Online Supplementary Figure S1). Al-

Haematologica | 109 January 2024

228

ARTICLE - First-line daratumumab in AL with Mayo stage IIIb though a clear conclusion cannot be drawn due to the small number of cases, when excluding patients treated with daratumumab, the data look similar to those previously published from the pre-daratumumab era. In our view, further studies should be performed to shed more light on this interesting point. Our data lead us to suggest that, with the inclusion of daratumumab as a firstline treatment and sequential addition of further components such as bortezomib, it may become necessary to define modified response criteria, possibly combining hematologic and organ response parameters.26 We found no relevant clonal or organ predictive variables that could influence OS. Only treatment with daratumumab had a significantly positive impact on OS compared to bortezomib and other regimens (Table 3). To illustrate differential influences of variables in different treatment groups we constructed a forest plot (Figure 5). Adjusted cutoff values of NTproBNP revealed an increased advantage up to about 25,000 pg/mL. At even higher NTproBNP levels up to 35,000 pg/mL, this advantage seemed to diminish slightly. These findings indicate that there may be a level at the top at which even the addition of daratumumab does not give advantages anymore, because in these patients the heart disease may already be so severe that even a hematologic response cannot turn the tide. Advanced kidney disease, i.e., renal stages II and III and especially proteinuria >5 g/24 h, reduced the benefit of adding daratumumab, confirming data from Kimmich et al. who showed that the antibody is excreted in the urine in the context of proteinuria and thus loses efficacy.14 Patients with a high dFLC (≥180 mg/L) benefit particularly from the addition of daratumumab. This is probably the reason that in our cohort even a dFLC ≥500 mg/L no longer predicted a poor outcome in univariate Cox regression analysis, as described by Basset et al.26 While there are clear therapy recommendations for patients with Mayo stages I-IIIa, supported by the findings of the ANDROMEDA study12 and the approval of daratumumab in combination with VCD, the situation is less clear for Mayo IIIb patients. Among the patients in our series with multiple myeloma, a relatively higher proportion received treatment with daratumumab (8 out of 17, 47%) due to its approval in combination with bortezomib and dexamethasone for myeloma. Conversely, in patients with monoclonal gammopathy, the proportion in the daratumumab group was lower (8 out of 40, 20%) (Table 2). Consequently, the daratumumab group had a higher representation of patients with advanced plasma cell disease. Despite this bias that may have put the daratumumab group at a disadvantage, patients treated with daratumumab demonstrated better response rates and survival, emphasizing the significance of the data. Our study has several limitations, most notably the retro-

S. Oubari et al.

spective design and the unavailability of a few clinical parameters or response data in some patients. A further selection bias may limit the information value of our study since not all Mayo IIIb patients may be referred to our treatment centers. However, the baseline characteristics in our study (Table 1) mostly correspond to the baseline characteristics in previous studies including Mayo IIIb patients.8,15,24,26 Another weakness of this study is that therapy-associated side effects were not systematically recorded. However, it can be said that in no case did therapy with daratumumab have to be reduced or discontinued due to side effects and no infection-related deaths occurred. In conclusion, our study demonstrates for the first time the superiority of an upfront treatment with daratumumab over the current standard-of-care with bortezomib in Mayo stage IIIb patients. The better survival rates in the group treated with daratumumab suggests its efficacy and safety profile in this vulnerable patient population. Further randomized trials should be conducted to confirm our data and to define the best sequence of daratumumab combination therapies. Disclosures HCR has received consulting and lecture fees from Abbvie, Roche, Novartis, Bristol Myers Squibb, AstraZeneca, Vertex, and Merck, has received research funding from AstraZeneca and Gilead Pharmaceuticals, and is a co-founder of CDL Therapeutics GmbH. TR reports speaker fees from AstraZeneca, Daiichy, Bayer, Novartis, and Abiomed outside the submitted work. AC has received consulting and/or lecture fees from Amgen, Janssen, Pfizer and Takeda and travel and congress participation grants from Janssen. SO has received travel and congress participation grants from Janssen. UH has received honoraria for talks from Janssen, Pfizer, Alnylam, and Akcea, has received financial support for congress participation from Janssen, Prothena, Pfizer, has sat on advisory boards for Pfizer, Prothena, and Janssen, and has received financial sponsorship of the Amyloidosis Registry from Prothena and Janssen. SS has received research support from Janssen, Prothena and Sanofi, has participated in advisory boards for Janssen, Telix, Sobo, and Prothena, has received honoraria from Janssen, Takeda, Pfizer and Prothena, and has received travel and congress participation grants from Janssen, Prothena, Celgene, Binding Site, and Jazz. HA has received honoraria from Amgen, BMS (Celgene), GSK, Janssen, Novartis, Sanofi, and Takeda. The remaining authors declare no competing financial interests. Contributions AC and SO conceived and designed the study. AC, SO, UH, SS, HA, RS, and MS provided study materials or patients and collected and assembled data. AC, SO, and SS analyzed and

Haematologica | 109 January 2024

229

ARTICLE - First-line daratumumab in AL with Mayo stage IIIb

S. Oubari et al.

interpreted data. All authors wrote the manuscript or revised it critically for important intellectual content.

acknowledge support from the Open Access Publication Fund of the University of Duisburg-Essen.

Funding This study was supported by a grant from the Deutsche Forschungsgemeinschaft to AC (grant number CA2420/2-1). We

Data-sharing statement For original data, please contact alexander.carpinteiro@ukessen.de.

References 1. Kyle RA, Linos A, Beard CM, et al. Incidence and natural history of primary systemic amyloidosis in Olmsted County, Minnesota, 1950 through 1989. Blood. 1992;79(7):1817-1822. 2. Falk RH, Comenzo RL, Skinner M. The systemic amyloidoses. N Engl J Med. 1997;337(13):898-909. 3. Merlini G, Dispenzieri A, Sanchorawala V, et al. Systemic immunoglobulin light chain amyloidosis. Nat Rev Dis Primers. 2018;4(1):38. 4. Dispenzieri A, Gertz MA, Kyle RA, et al. Serum cardiac troponins and N-terminal pro-brain natriuretic peptide: a staging system for primary systemic amyloidosis. J Clin Oncol. 2004;22(18):3751-3757. 5. Kumar S, Dispenzieri A, Lacy MQ, et al. Revised prognostic staging system for light chain amyloidosis incorporating cardiac biomarkers and serum free light chain measurements. J Clin Oncol. 2012;30(9):989-995. 6. Wechalekar AD, Schonland SO, Kastritis E, et al. A European collaborative study of treatment outcomes in 346 patients with cardiac stage III AL amyloidosis. Blood. 2013;121(17):3420-3427. 7. Oubari S, Naser E, Papathanasiou M, et al. Impact of time to diagnosis on Mayo stages, treatment outcome, and survival in patients with AL amyloidosis and cardiac involvement. Eur J Haematol. 2021;107(4):449-457. 8. Palladini G, Schonland S, Merlini G, et al. The management of light chain (AL) amyloidosis in Europe: clinical characteristics, treatment patterns, and efficacy outcomes between 2004 and 2018. Blood Cancer J. 2023;13(1):19. 9. Kyle RA, Bayrd ED. "Primary" systemic amyloidosis and myeloma. Discussion of relationship and review of 81 cases. Arch Intern Med. 1961;107:344-353. 10. Kyle RA, Greipp PR. Primary systemic amyloidosis: comparison of melphalan and prednisone versus placebo. Blood. 1978;52(4):818-827. 11. Palladini G, Milani P, Merlini G. Management of AL amyloidosis in 2020. Hematol Am Soc Hematol Educ Program. 2020;2020(1):363-371. 12. Kastritis E, Palladini G, Minnema MC, et al. Daratumumab-based treatment for immunoglobulin light-chain amyloidosis. N Engl J Med. 2021;385(1):46-58. 13. Minnema MC, Dispenzieri A, Merlini G, et al. Outcomes by cardiac stage in patients with newly diagnosed AL amyloidosis: phase 3 ANDROMEDA trial. JACC CardioOncol. 2022;4(4):474-487. 14. Kimmich CR, Terzer T, Benner A, et al. Daratumumab for systemic AL amyloidosis: prognostic factors and adverse outcome with nephrotic-range albuminuria. Blood. 2020;135(18):1517-1530. 15. Chakraborty R, Rosenbaum C, Kaur G, et al. First report of

outcomes in patients with stage IIIb AL amyloidosis treated with Dara-VCD front-line therapy. Br J Haematol. 2023;201(5):913-916. 16. Theodorakakou F, Briasoulis A, Fotiou D, et al. Outcomes for patients with systemic light chain amyloidosis and Mayo stage 3B disease. Hematol Oncol. 2023 Mar 27. doi: 10.1002/hon.3135. [Epub ahead of print] 17. Gertz MA, Comenzo R, Falk RH, et al. Definition of organ involvement and treatment response in immunoglobulin light chain amyloidosis (AL): a consensus opinion from the 10th International Symposium on Amyloid and Amyloidosis, Tours, France, 18-22 April 2004. Am J Hematol. 2005;79(4):319-328. 18. Dispenzieri A, Gertz MA, Kumar SK, et al. High sensitivity cardiac troponin T in patients with immunoglobulin light chain amyloidosis. Heart. 2014;100(5):383-388. 19. Palladini G, Hegenbart U, Milani P, et al. A staging system for renal outcome and early markers of renal response to chemotherapy in AL amyloidosis. Blood. 2014;124(15):2325-2332. 20. 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):4541-4549. 21. Palladini G, Schonland SO, Sanchorawala V, et al. Clarification on the definition of complete haematologic response in lightchain (AL) amyloidosis. Amyloid. 2021;28(1):1-2. 22. Palladini G, Barassi A, Klersy C, et al. The combination of highsensitivity cardiac troponin T (hs-cTnT) at presentation and changes in N-terminal natriuretic peptide type B (NT-proBNP) after chemotherapy best predicts survival in AL amyloidosis. Blood. 2010;116(18):3426-3430. 23. Muchtar E, Dispenzieri A, Wisniowski B, et al. Graded cardiac response criteria for patients with systemic light chain amyloidosis. J Clin Oncol. 2022;41(7):1393-1403. 24. Manwani R, Foard D, Mahmood S, et al. Rapid hematologic responses improve outcomes in patients with very advanced (stage IIIb) cardiac immunoglobulin light chain amyloidosis. Haematologica. 2018;103(4):e165-e168. 25. Kastritis E, Minnema MC, Dimopoulos MA, et al. Efficacy and safety of daratumumab monotherapy in newly diagnosed patients with stage 3B light chain amyloidosis: a phase 2 study by the European Myeloma Network. Blood. 2022;140(Suppl 1):1841-1843. 26. Basset M, Milani P, Foli A, et al. Early cardiac response is possible in stage IIIb cardiac AL amyloidosis and is associated with prolonged survival. Blood. 2022;140(18):1964-1971.

Haematologica | 109 January 2024

230

ARTICLE - Plasma Cell Disorders

DIS3 depletion in multiple myeloma causes extensive perturbation in cell cycle progression and centrosome amplification Vanessa K. Favasuli,1,2* Domenica Ronchetti,1* Ilaria Silvestris,1 Noemi Puccio,3,4 Giuseppina Fabbiano,5 Valentina Traini,5 Katia Todoerti,6 Silvia Erratico,7,8 Alessia Ciarrocchi,3 Valentina Fragliasso,3 Domenica Giannandrea,9 Francesca Tumiatti,10 Raffaella Chiaramonte,9 Yvan Torrente,7 Palma Finelli,10,11 Eugenio Morelli,2 Nikhil C. Munshi,2 Niccolò Bolli,1,5 Antonino Neri12 and Elisa Taiana5 1

Department of Oncology and Hemato-Oncology, University of Milan, Milan, Italy; 2Department of Medical Oncology, Jerome Lipper Multiple Myeloma Center, Dana-Farber Cancer Institute, Boston, MA, USA; 3Laboratory of Translational Research, Azienda USL-IRCCS di Reggio Emilia, Reggio Emilia, Italy; 4Clinical and Experimental Medicine PhD Program, University of Modena and Reggio Emilia, Modena, Italy; 5Hematology, Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan, Italy; 6Department of Pathology and Laboratory Medicine, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy; 7Stem Cell Laboratory, Department of Pathophysiology and Transplantation, University of Milan, Centro Dino Ferrari, Unit of Neurology, Fondazione IRCCS Cà Granda Ospedale Maggiore Policlinico, Milan, Italy; 8 Novystem Spa, Milan, Italy; 9Department of Health Sciences, University of Milan, Milan, Italy; 10 Medical Genetics Laboratory, Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan, Italy; 11Department of Medical Biotechnology and Translational Medicine, University of Milan, Segrate, Italy and 12Scientific Directorate, Azienda USL-IRCCS di Reggio Emilia, Reggio Emilia, Italy

Correspondence: A Neri antonino.neri@ausl.re.it Received: Accepted: Early view:

April 4, 2023. July 5, 2023. July 13, 2023.

*VKF and DR contributed equally as first authors.

Abstract DIS3 gene mutations occur in approximately 10% of patients with multiple myeloma (MM); furthermore, DIS3 expression can be affected by monosomy 13 and del(13q), found in roughly 40% of MM cases. Despite the high incidence of DIS3 mutations and deletions, the biological significance of DIS3 and its contribution to MM pathogenesis remain poorly understood. In this study we investigated the functional role of DIS3 in MM, by exploiting a loss-of-function approach in human MM cell lines. We found that DIS3 knockdown inhibits proliferation in MM cell lines and largely affects cell cycle progression of MM plasma cells, ultimately inducing a significant increase in the percentage of cells in the G0/G1 phase and a decrease in the S and G2/M phases. DIS3 plays an important role not only in the control of the MM plasma cell cycle, but also in the centrosome duplication cycle, which are strictly co-regulated in physiological conditions in the G1 phase. Indeed, DIS3 silencing leads to the formation of supernumerary centrosomes accompanied by the assembly of multipolar spindles during mitosis. In MM, centrosome amplification is present in about a third of patients and may represent a mechanism leading to genomic instability. These findings strongly prompt further studies investigating the relevance of DIS3 in the centrosome duplication process. Indeed, a combination of DIS3 defects and deficient spindle-assembly checkpoint can allow cells to progress through the cell cycle without proper chromosome segregation, generating aneuploid cells which ultimately lead to the development of MM.

Introduction Multiple myeloma (MM) is a hematologic malignancy that is still incurable despite the recent introduction of a large array of innovative therapies.1 MM is characterized by the abnormal proliferation of plasma cells (PC) in the bone marrow and has different clinical courses and a highly heterogeneous

genetic background with both structural chromosomal alterations and specific gene mutations affecting the expression and the activity of both putative oncogenes and tumor suppressor genes.2 Among the frequently mutated genes in MM, DIS3 has been reported to be mutated in roughly 10% of patients and to have a significant impact on clinical outcome.3-6 Despite the

Haematologica | 109 January 2024

231

ARTICLE - DIS3 affects centrosome duplication in myeloma detailed overview of DIS3 mutations, their functional consequences on the pathogenesis of MM remain largely unknown, to the point that it is still unclear whether DIS3 behaves as an oncogene or a tumor suppressor gene.7,8 Aberrant expression of DIS3 has been reported in different tumor types.7,9,10 Notably, monosomy 13 and del(13q), which occur in approximately 40% of MM cases, could affect DIS3 expression.11,12 DIS3 encodes for a highly conserved ribonuclease13-15 that endows the catalytic activity to the exosome, a multi-subunit complex that processes and degrades RNA for gene expression regulation, mRNA quality control, and small RNA processing.16,17 Moreover, studies in Schizosaccharomyces pombe have revealed functions of DIS3 in chromosome segregation,18,19 cell cycle progression,20,21 and spindle assembly.22 The involvement of DIS3 in cell cycle regulation has also been demonstrated in Drosophila melanogaster. In addition, DIS3 is required for the development of a multicellular organism participating in cell type-specific RNA turnover.23 In humans, DIS3 has been described to shape the RNA polymerase II transcriptome by degrading a variety of unwanted transcripts.24 Data concerning the functional role of DIS3 in myeloma are limited to a recent study showing that DIS3 depletion in different cell types, including malignant PC, causes a pervasive accumulation of DNA:RNA hybrids that induce genomic DNA double-strand breaks, eventually leading to genomic instability by increasing mutational load.25 In this study, we aimed to expand our knowledge on the biological role of DIS3 in MM by investigating the functional consequences of its depletion in myeloma cells. Our data indicate that DIS3 silencing causes marked perturbations in cell cycle progression and the process of mitosis.

V.K. Favasuli et al.

assess patterns of differential expression between two or multiple molecular groups. The Dunn test was used for pairwise comparisons. P values were corrected using the Benjamini-Hochberg method, and adjusted P values <0.05 were considered statistically significant. Survival analyses were performed as described in the Online Supplementary Methods. Multiple myeloma cell lines and drugs The AMO-1 cell line was kindly provided by Dr. C. Driessen (University of Tubingen, Germany). NCI-H929 and U266 cell lines were purchased from DSMZ, which certified authentication performed by short tandem repeat DNA typing. All human MM cell lines (HMCL) were immediately frozen and used from the original stock within 6 months. HMCL were cultured in RPMI-1640 medium (Gibco®, Life Technologies, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum, 50 U/mL penicillin, and 50 μg/mL streptomycin (Gibco®), at 37°C in 5% CO2 atmosphere, and routinely tested to rule out mycoplasma contamination. For cell cycle analyses, cells were synchronized with Synchroset kit (EuroClone, Milan, Italy). For mitotic analyses, cells were synchronized with RO-3306 and MG-132 (Selleckchem, Aurogene, Rome, Italy).

Primary patients’ cells CD138+ cells were isolated from the bone marrow aspirates of MM patients by Ficoll-Hypaque (Lonza Group, Basel, Switzerland) density gradient sedimentation, followed by antibody-mediated positive selection using anti-CD138 magnetic activated cell separation microbeads (Miltenyi Biotech, Gladbach, Germany). The purity of immunoselected cells was assessed by flow-cytometry analysis using a phycoerythrin-conjugated CD138 monoclonal antibody and Methods standard procedures. CD138+ cells from MM patients were Full details of quantitative real-time polymerase chain reac- seeded and cultured in RPMI-1640 medium (Gibco®, Life tions, cell cycle analysis and apoptosis, immunofluo- Technologies) supplemented with 20% fetal bovine serum rescence, gapmeR design and gymnotic delivery, proteomic (Lonza Group Ltd) and 1% penicillin/streptomycin (Gibco®, assays, and gene expression analysis are provided in the On- Life Technologies). line Supplementary Methods. Ethics approval and consent to participate Multi-omics data from the CoMMpass study Written informed consent to participation in the study was Multi-omics data about bone marrow MM samples at base- obtained from all patients in accordance with the Declarline (BM_1) were freely accessible from the Multiple Myeloma ation of Helsinki. The study was approved by the Ethical Research Foundation (MMRF) CoMMpass study (https://re- Committee of the Fondazione IRCCS Ca’ Granda Ospedale search.themmrf.org/) including more than 1,000 MM patients Maggiore Policlinico (N. 575, 03/29/2018). from several worldwide sites and retrieved from the Interim Analysis 15a (MMRF_CoMMpass_IA15a, accessed on 16 October 2020). Details about the molecular and clinical data of Results the CoMMpass cohort selected for the present study are provided in the Online Supplementary Methods. DIS3 expression in multiple myeloma correlates with molecular subtypes Statistical and survival analyses To gain insight into the role of DIS3 in MM, we started by Wilcoxon rank-sum and Kruskal-Wallis tests were applied to exploring its expression in pathological samples compared Haematologica | 109 January 2024

232

ARTICLE - DIS3 affects centrosome duplication in myeloma to normal controls. In detail, we investigated DIS3 expression levels in a proprietary, publicly available, RNA dataset profiled by microarrays that includes four healthy donors, 130 MM patients, 24 cases of primary plasma cell leukemia (PCL), and 12 cases of secondary PCL (GSE66293). Considering that DIS3 expression levels could be affected by the presence of chromosome 13 deletion (del13), we stratified patients based on the presence of this molecular lesion. We found that patients with MM, primary PCL and secondary PCL without del13 showed significantly higher DIS3 expression levels than normal controls; conversely, pathological samples with del13 displayed DIS3 expression levels comparable to those of normal controls (Figure 1A). Next, we focused on the expression pattern of DIS3 in MM by taking advantage of a large cohort of MM patients enrolled in the MMRF CoMMpass study. To this end, we evaluated DIS3 expression levels in PC from 774 MM pa-

V.K. Favasuli et al.

tients included in the CoMMpass dataset (Figure 1B). In detail, DIS3 expression spanned a limited range of estimated values (3.3-43.2; median: 13.16) of transcripts per million. MM patients with t(11;14) chromosomal translocation showed significantly higher DIS3 expression levels than those of patients with other translocations (namely t(4;14), MAF translocations, MYC, double translocations) or negative for any translocations. To assess DIS3 expression profiles in relation to major molecular aberrations in MM, we investigated 660 MM patients of the CoMMpass cohort for whom data on expression, non-synonymous somatic mutations, and copy number alterations, collected by RNA sequencing, whole exome sequencing and next-generation sequencing-based fluorescence in situ hybridization, respectively, were available (Online Supplementary Table S1). We observed that the presence of del13 was associated with a reduction of DIS3 expression levels. Interestingly,

Figure 1. DIS3 expression analysis in plasma cell dyscrasias. (A) Box plot of DIS3 expression level in healthy donors (N) and plasma cell (PC) dyscrasias (proprietary dataset, GSE66293). The four N are RNA samples from bone marrow PC purified (>90%) from normal individuals and purchased from Voden (Medical Instruments IT). Total RNA samples from highly purified bone marrow CD138+ PC were profiled by Gene 1.0 ST array. (B) Box plot of DIS3 expression level in main IGH translocation groups in 774 cases from the CoMMpass cohort. The Kruskal-Wallis test was applied to assess differences in expression levels between groups. The pairwise comparisons were performed by the Dunn test; statistically significant results are marked in red, bold in the tables. MM: multiple myeloma; pPCL: primary plasma cell leukemia; sPCL: secondary plasma cell leukemia; trx.MAF: MAF translocation, trx.MYC: MYC translocation, double.trx: presense of two translocations, trx.neg: absence of the considered translocations. Haematologica | 109 January 2024

233

ARTICLE - DIS3 affects centrosome duplication in myeloma significantly higher DIS3 expression levels were observed in MM patients carrying t(11;14), in particular in those without the del13 alteration (Online Supplementary Figures S1A and S2A). Significantly higher DIS3 expression levels were also found in MM patients carrying non-synonymous somatic mutations in the RAS/BRAF or FAM46C genes, whereas lower expression levels were evidenced in samples with 1q-gain, del13, t(4;14), or a DIS3 gene mutation (Online Supplementary Figure S1A). However, concerning patients with DIS3 mutations, we observed that samples with DIS3 hotspot mutations had higher DIS3 expression levels than those from MM patients carryng nonhotspot DIS3 mutations (Online Supplementary Figure S2B). No significant differences in DIS3 expression levels were observed in relation to del(17p)/TP53, t(6;14), MYC or MAF translocations, del(1p), the occurrence of non-synonymous somatic mutations in TRAF3 or TP53 genes, or hyperdiploid cases (Online Supplementary Figure S1B). Finally, in order to investigate the relevance of DIS3 expression levels in clinical outcome, we considered 767 MM patients from the CoMMpass dataset with available clinical data. High versus low expression groups were determined according to the median cutoff value for DIS3 expression level across the entire dataset. Our data showed that DIS3 expression levels did not affect clinical outcome in terms of either overall survival or progressionfree survival (Online Supplementary Figure S3). DIS3 knockdown inhibits proliferation in multiple myeloma cell lines and affects cell cycle progression and distribution To investigate the functional role of DIS3 in MM, we exploited a loss-of-function approach by using locked nucleic acid-gapmeR antisense oligonucleotides that trigger RNAse-H-dependent degradation of transcripts. Specifically, we designed three “in-house” sequences (gDIS3#2/13/15) able to recognize all three DIS3 isoforms (Online Supplementary Figure S4A) and tested them in the AMO-1 HMCL by using electroporation. Among them, the gDIS3#13 gapmeR showed the strongest silencing efficiency (nearly 70%) after 24 h (Online Supplementary Figure S4B). To achieve a more pronounced and prolonged knockdown (KD) effect, we optimized its gymnotic delivery;26 the striking downregulation of DIS3 at both the transcriptional and protein levels obtained in different HMCL (AMO-1, NCI-H929 and U266) (Figure 2A, B), suggested that this strategy was a valuable tool for subsequent investigations. In particular, we found that DIS3-KD in AMO-1, NCI-H929 and U266 cells was associated with a significant reduction of cell growth and an increase in apoptosis at 6 days after gDIS3#13 delivery (Figure 3A, B). Notably, HMCL showed the same phenotype at 48 h after electroporation when an RNA interference strategy was used (Online Supple-

V.K. Favasuli et al.

mentary Figure S5). Furthermore, DIS3-KD dramatically reduced (fold change 5-10) the clonogenic potential of MM cells (Figure 3C). Finally, we evaluated the biological effects of gDIS3#13 ex-vivo on primary CD138+ tumor cells purified from the bone marrow of four MM patients. DIS3KD led to an important alteration of cell morphology, causing cytoplasmic vacuolization, which suggests a suffering cell phenotype, and affected cell membrane integrity, converting it to one typical of apoptotic cells (Figure 3D), in agreement with the reduced viability of gDIS3#13treated primary cells (Online Supplementary Table S2). To inspect whether the inhibitory effects of DIS3 silencing on proliferative potential are associated with specific changes in cell cycle progression, we analyzed AMO-1 and NCI-H929 cells silenced with gDIS3 for 4 days. DIS3-KD in HMCL caused a significant increase in the number of cells in the G0/G1 phase of the cell cycle, and a decrease in the percentage of cells distributed in the S and G2/M phases, indicating G0/G1 cell cycle arrest (Online Supplementary Figure S6). To investigate the alterations in cell cycle progression in more detail, we synchronized HMCL using the Synchroset reagent, and followed cell cycle progression at different timepoints (Online Supplementary Figure S7) by collecting samples before blocking, and during and after release (Online Supplementary Figures S8-S10). In agreement with previous results, we found that AMO-1, NCI-H929 and U266 exhibited a modulation of the distribution of cell cycle phases after DIS3-KD, with a significantly increased percentage of G0/G1 phase events and a decrease of the S and G2/M phases (Figure 4A, B). To further confirm that DIS3-KD results in a perturbation of cell cycle progression in HMCL, we investigated proteins that have been reported to be associated with different cell cycle checkpoints (Figure 5). In line with the cytofluorimetric analysis showing cell cycle arrest in the G0/G1 phase, DIS3-KD cells showed a slight increase of CCNE1 protein levels, whereas there was a decrease in the expression levels of the cyclin A protein (CCNA2), whose accumulation creates a decision window to enter the S phase and to initiate DNA replication. In agreement with a reduced G2/M phase of the cell cycle, DIS3-KD cells displayed reductions of both cyclin B (CCNB1) total protein and its phosphorylated fraction (pCCNB1), whose accumulation in the cyclin A/B-CDK1 complex creates a second decision window.27,28 CCNB1 expression changes during the different phases of the cell cycle, reaching its maximum in G2/M transition. Cell entry into mitosis depends on the binding of CCNB1 to CDK1 to form the mitosis-promoting factor (MPF); CCNB1 phosphorylation promotes nuclear translocation of MPF and is important for the decisiveness and irreversibility of mitotic entry. However, MPF remains in the inactive state until phosphorylation of CDK1 (pCDK1). In our experimental condition, gDIS3-treated NCIH929 and U266 cells showed a consistent reduction of the

Haematologica | 109 January 2024

234

ARTICLE - DIS3 affects centrosome duplication in myeloma pCDK1 fraction; pCDK1 was not affected in the gDIS3treated AMO-1 cell line, although the total amount of CDK1 protein was reduced. We also investigated CDC20, which is a critical checkpoint effector of mitosis and whose overexpression is associated with several types of cancer, including MM.29,30 In our experiments, we observed an important decrease in the expression level of CDC20 protein but not of its phosphorylated fraction (pCDC20) in DIS3-KD cells. pCDC20 is important in the early steps of mitosis; in late anaphase and the G1 phase, CDC20 is dephosphorylated and targeted for degradation by the proteasome and is not expressed again until the S phase. Finally, we observed a strong downmodulation of histone H3 variant CENP-A in all DIS3-depleted cells tested; this protein plays a fundamental role in defining centromere identity and structure and is a critical protein for the proper formation of the mitotic spindle.31

V.K. Favasuli et al.

Transcriptional signature associated with DIS3 silencing in the NCI-H929 cell line In order to identify DIS3 downstream-related pathways in MM, we investigated gene expression profiling of synchronized NCI-H929 cells after 5 days of gDIS3#13 gymnotic delivery. Out of 55,540 globally analyzed genes, the expression of 4,032 genes resulted significantly modulated (false discovery rate <10%) by SAM analysis. These were virtually all downregulated (3,995 genes, 99%) in DIS3-silenced cells and principally involved protein-coding genes (2,608 genes, 65%) and to a lesser extent long non-coding RNA (464 long non-coding RNA, 12%) (Online Supplementary Table S3). As expected based on previous biological results, chromosome organization, chromatin and histone modification, and cell cycle checkpoint were recognized among the top 20 most significantly enriched Gene Ontology Biological Process terms (Online Supplementary Figure S11A). In addition, genes involved in serine/threonine protein kinase activity, with helicase function, cata-

Figure 2. DIS3 expression analysis in DIS3-depleted cells. (A) Quantitative real-time polymerase chain reaction analysis of DIS3 mRNA in AMO-1, NCI-H929 and U266 cells at the indicated timepoints. DIS3 mRNA expression is represented as 2−ΔΔCt relative to the GAPDH housekeeping gene and the scrambled condition at the same timepoint used as a calibrator. *P<0.05, **P<0.01, Student t test. (B) Western blot analyses of DIS3-KD in AMO-1, NCI-H929 and U266 cells. Western blot results are quantified and plotted in a histogram. SCR: scrambled condition. Haematologica | 109 January 2024

235

ARTICLE - DIS3 affects centrosome duplication in myeloma

V.K. Favasuli et al.

Continued on following page. Haematologica | 109 January 2024

236

ARTICLE - DIS3 affects centrosome duplication in myeloma

V.K. Favasuli et al.

Figure 3. Functional impact of DIS3 depletion in multiple myeloma cells. (A) Growth curves of AMO-1, NCI-H929 and U266 cells following DIS3 silencing. (B) Flow cytometry analyses of apoptosis in AMO-1, NCI-H929 and U266 cells 6 days after treatment with gDIS3 (5 μM). (C) Colony formation assay performed on AMO-1, NCI-H929 and U266 cells treated for 21 days with gDIS3; representative pictures of colonies at day 21 are also shown, **P<0.01, ***P<0.001, Student t test. (D) Rappresentative images of May-Grünwald Giemsa staining of two CD138+ primary tumors treated for 6 days with gDIS3 (50x magnification). SCR: scrambled condition; 7-AAD: 7 amino actinomycin D.

Figure 4. Cell cycle analysis in DIS3-depleted cells. (A) Cell cycle progression in synchronized AMO-1, NCI-H929, and U266 cells was monitored by propidium iodide staining and fluorescence activated cell sorting analysis of the DNA content of cells at 6 h after release. (B) The percentage of cell cycle distribution is represented in the histograms; the standard deviations of three replicates are reported. *P<0.05, **P<0.01, ***P<0.001, Student t test. SCR: scrambled condition.

lytic activity on RNA, or with tubulin binding function were evidenced in the top 20 remarkably enriched Gene Ontology Molecular Function terms under DIS3 silencing (Online Supplementary Figure S11B). In particular, numerous protein-coding genes that were significantly enriched in

multiple annotation categories linked to complex interactions, regarding chromosome organization and nuclear division, or molecular activities on nucleic acids and tubulin, resulted globally downregulated in DIS3-silenced cells (Online Supplementary Figure S12).

Haematologica | 109 January 2024

237

ARTICLE - DIS3 affects centrosome duplication in myeloma

V.K. Favasuli et al.

Figure 5. Western blot analyses of cell cycle checkpoint proteins. (A) Analysis performed in synchronized AMO-1, NCI-H929 and U266 cells 6 h after release. For NCI-H929 cells, pCDC20, CDC20 and CENP-A were evaluated on the same blot and thus with the same GAPDH control; to improve the overall symmetry of the figure and its comprehension, the same GAPDH panel has been reported twice. (B) Western blot results are quantified and plotted in a histogram. (C) Histogram of the modulation of the phosphorylated protein fraction upon gDIS3 treatment, calculated as a ratio of phosphorylated protein to total protein in gDIS3treated cells compared to the ratio of phosphorylated protein to total protein in scramble-treated ones. SCR: scramble condition. Haematologica | 109 January 2024

238

ARTICLE - DIS3 affects centrosome duplication in myeloma

V.K. Favasuli et al.

Continued on following page. Haematologica | 109 January 2024

239

ARTICLE - DIS3 affects centrosome duplication in myeloma

V.K. Favasuli et al.

Figure 6. Transcriptomic analyses in DIS3-depleted cells. Enrichment plots of selected gene set enrichment analysis (GSEA) gene sets significantly modulated in DIS3-silenced cells in comparison to scrambled NCI-H929 cells. GSEA was performed on global annotated protein-coding gene expression profiles (19,048 genes) and the most significant gene sets were selected under stringent conditions (nominal P value <0.05 and false discovery rate q value <5%). The enrichment score and nominal P values are reported for each plot. NES: normalized enrichment score; FDR: false discovery rate.

Finally, to find possible molecular subsets of genes that were coordinately modulated in DIS3-silenced cells in comparison to control NCI-H929 cells, gene set enrichment analysis was performed on global annotated protein coding gene expression profiles and the most significant gene sets were selected under stringent conditions (nominal P value <0.05 and false discovery rate q value <5%) (Online Supplementary Table S4). Among the most significantly upregulated gene sets in DIS3-silenced NCI-H929 cells, some were involved in degradation of extracellular matrix and DNA and histone epigenetic modifications, whereas cell cycle checkpoint, chromosome organization, RNA processing and degradation, DNA recombination and repair, regulation of TP53 activity and numerous cell signaling pathways resulted among the most significantly underexpressed gene sets in DIS3-silenced cells (Figure 6, Online Supplementary Table S4). Based on our evidence that DIS3 can regulate cell cycle, chromosome organization and DNA repair processes, we validated the expression of some genes involved in these processes by means of quantitative real-time polymerase chain reaction in DIS3-silenced AMO-1, NCI-H929, and U266 cell lines. In particular, we considered RAD51B and ARID5B, which are involved in homologous recombination and DNA repair processes; in addition, we investigated CCNB1, CDC20, the microtubule motor protein KIF14, the DNA topoisomerase TOP2A and the RNA polymerase POLR2H, all of which are involved in the cell cycle process. All these genes resulted downregulated upon DIS3 silencing in all cell lines (Figure 7A). The downregulation of these genes was also validated ex-vivo in DIS3-depleted primary CD138+ tumor cells, except for CCNB1 and CDC20, likely because purified PC hardly duplicate (Figure 7B). DIS3 depletion affects mitotic spindle organization and the geometry of multiple myeloma cells Considering previously reported data in yeast and Drosophila indicating that DIS3-KD can affect chromosome segregation and spindle assembly, and based on our evidence that DIS3-KD in HMCL leads to a decrease of CENPA, which is fundamental for the proper formation of the mitotic spindle, we investigated whether DIS3 depletion could affect mitotic spindle formation also in MM PC. To obtain a significant number of mitotic cells to analyze, DIS3-KD cells were synchronized at the M phase of the cell cycle (Figure 8A). In detail, cells were immunostained for DIS3, centrosomes (γ-tubulin), and mitotic spindle fibers (α-tubulin), and nuclei were counterstained with

DAPI. Notably, we found that DIS3 silencing leads to the formation of supernumerary centrosomes accompanied by important changes in mitotic spindle organization and geometry in HMCL. Hence, in contrast to control cells which displayed normal spindle localization and physiological bipolar cell division, DIS3-KD cells showed the presence of multipolar spindles during mitosis (Figure 8B, C, Online Supplementary Figure S13).

Discussion Aberrant expression of DIS3 has been described in different tumors types, including MM. Of note, the frequent and almost specific occurrence of DIS3 mutations observed in MM influences the clinical outcome of the patients.3 However, the pattern of DIS3 expression and its functional and pathogenic role in PC malignancies has not yet been clarified. In this study, we investigated the effects of DIS3 depletion in myeloma cells, providing insights into its putative role in the pathobiology of the disease. First of all, we explored DIS3 expression in patients affected by MM or PCL and found that pathological samples expressed higher DIS3 transcript levels than those of normal controls; indeed, even in patients carrying del13, despite their loss of heterozygosity, the amount of DIS3 transcript did not fall below the levels of normal controls. Our analysis showed that DIS3 expression levels did not influence the clinical outcome of MM in patients included in the CoMMpass dataset; however, this result could be affected by the fact that the group with low DIS3 expression level was largely enriched in cases with del13 and/or t(4;14), whereas the group with higher DIS3 expression level was enriched in patients with t(11;14) and/or hotspot DIS3 mutations (Online Supplementary Figure S2). Overall, these data suggest that the co-occurrence of different molecular alterations in MM patients could mask the specific oncogenic function of DIS3 expression in the disease.32 We, therefore, investigated the biological role of DIS3 in vitro by exploiting a loss-of-function approach in HMCL. Our experiments revealed that DIS3-KD largely affects cell cycle progression in MM cells, leading to a decrease in the fraction of cells distributed in the S and G2/M phases, and a consistent increase of cells in the G0/G1 phase. These findings strongly suggest that DIS3 plays an important role in the control of the G1 phase of the cell cycle. In agreement with this idea, we found that in the CoMMpass da-

Haematologica | 109 January 2024

240

ARTICLE - DIS3 affects centrosome duplication in myeloma taset, MM patients with t(11;14) and high expression levels of CCND1, which is necessary for the transition from the G1 to the S phase of the cell cycle, were those significantly associated with the highest DIS3 expression levels (Online Supplementary Figure S1). The prominent role of DIS3 in the G1 phase is also suggested by the fact that in different cell types, including HMCL, DIS3-KD gives rise to a pervasive accumulation of DNA:RNA hybrids.25 R-loops are three-stranded structures generated by the annealing of nascent transcripts to the template DNA strand and play important roles in physiological processes; however, uncontrolled hybrid production or lack of their immediate resolution through protein complexes, including RNases, helicases, and topoisomerases, can induce DNA damage and genome instability.33 Importantly, the inhibition of the key homologous recom-

V.K. Favasuli et al.

bination factor RAD51 leads to accumulation of DNA:RNA hybrids in the early G1 phase of the cell cycle,34 indicating that R-loop metabolism mainly occurs in the G1 phase. Our results showed that DIS3 plays an important role not only in the control of cell cycle in MM cells, but also in the centrosome duplication cycle, which are strictly coregulated in physiological conditions in the G1 phase of the cell cycle.35,36 Indeed, for the first time we identified that DIS3KD leads to the formation of supernumerary centrosomes which finally result in altered and multipolar mitotic spindles in the M phase of the cell cycle of MM cells. The requirement of DIS3 RNase activity for proper mitotic cell division and correct chromosome condensation has been already described in yeast and Drosophila melanogaster models.21,22,37-39 In addition, S. pombe DIS3 mutants have been shown to have elongated metaphase spindles and a

Figure 7. Molecular validation of the transcriptomic analysis. (A) Quantitative real-time polymerase chain reaction validation of the indicated genes in DIS3-depleted AMO-1, NCI-H929, and U266 human multiple myeloma cell lines; gene expression is presented as 2−ΔΔCt relative to the GAPDH housekeeping gene and the scrambled condition used as a calibrator. *P<0.05, **P<0.01, ***P<0.001, Student t test. (B) Ex-vivo validation of the indicated genes in DIS3-depleted primary CD138+ tumor cells purified from four multiple myeloma patients. SCR: scrambled condition. Haematologica | 109 January 2024

241

ARTICLE - DIS3 affects centrosome duplication in myeloma

V.K. Favasuli et al.

Figure 8. Images of metaphase spindles in NCI-H929 and AMO-1 cells treated with gDIS3 or scrambled gapmer. (A) Schematic representation of the experimental setup used to synchronize cells. (B) Representative images of multipolar spindles in both cell lines following DIS3-knockdown (right panel for each cell line). Scale bar, 5 μm. (C) Percentage of mitotic cells with defective spindles in NCI-H929 and AMO-1 cells treated with gDIS3 versus scrambled gapmer. IF: immmunofluorescence; SCR: scrambled condition. Haematologica | 109 January 2024

242

ARTICLE - DIS3 affects centrosome duplication in myeloma block in metaphase to anaphase transition.22 Smith et al. provided further evidence that DIS3 is involved in mitotic progression by demonstrating that perturbation of DIS3 in S. cerevisiae affects microtubule localization and structure.21 In MM PC, we found that DIS3 could impact mitotic spindle formation by affecting centrosome duplication. Numerical and/or structural centrosome amplification is a hallmark of solid tumors and hematologic malignancies and is often associated with aberrant tumor karyotypes and poor clinical outcomes.40,41 In MM, centrosome amplification is present in about a third of patients and may represent a mechanism leading to genomic instability.42 However, the clinical impact of centrosome amplification in MM is still debated and existing contradictory results may be explained by the clonal heterogeneity of MM in which each myeloma subclone with centrosome amplification may be potentially associated with different clinical behaviors influenced by complex processes including mitotic dysregulation, apoptotic failure and chromosome instability.43,44 The molecular mechanism by which DIS3-KD could induce formation of supernumerary centrosomes remains to be elucidated. These mechanisms might involve consecutive rounds of centrosome reproduction, or concurrent formation of daughter centrioles around the existing centrioles, or via de novo centrosome assembly independent of preexisting centrioles.40 Our transcriptional analysis revealed that genes involved in cell cycle checkpoint and chromosome organization were among the most significantly underexpressed ones in DIS3-silenced cells (Figure 6, Online Supplementary Table S4). Moreover, genes involved in serine/threonine protein kinase activity, with helicase function, catalytic activity on RNA, or with tubulin-binding function were evidenced in the top 20 remarkably enriched Gene Ontology Molecular Function terms upon DIS3 silencing (Online Supplementary Figure S11B). In particular, several protein-coding genes that were significantly enriched in multiple annotation categories linked in complex interactions, regarding chromosome organization and nuclear division, or molecular activities on nucleic acids and tubulin, resulted globally downregulated in DIS3-silenced cells. Specifically, the downregulation of KIF14, TOP2A and POLR2H, all involved in cell cycle processes, has been validated in vitro and ex vivo. Our transcriptomic analysis also showed that genes involved in DNA methylation were overexpressed in DIS3-si-

V.K. Favasuli et al.

lenced cells. This finding is in agreement with the observation that most of the mRNA expression level is reduced after DIS3-KD despite DIS3 having ribonuclease activity, suggesting that global gene downregulation may be a secondary effect of DIS3 deficiency on DNA methylation. Overall, our data prompt further studies to clarify the role of DIS3 in the process of centrosome duplication; indeed, a combination of DIS3 defects and deficient spindle-assembly checkpoint, already described in MM,45 can allow cells to progress through the cell cycle without proper chromosome segregation, thus generating aneuploid cells which may lead to the development of MM. Disclosures No conflicts of interest to disclose. Contributions VKF, ET, IS, GF, VT, NP, FT, KT, VF, and DR performed experiments and analyzed the data. VKF, ET, DG, and IS performed cytofluorimetric experiments. ET, VKF, and SE performed the confocal analysis, NB, NM, EM AC, RC, PF, and YT provided critical evaluation of experimental data and of the manuscript. DR, VKF, ET, and AN conceived the study and wrote the manuscript. Funding This work was financially supported by grants from the Associazione Italiana Ricerca sul Cancro (AIRC) to AN (IG24365), to ET (MFAG 2022-ID.27606) and to RC (IG20614). NB is funded by the European Research Council under the European Union’s Horizon 2020 research and innovation program (grant agreement n. 817997) and by the Associazione Italiana Ricerca sul Cancro (IG25739). YT acknowledges support from “Piattaforme cellulari per la Ricerca e lo Sviluppo di Terapie Avanzate in Life Science” - Fondo Europeo di Sviluppo Regionale 2014-2020 (POR FESR 20142020). VKF received a fellowship through the PhD program in Experimental Medicine of the University of Milan. NP received a fellowship through the PhD program in Clinical and Experimental Medicine of the University of Modena and Reggio Emilia. The work was partially founded by the Italian Ministry of Health (Current Research IRCCS). Data-sharing statement Data are available from the corresponding author upon reasonable request.

References 1. Munshi NC, Anderson KC. New strategies in the treatment of multiple myeloma. Clin Cancer Res. 2013;19(13):3337-3344. 2. Morgan GJ, Walker BA, Davies FE. The genetic architecture of multiple myeloma. Nat Rev Cancer. 2012;12(5):335-348. 3. Todoerti K, Ronchetti D, Favasuli V, et al. DIS3 mutations in

multiple myeloma impact the transcriptional signature and clinical outcome. Haematologica. 2022;107(4):921-932. 4. Lionetti M, Barbieri M, Todoerti K, et al. A compendium of DIS3 mutations and associated transcriptional signatures in plasma cell dyscrasias. Oncotarget. 2015;6(28):26129-26141.

Haematologica | 109 January 2024

243

ARTICLE - DIS3 affects centrosome duplication in myeloma 5. Boyle EM, Ashby C, Tytarenko RG, et al. BRAF and DIS3 mutations associate with adverse outcome in a long-term follow-up of patients with multiple myeloma. Clin Cancer Res. 2020;26(10):2422-2432. 6. Bolli N, Avet-Loiseau H, Wedge DC, et al. Heterogeneity of genomic evolution and mutational profiles in multiple myeloma. Nat Commun. 2014;5:2997. 7. de Groen FLM, Krijgsman O, Tijssen M, et al. Gene-dosage dependent overexpression at the 13q amplicon identifies DIS3 as candidate oncogene in colorectal cancer progression. Genes Chromosomes Cancer. 2014;53(4):339-348. 8. Weißbach S, Langer C, Puppe B, et al. The molecular spectrum and clinical impact of DIS3 mutations in multiple myeloma. Br J Haematol. 2015;169(1):57-70. 9. Ng D, Toure O, Wei MH, et al. Identification of a novel chromosome region, 13q21.33-q22.2, for susceptibility genes in familial chronic lymphocytic leukemia. Blood. 2007;109(3):916-925. 10. Rose AE, Poliseno L, Wang J, et al. Integrative genomics identifies molecular alterations that challenge the linear model of melanoma progression. Cancer Res. 2011;71(7):2561-2571. 11. Binder M, Rajkumar SV, Ketterling RP, et al. Prognostic implications of abnormalities of chromosome 13 and the presence of multiple cytogenetic high-risk abnormalities in newly diagnosed multiple myeloma. Blood Cancer J. 2017;7(9):e600. 12. Chiecchio L, Dagrada GP, Ibrahim AH, et al. Timing of acquisition of deletion 13 in plasma cell dyscrasias is dependent on genetic context. Haematologica. 2009;94(12):1708-1713. 13. Tomecki R, Kristiansen MS, Lykke-Andersen S, et al. The human core exosome interacts with differentially localized processive RNases: hDIS3 and hDIS3L. EMBO J. 2010;29(14):2342-2357. 14. Schneider C, Leung E, Brown J, Tollervey D. The N-terminal PIN domain of the exosome subunit Rrp44 harbors endonuclease activity and tethers Rrp44 to the yeast core exosome. Nucleic Acids Res. 2009;37(4):1127-1140. 15. Tomecki R, Drazkowska K, Kucinski I, et al. Multiple myelomaassociated hDIS3 mutations cause perturbations in cellular RNA metabolism and suggest hDIS3 PIN domain as a potential drug target. Nucleic Acids Res. 2014;42(2):1270-1290. 16. Mitchell P, Petfalski E, Shevchenko A, Mann M, Tollervey D. The exosome: a conserved eukaryotic RNA processing complex containing multiple 3′→5′ exoribonucleases. Cell. 1997;91(4):457-466. 17. Wasmuth EV, Lima CD. Exo- and endoribonucleolytic activities of yeast cytoplasmic and nuclear RNA exosomes are dependent on the noncatalytic core and central channel. Mol Cell. 2012;48(1):133-144. 18. Kinoshita N, Goebl M, Yanagida M. The fission yeast dis3+ gene encodes a 110-kDa essential protein implicated in mitotic control. Mol Cell Biol. 1991;11(12):5839-5847. 19. Ohkura H, Adachi Y, Kinoshita N, Niwa O, Toda T, Yanagida M. Coldsensitive and caffeine-supersensitive mutants of the Schizosaccharomyces pombe dis genes implicated in sister chromatid separation during mitosis. EMBO J. 1988;7(5):1465-1473. 20. Noguchi E, Hayashi N, Azuma Y, et al. Dis3, implicated in mitotic control, binds directly to Ran and enhances the GEF activity of RCC1. EMBO J. 1996;15(20):5595-5605. 21. Smith SB, Kiss DL, Turk E, Tartakoff AM, Andrulis ED. Pronounced and extensive microtubule defects in a Saccharomyces cerevisiae DIS3 mutant. Yeast Chichester Engl. 2011;28(11):755-769. 22. Murakami H, Goto DB, Toda T, et al. Ribonuclease activity of Dis3 is required for mitotic progression and provides a possible link between heterochromatin and kinetochore function. PloS One. 2007;2(3):e317. 23. Hou D, Ruiz M, Andrulis ED. The ribonuclease Dis3 is an essential

V.K. Favasuli et al. regulator of the developmental transcriptome. BMC Genomics. 2012;13:359. 24. Szczepińska T, Kalisiak K, Tomecki R, et al. DIS3 shapes the RNA polymerase II transcriptome in humans by degrading a variety of unwanted transcripts. Genome Res. 2015;25(11):1622-1633. 25. Gritti I, Basso V, Rinchai D, et al. Loss of ribonuclease DIS3 hampers genome integrity in myeloma by disrupting DNA:RNA hybrid metabolism. EMBO J. 2022;41(22):e108040. 26. Taiana E, Favasuli V, Ronchetti D, et al. Long non-coding RNA NEAT1 targeting impairs the DNA repair machinery and triggers antitumor activity in multiple myeloma. Leukemia. 2020;34(1):234-244. 27. Matthews HK, Bertoli C, de Bruin RAM. Cell cycle control in cancer. Nat Rev Mol Cell Biol. 2022;23(1):74-88. 28. Musacchio A, Salmon ED. The spindle-assembly checkpoint in space and time. Nat Rev Mol Cell Biol. 2007;8(5):379-393. 29. Bruno S, Ghelli Luserna di Rorà A, Napolitano R, Soverini S, Martinelli G, Simonetti G. CDC20 in and out of mitosis: a prognostic factor and therapeutic target in hematological malignancies. J Exp Clin Cancer Res CR. 2022;41(1):159. 30. Lub S, Maes A, Maes K, et al. Inhibiting the anaphase promoting complex/cyclosome induces a metaphase arrest and cell death in multiple myeloma cells. Oncotarget. 2016;7(4):4062-4076. 31. Pesenti ME, Raisch T, Conti D, et al. Structure of the human inner kinetochore CCAN complex and its significance for human centromere organization. Mol Cell. 2022;82(11):2113-2131. 32. Bolli N, Biancon G, Moarii M, et al. Analysis of the genomic landscape of multiple myeloma highlights novel prognostic markers and disease subgroups. Leukemia. 2018;32(12):2604-2616. 33. Santos-Pereira JM, Aguilera A. R loops: new modulators of genome dynamics and function. Nat Rev Genet. 2015;16(10):583-597. 34. Nascakova Z, Boleslavska B, Urban V, et al. RAD51 inhibition induces R-loop formation in early G1 phase of the cell cycle. Int J Mol Sci. 2021;22(7):3740. 35. Prigent C, Uzbekov R. Duplication and segregation of centrosomes during cell division. Cells. 2022;11(15):2445. 36. Lin M, Xie SS, Chan KY. An updated view on the centrosome as a cell cycle regulator. Cell Div. 2022;17(1):1. 37. Graham AC, Kiss DL, Andrulis ED. Core exosome-independent roles for Rrp6 in cell cycle progression. Mol Biol Cell. 2009;20(8):2242-2253. 38. Robinson SR, Oliver AW, Chevassut TJ, Newbury SF. The 3’ to 5’ exoribonuclease DIS3: from structure and mechanisms to biological functions and role in human disease. Biomolecules. 2015;5(3):1515-1539. 39. Snee MJ, Wilson WC, Zhu Y, et al. Collaborative control of cell cycle progression by the RNA exonuclease Dis3 and Ras is conserved across species. Genetics. 2016;203(2):749-762. 40. Mittal K, Kaur J, Jaczko M, et al. Centrosome amplification: a quantifiable cancer cell trait with prognostic value in solid malignancies. Cancer Metastasis Rev. 2021;40(1):319-339. 41. Krämer A, Neben K, Ho AD. Centrosome aberrations in hematological malignancies. Cell Biol Int. 2005;29(5):375-383. 42. Chng WJ, Fonseca R. Centrosomes and myeloma; aneuploidy and proliferation. Environ Mol Mutagen. 2009;50(8):697-707. 43. Dementyeva E, Kryukov F, Kubiczkova L, et al. Clinical implication of centrosome amplification and expression of centrosomal functional genes in multiple myeloma. J Transl Med. 2013;11:77. 44. Kryukova E, Kryukov F, Hajek R. Centrosome amplification and clonal evolution in multiple myeloma: short review. Crit Rev Oncol Hematol. 2016;98:116-121. 45. Díaz-Rodríguez E, Álvarez-Fernández S, Chen X, et al. Deficient spindle assembly checkpoint in multiple myeloma. PloS One. 2011;6(11):e27583.

Haematologica | 109 January 2024

244

ARTICLE - Plasma Cell Disorders

Elotuzumab plus pomalidomide and dexamethasone in relapsed/refractory multiple myeloma: a multicenter, retrospective, real-world experience with 200 cases outside of controlled clinical trials Massimo Gentile,1,2 Ernesto Vigna,1 Salvatore Palmieri,3 Monica Galli,4 Daniele Derudas,5 Roberto Mina,6 Roberta Della Pepa,7 Renato Zambello,8 Enrica Antonia Martino,1 Antonella Bruzzese,1 Silvia Mangiacavalli,9 Elena Zamagni,10,11 Catello Califano,12 Maurizio Musso,13 Concetta Conticello,14 Claudio Cerchione,15 Giuseppe Mele,16 Nicola Di Renzo,17 Massimo Offidani,18 Giuseppe Tarantini,19 Gloria Margiotta Casaluci,20 Angela Rago,21 Roberto Ria,22 Giuseppina Uccello,23 Gregorio Barilà,24 Gaetano Palumbo,25 Alessandra Pompa,26 Donatella Vincelli,27 Marino Brunori,28 Fabrizio Accardi,29 Valeria Amico,30 Angela Amendola,31 Raffaele Fontana,32 Velia Bongarzoni,33 Bernardo Rossini,34 Emilia Cotzia,35 Alessandro

Correspondence: M. Gentile massim.gentile@tiscali.it A. Neri antonino.neri@ausl.re.it Received: Accepted: Early view:

April 7, 2023. July 6, 2023. July 13, 2023.

Gozzetti,36 Rita Rizzi,37 Nicola Sgherza,37 Eleonora Ferretti,38 Giuseppe Bertuglia,6 Davide Nappi,15 Maria Teresa Petrucci,39 Francesco Di Raimondo,14 Antonino Neri,38 Fortunato Morabito40 and Pellegrino Musto37

Department of Onco-Hematology, Hematology Unit, Azienda Ospedaliera Annunziata, Cosenza;

Department of Pharmacy, Health and Nutritional Science, University of Calabria, Rende; 3Hematology

Unit, Ospedale Cardarelli, Naples; 4Hematology and Bone Marrow Transplant Unit, Azienda SocioSanitaria Territoriale-Papa Giovanni XXIII, Bergamo; 5Department of Hematology, Businco Hospital, Cagliari; 6Division of Hematology, AOU Città della Salute e della Scienza di Torino, University of Turin, Turin; 7Hematology Unit, Department of Clinical Medicine and Surgery, University of Naples “Federico II”, Naples; 8University of Padova, Department of Medicine, Hematology Unit and Veneto Institute of Molecular Medicine, Padova; 9Hematology Division, Department of Hematology-Oncology, IRCCS Fondazione Policlinico San Matteo, Pavia; 10IRCCS Azienda Ospedaliero-Universitaria di Bologna, Istituto di Ematologia “Seràgnoli”, Bologna; 11Dipartimento di Medicina Specialistica, Diagnostica e Sperimentale, Università di Bologna, Bologna; 12Onco-Hematology Unit, “A. Tortora” Hospital, Pagani; 13OncoHematology Unit and TMO UOC, Department of Oncology, Palermo; 14Division of Hematology, Azienda Policlinico-S. Marco, University of Catania, Catania; 15Hematology Unit, IRCCS Istituto Romagnolo per lo Studio dei Tumori (IRST) “Dino Amadori”, Meldola; 16Department of Hematology, Hospital Perrino, Brindisi; 17Department of Hematology, “Vito Fazzi” Hospital, Lecce; 18Hematology Unit, AOU Ospedali Riuniti di Ancona, Ancona; 19Hematology Unit, “Dimiccoli” Hospital, Barletta; 20Division of Hematology, Department of Translational Medicine, University of Eastern Piedmont, Novara; 21Haematology Unit, ASL Roma 1 Santo Spirito Hospital of Rome, Rome; 22Department of Biomedical Science, Internal Medicine “G. Baccelli”, Policlinico, University of Bari “Aldo Moro” Medical School, Bari; 23Hematology Department, “G. Garibaldi” Hospital, Catania; 24University of Padova, Department of Medicine, Hematology Unit, Veneto Institute of Molecular Medicine, Padova; 25Department of Hematology, Hospital University Riuniti, Foggia; 26Hematology Unit, Fondazione IRCCS Ca’ Granda, Ospedale Maggiore Policlinico, Milan; 27

Department of Hemato-Oncology and Radiotherapy, Hematology Unit, Great Metropolitan Hospital

“Bianchi-Melacrino-Morelli”, Reggio Calabria; 28Internal Medicine, Ospedale S. Croce, Azienda Ospedaliera Ospedali Riuniti Marche Nord, Fano; 29Department of Hematology I, Azienda Ospedaliera Ospedali Riuniti Villa Sofia-Cervello, Palermo; 30Hematology Unit, AORN “G Rummo”, Benevento; 31

Hematology Unit, Azienda Ospedaliera Regionale “San Carlo”, Potenza; 32Hematology and Transplant

Center, University Hospital “San Giovanni di Dio e Ruggi d’Aragona”, Salerno; 33Department of Hematology San Giovanni-Addolorata Hospital, Rome; 34Hematology, Ospedale Giovanni Paolo II°, Bari; 35

Section of Hematology, Ospedale “E. Muscatello-Augusta”, Siracusa; 36Hematology, Azienda

Ospedaliera Universitaria Senese, University of Siena, Siena; 37Hematology Section, Department of Precision and Regenerative Medicine and Ionian Area (DiMePRe-J), “Aldo Moro” University of Bari, Bari; 38

Scientific Directorate, Azienda USL-IRCCS di Reggio Emilia, Reggio Emilia; 39Department of

Translational and Precision Medicine, Hematology Azienda Policlinico Umberto I Sapienza University of Rome, Rome and 40Biotechnology Research Unit, AO of Cosenza, Cosenza, Italy

Haematologica | 109 January 2024

245

ARTICLE - An Italian real-world experience of EloPd for RRMM

M. Gentile et al.

Abstract In the ELOQUENT-3 trial, the combination of elotuzumab, pomalidomide and dexamethasone (EloPd) proved to have a superior clinical benefit over pomalidomide and dexamethasone with a manageable toxicity profile, leading to its approval for the treatment of patients with relapsed/refractory multiple myeloma (RRMM) who have received at least two prior therapies, including lenalidomide and a proteasome inhibitor. We report here a real-world experience of 200 cases of RRMM treated with EloPd in 35 Italian centers outside of clinical trials. In our dataset, the median number of prior lines of therapy was two, with 51% of cases undergoing autologous stem cell transplant and 73% having been exposed to daratumumab. After a median follow-up of 9 months, 126 patients had stopped EloPd, most of them (88.9%) because of disease progression. The overall response rate was 55.4%, a finding in line with the pivotal trial results. Regarding adverse events, the toxicity profile in our cohort was similar to that in the ELOQUENT-3 trial, with no significant differences between younger (<70 years) and older patients. The median progression-free survival was 7 months, which was shorter than that observed in ELOQUENT-3, probably because of the different clinical characteristics of the two cohorts. Interestingly, International Staging System stage III disease was associated with worse progression-free survival (hazard ratio=2.55). Finally, the median overall survival of our series was shorter than that observed in the ELOQUENT-3 trial (17.5 vs. 29.8 months). In conclusion, our real-world study confirms that EloPd is a safe and possible therapeutic choice for patients with RRMM who have received at least two prior therapies, including lenalidomide and a proteasome inhibitor.

Introduction The treatment landscape of multiple myeloma (MM) has changed dramatically over the years as the result of the introduction of several new drugs that have improved these patients’ survival.1,2 At present, proteasome inhibitors and immunomodulatory drugs are still the fundamental backbones of MM therapy. However, given the encouraging results from clinical trials, especially among double-refractory MM patients, monoclonal antibodies, a new class of drugs, are now being used with proteasome inhibitors and immunomodulatory drugs and are being incorporated in earlier lines of therapies.3 The use of triplet combinations in clinical practice allows for deeper and more sustained responses with an acceptable safety profile.3 Elotuzumab is a humanized immunoglobulin G1 immunostimulatory monoclonal antibody that is directed against signaling lymphocytic activation molecule F7 (SLAMF7).4 SLAMF7 is a glycoprotein expressed on myeloma cells and natural killer cells, which promote MM cell proliferation and survival.5 Thus, the mechanism of action of elotuzumab is the prevention of interactions that allow the growth and sustainment of neoplastic cells. Moreover, elotuzumab stimulates natural killer cells by strengthening their antibody-dependent cellular cytotoxicity.6,7 As found in in vitro models, this phenomenon is amplified when elotuzumab is combined with lenalidomide.8 It was hypothesized that a similar effect would be observed in patients with relapsed or refractory MM (RRMM). Indeed, based on the results from a phase III trial (ELOQUENT-2), elotuzumab was first approved by the Food and Drug Administration (FDA) in November 2015 and by the European Medicines Agency in January 2016 in combination with lenalidomide and dexamethasone for the treatment of MM patients who

had received at least one prior line of therapy.9 Our group confirmed the safety and efficacy of this combination in a cohort of RRMM treated outside clinical trials.10-13 Like lenalidomide, pomalidomide is an immunomodulatory drug that directly determines MM cell death and has immune-enhancing effects via binding to cereblon.14 However, compared to lenalidomide, pomalidomide demonstrated a more potent antineoplastic activity towards lenalidomide-resistant MM cell lines in vitro and in preclinical in vivo studies. Moreover, it was shown that the combination of elotuzumab with pomalidomide exerts synergistic antimyeloma effects.15 These results laid the groundwork for an in vivo combination. ELOQUENT-3, a multicenter, randomized, controlled, open-label, phase II trial, investigated the efficacy and safety of elotuzumab in combination with pomalidomide and dexamethasone (EloPd) compared to pomalidomide and dexamethasone in the setting of patients with RRMM who had been previously treated with lenalidomide and a proteosome inhibitor.16 After a follow-up of 45 months, the study demonstrated that the triplet combination still improved progression-free survival (PFS) and overall survival (OS), with a lower rate of adverse events compared to that in the control arm.17 Here, we present the outcomes of 200 heavily pre-treated MM patients who received EloPd outside of clinical trials to evaluate the safety and efficacy (response, PFS, OS, time to next treatment) of this triple combination in a real-world setting.

Methods Patients Data from a retrospective cohort of RRMM patients treated

Haematologica | 109 January 2024

246

ARTICLE - An Italian real-world experience of EloPd for RRMM

with EloPd in 35 Italian centers were collected for the purpose of this retrospective analysis. The databases contained clinical information such as age, gender, date of diagnosis, laboratory parameters, treatment history, and date of last follow-up or death extracted from clinical records at the time of inclusion and updated on an ongoing basis. The 35 databases included 200 consecutive patients with RRMM who received at least one cycle of EloPd as salvage treatment between October 2020 and December 2022. All patients were treated with EloPd according to marketing approval as previously described.16,17 Specifically, elotuzumab was given at a dose of 10 mg/kg i.v. on days 1, 8, 15, and 22 during the first two cycles and at a dose of 20 mg/kg once daily on day 1 of each following cycle. The dose of pomalidomide was 4 mg orally once daily on days 1 to 21 of each cycle, whereas that of dexamethasone was 40 mg (or 20 mg in patients aged older than 75 years) once weekly, except on days of elotuzumab administration, when patients received both oral (28 mg [or 8 mg in patients aged older than 75 years]) and intravenous (8 mg) dexamethasone. All patients were given premedication with diphenhydramine (25 to 50 mg) or its equivalent, ranitidine (50 mg) or its equivalent, and acetaminophen (650 to 1,000 mg) or its equivalent 30 to 90 minutes before the elotuzumab infusions. All patients received antibacterial, antiviral, and antithrombotic prophylaxis during treatment. EloPd was administered in 28-day cycles until disease progression, unacceptable toxicity, or withdrawal of consent. The time-to-event endpoints were PFS, OS, and time to next treatment. Safety profile and response were also evaluated for the purposes of the study. Response to treatment and disease progression were evaluated according to the International Myeloma Working Group (IMWG) criteria.18,19 Patients had to reach at least partial remission (PR) in order to be considered to have had a response. The Institutional Ethics Committee of each of the participating hospitals approved the study, which was conducted according to the principles of the Declaration of Helsinki and Good Clinical Practice guidelines. Statistical analysis Statistical comparisons for categorical variables were performed using two-way tables for the Fisher exact test and multi-way tables for the Pearson χ2 test. Multivariable ordinal regression analysis was used to examine the effects of potential confounders on the association between the best response and several variables that were statistically significant on univariable analysis by the Pearson χ2 test or Fisher exact test. PFS, OS and time to next treatment were analyzed using the Kaplan-Meier method. PFS was measured from the initiation of EloPd treatment until death from any cause or progression or last follow-up. OS was measured from the initiation of EloPd treatment until death from any cause or last follow-up The time to next treatment was measured from the initiation of EloPd

M. Gentile et al.

treatment to the earliest date of starting any subsequent therapy or last follow-up. The statistical significance of associations between individual variables and survival was calculated using the log-rank test. The prognostic impact of the outcome variable was investigated by univariable and multiple Cox regression analyses. Results are expressed as hazard ratios (HR) and 95% confidence intervals (95% CI). A P value <0.05 was considered statistically significant. STATA for Windows v.9 and SPSS Statistics v.21 were used to analyze the data.

Table 1. Main characteristics of the patients at baseline. Variable

N of patients (%)

Age <70 years ≥70 years

86 (43) 114 (57)

Sex Male Female

108 (54) 92 (46)

Paraprotein isotype Immunoglobulin G Immunoglobulin A Immunoglobulin D Immunoglobulin M Light chain only

121 (60.5) 44 (22) 3 (1.5) 2 (1) 30 (15)

Creatinine clearance ≥60 mL/min <60 mL/min

132 (66) 68 (34)

ISS stage I II III

61 (30.5) 86 (43) 53 (26.5)

Lactate dehydrogenase Normal Elevated

142 (71) 58 (29)

Previous lines of therapy 2 3 ≥4

101 (50.5) 57 (28.5) 42 (21)

Previous ASCT No Yes

99 (49.5) 101 (50.5)

Previous daratumumab No Yes

54 (27) 146 (73)

Refractory to lenalidomide No Yes

5 (2.5) 195 (97.5)

Disease status Biochemical relapse Symptomatic relapse Refractory to last treatment

30 (15) 94 (47) 76 (38)

FISH analysis available (N= 80) Standard risk High risk

43 (53.8) 37 (46.2)

ISS: International Staging System; ASCT: autologous stem cell transplantation; FISH: fluorescence in situ hybridization.

Haematologica | 109 January 2024

247

ARTICLE - An Italian real-world experience of EloPd for RRMM

Results Patients Overall, 200 RRMM patients treated with EloPd between October 2020 and December 2022 in 35 Italian centers entered this study. Their baseline characteristics are shown in Table 1. At the start of treatment with EloPd, 26.5% of patients had stage III disease according to the International Staging System (ISS), 30.5% were in ISS stage I, and 43% were in ISS stage II. Seventy-six cases (38%) had disease refractory to the previous line of therapy, a symptomatic relapse was observed in 94 patients (47%) and a biochemical relapse in 30 (15%); almost all cases (97.5%) were refractory to lenalidomide. Fifty-one patients (25.5%) showed mild renal impairment, while kidney function was severely compromised in 17 patients (8.5%). Before EloPd treatment, 101 patients (50.5%) had received two lines of therapy, approximately half of the patients (51%) had undergone autologous stem cell transplant (ASCT), while roughly three-quarters (73%) of patients had been exposed to daratumumab. All 146 patients who received daratumumab were refractory to this treatment. One hundred and eleven patients received EloPd immediately after a daratumumab-containing regimen, while 35 patients received other therapy schedules between the daratumumab-containing regimen and treatment with the EloPd regimen. Fluorescence in situ hybridization (FISH) data were available for 80 patients. Forty-three patients (53.8%) had favorable cytogenetic abnormalities, while 37 patients (46.2%) were categorized as being at high risk, because they harbored one of the following aberrations: t(4;14), t(14;16) and del(17p). Response evaluation At the last follow-up, 193 of the 200 enrolled patients were evaluable for response (7 cases had not yet completed the first cycle of therapy). Of these 193 patients, 107 (55.4%) reached at least partial remission (≥PR). More in detail, six (3.1%) achieved a complete remission (CR), 39 (20.2%) a very good partial response (VGPR), and 62 (32.1%) a PR. The median time to response was 1.8 months. The overall response rate (ORR) was statistically higher in patients who did not undergo ASCT (63.3% vs. 47.8%; P=0.032) (Table 2), while a trend towards statistical significance was observed in cases with ISS stage I (stage I=66.1%, stage II=54.9%, and stage III=44.2%; P=0.068), in those treated at biochemical relapse (biochemical relapse=75.9%, symptomatic relapse=52.8%, refractory disease=50.7; P=0.054) and in older patients (>70 years=48.8%, ≤70 years =36.4%; P=0.08) (Table 2). Gender, creatinine clearance, lactate dehydrogenase concentration, number of prior lines of therapy, and previous exposure to daratumumab did not affect the probability of achieving a response to EloPd (Table 2). No differences in ORR were observed between patients receiving EloPd immediately

M. Gentile et al.

after a daratumumab-containing regimen and those who received other schedules of therapy between a daratumumab-containing regimen and EloPd (ORR: 55% vs. 43%, respectively; P=0.42). Progression-free survival After a median follow-up of 9 months (range, 1-26), 121 patients (60.5%) out of 200 had experienced disease progression or died. The total number of deaths was 79 (39.5%). The median PFS was 7 months (95% CI: 5.8-8.2 months), and the 1-year probability of PFS was 33.6% (Figure 1A). Univariable analyses showed that ISS stage II (HR=1.61, 95% CI: 1.03-2.54; P=0.039), ISS stage III (HR=2.9, 95% CI: 1.774.75; P<0.0001), previous ASCT (HR=1.43, 95% CI: 1.05-2.05; P=0.05), previous daratumumab exposure (HR=1.72, 95% CI: 1.14-2.59; P=0.01) (Online Supplementary Figure S1A), symptomatic relapse (HR=2.02, 95% CI: 1.13-3.63; P=0.018) and refractory disease at the time of starting EloPd treatment (HR=1.86, 95% CI: 1.02-3.37; P=0.041) were associated with

Table 2. Association between overall response rate and main clinical-hematologic characteristics of multiple myeloma patients treated with elotuzumab plus pomaidomide and dexamethasone (N=193). ≥PR N (%)

<PR N (%)

Age ≥70 years >70 years

39 (36.4) 42 (48.8)

68 (63.6) 44 (51.2)

0.08

Sex Female Male

50 (56.8) 57 (54.3)

38 (43.2) 48 (45.7)

0.72

Creatinine clearance ≥60 mL/min <60 mL/min

67 (62.6) 40 (37.4)

59 (68.6) 27 (31.4)

0.38

ISS stage I II III

39 (66.1) 45 (54.9) 23 (44.2)

20 (33.9) 37 (45.1) 29 (55.8)

Lactate dehydrogenase Normal Elevated

72 (52.6) 35 (62.5)

65 (47.4) 21 (37.5)

0.2

Previous lines of therapy 2 >2

59 (60.2) 48 (50.5)

36 (36.7) 48 (52.2)

0.032

Previous ASCT No Yes

62 (63.3) 44 (47.8)

36 (36.7) 48 (52.2)

0.032

Previous daratumumab No Yes

31 (60.8) 76 (53.5)

20 (39.2) 66 (46.5)

0.37

Disease status Biochemical relapse Symptomatic relapse Refractory to last treatment

22 (75.9) 47 (52.8) 38 (50.7)

7 (24.1) 42 (47.2) 37 (49.3)

Variable

0.068

0.054

PR: partial response; ISS: International Staging System; ASCT: auto-logous stem cell transplantation.

Haematologica | 109 January 2024

248

ARTICLE - An Italian real-world experience of EloPd for RRMM

a significantly lower PFS rate (Table 3). No differences in PFS were observed between patients receiving EloPd immediately after a daratumumab-containing regimen and those who received other schedules of therapy between a daratumumab-containing regimen and EloPd (HR=1.34, 95% CI: 0.83-2.16; P=0.23). Notably, in the Cox multivariable analysis, only advanced ISS stage (III) maintained an independent prognostic impact on PFS (HR=2.55, 95% CI: 1.54-4.24; P<0.0001) (Table 3). Conversely, ISS stage II (HR=1.53, 95% CI: 0.97-2.44; P=0.69), previous ASCT (HR=1.35, 95% CI: 0.93-1.96; P=0.31), previous daratumumab exposure (HR=1.35, 95% CI: 0.9-2.08; P=0.17), symptomatic relapse (HR=1.69, 95% CI: 0.94-3.05; P=0.008) and refractory disease at the time of starting EloPd treatment (HR=1.49, 95% CI: 0.81-2.73; P=0.2) lost their independent predictive value on PFS.

M. Gentile et al.

ceiving EloPd immediately after a daratumumab-containing regimen and those who received other schedules of therapy between a daratumumab-containing regimen and EloPd (HR=1.19, 95% CI: 0.7-2.01; P=0.53). Notably, in the Cox multivariable analysis, advanced ISS stage (i.e., stage III) (HR=1.87, 95% CI: 1.16-3.02; P=0.01), symptomatic relapse (HR=2.5, 95% CI: 1.06-6.0; P=0.04) and refractory disease at the start of EloPd treatment (HR=2.4, 95% CI: 1.04-5.5; P=0.05) maintained an independent prognostic impact on the survival outcome (Table 4). In contrast, the effect of previous daratumumab treatment lost its independent prognostic significance on OS (HR=1.68, 95% CI: 0.98-2.88; P=0.06).

Overall survival The median OS was 17.5 months (95% CI: 28-40.2), and the 1-year probability of OS was 57.9% (95% CI: 12-23.2 months) (Figure 1B). Univariable analyses showed that ISS stage III (HR=2.46, 95% CI: 1.35-4.48; P=0.003), previous daratumumab exposure (HR=1.87, 95% CI: 1.1-3.19; P=0.02) (Online Supplementary Figure S1B), symptomatic relapse (HR=2.83, 95% CI: 1.19-6.7; P=0.018) and refractory disease (HR=2.56, 95% CI: 1.07-6.12; P=0.034) at the start of EloPd treatment were associated with a significantly shorter OS (Table 4). No differences in OS were observed between patients re-

Time to next treatment and subsequent therapy After discontinuation of EloPd therapy, 71 patients (35.5%) received subsequent treatment. The median time to the next treatment was 8.1 months (95% CI: 6.7-9.4), with a 1-year re-treating probability of 37.5% (Figure 1C). The type of subsequent treatment is shown in Table 5. Overall, 20 different salvage therapy regimens were used after discontinuation or failure of EloPd. Roughly one-third of patients (24 cases) received belantamab alone (23 cases) or in combination with isatuximab (1 patient), 24 patients (33.8%) were given a proteasome inhibitor-containing regimen (carfilzomib-based in 14 cases, bortezomib-based in 6 cases and ixazomib-based in 4 cases), while 13 patients (18.3%) received an anti-CD38-containing regimen (dara-

Figure 1. Kaplan-Meier curves for all 200 patients with relapsed or refractory multiple myeloma treated with the elotuzumab, pomalidomide and dexamethasone triple regimen. (A) Kaplan-Meier curve of progression-free survival. (B) Kaplan-Meier curve of overall survival. (C) Kaplan-Meier curve of time to next treatment.

Haematologica | 109 January 2024

249

ARTICLE - An Italian real-world experience of EloPd for RRMM

M. Gentile et al.

tumumab-based in 10 cases and isatuximab-based in 3 cases). Finally, ten patients (14.1%) received a subsequent chemotherapeutic regimen (6 cases were treated with a cyclophosphamide-based regimen, 3 with bendamustine, and 1 with melphalan).

(9.5%), and thrombocytopenia (9.5%), while infection rates and pneumonia were roughly 14% and 6.5%, respectively. The rate of adverse events was not significantly different between patients aged less or more than 70 years (data not shown).

Safety At the last database update, the median number of EloPd courses administered was five (range, 1-20). EloPd treatment had been withdrawn from a total of 126 (62.5%) patients by the cutoff date, mainly due to disease progression (112 cases). Of the remaining cases, nine patients discontinued therapy because of toxicity (6 infections and 3 cases of pomalidomide-related severe skin rash) and five patients died of causes unrelated to the therapy. Infusion reactions occurred at first administration of elotuzumab in 11 patients (5.5%, all grade 1 or 2) and were promptly resolved in all patients (no discontinuation of treatment reported). Major adverse events are presented in Table 6 and include grade 3 or 4 neutropenia (21.5%), anemia (11%), lymphocytopenia

Outcome analysis by cytogenetic risk Data on cytogenetic abnormalities were available for only 40% of cases (80/200). However, the analytical weight of this biomarker for prognosis, also emphasized by the Revised ISS (R-ISS),20 prompted us to carry out an ancillary analysis, conscious that the relatively low incidence of accessible cases could bias the statistical accuracy. When comparing the main characteristics of the patients in each group, the patients for whom cytogenetic information was available differed from the remaining cases only for a smaller proportion of patients with creatinine clearance <60 mL/min (Online Supplementary Table S1). No difference in ORR was observed between the high-risk and the standard-risk groups (54.1 vs. 53.7%, respectively;

Table 3. Univariable and multivariable analyses of progression-free survival. Univariable analysis

Variable

PFS at 12 months %

HR (95% CI)

Age ≤70 years >70 years

86 114

27.9 39.2

0.72 (0.5-1.03)

0.75

Gender Male Female

108 92

34.7 34.2

1.05 (0.74-1.49)

0.79

Creatinine clearance ≥60 mL/min <60 mL/min

132 68

31.2 41.3

0.9 (0.62-1.32)

0.6

ISS stage I II III

61 86 53

50.4 30.2 24.2

1.61 (1.03-2.54) 2.9 (1.77-4.75)

0.039 <0.0001

Lactate dehydrogenase Normal Elevated

142 58

33.7 33.1

0.99 (0.67-1.45)

0.95

Previous lines of therapy 2 >2

101 99

39.6 28.6

1.33 (0.93-1.9)

0.11

Previous ASCT No Yes

99 101

41.1 25.4

1.43 (1.05-2.05)

Previous daratumumab No Yes

54 146

46.9 29.2

Disease status Biochemical relapse Symptomatic relapse Refractory to last treatment

30 94 76

57.3 29.1 31.2

Multivariable analysis HR (95% CI)

1.53 (0.97-2.44) 2.55 (1.54-4.24)

0.69 <0.0001

0.05

1.35 (0.93-1.96)

0.31

1.72 (1.14-2.59)

0.01

1.35 (0.9-2.08)

0.17

2.02 (1.13-3.63) 1.86 (1.02-3.37)

0.018 0.041

1.69 (0.94-3.05) 1.49 (0.81-2.73)

0.08 0.2

PFS: progression-free survival; HR: hazard ratio; 95% CI: 95% confidence interval; ISS: International Staging System; ASCT: autologous stem cell transplantation.

Haematologica | 109 January 2024

250

ARTICLE - An Italian real-world experience of EloPd for RRMM

P=0.97). The two subgroups showed a non-statistically different PFS (1-year PFS; high-risk group vs. standard-risk: 28.4% vs. 44.7%, respectively; HR=1.34, 95% CI: 0.77-2.34; P=0.29) (Online Supplementary Figure S2A), while a trend towards statistical significance in terms of OS was observed in standard-risk patients (1-year OS; high-risk group vs. standard-risk: 50.1 vs. 75.1%; HR=2, 95% CI: 0.94-4.29; P=0.07) (Online Supplementary Figure S2B).

Discussion Elotuzumab, as monotherapy, was first evaluated in a phase 1I dose-finding study, which demonstrated the safety and tolerability of the drug at either 10 mg/kg or 20 mg/kg, but, at the same time, the absence of response, especially in the setting of heavily pre-treated patients.21 Given the enhanced antimyeloma activity in combination with other drugs within preclinical studies, elotuzumab was tested

M. Gentile et al.

in association with lenalidomide in a phase II study which showed better efficacy of the triplet regimen in the setting of relapsed-refractory patients.22 Those results were subsequently confirmed by the phase III ELOQUENT-2 trial23 and remain robust at a follow-up of 70 months.24 Recently, data from the ELOQUENT-3 trial showed that the addition of elotuzumab to pomalidomide and dexamethasone provided a significant clinical improvement, in terms of PFS and OS, over pomalidomide and dexamethasone, with a manageable toxicity profile in the treatment of RRMM patients who had received at least two prior therapies, including lenalidomide and a proteasome inhibitor.16,17 Furthermore, the addition of elotuzumab to pomalidomide and dexamethasone did not have a negative impact on health-related quality of life of MM patients.25 Based on the results of these trials, the FDA approved EloPd for this setting of MM patients. Here we have described an Italian real-world experience of the use of EloPd. To the best of our knowledge, our survey is the first real-world EloPd series. Real-world profiles

Table 4. Univariable and multivariable analyses of overall survival.

Variable

Univariable analysis

OS at 12 months %

HR (95% CI)

Multivariable analysis HR (95% CI)

1.87 (1.16-3.02)

0.01

Age ≤70 years >70 years

86 114

51.8 62.4

0.77 (0.49-1.21)

0.26

Gender Male Female

108 92

61.1 54.4

1.17 (0.75-1.83)

0.48

Creatinine clearance ≥60 mL/min <60 mL/min

132 68

59.1 55.1

1.02 (0.64-1.64)

0.91

ISS stage I II III

61 86 53

72.9 56.9 42.9

1.24 (0.7-2.22) 2.46 (1.35-4.48)

0.46 0.003

Lactate dehydrogenase Normal Elevated

142 58

59.7 53.4

1.31 (0.82-2.1)

0.26

Previous lines of therapy 2 >2

101 99

64.5 50.8

1.43 (0.91-2.22)

0.12

Previous ASCT No Yes

99 101

59.3 56.1

1.25 (0.8-1.96)

0.33

Previous daratumumab No Yes

54 146

69.7 50.6

1.87 (1.1-3.19)

0.02

1.68 (0.98-2.88)

0.06

Disease status Biochemical relapse Symptomatic relapse Refractory to last treatment

30 94 76

81 55.1 51.8

2.83 (1.19-6.7) 2.56 (1.07-6.12)

0.018 0.034

2.5 (1.06-6.0) 2.4 (1.04-5.5)

0.04 0.05

OS: overall survival; HR: hazard ratio; 95% CI: 95% confidence interval; ISS: International Staging System; ASCT: autologous stem cell transplantation.

Haematologica | 109 January 2024

251

ARTICLE - An Italian real-world experience of EloPd for RRMM

are rarely fully represented in randomized clinical trials, a factor that can complicate treatment decision-making. In this regard, aging is a critical problem in MM patients’ management because of its association with frailty, increased comorbidities, poor tolerability of treatment, and higher risk of complications.26 In our series, approximately onethird of patients were ≥75 years old and 8.5% had severe renal impairment (creatinine clearance <30 mL/min). In comparison, in the registration trial, 21.7% of the patients were elderly and a creatinine clearance <45 mL/min was an exclusion criterion. The ELOQUENT-3 trial,16 and its update,17 showed the safety of the triplet EloPd drug regimen. Although some caution should be exercised because of the retrospective nature of the present study, similar adverse event profiles were documented in our real-world cohort, except for a slightly higher incidence of neutropenia, possibly due to the aforementioned differences in age and prevalence of severe renal impairment. Nevertheless, the incidence of infections was comparable.16 Of note, no significant differences were documented in the incidences of adverse events between younger (<70 years) and older patients. The ORR in our real-world cohort was comparable to that in the ELOQUENT-3 trial (55.4% vs. 53%), with a similar number of patients reaching good quality responses,16 although the different clinical features of patients included in the two series should be taken into consideration (e.g., the median number of previous lines of therapies was 3 in the ELOQUENT-3 trial and 2 in our retrospective series) (Table 7). Interestingly, the only patients who showed a significantly lower response rate were those who had previously undergone ASCT. Conversely, there was a trend to a statistically significant higher response rate in patients with a low ISS stage and in those treated in biochemical relapse. These findings should be taken into consideration when choosing EloPd treatment. The median time to achieve the best response was similar in our study and in the ELOQUENT-3 trial,16 being 1.8 months and 2 months, respectively. PFS predictors should also be considered to reduce the chance of progression. In our series, the estimated median PFS was 7 months, shorter than the 10.3 months observed in the ELOQUENT-3 trial.16 This relatively poorer clinical outcome is possibly due to differences in baseline characteristics of patients in our real-world cohort and those in clinical trials (Table 7). Specifically, our cohort included a higher proportion of patients with advanced ISS stage (stage III) (26.5% vs. 11.7%) and a not negligible rate of patients with high-risk cytogenetics (46.2% vs. 10%) (Table 7), both categories having a poor prognosis. A multivariable model revealed that only ISS stage III was an independent predictor of shorter PFS. In our series, the median OS was shorter than that observed in the ELOQUENT-3 trial (17.5 vs. 29.8 months).17 Nevertheless, OS results should be considered somewhat

M. Gentile et al.

immature because of the relatively short follow-up. The differences in baseline characteristics between patients in our real-world cohort and those in the clinical trial could also have a negative impact on survival (Table 7). Again, at multivariable analysis, advanced ISS stage (stage

Table 5. Salvage therapy regimens after the elotuzumab, pomalidome and dexamethasone triple regimen. Salvage therapy regimen

N of cases (%)

Antibody drug-coniugates Belantamab Belantamab-isatuximab

24 (33.8) 23 (32.4) 1 (1.4)

Anti-CD38-containing regimens DVd Daratumumab DRd IsaKd

13 (18.3)* 5 (7) 3 (4.2) 2 (2.8) 3 (4.2)

PI-containing regimens Kd KRd K-Ctx-d Vd Vd-venetoclax Vd-PACE Vd-eftozanermin VMP Ixa-Rd Ixa-Ctx

24 (33.8) 12 (16.9) 2 (2.8) 1 (1.4) 1 (1.4) 1 (1.4) 1 (1.4) 1 (1.4) 1 (1.4) 2 (2.8) 2 (2.8)

Other therapies Bendamustine Ctx Ctx+Caelyx+d Melphalan

10 (14.1) 3 (4.2) 5 (7) 1 (1.4) 1 (1.4)

Considering the patient treated with belantamab-isatuximab: 14 (19.7%). DVd: daratumumab, bortezomib, dexamethasone; DRd: daratumumab, lenalidomide, dexamethasone; IsaKd: isatuximab, carfilzomib, dexamethasone; PI: proteasome inhibitor; Kd: carfilzomib, dexamethasone; KRd: carfilzomib, lenalidomide, dexamethasone: K-Ctx-d: carfilzomib cyclophosphamide, dexamethasone; Vd: bortezomib, dexamethasone; Vd-PACE: bortezomib, dexamethasone, cisplatin, doxorubicin, cyclophosphamide, etoposide; VMP: bortezomib, melphalan, prednisone; IxaRd: ixazomib, lenalidomide, dexamethasone; Ixa-Ctx: ixazomib cyclophosphamide; Ctx: cyclophosphamide; Ctx+Caelyx+d: cyclophosphamide, caelyx, dexamethasone.

Table 6. Incidence of serious adverse events among the patients treated with the elotuzumab, pomalidomide and dexamethasone triple regimen (N=200). Grade 3/4 adverse events Hematologic toxicities Lymphocytopenia Anemia Thrombocytopenia Neutropenia

19 (9.5) 22 (11) 19 (9.5) 43 (21.5)

Non-hematologic toxicities Infections Pneumonia Gastrointestinal toxicity

28 (14) 13 (6.5) 8 (4)

Haematologica | 109 January 2024

252

N of cases (%)

ARTICLE - An Italian real-world experience of EloPd for RRMM

M. Gentile et al.

Table 7. Comparison of the characteristics at baseline between the cohort of patients treated with the elotuzumab, pomalidomide and dexamethasone triple regimen in the real-world setting and those enrolled in the ELOQUENT-3 clinical trial. Variable

Real-world study

ELOQUENT-3 trial

Age ≥75 years, %

33.5

21.7

Creatinine clearance, % <45 mL/min <30 mL/min

20 8.5

0 0

ISS stage III, %

26.5

11.7

Elevated LDH, %

23.3

2 (2-9)

3 (2-8)

Previous ASCT, %

51.7

Prior daratumumab exposure, %

Refractory to lenalidomide, %

97.5

High-risk FISH findings, %

46.2

Previous lines of therapy, median (range)

ISS: International Staging System; LDH: lactate dehydrogenase; ASCT: autologous stem cell transplantation; FISH: fluorescence in situ hybridization.

III) showed an independent prognostic impact on the OS together with disease status at the start of EloPd therapy. In our cohort, PFS and OS were similar in both age groups (i.e., <70 years and ≥70 years), and EloPd showed a good safety profile even when used in the elderly (57% of our EloPd cohort), whose treatment is challenging because such patients are often frail, and have increased comorbidities, poor tolerability, and a higher risk of complications.26 There are two key reasons why the information on patients exposed to daratumumab is interesting. First of all, the data are lacking in the ELOQUENT-3 trial. Secondly, daratumumab-based therapy is currently the standard of care for most MM patients, both in the first- and in the second-line, enabling the evaluation of the impact of previous daratumumab treatment on the efficacy of EloPd in the real-world setting. In fact, in our experience, prior daratumumab treatment did not affect either the probability of achieving a response or outcome indicators in RRMM patients treated with EloPd. The IMWG consensus recommends using ISS stage and cytogenetic abnormalities to analyze OS risk stratification.27 Unfortunately, cytogenetic analysis is rarely performed in a real-world setting. Although we were conscious that the relatively low number of accessible cases (approximately 40%) might lead to incorrect statistical interpretations, the prognostic importance of FISH information, highlighted by the R-ISS,20 motivated us to conduct an additional investigation. In this respect, high-risk patients, defined as those with poor cytogenetics (t[4;14], t[14;16], or del[17p]), did not show a significantly shorter PFS or OS, although the low number of cases did not allow assessment of the independent prognostic value of this parameter in multivariable analysis. Among the study’s strengths, we highlight that the number of patients enrolled in our real-world study is more than three times greater than that of the cohort of patients enrolled in the EloPd arm (n=60) of the ELOQUENT-3 trial.

Furthermore, taking into account the growing number of patients receiving anti-CD38 monoclonal antibody in the early phase of treatment, data on the efficacy of EloPd in patients previously exposed to daratumumab represent an additional value coming from our retrospective observation, since this information has not yet been provided by a randomized clinical trial. Conversely, among the weaknesses, the follow-up time is relatively short to draw definitive conclusions about OS and the well-known biases associated with the retrospective nature of the study must be mentioned. In conclusion, our real-world data confirm the results obtained in the ELOQUENT-3 controlled clinical trial.16,17 EloPd is a safe and possible therapeutic choice for RRMM patients who have received at least two prior therapies, including lenalomide and a proteasome inhibitor. Notably, prior treatment with daratumumab did not have a negative impact of the efficacy of the EloPd triplet regimen. Several clinical trials are currently exploring the efficacy of elotuzumab in association with other antimyeloma drugs, such as iberdomide (CC-220),28 isatuximab29 and belantamab,30 in the setting of RRMM patients. Disclosures RZ has served on advisory boards for Bristol, Janssen, Sanofi, Takeda, Amgen, and GSK. SM has received honoraria from Bristol-Myers Squibb, Sanofi, AMGEN, GSK Takeda, and Janssen; and has served on advisory boards for Sanofi, Takeda, Bristol-Myers Squibb and Janssen. AG has received honoraria from Janssen, Amgen, Pfizer, and Takeda. MTP has received honoraria from Bristol-Myers Squibb, Sanofi, AMGEN, GSK, Takeda, and Janssen; has served on advisory boards for Celgene-Bristol-Myers Squibb, Sanofi, AMGEN, Sanofi, GSK, Takeda, Roche, Karyopharm, and Janssen; and has received support for attending meetings and/or travel from Janssen, Celgene-Bristol-Myers Squibb, AMGEN, Sanofi, and Takeda.

Haematologica | 109 January 2024

253

ARTICLE - An Italian real-world experience of EloPd for RRMM

Contributions: MG, EV, FDR, AN, FM, and PM designed the study. MG and FM performed the statistical analysis. SP, MGa, DD, RM, RDP, RZ, EAM, AB, SM, EZ, CCa, MM, CCo, CCe. GM, NDR, MO, GT, GMC, AR, RR, GU, GB, GP, AP, DV, MB, FA, VA, AA, RF, VB, BR, EC, AG, RRi, NS, EF, GB, DN, and MTP analyzed and interpreted data. MG, EV, FDR, AN, FM, and PM wrote the manuscript. All authors gave final approval.

M. Gentile et al.

Funding This study was partially supported by the Italian Ministry of Health- Ricerca Corrente annual program 2024. Data-sharing statement The data that support the findings of this study are available from the corresponding author upon reasonable request.

References 1. Kumar SK, Dispenzieri A, Lacy MQ, et al. Continued improvement in survival in multiple myeloma: changes in early mortality and outcomes in older patients. Leukemia. 2014;28(5):1122-1128. 2. Kumar SK, Rajkumar SV, Dispenzieri A, et al. Improved survival in multiple myeloma and the impact of novel therapies. Blood. 2008;111(5):2516-2520. 3. Nooka AK, Kaufman JL, Hofmeister CC, et al. Daratumumab in multiple myeloma. Cancer. 2019;125(14):2364-2382. 4. Bruzzese A, Martino EA, Vigna E, et al. Elotuzumab in multiple myeloma. Expert Opin Biol Ther. 2023;23(1):7-10. 5. Durer C, Durer S, Lee S, et al. Treatment of relapsed multiple myeloma: evidence-based recommendations. Blood Rev. 2020;39:100616. 6. Collins SM, Bakan CE, Swartzel GD, et al. Elotuzumab directly enhances NK cell cytotoxicity against myeloma via CS1 ligation: evidence for augmented NK cell function complementing ADCC. Cancer Immunol Immunother. 2013;62(12):1841-1849. 7. Hsi ED, Steinle R, Balasa B, et al. CS1, a potential new therapeutic antibody target for the treatment of multiple myeloma. Clin Cancer Res. 2008;14(9):2775-2784. 8. Balasa B, Yun R, Belmar NA, et al. Elotuzumab enhances natural killer cell activation and myeloma cell killing through interleukin-2 and TNF-α pathways. Cancer Immunol Immunother. 2015;64(1):61-73. 9. European Medicines. Agency Empliciti, INN-elotuzumab. https:// www.ema.europa.eu/en/documents/product-information/emplicitiepar-product-information_it.pdf Accessed on 6 June, 2023. 10. Gentile M, Specchia G, Derudas D, et al. Elotuzumab, lenalidomide, and dexamethasone as salvage therapy for patients with multiple myeloma: Italian, multicenter, retrospective clinical experience with 300 cases outside of controlled clinical trials. Haematologica. 2021;106(1):291-294. 11. Bruzzese A, Derudas D, Galli M, et al. Elotuzumab plus lenalidomide and dexamethasone in relapsed/refractory multiple myeloma: extended 3-year follow-up of a multicenter, retrospective clinical experience with 319 cases outside of controlled clinical trials. Hematol Oncol. 2022;40(4):704-715. 12. Morabito F, Zamagni E, Conticello C, et al. Adjusted comparison between elotuzumab and carfilzomib in combination with lenalidomide and dexamethasone as salvage therapy for multiple myeloma patients. Eur J Haematol. 2022;108(3):178-189. 13. Morabito F, Zamagni E, Conticello C, et al. Survival risk scores for real-life relapsed/refractory multiple myeloma patients receiving elotuzumab or carfilzomib in combination with lenalidomide and dexamethasone as salvage therapy: analysis of 919 cases outside clinical trials. Front Oncol. 2022;12:890376. 14. Uccello G, Petrungaro A, Mazzone C, et al. Pomalidomide in

multiple myeloma. Expert Opin Pharmacother. 2017;18(2):133-137. 15. Lopez-Girona A, Mendy D, Ito T, et al. Cereblon is a direct protein target for immunomodulatory and antiproliferative activities of lenalidomide and pomalidomide. Leukemia. 2012; 26(11):2326-2335. 16. Dimopoulos MA, Dytfeld D, Grosicki S, et al. Elotuzumab plus pomalidomide and dexamethasone for multiple myeloma. N Engl J Med 2018;379(19):1811-1822. 17. Dimopoulos MA, Dytfeld D, Grosicki S, et al. Elotuzumab plus pomalidomide and dexamethasone for relapsed/refractory multiple myeloma: final overall survival analysis from the randomized phase II ELOQUENT-3 trial. J Clin Oncol. 2023;41(3):568-578. 18. Durie BG, Harousseau JL, Miguel JS, et al. International uniform response criteria for multiple myeloma. Leukemia. 2006;20(9):1467-1473. 19. Rajkumar SV, Harousseau JL, Durie B, et al. Consensus recommendations for the uniform reporting of clinical trials: report of the International Myeloma Workshop Consensus Panel 1. Blood. 2011;117(18):4691-4695. 20. Palumbo A, Avet-Loiseau H, Oliva S, et al. Revised International Staging System for multiple meloma: a report from International Myeloma Working Group. J Clin Oncol. 2015;33(26):2863-2869. 21. Zonder JA, Mohrbacher AF, Singhal S, et al. A phase 1, multicenter, open-label, dose esclation study of elotuzumab in patients with advanced multiple myeloma. Blood. 2012; 120(3):552-559. 22. Richardson PG, Jagannath S, Moreau P, et al.; 1703 study investigators. Elotuzumab in combination with lenalidomide and dexamethasone in patients with relapsed multiple myeloma: final phase 2 results from the randomised, openlabel, phase 1b-2 dose-escalation study. Lancet Haematol. 2015;2(12):e516-527. 23. Lonial S, Dimopoulos M, Palumbo A, et al; ELOQUENT-2 investigators. Elotuzumab therapy for relapsed or refractory multiple myeloma. N Engl J Med. 2015;373(7):621-631. 24. Dimopoulos MA, Lonial S, White D, et al. Elotuzumab, lenalidomide, and dexamethasone in RRMM: final overall survival results from the phase 3 randomized ELOQUENT-2 study. Blood Cancer J. 2020;10(9):91. 25. Weisel K, Dimopoulos MA, San-Miguel J, et al. Impact of elotuzumab plus pomalidomide/dexamethasone on healthrelated quality of life for patients with relapsed/refractory multiple myeloma: final data from the phase 2 ELOQUENT-3 trial. Hemasphere. 2023;7(3):e843. 26. Diamond E, Lahoud OB, Landau H. Managing multiple myeloma in elderly patients. Leuk Lymphoma. 2018;59(6):1300-1311.

Haematologica | 109 January 2024

254

ARTICLE - An Italian real-world experience of EloPd for RRMM

27. Chng WJ, Dispenzieri A, Chim CS, et al. IMWG consensus on risk stratification in multiple myeloma. Leukemia. 2014;28(2):269-277. 28. A study of iberdomide (CC-220) in combination with elotuzumab and dexamethasone for relapsed/refractory multiple myeloma (CC-220). ClinicalTrials.gov identifier: NCT 05560399. 29. Isatuximab, pomalidomide, elotuzumab and dexamethasone in

relapsed and/or refractory multiple myeloma (IMPEDE). ClinicalTrials.gov identifier: NCT 04835129. 30. Novel combination of belantamab mafodotin and elotuzumab to enhance therapeutic efficacy in multiple myeloma. ClinicalTrials.gov identifier: NCT 05002816.

Haematologica | 109 January 2024

255

M. Gentile et al.

ARTICLE - Plasma Cell Disorders

S-adenosylmethionine biosynthesis is a targetable metabolic vulnerability in multiple myeloma Yanmeng Wang,1 Catharina Muylaert,1 Arne Wyns,1 Philip Vlummens,1,2 Kim De Veirman,1 Karin Vanderkerken,1 Esther Zaal,3 Celia Berkers,3 Jérome Moreaux,4,5,6 Elke De Bruyne1 and Eline Menu1 Department of Hematology and Immunology-Myeloma Center Brussels, Vrije Universiteit Brussel, Jette, Belgium; 2Department of Clinical Hematology, Ghent University Hospital, Ghent, Belgium; 3Utrecht Metabolism Expertise Center, Nieuw Gildestein, Utrecht, the Netherlands; 4Laboratory for Monitoring Innovative Therapies, Department of Biological Hematology, CHU Montpellier, Montpellier, France; 5Institute of Human Genetics, University of Montpellier, Montpellier, France and 6Institut Universitaire de France, Paris, France 1

Correspondence: E.D. Bruyne Elke.De.Bruyne@vub.be E. Menu Eline.Menu@vub.be Received: Accepted: Early view:

February 1, 2023. July 11, 2023 July 20, 2023.

Abstract Multiple myeloma (MM) is the second most prevalent hematologic malignancy and is incurable because of the inevitable development of drug resistance. Methionine adenosyltransferase 2α (MAT2A) is the primary producer of the methyl donor S-adenosylmethionine (SAM) and several studies have documented MAT2A deregulation in different solid cancers. As the role of MAT2A in MM has not been investigated yet, the aim of this study was to clarify the potential role and underlying molecular mechanisms of MAT2A in MM, exploring new therapeutic options to overcome drug resistance. By analyzing publicly available gene expression profiling data, MAT2A was found to be more highly expressed in patient-derived myeloma cells than in normal bone marrow plasma cells. The expression of MAT2A correlated with an unfavorable prognosis in relapsed patients. MAT2A inhibition in MM cells led to a reduction in intracellular SAM levels, which resulted in impaired cell viability and proliferation, and induction of apoptosis. Further mechanistic investigation demonstrated that MAT2A inhibition inactivated the mTOR-4EBP1 pathway, accompanied by a decrease in protein synthesis. MAT2A targeting in vivo with the small molecule compound FIDAS-5 was able to significantly reduce tumor burden in the 5TGM1 model. Finally, we found that MAT2A inhibition can synergistically enhance the anti-MM effect of the standard-of-care agent bortezomib on both MM cell lines and primary human CD138+ MM cells. In summary, we demonstrate that MAT2A inhibition reduces MM cell proliferation and survival by inhibiting mTOR-mediated protein synthesis. Moreover, our findings suggest that the MAT2A inhibitor FIDAS-5 could be a novel compound to improve bortezomib-based treatment of MM.

Introduction Multiple myeloma (MM) is a hematologic malignancy characterized by the accumulation of malignant plasma cells in the bone marrow. MM is the second most frequent hematologic disorder and accounts for 10% of all hematologic malignancies.1,2 Several drugs have already been approved to treat MM, including proteasome inhibitors, immunomodulatory drugs, steroids, histone deacetylase inhibitors and monoclonal antibodies.3 Bortezomib was the first proteasome inhibitor approved by the US Food and Drug Administration for the treatment of MM and is currently one of the standard-of-care agents for first-line treatment.4 However, despite the significant progress in treating MM patients, one of the major issues is the recur-

rence of the disease due to an incomplete eradication of myeloma cells.5,6 Thus, therapies capable of inducing durable elimination of MM cells are still needed. Recent studies provide evidence that dysregulation of cellular metabolism plays a central role in the pathogenesis of MM, including MM cell survival, growth, and drug resistance.7-9 Methionine adenosyltransferase (MAT) is a key regulator of cellular metabolism and catalyzes the reaction of L-methionine and adenosine triphosphate (ATP) to S-adenosylmethionine (SAM), which is an essential methyl donor.10 MAT includes MAT1A and MAT2A in mammals, which code for two different enzymes, MATI/III and MATII respectively. MAT1A is liver-specific, whereas MAT2A shows a wide distribution and is responsible for SAM synthesis in extrahepatic tissues.11 MAT2A is dysregulated in

Haematologica | 109 January 2024

256

ARTICLE - Synthesis of SAM is a target in multiple myeloma several cancer types, such as hepatic cancer, breast cancer, and colon cancer and it has been shown that silencing of MAT2A results in reduced cancer cell proliferation and cell death.12-14 However, so far, the role of MAT2A in MM has not been explored. A high rate of synthesis of immunoglobulins is a key feature of MM cells. Accordingly, MM cells are characterized by an expanded endoplasmic reticulum and partial activation of unfolded protein response pathways that contribute to the cells’ excessive protein production.15,16 The kinase mammalian target of rapamycin (mTOR) is the core component of two protein complexes, mTOR complex 1 (mTORC1) and mTOR complex 2, and is a member of the family of phosphoinositide 3-kinase (PI3K)-related kinases.17 The mTOR signaling pathway, which is regularly activated in tumors, not only plays an important role in tumor metabolism, but also regulates gene transcription and protein synthesis to regulate cell proliferation.18-20 The inhibitory eukaryotic translation initiation factor 4E binding proteins 1 and 2 (4EBP1 and 2) are phosphorylated by mTORC1, resulting in the release of the eukaryotic translation initiation factor 4E (eIF4E). By assisting in the formation of the translation initiation complex at the 5' ends of mRNA, the released eIF4E then enhances global translation.21,22 Furthermore, mTORC1 phosphorylates the ribosomal proteins p70 S6 kinase (p70S6K) and S6, which are necessary for the positive regulation of the eukaryotic initiation factors 4A and 4B to augment translation.23 MAT2A has been shown to impact different (patho)physiological processes (including protein synthesis) depending on the cell type.24 Whether this is the case for MM cells remains to be determined. In this study, we aimed to uncover the role of MAT2A in MM survival and response to drugs. We investigated the relationship between MAT2A expression and survival using publicly available gene expression profiling data. By using MAT2A-specific short interfering (si)RNA, we then studied the impact of MAT2A targeting on MM cell viability, proliferation, apoptosis, mTOR signaling and protein synthesis. To validate the therapeutic potential of targeting MAT2A further, we used the small molecule inhibitor FIDAS-5 to target MAT2A both in vitro and in vivo. Finally, we investigated whether the MAT2A inhibitor could increase the efficacy of bortezomib.

Methods Cells and cell culture Seven human myeloma cell lines (HMCL), ANBL6, AMO-1, JJN3, LP-1, OPM2, RPMI8226, and U266, and the human bone marrow stromal cell line HS-5 were obtained from the American Type Culture Collection. Two other HMCL, XG-2 and XG-7, were kindly provided by J. Moreaux (University of Montpellier, Montpellier, France). The HMCL and murine MM

Y. Wang et al.

cell line 5TGM1 were cultured in RPMI1640 medium (Thermo Fisher Scientific, Waltham, MA, USA) and HS-5 was cultured in DMEM (Thermo Fisher Scientific). All the cell cultures were supplemented with 100 U/mL penicillin/streptomycin, 2 mmol/L glutamine (Thermo Fisher Scientific), and 10% fetal calf serum (Hycone, Logan, UT, USA). Cells were cultured at 37°C in a humidified atmosphere of 5% CO2. All cell lines were authenticated by short-tandem repeat profiling and regularly tested to rule out contamination by mycoplasma. Compounds Bortezomib was purchased from Selleckchem (Munich, Germany), FIDAS-5 was purchased from Millipore Sigma (St. Louis, MO, USA) and MAT2A inhibitor 1 (Compound 196) was purchased from MedChemExpress (Monmouth Junction, NJ, USA). For in vitro studies, all compounds were dissolved in dimethylsulfoxide according to the manufacturer’s instructions. For the murine experiments, FIDAS-5 was dissolved in polyethylene glycol 400 (Sigma-Aldrich, St Louis, MO, USA). Murine experiments C57BL/KalwRij mice were purchased from Envigo Laboratories (Horst, the Netherlands). They were housed and treated following conditions approved by the Ethical Committee for Animal Experiments of the Vrije Universiteit Brussel (CEP N 20-281-6). On day 0, mice were injected intravenously with 1×106 5TGM1-eGFP-positive cells dissolved in 200 μL of RPMI-1640 medium. On day 3, mice were treated intraperitoneally with either vehicle or 20 mg/kg FIDAS-5 three times a week. After 28 days, all mice were sacrificed when the vehicle-treated mice had reached endstage. We isolated bone marrow from the spine and legs, and then lysed the red blood cells. Tumor burden in the bone marrow was assessed by determining the percentage of eGFP-positive cells by flow cytometry and the percentage plasmacytosis on cytospins using May-GrünwaldGiemsa staining. Additionally, serum protein electrophoresis was performed to assess the serum M spike. Statistical analysis All results presented in this study were acquired from at least three independent experiments and analyzed with GraphPad Prism 9.0 software (GraphPad Software Inc, La Jolla, CA, USA). The data represent the mean ± standard deviation, and results were analyzed using the Mann–Whitney U non-parametric test or one-way analysis of variance for multiple testing. Values of *P<0.05, **P<0.01, ***P<0.001 and ****P<0.0001 were considered statistically significant. Other methods The Online Supplementary Information contains details on the other methods.

Haematologica | 109 January 2024

257

ARTICLE - Synthesis of SAM is a target in multiple myeloma

Y. Wang et al.

Continued on following page. Haematologica | 109 January 2024

258

ARTICLE - Synthesis of SAM is a target in multiple myeloma

Y. Wang et al.

Figure 1. Expression and prognostic value of MAT2A in multiple myeloma. (A) Left side. MAT2A mRNA expression levels in normal bone marrow plasma cells (BMPC; N=22), cells from patients with monoclonal gammopathy of undetermined significance (MGUS; (N=44), or smoldering multiple myeloma (SMM; N=12), and multiple myeloma (MM) cells from newly diagnosed patients (N=345) from the UAMS TT2 cohort. Right side. MAT2A mRNA expression levels in normal bone marrow plasma cells (N=5), MGUS (N=5), primary MM cells from newly diagnosed patients (N=206), and human MM cell lines (N=42) from the Heidelberg-Montpellier cohort. ns, non-significant; *P<0.05; ****P<0.0001. (B) Prognostic value of MAT2A mRNA levels in terms of overall survival in newly diagnosed patients from the UAMS TT2 (N=256), and UAMS TT3 (N=158) cohort and in relapsed MM patients from the Mulligan cohort (N=264). Maxstat analysis was used to calculate the optimal separation of patients based on a cutoff value. (C) Overall survival and progression-free survival curves for newly diagnosed MM patients with high or low MAT2A RNA expression (MMRF CoMMpass trial, N=653). The Kaplan-Meier method was used to plot survival curves and group comparisons were made using the log-rank test. P<0.05 was considered as statistically different. (D, E) Basal mRNA (D) and protein (E) expression of MAT2A in nine human MM cell lines (ANBL6, AMO-1, JJN3, LP-1, OPM2, RPMI8226, U266, XG-2 and XG-7). HR: hazard ratio; OS: overall survival; PFS: progression-free survival.

Results High expression of MAT2A is associated with progression of multiple myeloma and a worse outcome To determine the role of MAT2A in MM, we first analyzed the mRNA expression levels of MAT2A in healthy donors and at different stages of MM disease using both publicly available and our own gene expression profiling data. We found that MAT2A expression was significantly upregulated in malignant plasma cells compared to normal bone marrow plasma cells and that MAT2A was even more highly expressed in HMCL (Figure 1A). By comparing MAT2A expression in the different molecular subgroups of MM, we observed a significantly elevated MAT2A mRNA level in the hyperdiploidy and the proliferation groups (Online Supplementary Figure S1A). Next, we investigated the relation between MAT2A mRNA expression and overall survival of MM patients using a cohort of newly diagnosed MM patients (TT2/TT3 cohort) and a cohort of relapsed MM patients (Mulligan cohort). In the cohort of relapsed patients, high MAT2A expression was found to be associated with a significantly worse overall survival, whereas no significant relation was observed in the newly diagnosed patients (Figure 1B). Nevertheless, after assessing the more recent RNA sequencing data from the CoMMpass trial, we did find a significant correlation between high mRNA expression of MAT2A and shorter progression-free survival and overall survival (Figure 1C). The prognostic value of high MAT2A gene expression was also validated in our own cohort of newly diagnosed MM patients in whom RNA sequencing was performed (Online Supplementary Figure S1B). We next determined MAT2A expression in a selected panel of HMCL using quantitative real-time polymerase chain reaction (qRT-PCR) and western blotting and confirmed that MAT2A is highly expressed in HMCL both on the mRNA and protein levels (Figure 1D, E). Together, these data show that MAT2A is highly expressed in MM patients and HMCL, correlating with an unfavorable prognosis. This indicates that MAT2A could be a novel promising target in MM. Silencing of MAT2A impairs multiple myeloma cell viability and proliferation and induces apoptosis To investigate the role of MAT2A in MM in vitro, we selected

the JJN3 and OPM2 cell lines for further in vitro experiments, as these cell lines showed differential MAT2A expression. MAT2A-specific siRNA transfection was performed to inhibit MAT2A gene expression in MM cells. The silencing efficiency was confirmed by detecting MAT2A expression using qRT-PCR and western blotting (Online Supplementary Figure S1C-E). MAT2A expression was inhibited by 40% in JJN3 cells and by 70% in OPM2 cells. Next, we evaluated the long-term effects of MAT2A inhibition, as we assumed that it would take some time for downstream effects to come into play.24 We found that silencing of MAT2A significantly reduced cell viability after 5 days, compared to scrambled siRNA, by 21% in JJN3 and 30% in OPM2 cells (Figure 2A). Furthermore, bromodeoxyuridine (BrdU) incorporation revealed that targeting MAT2A by siMAT2A was able to decrease proliferation in JJN3 cells from 55% to 45% and in OPM2 cells from 55% to 41% (Figure 2B; Online Supplementary Figure S2A). To better understand the effect on proliferation, we then analyzed cell cycle progression using propidium iodide staining. We found that the percentages of MM cells in the G0/G1 and G2 phases were modestly but significantly higher after MAT2A inhibition in the JJN3 and OPM2 cells, thereby confirming our previous results (Figure 2C; Online Supplementary Figure S2B). Finally, MAT2A interference also induced apoptosis in both cell lines, as evidenced by a significant increase in the number of annexin V and/or 7-amino-actinomycin D-positive cells (in JJN3 cells from 23% to 38% and in OPM2 cells from 11% to 18%) (Figure 2D; Online Supplementary Figure S2C) and increased cleavage of the apoptosis markers poly ADP ribose protein (PARP), caspase 9, and caspase 3 (Figure 2E; Online Supplementary Figure S3A). We further confirmed the anti-MM effect of MAT2A depletion by using OPM2 cells transduced with either MAT2A-short hairpin (sh)RNA or scrambled shRNA lentiviruses (Online Supplementary Figure S4A-F). Together, these data prove that MAT2A plays a crucial role in MM cell proliferation, cell cycle progression and survival. MAT2A depletion impairs mTOR-regulated protein synthesis Next, we investigated the underlying mechanisms contributing to the inhibition of MM cell proliferation and sur-

Haematologica | 109 January 2024

259

ARTICLE - Synthesis of SAM is a target in multiple myeloma

Y. Wang et al.

Figure 2. MAT2A inhibition by siMAT2A reduces multiple myeloma cell growth and survival in vitro. Human myeloma cell lines, JJN3 and OPM-2, were transfected with 20 nM siRNA against MAT2A. (A) Cell viability was measured by the CellTiter-Glo Luminescent cell viability assay after 5 days. (B) Cell proliferation was measured by a bromodeoxyuridine incorporation assay followed by flow cytometric analysis after 4 days. (C) The cell cycle was analyzed using propidium iodide followed by flow cytometric analysis after 4 days. (D) Cell apoptosis was measured using flow cytometry by staining for annexin V-FITC/7-amino-actinomycin D after 5 days. (E) The levels of markers of apoptosis, i.e., PARP, caspase 9 and caspase 3 enzymes, were measured by western blotting. Statistical significance was determined by oneway analysis of variance. *P≤0.05, **P≤0.01. si: short interfering; BrdU: bromodeoxyuridine; PARP: poly ADP ribose polymerase; cl: cleaved. Haematologica | 109 January 2024

260

ARTICLE - Synthesis of SAM is a target in multiple myeloma vival upon MAT2A blockade. Since the main biological function of MAT2A is the synthesis of the methyl donor SAM, we first determined SAM levels using liquid chromatography-mass spectrometry (LC-MS). We found that si-

Y. Wang et al.

lencing MAT2A did indeed lower SAM levels, suggesting inhibited SAM synthesis in OPM2 cells (Figure 3A). Importantly, recent research showed that SAM synthesis is required to support protein synthesis and tumor growth,

Figure 3. MAT2A interference inhibits protein synthesis by downregulating the mTOR pathway. (A) OPM2 cells were treated with 20 nM siMAT2A for 3 days and s-adenosyl-methionine concentration levels were measured by liquid chromatography-mass spectrometry. Significance was determined by one-way analysis of variance. (B) Western blot analysis of puromycin uptake after MAT2A knockdown for 4 days in JJN3 and OPM2 cells. (C) Western blot analysis of protein synthesis-related proteins, isolated after 4 days of MAT2A knockdown. ***P≤0.001. si: short interfering; α-puro: α-puromycin.

Haematologica | 109 January 2024

261

ARTICLE - Synthesis of SAM is a target in multiple myeloma

Y. Wang et al.

Continued on following page. Haematologica | 109 January 2024

262

ARTICLE - Synthesis of SAM is a target in multiple myeloma

Y. Wang et al.

Figure 4. MAT2A inhibition by FIDAS-5 reduces multiple myeloma cell survival in vitro. (A, B) JJN3 and OPM2 cells were treated with increasing doses of FIDAS-5 for 48 h (A) or 5 days (B) and cell viability was measured by a CellTiter-Glo Luminescent cell viability assay. (C) Cell proliferation was measured using a bromodeoxyuridine incorporation assay followed by flow cytometric analysis after treatment with FIDAS-5 (250 nM) for 4 days. (D) Cell cycle was analyzed using propidium iodide staining followed by flow cytometry after treatment with FIDAS-5 (250 nM) for 4 days. (E) Western blot analysis of cell cycle proteins, isolated after 4 days of FIDAS-5 (250 nM) treatment. (F) Cell apoptosis was measured using flow cytometry by staining for annexin VFITC/7-amino-actinomycin D after 5 days of treatment with increasing doses of FIDAS-5. (G) The protein levels of the apoptosis markers PARP, caspase 9 and caspase 3 and MCL-1 were measured by western blotting after treatment with FIDAS-5 (250 nM) for 5 days. All experiments were performed in conditioned medium. Statistical significance was determined by one-way analysis of variance or the Mann-Whitney U test. *P≤0.05, **P≤0.01, ***P≤0.001, ****P≤0.0001. BrdU: bromodeoxyuridine; PARP: poly ADP ribose polymerase; cl: cleaved.

mediated by the mTOR pathway.24 Therefore, we next evaluated whether MAT2A is also involved in controlling protein synthesis in MM cells by using the non-radioactive surface sensing of translation (SUnSET) method. After 4 days, we observed that silencing of MAT2A did indeed significantly impair protein synthesis in JJN3 and even more markedly so in OPM2 cells (Figure 3B; Online SupplementaryFigure S3B). Next, we investigated the mTOR pathways and found a significant decrease in p-p70 and p-4EBP1 levels in OPM2 and JJN3 cells after 4 days of MAT2A inhibition. We also observed a significant decrease of peIF4E in JJN3 cells and p-PRAS40 in OPM2 cells (Figure 3C; Online Supplementary Figure S3C). Similar results were obtained with OPM2-shMAT2A cells, in which we observed clear reductions of p-mTOR, p-p70, and p-4E-BP1 (Online Supplementary Figure S3G, H). These findings indicate that the inhibition of protein synthesis by targeting MAT2A is mainly mediated through mTOR pathways. FIDAS-5 restrains cell cycle progression and induces apoptosis of multiple myeloma cells Recently, several new inhibitors of MAT2A have been developed, allowing us to further validate the impact of MAT2A inhibition on MM cell survival. FIDAS-5 and MAT2A inhibitor 1 (or Compound 196) are two potent MAT2A inhibitors, inducing strong anti-cancer effects.25,26 Compared to Compound 196, the anti-tumor effect of FIDAS-5 is supported by more literature, and its in vivo pharmacokinetic properties and applicability have also been described. All tests with MAT2A inhibitors were performed with MM cells cultured in conditioned medium of HS-5 cells, which was used to mimic the bone marrow environment in vitro. We confirmed that this medium did not alter the expression of MAT2A in MM cells (Online Supplementary Figure S5). We treated MM cells (JJN3 and OPM2) for 2 days with either FIDAS-5 or Compound 196 and observed a dose-dependent decrease in cell viability and increase in apoptosis (Figure 4A; Online Supplementary Figure S6A-C). In line with genetic MAT2A inhibition, FIDAS-5 was also able to increase the levels of cleaved caspase-3, caspase-9 and PARP in MM cells (Online Supplementary Figure S6D, E). When the treatment period was increased to 5 days (long-term treatment), we found that

the dose of FIDAS-5 could be reduced to a range of 0.1-1 μM in both cell lines (Figure 4B). This dose was also found to be effective in four other MM cell lines, while the bone marrow stromal cell line HS-5 was only affected at about 4-fold higher doses (Online Supplementary Figure S7). Furthermore, targeting MAT2A with FIDAS-5 for 4 days significantly suppressed cell proliferation in both cell lines, as evidenced by a decrease in BrdU incorporation and a G0/G1 and G2 phase arrest (Figure 4C, D; Online Supplementary Figure S8A, B). In addition, we confirmed the effect of FIDAS-5 on cell cycle-related proteins by western blot analysis. The results showed that MAT2A inhibition with FIDAS-5 potently increased the expression of the cyclin-dependent kinase inhibitor 1 (p21Waf1/Cip1) and 1B (p27kip1) (Figure 4E; Online Supplementary Figure S8C), while strongly reducing the protein levels of the master cell cycle regulator MYC.27 Moreover, we found that longterm inhibition of MAT2A using FIDAS-5 also led to increased, dose-dependent apoptosis of MM cells (Figure 4F; Online Supplementary Figure S9A). On a protein level, again significant increases in cleavage of PARP, caspase 9, and caspase 3 were detected (Figure 4G; Online Supplementary Figure S9B). In addition, western blotting analysis revealed that FIDAS-5 induced cleavage of MCL-1 after 4 days in JJN3 and OPM2 cells (Figure 4G; Online Supplementary Figure S9B). These data confirm that MAT2A targeting leads to a reduction in proliferation and survival of MM cells. FIDAS-5 suppresses multiple myeloma cell protein synthesis through SAM synthesis To determine whether FIDAS-5 does indeed impair MM cell survival by inhibiting enzymatic activity of MAT2A, we next treated MAT2A-silenced and parental OPM2 cells with different doses of FIDAS-5. Briefly, OPM2 cells were transduced with siMAT2A or scrambled siRNA for 3 days. Next, the cells were refreshed and treated with FIDAS-5 for 2 additional days. The antitumor effect of FIDAS-5 on MAT2A-silenced cells was impaired compared to that of the cells transduced with scrambled siRNA (Figure 5A, B; Online Supplementary Figure S10A), indicating that FIDAS5 does inhibit MM cell survival partly by targeting MAT2A. Next, the two cell lines were further treated with 250 nM

Haematologica | 109 January 2024

263

ARTICLE - Synthesis of SAM is a target in multiple myeloma

Y. Wang et al.

Figure 5. Inhibition of MAT2A with FIDAS-5 impairs protein synthesis by downregulating the mTOR pathway. (A, B) siRNAmediated knockdown was established using 20 nM siMAT2A in OPM2 cells for 3 days, after which the cells were treated with 0.5, 1, or 2 μM FIDAS-5 for 48 h. Cell viability was evaluated using a CellTiter-Glo Luminescent Cell viability assay (A) and cell apoptosis was measured using flow cytometry by staining for annexin V-FITC/7-amino-actinomycin D (B). (C) Western blot analysis of protein synthesis-related proteins, isolated after 4 days of FIDAS-5 (250 nM) treatment. (D) Western blot analysis of puromycin uptake after FIDAS-5 (250 nM) treatment for 4 days. All experiments were performed in conditioned medium. Statistical significance was determined by one-way analysis of variance. *P≤0.05, **P≤0.01, ***P≤0.001. Ctl: control; si: short interfering; α-puro: α-puromycin.

Haematologica | 109 January 2024

264

ARTICLE - Synthesis of SAM is a target in multiple myeloma FIDAS-5 for 4 days and the key kinases/proteins related to the mTOR pathway were investigated. Similarly to previous results obtained with silencing MAT2A, we observed a clear decrease in phosphorylation of mTOR, PRAS40, p70, 4EBP1 and eIF4E in JJN3 and OPM2 cells (Figure 5C; Online Supplementary Figure S10B). Moreover, we observed a decrease in puromycin uptake in the same two cell lines (Figure 5D; Online Supplementary Figure S10C). Glycolysis is an important pathway for energy generation and

Y. Wang et al.

cell growth and is often associated with cancer, including MM.28 LC-MS-based metabolomics showed that long-term FIDAS-5 treatment resulted in lower glucose uptake and less lactate secretion, indicative of lower glycolytic activity (Figure 6). More specifically, glycolytic intermediates of upper glycolysis were increased after FIDAS-5 treatment, while intermediates of lower glycolysis were decreased, suggesting a blockage of glycolysis. In addition, FIDAS-5 treatment lowered tricarboxylic acid cycle intermediates,

Figure 6. Inhibition of MAT2A with FIDAS-5 blocks glycolysis and the tricarboxylic acid cycle. Media and cell samples were collected after 5 days of FIDAS-5 (250 nM) treatment, followed by liquid chromatography-mass spectrometry analysis of extracellular glucose and lactate and intracellular metabolites. For the analysis of extracellular metabolite, results represent % peak area ± standard deviation (SD) compared to that of cell-free media. For analysis of intracellular metabolites, data represent relative peak area of metabolite ± SD compared to that of the non-treated controls. Unpaired t tests were performed. **P≤0.01. Ctl: control; TCA: tricarboxylic acid. Haematologica | 109 January 2024

265

ARTICLE - Synthesis of SAM is a target in multiple myeloma such as α-ketoglutarate. These results indicate that FIDAS5 treatment impairs energy metabolism, which can influence cell growth.

Y. Wang et al.

Inhibition of MAT2A delays disease progression in vivo To evaluate whether the anti-tumor effects observed in vitro could be translated in vivo, we used the murine

D E

Figure 7. MAT2A inhibition by FIDAS-5 reduces tumor burden in vivo. (A, B) Murine multiple myeloma cells 5TGM1 were treated with increasing doses of FIDAS-5 for 5 days. Next, cell viability was measured by a CellTiter-Glo Luminescent cell viability assay (A), while cell apoptosis was measured using flow cytometry by staining for annexin VFITC/7-amino-actinomycin D (B). (C) Tumor burden was analyzed by assessing the percentage of eGFP-positive cells using flow cytometry, bone marrow plasmacytosis on cytosmears and the amount of M-protein in serum. (D, E) Western blot analysis of protein synthesis-related proteins. Bar graphs represent the pixel intensity as measured by Image Studio Lite v5.2. Statistical significance was determined by one-way analysis of variance or the Mann-Whitney U test. *P≤0.05, **P≤0.01, ***P≤0.001, ****P≤0.0001. Haematologica | 109 January 2024

266

ARTICLE - Synthesis of SAM is a target in multiple myeloma 5TGM1 MM model to investigate the effect of FIDAS-5 on tumor growth. We first confirmed that FIDAS-5 significantly inhibited cell viability and induced cell apoptosis of the murine MM cell line 5TGM1 (Figure 7A, B). Then, two groups of nine C57BL/KaLwRij mice were injected intravenously with 1×106 5TGM1-eGFP-positive cells. On day 3 after inoculation, mice were treated with either vehicle or 20 mg/kg FIDAS-5. When vehicle-treated mice reached end-stage disease, we evaluated effects on tumor burden by assessing the percentages of eGFP-positive cells and bone marrow plasmacytosis, and the amount of M-protein in the serum. In line with the in vitro data, the results showed that FIDAS-5 was able to significantly decrease tumor load as a single agent (Figure 7C). On a protein level, FIDAS-5 again significantly reduced p-mTOR, p-4EBP1 and MYC levels (Figure 7D, E). Importantly, following treatment, mice were also monitored for body weight loss and signs of toxicity. No significant body weight loss or overt signs of toxicity were observed (Online Supplementary Figure S11A). A repetition of the in vivo experiment had the same effect on mice, again confirming the therapeutic potential of FIDAS5 in MM (Online Supplementary Figure S11B-E). MAT2A inhibition sensitizes multiple myeloma cells to bortezomib Finally, we investigated whether MAT2A inhibition influences MM cell sensitivity towards the standard-of-care agent bortezomib. Inhibition of MAT2A either by FIDAS-5 or siMAT2A significantly enhanced the effect of bortezomib on cell viability in both cell lines (Figure 8A, B). Furthermore, we found strong synergy for the FIDAS-5 and bortezomib long-term combination in both JJN3 and OPM2 cells, with the highest synergy scores being 30.46 (bortezomib: 2.5 nM, FIDAS-5: 500 nM) and 34.41 (bortezomib: 1.25 nM, FIDAS-5: 125 nM), respectively (Online Supplementary Figure S12). We also noticed a significant decrease in BrdU incorporation compared to that for both single agents alone, accompanied by a further reduction in MYC and increase in p21Waf1/Cip1 protein levels. For the JJN3 cells, we also observed a further increase in p27kip1 levels (Figure 8C, D; Online Supplementary Figure S13). In addition, we found that both FIDAS-5 and siMAT2A could significantly enhance the bortezomib-mediated effect on apoptosis (Figure 8E, F; Online Supplementary Figure S4A, B) together with inducing a significant increase in cleaved PARP, cleaved caspase 9 and cleaved caspase 3 levels in both cell lines (Figure 8G; Online Supplementary Figure S14C). Moreover, the combination therapy was beneficial on human primary MM samples (CD138+ cells). The combination further reduced the viability of the MM cells, even though each patient’s sample responded to the treatments differently (Figure 8H). Taken together, these findings provide the rationale to test the anti-MM activity of bortezomib in combination with FIDAS-5 in patients.

Y. Wang et al.

Discussion Our study, for the first time, provides evidence supporting a key role for MAT2A in myeloma progression, thus constituting a promising novel target for the treatment of (relapsed) MM. Mechanistically, we demonstrated that MAT2A inhibition in MM cells triggers, at least in part, inhibition of the mTOR pathway leading to reduced protein synthesis, thereby inhibiting MM cell growth and survival. Using public gene expression profiling data, we showed that high MAT2A mRNA expression is associated with MM malignancy and adverse survival in MM patients. Previous studies showed that MAT2A is overexpressed in various tumors, such as hepatic cancer, breast cancer, glioma cells and colon cancer, and plays a significant role in the pathogenesis of those tumors.29-32 In MM, research by Janker et al. showed that MAT2A is upregulated in patients with a high tumor burden.33 Moreover, we recently demonstrated that MAT2A is upregulated in HMCL resistant to EZH2 inhibition.34 MAT2A may therefore be an interesting therapeutic target in MM and an indicator of poor prognosis. In order to study the function of MAT2A in MM cell biology, we conducted both genetic silencing experiments, in which we knocked down MAT2A using siRNA or shRNA, and chemical inhibition experiments using FIDAS-5 treatment. Inhibiting MAT2A strongly affected MM cell viability, proliferation and protein synthesis, resulting in apoptosis. To understand the mechanistic aspects, we first investigated the effect of MAT2A inhibition on SAM synthesis by mass spectrometry. We observed that silencing MAT2A is able to decrease SAM levels in OPM2 cells. As mentioned before, silencing the expression of MAT2A in solid tumors is also able to affect tumor growth. For example, in the HepG2 cell line, MAT2A silencing could lower intracellular SAM levels and restrict polyamine production, preventing leptin’s pro-survival signal, which is required for cell growth.35 In 2013, Zhang et al. showed that blocking MAT2A activity with FIDAS agents resulted in reduced SAM synthesis, which significantly inhibited the growth of colorectal cancer in vitro. Additionally, using cell-based assays, they found that FIDAS-5 is more potent than other family members and that FIDAS-5 was able to significantly inhibit the growth of tumors in vivo, with minimal differences in body weight of the animals treated.25 The anticancer activity of FIDAS-5 was also validated in gastric cancer cells, in which an obvious downregulation of SAM sensitized the gastric cancer cells to ferroptosis.36 We confirmed that MAT2A silencing attenuates the antitumor effect of FIDAS5, indicating that MAT2A is indeed a target of FIDAS-5. Additionally, SAM inhibition in tumors causes cell cycle arrest in the G1 and G2 phases, which we were able to detect in MM using both siMAT2A and FIDAS-5. We also demonstrated that MYC expression is suppressed by

Haematologica | 109 January 2024

267

ARTICLE - Synthesis of SAM is a target in multiple myeloma

Y. Wang et al.

E G

Figure 8. Inhibition of MAT2A sensitizes multiple myeloma cells to bortezomib. (A) Cells were transfected with 20 nM siRNA against MAT2A for 5 days. On day 3, cells were treated with bortezomib (JJN3: 1.5 nM, OPM2: 2.5 nM) for an additional 2 days. Cell viability was measured by a CellTiter-Glo Luminescent cell viability assay. (B) Cells were treated with FIDAS-5 (JJN3: 2.5 μM, OPM2: 10 μM) and bortezomib (JJN3: 1.5 nM, OPM2: 2.5 nM) for 48 h in conditioned medium and cell viability was measured by a CellTiter-Glo Luminescent cell viability assay. (C, D) Cells were treated with FIDAS-5 (JJN3: 2.5 μM, OPM2: 10 μM) and bortezomib Continued on following page. Haematologica | 109 January 2024

268

ARTICLE - Synthesis of SAM is a target in multiple myeloma

Y. Wang et al.

(JJN3: 1.5 nM, OPM2: 2.5 nM) for 24 h in conditioned medium. Next, cell proliferation was measured by a bromodeoxyuridine incorporation assay followed by flow cytometry (C) and cell cycle proteins were analyzed by western blot (D). (E, F) Cells were transfected with 20 nM siMAT2A (E) or treated with FIDAS-5 (JJN3: 2.5 μM, OPM2: 10 μM) in conditioned medium (F), combined with bortezomib (JJN3: 1.5 nM, OPM2: 2.5 nM). After 48 h, cell apoptosis was measured using flow cytometry by staining for annexin V-FITC/7-amino-actinomycin D. (G) The protein levels of the apoptosis markers PARP, caspase 9 and caspase 3 were measured by western blotting after treatment with FIDAS-5 (JJN3: 2.5 μM, OPM2: 10 μM) and bortezomib (JJN3: 1.5 nM, OPM2: 2.5 nM) for 48 h in conditioned medium. (H) CD138+ MM cells isolated from three patients’ samples were cultured for 24 h in conditioned medium and treated with bortezomib and FIDAS-5. Cell viability was measured by CellTiter-Glo assay. Statistical significance was determined by one-way analysis of variance. *P≤0.05, **P≤0.01, ***P≤0.001, ****P≤0.0001. BZ: bortezomib; BrdU: bromodeoxyuridine; Combo: combination of bortezomib and FIDAS-5; PARP: poly ADP ribose polymerase; cl: cleaved.

MAT2A inhibition. It is well known that MYC contributes significantly to a number of biological functions, including proliferation and cell cycle progression.37 Recently, it was discovered that MYC induces the expression of MAT2A, explaining why MAT2A is elevated in proliferative tissues, notably cancer.24 In colorectal cancer cells, FIDAS-5 decreased the expression of the oncogenes MYC and Cyclin D1, while increasing the expression of the cell cycle inhibitor p21Waf1/Cip125. A copious amount of protein must be synthesized in order for cells to survive and this is especially true for malignant plasma cells. According to our data, MAT2A seems to contribute to cell proliferation by supporting protein synthesis, since both MAT2A depletion and pharmacological inhibition of MAT2A activity resulted in a decrease of protein synthesis. This is in line with the study of Villa and colleagues, who demonstrated that MAT2A targeting inhibits protein synthesis, and tumor development by diminishing intracellular SAM levels. Moreover, we observed that MAT2A regulates translation via mTOR signaling. According to Gu et al., the SAM sensor SAMTOR forms a stable dimer with Gap Activity TOward Rags 1 (GATOR1) to limit mTORC1 activity in HEK293T cells when SAM levels are decreasing. In contrast, when sufficient SAM levels are present, SAM binds to SAMTOR thereby disassembling the SAMTOR-GATOR1 complex and thus activating mTORC1 signaling and triggering translation.38 As a key player in drug resistance and the pathogenesis of MM, the PI3K/AKT/mTOR pathway has been well described in MM.39 The progression of MM is decelerated and cell death is induced when p-p70, p-S6, p-4EBP1, and p-eIF4E are inhibited.16 mTOR is also a regulator of glycolysis, an important pathway for energy and cell growth, and our metabolomics data showed that MAT2A inhibition results in lower glycolytic activity, which could also result in lower cell proliferation. As a major methyl donor, SAM plays a pivotal role in the methylation of DNA, RNA and proteins that regulate fundamental cellular activities. In MM, hypermethylation of tumor suppressor genes is involved in the regulation of cell cycle progression, DNA repair, apoptosis, drug response and key signaling pathways.40 Therefore, investigating the involvement of MAT2A in the DNA methylation landscape of MM cells will be an interesting topic for future research. We also performed in vivo experiments using FIDAS-5 to

target MAT2A in order to analyze its therapeutic potential. In our study, we found that FIDAS-5, as a single treatment, was able to considerably reduce tumor burden in two separate experiments, without inducing obvious toxicity. Moreover, FIDAS-5 significantly decreased the levels of p-mTOR and p-4EBP1 in vivo, while p-PRAS40 and p-p70 also showed modest decreases in one experiment. Furthermore, there was a considerable decrease in the levels of MYC. Previous research revealed that FIDAS-5 dramatically delayed the development of HT29 tumor xenografts in nude mice, while only minimally altering the body weight of the animals.25 It could also decrease tumor growth in the CAL-51 cancer cell line-derived xenograft model.24 The effectiveness of bortezomib in MM was the first evidence that disrupting protein homeostasis may be used as a cancer treatment strategy.41 Currently, bortezomib is one of the main standard-of-care agents for the treatment of newly diagnosed MM patients.42 However, although the majority of the MM patients initially respond well to bortezomib treatment, most of them will eventually relapse. Since we demonstrated in our study that inhibiting MAT2A reduced protein synthesis, we hypothesized that the combination of MAT2A inhibition and bortezomib might have enhanced antitumor efficacy. In addition, enhanced glycolysis and tricarboxylic acid cycle activity have been associated with resistance to bortezomib, and FIDAS-5 inhibits both these pathways as well.43,44 Our data showed that the combination did indeed significantly decrease cell proliferation and synergistically inhibited cell viability compared to treatment with the two drugs as single agents. Furthermore, FIDAS-5 and bortezomib induced cell death in a caspase-dependent manner. Thus, combining FIDAS-5 with bortezomib may be a novel way to delay or even overcome bortezomib resistance in MM. The clinical phase I trial with AG-270, another MAT2A inhibitor with strong affinity, in patients with lymphoma or solid tumors (NCT03435250, clinical trials.gov) further supports the idea that targeting MAT2A could be clinically feasible. Disclosures No conflicts of interest to disclose. Contributions EM, EDB, and YW contributed to the conception, design and

Haematologica | 109 January 2024

269

ARTICLE - Synthesis of SAM is a target in multiple myeloma revision of the manuscript. YM and AW performed experiments. EDB, JM, CM, and PV provided bioinformatic analyses. EZ and CB performed the liquid chromatography-mass spectometry. EM, EDB, YW, CM, AW, AM, PV, KDV, KV, JM, EZ, and CB were responsible for writing, reviewing and editing the manuscript. All authors read and approved the final version of the manuscript. Acknowledgments The authors would like to thank Carine Seynaeve and Charlotte Van De Walle for their invaluable help.

Y. Wang et al.

Funding This study was supported by the China Scholarship Council (N. 01906280057), Vrije Universiteit Brussel (SRP48), Kom Op Tegen Kanker, the International Myeloma Foundation and Fonds Willy Gepts (UZ-Brussel). KDV is a post-doctoral fellow of FWO Vlaanderen (12I0921N). Data-sharing statement The data generated in this study are available upon request from the corresponding authors.

References 1. Rajkumar SV, Dimopoulos MA, Palumbo A, et al. International Myeloma Working Group updated criteria for the diagnosis of multiple myeloma. Lancet Oncol. 2014;15(12):e538-548. 2. Rajkumar SV, Kumar S. Multiple myeloma current treatment algorithms. Blood Cancer J. 2020;10(9):94. 3. Moreau P, Kumar SK, San Miguel J, et al. Treatment of relapsed and refractory multiple myeloma: recommendations from the International Myeloma Working Group. Lancet Oncol. 2021;22(3):e105-e118. 4. Cengiz Seval G, Beksac M. The safety of bortezomib for the treatment of multiple myeloma. Expert Opin Drug Saf. 2018;17(9):953-962. 5. Rajkumar SV. Multiple myeloma: 2020 update on diagnosis, riskstratification and management. Am J Hematol. 2020;95(5):548-567. 6. Sonneveld P. Management of multiple myeloma in the relapsed/refractory patient. Hematology Am Soc Hematol Educ Program. 2017;2017(1):508-517. 7. Hirschey MD, DeBerardinis RJ, Diehl AME, et al. Dysregulated metabolism contributes to oncogenesis. Semin Cancer Biol. 2015;35(Suppl):S129-S150. 8. Yang C, Ko B, Hensley CT, et al. Glutamine oxidation maintains the TCA cycle and cell survival during impaired mitochondrial pyruvate transport. Mol Cell. 2014;56(3):414-424. 9. Zaal EA, Berkers CR. The influence of metabolism on drug response in cancer. Front Oncol. 2018;8:500. 10. LeGros HL Jr, Halim AB, Geller AM, Kotb M. Cloning, expression, and functional characterization of the beta regulatory subunit of human methionine adenosyltransferase (MAT II). J Biol Chem. 2000;275(4):2359-2366. 11. Pajares MA, Markham GD. Methionine adenosyltransferase (Sadenosylmethionine synthetase). Adv Enzymol Relat Areas Mol Biol. 2011;78:449-521. 12. Yang H, Huang ZZ, Wang J, Lu SC. The role of c-Myb and Sp1 in the up-regulation of methionine adenosyltransferase 2A gene expression in human hepatocellular carcinoma. FASEB J. 2001;15(9):1507-1516. 13. Tomasi ML, Ryoo M, Skay A, et al. Polyamine and methionine adenosyltransferase 2A crosstalk in human colon and liver cancer. Exp Cell Res. 2013;319(12):1902-1911. 14. Liu Q, Liu L, Zhao Y, et al. Hypoxia induces genomic DNA demethylation through the activation of HIF-1alpha and transcriptional upregulation of MAT2A in hepatoma cells. Mol Cancer Ther. 2011;10(6):1113-1123. 15. Nikesitch N, Lee JM, Ling S, Roberts TL. Endoplasmic reticulum stress in the development of multiple myeloma and drug

resistance. Clin Transl Immunology. 2018;7(1):e1007. 16. Zismanov V, Attar-Schneider O, Lishner M, Heffez Aizenfeld R, Tartakover Matalon S, Drucker L. Multiple myeloma proteostasis can be targeted via translation initiation factor eIF4E. Int J Oncol. 2015;46(2):860-870. 17. Zhou H, Huang S. The complexes of mammalian target of rapamycin. Curr Protein Pept Sci. 2010;11(6):409-424. 18. Hsieh AC, Liu Y, Edlind MP, et al. The translational landscape of mTOR signalling steers cancer initiation and metastasis. Nature. 2012;485(7396):55-61. 19. Guri Y, Colombi M, Dazert E, et al. mTORC2 promotes tumorigenesis via lipid synthesis. Cancer Cell. 2017;32(6):807-823.e12. 20. Pollizzi KN, Patel CH, Sun IH, et al. mTORC1 and mTORC2 selectively regulate CD8(+) T cell differentiation. J Clin Invest. 2015;125(5):2090-2108. 21. Goldfinger M, Shmuel M, Benhamron S, Tirosh B. Protein synthesis in plasma cells is regulated by crosstalk between endoplasmic reticulum stress and mTOR signaling. Eur J Immunol. 2011;41(2):491-502. 22. Dreesen IA, Fussenegger M. Ectopic expression of human mTOR increases viability, robustness, cell size, proliferation, and antibody production of Chinese hamster ovary cells. Biotechnol Bioeng. 2011;108(4):853-866. 23. Kroczynska B, Kaur S, Katsoulidis E, et al. Interferon-dependent engagement of eukaryotic initiation factor 4B via S6 kinase (S6K)- and ribosomal protein S6K-mediated signals. Mol Cell Biol. 2009;29(10):2865-2875. 24. Villa E, Sahu U, O'Hara BP, et al. mTORC1 stimulates cell growth through SAM synthesis and m(6)A mRNA-dependent control of protein synthesis. Mol Cell. 2021;81(10):2076-2093.e9. 25. Zhang W, Sviripa V, Chen X, et al. Fluorinated N,Ndialkylaminostilbenes repress colon cancer by targeting methionine S-adenosyltransferase 2A. ACS Chem Biol. 2013;8(4):796-803. 26. Konteatis ZD, Sui Z, Travins JM, Ye Z, inventors. Inhibitors of cellular metabolic processes. US patent application 2018; WO/2018/045071. 27. Blomberg I, Hoffmann I. Ectopic expression of Cdc25A accelerates the G(1)/S transition and leads to premature activation of cyclin E- and cyclin A-dependent kinases. Mol Cell Biol. 1999;19(9):6183-6194. 28. Oudaert I, Van der Vreken A, Maes A, et al. Metabolic cross-talk within the bone marrow milieu: focus on multiple myeloma. Exp Hematol Oncol. 2022;11(1):49. 29. Cai J, Sun WM, Hwang JJ, Stain SC, Lu SC. Changes in S-

Haematologica | 109 January 2024

270

ARTICLE - Synthesis of SAM is a target in multiple myeloma adenosylmethionine synthetase in human liver cancer: molecular characterization and significance. Hepatology. 1996;24(5):1090-1097. 30. Simile MM, Peitta G, Tomasi ML, et al. MicroRNA-203 impacts on the growth, aggressiveness and prognosis of hepatocellular carcinoma by targeting MAT2A and MAT2B genes. Oncotarget. 2019;10(29):2835-2854. 31. Xu J, Wu D, Wang S, Wang Z. MAT2B expression correlates with poor prognosis in triple-negative breast cancer. Cancer Manag Res. 2019;11:5501-5511. 32. Chen H, Xia M, Lin M, et al. Role of methionine adenosyltransferase 2A and S-adenosylmethionine in mitogeninduced growth of human colon cancer cells. Gastroenterology. 2007;133(1):207-218. 33. Janker L, Mayer RL, Bileck A, et al. Metabolic, anti-apoptotic and immune evasion strategies of primary human myeloma cells indicate adaptations to hypoxia. Mol Cell Proteomics. 2019;18(5):936-953. 34. Nylund P, Atienza Parraga A, Haglof J, et al. A distinct metabolic response characterizes sensitivity to EZH2 inhibition in multiple myeloma. Cell Death Dis. 2021;12(2):167. 35. Ramani K, Yang H, Xia M, Ara AI, Mato JM, Lu SC. Leptin's mitogenic effect in human liver cancer cells requires induction of both methionine adenosyltransferase 2A and 2beta. Hepatology. 2008;47(2):521-531. 36. Ma M, Kong P, Huang Y, et al. Activation of MAT2A-ACSL3

pathway protects cells from ferroptosis in gastric cancer. Free Radic Biol Med. 2022;181:288-299. 37. Garcia-Gutierrez L, Delgado MD, Leon J. MYC oncogene contributions to release of cell cycle brakes. Genes (Basel). 2019;10(3):244. 38. Gu X, Orozco JM, Saxton RA, et al. SAMTOR is an Sadenosylmethionine sensor for the mTORC1 pathway. Science. 2017;358(6364):813-818. 39. Liu R, Chen Y, Liu G, et al. PI3K/AKT pathway as a key link modulates the multidrug resistance of cancers. Cell Death Dis. 2020;11(9):797. 40. Alzrigat M, Parraga AA, Jernberg-Wiklund H. Epigenetics in multiple myeloma: from mechanisms to therapy. Semin Cancer Biol. 2018;51:101-115. 41. Bianchi G, Anderson KC. Contribution of inhibition of protein catabolism in myeloma. Cancer J. 2019;25(1):11-18. 42. Yimer H, Melear J, Faber E, et al. Daratumumab, bortezomib, cyclophosphamide and dexamethasone in newly diagnosed and relapsed multiple myeloma: LYRA study. Br J Haematol. 2019;185(3):492-502. 43. Zaal EA, de Grooth HJ, Oudaert I, et al. Targeting coenzyme Q10 synthesis overcomes bortezomib resistance in multiple myeloma. Mol Omics. 2022;18(1):19-30. 44. Zaal EA, Wu W, Jansen G, Zweegman S, Cloos J, Berkers CR. Bortezomib resistance in multiple myeloma is associated with increased serine synthesis. Cancer Metab. 2017;5:7.

Haematologica | 109 January 2024

271

Y. Wang et al.

ARTICLE - Plasma Cell Disorders

Impact of viral hepatitis therapy in multiple myeloma and other monoclonal gammopathies linked to hepatitis B or C viruses Alba Rodríguez-García,1 Nicolas Mennesson,2 Gema Hernandez-Ibarburu,3,4 María Luz Morales,1 Laurent Garderet,5 Lorine Bouchereau,2 Sophie Allain-Maillet,2 Eric Piver,6,7 Irene Marbán,3 David Rubio,3,4 Edith Bigot-Corbel,2,8 Joaquín Martínez-López,1,9# María Linares1,10# and Sylvie Hermouet.2,11# Department of Translational Hematology, Instituto de Investigación Hospital 12 de Octubre (i+12), Hematological Malignancies Clinical Research Unit H120-CNIO, Madrid, Spain; 2Nantes Université, INSERM, Immunology and New Concepts in ImmunoTherapy (INCIT), UMR 1302 F44000 Nantes, France; 3Biomedical Informatics Group, Universidad Politécnica de Madrid, Madrid, Spain; 4TriNetX LLC, Madrid, Spain; 5Sorbonne Université-INSERM, UMR_S 938, Centre de Recherche Saint-Antoine-Team Hematopoietic and Leukemic Development, Assistance Publique-Hôpitaux de Paris, Hôpital Pitié Salpetrière, Département d'Hématologie et de Thérapie Cellulaire, F-75013 Paris, France; 6Laboratoire de Biochimie, CHU Tours, Tours, France; 7Inserm UMR1253, MAVIVH Tours, Tours, France; 8Laboratoire de Biochimie, CHU Nantes, F-44000 Nantes, France; 9Department of Medicine, Medicine School, Universidad Complutense de Madrid, Madrid, ES 28040, Spain; 10Department of Biochemistry and Molecular Biology, Pharmacy School, Universidad Complutense de Madrid, Madrid, ES 28040, Spain and 11Laboratoire d’Hématologie, CHU Nantes, F-44000 Nantes, France 1

Correspondence: S. Hermouet sylvie.hermouet@univ-nantes.fr M. Linares mlinares@ucm.es Received: Accepted: Early view:

March 7, 2023. May 10, 2023. May 18, 2023.

JM-L, ML and SH contributed equally as senior authors.

Abstract Subsets of multiple myeloma (MM) and monoclonal gammopathies of undetermined significance (MGUS) present with a monoclonal immunoglobulin specific for hepatitis C virus (HCV), thus are presumably HCV-driven, and antiviral treatment can lead to the disappearance of antigen stimulation and improved control of clonal plasma cells. Here we studied the role of hepatitis B virus (HBV) in the pathogenesis of MGUS and MM in 45 HBV-infected patients with monoclonal gammopathy. We analyzed the specificity of recognition of the monoclonal immunoglobulin of these patients and validated the efficacy of antiviral treatment (AVT). For 18 of 45 (40%) HBV-infected patients, the target of the monoclonal immunoglobulin was identified: the most frequent target was HBV (n=11), followed by other infectious pathogens (n=6) and glucosylsphingosine (n=1). Two patients whose monoclonal immunoglobulin targeted HBV (HBx and HBcAg), implying that their gammopathy was HBV-driven, received AVT and the gammopathy did not progress. AVT efficacy was then investigated in a large cohort of HBV-infected MM patients (n=1367) who received or did not receive anti-HBV treatments and compared to a cohort of HCV-infected MM patients (n=1220). AVT significantly improved patient probability of overall survival (P=0.016 for the HBV-positive cohort, P=0.005 for the HCV-positive cohort). Altogether, MGUS and MM disease can be HBV- or HCV-driven in infected patients, and the study demonstrates the importance of AVT in such patients.

Introduction Multiple myeloma (MM) is characterized by clonal expansion of transformed plasma cells in the bone marrow. MM is preceded by pre-malignant stages including asymptomatic monoclonal gammopathy of undetermined significance (MGUS) and/or smoldering myeloma (SMM).1-5 A minority of MGUS evolves over time toward overt MM. Unfortunately, despite great advances in the understanding and treatment of this disease, MM remains incurable.

The primary function of plasma cells is to produce immunoglobulins (Ig) that mediate humoral immunity against infection. When there is an infection, Ig-secreting plasma cells differentiate from the naïve B cells that recognize foreign antigens. This process takes place in the germinal centers of secondary lymphoid organs where B cells undergo proliferation and somatic hypermutations, followed by the selection of B cells with high antigen affinity. In MGUS, SMM and particularly MM, clonal plasma cells secrete large quantities of a single Ig (monoclonal Ig),

Haematologica | 109 January 2024

272

ARTICLE - Impact of viral hepatitis therapy in multiple myeloma which serves as a marker of the disease and triggers many of the symptoms.1-6 Latent infection and chronic antigen stimulation are now recognized as initial pathogenic events leading to chronic lymphocytic leukemia (CLL) and lymphoma.7-11 B-cell receptor (BCR) signaling is central to the specific recognition of Igs, suggesting that specific antigens could be involved in the development of CLL. For instance, CLL may express stereotypic BCR specific for β-(1,6)-glucan, a major component of yeasts and fungi of the microbiota.8 BCR stimulation from these antigens results in proliferation and increased cell survival. Several studies support chronic antigenic stimulation as a pathogenic mechanism in subsets of MGUS and MM.12-22 Associations between MM and viral infection, particularly hepatitis C virus (HCV), human immunodeficiency virus (HIV), or Epstein-Barr virus (EBV) were reported.12-20 In addition, Nair et al. identified glucosylsphingosine (GlcSph) as the target of monoclonal Ig in the context of Gaucher’s disease and also in sporadic gammopathies.21,22 This pathogenic model opened new possibilities of treatment for MGUS, SMM and MM: if the target of a patient’s monoclonal Ig is identified and can be eliminated, chronic antigenstimulation disappears, leading to the control of clonal plasma cells. The efficacy of target antigen reduction therapy has been proven, first, for MGUS and SMM linked to GlcSph, and recently, for MGUS and MM linked to HCV.23-25 In HCV-infected patients whose monoclonal Ig reacted against HCV, treating the HCV infection improved MGUS and MM disease.25 In the present study, we first explored the role of hepatitis B virus (HBV) in patients infected with HBV and diagnosed with monoclonal gammopathy. We analyzed the reactivity of the monoclonal Ig of patients against HBV, and the efficacy of treating the viral infection on MGUS and MM disease outcome. We found that for 36.7% of the 30 HBV-infected (HBV+) patients for whom the monoclonal Ig could be analyzed, the target was HBV, suggesting that HBV infection initiated the clonal gammopathy. The efficacy of antiviral therapy was confirmed in a second set of studies performed on large cohorts of MM patients with a history of HBV or HCV infection prior to the diagnosis of MM.

Methods Patients A first cohort included 45 patients with monoclonal gammopathy who had been infected with HBV. In this cohort, different assays were performed with the aim of identifying the target of the monoclonal Ig of patients. Thirty patients were followed at the Centres Hospitaliers Universitaires (CHU) of Paris Saint Antoine (n=26) or Tours (n=4), France, and 15 patients at the Hospital 12 de Oc-

A. Rodríguez-García et al.

tubre, Madrid, Spain. Most patients were diagnosed with MGUS (n=14) or MM (n=29). Biological and clinical data were available at the time of diagnosis for 31 patients (8 MGUS, 23 MM) (Online Supplementary Table S1). The studies were approved by our Institutional Review Boards and the patients provided written informed consent in accordance with the Declaration of Helsinki. MGUS or MM disease status was monitored by the quantification of monoclonal Ig or total Igs or free kappa/lambda light chain levels in serum. Protein and monoclonal Ig levels were visualized by serum protein electrophoresis and/or immunofixation electrophoresis.25,26 A second and a third cohort included patient data obtained from TriNetX, LLC (TriNetX) including data from subjects from healthcare organizations (HCO) all over the world with an MM diagnosed between the last 3 to 20 years from now, after a diagnosis of HBV or HCV infection, who had or had not received antiviral treatments. Cohorts were queried in TriNetX on March 22nd, 2022, and two large cohorts of MM patients found to have been infected by HBV (HBV+, n=1367) or HCV (HCV+, n=1220) were collected. Patients were classified into four groups: infected MM patients who received antiviral treatment (against HBV, n=175; or against HCV, n=179), and those who did not (n=1192 for HBV+ patients; n=1041 for HCV+ patients). The characteristics of these patients are summarized in Table 1, and the flowchart of the study is represented in Figure 1A. Separation of monoclonal Ig from non-clonal Ig For each patient, agarose gel electrophoresis and purification of the monoclonal Ig from other Ig present in serum samples were performed as published previously1720,27 and detailed in the Online Supplementary Appendix. Analysis of the specificity of antigen recognition of monoclonal Ig We and others previously reported that GlcSph, a glucolipid, and at least eight infectious pathogens could be the targets of monoclonal Ig from MGUS and MM patients.12-22 Thus three different assays were used to determine the target of monoclonal Ig of HBV+ patients, after purification of each monoclonal Ig : i) a GlcSph immunoblot assay, to determine whether a monoclonal Ig recognizes GlcSph; ii) the multiplex infectious antigen microarray (MIAA) assay, to determine whether purified monoclonal Ig recognize one infectious pathogen among different pathogens; and iii) dot blot assays designed to confirm that HBV or other pathogens, proteins or Ag, were recognized by the purified monoclonal Ig.17-20,27,28 All assays are detailed in the Online Supplementary Appendix. Statistical analysis Data analysis was performed by GraphPad Prism 6.01 software and for the large HBV+ and HCV+ MM cohorts with

Haematologica | 109 January 2024

273

ARTICLE - Impact of viral hepatitis therapy in multiple myeloma

A. Rodríguez-García et al.

Continued on following page. Haematologica | 109 January 2024

274

ARTICLE - Impact of viral hepatitis therapy in multiple myeloma

A. Rodríguez-García et al.

Figure 1. The effect of antiviral treatment in hepatitis C virus- and hepatitis B virus-infected multiple myeloma cohorts. (A) Study design and flowchart of hepatitis B virus (HBV)+ (orange) and hepatitis C virus (HCV)+ (blue) cohorts. A total of 1,367 (HBV+) or 1,192 (HCV+) patients from 49 out of the 73 healthcare organizations (HCO) of the TriNetX network were included in the cohort of patients diagnosed with multiple myeloma (MM) after HBV or HCV infection. Anti-viral treatments (AVT) were tenofovir disoproxil, lamivudine, peginterferon alfa-2a, interferon alfa-2b, tenofovir alafenamide, entecavir. Anti-HCV treatments were elbasvir, grazoprevir, glecaprevir, pibrentasvir, sofosbuvir, velpatasvir, voxilaprevir, or combinations of these drugs. (B) Survival analysis of the HBV+ and HCV+ cohorts. Number of patients and log-rank test in each cohort, with outcome and survival probability at the end of time window in HBV or HCV cohorts. df: degree of freedom. (C) Kaplan-Meier plots comparing overall survival since the time of MM diagnosis of patients with HBV or HCV infection who received AVT (purple) or not (green). Since no differences in age or sex among groups of AVT treated versus untreated patients were observed, analyses were performed without propensity score matching (HBV+ patients: P=0.270; HCV+ patients: P=0.466).

Table 1. Characteristics of patients diagnosed with multiple myeloma post hepatitis B virus or hepatitis C virus infection, depending on receiving or not receiving antiviral therapy. HBV+ Cohort

HCV+ Cohort

AVT

No AVT

AVT

No AVT

Sex N Male/female (%)

175 73.7

1,192 60.1

179 65.9

1,041 64.8

Age at diagnosis N Mean ± SD

175 59.8 ± 10.4

1,192 61 ± 13.1

179 59.5 ± 8.9

1,041 60.2 ± 12.7

Leukocytes (x109/L) N Mean ± SD

155 6.46 ± 5.58

1,077 12.6 ± 13.4

168 6.26 ± 4.79

882 13.6 ± 16.8

Hemoglobin (g/dL) N Mean ± SD

164 11.4 ± 3.08

1,125 11.1 ± 2.67

175 11.4 ± 2.57

946 10.9 ± 2.75

Platelets (x109/L) N Mean ± SD

166 158 ± 96.8

1,101 179 ± 102

176 161 ± 87.9

916 174 ± 100

Calcemia (mmol/L) N Mean ± SD

159 2.25 ± 0.23

1,091 2.25 ± 0.21

175 2.27 ± 0.20

890 2.23 ± 0.23

Creatinine (μmol/L) N Mean ± SD

164 173 ± 214

1,117 260 ± 906

175 250 ± 933

937 247 ± 715

Bone lesions N N, with lesions

175 61

1,192 449

179 78

1,041 379

N: number of patients; SD: Standard Deviation; AVT: antiviral treatment; HBV: hepatitis B virus; HCV: hepatitis C virus.

the TriNetX Platform. All analyses are described in the Online Supplementary Appendix.

Results Determination of the target of purified monoclonal Ig of hepatitis B virus-positive patients Serum samples were collected from 45 HBV+ patients with a monoclonal Ig (13 MGUS, 1 POEMS syndrome, 30 MM, 1 plasma cell leukemia [PCL]) and 3 control MM patients vaccinated against HBV. MGUS monoclonal Ig were 10 IgG, 1 IgA and 2 IgM. MM monoclonal Ig were 20 IgG, 8 IgA, 1 IgD, and light chains in one case. The POEMS patient had a mono-

clonal IgA and the PCL patient, a monoclonal IgG. The characteristics of HBV+ patients are presented in Online Supplementary Table S1. We were able to purify 30 monoclonal Ig (27 IgG, 2 IgA, 1 IgM) from 8 MGUS and 22 MM HBV-infected patients (Online Supplementary Figure S1). The specificity of recognition of the purified monoclonal Ig was analyzed using the GlcSph, MIAA and dot blot assays. These studies identified the target of 20/30 (66.7%) monoclonal Ig (18 IgG, 1 IgA, 1 IgM) from 18 HBV-infected patients and 2 patients vaccinated against HBV. Six of these patients were diagnosed with MGUS, and 14 with MM. The targets of monoclonal Ig were infectious pathogens for 18 patients (6 MGUS, 12 MM) and GlcSph for 2 MM patients (Table 2). These results were confirmed by

Haematologica | 109 January 2024

275

ARTICLE - Impact of viral hepatitis therapy in multiple myeloma

A. Rodríguez-García et al.

Figure 2. Results of the MIAA and dot blot assays of six hepatitis B virus-specific monoclonal Ig. For each patient (Pt), results obtained in parallel with the unseparated serum IgG (serum, left) and the patient’s purified monoclonal IgG (Mc IgG, right) using the MIAA assay are represented; results are shown as fluorescent intensity (FI). The FI values shown for each pathogen, Ag, protein or lysate, were obtained after subtraction of the fluorescent background (F-B) of each pathogen protein or lysate. Dotted Continued on following page. Haematologica | 109 January 2024

276

ARTICLE - Impact of viral hepatitis therapy in multiple myeloma

A. Rodríguez-García et al.

lines show thresholds of specific positivity defined for each viral pathogen or protein: 1,400 for Epstein-Barr virus (EBV), cytomegalovirus (CMV), varicella-zoster virus (VZV), hepatitis B virus (HBV) and B. burgdorferi (blue); 1000 for herpes simplex virus (HSV)-1 and HSV-2 (orange); 500 for hepatitis C virus (HCV) and T. gondii (green). Dots may be superimposed; horizontal bars represent means of results obtained for a pathogen (Ag, lysate). Experiments were performed in triplicate, repeated at least once. Inserts show the results of dot blot assays performed with purified recombinant HBV proteins (HBsAg, HBeAg, HBcAg, HBx) and water as control. The assays showed that the Mc IgG (or IgM, Pt M13) recognized a single HBV protein. Experiments were performed at least twice.

Table 2. Identified targets of purified monoclonal Ig from HBV-infected or -vaccinated patients with monoclonal gammopathy. Diagnosis

Mc Ig type

Anti-HBs/HBc/HBe Ab in serum

Target of Mc Ig

AVT

Pt M02

IgGλ

Anti-HBc, Anti-HBs

HSV-1

Yes

Pt M13

MGUS

IgMκ

HBx

Yes (entecavir)

Pt M14

IgGκ

Anti-HBc

H. pylori

Yes

Pt M18

MGUS

IgGλ

Anti-HBc, anti-HBe, HBsAg

HBcAg

Yes (entecavir)

Pt M25

MGUS

IgGκ

Anti-HBe*

EBV EBNA-1

Yes (entecavir)

Yes

Pt P01

IgGλ

Anti-HBc, anti-HBe*

GlcSph

Pt P04

MGUS

IgGκ

Anti-HBe, Anti-HBc

HBeAg

Pt P05

IgG

Anti-HBc

HBx

Pt P06

IgG

Anti-HBs, anti-HBc, anti-HBe*

HBx

Pt P07

IgGκ

Anti-HBs

HBx

Pt P08

MGUS

IgGλ

Anti-HBe*

EBV EBNA-1

Yes (unknown)

Pt P11

IgG

Anti-HBs,* anti-HBe*

EBV EBNA-1

Pt P12

IgGκ

Anti-HBe*

HSV-1 gG

Pt P14

IgGκ

Anti-HBc, anti-HBe*

HBx

Pt P16

IgGκ

Anti-HBc, anti-HBe*

HBx

Yes

Pt P23

IgGκ

Anti-HBe*

HBx

Pt P24

IgGκ

Anti-HBc

HBcAg

Pt P29

MGUS

IgGλ

Anti-HBs,* anti-HBe*

HBeAg

Pt P13

IgGκ

Anti-HBs

GlcSph

Pt P17

IgAκ

Anti-HBs

Coxsackievirus VP1

Patients

Decrease in Disease Mc Ig progression

Infected

Vaccinated

Information on the presence of anti-HBs/HBc/HBe Ab in serum was obtained from hospital laboratories, except when labelled * (data obtained from the MIAA assay). Ab: antibodies; AVT: antiviral therapy; CR: complete remission of MM; MM: multiple myeloma; MGUS: monoclonal gammopathy of undetermined significance; Mc Ig: monoclonal immunoglobulin; HBV: hepatitis B virus; EBV: Epstein Barr virus; EBNA: EpsteinBarr nuclear antigen; HSV-1: herpes simplex virus 1; H. pylori: Helicobacter pylori; GlcSph: glucosylsphingosine; ND: no data; NA: not applicable; Pt: patient.

immunoblot or dot blot assays (Figure 2, Figure 3, Online Supplementary Figures S2 and S3). To summarize, four infectious pathogens were recognized by monoclonal Ig from HBV-infected patients: HBV in 11 cases (4 MGUS, 7 MM), EBV in 3 cases (2 MGUS, 1 MM), H. pylori in one MM case, and HSV-1 for 2 MM patients. Thus, HBV was the most frequent target of monoclonal Ig from HBV-infected patients (11/30 analyzed, or 36.7%). The HBV proteins targeted were protein X (HBx) (n=7), antigen e (HBeAg) (n=2), and the core protein (HBc) (n=2). For other HBV-infected patients, the targets of monoclonal Ig were EBV EBNA-1 (n=3), HSV-1 gG (n=2), H. pylori (n=1), and GlcSph (n=1). In

addition, the purified monoclonal Ig from 3 MM patients (P13, P17, P18) vaccinated against HBV were also analyzed. The targets were GlcSph (P13), Coxsackievirus VP1 protein (P17), or unknown (P18). Altogether, the monoclonal Ig of 2 MGUS and 8 MM patients lacked an identified target after these analyses (Online Supplementary Figure S4). The targets of the monoclonal Ig of the PCL and POEMS patients were also not identified. As reported,19,20 the percentage of monoclonal Ig of unknown specificity was higher for MM (10/22 or 45.5%) than for MGUS (2/8 or 25%) patients, but in this cohort the difference was not significant (P=0.3118 by χ2 test).

Haematologica | 109 January 2024

277

ARTICLE - Impact of viral hepatitis therapy in multiple myeloma

A. Rodríguez-García et al.

Figure 3. Results of the MIAA and dot blot assays of monoclonal Ig that target an infectious pathogen other than hepatitis B virus. For each patient (Pt), results obtained in parallel with the unseparated serum IgG (serum, left) and the patient’s purified monoclonal IgG or IgA (Mc IgG or Mc IgA, right) using the MIAA assay are represented; results are shown as fluorescent intensity (FI). The FI values shown for each pathogen, Ag, protein or lysate, were obtained after subtraction of the fluorescent background Continued on following page. Haematologica | 109 January 2024

278

ARTICLE - Impact of viral hepatitis therapy in multiple myeloma

A. Rodríguez-García et al.

(F-B) of each pathogen protein or lysate. Dotted lines show thresholds of specific positivity defined for each viral pathogen or protein: 1,400 for Epstein-Barr virus (EBV), cytomegalovirus (CMV), varicella-zoster virus (VZV), hepatitis B virus (HBV) and B. burgdorferi (blue); 1000 for HSV-1 and HSV-2 (orange); 500 for hepatitis C virus (HCV), H. pylori and T. gondii (green). Dots may be superimposed; horizontal bars represent the means of results obtained for a pathogen (Ag, lysate). Experiments were performed in triplicate, repeated at least once. Inserts show the results of dot blot assays performed with either purified recombinant proteins from EBV, herpes simplex virus (HSV)-1 or Enterovirus VP1, and water or PBS used as control. Experiments were performed at least twice.

Table 3. Summary of the characteristics of the cohorts in the hepatitis B virus analysis before and after matching. Cohort

Mean ± SD

N of patients

% Cohort

P 0.270

Age at Index

59.8 ± 10.4

175

100

Age at Index

61.0 ± 13.1

1,192

100

Cohort 1 is Global multiple myeloma (MM) post hepatitis B virus (HBV) infection with antiviral therapy (AVT), and cohort 2 is Global MM post HBV infection without AVT. SD: Standard Deviation.

Table 4. Summary of the characteristics of the cohorts in the HCV analysis before and after matching. Cohort

Mean ± SD

N of patients

% Cohort

P 0.466

Age at Index

59.5 ± 8.9

179

100

Age at Index

60.2 ± 12.7

1,041

100

Cohort 1 is Global multiple myeloma (MM) post hepatitis C virus (HCV) infection with antiviral therapy (AVT), and cohort 2 is Global MM post HCV infection without AVT. SD: Standard Deviation.

Evolution of monoclonal gammopathies of undetermined significance and multiple myeloma disease in patients with a hepatitis B virus-specific monoclonal Ig Patient data and disease evolution were analyzed according to the target of the monoclonal Ig: HBV, other target, unknown target. For MGUS patients, HBV infection was anterior to the diagnosis of MGUS in all cases, and confirmed by the presence of antibodies against HBc, HBe or HBs (Table 2). MGUS patients P04 and P29, who presented with a monoclonal Ig specific for HBeAg, were both young (27 and 42 years old); unfortunately, follow-up information was not available for these patients. MGUS patients M13, M18, M25 and P08 received antiviral treatment. No progression was observed for patients M13 and M18, who both had a monoclonal Ig specific for a HBV protein (HBx for Pt M13, HBcAg for Pt M18). MGUS patients M25 and P08 had a monoclonal Ig specific for EBV EBNA-1, and antiHBV treatment had no effect on their gammopathy. Patient M25 showed a strong increase in the amount of monoclonal Ig, and patient P08 progressed toward MM, now in third relapse. Seven MM patients had a HBV-specific monoclonal Ig. Those patients for whom information was available presented at diagnosis with a low burden of MM disease (<15% plasma cells in bone marrow, β2-microglobulin <3.5 mg/L, hemoglobin (Hb) ≥10 g/dL, creatinine <100 μmol/L, ISS stage I-II), compared to patients with a monoclonal Ig of unknown specificity (Online Supplementary Table S2); however, differences were not significant. Unfortunately,

none of these 7 patients had received antiviral treatment. After MM therapy, 2 patients (P14, P24) achieved complete remission (CR) for at least one year, and one patient (P16) progressed. Overall survival of multiple myeloma patients who received anti-hepatitis B virus treatment We then analyzed the TriNetX cohort of 1367 HBV+ MM patients. The main characteristics of patients treated (or not) with antiviral drugs are summarized in Table 3; there was no significant difference between the treated and untreated cohorts. We analyzed the evolution of MM disease in patients post HBV infection: those who received antiviral treatment (tenofovir disoproxil, lamivudine, peginterferon alfa-2a, interferon alfa-2b, tenofovir alafenamide, entecavir) had an overall survival probability at 3 years of 77.91%, compared to 68.41% for patients who did not receive antiviral treatment (Figure 1B, C); this difference was statistically significant by log-rank test (P=0.016). Since we did not observe differences in age or sex among the two groups of patients (P=0.270), analyses were performed without propensity score matching. Overall survival of multiple myeloma patients who received anti-hepatitis C virus treatment For comparison, we analyzed the evolution of MM disease in 1,220 HCV-infected MM patients. Table 4 shows patients' characteristics; again, there was no significant difference between patients treated or not with antiviral drugs. MM patients who received anti-HCV treatments (elbasvir, gra-

Haematologica | 109 January 2024

279

ARTICLE - Impact of viral hepatitis therapy in multiple myeloma zoprevir, glecaprevir, pibrentasvir, sofosbuvir, velpatasvir, voxilaprevir, or combinations of some of these drugs) had a high overall survival probability at 3 years (80.46%) compared to patients who did not (70.78%) (P=0.005 by logrank test) (Figure 1B, C). Analyses were performed without propensity score matching since there was no difference in age or sex among the two groups of patients (P=0.466).

Discussion For significant subsets of monoclonal gammopathies, including MM, the patient’s monoclonal Ig specifically recognizes an infectious pathogen, which implies that the clonal gammopathy was initiated by chronic stimulation by an antigen from this pathogen.12,13,17-20,29 Moreover, patients with HCV-driven MGUS or MM (when the monoclonal Ig targeted HCV) benefited from AVT: MGUS or MM disease evolution improved after anti-HCV therapy, including a refractory MM for whom antiviral therapy led to long-term complete remission.25 The present study strengthens and extends these findings to MGUS and MM linked to HBV: for over a third of HBV+ patients whose monoclonal Ig could be studied, the monoclonal Ig targets HBV, which implies that HBV initiated the gammopathy in these individuals. Moreover, we demonstrate that AVT improves the survival of HBV+ MM patients. Our study also confirms in a large cohort that antiviral treatment improves MM outcome for HCV-infected patients.24,25 The targets of monoclonal Ig from HBV+ MGUS and MM patients could be studied only in the first cohort of 45 patients, for whom 30 monoclonal Ig could be purified and analyzed. For 23.3% (7/30) HBV+ MGUS and MM patients, the target of the purified monoclonal Ig was EBV, HSV-1, H. pylori or GlcSph; thus the gammopathy of these patients was not linked to HBV. For 36.7% (11/30) HBV+ patients (4 MGUS, 7 MM), the monoclonal Ig specifically recognized a HBV protein, implying a role for HBV in the initiation of the gammopathy. These findings support early prescription of AVT, at the MGUS stage whenever possible. In western Europe, HBV-initiated gammopathies may not represent a large fraction of MGUS and MM, but the implication of HBV in MM pathogenesis may be more frequent in parts of the world where HBV infection is endemic (Eastern Europe, Asia, the Middle East, sub-Saharan Africa, South America).30 Previous studies showed that ≥80% monoclonal Ig of HCV+ patients target HCV, especially the strongly immunogenic and oncogenic core protein. Regarding HBV, patients with controlled, inactive HBV infection present with antibodies directed at HBe, HBc, and HBs. Consistent with ancient, controlled HBV infection, at the time of MGUS or MM diagnosis patients had polyclonal antibodies against HBV (anti-HBe, HBc and/or HBs) in serum. However, mono-

A. Rodríguez-García et al.

clonal Ig of HBV+ MM patients frequently targeted a lesser known viral protein, non-structural HBx. The genome of HBV, a DNA virus, contains four overlapping open reading frames (ORF) named S, P, C and X. The S ORF produces envelope proteins including HBsAg; the P and C ORF encode the HBV polymerase, core protein (HBc), and the HBeAg; and X ORF encodes Hbx.31,32 HBx is commonly detected in hepatocytes of HBV-infected patients, and its expression correlates with viral replication.33 Indeed, Hbx plays a central role in the initiation and maintenance of HBV replication, but also inhibits the host’s immune response to the virus, and facilitates the development of hepatocellular carcinoma (HCC).34,35 The presence of anti-HBx IgG in the serum of individuals with chronic HBV infection is relatively frequent, and associated with the presence of anti-HBc IgG.36,37 Several groups reported anti-HBx IgG in the serum of as many as 40.6% HBV-infected patients with liver cirrhosis (LC), and 5-34.4% of HBV-infected patients with HCC.36,37 Thus antiHBx IgG can be considered as markers of viral replication and HBV-mediated development of LC and HCC.38 Chronic production of a large amount of clonal anti-HBx antibody could help control HBV infection. In this study, biological and clinical information was available for 3 MM patients with HBV-specific monoclonal Ig: at the time of MM diagnosis, none presented signs of active liver disease, and they had a mild form of MM disease (β2-microglobulin <3.5 mg/L, Hb ≥10 g/L, calcemia <2.6 mmol/L, no bone lesions, ISS I-II). Unfortunately, the evolution of these patients was unknown. More studies are necessary to establish whether the production of large amounts of anti-HBx IgG has beneficial effects on HBV infection and hepatic disease. The second part of our studies was designed to investigate the effect of AVT on MM disease for HBV-infected MM patients, compared to HCV-infected MM patients. In both cohorts, individuals who received AVT fared significantly better after 1,000 days of evolution than MM patients whose HBV or HCV infection was not treated. The best effect of AVT was observed for HCV+ patients, possibly because the anti-HBV treatments in this retrospective study do not totally eliminate HBV. HBV covalently closed circular DNA (cccDNA) remains in hepatocytes, and cccDNA persistence represents a therapeutic barrier in curing HBV infection.39-41 Another reason is that MM is frequently driven by HCV in HCV+ patients (with >80% of monoclonal Ig targeting HCV), whereas MM is presumably HBV-driven in “only” 36.7% HBV+ patients. One limitation of this study is the lack of serum samples for the TriNetX cohorts, which made identifying the target of the monoclonal Ig in these cohorts impossible. Altogether, clonal gammopathies of HBV-infected patients are frequently HBV-driven. This study demonstrates the importance of antiviral treatments for patients with HBVor HCV-driven clonal gammopathies, including MM. Anti-

Haematologica | 109 January 2024

280

ARTICLE - Impact of viral hepatitis therapy in multiple myeloma viral therapy should be prescribed as early as possible in the development of HBV- or HCV-linked clonal gammopathies, ideally at the MGUS stage. Patients with virusinitiated MGUS or MM can be identified via the analysis of the target of their monoclonal Ig. Disclosures No conflicts of interest to disclose. Contributions ARG, ML, JML, SH and EBC designed the research, analyzed data, and wrote the initial manuscript draft. NM, LB, SAM, EBC and ARG performed experiments and edited the manuscript. EP, LG, GHI, MLM, IM, DR and JML contributed patient samples and clinical data, and edited the manuscript. SH, ML and JML provided resources and financial support, and edited the manuscript. All authors gave final approval of the version to be submitted to publication and agreed to be accountable for all aspects of the work.

A. Rodríguez-García et al.

Acknowledgments We are particularly indebted to all the patients who participated in the study. Funding This work was supported by grants from the Ligue Nationale contre le Cancer (Comités Départementaux 44, 56, 29, 85, 35) (to SH), and a grant from the Spanish Society of Hematology (to ARG). We thank the Instituto de Investigación Hospital 12 de Octubre (i+12), CIBERONC, AECC (Accelerator Award and Ideas Semilla), and the CRIS foundation for their help. Data-sharing statement All data generated or analyzed during this study are included in this published article and its Online Supplementary Appendix. All source data are available on request to the corresponding authors.

References Hermouet S. Hepatitis C virus (HCV) infection, monoclonal immunoglobulin specific for HCV core protein, and plasma-cell malignancy. Blood. 2008;112(10):4357-4358. 14. McShane CM, Murray LJ, Engels EA, Landgren O, Anderson LA. Common community-acquired infections and subsequent risk of multiple myeloma: a population-based study: infections and multiple myeloma. Int J Cancer. 2014;134(7):1734-1740. 15. Li Y, Li Y, Zhang L, Li W. Hepatitis C virus infection and risk of multiple myeloma: evidence from a meta-analysis based on 17 case-control studies. J Viral Hepat. 2017;24(12):1151-1159. 16. Yan J, Wang J, Zhang W, Chen M, Chen J, Liu W. Solitary plasmacytoma associated with Epstein-Barr virus: a clinicopathologic, cytogenetic study and literature review. Ann Diagn Pathol. 2017;27:1-6. 17. Bosseboeuf A, Feron D, Tallet A, et al. Monoclonal IgG in MGUS and multiple myeloma targets infectious pathogens. JCI Insight. 2017;2(19):e95367. 18. Bosseboeuf A, Seillier C, Mennesson N, et al. Analysis of the targets and glycosylation of monoclonal IgAs from MGUS and myeloma patients. Front Immunol. 2020;11:854. 19. Bosseboeuf A, Mennesson N, Allain-Maillet S, et al. Characteristics of MGUS and multiple myeloma according to the target of monoclonal immunoglobulins, glucosylsphingosine, or Epstein-Barr virus EBNA-1. Cancers (Basel). 2020;12(5):1254. 20. Harb J, Mennesson N, Lepetit C, et al. Comparison of monoclonal gammopathies linked to poliovirus or coxsackievirus vs. other infectious pathogens. Cells. 2021;10(2):438. 21. Nair S, Branagan AR, Liu J, Boddupalli CS, Mistry PK, Dhodapkar MV. Clonal immunoglobulin against lysolipids in the origin of myeloma. N Engl J Med. 2016;374(6):555-561. 22. Nair S, Sng J, Boddupalli CS, et al. Antigen-mediated regulation in monoclonal gammopathies and myeloma. JCI Insight. 2018;3(8):e98259. 23. Nair S, Bar N, Xu ML, Dhodapkar M, Mistry PK.

1. Palumbo A, Anderson KC. Multiple myeloma. N Engl J Med. 2011;364(11):1046-1060. 2. Rajkumar SV, Dimopoulos MA, Palumbo A, et al. International Myeloma Working Group updated criteria for the diagnosis of multiple myeloma. Lancet Oncol. 2014;15(12):e538-e548. 3. Dhodapkar MV. MGUS to myeloma: a mysterious gammopathy of underexplored significance. Blood. 2016;128(23):2599-2606. 4. Kyle RA, Larson DR, Therneau TM, et al. Long-term follow-up of monoclonal gammopathy of undetermined significance. N Engl J Med. 2018;378(3):241-249. 5. Boyle EM, Davies FE, Leleu X, Morgan GJ. Understanding the multiple biological aspects leading to myeloma. Haematologica. 2014;99(4):605-612. 6. Guillerey C, Nakamura K, Vuckovic S, Hill GR, Smyth MJ. Immune responses in multiple myeloma: role of the natural immune surveillance and potential of immunotherapies. Cell Mol Life Sci. 2016;73(8):1569-1589. 7. Morrison VA. Infections in patients with leukemia and lymphoma. In: Stosor V, Zembower TR, editors. Infectious complications in cancer patients. Cham: Springer International Publishing; 2014. p. 319-349. 8. Hoogeboom R, van Kessel KPM, Hochstenbach F, et al. A mutated B cell chronic lymphocytic leukemia subset that recognizes and responds to fungi. J Exp Med. 2013;210(1):59-70. 9. Seifert M, Scholtysik R, Küppers R. Origin and pathogenesis of B cell lymphomas. In: Küppers R, editor. Lymphoma. New York, NY: Springer New York; 2019. p. 1-33. 10. Stevenson FK, Forconi F, Kipps TJ. Exploring the pathways to chronic lymphocytic leukemia. Blood. 2021;138(10):827-835. 11. Stevens WBC, Netea MG, Kater AP, van der Velden WJFM. Trained immunity: consequences for lymphoid malignancies. Haematologica. 2016;101(12):1460-1468. 12. Hermouet S, Corre I, Gassin M, Bigot-Corbel E, Sutton CA, Casey JW. Hepatitis C virus, human herpesvirus 8, and the development of plasma-cell leukemia. N Engl J Med. 2003;348(2):178-179. 13. Bigot-Corbel E, Gassin M, Corre I, Le Carrer D, Delaroche O,

Haematologica | 109 January 2024

281

ARTICLE - Impact of viral hepatitis therapy in multiple myeloma Glucosylsphingosine but not Saposin C, is the target antigen in Gaucher disease-associated gammopathy. Mol Genet Metab. 2020;129(4):286-291. 24. Panfilio S, D’Urso P, Annechini G, et al. Regression of a case of multiple myeloma with antiviral treatment in a patient with chronic HCV infection. Leuk Res Rep. 2013;2(1):39-40. 25. Rodríguez-García A, Linares M, Morales ML, et al. Efficacy of antiviral treatment in hepatitis C virus (HCV)-driven monoclonal gammopathies including myeloma. Front Immunol. 2022;12:797209. 26. Kumar S, Paiva B, Anderson KC, et al. International Myeloma Working Group consensus criteria for response and minimal residual disease assessment in multiple myeloma. Lancet Oncol. 2016;17(8):e328-e346. 27. Bosseboeuf A, Allain-Maillet S, Mennesson N, et al. Proinflammatory state in monoclonal gammopathy of undetermined significance and in multiple myeloma is characterized by low sialylation of pathogen-specific and other monoclonal immunoglobulins. Front Immunol. 2017;8:1347. 28. Feron D, Charlier C, Gourain V, et al. Multiplexed infectious protein microarray immunoassay suitable for the study of the specificity of monoclonal immunoglobulins. Anal Biochem. 2013;433(2):202-209. 29. Linares M, Hermouet S. Editorial: the role of microorganisms in multiple myeloma. Front Immunol. 2022;13:960829. 30. Im YR, Jagdish R, Leith D, et al. Prevalence of occult hepatitis B virus infection in adults: a systematic review and meta-analysis. Lancet Gastroenterol Hepatol. 2022;7(10):932-942. 31. Zhao F, Xie X, Tan X, et al. The functions of hepatitis B virus encoding proteins: viral persistence and liver pathogenesis. Front Immunol. 2021;12:691766. 32. Feitelson MA, Bonamassa B, Arzumanyan A. The roles of hepatitis B virus-encoded X protein in virus replication and the

pathogenesis of chronic liver disease. Expert Opin Ther Targets. 2014;18(3):293-306. 33. Slagle BL, Bouchard MJ. Role of HBx in hepatitis B virus persistence and its therapeutic implications. Curr Opin Virol. 2018;30:32-38. 34. You H, Qin S, Zhang F, et al. Regulation of pattern-recognition receptor signaling by HBX during hepatitis B virus infection. Front Immunol. 2022;13:829923. 35. Moraleda G, Saputelli J, Aldrich CE, Averett D, Condreay L, Mason WS. Lack of effect of antiviral therapy in nondividing hepatocyte cultures on the closed circular DNA of woodchuck hepatitis virus. J Virol. 1997;71(12):9392-9399. 36. Ogawa E, Nakamuta M, Koyanagi T, et al. Sequential HBV treatment with tenofovir alafenamide for patients with chronic hepatitis B: week 96 results from a real-world, multicenter cohort study. Hepatol Int. 2022;16(2):282-293. 37. Tu T, Zehnder B, Wettengel JM, et al. Mitosis of hepatitis B virus-infected cells in vitro results in uninfected daughter cells. JHEP Rep. 2022;4(9):100514. 38. Zhang H, Wu L-Y, Zhang S, et al. Anti-hepatitis B virus X protein in sera is one of the markers of development of liver cirrhosis and liver cancer mediated by HBV. J Biomed Biotechnol. 2009;2009:289068. 39. Rajoriya N, Combet C, Zoulim F, Janssen HLA. How viral genetic variants and genotypes influence disease and treatment outcome of chronic hepatitis B. Time for an individualised approach? J Hepatol. 2017;67(6):1281-1297. 40. Levrero M, Stemler M, Pasquinelli C, et al. Significance of antiHBx antibodies in hepatitis B virus infection. Hepatology. 1991;13(1):143-149. 41. Hoare J, Henkler F, Dowling JJ, et al. Subcellular localisation of the X protein in HBV infected hepatocytes. J Med Virol. 2001;64(4):419-426.

Haematologica | 109 January 2024

282

A. Rodríguez-García et al.

LETTER TO THE EDITOR

SARS-CoV-2 vaccination in 361 non-transplanted patients with aplastic anemia and/or paroxysmal nocturnal hemoglobinuria SARS-CoV-2, first recognized in December 2019 in Wuhan, China, was responsible for the largest global pandemic in recent history. Patients with hematological disease, including aplastic anemia (AA) and/or paroxysmal nocturnal hemoglobinuria (PNH) were considered potentially vulnerable, being advised to take precautions such as home isolation to reduce infection risk.1 There are also often concerns about vaccination in this patient group, due to case reports of vaccination e.g., flu vaccination causing AA de novo, or disease relapse.2-4 In addition, vaccination of patients with PNH, a highly hemolytic and thrombotic condition usually treated with complement inhibition, may cause potential issues of overwhelming hemolysis or thrombotic complications.5 SARS-CoV-2 posed a potential life-threatening health risk to patients with AA and/or PNH, thus consensus advice was to proceed with vaccination in this cohort.6 The Severe Aplastic Anemia Working Party and Infectious Diseases Working Party of European Society for Blood and Marrow Transplantation (EBMT) undertook an observational study of SARS-CoV-2 vaccination outcome to guide clinical practice which we report here, with the largest cohort of patients to date with aplastic anemia and/or PNH receiving SARS-Cov-2 vaccination. Disease-related outcome, vaccine complications including overall survival, comparison between those on immunosuppression and those not on immunosuppression and PNH-related complications were considered. All EBMT centers were invited to participate in this prospective observational study. Non-transplanted adult patients with AA and/or PNH invited to receive a SARS-CoV-2 vaccine were included. Patients consented to share their data with EBMT. Data was collected from January 2021 to September 2022 and included disease status, vaccinations received, vaccination side effects, blood parameters 3, 6 and 12 months post-vaccination with disease status at these time points, and SARS-CoV-2 infection occurrence. Four hundred and fifty-seven patients were included from 20 centers; 361 patients received at least one vaccination and were included for detailed analysis. See Table 1 for baseline characteristics. One hundred and thirty-nine patients were on active immunosuppression at first vaccination, and 120 patients on complement inhibitors with the majority on C5 inhibition (90.8%) (see Table 1).

Treatment received within 1 year of first vaccination included PNH treatment only (99/361), AA treatment only (149/361), PNH and AA treatment concurrently (30/361) and no treatment for either disorder (83/361). Three hundred and forty-seven patients had at least two vaccinations, with a mean of 45.51 days between vaccine one and vaccine two. Side effects within 5 days post vaccination were experienced by 117 of 361 patients (32.4%), most commonly pyrexia, myalgia, headache and/or fatigue or a combination of the four symptoms (101/117). Two patients had pancytopenia, which recovered and was not considered AA relapse with disease status stable at 6 and 12 months with partial remission (1/2) and complete re-

Table 1. Baseline characteristics. Baseline characteristic Median age at vaccination in years (IQR) Sex: female, N (%) PNH only (no AA), N (%) AA disease status at time of vaccination, N (%) Complete remission Partial remission Relapsed/refractory/progression Stable disease Not applicable Treatment status at time of vaccination, N (%) AA on IST AA on non-IST treatment No treatment for AA or PNH PNH and AA on treatment for both* C5 inhibitor Factor B inhibitor Anticoagulation PNH on complement inhibitor C5 inhibitor C3 inhibitor Factor B inhibitor Number of vaccines administered, N (%) 1 2 3 >3 Mean time in days between vaccines (IQR)

Patients within study N=361 55.2 (38.1-69.2) 206 (57) 52 (14) N=309 102 (33) 158 (51.1) 19 (6.1) 22 (7.1) 8 (2.6) N=361 117 (32.4) 27 (7.4) 97 (26.8) 24 (6.6) 21 (87.5) 2 (8.3) 1 (4.2) 96 (26.5) 88 (91.7) 3 (3.1) 5 (5.2) 14 (3.9) 110 (30.5) 225 (62.3) 12 (3.3) 76 (30-186)

*Twenty-two patients were on immunosuppressive therapy (IST) treatment for aplastic anemia (AA) and complement inhibition. IQR: interquartile range; PNH: paroxysmal nocturnal hemoglobinuria.

Haematologica | 109 January 2024

283

LETTER TO THE EDITOR mission (1/2). For patients with PNH, PNH-related side effects were experienced by 19 (14.7%), including four complement inhibitor-treated patients experiencing a breakthrough hemolysis event. One patient on complement inhibition and calcineurin inhibitor immunosuppression who received ChAdOx1-S vaccine experienced a pulmonary embolism 16 days post vaccination. Platelet count at vaccination was 161x109/L, 146x109/L at presentation of thrombosis, and 100x109/L 3 days post thrombosis. D dimer were not taken at the time of thrombosis. This was considered vaccine-induced pulmonary embolism and no further SARS-CoV-2 vaccinations were undertaken. Disease relapse: nine patients experienced AA relapse with six patients undergoing a concomitant reduction in immunosuppression, one with unknown circumstances, and one due to myelodysplastic syndrome transformation. One relapse considered possibly vaccine-related with platelet count reduction from 72x109/L to 12x109/L 1 week post vaccination. This patient responded to oxymethalone dose increase within 4 weeks, although this could be considered vaccine induced thrombocytopenia. Twelve-month overall survival was 98% (95% confidence interval [CI]: 97-100), with no difference in outcome of those on active immunosuppression compared to patients not on active immunosuppression with 12-month survival of 99% in patients not on immunosuppression and 98% in patients taking immunosuppression at first vaccination; P=0.5. There were six deaths within 1 year after the first vaccination, none related to SARS-CoV-2 vaccination or infection (3 due to aplastic anemia relapse, progression or refractory disease, 1 due to bacterial infection, 1 due to diffuse large B-cell lymphoma and 1 due to brain hemorrhage). Blood counts remained stable up to 1 year post vaccination with minimal change (see Table 2). Cumulative incidence of SARS-CoV-2 12 months post vaccination was 16% (95% CI: 12-20) (53/361), with 76.9% of patients (40/53) experiencing symptoms which were generally mild. No patients required oxygen or had pulmonary findings on chest X-ray (39/40, 1 missing data point) and no pa-

tients died from Covid-19. We report the largest vaccine study to date in non-transplant patients with AA and/or PNH. The consensus guideline for patients with AA and/or PNH recommended to proceed with vaccination, due to significant concerns about disease-related complications associated with potential SARS-CoV-2 infection.6 There are case reports of de novo AA in patients testing positive for SARS-CoV-2,7 and of SARS-CoV-2 causing reduction in blood counts or new transfusion requirements in patients with known AA, although insufficient to cause relapse. PNH related complications post infection were also reported with hemolysis.1,8,9 Case reports of patients developing de novo or relapsing AA caused by different vaccines e.g., flu vaccine, hepatitis vaccines, has led to caution in vaccinating this cohort of patients. Case reports are also available following SARS-CoV-2 vaccination, irrespective of vaccine used, with patients presenting with de novo AA several weeks after vaccination.10,11 We did not identify any newly diagnosed patients with AA within this study, as patients included were those already diagnosed. Country-wide vaccine observational data are required to assess this without reporting bias, such as Sweden’s CoVacSafe-SE study, although registry level ‘big’ data have advantages and disadvantages.12 Röth et al. report four patients who had AA relapse at median 77 days post vaccination, with mainly platelet counts affected. All four patients were on calcineurin inhibitors, and all needed treatment for relapsed AA.13 Within our cohort, nine patients relapsed within 12 months postvaccination, the majority associated with immunosuppression withdrawal and thus cannot be attributed to vaccination. Our cohort had 38.5% of patients on active immunosuppression at time of vaccination, with no outcome difference compared to patients not on immunosuppression. Rajput et al. report similar outcomes, with 30% (15/50) of patients on immunosuppression at vaccination, and 94% of patients

Table 2. Blood results at 3, 6 and 12 months post vaccination. Baseline, N=361

Hb, g/L WBC, x109/L Neutrophils, x109/L Platelets, x109/L

Median (IQR) 116.5 (98-129.8) 4 (2.9-5.2) 2 (1.4-2.9) 123 (74-171)

Missing N (%) 3 (0.8) 8 (2.2) 5 (1.4) 4 (1.1)

3 Months, N=361 Median (IQR) 114 (96-128) 4 (3-5) 2 (1.4-2.7) 122.5 (73.8-169.2)

Missing N (%) 38 (10) 42 (11.6) 44 (12.2) 41 (11.4)

IQR: interquartile range; Hb: hemoglobin; WBC: white blood count. Haematologica | 109 January 2024

284

6 Months, N=361 Median (IQR) 115.5 (97-130) 3.9 (3-5.1) 2 (1.4-2.9) 122 (79-172)

Missing N (%) 25 (6.9) 30 (8.3) 30 (8.3) 27 (7.5)

12 Months, N=361 Median (IQR) 119 (101-132) 4.1 (3.2-5) 2 (1.5-2.8) 128 (86-174.8)

Missing N (%) 56 (15.5) 60 (16.6) 61 (16.9) 55 (15.2)

LETTER TO THE EDITOR having no disease status change post vaccination. Six percent (3/50) patients relapsed post vaccination, although similar to our cohort, two of the three patients had discontinued AA treatment 2 and 30 days prior to vaccination, raising the question of whether this was vaccine-related relapse or due to withdrawal of immunosuppresion.14 Potential considerations would be to suspend immunosuppression withdrawal during a vaccination program. PNH-related complications in those on complement inhibitors in our cohort were moderate, similar reports conclude up to a 14% breakthrough hemolysis risk in patients undergoing vaccination.8,15 This is manageable, but requires physician vigilance and patient education. Advising patients with PNH to receive vaccines within the first half of complement inhibitor treatment schedule is preferable to reduce risks of breakthrough hemolysis e.g., within 1 week post eculizumab, within 4 weeks post ravulizumab. There was a high proportion of patients in this cohort experiencing SARS-CoV-2 infection following vaccination, the majority of whom were symptomatic. Reassuringly, there were no deaths from COVID-19 in this large 361 patient cohort, supporting vaccination for patients with AA and/or PNH. There were some study limitations: analyzing vaccine response in patients with AA and/or PNH was not done, as not all participants had access to antibody testing. However, several groups have undertaken vaccine response studies in this patient cohort, demonstrating similar antibody response to healthy volunteers after the second SARS-CoV-2 vaccine irrespective of immunosuppression use.8,14 In the current prospective study, some data points were missing, and a cohort of non-vaccinated patients would have been useful for comparison. In conclusion, we report the largest study to date assessing non-transplant patients with AA and/or PNH receiving vaccination for SARS‐CoV‐2 infection. Vaccination was well tolerated, with low complication rates and low relapse risk of AA. There was no difference in outcome of those on active immunosuppression compared to patients who were not on active immunosuppression and no deaths from SARS‐CoV‐2 infection were observed post vaccination. This cohort would suggest clinicians and patients should be reassured regarding vaccination for SARS-CoV-2 in patients with AA and/or PNH.

Gavriilaki,19 Francesco Lanza,20 Corentin Orvain,21 Antonio Maria Risitano,22 Rafael de la Camara23 and Régis Peffault de Latour24 Department of Hematology, St James University Hospitals, Leeds,

UK; 2EBMT Statistical Unit, Leiden, the Netherlands; 3EBMT Leiden Study Unit, Leiden, the Netherlands; 4RM Gorbacheva Research Institute, Pavlov University, St. Petersburg, Russian Federation; 5

Department of Hematology, Leiden University Medical Center,

Leiden, the Netherlands; 6SC Ematologia, Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan, Italy; 7Department of Hematology, University Hospital of Wales, Cardiff, UK; 8National Institute of Blood Disease & Bone Marrow Transplantation, Karachi, Pakistan; 9Department of Hematology, University Hospital Center Zagreb, Zagreb, Croatia; 10Dipartimento di Diagnostica per Immagini, Radioterapia Oncologica ed Ematologia, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, Italy; 11Department of Hematology, Medical University of Warsaw, Warsaw, Poland; 12Institut Català d'Oncologia-Hospital Universitari Germans Trias i Pujol; Josep Carreras Leukemia Research Institute, Barcelona, Spain; 13Azienda Ospedaliero-Universitaria Maggiore della Carità and Translational Medicine Department, University of Eastern Piedmont, Novara, Italy; 14

Department of Hematology, Nice University Hospital, Cote d’Azur

University, Nice, France; 15Department of Hematology, Copernicus Hospital, Lodz, Poland; 16Department of Hematology, Oncology, Hemostaseology and Stem Cell Transplantation, RWTH Aachen University, Aachen, Germany; 17University Division of Hematology, A.O. Ordine Mauriziano, Torino, Italy; 18Department of Hematology, University Hospital Basel, Basel, Switzerland; 19Department of Hematology, George Papanicolaou General Hospital, Thessaloniki, Greece; 20Metropolitan Transplant Network, Hospital Santa Maria delle Croci, Ravenna, Italy; 21Centre Hospitalier Universitaire d’Angers, Angers, France; 22Department of Hematology, AORN Moscati, Avellino, Italy; 23Department of Hematology, Hospital de la Princesa, Madrid, Spain and 24Clinical Investigations Center, Hôpital Saint-Louis, Paris, France Correspondence: M GRIFFIN - m.griffin@nhs.net https://doi.org/10.3324/haematol.2023.283863 Received: July 6, 2023. Accepted: August 10, 2023. Early view: August 17, 2023. ©2024 Ferrata Storti Foundation Published under a CC BY-NC license

Authors

Disclosures 1

Morag Griffin, Dirk-Jan Eikema, Inge Verheggen, Alexander Kulagin,

MG is on the advisory board and/or has received speaker fees from

Jennifer M-L. Tjon,5 Bruno Fattizzo,6 Wendy Ingram,7 Uzma Zaidi,8

Alexion AstraZeneca rare disease, Sobi, Biocryst, Novartis and Amgen;

Lana Desnica,9 Sabrina Giammarco,10 Joanna Drozd-Sokolowska,11

consults for Regeneron and Biocryst. BF consults for and has received

Blanca Xicoy,12 Andrea Patriarca,13 Michael Loschi,14 Anna Szmigielska-

speaker bureau fees from Alexion AstraZeneca rare disease, Amgen,

Kaplon, Fabian Beier, Alessandro Cignetti, Beatrice Drexler, Eleni

Annexon, Novartis, Janssen and Sobi. WI is on the advisory board of

Haematologica | 109 January 2024

285

LETTER TO THE EDITOR Celgene, Novartis, Gilead, Alexion and Sobi; has received honoraria from

reviewed and edited the manuscript. IV collated the data, reviewed

Takeda and Medac. JS has received honoraria and/or speakers fees

and edited the manuscript. All other authors collected the data,

from AstraZeneca, Swixx, AbbVie, Janssen, Sanofi and Servie; consults

reviewed and edited the manuscript

for AstraZeneca and Roche. AR has received honoraria/speakers fees from Alexion, Pfizer, Apellis, Sobi and Novartis; is on the advisory board

Acknowledgments

of Alexion, Roche, Sobi, Achillion, Apellis and Novartis; has received

We thank the support of the Severe Aplastic Anemia and Infectious

research funding from Alnylam: RdL has received lecture and research

Disease working parties of the EBMT and the patients for

grants from Alexion, AstraZeneca, Novartis and Pfizer. All other authors

participation.

have no conflicts of interest to disclose. Data-sharing statement Contributions

Datasets are maintained in the EBMT electronic database; data

MG, AR, RD, RC devised the study, collected data, reviewed the

requests will be considered on request from the corresponding

data, wrote and edited the manuscript. DE analyzed the data,

author and the Severe Aplastic Anaemia Working Party of EBMT.

References 1. Severe Aplastic Anemia Working Party European Group for Blood and Marrow Transplantation (EBMT). COVID-19 - bone marrow failure and PNH recommendations on behalf of the Severe Aplastic Anemia Working Party, European Group for Blood and Marrow Transplantation (EBMT) 2021: https://www.ebmt.org/sites/default/files/202003/SAAWP_COVID_Recommendations.pdf. Accessed 1st June 2023. 2. Angelini P, Kavadas F, Sharma N, et al. Aplastic anemia following varicella vaccine. Pediatr Infect Dis J. 2009;28(8):746-748. 3. Hendry CL, Sivakumaran M, Marsh J, Gordon-Smith EC. Relapse of severe aplastic anaemia after influenza immunization. Br J Haematol. 2002;119(1):283-284. 4. Donnini I, Scappini B, Guidi S, Longo G, Bosi A. Acquired severe aplastic anemia after H1N1 influenza virus vaccination successfully treated with allogeneic bone marrow transplantation. Ann Hematol. 2012;91(3):475-476. 5. Arnold A, Kelly R, Munir T, et al. Thrombotic events with neisseria meningitidis vaccination in patients with paroxysmal nocturnal hemoglobinuria, UK experience. Blood. 2020;136(Suppl 1):S35-36. 6. COVID-19 vaccine in patients with haematological disorders. British Society for Haematology. https://b-sh.org.uk/media/19195/haematology-covid-19-v10-vaccination-sta tement-231220.pdf. Accessed 1st June 2023. 7. Avenoso D, Marsh JC, Potter V et al. SARS-CoV-2 infection in aplastic anemia. Haematologica. 2021;107(2):541-543.

8. Pike A, Muus P, Munir T, et al. COVID-19 infection in patients on anti-complement therapy: the Leeds National Paroxysmal Nocturnal Haemoglobinuria service experience. Br J Haematol. 2020;191(1):e1-e4. 9. Barcellini W, Fattizzo B, Giannotta JA, et al. COVID-19 in patients with paroxysmal nocturnal haemoglobinuria: an Italian multicentre survey. Br J Haematol. 2021;194(5):854-856. 10. Chen CY, Chen TT, Hsieh CY, Lien MY, Yeh SP, Chen CC. Case reports of management of aplastic anemia after COVID-19 vaccination: a single institute experience in Taiwan. Int J Hematol. 2023;117(1):149-152. 11. Cecchi N, Giannotta JA, Barcellini W, Fattizzo B. A case of severe aplastic anaemia after SARS-CoV-2 vaccination. Br J Haematol. 2022;196(6):1334-1336. 12. Ljung R, Sundström A, Grünewald M, et al. The profile of the COvid-19 VACcination register SAFEty study in Sweden (CoVacSafe-SE). Ups J Med Sci. 2021;126:e8136. 13. Röth A, Bertram S, Schroeder T, et al. Acquired aplastic anemia following SARS-CoV-2 vaccination. Eur J Haematol. 2022;109(2):186-194. 14. Rajput RV, Ma X, Boswell KL, et al. Clinical outcomes and immune responses to SARS-CoV-2 vaccination in severe aplastic anaemia. Br J Haematol. 2022;199(5):679-687. 15. Giannotta JA, Fattizzo B, Bortolotti M, et al. SARS-CoV-2 vaccination in patients with paroxysmal nocturnal hemoglobinuria: an Italian multicenter survey. Am J Hematol. 2022;97(7):E229-E232.

Haematologica | 109 January 2024

286

LETTER TO THE EDITOR

Molecular predictors of response and survival following IDH1/2 inhibitor monotherapy in acute myeloid leukemia Mutations in the isocitrate dehydrogenase gene (IDH1/2) are present in approximately 20% of patients with acute myeloid leukemia (AML), with a higher incidence in older patients.1 The prognostic impact of IDH1/2 mutations is context-dependent, given their frequent association with diploid or other intermediate-risk karyotype, FLT3-internal tandem duplications (ITD) and NPM1 mutations.2 Three IDH inhibitors are currently Food and Drug Administration-approved for IDH-mutated AML; ivosidenib and olutasidenib (IDH1 inhibitors) and enasidenib (IDH2 inhibitor). Ivosidenib ± azacitidine is approved for frontline therapy and all three IDH inhibitors for relapsed/refractory disease. Complete remission with (CR) or without (CRi) blood count recovery rates in newly diagnosed IDH-mutated AML patients treated with ivosidenib or enasidenib were reported at 42.4%,3 and 21%,4 respectively; the corresponding CR/CRi rate in the relapsed/refractory setting were 30.4% and 26.6%.5,6 Hypomethylating agent and venetoclax combination therapy (HMA-Ven) is also a therapeutic consideration for elderly/unfit IDH-mutated AML.7 There is limited comparative data on the outcomes of patients treated with IDH inhibitors versus HMA-Ven. We have previously described our experience with HMA-Ven in treatment-naïve and relapsed/refractory AML and identified molecular predictors of response and survival.8,9 In the current study, our primary objective was to determine the impact of mutations and karyotype on response and survival in IDH-mutated AML patients receiving IDH1/2 inhibitor monotherapy in routine clinical practice and retrospectively compare the findings with those of IDHmutated patients treated with HMA-Ven. The current study includes a total of 59 consecutive patients with IDH-mutated AML treated with single-agent IDH1/2 inhibitor (ivosidenib or enasidenib), outside of clinical trials, at the Mayo Clinic (Rochester MN, Jacksonville FL, Scottsdale, AZ), between 2017 and 2023. Study patients were retrospectively recruited after Institutional Review Board approval. Cytogenetic and molecular studies were performed at the time of diagnosis by conventional karyotype and next-generation sequencing (NGS) of a 42gene panel, respectively. Four-gene panel NGS (FLT3, IDH1/2 and TP53) was obtained at relapse. Disease risk and response were assessed according to the 2022 European LeukemiaNet (ELN) criteria.10 Timing of response assessment was based on treating physician discretion. Determinants of treatment response were assessed by Χ2 or Fisher’s exact test for nominal data and Wilcoxon rank sum test for continuous variables. Overall survival was

calculated from the time of initiation of IDH1/2 inhibitor to last follow-up or death without censoring for transplant and evaluated by the Kaplan–Meier method. Analyses were performed using JMP Pro 16.0.0 software package, SAS Institute, Cary, NC. A total of 59 patients with IDH-mutated AML (median age 74 years, range 54-91; 59% males; 54% secondary or therapy-related) received ivosidenib (n=16) or enasidenib (n=43), of which 11 (19%) and 48 (81%) patients were treated in the frontline and relapsed/refractory setting, respectively. Patients with relapsed/refractory disease had received either one (n=24), two (n=15), three (n=6), or four (n=3) prior therapies which included cytarabine + idarubicin (7+3) (n=24), HMA-Ven (n=11), mitoxantrone, etoposide, cytarabine (MEC) (n=5), liposomal daunorubicin/cytarabine (n=4), cladribine, cytarabine, granulocyte colony-stimulating factor (G-CSF), mitoxantrone (CLAG-M)/fludarabine, cytarabine, G-CSF, idarubicin (FLAG-IDA) (n=4), and 7+3+ midostaurin (n=3). Seven patients had relapsed following allogeneic hematopoietic stem cell transplant (alloHSCT). ELN cytogenetic risk was evaluable in 57 patients and included intermediate (75%, n=43) or adverse (25%, n=14) risk. Mutations detected included DNMT3A in 22 patients (37%), SRSF2 in 21 (36%), RUNX1 in 15 (25%), ASXL1 in 12 (20%), BCOR in seven (12%), NPM1 in seven (12%), K/NRAS in six (10%), FLT3-ITD in six (10%), and TP53 in four (7%). A comparison of IDH1- versus IDH2-mutated patients revealed a higher incidence of RUNX1 mutations (44% vs. 19%; P=0.05) and adverse karyotype (50% vs. 14%; P=0.01) in IDH1-mutated patients. Treatment-emergent toxicities included differentiation syndrome (n=17), hyperbilirubinemia (n=11), and Qtc prolongation (n=6); treatment was discontinued due to toxicity in six patients. Table 1 provides information regarding patient characteristics at the time of initiation of IDH inhibitor, response rates, and overall outcome. Fifteen (25%) patients achieved CR (n=10; 17%) or CRi (n=5; 8%); median time to response was 2.2 months (range, 1.07.1) and median response duration 3.6 months (range, 1.033). In addition, two (3%) patients experienced partial remission and eight (14%) hematological improvement. Measurable residual disease (MRD) assessement was performed in a subset of patients; MRD negativity was confirmed in three (75%) of four patients, evaluable by multiparametric flow cytometry, and in one (17%) of six patients, evaluable by IDH mutation analysis. Among the 15 patients with CR/CRi, relapse was documented in four (27%), and six (40%) patients were bridged to alloHSCT.

Haematologica | 109 January 2024

287

LETTER TO THE EDITOR Table 1. Clinical characteristics at time of treatment initiation with single agent IDH1/2 inhibitor for 59 patients with acute myeloid leukemia stratified by achievement of complete response or complete response with incomplete count recovery. Variables

All patients N=59

Patients in CR/CRi N=15 (25%)

Patients not in CR/CRi N=4 (75%)

P/Multivariate P

Age in years, median (range) >60 years, N (%)

74 (54-91) 53 (90)

71 (54-84) 12 (80)

74 (55-91) 41 (93)

0.37 0.17

Male, N (%)

35 (59)

9 (60)

26 (59)

0.95

AML type, N (%) De novo Secondary or therapy-related

27 (46) 32 (54)

6 (40) 9 (60)

21 (48) 23 (52)

0.60

Treatment setting, N (%) Upfront Relapsed/refractory Number of prior treatments, median (range)

11 (19) 48 (81) 1 (1-4)

3 (20) 12 (80) 1.5 (1-4)

8 (18) 36 (82) 1 (1-4)

0.88

16 (27) 43 (73) N=52 38 (5-50)

1 (7) 14 (93) N=14 38 (10-50)

15 (34) 29 (66) N=38 36 (5-49)

0.02

Hemoglobin, g/dL, median (range)

8.8 (6-14.7)

8.6 (7.5-14.7)

8.8 (6-12.8)

0.88

Leukocyte count, x109/L, median (range) Leukocyte count >2x109/L*, N (%)

1.92 (0.3-71.9) 27 (46)

1.05 (0.3-6.5) 3 (20)

2.35 (0.3-71.9) 24 (55)

0.01/0.06 0.02/0.12

Platelet count, x109/L, median (range)

53 (5-510)

74 (12-360)

46 (5-510)

0.79

Circulating blasts %, median (range) Circulating blasts >5%,* N (%)

10 (0-83) 32 (54)

2 (0-30) 4 (27)

16 (0-83) 28 (64)

0.01/0.09 0.01/0.07

2022 ELN cytogenetic risk stratification, N (%) Intermediate Adverse

N=57 43 (75) 14 (25)

N=15 14 (93) 1 (7)

N=42 29 (69) 13 (31)

0.04/0.01

Mutations on NGS, N (%) DNMT3A SRSF2 RUNX1 ASXL1 BCOR NPM1 K/NRAS FLT3-ITD TP53 TET2

22 (37) 21 (36) 15 (25) 12 (20) 7 (12) 7 (12) 6 (10) 6 (10) 4 (7) 4 (7)

8 (53) 5(33) 7 (47) 1 (7) 5 (33) 3 (20) 2 (13) 2 (13) 0(0) 1 (7)

14 (32) 16 (36) 8 (18) 11 (25) 2 (5) 4 (9) 4 (9) 4 (9) 4 (9) 3 (7)

0.14 0.83 0.03/0.04 0.09 0.01/0.01 0.28 0.65 0.65 0.12 1.0

IDH mutation, N (%) IDH-1 IDH-2 IDH (VAF), median (range)

0.86

0.38

Cut-off was determined by receiving operating characteristic (ROC) analysis. CR: complete response; Cri: CR with incomplete count recovery; AML: acute myeloid leukemia; HMA: hypomethylating agent; Ven: venetoclax; VAF: variant allele frequency; ELN: European LeukemiaNet; NGS: next-generation sequencing; ITD: internal tandem dublication.

Four of the five remainder patients are alive and in continuing response for a median duration of 22.2 months (range, 2.1-44.4), while one patient in ongoing response for 6.9 months, has died from sepsis. An additional two patients, one with hematological response and one non-responder, proceeded to alloHSCT following HMA-Ven and cladribine-cytarabine-Ven, respectively. CR/CRi rates were higher with IDH2 versus IDH1 mutation (33% vs. 6%; P=0.02). Among IDH2-mutated patients, CR/CRi was more likely with R140 versus R170 mutation (14/34, 41% vs. 0/6, 0%; P=0.02). In addition, CR/CRi rates were higher with BCOR mutation (CR/CRi 71% vs. 19%; P=0.01), and RUNX1

mutation (47% vs. 18%; P=0.03); CR/CRi rates were lower in the presence of ELN adverse karyotype (7% vs. 33%; P=0.04), ASXL1 (8% vs. 30%; P=0.09) or TP53 mutations (0% vs. 27%; P=0.12). Multivariable analysis confirmed superior response in the presence of BCOR (P=0.01; overall response to odds ratio [OR]=19.5) or RUNX1 mutations (P=0.04; OR 5) and inferior response in the presence of ELN adverse karyotype (P=0.01; OR=13.5). CR/CRi rates were not significantly different in the frontline versus relapsed/refractory setting (27% vs. 25%; P=0.88), de novo versus secondary AML (22% vs. 29%; P=0.55), presence or absence of NPM1 (43% vs. 23%; P=0.28), FLT3-ITD (33% vs.

Haematologica | 109 January 2024

288

LETTER TO THE EDITOR Table 2. Predictors of complete response or complete response with incomplete count recovery and post-treatment survival in 59 patients with acute myeloid leukemia treated with IDH1/2 inhibitor monotherapy. Variables Age

CR/CRi Univariate P CR/CRi rates

CR/CRi Multivariate P/OR

0.37

Overall survival Univariate P Median survival

Overall survival Multivariate P HR (95% CI)

0.27

0.04 7% vs. 33% Presence vs. absence

0.01/13.5

ELN adverse karyotype

0.04 9 vs. 11 months Presence vs. absence

<0.01 3.2 (1.5-6.7)

0.01 71% vs 19% Presence vs. absence

0.01/19.5

BCOR mutation

0.01 9.5 vs. NR Absence vs. presence

<0.01 4.8 (1.4-16.5)

0.04/5.0

RUNX1 mutation

0.03 47% vs. 18% Presence vs. absence

0.05 9.5 vs. 14.6 months Absence vs. presence

0.04 2.4 (1.1-5.4)

ASXL1 mutation

0.09 8% vs. 30% Presence vs. absence

NPM1 mutation

0.28

0.07

FLT3-ITD mutation

0.65

0.42

0.09 -

TP53 mutation

0.12 0% vs. 27% Presence vs. absence

0.15 -

K/NRAS mutation

0.65

0.87

Presence of CR/CRi

<0.01 NR vs. 9.3 months

AlloHSCT after IDH inhibitor

<0.01 NR vs. 9.5 months

NR: not reached; NA: not applicable; CR: complete response; Cri: CR with incomplete count recovery; OR: odds ratio; HR: hazard ratio; CI: confidence interval; ELN: European LeukemiaNet; ITD: internal tandem duplication; alloHSCT: allogeneic stem cell transplant.

25%; P=0.65), DNMT3A (36% vs. 19%; P=0.14), TET2 (25% vs. 25%; P=1.0), SRSF2 (24% vs. 26%; P=0.83), or K/NRAS (33% vs. 25%; P=0.65) mutations (Table 2). Response was also not influenced by IDH mutation variant allele frequency (median-38%; range, 5-50; P=0.38) and number of prior therapies (P=0.86). Only one (9%) of 11 patients previously treated with HMA-Ven (n=8, frontline HMA-Ven) responded to IDH inhibitor therapy compared to 14 of 48 (29%) patients without prior HMA-Ven exposure (P=0.13). At a median follow-up of 9.5 months (range, 0.4-70), from initiation of IDH inhibitor, 43 (73%) patients have died and eight (14%) underwent alloHSCT. Median survival following IDH inhibitor therapy was 10.4 months (range, 3.4-33.1) and superior in IDH2- versus IDH1-mutated patients (13.1 vs. 5.1 months; P<0.01), in the presence versus absence of CR/CRi (not reached [NR] vs. 9.3 months; P<0.01) and in patients receiving alloHSCT (NR vs. 9.5 months; P<0.01). The survival differences in IDH1- versus IDH2-mutated patients remained significant after accounting for karyotype, mutations, response and alloHSCT. Univariate survival analysis identified adverse karyotype (P=0.04), absence of BCOR (P=0.01) and absence of RUNX1 mutations (P=0.05) as predictors of inferior survival. Multivariable analysis,

adjusted for alloHSCT, confirmed the survival impact of BCOR and RUNX1 mutations and adverse karyotype; corresponding hazard ratio (HR) and 95% confidence interval (CI) were HR=4.8, 95% CI: 1.4-16.5; HR=2.4, 95% CI: 1.1-5.4; HR=3.2, 95% CI: 1.5-6.7 (Table 2). A three-tiered survival model was subsequently generated by using HR-weighted risk point assignment; two points for absence of BCOR mutation and one point each for absence of RUNX1 mutation and presence of adverse karyotype, resulting in high (4 points, n=7; median survival 2.4 months), intermediate (2-3 points, n=45, median 11 months) and low (0–1 point, n=5; median NR) risk categories (P<0.01; Figure 1A). The aforementioned observations were confirmed when survival analysis was restricted to IDH2-mutated patients; the small number of IDH1-mutated patients precluded similar analysis. Response rates and survival in the 59 IDH-mutated patients treated with IDH inhibitors were retrospectively compared to IDH-mutated Mayo Clinic patients treated with HMA-Ven (n=32; median age 69 years; treatmentnaïve n=20 and relapsed/refractory n=12); CR/CRi rates were not different in HMA-Ven treated frontline versus relapsed/refractory patients (70% vs. 75%; P=0.76). Similarly,

Haematologica | 109 January 2024

289

LETTER TO THE EDITOR CR/CRi rates were comparable in IDH inhibitor-treated frontline versus relapsed/refractory patients (27% vs. 25%; P=0.87). CR/CRi rates were superior with HMA-Ven compared to IDH inhibitor (72% vs. 25%; P<0.01) while survival was similar between the two treatment regimens (17.8 vs. 10.4 months; P=0.24; Figure 1B). CR/CRi with HMA-Ven was higher in the presence of SRSF2 mutation (100% vs. 65%; P=0.03) while not influenced by adverse karyotype (63% vs. 73%; P=0.59) or RUNX1 (57% vs. 76%; P=0.34) or BCOR

mutations (100% vs. 69%; P=0.14). The current study confirms activity of IDH inhibitor monotherapy in treatment-naïve and relapsed/refractory IDHmutated AML3-6 and unveils unique molecular predictors of response and survival. The study also provides comparative information in IDH-mutated patients treated with HMA-Ven. Salient observations include the favorable impact of BCOR and RUNX1 mutations on IDH inhibitor treatment response and survival and were most apparent in

Figure 1. Overall survival in patients with IDH1/2-mutated acute myeloid leukemia. (A) Overall survival in 57 patients with acute myelod leukemia (AML) treated with IDH inhibitor, stratified by hazard ratio (HR)-weighted scoring system, absence of BCOR mutation, HR=4.8, 95% confidence interval (CI): 1.4-16.5; presence of adverse karyotype, HR=3.2, 95% CI: 1.56.7; and absence of RUNX1 mutation, HR=2.4, 95% CI: 1.1-5.4, allocating 2 adverse points for absence of BCOR mutation, 1 adverse point for adverse karyotype, and 1 adverse point for absence of RUNX1 mutation. Median overall survival stratified by low risk (0-1 points), intermediate risk (2-3 points) and high risk (4 points) is shown. (B) Overall survival in 91 patients with IDH-mutated AML treated with either IDH inhibitor (N=59) or hypomethylating agent + venetoclax (N=32). alloHSCT: allogeneic stem cell transplantation; RR: relapsed refractoryrefractory; yr: years.

Haematologica | 109 January 2024

290

LETTER TO THE EDITOR IDH-2 mutated patients. In previously reported enasidenib clinical trials, responses were negatively affected by the presence of FLT3 (overall response rate [ORR] 8.3%) or NRAS (ORR 29%) mutations, and high co-mutational burden (≥ 6 vs. ≤3 mutations; ORR 31% vs. 55%);11 in addition, as was the case in the current study, adverse karyotype was associated with inferior response (ORR 18% vs. 46%) and survival (ORR 7 vs. 9.3 months).11 In ivosidenibtreated patients, treatment response was lower in the presence of receptor tyrosine kinase pathway mutations (CR/CRh 7% vs. 43%) and higher with JAK2 mutation (CR/CRh 64% vs. 32%).12 The observations from the current study are particularly relevant in light of the adverse prognosis assigned to BCOR and RUNX1 mutations, in the latest ELN 2022 risk stratification, which was, however, derived from intensively treated patients.10 Our findings differ from a prior study on IDH-mutated AML patients treated with IDH inhibitors (n=60), in which RUNX1 mutation was associated with inferior CR rate, in addition, among patients in whom pre-treatment and relapsed samples were analyzed (n=18), BCOR and RUNX1 mutations were frequently acquired at the time of relapse in four and three cases, respectively.13 The discrepancies stem from key differences in the study populations, unlike the current study, the former study included patients with myelodysplastic syndrome and chronic myelomonocytic leukemia (n=5) and was enriched with patients harboring complex karyotype (22% vs. 11%).13 It should be noted that response prediction is different than survival prediction and frequency of mutations at time of relapse might actually suggest that a higher proportion of patients with the specific mutations achieved response and thus were at risk for relapse. In other words, one has to respond first, in order to relapse. Regardless, the current study provides a practical prognostic model for use in IDH-mutated patients with AML receiving IDH inhibitor therapy. The study also suggests superiority of HMA-Ven to IDH inhibitor, in IDH-mutated AML, irrespective of karyotype or mutational profile. The favorable responses seen with HMA-Ven were also evident in a prior study of 81 IDH1/2-mutated AML patients treated with HMA-Ven with reported CR/CRi rate of 79%.14 Recently, “doublet therapy” with IDH inhibitor + HMA and “triplet therapy” with IDH inhibitor + Ven + HMA have garnered interest due to higher responses with composite CR rates of 53%,15 and 90%,16 respectively. Nonetheless,

controlled studies are needed to determine the optimal combination therapy and associated molecular determinants of response and survival.

Authors Naseema Gangat,1* Kristen McCullough,1* Maymona Abdelmagid,1 Omer Karrar,1 Marissa Powell,2 Aref Al-Kali,1 Hassan Alkhateeb,1 Kebede Begna,1 Abhishek Mangaonkar,1 Antoine Saliba,1 Mehrdad Hefazi Torghabeh,1 Mark Litzow,1 William Hogan,1 Mithun Shah,1 Mrinal Patnaik,1 Animesh Pardanani,1 Talha Badar,3 James Foran,3 Jeanne Palmer,4 Lisa Sproat,4 Cecilia Arana Yi4 and Ayalew Tefferi1 Division of Hematology, Mayo Clinic, Rochester, MN; 2Henry Ford

Health, Michigan, MI; 3Division of Hematology, Mayo Clinic, Jacksonville, FL and 4Division of Hematology, Mayo Clinic, Scottsdale, AZ, USA NG and KM contributed equally as first authors.

Correspondence: N. GANGAT - gangat.naseema@mayo.edu https://doi.org/10.3324/haematol.2023.283732 Received: June 11, 2023. Accepted: July 27, 2023. Early view: August 3, 2023. ©2024 Ferrata Storti Foundation Published under a CC BY-NC license Disclosures No conflicts of interest to disclose. Contributions KM, NG and AT designed the study, collected data, performed analysis and co-wrote the paper. MA, OK and MP collected data. AA, HA, KHB, AM, AS, MH, MRL, WH, MS, MMP, AP, TB, JF, JP, LS and CA contributed patients. All authors reviewed and approved the final draft of the paper. Data-sharing statement Data will be shared upon reasonable request to the corresponding author.

References 1. Issa GC, DiNardo CD. Acute myeloid leukemia with IDH1 and IDH2 mutations: 2021 treatment algorithm. Blood Cancer J. 2021;11(6):107. 2. DiNardo CD, Ravandi F, Agresta S, et al. Characteristics, clinical

outcome, and prognostic significance of IDH mutations in AML. Am J Hematol. 2015;90(8):732-736. 3. Roboz GJ, DiNardo CD, Stein EM, et al. Ivosidenib induces deep durable remissions in patients with newly diagnosed IDH1-

Haematologica | 109 January 2024

291

LETTER TO THE EDITOR mutant acute myeloid leukemia. Blood. 2020;135(7):463-471. 4. Pollyea DA, Tallman MS, de Botton S, et al. Enasidenib, an inhibitor of mutant IDH2 proteins, induces durable remissions in older patients with newly diagnosed acute myeloid leukemia. Leukemia. 2019;33(11):2575-2584. 5. DiNardo CD, Stein EM, de Botton S, et al. Durable remissions with ivosidenib in IDH1-mutated relapsed or refractory AML. N Engl J Med. 2018;378(25):2386-2398. 6. Stein EM, DiNardo CD, Pollyea DA, et al. Enasidenib in mutant IDH2 relapsed or refractory acute myeloid leukemia. Blood. 2017;130(6):722-731. 7. Kantarjian H, Kadia T, DiNardo C, et al. Acute myeloid leukemia: current progress and future directions. Blood Cancer J. 2021;11(2):41. 8. Gangat N, Johnson I, McCullough K, et al. Molecular predictors of response to venetoclax plus hypomethylating agent in treatment-naive acute myeloid leukemia. Haematologica. 2022;107(10):2501-2505. 9. Johnson IM, Ilyas R, McCullough K, et al. Molecular predictors of response and survival in patients with relapsed/refractory acute myeloid leukemia lollowing venetoclax plus hypomethylating agent therapy. Blood. 2022;140(Suppl 1):S3233-3234. 10. Arber DA, Orazi A, Hasserjian RP, et al. International Consensus

Classification of Myeloid Neoplasms and Acute Leukemias: integrating morphologic, clinical, and genomic data. Blood. 2022;140(11):1200-1228. 11. Stein EM, DiNardo CD, Fathi AT, et al. Molecular remission and response patterns in patients with mutant-IDH2 acute myeloid leukemia treated with enasidenib. Blood. 2019;133(7):676-687. 12. Choe S, Wang H, DiNardo CD, et al. Molecular mechanisms mediating relapse following ivosidenib monotherapy in IDH1mutant relapsed or refractory AML. Blood Adv. 2020;4(9):1894-1905. 13. Wang F, Morita K, DiNardo CD, et al. Leukemia stemness and co-occurring mutations drive resistance to IDH inhibitors in acute myeloid leukemia. Nat Commun. 2021;12(1):2607. 14. Pollyea DA, DiNardo CD, Arellano ML, et al. Impact of venetoclax and azacitidine in treatment-naïve patients with acute myeloid leukemia and IDH1/2 mutations. Clin Cancer Res. 2022;28(13):2753-2761. 15. Montesinos P, Recher C, Vives S, et al. Ivosidenib and azacitidine in IDH1-mutated acute myeloid leukemia. N Engl J Med. 2022;386(16):1519-1531. 16. Lachowiez CA, Loghavi S, Zeng Z, et al. A phase Ib/II study of ivosidenib with venetoclax +/- azacitidine in IDH1-mutated myeloid malignancies. Blood Cancer Discov. 2023;4(4):276-293.

Haematologica | 109 January 2024

292

LETTER TO THE EDITOR

Myeloid lineage switch in KMT2A-rearranged acute lymphoblastic leukemia treated with lymphoid lineagedirected therapies Leukemias usually commit to a specific lineage in their development. However, lineage switch (LS), which is the transformation of the original leukemic clone into a different cellular lineage, is associated with a poor survival and usually occurs in leukemias with rearrangements of the KMT2A gene (KMT2Ar), which already convey a dismal prognosis.1,2 LS is rare, but its incidence may increase with the use of lymphoid lineage-targeting treatments, such as CD19-directed therapies.3–7 Herein, we report one of the largest case series of patients with B-cell acute lympho-

blastic leukemia (B-ALL) with KMT2Ar that had LS to a myeloid phenotype, all after receiving CD19- or CD22-directed therapy. The patients’ baseline characteristics and the evolution of their diseases are detailed in Table 1 and Figure 1. Immunophenotype changes between diagnosis and LS are detailed in Figure 2. Detailed information regarding mutations tested in these patients are detailed in the Online Supplementary Tables S1 and S2. This study received ethics approval from the University of Texas MD Anderson Cancer Institutional Review Board (IRB) and was

Table 1. Characteristics of each patient’s B-cell acute lymphoblastic leukemia (at diagnosis) and acute myeloid leukemia (after lineage switch). Patient (sex, age in years)

#1 (F, 43)

Acute lymphoblastic leukemia Cytogenetics and FISH 46,XX,t(4;11)(q21;q23)[12]/ 47,idem,+8[2]/46,XX[6]

Acute myeloid leukemia (LS)

Mutations

Cytogenetics and FISH

CNS

TP53 p.Y163SA

61~65,XX,-X,+add(1)(p13),3,der(4)t(4;11)(q21;q23), t(4;11),+del(6)(q15q21),-10,-11,-17,-17,17,+1~6mar[cp12]/46,XX[8]

Mutations

TP53 p.Y163S & p.I251fsA

FISH positive for KMT2Ar FISH positive for KMT2Ar

#2 (M, 34)

46,XY,t(4;11)(q21;q23)[11]/ 46,idem,add(18)(p11.2)[8]/ 46,XY[1]

TP53 p.R282PA

FISH positive for KMT2Ar

#3 (F, 27)

#4 (M, 0)

#5 (F, 60)

46,XY,t(4;11)(q21;q23) [13]/46,XX[7] Not assessed

37~45,XY,add(2)(q33),add(3)(p13), add(3)(p21),t(4;11)(q21;q23), add(9)(p22),-14,add(15)(q24),16,add(17)(p11.2),-18,i(18)(q10),-19,21,+1~5mar[cp10]/46,XY[10] FISH positive for KMT2Ar 53,XX,+der(4)t(4;11)(q21;q23),t(4;11),+6,+7,+8, +13,+19,+20[6]//46,XY[14]

FISH positive for KMT2Ar 46,XY,t(4;11)(q21;q23)[15]/46,XY[5] FISH positive for KMT2Ar 46,XX,t(11;19)(q23;p13.3)[14]/46,XX[6] FISH positive for KMT2Ar

TP53 p.R282PA

Not assessed

FISH positive for KMT2Ar CCND3, FLT3, CD28

KRAS p.G12DB

Cytogenetics not available + FISH positive for KMT2Ar

PTPN11 p.D61VA

46,XX,+i(8)(q10),t(11;19)(q23;p13.3),16[12]/46,XX,+i(8)(q10),t(11;19),add(14)(p11.2), -16[cp6]/47,XX,+X,t(11;19)[2] Not assessed FISH not available

#6 (M, 44)

46,XY,t(4;11)(q21;q23) [20] Not available

84~87,XXYY,-1,der(4)t(4;11)(q21;q23),t(4;11),5,-10,-11,-21[cp20]

FISH positive for KMT2Ar

NRAS p.Q61HA

FISH positive for KMT2Ar

Mutation/s detected using a 81-gene next-generation sequencing panel. BMutation/s detected using a 28-gene next-generation sequencing panel. Panel gene coverage is detailed in the Online Supplementary Appendix. LS: lineage switch; CNS: central nervous system involvement; FISH: fluorescence in situ hybridization; KMT2Ar: KMT2A rearrangement; F: female; M: male. A

Haematologica | 109 January 2024

293

LETTER TO THE EDITOR conducted in accordance with the Declaration of Helsinki. Patient 1 was a 43-year-old woman diagnosed with B-ALL with t(4:11)/KMT2A::AFF1 rearrangement. The patient received treatment with hyper-CVAD (hyper-fractionated cyclophosphamide, vincristine, doxorubicin and dexamethasone). She achieved complete remission (CR) but had disease bone marrow relapse after the eighth course. She then received mini-hyper-CVD (mini-hyper-fractionated cyclophosphamide, vincristine, and dexamethasone) with inotuzumab but did not achieve CR. A second salvage therapy consisting of fludarabine, idarubicin, cytarabine, pegylated asparaginase, and blinatumomab elicited CR, but the patient had a relapse with LS after the second cycle of blinatumomab. She then received salvage chemotherapy with venetoclax and later azacytidine plus ipilimumab and nivolumab but had no response.

Patient 2 was a 34-year-old man diagnosed with B-ALL with t(4:11)/KMT2A::AFF1 rearrangement. He received hyper-CVAD and initially achieved CR with negative measurable residual disease by multiparameter flow cytometry (MRDfc) but had relapse after the fourth course. He then received three courses of dose-reduced hyper-CVD with inotuzumab and achieved CR with negative MRDfc before proceeding to allogeneic stem cell transplantation (allo-SCT). Two months after allo-SCT, the patient had a mixed relapse with 57% lymphoid blasts in the bone marrow and two myeloid sarcomas in the left tonsillar region. He was treated with fludarabine, idarubicin, cytarabine, venetoclax, and blinatumomab and initially had CR with negative MRDfc and a reduction in sarcoma size. However, his disease progressed rapidly; a full LS with full bone marrow involvement by myeloid blasts was accompanied

Figure 1. Summary of therapies received. Patient treatments and responses over time. Blue indicates therapies for B-cell acute lymphoblastic leukemia; green indicates therapies for acute myeloid leukemia after lineage switch (LS; red). Allo-SCT: allogeneic stem cell transplantation; AraC: cytarabine; Aza: azacytidine; Blina: blinatumomab; CIA: clofarabine, idarubicin, and cytarabine; CLAG: cladribine, cytarabine, and granulocyte colony-stimulating factor; CLIA: cladribine, idarubicin, and cytarabine; CR: complete remission; Dauno: daunorubicin; Dec: decitabine; FIA: fludarabine, idarubicin, and cytarabine; HCVAD: hyperfractionated cyclophosphamide, vincristine, doxorubicin, and dexamethasone; IDA: idarubicin; iMen: menin inhibitor; Ino: inotuzumab ozogamicin; Ipi: ipilumumab; mHCVD: mini-hyperfractionated cyclophosphamide, vincristine, and dexamethasone; MRDfc: minimal residual disease by flow cytometry; Nivo: nivolumab; NR: no response; Peg: pegylated asparaginase; VCR: vincristine; Ven: venetoclax; VP16: etoposide. Haematologica | 109 January 2024

294

LETTER TO THE EDITOR by clinical deterioration, and the patient received no further treatment. Patient 3 was a 27-year-old woman diagnosed with B-ALL with t(4:11)/KMT2A::AFF1 rearrangement. The patient initially received induction therapy based on the CALGB 10403 pediatric protocol and then received dose-dense mini-hyper-CVD with inotuzumab and blinatumomab. After three cycles of therapy, she achieved CR with negative MRDfc and proceeded to allo-SCT. Five months after allo-SCT, she had a relapse with LS with multiple myeloid sarcomas without bone marrow involvement. She received FLAG-IDA (fludarabine, cytarabine, idarubicin, and G-CSF) with venetoclax but had no response. The patient is currently enrolled in a clinical trial combining chemotherapy with a menin inhibitor; her sarcomas shrunk after the first cycle of treatment. Patient 4 was a 2-month-old male diagnosed with B-ALL with t(4:11)/KMT2A::AFF1 rearrangement. The patient was treated with the AALL0631 pediatric protocol but had persistent MRDfc and received one course of blinatumomab. He then relapsed with LS. He received cytarabine with

etoposide and achieved CR. He started maintenance therapy with cytarabine and venetoclax as a bridge to alloSCT but had relapse with LS. The patient was enrolled in a clinical trial with a menin inhibitor but did not achieve response. Patient 5 was a 60-year-old woman diagnosed with B-ALL with 46,XX,t(11;19)(q23;p13.3)/KMT2A::MLLT1 rearrangement. The patient received hyper-CVAD but had progressive disease after the first course and, therefore, received blinatumomab, which elicited CR with negative MRDfc. However, the patient had a relapse and was enrolled in a clinical trial of the CD19-directed antibody–drug conjugate ADCT-402. After one course, the patient had a relapse with LS and received no further therapy. Patient 6 was a 44-year-old man diagnosed with B-ALL with t(4:11)/KMT2A::AFF1 rearrangement. The patient received three courses of hyper-CVAD; he initially achieved CR with MRDfc but had a relapse after the last cycle. He then received one cycle of blinatumomab and had a relapse with LS. He was treated with decitabine plus chemotherapy and venetoclax and had CR after two

Figure 2. Immunophenotypic changes between diagnosis and lineage switch. Immunophenotypes of the leukemic cells at different time points. LS: lineage switch; AML: acute myeloid leukemia; ALL: acute lymphoblastic leukemia. Haematologica | 109 January 2024

295

LETTER TO THE EDITOR cycles before proceeding to allo-SCT. However, he had a relapse 30 days after allo-SCT. He received decitabine, venetoclax, and sorafenib but did not achieve CR. LS is rare, occurring in about 6% of all relapsed leukemias in childhood.2 In our cohort, this phenomenon occurred in 7% of patients with KMT2Ar B-ALL treated in our institution. LS occurs primarily in KMT2Ar leukemias, usually in pediatric patients with ALL with t(4;11)/KMT2A::AFF1 rearrangement,4–6 this last characteristic present in five of the six patients presented here. Patients with LS have poor prognosis, and current treatments do not have a high rate of success.4 Therefore, a better understanding of the mechanisms of LS is needed to develop strategies to prevent and treat it. The specific mechanism of LS in KMT2Ar leukemias is unclear. Potential mechanisms include bipotential progenitors, cellular reprograming, de-differentiation, and clonal selection.8 Tirtakusuma et al.,9 investigating the mechanisms of LS in KMT2A::AFF1 leukemias, found that LS is related to the rewiring of gene regulatory networks and to changes in chromatin accessibility, with myeloid relapses recurrently associated with CHD4 gene abnormalities. This suggest that LS is driven and maintained by epigenetic dysregulation. Other studies have shown that the lineage fate of these leukemias is influenced by the bone marrow microenvironment.10 Of note, four of five patients with available cytogenetics at LS acquired a complex karyotype besides KMT2A rearrangement, two of them with a detectable TP53 mutation. A biological hypothesis to be validated is that this evident genetic instability that drives these chromosome abnormalities could also favor a phenotype switch under the selective pressure or B-cell-directed therapies. LS has become a topic of interest among leukemia researchers because it provides a mechanism by which BALL cells can escape directed immunotherapies, which requires to switch the highly effective lymphoid-lineage directed therapy to myeloid lineage-directed therapy. Recent studies have shown that in patients with KMT2Ar leukemias, LS from B-ALL to AML can occur after treatment with blinatumomab or chimeric antigen receptor T cells.3,7,11,12 This is important because the increased use of these treatments, which have selective pressure against lymphoid antigens, may increase the rate of LS. In our cohort, all patients received at least one B-cell–directed immunotherapy, against either CD19 or CD22, and these antigens had disappeared by the time of LS. However, whether these therapies can increase the rate of LS is unclear; for example, a recent study in a large cohort of pediatric patients receiving blinatumomab has not detected any LS.13 It is likely that leukemias presenting with LS have high lineage promiscuity, and B-cell-directed therapies are highly effective at eliminating B-ALL subclones, leaving the myeloid-committed leukemic cells a beneficial environment in

which to proliferate. Continuing to monitor LS rates among patients receiving B-cell–directed therapies, especially those who have leukemias with KMT2Ar, is crucial. Regardless of their phenotype, the leukemias reported herein had the genetic hallmarks of KMT2Ar, which made them potentially sensitive to targeted therapy with menin inhibitors. These recently developed compounds interfere with the interaction between the KMT2A-rearranged protein and its cofactor, menin, thereby impeding the abnormal expression of the HOXA genes responsible for maintaining leukemic cells.14 The KMT2A rearrangement is maintained at the time of LS, therefore there is a rationale for testing the efficacy of menin inhibitors in this situation.15 In our cohort, two patients with relapsed AML arising from B-ALL received menin inhibitor; as of this writing, one has had a response whereas the other did not respond. It is warranted to further explore the efficacy of menin inhibition in KMT2A rearranged leukemias presenting LS. In conclusion, LS is a rare but devastating scenario. Studies to determine whether novel directed therapies against Blineage antigens can increase the incidence of LS in these patients are warranted. Combination therapy of menin inhibitors with targeted therapies could change the treatment paradigm for patients with leukemias with LS.

Authors Alex Bataller,1* Tareq Abuasab,1* David McCall,2 Wei Wang,3 Branko Cuglievan,2 Ghayas C. Issa,1 Elias Jabbour,1 Nicholas Short,1 Courtney D. DiNardo,1 Guilin Tang,3 Guillermo Garcia-Manero,1 Hagop M. Kantarjian1 and Koji Sasaki1 Department of Leukemia, 2Department of Pediatrics and

Department of Hematopathology, the University of Texas MD

Anderson Cancer Center, Houston, TX, USA AB and TA contributed equally as first authors.

Correspondence: K. SASAKI - ksasaki1@mdanderson.org https://doi.org/10.3324/haematol.2023.283705 Received: June 7, 2023. Accepted: August 23, 2023. Early view: August 31, 2023. ©2024 Ferrata Storti Foundation Published under a CC BY-NC license Disclosures GCI discloses consultancy from Novartis, Kura Oncology and

Haematologica | 109 January 2024

296

LETTER TO THE EDITOR Nuprobe; and research funding from Celgene, Kura Oncology,

discloses research funding from Abbvie, Amgen, Ascentage, BMS,

Syndax, Merck, Cullinan and Novartis. EJ discloses research funding

Daiichi Sankyo, Immunogen, Jazz Pharmaceuticals and Novartis; and

from Amgen, Pfizer, Abbvie, Adaptive Biotechnologies, Astex and

honoraria from Abbvie, Amgen, Amphista, Ascentage, Astellas,

Ascentage; and consultancy from Amgen, Pfizer, Abbvie, Takeda,

Biologix, Curis, Ipsen Biopharmaceuticals, KAHR Medical, Novartis,

Adaptive Biotechnologies, Astex, Ascentage, Genentech, Novartis,

Pfizer, Precision Biosciences, Shenzhen Target Rx and Takeda. KS

BMS, Jazz Pharmaceuticals, Hikma Pharmaceuticals and Incyte. NS

discloses research funding from Novartis; consultancy from Novartis;

discloses consultancy from Takeda Oncology, AstraZeneca, Amgen,

honoraria from Otsuka Pharmaceuticals; and membership on an

Novartis and Pfizer; research funding from Takeda Oncology,

entity’s Board of Directors or advisory committees from Novartis,

Astellas and Stemline Therapeutics; and honoraria from Amgen. CDD

Pfizer and Daiichi-Sankyo.

discloses membership on an entity’s Board of Directors or advisory committees from GenMab, GlaxoSmithKline, Kura and Notable Labs;

Contributions

honoraria from Kura, Astellas, Bluebird Bio, Bristol Myers Squibb,

KS, HMK, AB and TA conceived and designed the study. AB, TA, DM

Foghorn, ImmuneOnc, Novartis, Takeda, Gilead and Jazz

and BC collected the clinical data. WW and GT provided and

Pharmaceuticals; is a current holder of stock options in a privately-

interpreted biological data. AB elaborated the figures. GCI, EJ, NS,

held company from Notable Labs; consultancy from Abbvie, Servier;

CDD, GG-M, HMK and KS provided clinical care to the patients. AB

and research funding from Servier, Bristol Myers Squibb, Foghorn,

and TA wrote the manuscript. All authors contributed intellectually

ImmuneOnc, LOXO, Astex, Cleave and Forma. GG-M discloses

to the study and critically reviewed and edited the manuscript.

research funding from Astex Pharmaceuticals, Novartis, Abbvie, BMS, Genentech, Aprea Therapeutics, Curis and Gilead Sciences;

Data-sharing statement

consultancy from Astex Pharmaceuticals, Acceleron Pharma and

The data used for this study is not publicly available in order to pro-

BMS; and honoraria from Astex Pharmaceuticals, Acceleron Pharma,

tect patient confidentiality. Reasonable requests for de-identified

Abbvie, Novartis, Gilead Sciences, Curis, Genentech and BMS. HMK

data should be directed to the corresponding author.

References 1. Winters AC, Bernt KM. MLL-rearranged leukemias - an update on science and clinical approaches. Front Pediatr. 2017;5:4. 2. Stass S, Mirro J, Melvin S, Pui CH, Murphy SB, Williams D. Lineage switch in acute leukemia. Blood. 1984;64(3):701-706. 3. Gardner R, Wu D, Cherian S, et al. Acquisition of a CD19-negative myeloid phenotype allows immune escape of MLL-rearranged B-ALL from CD19 CAR-T-cell therapy. Blood. 2016;127(20):2406-2410. 4. Takeda R, Yokoyama K, Fukuyama T, et al. Repeated lineage switches in an elderly case of refractory B-cell acute lymphoblastic leukemia with MLL gene amplification: a case report and literature review. Front Oncol. 2022;12:799982. 5. Yost CM, Gaikwad AS, Parsons DW, et al. Aberrant leukemiaassociated immunophenotype as potential harbinger of lineage switch in KMT2A-rearranged leukemia: a case series. Leuk Lymphoma. 2020;61(14):3515-3518. 6. Rossi JG, Bernasconi AR, Alonso CN, et al. Lineage switch in childhood acute leukemia: an unusual event with poor outcome. Am J Hematol. 2012;87(9):890-897. 7. Rayes A, McMasters RL, O’Brien MM. Lineage switch in MLLrearranged infant leukemia following CD19-directed therapy. Pediatr Blood Cancer. 2016;63(6):1113-1115. 8. Dorantes-Acosta E, Pelayo R. Lineage Switching in acute leukemias: a consequence of stem cell plasticity? Bone Marrow

Res. 2012;2012:406796. 9. Tirtakusuma R, Szoltysek K, Milne P, et al. Epigenetic regulator genes direct lineage switching in MLL/AF4 leukemia. Blood. 2022;140(17):1875-1890. 10. Hu T, Murdaugh R, Nakada D. Transcriptional and microenvironmental regulation of lineage ambiguity in leukemia. Front Oncol. 2017;7:268. 11. Lamble AJ, Myers RM, Taraseviciute A, et al. KMT2A rearrangements are associated with lineage switch following CD19 targeting CAR T-cell therapy. Blood. 2021;138(Suppl 1):S256. 12. Wölfl M, Rasche M, Eyrich M, Schmid R, Reinhardt D, Schlegel PG. Spontaneous reversion of a lineage switch following an initial blinatumomab-induced ALL-to-AML switch in MLLrearranged infant ALL. Blood Adv. 2018;2(12):1382-1385. 13. van der Sluis IM, de Lorenzo P, Kotecha RS, et al. Blinatumomab added to chemotherapy in infant lymphoblastic leukemia. N Engl J Med. 2023;388(17):1572-1581. 14. Issa GC, Ravandi F, DiNardo CD, Jabbour E, Kantarjian HM, Andreeff M. Therapeutic implications of menin inhibition in acute leukemias. Leukemia. 2021;35(9):2482-2495. 15. Issa GC, Aldoss I, DiPersio J, et al. The menin inhibitor revumenib in KMT2A-rearranged or NPM1-mutant leukaemia. Nature. 2023;615(7954):920-924.

Haematologica | 109 January 2024

297

LETTER TO THE EDITOR

Red cell distribution width as a new prognostic biomarker in refractory chronic graft-versus-host disease Chronic graft-versus-host disease (cGvHD) is an important immunologic complication of allogeneic hematopoietic cell transplantation (HCT) that can affect multiple organs and lead to increased morbidity and mortality.1,2 Several adverse factors for survival in patients with cGvHD have been identified.3,4 Despite significant progress in the classifications of cGvHD, there is a need for improved and practical prognostication systems to address interpatient variability and achieve the goal of personalized treatment.5 In this study, we introduce red cell distribution width (RDW) as a new biomarker candidate that could enhance the development of personalized treatment strategies and risk stratifications for cGvHD. Red cell distribution width is a measure of the variation of the mean corpuscular volume of red blood cells (RBC) and is routinely reported in most standard complete blood count results.6 Although RDW is affected to some degree by any type of anemia, the etiology of variation in RDW is considered to be multifactorial and is not completely understood.7 Multiple epidemiological and longitudinal observational studies have shown that higher RDW values are independently associated with increased all-cause, cardiovascular and cancer-related mortality.7-9 Age, inflammation and the presence of clonal hematopoiesis are associated with increased RDW, but the association of RDW with mortality is independent of these factors.6-8,10 Red cell distribution width has not been previously evaluated in patients with cGvHD. The primary objective of this study was to determine if an RDW cut-off value is associated with overall survival (OS) in patients with moderate or severe cGvHD enrolled on an observational, cross-sectional study (NCT00092235) at the National Institutes of Health (NIH),11 and to estimate the 5-year OS probability in patients with RDW values above and below the specified cut-off. Secondary objectives were to determine 5year relapse-free survival, cumulative incidence of relapse, non-relapse mortality (NRM), and significant associations of RDW with clinical variables and other biomarkers in patients with RDW values above and below the established cut-off. All variables were prespecified in an analysis plan. Selected biomarkers of interest were divided into quartiles and the association of the four groups with OS was determined to establish binary cut-offs. Probabilities of survival were determined by the Kaplan-Meier method, with log-rank tests used to determine the difference in survival between groups. All P values are twotailed. Statistical analyses were performed using SAS version 9.4 and GraphPad Prism 9.3.

The characteristics of the 404 patients included in this study, who were enrolled from October 2004 to December 2018, are presented in Online Supplementary Table S1. Patients were enrolled at a median of three years after HCT and two years after diagnosis of cGvHD, and underwent a one-week comprehensive multidisciplinary evaluation including laboratory collection at this single time point.11 All patients provided written informed consent to participate in the study, which was approved by the National Cancer Institute Institutional Review Board. The trial was registered at ClinicalTrials.gov (NCT00092235). The median follow-up of the total cohort, calculated by the reverse Kaplan-Meier method, was 12 years (interquartile range [IQR]: 8.4-14.5 years) but the median OS was not reached (estimated 15-year OS: 54.1% [95% Confidence Interval (CI): 47.7-60.1%]). Median RDW of the total study cohort was 14.5% (IQR: 13.4-16.0%). In univariate Cox regression, RDW (%) displayed an association with survival as a continuous variable (Hazard Ratio [HR]=1.09; 95%CI: 1.03-1.14; P=0.004). However, we identified a clear split in survival curves between the first quartile (<13.4%) and the remaining overlapping quartiles of RDW (Figure 1) distinguishing groups with statistically different OS (P=0.0024, adjusted for 3 cut-points). The corresponding 5-year OS probabilities were 84.1% (95%CI: 75.3-89.9%) and 69.1% (95%CI: 63.674.0%) for patients with RDW <13.4% and ≥13.4%, respectively. Information on the cause of death after excluding those with non-malignant reasons for transplant was available in 286 patients (Online Supplementary Figure S1). Patients with RDW ≥13.4% had a higher 5-year NRM of 20.8% (95%CI: 15.3-26.7%) compared to 7.6% (95%CI: 2.815.7%) among patients with RDW <13.4% (P=0.0082 for the full curves). The absolute difference in survival between the two RDW groups was similar to Karnofsky performance status (KPS) <80% versus ≥80% and slightly lower than estimated glomerular filtration rate (eGFR) <45 versus ≥45 mL/min/1.73m2 but not comparable to NIH lung score 3 versus 0-2, the latter representing the most serious risk factor for mortality in our cohort (Figure 1). Patients with an RDW value ≥13.4% were more likely to have received a peripheral blood stem cell graft (PBSC) but there were no between-group differences in age, sex, donor HLA-match or kidney function. Anemia, defined by a hemoglobin (Hb) level below 11 g/dL, was present in 21% (85/402) of patients, and RDW ≥13.4% was strongly associated with lower Hb and higher lactate dehydrogenase (LDH).

Haematologica | 109 January 2024

298

LETTER TO THE EDITOR

E Figure 1. Probability of overall survival according to red cell distribution width and other risk factors. The separation of red cell distribution width (RDW) values into quartiles (A) demonstrates a clear split in the overall survival (OS) probability between the first quartile (RDW <13.4%) and the remaining quartiles, which overlap each other (B) (see also Table 1). Other previously described risk factors of adverse survival such as Karnofsky performance score <80% (C), National Institutes of Health lung score 3 (D), and moderately to severely decreased kidney function (E) defined by estimated glomerular filtration rate (eGFR) <45 mL/min/1.73m2 (calculated according to Chronic Kidney Disease Epidemiology Collaboration 2021 equation) are presented for comparison of impact on OS. *P value adjusted to account for three implicit cutpoints that could have been used when dividing the groups into quartiles.

In univariate analyses (Online Supplementary Table S1), RDW ≥13.4% was strongly associated with indicators of increased cGvHD severity and activity such as KPS <80%, cGvHD global score severe, increasing involvement of skin and mouth, higher number of prior lines of GvHD therapy, and higher number of concurrent GvHD treating agents at the time of enrollment. Patients with an RDW value ≥13.4% received steroids more frequently at enrollment than those with RDW <13.4% (70% vs. 47%; P<0.0001). In addi-

tion, patients with RDW ≥13.4% receiving steroids had a higher prednisone equivalent dose compared to those with RDW <13.4% (median mg/kg/day: 0.26 [IQR: 0.13-0.5] vs. 0.15 [IQR: 0.07-0.36]; P=0.008). Biomarkers of inflammation,12 including high sensitivity CRP, C3, C4, and erythrocyte sedimentation rate were elevated in patients with RDW ≥13.4%, but no differences were found in ferritin or interleukin 6 levels. The multiple logistic regression established oral cGvHD

Haematologica | 109 January 2024

299

LETTER TO THE EDITOR (binary; Odds Ratio [OR]=1.77), mammalian target of rapamycin (mTOR) inhibitor use (binary; OR=3.49), LDH (continuous; OR=1.01), Hb (continuous; OR=0.68), and absolute neutrophil count (ANC; continuous; OR=1.21) as independent factors associated with RDW ≥13.4% (Online Supplementary Table S1). In this cohort, there was no statistical difference in OS according to age, sex, stem cell source (bone marrow vs. PBSC; P=0.17) and NIH global score moderate versus severe (P=0.071). The analysis of laboratory biomarkers4,12 revealed notable associations with survival. While platelet count was not associated with OS, we observed important survival differences among patients who were divided into groups based on their albumin, CRP and immunoglobulin G (IgG) levels.

Albumin (g/dL) was strongly associated with longer OS as a continuous variable (HR: 0.42 [univariate Cox regression]; 95%CI: 0.31-0.58; P<0.001) and patients with an albumin level <3.45 g/dL had the poorest survival (Figure 2). Increasing CRP (mg/L) was linked to shorter OS (HR: 1.01; 95%CI: 1.00-1.02; P=0.009) and two groups divided at 7.6 mg/L could be distinguished (Online Supplementary Figure S2). IgG showed a distinctly binary cut-off at 6.25 g/L (Figure 2) and no correlation with OS as a continuous variable (HR: 0.97; 95%CI: 0.93-1.00; P=0.06). Absolute eosinophil count and absolute lymphocyte count displayed a weak association with OS but no significant threshold could be identified. C3 and C4 showed no association with OS. The Cox proportional hazards model confirmed that RDW ≥13.4%, albumin <3.45 mg/dL, IgG <6.25 g/L, KPS <80%,

Figure 2. Multivariable Cox model including other laboratory biomarkers. Inspection of the albumin quartiles revealed three distinct groups (A) with different overall survival (OS) probabilities whereas immunoglobulin G (IgG) (B) could be divided into two groups using a binary cut-off. Cox proportional hazards models for OS were built to evaluate the association of red cell distribution width (RDW) on OS, controlling for prespecified variables with P<0.10 by the log-rank test, and with the final model determined by backward selection. The final Cox model for OS (C) shows that RDW ≥13.4%, albumin <3.45 g/dL, and IgG <6.25 g/L are independent adverse risk factors for survival leading to different survival probabilities (D) depending on the number of any of these risk factors identified (i.e., 0, 1, 2, or 3 of these traits). *P value adjusted to account for three implicit cut-points that could have been used when dividing the groups into quartiles. KPS: Karnofsky performance score; eGFR: estimated glomerular filtration rate; NIH: National Institutes of Health. Haematologica | 109 January 2024

300

LETTER TO THE EDITOR Table 1. Cox proportional hazards model for overall survival. RDW

Univariate HR (95% CI)

Multivariable HR (95% CI)

1st quartile (<13.4%)

101

1.00 [ref]

N/A

1.00 [ref]

N/A

2nd quartile (13.4-14.5%)

104

2.15 (1.34-3.43)

0.0014

2.14 (1.34-3.42)

0.0014

3rd quartile (14.6-16.0%)

102

1.75 (1.08-2.86)

0.0246

1.71 (1.05-2.77)

0.0306

4th quartile (≥16.1%)

2.24 (1.38-3.64)

0.0011

2.12 (1.32-3.42)

0.0020

CI: Confidence Interval; HR: Hazard Ratio; N: number; N/A: not applicable; RDW: red cell distribution width; ref: reference value.

eGFR <45 mL/min/1.73m2 and lung cGvHD NIH score 3 were independent risks factors for increased mortality (Figure 2). Furthermore, there was no linear relationship between RDW quartiles and the risk of death, with similar HR observed for other quartiles (Table 1). Using only lab parameters from the Cox model, RDW ≥13.4% together with albumin <3.45 mg/dL and IgG <6.25 g/L allowed for estimation of different survival probabilities (Figure 2). The association of RDW ≥13.4% with acute-phase proteins such as C-reactive protein (CRP) in the univariate analysis, and increased ANC in the multivariable regression analysis, suggests that inflammation may be a contributor to RDW elevation and the associated reduction in Hb. While hypoalbuminemia can also result from malnutrition and renal or hepatic disease, the drop in albumin concentration and association with worse OS is likely due to chronic inflammation from cGvHD, which was independent of moderate or severe renal function impairment. Albumin production by hepatocytes is known to be inhibited by proinflammatory cytokines13 and appears to be a more reliable marker of chronic systemic inflammation in cGvHD patients compared to CRP, which could be more rapidly influenced by transient infections. The association of RDW ≥13.4% with mTOR inhibitor therapy cannot be explained by a direct side-effect of this class of drugs. Contrary to expectations, a placebo-controlled randomized trial involving 25 healthy older adults (age 70-95 years) observed a decrease in RDW during sirolimus treatment.14 This suggests that the association may be attributed to higher cGvHD activity, necessitating treatment with these agents. The association of RDW ≥13.4% with elevated LDH in multiple regression analysis is unclear. LDH elevation at this stage after HCT is rarely explained by relapsing disease. Possible causes include inflammation, disturbed erythropoiesis, tissue damage from cGvHD, drugs, and subclinical transplant-associated thrombotic microangiopathy or hemolytic anemia.15 Patients in this population

infrequently require transfusions but no systematic information regarding transfusion support was available. The cross-sectional design of the study does not allow us to assess changes in the measured variables over time. Survivorship bias might have affected the biomarker cutoff values because patients with extreme values could have died before enrollment in this trial. The identified biomarker thresholds should be interpreted in the context of the severely affected cGvHD population included in this study. Further studies, ideally clinical trials, are required to replicate these observations and confirm the prognostic validity of the RDW. These findings, however, strongly support the conclusion that RDW values beyond the normal range are associated with an unfavorable outcome.

Authors Eduard Schulz,1,2 Filip Pirsl,1 Noa G. Holtzman,1,2 David Beshensky,1 Edward W. Cowen,3 Sandra A. Mitchell,4 Seth M. Steinberg5 and Steven Z. Pavletic1,2 Center for Cancer Research, National Cancer Institute, National

Institutes of Health, Bethesda; 2Myeloid Malignancies Program, National Institutes of Health, Bethesda; 3Dermatology Branch, National Institute of Arthritis and Musculoskeletal and Skin Diseases, National Institutes of Health, Bethesda; 4Outcomes Research Branch, Division of Cancer Control and Population Sciences, National Cancer Institute, Rockville and 5Biostatistics and Data Management Section, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA Correspondence: S. PAVLETIC - pavletis@mail.nih.gov https://doi.org/10.3324/haematol.2023.283646

Haematologica | 109 January 2024

301

LETTER TO THE EDITOR Received: May 30, 2023.

wrote the original draft. All authors wrote, reviewed and edited the

Accepted: August 10, 2023.

final version for publication.

Part of this work was presented at the 64th Annual Meeting of the American Society of Hematology 2022.

Disclosures ES received honoraria from Amgen. All other authors have no

Funding

conflicts of interest to disclose. The views expressed here do not

This work was supported by funding from the Intramural Research

represent the official views of the National Institutes of Health or

Program, National Institutes of Health, National Cancer Institute,

the United States Government.

Center for Cancer Research.

Contributions

Data-sharing statement

ES, FP, NGH, SMS and SZP are responsible for the study concept. FP

The de-identified data set will be provided to researchers who pro-

and DB collected the data. SMS and ES performed the statistical

vide a methodologically sound proposal.

analysis. ES, NGH, EWC, SAM, SMS and SZP interpreted the data. ES

References 1. Holtzman NG, Pavletic SZ. The clinical landscape of chronic graft-versus-host disease management in 2021. Br J Haematol. 2022;196(4):830-848. 2. Palmer J, Chai X, Pidala J, et al. Predictors of survival, nonrelapse mortality, and failure-free survival in patients treated for chronic graft-versus-host disease. Blood. 2016;127(1):160-166. 3. Arora M, Klein JP, Weisdorf DJ, et al. Chronic GVHD risk score: a Center for International Blood and Marrow Transplant Research analysis. Blood. 2011;117(24):6714-6720. 4. Ayuk F, Veit R, Zabelina T, et al. Prognostic factors for survival of patients with newly diagnosed chronic GVHD according to NIH criteria. Ann Hematol. 2015;94(10):1727-1732. 5. Buxbaum NP, Socie G, Hill GR, et al. Chronic GvHD NIH Consensus Project Biology Task Force: evolving path to personalized treatment of chronic GvHD. Blood Adv. 2023;7(17):4886-4902. 6. Hoffmann JJ, Nabbe KC, van den Broek NM. Effect of age and gender on reference intervals of red blood cell distribution width (RDW) and mean red cell volume (MCV). Clin Chem Lab Med. 2015;53(12):2015-2019. 7. Horne BD, Muhlestein JB, Bennett ST, et al. Extreme erythrocyte macrocytic and microcytic percentages are highly predictive of morbidity and mortality. JCI Insight. 2018;3(14):e120183. 8. Perlstein TS, Weuve J, Pfeffer MA, Beckman JA. Red blood cell distribution width and mortality risk in a community-based

prospective cohort. Arch Intern Med. 2009;169(6):588-594. 9. Riedl J, Posch F, Konigsbrugge O, et al. Red cell distribution width and other red blood cell parameters in patients with cancer: association with risk of venous thromboembolism and mortality. PLoS One. 2014;9(10):e111440. 10. Jaiswal S, Fontanillas P, Flannick J, et al. Age-related clonal hematopoiesis associated with adverse outcomes. N Engl J Med. 2014;371(26):2488-2498. 11. Goklemez S, Im AP, Cao L, et al. Clinical characteristics and cytokine biomarkers in patients with chronic graft-vs-host disease persisting seven or more years after diagnosis. Am J Hematol. 2020;95(4):387-394. 12. Grkovic L, Baird K, Steinberg SM, et al. Clinical laboratory markers of inflammation as determinants of chronic graftversus-host disease activity and NIH global severity. Leukemia. 2012;26(4):633-643. 13. Mantovani A, Garlanda C. Humoral innate immunity and acutephase proteins. N Engl J Med. 2023;388(5):439-452. 14. Kraig E, Linehan LA, Liang H, et al. A randomized control trial to establish the feasibility and safety of rapamycin treatment in an older human cohort: immunological, physical performance, and cognitive effects. Exp Gerontol. 2018;105:53-69. 15. Young JA, Pallas CR, Knovich MA. Transplant-associated thrombotic microangiopathy: theoretical considerations and a practical approach to an unrefined diagnosis. Bone Marrow Transplant. 2021;56(8):1805-1817.

Haematologica | 109 January 2024

302

LETTER TO THE EDITOR

Impaired T-cell response to mRNA vaccination heralds risk of COVID-19 in long-term allogeneic hematopoietic stem cell transplantation survivors Initial reports suggested a COVID-19-associated mortality rate approaching 30% among allogeneic hematopoietic stem cell transplant (allo-HSCT) recipients.1 However, patient outcomes have improved with the introduction of effective vaccines and advancements in patient management, such as rapid diagnostic testing and antiviral therapies. Although new Omicron virus variants show increased transmissibility, they have also been associated with reduced risk of severe disease among vaccinated individuals.2 Recent data suggests that mortality in fully vaccinated allo-HSCT recipients has decreased to as low as 1%.3 Our previous findings show that a weak T-cell response following the first two doses of vaccination leads to reduced humoral immunity4 and that allo-HSCT recipients who experience vaccine-related adverse events demonstrate augmented T-cell responsiveness.5 While various studies have evaluated COVID-19 vaccination response in this patient cohort,6,7 to our knowledge, this is the first re-

port of allo-HSCT recipients with poor T-cell response following two doses of mRNA vaccine being at increased risk of acquiring subsequent COVID-19. This study was conducted between March 2021 and November 2022 at the Sahlgrenska University Hospital in Gothenburg, Sweden (CONSORT diagram in Figure 1). AlloHSCT recipients without prior history of COVID-19 (n=50), identified in local transplant registries in the Region Västra Götaland (population of approximately 1.7 million), were recruited to participate in this sub-study within the DurIRVac study. All participants gave written informed consent before enrollment. Enrolled allo-HSCT recipients fulfilled the following predetermined criteria: (i) to be at least 3 months post allo-HSCT, (ii) to not have received rituximab during the last 6 months, and (iii) to be without uncontrolled acute graft-versus-host disease (GvHD) (grade 34), all according to EBMT guidelines for COVID-19 vaccination (version 7, 2021). Patients were sampled before

Figure 1. CONSORT-diagram for allogeneic stem cell transplantation recipients in the study. 1Previously COVID-19 vaccinated (N=13); recent treatment with rituximab (N=4); hesitant to COVID-19 vaccine (N=2), current infection (N=2), declined participation (N=3). 2Previously polymerase chain reaction (PCR)-confirmed COVID-19 infection (N=5) or seropositive at baseline (N=3).3PCR-positive between dose 1 and 2 (N=1); dead due to complications of severe graft-versushost disease (N=1); relapse of underlying disease (N=3); lost to follow-up (N=4). 4At least 2 blood samples available at the first 2 vaccine doses and data available on occurrence of COVID-19. 5Complete data available on interleukin-2 (N=27) and anti-nucleocapsid immunoglobulin G (N=28). Haematologica | 109 January 2024

303

LETTER TO THE EDITOR and after the first two vaccine doses and then every 6 months. The cohort mainly consisted of long-term survivors with a median time following transplant of 90 months (range, 10-330), with no participant receiving any COVID vaccine dose prior to transplantation. The DurIRVac study was approved by the Swedish Ethical Review Authority (Etikprövningsmyndigheten; permit no. 2020-03276, 202100374 and 2021-00539) and by the Swedish Medical Products Agency (Dnr: 5.1-2021-11118) and has been registered at the European Union Drug Regulating Authorities Clinical Trials Database (EudraCT no. 2021-000349-42). The researchers in this study had neither an influence on which mRNA COVID-19 vaccine (mRNA-1273 Moderna Spikevax® or BNT162b2 Pfizer-BioBTech Comirnaty®) participants received nor on the number or time point of vaccinations, as doses were administered in accordance with regional prioritization. Chemiluminescent microparticle immunoassays (CMIA) were performed on serum using the automated Alinity system for quantitative measurement of immunoglobulin (Ig)G antibodies against the receptor-binding-domain (RBD) of the spike protein of SARS-CoV-2 (SARS-CoV-2 IgG II Quant, Abbott, Abbott Park, Illinois, USA) with levels reported in the World Health Organization international standard binding antibody units (BAU)/mL (quantitative detection range: 14-5,680 BAU/mL or equivalent to 1.14-3.75 log10 BAU/mL; samples reaching 5,680 BAU/mL were diluted with seronegative serum and reanalyzed). Anti-nucleocapsid antibodies were analyzed at baseline and during follow-up sampling (November 2022). Peripheral blood, collected in vacutainer lithium-heparin tubes (BD, Plymouth, UK), was stimulated with 15-mer peptides with 11-amino acid overlap spanning the S1 domain of the SARS-CoV-2 surface glycoprotein as described previously.8,9 Plasma from whole blood cultures was recovered and stored at -80°C until analysis of IL-2. Levels of interleukin (IL)-2 in whole blood supernatants were determined by the FirePlex®-96 Key Cytokines Immunoassay (antibody [ab] 243549 and ab 285173, Abcam) according to the manufacturer’s instructions. Samples were acquired on a BD LSR Fortessa (BD) and analyzed with FirePlex Analysis Workbench (Abcam). S1-specific responses were calculated by subtracting levels of IL-2 in unstimulated control samples from those in S1-stimulated samples. The S1 peptide-induced IL-2 has been shown to be contributed foremost by T cells, in particular CD4+ T cells.10 All 41 patients were COVID-19-naïve 5 months after the second vaccine dose as per polymerase chain reaction (PCR), COVID-19 antigen test, analysis of antibodies against the nucleocapsid protein of SARS-CoV-2 (SARS-CoV-2 IgG, Abbott, Abbott Park, Illinois, USA), review of medical records, or evaluation of participant self-reporting questionnaire. During the follow-up period after receiving the initial two

vaccine doses, the patients were divided into two groups based on whether they had subsequent breakthrough COVID-19 infection or not (baseline characteristics detailed in Table 1). All 12 documented COVID-19 infections had detectable SARS-CoV-2 RNA (n=8), reactive COVID antigen test (n=3), and/or presence of antibodies against the nucleocapsid (n=2) in a sample obtained after the second vaccine dose. No severe breakthrough infections, i.e., requiring hospitalization or leading to death, were observed during the study. The Omicron variants (B.1.1.529) BA.1 and BA.2 dominated the local circulation during this study. At 5 months after the second vaccine dose, the group of patients who experienced subsequent breakthrough COVID-19 infection (n=12) had a significantly lower median level of S1-induced IL-2 (52 pg/mL; range 7-141) compared to the group without later breakthrough infection (n=27) (174 pg/mL; range, 1-2,344; P=0.016, Mann-Whitney) (Figure 2). As baseline characteristics stratified according to COVID-19 breakthrough infection demonstrated a significant difference in the mRNA vaccine used (i.e., Moderna vs. Pfizer-BioNTech) for primary vaccination, this association between T-cell response 5 months after the second vaccination and risk of infection was confirmed in a subgroup of the cohort consisting of those <65 years of age, where all patients received BNT162b2 (Pfizer-BioBTech Comirnaty®) and were 11 of 12 infections occurred (Online Supplementary Figure S1). A receiver operating characteristics (ROC) analysis determined 145 pg/mL of S1-induced IL-2 as a cut-off level that best discriminated patients at risk for subsequent breakthrough infections. No patients with S1induced IL-2 levels above this value after two doses of vaccination experienced breakthrough infection, compared to 52% of those with levels below (P<0.001, Fischer's exact test). The antibody levels against RBD within the spike protein after two vaccine doses did not differ significantly between patients who experienced breakthrough infection and those who did not. The median antibody levels were 207 BAU/mL; range, 0-1,778 (2.32 log10 BAU/mL; range <1-3.25) in the group with subsequent COVID-19 versus 479 BAU/mL; range, 0-1413 (2.68 log10 BAU/mL; range <1-3.15) in the group without subsequent infection (P=0.89, MannWhitney; Figure 2). However, there was a significant correlation between antibody levels and IL-2 response following vaccination (r=0.51; P=0.001, Spearman's Ρ). Of the 41 patients, 15 were above or equal to the age of 65 years. This elderly portion of the cohort differed in two ways from younger allo-HSCT recipients; only one of 15 patients had a breakthrough infection, and all were vaccinated with mRNA-1273 (Moderna Spikevax®). In contrast, all study participants below 65 years of age were vaccinated with BNT162b2 (Pfizer-BioBTech Comirnaty®), and 11 of 26 experienced breakthrough infection. A subgroup analysis showed that, also in the latter younger cohort, high levels

Haematologica | 109 January 2024

304

LETTER TO THE EDITOR Table 1. Baseline characteristics of allogeneic stem cell transplantation recipients stratified by COVID-19 breakthrough infection status after receiving two mRNA vaccine doses and analyzing the complete sample data.

Characteristics

All patients N=41

COVID-19† following 2nd vaccination N=12

Age in years, median (range)

54 (29-78)

46.5 (29-78)

62 (40-75)

0.16*

21 (51) 20 (49)

5 (42) 7 (58)

16 (55) 13 (45)

0.51**

Vaccine Moderna/Pfizer-BioNTech, N (%)

15/26 (37/63)

1/11 (8/92)

14/15 (48/52)

0.03**

Time in months from transplant to first vaccine dose, median (range)

90 (10-330)

108 (13-330)

89 (10-246)

0.58*

Conditioning intensity, N (%) RIC/MAC

20/21 (49/51)

5/7 (42/58)

15/14 (52/48)

0.73**

Donor source, N (%) RD /URD

10/31 (24/76)

4/8 (33/67)

6/23 (21/79)

0.44**

Diagnosis, N (%) AML CML MDS Other

18 (44) 7 (17) 4 (10) 12 (29)

7 (44) 2 (13) 1 (6) 6 (38)

11 (44) 5 (20) 3 (12) 6 (24)

1.0** 0.31** 1.0** 0.13**

GvHD, N (%) Yes No

15 (38) 25 (63)

4 (33) 8 (67)

11 (39) 17 (61)

1.0**

Ongoing immunosuppression, N (%) Yes No

7 (18) 33 (83)

2 (17) 10 (83)

5 (18) 23 (82)

1.0**

Sex, N (%) Male Female

No COVID-19† following 2nd vaccination N=29

Definition: positive SARS-CoV-2 PCR, antigen test and or antibodies to nucleocapsid antigen. The presence of graft-versus-host disease (GvHD) was assessed at the visit prior to the first vaccine dose. Ongoing immunosuppression was registered if present at first vaccination. Statistics using Mann-Whitney U test (*) or Fischer´s exact test (**). Statistical analyses were performed using SPSS statistical software package (version 28). P values are designated as follows: *P<0.05 and **P<0.01. All indicated P values are two-sided. RIC: reduced intensity conditioning; MAC: maximal intensity conditioning; RD: related donor; URD: unrelated donor; AML: acute myeloid leukemia; CML: chronic myeloid leukemia; MDS: myelodysplastic syndrome.

†

of SARS-CoV-2 specific T cells, but not antibodies, following dual vaccination were protective against subsequent infection (Online Supplemental Figure S1). Notably, we did not observe significant differences in the immune responses elicited by two different vaccines (495 vs. 434 BAU/mL [2.69 vs. 2.64 log10 BAU/mL]; P=0.8; and 167 vs. 88 IL-2 pg/mL; P=0.3; following 2 doses of mRNA-1273 and BNT162b2, respectively). This finding suggests that the lower incidence of COVID-19 among individuals over the age of 65 might be attributed to their higher level of adherence to social distancing and other preventive measures, rather than differences regarding vaccine-induced immune responses. The immunogenic mRNA COVID-19 vaccines elicit robust innate and adaptive immune responses.11 Although T and B cells crosstalk, they have partially independent roles.10 Higher neutralizing antibody levels against SARS-CoV-2 in healthy adults have been associated with protection against infection.12 However, in the Omicron era, studies have reported a lack of association between antibody concentrations and the incidence of infection,13 which was sup-

ported by our findings. The magnitude of the T-cell response has been shown to negatively correlate with disease severity.14 However, an absolute protective effect against infection due to high T-cell activity in the absence of antibodies has not been established. A recent study of 572 vaccinated allo-HSCT recipients found a 1-year cumulative incidence of breakthrough infection of 15%, with antibody titers associated with both breakthrough infections and disease severity, though data on T-cell response were not reported.15 As an increasing number of T-cell assays for SARS-CoV-2 (e.g., Qiagen QuantiFERON® SARS-CoV-2, Hyris’ T-Cell Test, T-SPOT®.COVID, etc.) are becoming commercially available as complements to already accessible COVID-19 antibody tests, our findings might be easier to generalize to other healthcare settings. Despite the small sample size, which may have resulted in some numerically very different results being statistically similar, and the inevitable risk of not detecting asymptomatic COVID-19 despite analyzing anti-nucleocapsid antibodies at baseline and follow-up, we conclude that a strong T-cell response might protect against acquiring

Haematologica | 109 January 2024

305

LETTER TO THE EDITOR Figure 2. Humoral and T-cell-mediated immune responses 5 months after second mRNA COVID vaccination in allogeneic stem cell transplantation recipients stratified according to subsequent SARS-CoV-2 infection or not. Spike 1 (S1) specific immune responses in allogeneic stem cell transplantation recipients 3 months after the second vaccine dose demonstrated as T-cell production of interleukin (IL)-2 in supernatant plasma following stimulation of whole blood with S1 peptides and immunoglobulin (IgG) antibody levels in serum against the receptor-binding domain (RBD) in S1. Statistical comparisons were calculated by Mann-Whitney U test. Statistical analyses were performed using GraphPad Prism software (version 9). P value is twosided and designated as *P<0.05. BAU: standard binding antibody units.

breakthrough COVID-19. Additionally, antibody and T-cell responses following mRNA COVID vaccination were correlated in allo-HSCT recipients. This study suggests that antigen-specific T cells play a pivotal role in protection against COVID-19 and warrants future studies of their protective role against COVID-19 and infection with other respiratory viruses.

Authors

Contributions ML, AM and KH were responsible for designing and writing the protocol,

Sigrun Einarsdottir, Jesper Waldenström, Andreas Törnell, Johan

conducting the study, extracting and analyzing data, interpreting

Ringlander,

results, writing the manuscript, and updating the reference list. SE

2,4

Joakim B. Stenbäck,

2,4

Sebastian Malmström, Kristoffer 4

conducted the study, included all patients, participated in the

Hellstrand,2,4 Anna Martner3 and Martin Lagging2,4

extracting and analyzing of data, updated the reference list, and wrote Department of Hematology and Coagulation, Institute of Medicine,

the first draft of the manuscript. JW extracted and analyzed data,

Sahlgrenska Academy, University of Gothenburg; Department of

performed statistical analyses, created the figure and table, and

Infectious Diseases, Institute of Biomedicine, Sahlgrenska Academy,

participated in writing the manuscript. AM, AT, JR, JS and SM were

University of Gothenburg; TIMM Laboratory, Sahlgrenska Center for

responsible for the T-cell analyses, participated in the extraction and

Cancer Research, Department of Microbiology and Immunology,

analysis of data and in writing of the manuscript. All authors have read

Institute of Biomedicine, Sahlgrenska Academy, University of

and approved the last version of the manuscript.

Gothenburg and 4Region Västra Götaland, Sahlgrenska University Hospital, Department of Clinical Microbiology, Gothenburg, Sweden

Funding This work was supported by the Swedish Medical Research Council

SE and JW contributed equally as first authors.

(Vetenskapsrådet; grant no. 2021-04779), and the Gothenburg Society of Medicine (GLS-961772).

Correspondence: M. LAGGING - martin.lagging@medfak.gu.se

Data-sharing statement For original data, please contact the corresponding author. As per

https://doi.org/10.3324/haematol.2023.283551

Swedish law, individual participant data will not be shared. Haematologica | 109 January 2024

306

LETTER TO THE EDITOR

References 1. Sharma A, Bhatt NS, St Martin A, et al. Clinical characteristics and outcomes of COVID-19 in haematopoietic stem-cell transplantation recipients: an observational cohort study. Lancet Haematol. 2021;8(3):e185-e193. 2. Ren SY, Wang WB, Gao RD, Zhou AM. Omicron variant (B.1.1.529) of SARS-CoV-2: mutation, infectivity, transmission, and vaccine resistance. World J Clin Cases. 2022;10(1):1-11. 3. Ljungman P, Tridello G, Piñana JL, et al. Improved outcomes over time and higher mortality in CMV seropositive allogeneic stem cell transplantation patients with COVID-19; an infectious disease working party study from the European Society for Blood and Marrow Transplantation registry. Front Immunol. 2023;14:1125824. 4. Einarsdottir S, Martner A, Waldenström J, et al. Deficiency of SARS-CoV-2 T-cell responses after vaccination in long-term allo-HSCT survivors translates into abated humoral immunity. Blood Adv. 2022;6(9):2723-2730. 5. Wiktorin HG, Einarsdottir S, Törnell A, et al. COVID-19 vaccineinduced adverse events predict immunogenicity among recipients of allogeneic haematopoietic stem cell transplantation. Haematologica. 2022;107(10):2492-2495. 6. Dhakal B, Abedin S, Fenske T, et al. Response to SARS-CoV-2 vaccination in patients after hematopoietic cell transplantation and CAR T-cell therapy. Blood. 2021;138(14):1278-1281. 7. Einarsdottir S, Martner A, Nicklasson M, et al. Reduced immunogenicity of a third COVID-19 vaccination among recipients of allogeneic hematopoietic stem cell transplantation. Haematologica. 2022;107(6):1479-1482. 8. Törnell A, Grauers Wiktorin H, Ringlander J, et al. Rapid cytokine

release assays for analysis of SARS-CoV-2-specific T cells in whole blood. J Infect Dis. 2022;226(2):208-216. 9. Martner A, Grauers Wiktorin H, Törnell A, et al. Transient and durable T cell reactivity after COVID-19. Proc Natl Acad Sci U S A. 2022;119(30):e2203659119. 10. Törnell A, Grauers Wiktorin H, Ringlander J, et al. Induction and subsequent decline of S1-specific T cell reactivity after COVID19 vaccination. Clin Immunol. 2023;248:109248. 11. Li C, Lee A, Grigoryan L, et al. Mechanisms of innate and adaptive immunity to the Pfizer-BioNTech BNT162b2 vaccine. Nat Immunol. 2022;23(4):543-555. 12. Khoury DS, Cromer D, Reynaldi A, et al. Neutralizing antibody levels are highly predictive of immune protection from symptomatic SARS-CoV-2 infection. Nat Med. 2021;27(7):1205-1211. 13. Smoot K, Yang J, Tacker DH, et al. Persistence and protective potential of SARS-CoV-2 antibody levels after COVID-19 vaccination in a West Virginia nursing home cohort. JAMA Netw Open. 2022;5(9):e2231334. 14. Rydyznski Moderbacher C, Ramirez SI, Dan JM, et al. Antigenspecific adaptive immunity to SARS-CoV-2 in acute COVID-19 and associations with age and disease severity. Cell. 2020;183(4):996-1012. 15. Piñana JL, Martino R, Vazquez L, et al. SARS-CoV-2-reactive antibody waning, booster effect and breakthrough SARS-CoV-2 infection in hematopoietic stem cell transplant and cell therapy recipients at one year after vaccination. Bone Marrow Transplant. 2023;58(5):567-580.

Haematologica | 109 January 2024

307

LETTER TO THE EDITOR

Colchicine reduces inflammation in a humanized transgenic murine model of sickle cell disease Sickle cell diseases (SCD) are caused by a mutation in the β-globin gene that causes deoxygenated hemoglobin to polymerize, resulting in rigid, fragile, sickle-shaped red blood cells.1,2 Prominent clinical features are hemolytic anemia; hyperalgesia; chronic and acute pain; wide-spread organ damage; and premature death (USA median age ~50 years). SCD pathobiology involves oxidative stress, ischemia/reperfusion injury, and activation of the innate immune system with sterile inflammation and “cytokine storm”. Inflammation in SCD is chronic with acute exacerbations, and complex with many cellular and molecular components. Among these, mast cell activation and interleukin (IL)-6 and C-reactive protein (CRP) are chronically and acutely increased in SCD.3,4 CRP contributes to inflammation through the activation of the complement system, a component of innate immunity. CRP has been used to monitor both disease activity and response to therapy in many inflammatory conditions involving both innate and acquired immunity.5 Colchicine is a drug used for the treatment or prevention of symptoms of gout and other non-infectious inflammatory conditions. Recently, it has been shown to improve outcomes in atherosclerotic cardiovascular disease (ASCVD) in which, like SCD, inflammation is a secondary pathological process.6 Colchicine effects in ASCVD and other conditions include reductions in IL-6 and CRP. Among five selected agents or agent classes of anti-inflammatory drugs - colchicine, non-steroidal anti-inflammatory drugs, corticosteroids, canakinumab, and imatinib - only colchicine has been shown to inhibit each of four selected inflammatory mechanisms - inflammatory cytokines, inflammasomes, mast cell (MC) activation, and neutrophil adhesion and superoxide release. Among the five anti-inflammatory agents or agent classes, only colchicine is a tubulin inhibitor (TI). Among TI, colchicine has unique properties.7 Tubulin is a dynamic structural protein, and TI are classified as stabilizers (sTI) or destabilizers (dTI). dTI are associated with five binding sites named by an agent or agent class: colchicine, taxanes, laulimalide, epothilones, and vinca alkaloids. Among TI, only colchicine is used as an anti-inflammatory agent, with all others targeted to various types of cancer. We hypothesized that colchicine might reduce inflammation and improve the clinical course of SCD. We sought evidence for this by examining the effect of colchicine on inflammation in the HbSS-BERK humanized transgenic mouse model of SCD. These sickle mice are knockout for murine α and β globins and express human α and βS (>99% of total β) globins on a mixed genetic background. They

have severe disease with hyperalgesia, inflammation, extensive organ damage, and pathology consistent with SCD; in contrast, HbAA-BERK mice that express normal human α and β globins do not exhibit hyperalgesia or organ pathology. Only female mice were studied, as male mice are fragile with less tolerance to experimental manipulations, rendering them unsuitable for this sort of study. Female ~4-month-old mice were treated with colchicine 100 μg/kg body weight (Tocris Bioscience, Bristol, United Kingdom) or vehicle (phosphate-buffered saline) intraperitoneally once daily for 14 days. Hyperalgesia testing was performed at 1 hour (h) and 24 h after the first treatment and on day 7 and day 14.8 Then mice were humanely euthanized; blood was collected by cardiac puncture; and 4 mm diameter dorsal skin punch biopsies were collected for mast cell analysis9 or incubated in culture media for 24 h to analyze secreted inflammatory cytokines. All experiments were performed following protocols approved by the Institutional Animal Care and Use Committee. Data were analyzed using GraphPad Prism software version 9.3.1 (GraphPad Prism Inc., San Diego, CA). Serum amyloid protein (SAP) is a murine acute-phase reactant with extensive (60-70%) sequence homology with human CRP; it is commonly used as a surrogate for CRP in mouse studies.10 Consistent with our hypothesis, plasma IL-6 levels were ~67% lower (~130 vs. ~48 μg/g protein; P<0.01), and plasma SAP levels were ~45% lower (~33 vs. ~18 μg/g protein; P<0.01) in colchicine-treated mice compared to vehicle (Figure 1A). Consistent with our observations of reduced systemic inflammation with colchicine treatment, we observed that skin-conditioned media from colchicine-treated mice showed significantly reduced levels of three pro-inflammatory cytokines compared to skin-conditioned media from vehicle-treated mice: granulocyte-macrophage colony-stimulating factor (GM-CSF), 40% reduction (P<0.05); IL-3, 40% reduction (P<0.01); and interferon γ (IFNγ), 20% reduction (P<0.05) (Figure 1B). Of note, GM-CSF can antagonize the therapeutic increase in fetal hemoglobin expression caused by hydroxyurea (HU), the most effective SCD disease-modifying drug (DMD).11 Thus, colchicine-induced reduction in GM-CSF might enhance HU effectiveness. Conditioned media levels of the anti-inflammatory cytokine IL-10 were reduced by 20% (P<0.05). We speculate that this reflects downregulation of the inflammatory loop that stimulates IL-10 production. Conditioned media cytokine levels not affected by colchicine were IL-1α, IL1β, IL-2, IL-4, IL-5, IL-6, IL-12, IL-17, monocyte chemo-at-

Haematologica | 109 January 2024

308

LETTER TO THE EDITOR

Figure 1. Colchicine attenuates inflammation in sickle mice. Female HbSS-BERK sickle mice were treated daily with 100 μg/kg/day colchicine or vehicle (phosphate-buffered saline). After 14 days of treatment, mice were euthanized, blood collected, and punch biopsies of dorsal skin were incubated in culture media for 24 hours. Data are expressed as mean ± standard error of the mean. (A) Plasma interleukin (IL)-6 and serum amyloid protein in vehicle- and colchicine-treated mice. (B) Concentrations in skin-conditioned media of IL-1α, IL-1β, IL-2, IL-3, IL-4, IL-5, IL-6, IL-10, IL-12, IL-17, monocyte chemo-attractant protein 1, tumor necrosis factor α, macrophage inflammatory protein-1α, and regulated on activation, normal T-cell expressed and secreted protein as percentage of vehicle-treated control. Cytokines were analyzed using a microplate-based array (Quansys Biosciences, Logan, UT, USA), a sandwich enzyme-linked immunosorbent assay in microscale. Data expressed as colchicine-treated as percent of control. SAP: serum amyloid protein; GM-CSF: granulocyte-macrophage colony-stimulating factor.

tractant protein 1, tumor necrosis factor α, macrophage inflammatory protein-1 α, and regulated on activation, normal T-cell expressed and secreted protein. All of these are elevated in human SCD plasma,12,13 but we have no human SCD plasma samples from patients exposed to colchicine for comparison. Previous studies in sickle mice indicate that disease-associated MC activation results in the release of substance P, tryptase, and multiple inflammatory cytokines from skin and dorsal root ganglia that leads to neurogenic inflammation and nociceptor activation.8 Colchicine reduced skin MC degranulation by ~75% (P<0.0001), and we attribute reductions in GM-CSF, IL-3, and IFNγ to this effect (Figure 2). We did not observe changes in circulating blood cell parameters (see the Online Supplementary Figure S1). Although colchicine can cause myelosuppression, it does not cause significant myelosuppression in SCD or other conditions at currently recommended clinical doses.6 We also

did not observe changes in hyperalgesia (Online Supplementary Figure S2). Given the chronic nature of inflammation and tissue damage in SCD, we speculate that starting colchicine treatment in younger mice and continuing dosing for a longer interval might attenuate the development of inflammation and consequent organ damage. The current SCD drug armamentarium is inadequate, and introduction of a drug that targets SCD inflammation might improve outcomes. Although the four DMD approved in the USA for the treatment of SCD - hydroxyurea, voxelotor, L-glutamine, and crizanlizumab - have some secondary anti-inflammatory effects, significant inflammation persists in persons with SCD despite taking one or more of these drugs.14 Non-steroidal anti-inflammatory agents are commonly used to treat pain in SCD, but they have not been shown to have disease-modifying effects in SCD. For unclear reasons, corticosteroids have mixed effects, and in general are avoided. In a small trial, the anti-

Haematologica | 109 January 2024

309

LETTER TO THE EDITOR

Figure 2. Colchicine inhibits mast cell degranulation in the skin of transgenic humanized sickle mice. (A) Black arrow: degranulated mast cell (MC). (B) Yellow arrow: intact MC. Toluidine blue stain was prepared by dissolving 0.25 g toluidine blue (SigmaAldrich) in 35 mL distilled water, 15 mL ethanol 100%, and 1 mL HCL. After deparaffinization, the skin sections were incubated in toluidine blue for 1 minute at room temperature, washed with distilled water, and air-dried. The stained specimens were observed under an Olympus Microscope BH-2. MC were recognized by red-purple metachromatic staining color on a blue background. MC were counted in 20 fields, 4 fields per slide, and expressed as total MC number, number of degranulated MC, and percentage of degranulated cells. Degranulated MC were defined as cells associated with ≥8 granules outside the cell membrane. Analyzed with unpaired t test, two-tailed in GraphPad Prism software; software (P<0.0001); one outlier in the colchicine-treated cohort was detected by Grubb's test. NS: not significant

IL-1β antibody, canakinumab, showed a nominal reduction in CRP and trends toward improvement in multiple clinical endpoints, but none of these were statistically significant, and the indication has not been pursued. We are aware of only one other study of colchicine in SCD, in which colchicine was shown to attenuate inflammation and cardiac damage in a transgenic mouse model.15 Reduced inflammation due to colchicine treatment might result in reduced pain, preservation of end-organ function, and prolonged survival in SCD. As colchicine has a unique mechanism of action and minimal side-effects at currently recommended clinical doses, patients might benefit

from combination of colchicine with one or more of the established DMD. Further research is warranted.

Authors Raghda T. Fouda,1 Hemanth M. Cherukury,1 Stacy B. Kiven,1 Natalie R. Garcia,1 Donovan A. Argueta,1 Graham J. Velasco,2 Kalpna Gupta1,3# and John D. Roberts4# Division of Hematology/Oncology, Department of Medicine,

Haematologica | 109 January 2024

310

LETTER TO THE EDITOR University of California, Irvine, CA; 2Pathology Department, VA Long

Other authors have no conflicts of interest to disclose.

Beach Medical Center, Long Beach, CA; 3Division of Hematology, Oncology and Transplantation, Department of Medicine, University

Contributions

of Minnesota, Minneapolis, MN and Department of Internal

RTF conducted experiments, mast cell staining and analysis,

Medicine, Yale University, New Haven, CT, USA

analyzed data, prepared the data for publication, prepared figures

and the visual abstract. HMC conducted hyperalgesia experiments. SBK bred, phenotyped and prepared sickle mice. NRG conducted

KG and JDR contributed equally as senior authors.

the ELISA assays. DAA participated in experimental design and Correspondence:

cytokine analysis. GJV analyzed histopathology. KG participated in

J.D. ROBERTS - john.d.roberts@yale.edu

experimental design, supervision of the study, data analysis and

K. GUOTA - kalpnag@hs.uci.edu

interpretation, and manuscript editing. JDR proposed the study, participated in the experimental design and interpretation of

https://doi.org/10.3324/haematol.2023.283377

results, and wrote the manuscript.

Received: May 3, 2023.

Funding

Accepted: July 28, 2023.

This research was supported by the National Institutes of Health

Early view: August 3, 2023.

grant R01 HL147562 and Susan Samueli Integrative Health Institute scholar award (to KG); Diversity Supplement R01HL147562-03S (to

SBK), University of California President’s and Gianini Foundation

Published under a CC BY-NC license

post-doctoral fellowship (to DAA); ASH graduate hematology award to HMC; and institutional funds (to JDR). The content is solely the

Disclosures

responsibility of the authors and does not necessarily represent the

KG has received honoraria from Novartis and CSL Behring; and

official views of the National Institutes of Health.

research grants from Cyclerion, 1910 Genetics, Novartis, Grifols, Zilker LLC, UCI Foundation and SCIRE Foundation. DAA has received

Data-sharing statement

honoraria from Cyclerion and Cayenne Foundation. JDR has been a

Original data can be made available on reasonable request to a

consultant for VeriMed Healthcare Network, Pfizer and Magellan.

corresponding author.

References 1. Ware RE, de Montalembert M, Tshilolo L, Abboud MR. Sickle cell disease. Lancet. 2017:(17);193-199. 2. DeBaun MR, Ghafuri DL, et al. Decreased median survival of adults with sickle cell disease after adjusting for left truncation bias: a pooled analysis. Blood. 2019;133(6):615-617. 3. de Almeida CB, Kato GJ, Conran N. Inflammation and sickle cell anemia. In: Costa FF, Conran N, editors. Sickle Cell Anemia: From Basic Science to Clinical Practice. Springer International Publishing. 2016. p. 177-211. 4. Pittman DD, Hines PC, Beidler D, et al. Evaluation of longitudinal pain study in sickle cell disease (ELIPSIS) by patient-reported outcomes, actigraphy, and biomarkers. Blood. 2021;137(15):2010-2020. 5. Banait T, Wanjari A, Danade V, Banait S, Jain J. Role of highsensitivity C-reactive protein (Hs-CRP) in non-communicable diseases: a review. Cureus. 2022;14(10):e30225. 6. Deftereos SG, Beerkens FJ, Shah B, et al. Colchicine in cardiovascular disease: in-depth review. Circulation. 2022;145(1):61-78. 7. Škubník J, Jurášek M, Ruml T, Rimpelová S. Mitotic poisons in research and medicine. Molecules. 2020;25(20):4632. 8. Cain DM, Vang D, Simone DA, Hebbel RP, Gupta K. Mouse models for studying pain in sickle disease: Effects of strain,

age, and acuteness. Br J Haematology. 2012;156(4):535-544. 9. Vincent L, Vang D, Nguyen J, Gupta M, et al. Mast cell activation contributes to sickle cell pathobiology and pain in mice. Blood. 2013;122(11):1853-1862. 10. Vang D, Paul JA, Nguyen J, et al. Small-molecule nociceptin receptor agonist ameliorates mast cell activation and pain in sickle mice. Haematologica. 2015;100(12):1517-1525. 11. Ikuta T, Adekile AD, Gutsaeva DR, et al The proinflammatory cytokine GM-CSF downregulates fetal hemoglobin expression by attenuating the cAMP-dependent pathway in sickle cell disease. Blood Cells Mol Dis. 2011;47(4):235-242. 12. Silva-Junior AL, Garcia NP, Cardoso EC, et al. Immunological hallmarks of inflammatory status in vaso-occlusive crisis of sickle cell anemia patients. Front Immunol. 2021;12:559925. 13. Qari MH, Dier U, Mousa SA. Biomarkers of inflammation, growth factor, and coagulation activation in patients with sickle cell disease. Clin Appl Thromb Hemost. 2011;18(2):195-200. 14. Lee MT, Ogu UO. Sickle cell disease in the new era: advances in drug treatment. Transfus Apher Sci. 2022;61(5):103555. 15. Federti E, Iatcenko I, Ghigo A, Mattè A, et al. Colchicine protects against cardiomyopathy in humanized mouse model for sickle cell disease. Blood. 2022;140(Suppl 1):S2513-2514.

Haematologica | 109 January 2024

311

LETTER TO THE EDITOR

Optimizing transplantation procedures through identification of prognostic factors in second remission for children with acute myeloid leukemia with no prior history of transplant Children with acute myeloid leukemia (AML) who present with a high likelihood of relapse are indicated for hematopoietic stem cell transplantation (HSCT) as a consolidation therapy in the first complete remission (CR1).1 However, the proportion of patients undergoing HSCT in the CR1 in recently published clinical trials varied significantly, ranging from 8% to 29%.1,2 Therefore, the patient population that would benefit from HSCT in the CR1 remains a topic of ongoing debate.1,2 In marked contrast to the CR1, nearly all research groups offer HSCT for all relapsed children in the second complete remission (CR2).1,2 Previous studies have identified various prognostic indicators in pediatric patients with relapsed AML, including a history of HSCT, the year of HSCT, the duration of CR1, and achieving the CR2 prior to HSCT.3,4 In addition, it is necessary to identify other modifiable prognostic factors associated with HSCT, such as donor type and conditioning regimen. Furthermore, given the inherent heterogeneity of relapsed AML, it is crucial to identify prognostic factors specifically in patients who underwent their first HSCT in their CR2, as this population is expected to benefit from HSCT. In order to evaluate the characteristics and prognostic factors of children with AML undergoing their first allogenic HSCT in their CR2, data from 225 patients were collected through the Transplant Registry Unified Management Program (TRUMP).5 The inclusion criteria for patients were as follows: (i) de novo non-M3 AML, (ii) age <16 years and in the CR2 at the time of undergoing allogeneic HSCT, (iii) no prior history of HSCT, and (iv) HSCT performed between 2000 and 2019. Patients with Down syndrome (DS) were excluded. Performance status was applied as defined by the Eastern Cooperative Oncology Group. The myeloablative conditioning (MAC) regimen was defined in a hierarchical manner as follows: (i) a totalbody irradiation (TBI)-based regimen, which included >8 Gy TBI; (ii) a busulfan-based regimen, which included ≥9 mg/kg busulfan; or (iii) a melphalan-based regimen, which included >140 mg/m2 melphalan. All other regimens were analyzed as reduced-intensity conditioning regimens.6 High-risk cytogenetics/genetics were defined as previously described.7 The patients or their parents provided written consent to undergo transplantation and for the use of medical records for research in accordance with

the Declaration of Helsinki. The study was approved by the Data Management Committee of the TRUMP and the Institutional Ethics Committee of Okayama University (2205-004). Probabilities of overall survival (OS) and event-free survival (EFS) were calculated using Kaplan– Meier estimators. EFS was defined as survival in continuous CR after HSCT. Relapse was defined by morphological relapse of ≥ 5% blasts in bone marrow or extramedullary relapse. Non-relapse mortality (NRM) was defined as death from any cause other than relapse. Competing events were death without relapse for hematological relapse, and hematological relapse for NRM. Univariate analysis was performed using the log-rank test, and multivariate analysis was conducted using the Cox proportional hazard regression model. Factors that were known to affect patient outcomes after HSCT were entered into the multivariate analysis, including the age at HSCT, duration of CR1, year of HSCT, conditioning regimen, donor source, and human leukocyte antigen disparities. French-American-British (FAB) classifications and cytogenetic/genetic classifications were also entered based on the univariate analysis results. All statistical analyses were performed using EZR (Version 1.54. Saitama Medical Center, Jichi Medical University, Saitama, Japan), a graphical user interface for R (Version 4.2.2. The R Foundation for Statistical Computing, Vienna, Austria).8 P<0.05 was considered significant for all analyses. Table 1 includes the characteristics of the 225 patients in this study. The median age at HSCT was 8 years old (range, 1–15), and the median follow-up period for survivors was 7.95 years (range, 0.18–19.2). The outcomes of the included patients are summarized in Figure 1A. The univariate analysis for EFS identified the duration of CR1, FAB classifications, cytogenetic/genetic classifications, and conditioning regimens as prognostic factors (the last one non-significant), while the duration of CR1 was unknown for 26.2% of patients (Table 1). We also performed univariate analyses for OS, the cumulative incidence of relapse, and the cumulative incidence of NRM (Online Supplementary Table S1). Among conditioning regimens, apart from “other MAC”, which was administered for only three patients, busulfan/melphalan-based MAC showed the best outcomes, with a 5-year-EFS (5yEFS) of 75.8% (95% confidence interval [CI]: 58.5–86.7),

Haematologica | 109 January 2024

312

LETTER TO THE EDITOR Table 1. Univariate analysis for event-free survival. Factor Age in years at HSCT

Duration of the first CR

Interval from the CR2 to HSCT

FAB classification

Year of HSCT Cytogenetics/genetics

Conditioning regimen

GvHD prophylaxis

HLA disparities

Donor source

Donor-recipient sex match

Group

N (%)

5y-EFS (95% CI)

0-4

73 (32.4)

58.4 (45.8-69.0)

0.238

5-9

62 (27.6)

67.7 (53.6-78.4)

10-15

90 (40.0)

61.1 (49.8-70.6)

0-1

189 (84.0)

63.1 (55.5-69.8)

2-4

8 (3.6)

62.5 (22.9-86.1)

28 (12.4)

55.8 (35.5-72.0)

≤1 year

23 (10.2)

36.0 (17.0-55.6)

>1 year

143 (63.6)

68.9 (60.1-76.1)

59 (26.2)

55.9 (42.0-67.7)

≤79 days

87 (38.7)

57.2 (45.3-67.4)

>79 days

82 (36.4)

69.2 (57.6-78.3)

56 (24.9)

59.0 (44.6-70.8)

9 (4.0)

60.0 (19.5-85.2)

31 (13.8)

64.2 (44.7-78.4)

92 (40.9)

66.4 (55.3-75.4)

32 (14.2)

71.7 (52.7-84.2)

34 (15.1)

59.2 (38.8-74.8)

5 (2.2)

53.3 (6.8-86.3)

16 (7.1)

18.8 (4.6-40.2)

Others

2 (0.9)

50.0 (0.6-91.0)

4 (1.8)

2000-2009

126 (56.0)

60.4 (51.2-68.4)

2010-2019

99 (44.0)

64.0 (52.6-73.4)

CBF

78 (34.7)

72.2 (60.6-80.9)

11q23*

22 (9.8)

43.8 (21.4-64.3)

HR cytogenetics**

25 (11.1)

36.4 (17.8-55.3)

Others

100 (44.4)

63.8 (52.9-72.9)

TBI/Cy based MAC

81 (36.0)

49.5 (38.0-59.9)

Bu/Cy based MAC

11 (4.9)

60.0 (25.3-82.7)

TBI/Mel based MAC

48 (21.3)

66.2 (50.8-77.7)

Bu/Mel based MAC

40 (17.8)

75.8 (58.5-86.7)

Mel based MAC

27 (12.0)

75.3 (48.7-89.4)

other MAC

3 (1.3)

100.0 (NA)

RIC

15 (6.7)

52.4 (22.0-75.9)

CSA ± MTX

71 (31.6)

56.0 (43.3-66.9)

TAC ± MTX

146 (64.9)

64.3 (55.5-71.9)

8 (3.6)

75.0 (31.5-93.1)

78 (34.7)

60.7 (48.5-70.9)

71 (31.6)

64.1 (51.0-74.6)

47 (20.9)

62.3 (45.5-75.2)

29 (12.9)

58.2 (38.3-73.8)

Rel-BM

55 (24.4)

54.2 (39.5-66.8)

Rel-PB

21 (9.3)

61.1 (34.5-79.6)

UR-BM

70 (31.1)

67.7 (54.8-77.7)

UR-CB

79 (35.1)

62.0 (49.7-72.1)

Match

104 (46.2)

59.4 (49.0-68.5)

Male to Female

60 (26.7)

59.7 (45.2-71.4)

Female to Male

44 (19.6)

59.5 (42.1-73.2)

17 (7.6)

88.2 (60.6-96.9)

0.785

0.001

0.320

0.005

0.455 0.003

0.088

0.642

0.848

0.391

0.204

Continued on following page. Haematologica | 109 January 2024

313

LETTER TO THE EDITOR

*The 11q23 group included those with no information on the rearrangement partner of KMT2A or those who did not meet the criteria of highrisk (HR) cytogenetics; 2 patients with t(11;19)(q23;p13.1) and 1 patient with t(9;11)(p22:q23). **HR cytogenetics found in the current study included 20 patients with complex karyotypes, 2 patients each with t(6;11)(q27;q23) or 7-/7q-, and 1 patient each with t(9;22)(q34;q11.2), t(7;12)(q36;p13), or FLT3-internal tandem duplication. Among them, 1 patient had both a complex karyotype and deletion of chromosome 7, and another patient had both a complex karyotype and t(7;12)(q36;p13). 5y-EFS: 5-year event-free survival; CI: confidence interval; HSCT: hematopoietic stem cell transplantation; PS: performance status; CR: complete remission; FAB: French-American-British; CBF: core-binding factor; TBI: total-body irradiation; Cy: cyclophosphamide; Mel: melphalan; MAC: myeloablative conditioning; RIC: reduced-intensity conditioning; CSA: cyclosporine A; TAC: tacrolimus; MTX: methotrexate; HLA: human leukocyte antigen; Rel: related; UR: unrelated; BM: bone marrow; PB: peripheral blood; CB: cord blood; GvHD: graft-versus-host disease; NA: not available.

Table 2. Multivariate analysis for event-free survival. Factor Age in years at HSCT

Duration of the first CR Year of HSCT Conditioning regimen Donor source

HLA disparities

FAB Cytogenetics

0-4

Hazard ratio (95% CI) ref

5-9

0.83 (0.37-1.85)

0.652

10-15

0.98 (0.42-2.31)

0.963

≤1 year

ref

>1 year

0.35 (0.18-0.70)

2000-2009

ref

2010-2019

1.09 (0.61-1.96)

Others

ref

Mel-containing non-TBI MAC

0.42 (0.20-0.86)

Rel-BM

ref

Rel-PB

1.27 (0.47-3.44)

0.643

UR-BM

0.56 (0.22-1.41)

0.219

UR-CB

0.64 (0.32-1.28)

0.207

ref

0.80 (0.39-1.63)

0.533

0.66 (0.30-1.47)

0.312

Others

ref

2.57 (1.06-6.20)

Others

ref

CBF

0.66 (0.30-1.41)

0.278

11q23*

1.62 (0.67-3.93)

0.286

HR cytogenetics**

1.81 (0.75-4.35)

0.186

0.003 0.769 0.018

0.036

*The 11q23 group included those with no information on the rearrangement partner of KMT2A or those who did not meet the criteria of highrisk (HR) cytogenetics; 2 patients with t(11;19)(q23;p13.1) and 1 patient with t(9;11)(p22:q23). **HR cytogenetics found in the current study included 20 patients with complex karyotypes, 2 patients each with t(6;11)(q27;q23) or 7-/7q-, and 1 patient each with t(9;22)(q34;q11.2), t(7;12)(q36;p13), or FLT3-internal tandem duplication. Among them, 1 patient had both a complex karyotype and deletion of chromosome 7, and another patient had both a complex karyotype and t(7;12)(q36;p13). CI: confidence interval; HSCT: hematopoietic stem cell transplantation; CR: complete remission; Mel: melphalan; TBI: total-body irradiation; MAC: myeloablative conditioning; Rel: related; UR: unrelated; BM: bone marrow; PB: peripheral blood; CB: cord blood; HLA: human leukocyte antigen; FAB: French-American-British; CBF: core-binding factor; NA: not available; ref: reference.

followed by melphalan-based non-TBI MAC, with a 5y-EFS of 75.3% (95% CI: 48.7–89.4; Table 1). The detailed distributions among melphalan-based non-TBI MAC and busulfan/melphalan-based MAC regimens are presented in the Online Supplementary Table S2. The 5y-EFS of the patients receiving chemotherapy-based MAC was significantly superior to that of patients receiving TBI-based MAC (74.4% vs. 56.2%, respectively; P=0.010). The multivariate analysis for EFS identified a duration of CR1 of more than 1 year (hazard ratio [HR] =0.35; P=0.003) and a melphalan-containing non-TBI MAC regimen

(HR=0.42; P=0.018) as independent favorable prognostic factors, while M7 was identified as an independent adverse prognostic factor (HR=2.57; P=0.036; Table 2). These results were similar when conditioning regimens were classified into three groups (TBI-MAC, chemotherapybased MAC, and reduced intensity conditioning; Online Supplementary Table S3). This study focused specifically on patients who had received chemotherapy alone in their CR1, and all included patients underwent HSCT for the first time after achieving their CR2. This finding implies that patients in this cohort

Haematologica | 109 January 2024

314

LETTER TO THE EDITOR

had a viable opportunity for treatment intensification with HSCT in their CR1. Additionally by identifying favorable prognostic factors related to HSCT, we can further optimize HSCT parameters in the CR2 as well. Our analysis revealed that the outcomes for patients diagnosed with AML M7 who relapsed after chemotherapy alone were particularly poor, even after achieving CR2. Recently, Hama et al.9 demonstrated that among those with non-DS M7, patients who underwent HSCT in their CR1 showed significantly better survival than those who underwent HSCT in their CR2 or those who did not achieve a CR. In the current study, the poor outcomes for those with M7 were reproduced in a more recent cohort (Figure 1B). Furthermore, non-DS M7 is a heterogeneous disease, and HSCT in the CR1 is recommended at least for high-risk groups in non-DS M7, such as those with CBFA2T3::GLIS2 or NUP98::KDM5A.10 A prospective trial to evaluate the efficacy of HSCT in the CR1 for this population is currently underway in Japan (JCCG AML-20 study: jRCTs041210015). In pediatric patients with AML, chemotherapy-based MAC regimens have been reported to be as effective, or even more effective, than TBI-MAC regimens.11 Our analysis further demonstrated that chemotherapy-based MAC showed comparable rates of relapse and reduced rates of NRM, resulting in improved EFS in the univariate and multivariate analyses compared with TBI-based MAC (Online Supplementary Tables S1 and S3). Excellent outcomes of chemotherapy-based MAC regimens containing melphalan have been demonstrated in previous studies,12–14 and these regimens have been associated with a superior outcome compared with other chemotherapy-based regimens11,15 for children with AML. In fact, European groups have adopted the use of busulfan, cyclophosphamide, and melphalan as their standard MAC regimen for pediatric AML.12 Our results also demonstrated that patients receiving the melphalan-containing non-TBI MAC regimen achieved superior outcomes compared with those of patients receiving other regimens in their CR2 (Table 2); however, none of these patients received cyclophosphamide (Online Supplementary Table S2). Further research is needed to determine the optimal combination of agents to add to melphalan in the treatment of relapsed AML. Along with adverse prognostic factors, we also identified several factors that did not alter outcomes. The year of HSCT was previously suggested to be a prognostic factor,4 but this was not observed in our study. Furthermore, as in previous reports,3,12 alternative-donor HSCT was feasible for these populations (Table 1; Online Supplementary Table S1). This information is also valuable for performing HSCT in a timely manner after achieving the CR2. This study has several limitations. First, we identified non-DS M7 as an unfavorable prognostic factor, but we

did not have information on the underlying genetic aberrations,10 and further studies including genetic information are warranted. Second, as a retrospective study, there may be selection bias among groups. We performed a multivariate analysis to adjust for the effects of known prognostic factors, including the age at diagnosis, duration of CR1, year of HSCT, donor sources, and cytogen-

Figure 1. The survival outcomes of the included patients. (A) The outcomes of all included patients. (B) Event-free survival of patients with French-American-British (FAB) M7 disease and patients with other types of acute myeloid leukemia (AML). 5yEFS: 5-year event-free survival; OS: overall survival; CIR: cumulative incidence of relapse; CINRM: cumulative incidence of non-relapse mortality.

Haematologica | 109 January 2024

315

LETTER TO THE EDITOR etic/genetic abnormalities. We identified melphalan-containing non-TBI MAC as being superior to other regimens (Table 2), and chemotherapy-based MAC was superior to TBI-based MAC (Online Supplementary Table S3). However, other potential prognostic factors, such as the details of pre-HSCT treatment or pre-HSCT minimal residual disease status, were not included due to a lack of information. Third, the current practice in Japan differs from those in other countries in several ways, such as more common usage of cord blood or busulfan in the context of chemotherapy-based regimens. As such, careful consideration should be given to apply these results in the practices of other countries. Fourth, it should be noted that the duration of CR1 was unknown for 26.2% of patients. In summary, patients with relapsed non-DS M7 exhibited poor outcomes after HSCT in their CR2, underscoring the need to identify high-risk subpopulations within this group and to consider offering HSCT in the CR1. Furthermore, melphalan-containing non-TBI MAC regimens demonstrated superior outcomes compared with other conditioning regimens, and alternative-donor HSCT was a viable option for the population included in this study. These findings can aid in the optimization of HSCT strategies for children with AML in both the CR1 and CR2.

Cross Aichi Medical Center Nagoya First Hospital, Nagoya; 10 11

Department of Pediatrics, Hokkaido University Hospital, Sapporo;

Department of Hematology and Oncology, Shizuoka Children's

Hospital, Shizuoka; 12Department of Pediatrics, National Hospital Organization Kyushu Cancer Center, Fukuoka; 13Department of Pediatrics, Graduate School of Medicine, Kyoto University, Kyoto; 14

Division of Pediatric Hematology and Oncology, Ibaraki Children's

Hospital, Mito; 15Department of Pediatrics, Osaka International Cancer Institute, Osaka; 16Department of Pediatrics, Tokyo Metropolitan Cancer and Infectious Disease Center, Komagome Hospital, Tokyo; 17Department of Pediatrics, Graduate School of Medicine, Chiba University, Chiba; 18Japanese Data Center for Hematopoietic Cell Transplantation, Nagakute; 19Department of Registry Science for Transplant and Cellular Therapy, Aichi Medical University School of Medicine, Nagakute and 20Department of Pediatrics, Graduate School of Medical and Dental Sciences, Kagoshima University, Kagoshima, Japan Correspondence: H. ISHIDA - hiishida1218@okayama-u.ac.jp https://doi.org/10.3324/haematol.2023.283203 Received: March 22, 2023. Accepted: June 30, 2023. Early view: July 20, 2023.

Authors

Hisashi Ishida,1 Shin-ichi Tsujimoto,2 Daisuke Hasegawa,3 Hirotoshi 4

Disclosures

Sakaguchi, Shohei Yamamoto, Masakatsu Yanagimachi, Katsuyoshi 7

No conflicts of interest to disclose.

Koh, Akihiro Watanabe, Asahito Hama, Yuko Cho, Kenichiro Watanabe,11 Maiko Noguchi,12 Masanobu Takeuchi,2 Junko Takita,13 1

Contributions

Kana Washio, Keisuke Kato, Takashi Koike, Yoshiko Hashii, Ken 16

18,19

Tabuchi, Moeko Hino, Yoshiko Atsuta

and Yasuhiro Okamoto

HI designed the study and analyzed the data. HI, ST, DH, HS, SY and

YO wrote the manuscript. MY, KaK, AW, AH, YC, KeW, MN, MT, JT, KaW, KeK, TK, YH, KT and MH collected patient data. HI and YA

Department of Pediatrics, Okayama University Hospital, Okayama;

performed the statistical analysis. All authors revised the

manuscript and approved the final version.

Department of Pediatrics, Yokohama City University Graduate 3

School of Medicine, Yokohama; Department of Pediatrics, St. Luke's International Hospital, Tokyo; 4Children's Cancer Center, National

Acknowledgments

Center for Child Health and Development, Tokyo; 5Department of

The authors are grateful for the assistance and cooperation of the

Pediatrics, Tokai University, Kanagawa; Division of

staff members of the collaborating institutes of the Japan Society

Hematology/Oncology, Kanagawa Children's Medical Center,

for Transplantation and Cellular Therapy.

Yokohama, Kanagawa; Department of Hematology and Oncology, Saitama Children's Medical Center, Saitama; 8Department of 9

Data-sharing statement

Pediatrics, Niigata Cancer Center Hospital, Niigata; Department of

Original data presented in this manuscript are available upon

Hematology and Oncology, Children's Medical Center, Japanese Red

request addressed to the corresponding author.

References 1. Zwaan CM, Kolb EA, Reinhardt D, et al. Collaborative efforts driving progress in pediatric acute myeloid leukemia. J Clin

Oncol. 2015;33(27):2949-2962. 2. Hasle H. A critical review of which children with acute myeloid

Haematologica | 109 January 2024

316

LETTER TO THE EDITOR leukaemia need stem cell procedures. Br J Haematol. 2014;166(1):23-33. 3. Karlsson L, Forestier E, Hasle H, et al. Outcome after intensive reinduction therapy and allogeneic stem cell transplant in paediatric relapsed acute myeloid leukaemia. Br J Haematol. 2017;178(4):592-602. 4. Egan G, Tasian SK. Relapsed pediatric acute myeloid leukaemia: state-of-the-art in 2023. Haematologica. 2023;108(9):2275-2288. 5. Atsuta Y. Introduction of Transplant Registry Unified Management Program 2 (TRUMP2): scripts for TRUMP data analyses, part I (variables other than HLA-related data). Int J Hematol. 2016;103(1):3-10. 6. Giralt S, Ballen K, Rizzo D, et al. Reduced-Intensity 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. 7. Creutzig U, van den Heuvel-Eibrink MM, Gibson B, et al. Diagnosis and management of acute myeloid leukemia in children and adolescents: recommendations from an international expert panel. Blood. 2012;120(16):3187-3205. 8. Kanda Y. Investigation of the freely available easy-to-use software ‘EZR’ for medical statistics. Bone Marrow Transplant. 2013;48(3):452-458. 9. Hama A, Taga T, Tomizawa D, et al. Haematopoietic cell transplantation for children with acute megakaryoblastic leukaemia without Down syndrome. Br J Haematol. 2023;201(4):747-756.

10. De Rooij JDE, Branstetter C, Ma J, et al. Pediatric non-Down syndrome acute megakaryoblastic leukemia is characterized by distinct genomic subsets with varying outcomes. Nat Genet. 2017;49(3):451-456. 11. Lucchini G, Labopin M, Beohou E, et al. Impact of conditioning regimen on outcomes for children with acute myeloid leukemia undergoing transplantation in first complete remission. An analysis on behalf of the Pediatric Disease Working Party of the European Group for Blood and Marrow Transplantation. Biol Blood Marrow Transplant. 2017;23(3):467-474. 12. Locatelli F, Masetti R, Rondelli R, et al. Outcome of children with high-risk acute myeloid leukemia given autologous or allogeneic hematopoietic cell transplantation in the aieop AML2002/01 study. Bone Marrow Transplant. 2015;50(2):181-188. 13. Beier R, Albert MH, Bader P, et al. Allo-SCT using BU, CY and melphalan for children with AML in second CR. Bone Marrow Transplant. 2013;48(5):651-656. 14. Versluys AB, Boelens JJ, Pronk C, et al. Hematopoietic cell transplant in pediatric acute myeloid leukemia after similar upfront therapy; a comparison of conditioning regimens. Bone Marrow Transplant. 2021;56(6):1426-1432. 15. Gorin N-C, Labopin M, Czerw T, et al. Autologous stem cell transplantation for adult acute myelocytic leukemia in first remission-Better outcomes after busulfan and melphalan compared with busulfan and cyclophosphamide: a retrospective study from the Acute Leukemia Working Party of the Europe. Cancer. 2017;123(5):824-831.

Haematologica | 109 January 2024

317

LETTER TO THE EDITOR

In-depth cytogenetic and immunohistochemical analysis in a real world cohort - reconsidering the role of primary and secondary aberrations in multiple myeloma Multiple myeloma (MM) arises from terminally differentiated antibody-producing B cells and evolves from monoclonal gammopathy of undetermined significance (MGUS) to overt disease through a sequence of genetic changes. These can be diverse and subdivided into primary and secondary events based on their occurrence. It is postulated that oncogenesis in MM is driven by two different almost exclusive primary genetic events (hyperdiploidy and immunoglobulin [Ig] H-translocation) and that clonal evolution occurs with the acquisition of different secondary chromosomal aberrations, e.g., monosomies, deletions, duplications and mutations.1 Via routine fluorescence in situ hybridization (FISH), hyperdiploidy is observed in ~50% of MM patients, whereas in the absence, MM is driven by class-switch recombination and somatic hypermutation during B-cell maturation in the loci encoding Ig on chromosome 14 resulting in chromosomal translocations.2 Monosomy 13/del13q14 is discussed as a possibly third type of primary aberration and appears in ~45% of MM patients.3 Some describe it as an independent MM-initiating event. Chesi et al.4 classified del13q as an early event, appearing in germinal centers, in MGUS and intramedullary MM. This finding was updated based on a mouse model, indicating that the loss of a particular gene on chromosome 13, Mir15a/Mir-16-1, regulates MM-progression in a copy numberdependent way, with a more aggressive tumor course in mice lacking both copies of the Mir-15a/Mir-16-1 cluster.4-8 A potential underlying mechanism may be the overexpression of cyclins D1-3 that are physiologically inhibited by the Mir-15a/Mir-16-1 genes. Cyclin D upregulation has been discussed as a unifying MM-initiating event by mediating the cell cycle transition from G1- to S-phase.9,10 Due to the ambiguous description of primary and secondary aberrations occurring alone or combined, and the varying role of monosomy 13/del13q14, further analyses seemed warranted. Therefore, we performed this exploratory, single center study at our Comprehensive Cancer Center Freiburg (CCCF) from 2013 to 2020. Of 338 patients, 29 had to be excluded, who did not fulfill the inclusion criteria, namely i) bearing the diagnosis of MM and ii) FISH diagnostics being routinely performed at initial diagnosis/initial presentation (ID). Thus, FISH data of 309 patients were prospectively assessed, with fully documented clinical data being associated with FISH and cyclin D1-3 results. Patients were assigned to four subgroups (SG); SG1: with only hyperdiploidy/IgH-translocations (n=74, 24%), SG2: without hyperdiploidy/IgH-transloca-

tions, but other cytogenetic aberrations (oCA) (n=40, 13%), SG3: both primary and secondary aberrations (n=182, 59%) and SG4: no aberrations detectable (n=13, 4%; Table 1). In depth cytogenetic, immunohistochemical and statistical analyses were carried out with SG2. Interphase FISH was routinely performed on CD138+ bone marrow (BM) plasma cells (PC) from ‘first pull’ BM aspirates using an extended FISH panel from MetaSystem Probes, Germany as previously described.11 An aberration was rated at a 10% cutoff-value.12 Progression-free survival (PFS) was defined as the time from ID to MM recurrence or death and overall survival (OS) from ID to death from any cause. Observation times for patients alive/without disease progression at the time of the analysis were censored at last follow-up (01/2023). Survival probabilities were estimated using the Kaplan-Meier method and compared with the log-rank test. All data were analyzed with SAS 9.4 (SAS Institute, Inc. Cary, North Carolina) with a P value of <0.05 considered as statistically significant. BM tissue sections (4 μm) were immunohistochemically analyzed as described.13 Samples were incubated with cyclin D1-3 antibodies. The staining patterns were independently and blindly assessed (SW with pathologists SC and M-AC). The study was performed according to the guidelines of the Declaration of Helsinki and Good Clinical Practice. All patients gave their written informed consent for institutionally initiated research studies conforming to the Institutional Review Board guidelines. The trial protocol was approved by the ethics committee of the University of Freiburg (EV 22-1525-S1) and registered at https://frks.uniklinik-freiburg.de/FRKS004358. Patients’ characteristics of the entire cohort and SG1-4 are summarized in Table 1. Our entire cohort revealed a median age of 65 years, Karnofsky Performance Status (KPS) of 80% and MM paraprotein types with preponderance of IgGκ-MM, with advanced International Staging System (ISS), Revised-ISS (R-ISS), Revised-Myeloma Comorbidity Index (R-MCI) and proteasome inhibitor (PI)based treatment in the majority of patients, reflecting a typical real-world cohort. SG2 showed light chain (LC)-MM frequencies of 35%, more λ-LC types (47%), increased ISS/R-ISS stage III (38%/25%), R-MCI frail patients (25%) and intense treatment: 75% received at least one autologous stem cell transplantation (ASCT) or even allogeneic SCT (allo-SCT). The best remission status during the disease course was rewarding: 42% achieved a complete remission.

Haematologica | 109 January 2024

318

84 (27)/160 (52)/65 (21)

R-MCI: fit, 0-3/intermediate-fit, 4-6/frail, 7-9

Haematologica | 109 January 2024

319 198 (64)/111 (36)

SCT: Yes/No

105 (0-431)

OS in months, median (range)

37 (93)

11 (28)/19 (47)/10 (25)

4 (10)/26 (65)/10 (25)

10 (24)/15 (38)/15 (38)

21 (53):19 (47)/0/0

17 (43)/9 (22)/0 14 (35)/0/0

90 (20-100)

23 (58)/17 (42)

62/63 (46-89)

Subgroup 2, N=40

88 (48)

35 (19)/103 (57)/44 (24)

29 (16)/109 (60)/44 (24)

42 (23)/52 (29)/88 (48)

109 (60): 68 (37)/3 (2)/2 (1)

93 (51)/33 (18)/1 (0.5) 52 (29)/2 (1)/1 (0.5)

80 (30-100)

104 (57)/78 (43)

66/65 (37-87)

Subgroup 3, N=182

117 (1-431)

49 (1-194)

56 (76)/ 18(24)

3 (4)/0/9 (12)/28 (38)/22 (30)/9 (12)/3 (4)

47 (64)/27 (36)

65 (88)/ 4 (6)/1 (1)/1 (1)/3 (4)

117 (0-288)

41 (0-288)

103 (0-387)

36 (0-149)

123 (68)/59 (32)

2 (1)/7 (4)/37 (20)/65 (36)/53 (29)/15 (8)/3 (2)

1 (2)/8 (20)/17 (42)/ (18)/2 (5)/3 (8)/2 (5) 27 (68)/13 (32)

109 (60)/73 (40)

160 (88)/ 10 (6)/4 (2)/6 (3)/2 (1)

30 (75)/10 (25)

31 (78)/5 (12)/3 (8)/0/1 (2)

3 (4)/17 (23)/38 (51)/16 (22) 1 (2)/17 (43)/9 (22)/13 (33) 2 (1)/91 (50)/32 (18)/57 (31)

29 (39)/35 (47)/10 (14)

24 (32)/42 (57)/8 (11)

32 (43)/25 (34)/17 (23)

52 (70):20 (27)/2 (3)/0

55 (75)/5 (7)/0 12 (16)/1 (1)/1 (1)

80 (30-100)

52 (70)/22 (30)

68/66 (29-86)

Subgroup 1, N=74

73 (14-88)

44 (7-72)

8 (62)/5 (38)

0/0/4 (31)/3 (23)/3 (23)/3 (23) /0

12 (92)/1 (8)

12 (92)/ 1 (8)/0/0/0

0/6 (46)/4 (31)/3 (23)

9 (69)/3 (23)/1 (8)

4 (31)/8 (62)/1 (7)

5 (39)/6 (46)/2 (15)

7 (54):6 (46)/0/0

8 (61)/1 (8)/0 4 (31)/0/0

80 (60-100)

11 (85)/2 (15)

58/55 (28-75)

Subgroup 4, N=13

Interphase fluorescence in situ hybridization (FISH) was routinely performed on CD138+ bone marrow (BM) plasma cells (PC) from ‘first pull’ BM aspirates using an extended FISH panel from MetaSystem Probes, Germany as described.11 [XL t(4;14) FGFR3/IGH DF; XL t(6;14) CCND3/IGH DF; XL t(11;14) MYEOV/IGH DF; XL t(14;16) IGH/MAF DF; XL t(14;20) IGH/MAFB DF; XL 5p15/9q22/15q22 hyperdiploidy, XCE 3/7/17; XL DLEU/LAMP; XL 19p/19q del; XL RUNX 1; XL CDKN2C/CKS1B, XL MYC BA, XCE 8; XL IGH BA; XL TP53/NF1; XL DLEU/TP53]. ID: initial diagnosis; KPS: Karnofsky Performance Status; MM: multiple myeloma; LC: light chain: ISS: International Staging system; R-ISS: Revised ISS; R-MCI: Revised Myeloma Comorbidity Index; ImiD: immunomodulatory drugs; SCT: stem cell transplantation; CR: complete remission; vgPR: very good partial response; PR: partial remission; SD: stable disease; PD: progressive disease; pts: patients; PFS: progression-free survival; OS: overall survival.

38 (0-288)

214 (69)/ 95(31)

PFS in months, median (range)

Pts alive: Yes/No

6 (2)/9 (3)/58 (18)/113 (37)/85 (28)/29 (9)/9 (3)

268 (87)/ 20 (6)/8 (3)/7 (2)/6 (2)

First-line therapy by substance class: Proteasome-inhibitors/ alkylants/antibodies/IMiD/ Ø therapy

Best response: Ø therapy/ID/CR/ vgPR/PR/SD/PD

6 (2)/152 (49)/62 (20)/89 (29)

x-line therapy: 0/1/2/≥3

Therapy/outcome, N (%)

125 (41)

62 (20)/183 (59)/64 (21)

R-ISS: I/II/III

Frequency monosomy13/del13q14

89 (29)/98 (32)/122 (39)

ISS: I/II/III

Risk stratification, N (%)

κ:λ/both/asecretory LC type

IgG/IgA/IgD LC only/asecretory / biclonal

173 (56)/48 (15)/1 (0.3) 82 (27)/3 (1)/2 (0.7) 189 (61):113 (37)/5 (1.5)/2 (0.5)

80 (20-100)

KPS %, median (range)

Type of MM, N (%)

190 (61)/119 (39)

65/64 (28-89)

All patients, N=309

Sex: male/female, N (%)

Age in years at Initial diagnosis, median/mean (range)

Demographic data

Table 1. Patient characteristics of all patients and subgroups (SG) 1-4 with only hyperdiploidy/IgH-translocations (SG1) versus no hyperdiploidy/IgH-translocation, but other cytogenetic aberrations (SG2), both primary and secondary aberrations (SG3) versus no aberrations detectable (SG4).

LETTER TO THE EDITOR

LETTER TO THE EDITOR Monosomy 13/del13q14 was present in our entire cohort in 41% and in SG1-4 only in SG2 and SG3 in 93% and 48%, respectively (Table 1). Detailed cytogenetic analyses were carried out in SG2 since these patients were of particular interest (Figure 1A): apart from monosomy 13/del13q14 in 37 of 40 patients (93%), gain1q21 was present in 21 of 40 (53%), del14p32 in 19 of 40 (48%), del17p13/monosomy 17 in ten of 40 (25%), monosomy 8 in eight of 40 (20%), monosomy 1/del1p32 in ten of 40 (25%), myc-A/-R in nine of 40 (22%), single trisomies in nine of 40 (22%) and other deletions in nine of 40 patients (22%). PFS and OS were determined via Kaplan-Meier curves. Figure 1B depicts the PFS of SG1-4 with 49, 41, 36 and 44 months, respectively (P=0.4214). OS of SG1-4 is visualized in Figure 1C, showing no statistical significance (P=0.2985), albeit median OS in SG1 and 2 were 117 months, whilst in SG3 and SG4 only 103 and 73 months, respectively. In order to determine, whether - based on their monosomy 13/del13q14 status – SG2 patients showed differences in demographics, MM-types, ISS/R-ISS, R-MCI and outcome, we compared three subgroups as outlined in the Online Supplementary Table S1. Patients with monosomy 13/del13q14 alone (i) were the youngest, all males, fit via KPS and R-MCI, had predominantly LC-only and κLC-types. All patients remain alive and showed median PFS of 43 months and OS of not reached. Patients with monosomy 13/del13q14 and ≥1 additional cytogenetic event (ii) were older, showed a median KPS of 80% and were frail in 26%. ISS and R-ISS II/III were predominant in 73% and 88%, respectively. Median PFS and OS in this group were 41 and 116 months, respectively. Patients with exclusively secondary aberrations (iii) were the oldest, frail in 33%, and had advanced ISS/R-ISS II/III. Whilst two of three patients remain alive, median PFS and OS were dismal with 13 and 21 months, respectively. Due to the individual cytogenetic aberrations of SG2 and the varying role of monosomy 13/del13q14, SG2 was dissolved and regrouped in the remaining three SG1, SG3 and SG4, based on their monosomy 13/del13q14-status (Online Supplementary Table S1): patients with monosomy 13/del13q14 alone (i) were included in SG1 (IgH-translocation alone). Patients with monosomy 13/del13q14 with ≥1 additional cytogenetic event (ii) were regrouped in SG3 (both primary and secondary aberrations) and patients with exclusively secondary aberrations (iii) in SG4. This led to new SG1.1, SG3.1 and SG4.1 (n=77, 216, 16, respectively). The median PFS was 46, 37 and 38 months, respectively, which revealed a 9-month difference between SG1.1 and SG3.1, but did not reach significance (P=0.3628; Figure 1D). The median OS was 117, 105 and 65 months, respectively (Figure 1E), also failing to reach significance (P=0.0680). When PFS and OS in SG1.1 (n=77) versus merged SG3.1 and SG4.1 (n=232) were compared, P values of 0.1544 and 0.0262, respectively, were obtained (Figure 1F, G). Thus, patients with

hyperdiploidy/IgH-translocation/monosomy13/del13q14status alone (primary aberrations) versus those with primary and secondary aberrations (including patients with monosomy 13/del13q14 and ≥1 additional cytogenetic event and those with exclusively secondary aberrations) largely differed in OS (Figure 1G). In order to investigate the role of cyclin D1, D2 and D3 (Online Supplementary Table S2), immunohistochemistry (IHC) was performed in SG2 and SG4 (control) patients. Thereof, 32 of 40 SG2 patients were analyzed and compared to ten of 13 SG4 patients (n=11 could not be assessed due to missing BM samples). Figure 2A depicts the underlying mechanism possibly leading to overexpression of cyclins D1-3 based on the loss of a specific gene on chromosome 13q14, Mir-15a/Mir-16-1, which physiologically inhibits cyclins D1-3 in the cell cycle and transition from G1- to S-phase. Figure 2B shows representative BM sections with intense nuclear cyclin D1 and D2 and cytoplasmic D3 expression. Cyclin D1-3 in the control (SG4) revealed an upregulation of cyclin D3 in five of ten patients (50%) alone (Figure 2C), and in SG2 patients in one of 32 (3%), three of 32 (9%) and eight of 32 (25%), respectively (Figure 2D). Figure 2E allied the upregulation of cyclins D1-3 in SG2 patients with their individual cytogenetic aberrations, indicating a heterogeneous distribution without a distinct cyclin D upregulation pattern. Monosomy 13/del13q14 alone occurred in three patients (patient #1-3), but in almost all others in conjunction with various additional aberrations. They revealed a cyclin D3 positivity in one patient as well as negativity for cyclins D1-3 in two others. Monosomy 13/del13q14 with one, two, three, four or five additional cytogenetic aberrations occurred in seven, 12, seven, six and two patients, respectively (#4-37; n=34). The heterogeneous distribution in cyclin D13 upregulation was confirmed in this subgroup. Of interest, three patients (#38-40) with exclusively secondary aberrations, i.e., gain1q21, myc-A/-R, del17p13/monosomy 17 or del14q32, showed cyclin D3 upregulation in one patient, but did not express cyclins D1-3 in both others. Consequently, no correlation between cytogenetics of patients without hyperdiploidy/IgH-translocations, but oCA (SG2) and cyclins D1-3 via IHC became apparent. In summary, we initiated this study after observing symptomatic MM patients showing no hyperdiploidy/IgH-translocation, but oCA (termed SG2) over a 7-year period. Our interest was to address their frequency, MM-initiating events and outcome data. We found a remarkable fraction of patients without hyperdiploidy/IgH-translocations, but oCA (13%) who showed a predominance of monosomy 13/del13q14. Notable was that only few SG2 patients (3/309 i.e., 1%) did reveal an impaired PFS and OS, who showed no monosomy 13/del13q14, but secondary aberrations alone, suggesting them to belong to another genetic entity as compared to those with monosomy 13/del13q14 and secondary events (Online Supplementary Table 1). No distinct

Haematologica | 109 January 2024

320

LETTER TO THE EDITOR

Figure 1. Comparison of progression-free survival and overall survival. (A) Detailed cytogenetic aberration numbers of respective interphase fluorescence in situ hybridization subset of patients without IgH-translocations, but other aberrations. (B) Progression-free survival (PFS) of subgroup (SG)1 versus SG2 versus SG3 versus SG4 and (C) overall survival (OS) of SG1 versus SG2 versus SG3 versus SG4. (D) PFS of SG1.1 versus SG3.1 versus SG4.1 and (E) OS of SG1.1 versus SG3.1 versus SG4.1. (F) PFS of SG1.1 versus merged SG3.1+4.1 and (G) OS of SG1.1 versus merged SG3.1+4.1.

Haematologica | 109 January 2024

321

LETTER TO THE EDITOR

Continued on following page. Haematologica | 109 January 2024

322

LETTER TO THE EDITOR Figure 2. Immunohistochemical analysis of subgroups 2 and 4 (control group). (A) Relation between monosomy 13/del13q14, Mir15a/Mir-16-1 and cyclin D1-3 upregulation. (B) Photomicrographs (x20) showing positive and negative immunohistochemical staining patterns for cyclin D1, D2 and D3. (C) Cyclin D1-3 overexpression via immunohistochemistry (IHC) in subgroups (SG)4 and SG2. (D) cyclin D1-3 expression in relation to the cytogenetics of SG2.

pattern in cyclin D1-3 upregulation could be observed and, therefore, no clear association of cyclin D1-3 upregulation via inhibited Mir-15a/Mir-16-1 due to monosomy 13/del13q14 occurrence (Figure 2E). Monosomy 13/del13q14 has been discussed as a third MMinitiating event, in addition to hyperdiploidy or IgH-translocations.4,9,14-16 Del13 with loss of the microRNA Mir-15a/Mir-16-1, located on 13q14, was described as an early event in MM via a refined murine model.15 Due to the loss of Mir-15a/Mir-16-1, proteins mediating the transition from G1to S-phase in the cell cycle, like cyclin D1-3, were discussed to be overexpressed.8 Deregulation of cyclin D1-3 had been proposed as a unifying oncogenic event.9,17 We performed IHC in SG2 (and SG4 [control]) to determine cyclin D1-3, but observed upregulation of cyclins D1-3 in only 37% of SG2 and 50% in SG4 patients. Thus, deletion of 13q14 or the absence of the entire chromosome 13 with loss of Mir-15a/Mir16-1 did not mandatorily lead to overexpression of cyclins D1-3 and cyclin D upregulation was not the unifying MMinitiating event in our cohort. Additional analyses seem warranted to further elucidate the MM-initiating mechanisms in this subgroup, as chromosome 13 includes other driver genes, e.g., Dis3 or Rb1. Strengths of our analysis were the meticulous examination of newly diagnosed MM patients with clinical data and follow-up information. Our median observation period of 7 years was substantial, therefore, our genetic subgroup and outcome data robust. Cytogenetics were analyzed according to the S3-MM-guideline18 with an extended FISH panel. Even more in-depth cytogenetic analyses can now be performed via whole genome sequencing (WGS) to verify our findings, thereby detecting additional driver genes. Nevertheless, the accumulation of monosomy 13/del13q14 in SG2 was remarkable and as WGS is not necessarily performed in clinical routine yet, but patients without hyperdiploidy/IgH-translocations, but oCA and predominance of monosomy 13/del13q14 occurred quite frequently, an accessible risk stratification and outcome analysis seemed of interest. Limitations of our study were the single institution approach, yet due to strict inclusion criteria regarding patients’ and therapy data, all patients included provided infinitely detailed information. Another criticism could be the heterogeneity of patients as typical for tertiary centers. Since our CCCF-treated patient population was relatively young and the majority received ASCT, we refrained from non- versus ASCT-based subgroup analyses, but considered all patients as one group. Besides, one could criticize the use of other than bortezomib-cyclo-

phosphamide-dexamethasone (VCD)-induction in rare subgroups, however, we focused on cytogenetics in this study, therefore, we refrained from further subgroup analyses than those already extensively performed. Due to limited patients, it was not reasonable to further distinguish between monosomy 13 and del13q14, although those events can be associated with a different prognosis.19 Certainly, SG2 with 13% of patients is a proportion that needs confirmation in even larger cohorts (i.e., within DSMM/GMMG, IFM, MRC or other study cohorts) before any definite assumptions are drawn. In conclusion, patients without hyperdiploidy/IgH-translocations, but oCA detected via FISH occurred in >10% of our patients presenting with newly diagnosed MM. Most of SG2 patients showed a predominance of monosomy 13/del13q14, suggesting that monosomy 13/del13q14 may act as a MM-initiating event in the absence of other primary events. Expanded analyses, preferably in even larger cohorts should verify our data and further elucidate the impact of hyperdiploidy, IgH-translocations or monosomy 13/del13q14 versus secondary aberrations and their underlying mechanisms.

Authors Sina Wenger,1,2 Milena Pantic,1,2 Gabriele Ihorst,2,3 Stepan Cysar,4 Marc-Antoine Calba,4 Christine Greil,1,2 Ralph Wäsch1,2 and Monika Engelhardt1,2 Department of Medicine I Hematology and Oncology, Faculty of

Medicine and Medical Center, University of Freiburg (UKF); 2Tumor Center Freiburg, Comprehensive Cancer Center Freiburg (CCCF); Clinical Trials Unit, Faculty of Medicine, University of Freiburg and

Department of Pathology, Faculty of Medicine, University of

Freiburg, Freiburg, Germany Correspondence: M. ENGELHARDT - monika.engelhardt@uniklinik-freiburg.de https://doi.org/10.3324/haematol.2023.283142 Received: March 17, 2023. Accepted: August 9, 2023. Early view: August 17, 2023. ©2024 Ferrata Storti Foundation Published under a CC BY-NC license

Haematologica | 109 January 2024

323

LETTER TO THE EDITOR Disclosures

Würzburg), Dr C. Miething (UKF), Dr M. Rassner (UKF, now postdoc in

No conflicts of interest to disclose.

Japan). We are also very thankful to all AG Engelhardt & Wäsch group members for their utmost MM enthusiasm, feedback and revision

Contributions

input. We thank the Center for biobanking (FREEZE-Biobank) for their

The results of this work are based on the results of the dissertation

support. We acknowledge the interdisciplinary multiple myeloma

of SW. SW acquired und analyzed the data, and both SW and ME

(MM) tumor board group of our CCCF, Prof Dr G. Herget, Dr H. Schäfer,

wrote the paper. MP provided the idea for this work and conducted

PD Dr J. Neubauer, Prof Dr Dr R. Schmelzeisen, Prof Dr K. Nelson, Prof

the FISH analyses. BVK assisted in the immunohistochemical

Dr G. Walz, and Prof Dr W. Kühn for their outstanding support. We

analysis. SW, SC and M-AC blindly assessed all bone marrow

greatly acknowledge the anonymous reviewers for their input and

samples. GI performed the statistical analysis. All authors discussed

recommendations. We are also indebted to IMWG, EMN and

the results and contributed to the final manuscript.

DSMM/GMMG experts for their insights on this data and manuscript, and thank all MM patients who participated in this study.

Acknowledgments The authors thank distinguished IMWG, EMN, DSMM and GMMG

Data-sharing statement

experts and friends. We are specially obliged to Prof Dr J. Duyster

The data that supports the findings of this study are available from

(Freiburg, UKF), Dr Dr J. Jung (TU München), Dr J. Waldschmidt (UK

the corresponding author upon reasonable request.

References 1. Rasche L, Kortüm K, Raab M, Weinhold N. The Impact of tumor heterogeneity on diagnostics and novel therapeutic strategies in multiple myeloma. Int J Mol Sci. 2019;20(5):1248. 2. Saxe D, Seo E-J, Beaulieu Bergeron M, Han J-Y. Recent advances in cytogenetic characterization of multiple myeloma. Int J Lab Hematol. 2019;41(1):5-14. 3. Avet-Loiseau H, Attal M, Moreau P, et al. Genetic abnormalities and survival in multiple myeloma: the experience of the Intergroupe Francophone du Myélome. Blood. 2007;109(8):3489-3495. 4. Chesi M, Bergsagel PL. Advances in the pathogenesis and diagnosis of multiple myeloma. Int J Lab Hematol. 37(Suppl 1):S108-114. 5. Chesi M, Stein CK, Garbitt VM, et al. Monosomic loss of MIR15A/MIR16-1 is a driver of multiple myeloma proliferation and disease progression. Blood Cancer Discov. 2020;1(1):68-81. 6. Walker BA. The chromosome 13 conundrum in multiple myeloma. Blood Cancer Discov. 2020;1(1):16-17. 7. Li F, Xu Y, Deng S, et al. MicroRNA-15a/16-1 cluster located at chromosome 13q14 is down-regulated but displays different expression pattern and prognostic significance in multiple myeloma. Oncotarget. 2015;6(35):38270-38282. 8. Roccaro AM, Sacco A, Thompson B, et al. MicroRNAs 15a and 16 regulate tumor proliferation in multiple myeloma. Blood. 2009;113(26):6669-6680. 9. Bergsagel PL, Kuehl WM, Zhan F, Sawyer J, Barlogie B, Shaughnessy J. Cyclin D dysregulation: an early and unifying pathogenic event in multiple myeloma. Blood. 2005;106(1):296-303. 10. Bergsagel PL, Kuehl WM. Critical roles for immunoglobulin translocations and cyclin D dysregulation in multiple myeloma. Immunol Rev. 2003;194:96-104. 11. Engelhardt M, Domm A-S, Dold SM, et al. A concise revised Myeloma Comorbidity Index as a valid prognostic instrument in a large cohort of 801 multiple myeloma patients.

Haematologica. 2017;102(5):910-921. 12. Ross FM, Avet-Loiseau H, Ameye G, et al. Report from the European Myeloma Network on interphase FISH in multiple myeloma and related disorders. Haematologica. 2012;97(8):1272-1277. 13. Mansoor A, Akhter A, Pournazari P, et al. Protein expression for novel prognostic markers (cyclins D1, D2, D3, B1, B2, ITGβ7, FGFR3, PAX5) correlate with previously reported gene expression profile patterns in plasma cell myeloma. Appl Immunohistochem Mol Morphol. 2015;23(5):327-333. 14. Manier S, Salem KZ, Park J, Landau DA, Getz G, Ghobrial IM. Genomic complexity of multiple myeloma and its clinical implications. Nat Rev Clin Oncol. 2017;14(2):100-113. 15. Chesi M, Stein CK, Garbitt VM, et al. Monosomic loss of MIR15A/MIR16-1 Is a driver of multiple myeloma proliferation and disease progression. Blood Cancer Discov. 2020;1(1):68-81. 16. Landgren O. Monoclonal gammopathy of undetermined significance and smoldering multiple myeloma: biological insights and early treatment strategies. Hematol Am Soc Hematol Educ Program. 2013;2013:478-487. 17. Thanh-Truc Ngo B, Felthaus J, Hein M, et al. Monitoring bortezomib therapy in multiple myeloma: screening of cyclin D1, D2, and D3 via reliable real-time polymerase chain reaction and association with clinico-pathological features and outcome. Leuk Lymphoma. 2010;51(9):1632-1642. 18. Piechotta V, Skoetz N, Engelhardt M, Einsele H, Goldschmidt H, Scheid C. Patients with multiple myeloma or monoclonal gammopathy of undetermined significance. Dtsch Arztebl Int. 2022;119(14):253-260. 19. Binder M, Rajkumar SV, Ketterling RP, et al. Prognostic implications of abnormalities of chromosome 13 and the presence of multiple cytogenetic high-risk abnormalities in newly diagnosed multiple myeloma. Blood Cancer J. 2017;7(9):e600.

Haematologica | 109 January 2024

324

LETTER TO THE EDITOR

Inflammation is predictive of outcome in Waldenström macroglobulinemia treated by Bruton tyrosine kinase inhibitors: a multicentric real-life study Waldenström macroglobulinemia (WM) is a chronic indolent mature B-cell lymphoma characterized by bone marrow infiltration of lymphoplasmacytic cells and a monoclonal immunoglobulin M.1 The most frequent somatic anomaly in WM is a gain of function mutation of MYD88 (MYD88 L265P), present in 90% of WM at diagnosis leading to constitutive activation of NF-kB and JAK/STAT pathways,1 followed by 6q deletion (del6q, 30-55%)2,3 and CXCR4 gain-of-function mutations (30-40%).3 The therapy is based on chemo-immunotherapy (CIT) or Bruton tyrosine kinase inhibitors (BTKi). The numerous prognostic scores, specific of WM, did not impact the therapeutic choice in clinical practice.4,5 Based on WM and inflammatory disorder association, the prognosis impact of inflammation in WM during CIT was recently highlighted by two independent teams.6,7 In the first cohort, inflammatory syndrome (C-reactive protein [CRP] ≥20 mg/L) was present in one third of the patients.6 This inflammatory syndrome decreased during treatment but was associated with a shorter time to next treatment (TTNT, 1.6 years vs. 4.8 years; P<0.001). In the second cohort, inflammation (CRP >5 mg/L) was associated with more frequent del6q and, more frequent need for treatment initiation, inflammation decrease during CIT and a trend towards poorer progression-free survival (PFS) and overall survival (OS).7 Importantly, the outcome of inflammatory WM (iWM) was evaluated during CIT, whereas data about BTKi therapy and inflammation were lacking. Therefore, we performed a multicentric cohort to evaluate the impact of BTKi on iWM. In this real-life cohort, the inflammation positively impacts the prognosis of WM with BTKi. The pooled cohort used in this study was based on published WM cohorts from Saint Louis6 (n=268) and Pitie-Salpetriere Hospitals (n=270)7 with the addition of a Necker cohort Hospital (n=110). MYD88 L265P and CXCR4 S338X mutations were evaluated by restriction-fragment-length polymorphism and allele-specific polymerase chain reaction or targeted next-generation sequencing, respectively, as previously described.8 iWM was defined by the presence of two CRP measures ≥20 mg/L without other causes to explain inflammatory syndrome (e.g., infection, inflammatory complication). We excluded patients without CRP measurement before treatment. The CIT cohort at second line used as control was extracted from the Saint-Louis cohort.6 The response was assessed according to the sixth International Workshop on Waldenström's Macroglobulinemia.9

Patient data were obtained in conformity with the Declaration of Helsinki and registered by the Assistance-PubliqueHôpitaux-de-Paris data protection office. OS was defined as the time from BTKi initiation to death from all causes. PFS was defined as the time from BTKi initiation and its discontinuation for any reason (progression or toxicity) or death. TTNT was defined as the time from BTKi or CIT initiation to disease progression requiring treatment or death from any cause. All quantitative variables were described using medians (quartiles), while qualitative variables were described by frequencies (percentage). Categorical and quantitative data were compared using Fisher’s exact test and Student t test, respectively. Kaplan–Meier curves were plotted for survival, and data for the various groups were compared using the log-rank test. Maximally selected-rank statistics were used to select the best cutoff for CRP based on TTNT survival. Statistical analyses were obtained using R 4.0.4. Among 648 WM patients from the pooled cohort, 474 (73%) WM patients received CIT including 398 (84%) with multiple CRP measures before treatment (Online Supplementary Table S1A). We confirmed the inferior outcome after CIT of iWM, defined by CRP ≥20 mg/L, based on TTNT at three levels (Online Supplementary Figure S1A) and maximally selected-rank statistics (Online Supplementary Figure S1B). CXCR4 mutation was not associated with TTNT (hazard ratio [HR] =1; 95% confidence interval [CI]: 0.65-1.6; P=0.96). iWM was associated with more del6q (49% vs. 20%; P<0.001; Online Supplementary Figure S1C) and less CXCR4 mutation than non-iWM (15% vs. 34%; P=0.02; Online Supplementary Figure S1D). Seventy-eight WM patients who received BTKi between 2015 and 2022 were included (Figure 1A). After one exclusion for lack of CRP measurement, 42 (54%) patients were classified as iWM as previously described6 (Figure 1B). Among iWM, the median CRP before BTKi initiation was 37 mg/L (interquatile range [IQR], 26-60). The median age at WM diagnosis was 65 years (IQR, 56-73; Table 1). Median follow-up after BTKi initiation was 3.3 years (IQR, 1.7-5.9). Thirty-one (40%) patients had full characterization of MYD88/CXCR4/del6q. Ninety percent of the patients had MYD88 mutation. CXCR4 mutation was less frequent in iWM than in non-iWM (17% [4/24] vs. 50% [10/20]; P=0.04). Del6q was twice more frequent in iWM than non-iWM without reaching statistical significance (58% [14/24] vs. 33% [6/18]; P=0.19). No difference was observed for characteristics at

Haematologica | 109 January 2024

325

LETTER TO THE EDITOR diagnosis, first-line choice or BTKi initiation between iWM and non-iWM. Among BTKi, most patients received Ibrutinib. Half of the patients received BTKi on second line.

Concerning hematologic response, the overall response rate (ORR = partial response/very good partial response/ complete response [CR]) was superior in iWM than non-

C D

Figure 1. Inflammatory Waldenström macroglobulinemia (iWM) was associated with a higher response to BTK inhibitors than noniWM. Flowchart of the study. Seventy-eight patients received Bruton tyrosine kinase inhibitors (BTKi) for WM, including 42 (54%) with inflammatory syndrome (C-reactive protein [CRP] ≥20 mg/L). (B) Histogram and density plot for CRP distribution at BTKi initiation with the cutoff 20 mg/L. Two peaks among patients were observed at 0-5 mg/L and at 20 mg/L. (C) Hematologic response based on the sixth International Workshop on Waldenström's Macroglobulinemia criteria between non-iWM and iWM patients. Green bar represents overall response rate (= partial response [PR]/very good partial response [VGPR]/complete response [CR] response), the olive green bar for minimal response (MR) and orange bar for progressive disease (PD)/stable disease (SD). (D) Inflammation evolution in iWM (N=37) during BTKi treatment. Orange lines show patients who had SD/PD. Green lines show patients who had MR/VGPR/PR/CR. For SD/PD patients (N=5) in iWM, two patients had persistent CRP ≥20 mg/L, but received less than 4 months of BTKi. Among patients with clinical signs (N=17), all patients had B-symptom regression except for 2 iWM (1 with hematologic SD and 1 with VGPR but high CRP level at 19 mg/L who had a progression at 9 months) and 1 non-iWM patient who developed diffuse large B-cell lymphoma 3 months after BTKi introduction. Haematologica | 109 January 2024

326

LETTER TO THE EDITOR iWM (82% vs. 52%; P=0.02; Figure 1C). Regarding the inflammation syndrome kinetic, all iWM patients who reached minimal response (MR) or better on BTKi had a nadir of CRP <20 mg/L (Figure 1D). Sixteen patients (43%) obtained normalization of CRP (<5 mg/L). In addition, 91% (34/37) obtained a 50% decrease or more in their CRP level. Among patients with B-symptoms, 82% (14/17) had regression during BTKi treatment. Furthermore, iWM patients had better PFS than non-iWM upon BTKi treatment (Figure 2A; median: 4 years vs. 2.4

years; P=0.0025). The leading cause of BTKi discontinuation was progression for non-iWM (n=11, 51%) and BTKi toxicity for iWM (n=8, 47%; Table 1). The cumulative incidence of progression was superior for non-iWM than iWM (4years: 43% vs. 21%; P=0.05; Figure 2B). There was no difference in discontinuation due to BTKi toxicity between non-iWM and iWM (4 years: 37% vs. 34%; P=0.11). Also, the TTNT survival was superior for iWM than non-iWM (median: 4 vs. 2.6 years; P=0.008; Figure 2C). TTNT survival and maximally selected rank statistics confirmed the rel-

Table 1. Initial characteristics of non-inflammatory Waldenström macroglobulinemia and inflammatory Waldenström macroglobulinemia patients. Non-iWM N=35

iWM N=42

21 (60)

29 (69)

0.556

63 (55-75)

66 (59-72)

0.449

At diagnosis Men, N (%) Age in years at WM diagnosis, median (IQR) Prior inflammatory disease, N (%)

4 (11)

2 (5)

0.402

κ/λ, N (%)

30/5 (85/15)

38/4 (90/10)

0.724

ISSWM score, N (% of the complete case) Low Intermediate High

14/35 (40) 3 (21) 5 (36) 6 (43)

19/42 (45) 1 (5) 3 (16) 15 (79)

0.107

LDH >ULN, N (%)

5/26 (19)

11/31 (36)

0.251

6 (19)

4 (12)

0.511

MYD88 mutation, N=54, 70%, N (%)

21/23 (91)

28/31 (90)

1.000

CXCR4 mutation, N=44, 57%, N (%)

10/20 (50)

4/24 (17)

0.041

Deletion 6q, N=42, 55%, N (%)

6/18 (33)

14/24 (58)

0.190

First line treatment, N (%) Rituximab + alkylating agent Rituximab monotherapy Chlorambucil BTKi Others

24 (69) 0 6 (17) 1 (3) 4 (11)

26 (62) 3 (7) 10 (24) 0 3 (7)

First line response,$ N (% complete case) PD/SD MR PR/VGPR/CR

33/35 (94) 9 (27) 5 (15) 19 (58)

37/42 (88) 6 (16) 6 (16) 25 (68)

0.526

73 (63-82)

75 (70-80)

0.344

B-symptoms before BTKi initiation, N (%)

4 (11)

13 (31)

0.075

Delay in years WM diagnosis - BTKi initiation, median (IQR)

7 (4-11)

8 (4-12)

0.375

2019 (2017-2020)

2018 (2017-2020)

0.580

Line of BTKi initiation, N (%) 1 2 3 4+

1 (3) 18 (51) 10 (29) 6 (17)

0 21 (50) 9 (21) 12 (29)

Ibrutinib/other BTKi, N (%)

34/1* (97/3)

41/1** (98/2)

Platelets <100x10 /L, N (%)

0.176

BTKi initiation Age in years at BTKi initiation, median (IQR)

Year of BTKi initiation, median (IQR)

0.457

0.706

Acalabrutinib; **zanubritinib; $based on IMWW-6; #2 inactive type 2 cryoglobulinemia, 1 inactive giant cell arteritis, 1 inactive ulcerative colitis; 1 inactive bullous dermatosis, 1 scleroderma. BTKi: Bruton tyrosine kinase inhibitors; CR: complete response; ISSWM: International Prognosis Score System for Waldenström Macroglobulinemia; MR: minimal response; PD: progressive disease; PR: partial response; SD: stable disease; VGPR: very good partial response; iWM: inflammatory Waldenström macroglobulinemia; LDH: lactate dehydrogenase; ULN: upper limit of normal; IQR: interquartile range.

Haematologica | 109 January 2024

327

LETTER TO THE EDITOR evance of the 20 mg/L CRP cutoff (Online Supplementary Figure 2A, B). In univariate analysis, TTNT was associated positively with inflammatory syndrome (HR=0.43; 95% CI: 0.22-0.81; P=0.01), negatively with CXCR4 mutations (HR=3.8; 95% CI: 1.5-9.6; P=0.01) and platelets <100x109/L (HR=2.38; 95% CI: 1.02-5.55; P=0.044; Online Supplementary Table 1B). Del6q was not associated with TTNT (HR=1.3; 95% CI: 0.49-3.4; P=0.60). Multivariate analysis was not performed because of a low number of events (n=18). No difference was observed for OS (4 years: 75% vs. 66%; P=0.15; Online Supplementary Figure 2C).

In order to evaluate the impact of inflammation in the current recommendation of BTKi in second line, TTNT survival analysis was performed with a focus on the second line of treatment (CIT or BTKi) between iWM and non-iWM (Figure 2D; P=0.012). No difference was observed for demographic/disease characteristics between BTKi and CIT cohorts (data not shown). Among patients receiving CIT at second line, most patients received alkylating agents +/- rituximab (55%) followed by chlorambucil (22%). iWM treated with BTKi (51% at 4 years) had better survival than BTKi treated non-IWM (22%), whereas 4-year survival in patients

Figure 2. Inflammatory Waldenström macroglobulinemia (iWM) had an improvement of time to next treatment during BTK inhibitor treatment than non-iWM. (A) Progression-free survival (PFS) between non-iWM (blue) and iWM (red). (B) Cumulative incidence of the reason for Bruton tyrosine kinase inhibitor (BTKi) discontinuation (progression in a straight line and toxicity in a dotted line) between non-iWM (blue) and iWM (red). (C) Time to next treatment (TTNT) after BTKi for all cohorts between noniWM (blue) and iWM (red) (D) TTNT after the 2nd line of treatment (BTKi in a straight line and chemo-immunotherapy [CIT] in dotted line) in non-iWM (blue) and iWM (red). Haematologica | 109 January 2024

328

LETTER TO THE EDITOR treated with CIT was 16% in iWM and 29% in non-iWM. To the best of our knowledge, this is the first study evaluating inflammation and response to BTKi in WM. We validated a cutoff of 20 mg/L of CRP for iWM definition and the association with poorer outcomes after CIT. However, we showed here that iWM with BTKi treatment had better hematological response and TTNT than non-iWM. In addition, the inflammatory syndrome decreased during the hematological response in BTKi-treated iWM. Several explanations could exist for these improved hematological response and survival. First, reduced inflammation via decreased pro-inflammatory cytokines after BTKi was described for SARS-COV-2,10 chronic graft-versus-host disease11 and Schnitzler syndrome.12 Second, BTK is expressed by malignant B cells13 but also by macrophages or monocytes.14 We can hypothesize that decreased inflammation mediated by BTKi might be related to the action on tumoral cells15 and/or microenvironment.14 Also, 6q chromosome contains an inhibitor of BTK.3 Deletion of 6q, more frequent in iWM, could thus lead to BTK activation, corrected by BTKi treatment. CXCR4 mutation was less frequent in iWM than non-iWM (17% vs. 50%) and could explain a part of differential prognosis to BTKi. Additional study about the inflammation origin in WM would be required to understand BTKi action in iWM. One limit of our study is the small sample size for genetic analysis that limits the evaluation of interaction between CXCR4 mutation, del6q and inflammation. CXCR4 especially with high clonality (>25%), is the main adverse factor during BTKi therapy in WM.16,17 Related to the partial evaluation of CXCR4 mutation (52% of the cohort) without any clonality analysis, additional studies are necessary to evaluate CXCR4 mutation and inflammation prognosis role independently. Nevertheless, assessment of CRP could be easier, quicker and cheaper to perform than evaluation of CXCR4 mutational status and clonality analysis on bone marrow samples. Also, the retrospective design of our study is a limit, and we will need to confirm our findings in prospective large multicentric cohorts but also to reanalyze clinical trials of BTKi in WM in light of our results.16 In summary, inflammation appears to have a positive impact on the clinical outcome of WM patients on BTKi therapy. Thus, this study supports the use of BTKi in patients with iWM. On the other hand, inflammation could represent a novel biomarker for predicting the effects of BTKi in WM patients that can be easily and quickly evaluated in prospective cohorts and clinical trials.

Sylvain Choquet,3 Frederic Davi,5 Florence Nguyen-Khac,5,6 Wendy Cuccuini,7 Morgane Cheminant,4 Clotilde Bravetti,5 Gregory Lazarian,8 Sophie Kaltenbach,9 Olivier Hermine,4 Damien Roos-Weil,3 Marion Espéli,1# Karl Balabanian1# and Bertrand Arnulf2,10# 1

INSERM U1160 EMiLy, Institut de Recherche Saint-Louis, Université

Paris-Cité; 2Department of Immuno Hematology, Hospital SaintLouis, Assistance Publique Hopitaux de Paris; 3Department of Hematology, Sorbonne Université, Hospital Pitie-Salpetriere, Assistance Publique Hopitaux de Paris; 4Department of Hematology, Hospital Necker, Assistance Publique Hopitaux de Paris; 5Laboratory of Hematology, Hospital Pitie-Salpetriere, Assistance Publique Hopitaux de Paris; 6Centre de Recherche des Cordeliers, Sorbonne Université, Université Paris Cité, INSERM UMRS1138, Drug Resistance in Hematological Malignancies Team; 7

Laboratory of Cytogenetic, Hospital Saint Louis, Assistance

Publique Hopitaux de Paris; 8Laboratory of Hematology, Hospital Avicenne, Assistance Publique Hopitaux de Paris; 9Laboratory of Hematology, Hospital Necker, Assistance Publique Hopitaux de Paris and 10Department of Immunology, University of Paris Cité, Paris, France NF and DE contributed equally. ME, KB and BA contributed equally as senior authors.

Correspondence: P.-E. DEBUREAUX - Pierre-edouard.debureaux@aphp.fr B. ARNULF - Bertrand.arnulf@aphp.fr https://doi.org/10.3324/haematol.2023.283141 Received: March 16, 2023. Accepted: August 7, 2023. Early view: August 17, 2023. ©2024 Ferrata Storti Foundation Published under a CC BY-NC license Disclosures No conflicts of interest to disclose. Contributions P-ED, BA, DRW, ME and KB developed the concept and designed the study. P-ED, NF, DE, SH, BR, LF, MC, FD, FNK, WC, CB, GL, SK, OH, DRW and BA collected and assembled data. P-ED, BA, AT, DRW, OH, ME and KB analyzed and interpreted data. P-ED, NF, DE, SH, BR, LF, DRW, MC, OH, AT and BA took care of patients. P-ED, AT, DRW, BA, ME and KB wrote the article. All authors read and approved the final version of this manuscript.

Authors

Acknowledgments The authors thank the patients with WM and their families. The

Pierre-Edouard Debureaux, Nathalie Forgeard, 1,2

2,3+

Dikelele Elessa,

Stéphanie Harel, Laurent Frenzel, Bruno Royer, Alexis Talbot, 2

authors thank the nurses and physicians who cared for these patients and their families.

Haematologica | 109 January 2024

329

LETTER TO THE EDITOR Funding

www.opale.org; email: contact@opale.org). KB was a recipient of an

PED received financial support from ITMO Cancer of Aviesan within

FRM grant (EQU202203014627).

the framework of the 2021-2030 Cancer Control Strategy on funds administered by INSERM. The "EMiLy" U1160 INSERM unit is a

Data-sharing statement

member of the OPALE Carnot Institute, the Organization for

The data that support the findings of this study are available from

Partnerships in Leukemia (Institut Carnot OPALE, Institut de

the corresponding author upon reasonable request.

Recherche Saint-Louis, Hôpital Saint-Louis, Paris, France (htpp:

References 1. Quintanilla-Martinez L. The 2016 updated WHO classification of lymphoid neoplasias. Hematol Oncol. 2017;35(Suppl 1):S137-45. 2. Nguyen-Khac F, Lambert J, Chapiro E, et al. Chromosomal aberrations and their prognostic value in a series of 174 untreated patients with Waldenström’s macroglobulinemia. Haematologica. 2013;98(4):649-654. 3. Guerrera ML, Tsakmaklis N, Xu L, et al. MYD88 mutated and wild-type Waldenström’s Macroglobulinemia: characterization of chromosome 6q gene losses and their mutual exclusivity with mutations in CXCR4. Haematologica. 2018;103(9):e408-e411. 4. Morel P, Duhamel A, Gobbi P, et al. International prognostic scoring system for Waldenstrom macroglobulinemia. Blood. 2009;113(18):4163-4170. 5. Kastritis E, Morel P, Duhamel A, et al. A revised international prognostic score system for Waldenström’s macroglobulinemia. Leukemia. 2019;33(11):2654-2661. 6. Elessa D, Debureaux P, Villesuzanne C, et al. Inflammatory Waldenström’s macroglobulinaemia: a French monocentric retrospective study of 67 patients. Br J Haematol. 2022;197(6):728-735. 7. Forgeard N, Baron M, Caron J, et al. Inflammation in Waldenström macroglobulinemia is associated with 6q deletion and need for treatment initiation. Haematologica. 2022;107(11):2720-2724. 8. Krzisch D, Guedes N, Boccon-Gibod C, et al. Cytogenetic and molecular abnormalities in Waldenström’s macroglobulinemia patients: correlations and prognostic impact. Am J Hematol. 2021;96(12):1569-1579. 9. 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. 10. Treon SP, Castillo JJ, Skarbnik AP, et al. The BTK inhibitor ibrutinib may protect against pulmonary injury in COVID-19– infected patients. Blood. 2020;135(21):1912-1915. 11. Miklos D, Cutler CS, Arora M, et al. Ibrutinib for chronic graftversus-host disease after failure of prior therapy. Blood. 2017;130(21):2243-2250. 12. Claves F, Siest R, Lefebvre C, Valmary-Degano S, Carras S. Dramatic efficacy of ibrutinib in a Schnitzler syndrome case with Indolent lymphoma. J Clin Immunol. 2021;41(6):1380-1383. 13. Castillo JJ, Buske C, Trotman J, Sarosiek S, Treon SP. Bruton tyrosine kinase inhibitors in the management of Waldenström macroglobulinemia. Am J Hematol. 2023;98(2):338-347. 14. Liu X, Pichulik T, Wolz O-O, et al. Human NACHT, LRR, and PYD domain–containing protein 3 (NLRP3) inflammasome activity is regulated by and potentially targetable through Bruton tyrosine kinase. J Allergy Clin Immunol. 2017;140(4):1054-1067. 15. Hodge LS, Ziesmer SC, Yang Z-Z, Secreto FJ, Novak AJ, Ansell SM. Constitutive activation of STAT5A and STAT5B regulates IgM secretion in Waldenström’s macroglobulinemia. Blood. 2014;123(7):1055-1058. 16. Buske C, Tedeschi A, Trotman J, et al. Five-year follow-up of ibrutinib plus rituximab vs placebo plus rituximab for Waldenstrom’s macroglobulinemia: final analysis from the randomized phase 3 INNOVATE study. Blood. 2020;136(Suppl 1):S24-26. 17. 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.

Haematologica | 109 January 2024

330

LETTER TO THE EDITOR

Actin-bundling protein L-plastin promotes megakaryocyte rigidity and dampens proplatelet formation Normal hemostasis requires an adequate number of normally functioning platelets. A better understanding of platelet biogenesis may facilitate the expanding effort to manufacture platelets in vitro.1 Platelets develop from mature bone marrow megakaryocytes (MK) that derive from megakaryocyte progenitor (MKP) cells.2 In the classic model of bone marrow megakaryopoiesis, hematopoietic stem cells (HSC) in the proliferative niche give rise to MK-erythroid progenitors that differentiate into MKP during migration to the vascular nice.3 The cytoplasm of mature MK has an abundant invaginated membrane system (IMS) that is contiguous with the plasma membrane.4,5 During the late stages of thrombopoiesis, the IMS begins extrude from the MK plasma membrane and form long extensions of proplatelets (PP) that eventually give rise to platelets released in the vasculature.6,7 Cell migration and signaling requires alterations in the organization and thickness of the cell cortex, which is comprised of a network of overlapping, bundled and crosslinked actin filaments.8 Filamentous actin (F-actin) organization regulates MK shape, MKP migration, formation of the IMS, early stage of PP formation (PPF), and later stages of PP branching. Despite this understanding of various stages of MK differentiation, the molecular mechanisms that signal the MK to initiate PPF are not well understood. Recently, our laboratory used an unbiased transcriptomic approach to identify an actin bundling protein, L-plastin, that decreased during MK differentiation, correlating with the onset of PPF.9 Using CRISPR/Cas to generate CD34+-derived human MK deficient in L-plastin, we found impaired MK migration in response to SDF-1α chemoattractant, increased podosome formation per MK, increased IMS and increased PPF. F-actin polymerization and depolymerization (i.e., “re-organization”) are dynamic reactions central to MK shape, migration, IMS formation, etc. F-actin also facilitates degradation of the extracellular matrix at sites of podosome formation.10 The MK cytoskeleton, comprised of a dense network of actin filaments and microtubules, is located at the periphery of the cell cytoplasm, below the plasma membrane. The manner by which the loose lipid-rich IMS is able to penetrate the dense actin cytoskeleton to initiate PPF has not been addressed. The prior work on MK L-plastin was limited by inter-individual variation affecting platelet number and artifacts of in vitro culturing.9 Because MK L-plastin has not been studied in vivo, and because L-plastin is

expected to increase the density of the MK actin cortex via actin bundling, we used L-plastin-deficient mice11 to assess the effects of L-plastin on in vivo hemostasis and thrombosis, determine the intracellular location of MK L-plastin, and assess MK membrane rigidity. We report that MK L-plastin colocalizes with actin in wildtype MK, and loss of MK L-plastin reduces MK membrane rigidity and increases MK PPF, providing a novel mechanism for regulating platelet formation. We first confirmed L-plastin protein expression in platelets from wild-type mice and littermate knockout mice with a disrupted Lcp1 gene lacked platelet L-plastin protein (Online Supplementary Figure S1A). Initial experiments addressed the effect of in vivo L-plastin deficiency on screening assays for hemostasis and thrombosis. Tail bleeding times and survival times after a pulmonary embolism model were shorter in L-plastin expression mice (Figure 1A, B), suggesting a prothrombotic phenotype in the knockout animals. Although this mouse strain is a global knockout, L-plastin expressionis limited to hematopoietic cells, and prior studies have focused on L-plastin effects in leukocytes.12 We, therefore, sought to investigate the role of L-plastin in thrombopoiesis using the L-plastin-deficient mice. Compared to wild-type mice, peripheral blood platelet counts in Lcp1-/- mice were modestly but significantly higher, with no difference in mean platelet volume (Figure 1C, D). Under steady state conditions and compared to wild-type mice, Lcp1-/- mice have normal red blood cell counts and modestly increased white blood cell counts (Online Supplementary Figure S1B, C), but no differences in body weight. Both male and female mice were studied and no effect of sex on the platelet count difference was observed. Note that the comparisons in this study were between inbred strains of Lcp1 wild-type and Lcp1-/- littermates, so we would not expect environmental or other genetic confounding factors. These in vivo data support the use of this Lcp1-/mouse line for studies of L-plastin in the MK/platelet lineage. No difference was observed between the number of acetylcholinesterase-positive MK in wild-type and Lcp1/mice (Online Supplementary Figure S2), suggesting Lplastin may not affect megakaryopoiesis. However, compared to wild-type mice, Lcp1 -/- mice had significantly more PP-forming MK (Figure 2A, B), consistent with an inhibitory role of L-plastin in platelet production. 9 We examined the subcellular distribution of L-

Haematologica | 109 January 2024

331

LETTER TO THE EDITOR

Figure 1. In vivo hemostasis and thrombosis models. (A) Tail-bleeding time for wild-type (WT) Lcp1+/+ and Lcp1-/- littermate mice (Lcp1+/+ N=10, Lcp1-/- N=10). (B) Pulmonary thromboembolism was induced by injecting Lcp1+/+ or Lcp1-/- mice with collagen and epinephrine. Time to death was monitored and plotted (Lcp1+/+ N=10, Lcp1-/- N=9). (C) Platelet count and (D) mean platelet volume (MPV) in the peripheral blood from littermate Lcp1+/+ and Lcp1-/- mice of both sexes. Each dot/circle is an individual mouse (N=19-20).

plastin in MK using immunofluorescent confocal microscopy. Figure 2C demonstrates co-localization between L-plastin and F-actin in MK adherent to both poly-Llysine and fibrinogen. A prominent distribution of Lplastin was observed below the plasma membrane of MK. Similar results were obtained with human CD34+derived MK (Online Supplementary Figure S3). Taken together, these data suggest a hypothesis whereby L-plastin-mediated F-actin bundling results in a denser MK actin cortex that impedes the IMS protrusion necessary to initiate PPF. However, we cannot exclude an increase in lifespan as an additional contributor to the increased platelet count in the Lcp1-/- mice. We have previously shown that L-plastin expression decreases with MK maturation. 9 A corollary of the above hypothesis is that lower levels of L-plastin in mature MK cause reduced actin bundling and hence, a less stiff MK actin cytoskeleton. Because integral membrane proteins directly or indirectly attach the plasma membrane to the outermost layer of actin filaments of the cortical cytoskeleton,13 we tested this corollary by measuring MK cytoskeleton stiffness using the Biomembrane Force Probe (BFP) tether assay (illustrated in Figure 3A, B).14,15 Compared to wild-type MK, L-plas-

tin-deficient MK were significantly less rigid (Figure 3C). Lastly, we used CRISPR to knockdown expression in cultured human MK and observed a similar relationship between L-plastin and membrane rigidity (Figure 3D). The molecular switches that tell the mature MK to begin PPF are incompletely understood. Actin, myosin heavy and light chains, tubulin, kinases, and phosphatases are all critical for this process. Bury et al. have reviewed inherited causes of thrombocytopenia, and listed 11 genes that regulate PPF.16 Based on the rationale that changes in gene expression during MK differentiation might provide clues to genes regulating normal PPF, we previously performed RNA sequencing on biologic replicates of day 6, day 9 and day 13 cord blood-differentiated Mk,9 and found numerous candidate genes, most of which increased as MK began to form PP. Like ROCK1,17 L-plastin is one of the few inhibitors of PPF described to date. In summary, we measured MK cytoskeleton stiffness for the first time and found an important role of the actin cytoskeleton stiffness in PPF. We showed that i) mice deficient in the actin-bundling protein L-plastin have an increased MK PPF and higher platelet counts,

Haematologica | 109 January 2024

332

LETTER TO THE EDITOR

Figure 2. L-plastin inhibits megakaryocyte proplatelet formation and localizes with F-actin. (A) Representative images of megakaryocyte (MK)-forming proplatelets (PP) from both Lcp1+/+ and Lcp1-/- mice. Mouse bone marrow cells were stimulated with 50 ng/mL thrombopoietin for 3-4 days. MK were enriched using bovine serum albumin gradient and cultured for another 16-24 hours and examined by light transmission (LT) microscopy. (B) Quantification of PP-forming MK. PP-forming MK were defined by at least 1 filamentous pseudopod observed by light microscopy with a 20x objective of day 3-4 cultured cells. Scoring was blinded as to Lcp1 genotype, and the percent of total MK number plotted. Three or more images from each of 3 pairs of littermates Lcp1+/+ and Lcp1-/- mice were counted and the average percentage per mouse was plotted. (C) Representative immunofluorescent confocal images of C57BL/6 MK stained for DNA (DAPI), F-actin (magenta) and L-plastin (green). F-actin/L-plastin co-localization is white. MK seeded on poly-L-lysine (PLL) or fibrinogen (FGN). PPF: proplatelet formation.

and ii) mouse and human MK L-plastin levels strongly correlate with MK membrane rigidity. Since higher MK membrane rigidity likely correlates with actin cytoskeletal density, higher levels of L-plastin in immature MK could prevent premature release of platelets in the

bone marrow. At the mature stage of MK differentiation, L-plastin levels have decreased, which is expected to reduce actin bundling and actin cytoskeletal rigidity, and facilitate initiation of membrane PP protrusions (Figure 3E). Bisaria et al. recently showed that newly

Haematologica | 109 January 2024

333

LETTER TO THE EDITOR

Continued on following page. Haematologica | 109 January 2024

334

LETTER TO THE EDITOR Figure 3. L-plastin maintains megakaryocyte membrane stiffness. (A) Illustration of the Biomembrane Force Probe (BFP) tether assay for the measurement of cell stiffness. Borosilicate beads labeled with mouse anti-CD41 antibody were attached to the apex of red blood cell (RBC) at the tip of a micropipette (left). A single megakaryocyte (MK) held by a micropipette (right) approaches the antibody-coated beads and is allowed to attach. Once attached, the MK pipette is pulled back under a constant speed. The cell stiffness is calculated using the pulling force required (= the elastic deformation phase). (B) Representative BFP tether assay readout plot. During the initial approaching stage, no force was detected. During the contact phase, the MK squeezes the bead and shows negative force. During the retraction phase, force is rapidly generated in the linear elastic deformation phase. Cell stiffness was calculated from the elastic deformation phase using the customized package run by LabVIEW (National Instruments). The experiment terminates when the MK detaches from the anti-CD41 coated bead. (C) The BFP tether assay MK membrane stiffness in picoNewtons (pN) per nanometer (nm) for 3 pairs of mice (wild-type [WT] 1-3 and Lcp1-/- 1-3). Each circle represents a single MK. (D) Similar to panel (C), the BFP tether assay was used to assess human MK membrane rigidity in CD34+-derived MK treated without (control 1-3) and with (knockout [KD] 1-3) CRISPR-mediated knockdown of L-plastin. N=3 pairs of different cord blood donor cells.

forming lamellipodial protrusions form in areas of low membrane proximal F-actin density,18 and it will be of interest for future studies to assess the effect of Lplastin-mediated actin bundling on the ability of cofilin to depolymerize F-actin.

Authors

No conflicts of interest to disclose. Contributions

Li Guo,1-4 Shancy Jacob,1 Bhanu Kanth Manne,1 E. Motunrayo

LG, SJ, BKM, EKM, SG, XW, DM, EAT, IP, YK, CB and SB performed

Kolawole,5 Siqi Guo,1,6 Xiang Wang,7 Darian Murray,1 Emilia A.

research. LG, SB and PFB wrote the manuscript. MTR, BE and CM provided scientific expertise.

Tugolukova, Irina Portier, Yasuhiro Kosaka, Cindy Barba, Matthew T. 1

Rondina,

1,5,8,9

Brian Evavold, Celeste Morley, 5

10,11

Seema Bhatlekar and 1#

Acknowledgments

Paul F. Bray

1,2#

The authors wish to acknowledge experiment technical support from 1

Program in Molecular Medicine, Department of Internal Medicine,

Jesse W. Rowley, Frederik Denorme, Mark J. Cody, Neal D. Tolley,

University of Utah, Salt Lake City, UT; 2Division of Hematology and

Marina Tristao and Meenakshi Banerjee, editorial help from Neal D.

Hematologic Malignancies, Department of Internal Medicine,

Tolley, and figure preparation expertise from Diana Lim and Nikita

University of Utah, Salt Lake City, UT; 3Bloodworks Research Institute,

Abraham, Molecular Medicine Program, University of Utah, as well as

Seattle, WA; Hematology Division, University of Washington, Seattle,

the technical support from Xue Lin, Washington University in St.

WA; Department of Pathology, University of Utah, Salt Lake City, UT;

Louis.

Applied Mathematics, Department of Mathematics, University of

Utah, Salt Lake City, UT; 7HSC Cell Imaging Core, School of Medicine,

Funding

University of Utah, Salt Lake City, UT; Department of Internal

This study was supported by grants from the National Institutes of

Medicine, University of Utah, Salt Lake City, UT; 9George E. Wahlen

Health (R01HL141424, K24HL155856, R01HL142804, R01AG048022,

Department of Veterans Affairs Medical Center, Department of

R56AG059877, and R01HL130541, R01AI169835, R01AI104732,) and the

Internal Medicine, and Geriatric Research, Education, and Clinical

Division of Hematology and Hematologic Malignancies at the University

Center (GRECC), Salt Lake City, UT; Pediatrics, Infectious Diseases,

of Utah. This work was also supported by Merit Review Award Number

Washington University, St. Louis, MO and Pathology and

I01 CX001696 from the US Department of Veterans Affairs Clinical

Immunology, Washington University, St. Louis, MO, USA

Sciences R&D (CSRD). In addition, this work was supported by

American Heart Association Career Development Award (No. 938717 to SB and PFB contributed equally as senior authors.

LG) and American Heart Association Postdoctoral Fellowship (2022Post906231 to IP). This material is the result of work supported

Correspondence:

with resources at the George E. Wahlen VA Medical Center, Salt Lake

P. F. BRAY - Paul.bray@hsc.utah.edu

City, Utah. The contents do not represent the views of the US

S. BHATLEKAR - saseems@gmail.com

Department of Veterans Affairs or the United States Government. The authors thank the University of Utah Flow Cytometry Facility in addition

https://doi.org/10.3324/haematol.2023.283016

to the National Cancer Institute through Award No. 5P30CA042014-24.

Received: February 24, 2023.

Data-sharing statement

Accepted: July 6, 2023.

The data and experimental protocol are available upon request.

Haematologica | 109 January 2024

335

LETTER TO THE EDITOR

References 1. Chen SJ, Sugimoto N, Eto K. Ex vivo manufacturing of platelets: beyond the first-in-human clinical trial using autologous iPSCplatelets. Int J Hematol. 2023;117(3):349-355. 2. Machlus KR, Italiano JE Jr. Megakaryocyte development and platelet formation. In: Michelson AD, editor. Platelets. 4th ed. London, UK. Academic Press; 2019. Chapter 2; p. 25-46. 3. Woolthuis CM, Park CY. Hematopoietic stem/progenitor cell commitment to the megakaryocyte lineage. Blood. 2016;127(10):1242-1248. 4. Eckly A, Heijnen H, Pertuy F, et al. Biogenesis of the demarcation membrane system (DMS) in megakaryocytes. Blood. 2014;123(6):921-930. 5. Antkowiak A, Viaud J, Severin S, et al. Cdc42-dependent F-actin dynamics drive structuration of the demarcation membrane system in megakaryocytes. J Thromb Haemost. 2016;14(6):1268-1284. 6. Hamada T, Mohle R, Hesselgesser J, et al. Transendothelial migration of megakaryocytes in response to stromal cellderived factor 1 (SDF-1) enhances platelet formation. J Exp Med. 1998;188(3):539-548. 7. Eckly A, Scandola C, Oprescu A, et al. Megakaryocytes use in vivo podosome-like structures working collectively to penetrate the endothelial barrier of bone marrow sinusoids. J Thromb Haemost. 2020;18(11):2987-3001. 8. Chalut KJ, Paluch EK. The actin cortex: a bridge between cell shape and function. Dev Cell. 2016;38(6):571-573. 9. Bhatlekar S, Manne BK, Basak I, et al. miR-125a-5p regulates megakaryocyte proplatelet formation via the actin-bundling protein L-plastin. Blood. 2020;136(15):1760-1772.

10. Schachtner H, Calaminus SD, Sinclair A, et al. Megakaryocytes assemble podosomes that degrade matrix and protrude through basement membrane. Blood. 2013;121(13):2542-2552. 11. Chen H, Mocsai A, Zhang H, et al. Role for plastin in host defense distinguishes integrin signaling from cell adhesion and spreading. Immunity. 2003;19(1):95-104. 12. Morley SC. The actin-bundling protein L-plastin supports T-cell motility and activation. Immunol Rev. 2013;256(1):48-62. 13. Fritzsche M, Thorogate R, Charras G. Quantitative analysis of ezrin turnover dynamics in the actin cortex. Biophys J. 2014;106(2):343-353. 14. Evans E, Heinrich V, Leung A, Kinoshita K. Nano- to microscale dynamics of P-selectin detachment from leukocyte interfaces. I. Membrane separation from the cytoskeleton. Biophys J. 2005;88(3):2288-2298. 15. Heinrich V, Leung A, Evans E. Nano- to microscale dynamics of P-selectin detachment from leukocyte interfaces. II. Tether flow terminated by P-selectin dissociation from PSGL-1. Biophys J. 2005;88(3):2299-2308. 16. Bury L, Falcinelli E, Gresele P. Learning the ropes of platelet count regulation: inherited thrombocytopenias. J Clin Med. 2021;10(3):533. 17. Chang Y, Auradé F, Larbret F, et al Proplatelet formation is regulated by the Rho/ROCK pathway. Blood. 2007;109(10):42294236. 18. Bisaria A, Hayer A, Garbett D, Cohen D, Meyer T. Membraneproximal F-actin restricts local membrane protrusions and directs cell migration. Science. 2020;368(6496):1205-1210.

Haematologica | 109 January 2024

336

LETTER TO THE EDITOR

An accessible patient-derived xenograft model of low-risk myelodysplastic syndromes Myelodysplastic syndromes (MDS) are among the most common myeloid malignancies.1 They encompass a spectrum of clonal bone marrow (BM) failure diseases characterized by ineffective hematopoiesis and progression to acute myeloid leukemia (AML).2 The clinical course of MDS varies widely from indolent, requiring only monitoring/supportive care, to aggressive disease with AML-like biology.1 Recurrent somatic mutations contribute to the pathophysiology, being recently incorporated into prognostic scoring systems.2,3 Akin to targeted therapy in AML, somatic mutations might be therapeutically targeted in MDS. Unfortunately, testing this hypothesis has been challenging due to the lack of preclinical models that truly recapitulate disease biology, particularly in lower-risk disease. This limitation has resulted in a dearth of novel targeted therapeutics and slow drug development in the field.4 Genetically engineered mouse models of recurrent somatic mutations present in MDS hold promise for advances in understanding the biology of these diseases. However, given that introns are generally not conserved between mice and humans, there is potential for significant differences in intron retention and cryptic splicing between human and mouse hematopoietic cells bearing spliceosome mutations.4 The lack of readily available patient-derived xenograft (PDX) models remains the major barrier to drug development. NSG/NSG-S mice lack functional B, T, and natural killer cells as well as immunoglobulins,5 and are unreliable for establishing PDX of low-risk MDS.4 Our previous experience with conditioning regimens based exclusively on radiotherapy did not result in engraftment of MDS cells (data not shown). Recently, PDX models of MDS based on intra-bone co-injection of human mesenchymal stromal cells6 or subcutaneous generation of ossicles using human mesenchymal stromal cells7 have mitigated this shortcoming. Moreover, intra-hepatic injection of OKT3-depleted MDS BM into double-irradiated newborn MISTRG mice resulted in robust and reliable engraftment.8 However, these models are not widely available and require technical skills that reduce their utility. It would, therefore, be opportune to develop a PDX model that is more suitable for testing novel targeted therapies and that allows for high throughput drug screening in low-risk diseases. MISTRG mice are similar to other xenograft models in the sense that they are immunocompromised and express human cytokines (knocked-in in the case of MISTRG mice). Nevertheless, one key feature of MISRTG mice is their expression of human SIRPα. Their ability to recognize human CD47 (the “don’t eat me” signal) may explain the

more reliable engraftment of human cells in these mice.9 One way to reproduce this feature in NSG/NSG-S mice is by depleting macrophages with clodronate. Intraperitoneal administration of 100 μL of clodronate liposomes (FormuMax) to 6to 8-week-old mice (NOD.Cgscid tm1Wjl Prkdc Il2rg /NSG-Tg(CMV-IL3,CSF2,KITLG)1Eav/MloySz J-NSG-S, prior to Jackson Laboratories; Jackson Laboratories), led to an 83% reduction of macrophages in the BM, and >99% reduction in the spleen after 48 h (Online Supplementary Figure S1A, B). The depletion persisted 7 days later, with 97% reduction in the BM, and still >99% in the spleen (Online Supplementary Figure S1A, B). In contrast, administration of 2 Gy of radiotherapy (classically used for conditioning) did not affect the presence of macrophages in the BM or spleen (Online Supplementary Figure S1A, B). To test whether depletion of macrophages enabled engraftment of patient-derived MDS cells, we conditioned 6- to 8-week-old NSG/NSG-S mice with 100 μL of clodronate liposomes, on day -2 and irradiated them with 2 Gy on day 0. Six hours later, we injected CD34+ cells, via the tail vein; the cells were selected, as previously published,7 from the BM of a patient with low-risk MDS (Figure 1A). At the time of sample collection, the patient had 12% CD34+ blasts (Table 1). The BM samples came from MDS patients who had consented to participate in this study in accordance with the Declaration of Helsinki. The research protocol was approved by the Johns Hopkins Institutional Review Board. Flow-cytometry analysis of peripheral blood collected from transplanted mice 1, 2, and 3 months after transplantation showed the presence of human leukocytes (mouse CD45–[mCD45–] human CD45+ [hCD45+]) (Figure 1B). At 3 months after transplantation, NSG mice showed reliable engraftment in the BM (Online Supplementary Figure S1C). NSG-S mice demonstrated consistent engraftment in the peripheral blood, spleen, and BM (Online Supplementary Figure S1C, Table 1). Furthermore, NSG-S mice showed detectable CD34+ leukocytes (mCD45–hCD45+hCD34+) as well as human erythroid cells (mCD45–hCD45–glycophorin A+ [GPA+]) in the BM at 3 months after transplantation. Using hCD45 beads and a magnetic column (Miltenyi), we selected hCD45+ cells from the BM of these mice. hCD45+ cells were pooled together and DNA was extracted using the Quick-DNA Miniprep Plus kit (Zymo Research). Libraries were prepared using a xGen Prism DNA Library Prep Kit (IDT) and IDT hybridization capture-based targeted duplex sequencing was done on a NOVASeq6000 platform in an SP flow cell. Data analyzed using DRAGEN Enrichment (v4.0.3) showed

Haematologica | 109 January 2024

337

LETTER TO THE EDITOR

Figure 1. Clodronate and radiotherapy conditioning allow for reconstitution of human low-risk myelodysplastic syndrome hematopoiesis in NSG-S mice. (A) Graphic representation of the mouse engraftment technique and assessment. (B) Representative flow cytometry analysis of the human graft in an NSG-S mouse. Data were analyzed using BD FACSDiva software v8.0.1. i.p.: intraperitoneal; XRT: radiotherapy; mo: months; SSC: side scatter. Haematologica | 109 January 2024

338

Haematologica | 109 January 2024

339

p.K634N

EZH2

p.V253

ZRSR2

7.9± 3.22± 2.28± 3.09 1.13 0.86

100x103/ mouse (5)

Patient 3 IPSS-R Aza

8.8

12.4

9.4

–

hCD45 mCD45 mean±SD

–

299

338

1.36

2.45

ANC, x109/L

% hCD45 mean±SD

Spleen – final

3.5 points/intermediate risk

1 point/very low risk

3 points/low risk

IPSS-R score/category

hCD45–mCD45– mean±SD

0.45/moderate high risk

-2.3/very low risk

0.23/moderate high risk

IPSS-M score/category

23.1± 4.35

12.7± 3.38 0

NA 0.11± 0.06

2.1± 0.6

9.16± 96.83 0.38± 0.57± 9.99 ± 2.5 0.37 0.17

0.1

62.6± 4.45± 19.6 4.05

74.6± 26.5± 2.93 4.2

1.28± 0.77

0.1± 0

17.6± 0.45± 0.5± 0.1 6.92 0.15

2.46± 0.63± 16.4± 0.47± 0.6± 3.27 0.62 12.93 0.05 0.1

35.65± NA 27.22 0.02± 90.7± 0.03± 0.05± 0.03± 0.03± 12.86± 0.01 3.7 0.01 0.03 0.04 0.04 1.65

48.7± 0.1± 0.5 0

38.2± 26.5± 4.02 4.2

88.03 0.01± 0.01± ± 4.7 0 0.005

0.25± 0.05

0.17± 0.09

0.02± 0.01

5 mo CD20+/ CD20+/ CD71+ CD71+ + + + + + + + + + + - D44 Total CD41+ Total CD41 CD3 CD33 CD3 CD33 CD71 GPA CD41 CD71 GPA CD41+ CD19+ CD19+ GPA+ GPA+ Aza1

8.25± 11.45± 55.4± 0.1± 12.51 19.79 0 0.95

No 2.56± Aza 0.87

4.5 mo - D17 Aza1

<2%

BM Hemoglo- Platelets, blasts, % bin, g/dL x109/L

Bone marrow – final % hCD45 mean±SD

46,XY

46, XY

47,XY+8[4]/ 46, XY[17]

Karyotype

NGS: next-generation sequencing; AA: amino acids; VAF: variant allele frequency; BM: bone marrow; ANC: absolute neutrophil count; IPSS-R: Revised International Prognostic Scoring System; IPSS-M: Molecular International Prognostic Scoring System; SD: standard deviation; h: human; m: mouse; mo: months; NA: not available; Aza: azacitidine.

9.9± 2.68

6.5± 0.64

14± 3.25

75x103/ mouse (3)

Patient 2 Fresh

11.3

7.6

10.4± 3.4

2x106/ mouse (3)

1 mo 2 mo 3 mo

% hCD45 mean±SD

Peripheral blood

48.2

p.Q1030fs

TET2 89.7

20.4

p.C28fs

PHF6

Sample N of cells type (N of mice)

27.7

p.G646fs

ASXL1 51.2

26.7

p.F683fs

36.1

p.K700E

SF3B1

TET2

48.9

p.T620M

PLCG2 40.2

p.D1214N

NUP98

p.T28fs

41.7

p.A91G

NRAS

KDM2B

46.2

p.F1005fs

ASXL1 p.R955H

5.8

p.G646fs

ASXL1

NOTCH1

VAF 41.1

AA change p.S34F

Gene U2AF1

NGS data for patients’ clinical samples

Patient 1 Fresh

Sample

Patient 3

Patient 1

Patient 2

Age in years

Sample

Table 1. Engraftment of primary low-risk myelodysplastic syndrome samples into NSG-S mice, after macrophage depletion, is reliable regardless of the cytogenetics or type of sample used. The table shows the patients’ characteristics for the samples used for engraftment (top) and human cell populations and percentages in the peripheral blood (at different timepoints), bone marrow and spleen of NSG-S mice (bottom).

LETTER TO THE EDITOR

B C

E D

Figure 2. Engrafted low-risk myelodysplastic syndrome cells maintain the cytogenetics (and their functional impact) of the original patients’ samples. (A) Comparison between the mutational profiles identified in the clinical report of patient 1 versus those in an analysis of CD34+ cells sorted before transplantation and pooled human CD45+ (hCD45+) cells sorted from mouse bone marrow (BM) 3 months after transplantation. (B) Comparison between the mutational profiles in the CD34+ cells sorted from patient 2 before transplantation and the pooled hCD45+ cells isolated from the BM of mice engrafted with cells from patient 2, at 3 months after transplantation. (C) Splicing assay showing abnormal intron retention in the hCD45+ cells of patient 2 prior to engraftment (MDS preENG, column 2) and those isolated from the BM of mice engrafted with cells from this patient at 3 months after transplantation (MDS postENG, column 3) as predicted in SF3B1-mutant cells. (D) Comparison between the mutational profile of CD34+ cells sorted from patient 3 before transplantation and the pooled hCD45+ cells sorted from the BM of mice engrafted with cells from patient 3 at 6 months after transplantation. (E) Ratio of cryptic (abnormal intron retention) to canonical LTZR1 splicing in the pooled hCD45+ cells isolated from the BM of mice engrafted with cells from patient 3 as predicted in ZRSR1-mutant cells. Each dot represents an independent experiment, horizontal lines represent the mean of four experiments. VAF: variant allele frequency; NBM: normal bone marrow; MDS: myelodysplastic syndrome; Aza: azacitidine. Haematologica | 109 January 2024

340

LETTER TO THE EDITOR that the selected hCD45+ cells had a similar clonal architecture to that of the transplanted CD34+ cells and the clinical samples (Figure 2A). Secondary transplantation using 106 hCD45+ cells isolated from the BM of NSG/NSGS mice resulted in detectable human engraftment, albeit at lower levels compared to those in primary recipients (data not shown). We tested the reproducibility of the NSG-S model using various conditions: different CD34+ numbers, fresh versus frozen cells and clinical specimens from low-risk MDS patients with different molecular profiles (Table 1). Human leukocytes were present in the peripheral blood of all mice at 1 (0.52%-14%) and 3 months (3.1%-20.7%) after transplantation (Table 1). Similarly, we observed successful engraftment in the BM (18.1%-99.3%) and spleen (8.1%-31.3%) at 3-6 months after transplantation (engraftment levels for each individual mouse are shown in Online Supplementary Figure S1C). Human cells were mostly CD33+ myeloid cells (38.2%-90.7% of hCD45+ cells). B-cell engraftment, as detected by the presence of CD19+/CD20+ cells (0.11%-2.1% of hCD45+) was present in a subset of animals. No mice showed CD3+ T- cell engraftment. Human erythroid engraftment is particularly difficult to achieve in xenograft models, as erythroid precursor cells are constantly removed by host macrophages. The depletion of macrophages with clodronate resulted in erythropoietic cell engraftment including hCD45–hCD71+ and hCD45–hCD71+GPA+ cells in the BM and spleen of recipient mice (Figure 1B, Table 1). In addition, hCD45+hCD41+ (0.14%-0.47% of hCD45+) and hCD45–hCD41+ (0.1%-0.63% of mCD45–hCD45–) megakaryocyte lineage cells were present in the BM and spleen. The clonal architecture of engrafted cells recapitulated that of the clinical BM samples and enriched CD34+ cells (Figure 2A, B, D). In lieu of morphological analysis of engrafted human cells, we investigated whether these cells showed preserved cryptic intron retention, similarly to the original sample. To this end, we extracted mRNA (from the pooled hCD45+ cells sorted from the mouse BM) using the Monarch Total RNA Miniprep Kit and synthesized cDNA with the High-Capacity cDNA Reverse Transcription Kit in order to do splicing assays of the samples with mutated splicing factors (SF3B1 and ZRSR2). For the SF3B1 splicing assay, an isoform-competitive endpoint polymerase chain reaction for canonical cryptic acceptors in exon 2 of TMEM14C was performed with the primers GACACCTCGCAGTCATTCCT and TGATCCCACCAGAAGCAACC. For the ZRSR1 splicing assay, quantitative polymerase chain reaction was performed using the KAPA SYBR FAST qPCR Master Mix (2X) Kit with the primers CCCGCTCCAGCTACTTTGAA and CAAACAAGTAGAGCGAGTCCT for canonical LZTR1 and primers CCCGCTCCAGCTACTTTGAA and AGTTCACTGGGGAGTGAGGAT for LZTR1 intron 18 retention. The expression ratio was calculated with 2^-(canonical cycle threshold–cryptic cycle threshold). mRNA analysis of engrafted MDS cells with

SF3B1 (patient 2) and ZRSR2 (patient 3) mutations validated the persistence of physiologically impactful10,11 mis-splicing events in TMEM14C and LZTR1, respectively (Figure 2C, E). For further exploration of disease relevance, after confirming engraftment, PDX derived from patient 3 were treated with either azacitidine (5 daily injections of 5 mg/kg, 2 cycles, 4 weeks apart) or saline. PDX treated with azacitidine showed decreased levels of BM engraftment at 6 months after transplantation (4 weeks after the completion of the 2nd cycle) (Table 1). Although not definitive (because of the small number of mice), these results serve as proof of concept of the model’s applicability to study therapeutic interventions. Recent years have seen the emergence of new technologies for augmentation of murine modeling of MDS, including genetically engineered mouse models and innovative PDX models. The PDX model described here complements currently available analytic tools in MDS. It makes use of relatively inexpensive and widely available reagents (i.e., NSG-S mice, clodronate, and radiation) and engraftment is tracked via non-invasive methods. Given the simplicity of its workflow, with no need for other moving parts (e.g., timed pregnancies, pre-engraftment of other cell populations or bone fragments), the use of viably frozen CD34+ cells and easy conditioning regimen followed by intravenous injection, this model is suitable for large-scale drug testing in low-grade MDS. Typically, 2030 mL of BM aspirate from a low-grade MDS sample may be sufficient to generate 50-60 PDX mice. Such scaling allows for testing of single-agent regimens (e.g., spliceosome inhibitors) and combination therapies (e.g., with hypomethylating agents) thus informing early-stage clinical development.

Authors Patric Teodorescu,1,2 Sergiu Pasca,1 Inyoung Choi,1 Cynthia Shams,1 W. Brian Dalton,1 Lukasz P. Gondek,1 Amy E. DeZern1 and Gabriel Ghiaur1 Sidney Kimmel Comprehensive Cancer Center, Johns Hopkins

University School of Medicine, Baltimore, MD, USA and 2

Universitatea de Medicina si Farmacie “Iuliu Hatieganu”, Cluj-

Napoca, Romania Correspondence: G. GHIAUR - gghiaur1@jhmi.edu https://doi.org/10.3324/haematol.2023.282967 Received: March 6, 2023. Accepted: June 29, 2023. Early view: July 6, 2023.

Haematologica | 109 January 2024

341

analyzed the sequencing experiments. AED contributed patients’ samples. GG designed the project. All authors critically reviewed the

Published under a CC BY-NC license

manuscript.

Disclosures

Funding

CG has received research support from Menarini Ricerche, Abbie,

The work was funded by R01 HL159306, and P01 CA225618, P30

Inc. and serves on an advisory board for Syros, Inc. WBD has

CA006973ASH, Bridge Award (to GG), a Johns Hopkins Hematological

received research support from Abbvie, Inc.

Malignancy Bone Marrow Transplant Pilot Grant (to AED and GG), and NCI R01 CA253981 (to AED and GG).

Contributions PT wrote the manuscript, performed research, and analyzed data.

Data-sharing statement

SP analyzed sequencing data. IC designed, performed and analyzed

The essential data supporting our findings are present within the ar-

the intron retention experiments. SC performed library preparation

ticle and Online Supplementary Material. The corresponding author

for sequencing experiments. WBD designed, performed and

can, upon reasonable request, provide raw data and additional details.

analyzed the splicing experiments. LPG designed, performed and

All shared data will be anonymized for privacy.

References 1. Zeidan AM, Shallis RM, Wang R, Davidoff A, Ma X. Epidemiology of myelodysplastic syndromes: why characterizing the beast is a prerequisite to taming it. Blood Rev. 2019;34:1-15. 2. Sperling AS, Gibson CJ, Ebert BL. The genetics of myelodysplastic syndrome: from clonal haematopoiesis to secondary leukaemia. Nat Rev Cancer. 2017;17(1):5-19. 3. Bernard E, Tuechler H, Greenberg PL, et al. Molecular International Prognostic Scoring System for myelodysplastic syndromes. NEJM Evidence. 2022;1(7). 4. Liu W, Teodorescu P, Halene S, Ghiaur G. The coming of age of preclinical models of MDS. Front Oncol. 2022;12:815037. 5. Shultz LD, Lyons BL, Burzenski LM, et al. Human lymphoid and myeloid cell development in NOD/LtSz-scid IL2R gamma null mice engrafted with mobilized human hemopoietic stem cells. J Immunol. 2005;174(10):6477-6489. 6. Medyouf H, Mossner M, Jann JC, et al. Myelodysplastic cells in patients reprogram mesenchymal stromal cells to establish a

transplantable stem cell niche disease unit. Cell Stem Cell. 2014;14(6):824-837. 7. Altrock E, Sens-Albert C, Jann JC, et al. Humanized threedimensional scaffold xenotransplantation models for myelodysplastic syndromes. Exp Hematol. 2022;107:38-50. 8. Song Y, Rongvaux A, Taylor A, et al. A highly efficient and faithful MDS patient-derived xenotransplantation model for pre-clinical studies. Nat Commun. 2019;10(1):366. 9. Rongvaux A, Willinger T, Martinek J, et al. Development and function of human innate immune cells in a humanized mouse model. Nat Biotechnol. 2014;32(4):364-372. 10. Clough CA, Pangallo J, Sarchi M, et al. Coordinated missplicing of TMEM14C and ABCB7 causes ring sideroblast formation in SF3B1mutant myelodysplastic syndrome. Blood. 2022;139(13):2038-2049. 11. Inoue D, Polaski JT, Taylor J, et al. Minor intron retention drives clonal hematopoietic disorders and diverse cancer predisposition. Nat Genet. 2021;53(5):707-718.

Haematologica | 109 January 2024

342

LETTER TO THE EDITOR

Final results and overall survival data from a phase II study of acalabrutinib monotherapy in patients with relapsed/refractory mantle cell lymphoma, including those with poor prognostic factors Mantle cell lymphoma (MCL) is an aggressive, rare, B-cell non-Hodgkin lymphoma often characterized by recurrent relapses after initial therapy.1 Blastoid and pleomorphic variants, representing <20% of MCL cases, are considered highrisk and are associated with a high proliferation index and poor prognoses.2 Patients with these variants rarely achieve durable remission or prolonged clinical outcomes with standard chemotherapies.2 High-risk Mantle Cell Lymphoma International Prognostic Index (MIPI) score and Ki-67 index >30% are also poor prognostic factors in MCL.1 The 5-year overall survival (OS) rate for patients with high-risk MIPI score has been reported to be 35% compared with 81% and 63% for those with low- and intermediate-risk scores, respectively.3 Ibrutinib, a Bruton tyrosine kinase (BTK) inhibitor, demonstrated favorable responses in patients with relapsed/refractory (R/R) MCL, but has been associated with less activity in patients with blastoid morphology.4,5 Acalabrutinib is a next-generation, potent, highly selective BTK inhibitor approved for the treatment of patients with R/R MCL who have received ≥1 prior therapy. Previous updates on this study (ACE-LY-004; clinicaltrials gov. Identifier: NCT02213926) demonstrated that the safety and efficacy of acalabrutinib were maintained over time6-8 and using a data cutoff of February 24, 2020, median OS had not yet been reached.8 Here, we report the final efficacy results using a data cutoff of December 4, 2020, which includes updated OS data and long-term safety findings in patients with R/R MCL, including patients with blastoid/pleomorphic morphology, high-risk MIPI score, and Ki-67 index >50%. Notably, the median follow-up of 38.1 months was the same at the February 24, 2020, and December 4, 2020, data cutoff dates because the patients who remained in the study had time on study greater than the median; the median follow-up of patients (n=65) who were alive at the time of analysis was 54.7 months. ACE-LY-004 was an open-label, multicenter, single-arm phase II study of acalabrutinib in patients with R/R MCL. Acalabrutinib 100 mg was administered orally twice daily until progressive disease or unacceptable toxicity. Patients ≥18 years with confirmed MCL who relapsed after, or were refractory to, one to five previous therapies were included. The primary endpoint was investigator-assessed overall response rate (ORR), defined as the proportion of patients who achieved partial response (PR) or complete response

(CR) according to the Lugano response criteria for nonHodgkin lymphoma.9 Secondary endpoints included investigator-assessed duration of response (DOR), progression-free survival (PFS), OS, and safety. More details on the study design have been previously published.6 A total of 124 patients were enrolled; the median age was 68 years. Many patients had high disease burden, including 37.1% of patients with bulky lymph nodes of ≥5 cm and 71.8% of patients with extranodal involvement. Other key risk factors indicative of poor prognosis were blastoid/pleomorphic morphology in 26 (21.0%) patients, high-risk MIPI score in 21 (16.9%) patients, and Ki-67 index ≥50% in 32 (25.5%) patients. About a quarter (26.6%) of patients had a history of cardiac disorders, including atrial fibrillation (6.5%). Almost half (45.2%) had a history of hypertension. Other patient characteristics were previously reported.7 The median number of prior therapies was two (range, 1–5); 59 (47.6%) patients had one prior therapy and 65 (52.4%) patients had ≥2 prior therapies. Fifty-four (43.5%) patients received acalabrutinib for >24 months, including 14 (11.3%) patients who received treatment for >60 months. Eighteen (14.5%) patients remained on treatment; the median dose intensity for all patients was 98.6% (Online Supplementary Table S1). In the total population, ORR and CR rates were 81.5% (95% confidence interval [CI]: 73.5-87.9) and 47.6% (95% CI: 38.5-56.7), respectively. With a median follow-up of 38.1 months, median DOR (mDOR) and median PFS (mPFS) were 28.6 months (95% CI: 17.5-39.1) and 22.0 months (95% CI: 16.6-33.3), respectively (Figure 1A). The estimated median OS (mOS) was 59.2 months (95% CI: 36.5-not evaluable [NE]; Figure 1C; Table 1) and the estimated 5year OS rate was 49.5% (95% CI: 40.1-58.2). Efficacy outcomes by subgroups of patients with blastoid/pleomorphic morphology, high-risk MIPI score, Ki-67 index, and response were also reported (Table 1; Online Supplementary Figure S1). In the subgroup of 26 patients with blastoid/pleomorphic MCL, mPFS and mOS were 15.2 months (95% CI: 3.7-30.4; Figure 1B) and 36.3 months (95% CI: 12.4-NE; Figure 1D), respectively, with an ORR of 80.8% (95% CI: 60.6-93.4), similar to the overall population. The adverse event (AE) profile remained consistent with the known acalabrutinib safety profile (Online Supplemen-

Haematologica | 109 January 2024

343

LETTER TO THE EDITOR tary Table S2). Median treatment exposure was 17.5 months (range, 0.1–65.3). Grade ≥3 AE were reported in 82 (66.1%) patients. Serious AE were reported in 62 (50.0%) patients, with the most common being pneumonia (8

[6.5%]). Serious grade ≥3 AE were observed in 61 (49.2%) patients. The highest incidence of grade ≥3 AE and anygrade serious AE occurred in the first year of acalabrutinib therapy (Figure 2A, B).

Continued on following page. Haematologica | 109 January 2024

344

LETTER TO THE EDITOR

Figure 1. Progression-free survival and overall survival in the total population (A, C) and in a subgroup of patients with blastoid and pleomorphic disease (B, D). A total of 59 out of 124 patients died (47.6%); causes include disease progression (N=40), adverse event (N=6 [1 patient experienced treatment-related myelodysplastic syndrome and died from pneumonia]), other cause (N=6), or unknown cause (N=7). Deaths due to “other cause” include secondary acute myeloid leukemia (N=2), lung cancer (N=1), pneumonia (N=1), intestinal obstruction (N=1), and graft-versus-host disease (N=1). Deaths due to “unknown cause” include a patient who died at home (N=1) and who had been qualified to next line of therapy (N=1). PFS: progression-free survival; CI: confidence interval; NE: not estimable; OS: overall survival.

Selected AE of clinical interest were atrial fibrillation (any grade, n=3 [2.4%]; no grade 3/4), hypertension (any grade, n=5 [4.0%]; grade 3/4, n=2 [1.6%]), major hemorrhage (n=5 [4.0%], all grade 3/4), and infections (any grade, n=84 [67.7%]; grade 3/4, n=21 [16.9%]). One of the five patients reporting major hemorrhage had previously experienced treatment-related atrial fibrillation (grade 2, 1030 days after initiating acalabrutinib), for which anticoagulation was initiated; 16 days later the patient suffered a treatment-related subdural hematoma (grade 4, day 1,046), prompting discontinuation of acalabrutinib. Of the three patients experiencing atrial fibrillation, two had a history of hypertension; one received aspirin for cardiac prophylaxis, and the other patient is described above (received anticoagulants). The cumulative incidence of atrial fibrillation (any grade) and major hemorrhage (any grade) remained low over time (Figure 2C). Major hemorrhage (grade 2 hematuria) and infection (grade 2 sinusitis) led to dose reduction in one patient each. One additional patient had an acalabrutinib dose reduction due to an AE of fatigue. Patients discontinued acalabrutinib mainly due to disease progression (62.1%) or AE (12.1%; Online Supplementary Table S1). Fifteen patients discontinued due to AE; in nine of these patients, AE were deemed related to acalabrutinib by the treating physician (diffuse large B-cell lymphoma [DLBCL], thrombocytopenia, myelodysplastic syndrome [MDS], leukocytosis, melanoma, cardiac arrythmia by atrial fibrillation and subdural hematoma, pulmonary fibrosis, hemorrhagic bullae and petechiae, and skin rash). All three patients with reported second primary malignancies related to acala-

brutinib had received chemotherapy prior to study enrollment (R-CHOP [rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone] plus lenalidomide in the patient with DLBCL, hyper-CVAD [cyclophosphamide, vincristine, doxorubicin, and dexamethasone] therapy with alternating methotrexate/cytarabine in the patient with MDS, and bendamustine plus rituximab in the patient with melanoma). Death was reported in 59 patients; 40 (32.3%), six (4.8%), and 13 (10.5%) patients died due to disease progression, AE, and unknown or other causes, respectively. Five deaths due to AE were not considered to be related to acalabrutinib (MDS, aortic stenosis, pulmonary embolism, non-small cell lung cancer, and suicide), one of which occurred during the post-treatment follow-up period (MDS). The only treatment-related death in this study was in a patient who died of pneumonia. This patient discontinued acalabrutinib (day 937) due to treatment-related grade 4 MDS (mentioned above) and died 185 days (day 1,112) after the last dose of acalabrutinib. There were 71 (57.3%) patients who received subsequent anticancer therapy post acalabrutinib treatment. Bendamustine with rituximab was used in 16 (12.9%) patients, radiotherapy was used in nine (7.3%) patients, and ibrutinib was used in eight (6.5%) patients. The final results of this study demonstrated that efficacy and safety of acalabrutinib were maintained compared with data from a previous analysis (median follow-up of 26 months).7 Acalabrutinib also demonstrated efficacy in patients with blastoid/pleomorphic morphology, high-risk MIPI score, and Ki-67 index >30% and >50%, which are poor prognostic factors in MCL.

Haematologica | 109 January 2024

345

LETTER TO THE EDITOR Table 1. Progression-free survival, overall survival, response, and duration of response in the total population. All patients N=124

Patients with blastoid/ pleomorphic MCL N=26

Patients with high-risk MIPI N=21

Patients with Ki-67 Index of >50% N=26

PFS in months, median (95% CI)a

22.0 (16.6-33.3)

15.2 (3.7-30.4)

5.7 (2.2-13.7)

5.8 (3.5-33.3)

OS in months, median (95% CI)b

59.2 (36.5-NE)

36.3 (12.4-NE)

17.6 (6.6-24.4)

22.3 (12.1-NE)

101 (81.5) 59 (47.6) 42 (33.9) 10 (8.1) 10 (8.1) 3 (2.4)

21 (80.8) 10 (38.5) 11 (42.3) 4 (15.4) 1 (3.8) 0

12 (57.1) 5 (23.8) 7 (33.3) 2 (9.5) 6 (28.6) 1 (4.8)

16 (61.5) 13 (50.0) 3 (11.5) 6 (23.1) 3 (11.5) 1 (3.8)

(N=101) 28.6 (17.5-39.1)

(N=21) 13.5 (1.9-28.6)

(N=12) 9.7 (3.6-25.7)

(N=16) 28.6 (3.6-NE)

ORR, N (%) CR PR SD PD NEc DOR in months, median (95% CI)d Overall survival by subgroup

Ki-67 index ≤30% N=44

Ki-67 index >30% N=52

Median OS in months (95% CI)e

NE (36.5-NE)

27.9 (16.6, NE)

60-month OS rate, % (95% CI)

57.9 (41.0-71.5)

43.0 (29.0-56.2)

Ki-67 index ≤50% N=70

Ki-67 index >50% N=26

Median OS in months (95% CI)e

NE (36.5-NE)

22.3 (12.1-NE)

60-month OS rate, % (95% CI)

56.0 (43.1-67.0)

33.5 (16.0-52.0)

Complete response N=59

Partial response N=42

Median OS in months (95% CI)e

NE (NE-NE)

37.8 (24.4-NE)

60-month OS rate, % (95% CI)

70.7 (57.2-80.7)

34.7 (19.6-50.4)

Progression-free survival (PFS) is calculated as the number of months from first dose date to the date of first event (disease progression or death) or censoring prior to the data cutoff. bOverall survival (OS) is calculated as the number of months from the first dose date to the date of death or censoring. cIncludes patients without any adequate post-baseline disease assessment. d Duration of response is calculated as the number of months from first documented response to the date of first event (disease progression or death) or censoring prior to the data cutoff. eCalculated as the number of months from the first dose date to the date of death or censoring. MCL: mantle cell lymphoma; MIPI: Mantle Cell Lymphoma International Prognostic Index; CI: confidence interval; CR: complete response; DOR: duration of response; NE: not evaluable; ORR: overall response rate; PD: progressive disease; PFS: progression-free survival; PR: partial response; SD: stable disease. a

In our analysis of acalabrutinib treatment in 124 patients, mDOR, mPFS, and mOS were 28.6 months, 22.0 months, and 59.2 months, respectively, after a median follow-up of 38.1 months, which are longer than those reported in a pooled analysis of ibrutinib. In the analysis of ibrutinib, treatment in 370 patients with R/R MCL (median of 2 prior lines of therapy) resulted in mDOR, mPFS, and mOS of 21.8 months, 12.5 months, and 26.7 months, respectively, after a median follow-up of 41.4 months.5,10 Similarly, in a pooled analysis of zanubrutinib treatment in 112 patients with R/R MCL (median of 2 prior lines of therapy), mDOR,

mPFS, and mOS were 24.9 months, 25.8 months, and 38.2 months, respectively, after a median follow-up of 24.9 months, with the majority of data from a study conducted in Chinese patients.11,12 Notably, the proportion of patients with blastoid and/or pleomorphic histology in the ibrutinib and zanubrutinib analyses compared with our analysis was 12% (n=44) and 13% (n=14), respectively, versus 21% (n=26), and the proportion of patients with high-risk MIPI scores was 32% and 21%, respectively, versus 17%. The patients with blastoid/pleomorphic morphology in this analysis of acalabrutinib achieved an ORR (80.8%) as high

Haematologica | 109 January 2024

346

LETTER TO THE EDITOR as that in the overall study population, with prolonged mPFS and mOS, a finding that was not observed in other high-risk subgroups such as those with high-risk MIPI scores or Ki-67 index >50%. The patients with blastoid morphology achieved an ORR of 50.0% in the ibrutinib analysis and 66.7% in the zanubrutinib study in Chinese patients.12 Although there were no reports of grade ≥3 atrial fibrillation (any grade, 2.4%) and a low incidence of grade ≥3 major hemorrhage (4.0%) in the present analysis, the ibrutinib analysis showed an overall incidence of grade ≥3 atrial fibrillation/flutter of 6.5% (any grade, 11.4%) and an incidence of grade ≥3 hemorrhage of 5.9%. In the zanubrutinib analysis, the incidence of any-grade atrial fibrillation/flutter was 1.8% (grade ≥3, 0.89%) and the

incidence of any-grade major hemorrhage was 5.4%. Although these analyses all evaluated efficacy and safety of BTK inhibitors in R/R MCL, cross-trial comparisons are limited by differences in trial design, heterogeneity in patient populations, and the changing treatment landscape across the periods of enrollment for the various studies. The final results of this study with 38.1 months median follow-up support the use of acalabrutinib in patients with R/R MCL, including those with high-risk features such as blastoid/pleomorphic variants, high-risk MIPI score, and Ki-67 index >50%, and demonstrated a consistent safety profile. Ongoing studies are evaluating acalabrutinib in combination with other agents in this difficult-to-treat aggressive lymphoma.

Continued on following page. Haematologica | 109 January 2024

347

LETTER TO THE EDITOR

Figure 2. Incidence of adverse events (AE) over time for grade ≥3 AE (A), any-grade serious AE (B), and any-grade select AE of interest (C; atrial fibrillation and major hemorrhage). A patient with multiple severity grades for a given AE was counted only once under the maximum severity. Multiple onsets of the same AE within a specific yearly interval were counted once, and the same AE term continuing across several yearly intervals was counted in each of the intervals; y: years.

Authors

Wales, Australia; 14Dana Farber Cancer Institute, Harvard Medical School, Boston, MA, USA; 15Amsterdam University Medical Center, on behalf of the HOVON Lunenburg Lymphoma Phase I/II Consortium,

Steven Le Gouill, Monika Długosz-Danecka, Simon Rule, 1

3,4

Pier Luigi

Amsterdam, the Netherlands; 16Copernicus Memorial Hospital, Medical

Zinzani,5,6 Andre Goy,7 Stephen D. Smith,8 Jeanette K. Doorduijn,9

University of Lodz, Lodz, Poland; 17MD Anderson Cancer Center,

Carlos Panizo,10 Bijal D. Shah,11 Andrew J. Davies,12 Richard Eek,13 Eric

University of Texas, Houston, TX, USA; 18AstraZeneca, Gaithersburg,

Jacobsen,14 Arnon P. Kater,15 Tadeusz Robak,16 Preetesh Jain,17 Roser

MD, USA and 19AstraZeneca, South San Francisco, CA, USA

Calvo,18 Lin Tao19 and Michael Wang17 Correspondence: Institut Curie, Paris, France; 2Maria Sklodowska-Curie National

M. WANG - miwang@mdanderson.org

Research Institute of Oncology, Krakow, Poland; 3Plymouth University Medical School, Plymouth, UK; 4AstraZeneca, Mississauga, Ontario,

https://doi.org/10.3324/haematol.2022.282469

Canada; IRCCS Azienda Ospedaliero-Universitaria di Bologna Istituto 5

di Ematologia “Seràgnoli”, Bologna, Italy; 6Dipartimento di Medicina

Received: December 5, 2022.

Specialistica, Diagnostica e Sperimentale Università di Bologna,

Accepted: July 7, 2023.

Bologna, Italy; John Theurer Cancer Center, Hackensack University

Early view: July 20, 2023.

Medical Center, Hackensack, NJ, USA; University of Washington, Fred 8

Hutchinson Cancer Center, Seattle, WA, USA; 9Erasmus MC Cancer

Institute, on behalf of the HOVON Lunenburg Lymphoma Phase I/II

Published under a CC BY-NC license

Consortium, Rotterdam, the Netherlands; Hospital Universitario 10

Donostia, San Sebastián, Spain; 11Moffitt Cancer Center, Tampa, FL,

Disclosures

USA; NIHR/Cancer Research UK Experimental Cancer Medicines

MD-D discloses consultancy or advisory role at AbbVie, Roche,

Center, University of Southampton, Faculty of Medicine,

Servier and Takeda. SR discloses research funding from Janssen

Southampton, UK; Border Medical Oncology, Albury, New South

Oncology; employment at VelosBio and AstraZeneca. PLZ discloses

Haematologica | 109 January 2024

348

LETTER TO THE EDITOR consultancy or advisory role at Verastem, Celltrion, Gilead, Janssen-

research funding from Janssen, AstraZeneca, Roche/Genentech and

Cilag, Bristol Meyers Squibb, Servier, Sandoz, Merck Sharp & Dohme,

Bristol Meyers Squibb. TR is employed by the Medical University of

TG Therapeutics, Takeda, Roche, EUSA Pharma, Kyowa Kirin, Sanofi

Lodz; has received research funding from Acerta, Roche, Janssen,

and ADC Therapeutics; is part of the speakers’ bureau of Verastem,

AbbVie, Novartis, BioGene, AstraZeneca, Pharmacyclics, UCB, Pfizer,

Celltrion, Gilead, Janssen-Cilag, Bristol Meyers Squibb, Servier, Merck

MorphoSys, UTX-TGR, GSK and Bristol Myers Squibb; has received

Sharp & Dohme, TG Therapeutics, Takeda, Roche, EUSA Pharma and

travel and accommodation expenses from Roche, Janssen and

Kyowa Kirin. AG discloses grant/research support from Acerta,

AbbVie; has received honoraria from Sandoz, UCB, Novartis,

AstraZeneca, Bristol Meyers Squibb, Celgene, Genentech, Genentech-

Octapharma, BioGene, AstraZeneca and Pharmacyclics; discloses

Hoffman La Roche, Hoffman La Roche, Infinity, Janssen, Karyopharm,

consultancy for Sandoz, Takeda and Momenta. PJ discloses

Kite Pharma, MorphoSys, Pharmacyclics, Seattle Genetics and

consultancy for Eli Lilly, Incyte and Kite Pharma; has received

Verastem; is a stock/shareholder of COTA, Alloplex, Resilience and

research funding from BeiGene, AstraZeneca, Kite Pharma and Eli

Genomic Testing Cooperative, acts as a consultant for Pharmacyclics,

Lilly. RC is employed by and is an equity holder of AstraZeneca. LT is

Janssen Global Services, Janssen Biotech, Clinical Advances in

employed by Acerta and discloses previous employment by Clindata

Hematology & Oncology, Elsevier’s Practice Update Oncology, OncLive

Insight. MW discloses consultancy for AbbVie, Acerta Pharma, ADC

Peer Exchange, Celgene, Kite Pharma, Medscape, Michael J.

Therapeutics America, AstraZeneca, Be Biopharma, BeiGene,

Hennessey Associates, Novartis and Xcenda; is on the advisory board

BioInvent, Deciphera, DTRM Biopharma (Cayman) Ltd., Genentech,

of Kite Pharma, Gilead, Pharmacyclics, Elsevier’s Practice Update

InnoCare, Janssen, Kite Pharma, Leukemia & Lymphoma Society, Lilly,

Oncology and Janssen Biotech; is on the board of directors of COTA,

Merck, Milken Institute, Oncternal, Parexel, PeproMene Bio,

Resilience and Genomic Testing Cooperative; discloses consulting

Pharmacyclics and VelosBio; has received research funding from

faculty at Roswell Park, Michael J. Hennessey Associates and

Acerta Pharma, AstraZeneca, BeiGene, BioInvent, Celgene, Genmab,

Physicians Education Resource; is on the scientific advisory board of

Genentech, Innocare, Janssen, Juno Therapeutics, Kite Pharma, Lilly,

Alloplex, Bristol Meyers Squibb and Vincerx; is on the steering

Loxo Oncology, Molecular Templates, Oncternal, Pharmacyclics,

committee of Pharmacyclics and Janssen Biotech. SDS discloses

VelosBio and Vincerx; has received honoraria from AbbVie, Acerta

research funding from ADC Therapeutics, AstraZeneca, Bayer,

Pharma, AstraZeneca, Bantam Pharmaceutical, BeiGene, BioInvent,

BeiGene, Ayala, Bristol Meyers Squibb, De Novo Biopharma,

Dava Oncology, Eastern Virginia Medical School, IDEOlogy Health,

Enterome, Genentech, Ignyta, Incyte, Kymera, Merck, MorphoSys,

Janssen, Kite Pharma, Leukemia & Lymphoma Society, TS Oncology,

Nanjing Pharmaceutical Co. Ltd, Portola and Viracta Therapeutics;

Medscape, Meeting Minds Experts, MD Education, MJH Life Sciences,

discloses consultancy for ADC Therapeutics, AstraZeneca, BeiGene,

Merck, Moffit Cancer Center, Oncology Specialty Group, OncLive,

Epizyme, Karyopharm, Kite Pharma, Incyte, Numan Therapeutics AG,

Pharmacyclics, Physicians Education Resources, Practice Point

AbbVie and Coherus Biosciences. JKD discloses travel and

Communications and Studio ER Congressi. All other authors have no

accommodation expenses from Roche; is on the advisory board of

conflicts of interest to disclose.

Loxo Oncology. CP discloses to be on the speakers bureau of Bristol Myers Squibb, Roche and Janssen; is on the board of directors or

Contributions

advisory committees of Janssen and Roche; discloses

SR and MW designed the study. SLG, SR, PLZ, AG, SDS, JKD, CP,

accommodation expenses from Kite Pharma and AbbVie. BDS is

BDS, AJD, RE, EJ, APK, TR and MW acted as study investigators.

employed by Moffitt Cancer Center; is on the board of directors or

SLG, MDD, SR, PLZ, AG, SDS, JKD, CP, BDS, AJD, RE, EJ, APK, TR and

advisory committees of NCCN; is the vice-chair of the Acute

MW enrolled patients. MDD, AG, JKD, CP, AJD, TR, MW collected and

Lymphoblastic Leukemia Working Group; has received research

assembled data. SR, AG, RC, LT and MW analyzed data. SLG, SR, AG,

funding from Kite Pharma, Gilead, Jazz and Incyte; has received

PJ, RC and MW prepared the manuscript. All authors interpreted

honoraria from Kite Pharma, Gilead, Celgene, Juno, Bristol Meyers

data, revised and reviewed the manuscript and gave their final

Squibb, Novartis, Pfizer, Amgen, Spectrum, Acrotech, Precision

approval of the manuscript.

Biosciences, BeiGene, AstraZeneca, Pharmacyclics, Jansen and Adaptive; has received travel and accommodation expenses from

Acknowledgments

Kite Pharma, Gilead, Precision Biosciences, Novartis and AstraZeneca.

The authors would like to acknowledge Helen Yang of AstraZeneca

AJD discloses consultanty or advisory role at Roche, Celgene,

for biostatistics support.

Gilead/Kite Pharma, Acerta/AstraZeneca, Regeneron, Incyte, Takeda, VelosBio, MorphoSys AG and AbbVie; has received honoraria from

Funding

Roche, Celgene, Gilead/Kite Pharma, Takeda, MorphoSys AG and

This work was funded by AstraZeneca. Writing and editorial

Janssen; has received research funding from Roche, Acerta Pharma,

assistance were provided by Tracy Diaz, PhD, and Sarah Huh,

AstraZeneca, Celgene, Gilead/Kite Pharma, ADC Therapeutics and

PharmD, of Peloton Advantage, an OPEN Health company, and

Janssen; discloses travel grants from Roche and Takeda. EJ discloses

funded by AstraZeneca.

consultancy for Acerta, AstraZeneca and Merck; has received research funding from Merck, Pharmacyclics, F. Hoffmann-LaRoche

Data-sharing statement

and Novartis; has received honoraria from Takeda. APK has received

Data underlying the findings described in this manuscript may be obtained

Haematologica | 109 January 2024

349

LETTER TO THE EDITOR in accordance with AstraZeneca’s data sharing policy described at

through Vivli at https://vivli.org/members/enquiries-about-studies-not-

https://astrazenecagrouptrials.pharmacm.com/ST/Submission/Disclosure

listed-on-the-vivli-platform/. AstraZeneca Vivli member page is also

.Data for studies directly listed on Vivli can be requested through Vivli

available outlining further details:

at www.vivli.org. Data for studies not listed on Vivli could be requested

https://vivli.org/ourmember/astrazeneca/.

References 1. Jain P, Wang ML. Mantle cell lymphoma in 2022 - a comprehensive update on molecular pathogenesis, risk stratification, clinical approach, and current and novel treatments. Am J Hematol. 2022;97(5):638-656. 2. Jain P, Wang M. Blastoid mantle cell lymphoma. Hematol Oncol Clin North Am. 2020;34(5):941-956. 3. Hoster E, Klapper W, Hermine O, et al. Confirmation of the mantle-cell lymphoma International Prognostic Index in randomized trials of the European Mantle-Cell Lymphoma Network. J Clin Oncol. 2014;32(13):1338-1346. 4. Wang ML, Rule S, Martin P, et al. Targeting BTK with ibrutinib in relapsed or refractory mantle-cell lymphoma. N Engl J Med. 2013;369(6):507-516. 5. Rule S, Dreyling M, Goy A, et al. Outcomes in 370 patients with mantle cell lymphoma treated with ibrutinib: a pooled analysis from three open-label studies. Br J Haematol. 2017;179(3):430-438. 6. Wang M, Rule S, Zinzani PL, et al. Acalabrutinib in relapsed or refractory mantle cell lymphoma (ACE-LY-004): a single-arm, multicentre, phase 2 trial. Lancet. 2018;391(10121):659-667. 7. Wang M, Rule S, Zinzani PL, et al. Durable response with single-

agent acalabrutinib in patients with relapsed or refractory mantle cell lymphoma. Leukemia. 2019;33(11):2762-2766. 8. Wang M, Rule S, Zinzani PL, et al. Acalabrutinib monotherapy in patients with relapsed/refractory mantle cell lymphoma: longterm efficacy and safety results from a phase 2 study. Blood. 2020;(Suppl 1):S38-39. 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. Rule S, Dreyling M, Goy A, et al. Ibrutinib for the treatment of relapsed/refractory mantle cell lymphoma: extended 3.5-year follow up from a pooled analysis. Haematologica. 2019;104(5):e211-e214. 11. Zhou K, Zou D, Zhou J, et al. Zanubrutinib monotherapy in relapsed/refractory mantle cell lymphoma: a pooled analysis of two clinical trials. J Hematol Oncol. 2021;14(1):167. 12. Song Y, Zhou K, Zou D, et al. Zanubrutinib in relapsed/refractory mantle cell lymphoma: long-term efficacy and safety results from a phase 2 study. Blood. 2022;139(21):3148-3158.

Haematologica | 109 January 2024

350

LETTER TO THE EDITOR

Germline loss-of-function BRCA1 and BRCA2 mutations and risk of de novo hematopoietic malignancies Breast cancer susceptibility genes type one (BRCA1) and type two (BRCA2) encode tumor suppressor proteins involved in homologous recombination in response to DNA double-strand breaks.1 Germline mutations resulting in loss of BRCA1/2 function prevent cells from undergoing normal homologous recombination in the setting of double-strand breaks, leading to somatic DNA mutations that arise from alternative error-prone DNA repair mechanisms such as non-homologous end joining.1 Germline loss-of-function mutations in BRCA1 occur in about 0.07-0.09% of the population at birth, whereas BRCA2 mutations occur in approximately 0.14-0.22%.2,3 Development of hereditary breast cancer is strongly associated with these germline mutations, and the penetrance for breast cancer by the age of 80 is approximately 48% for carriers of BRCA1 germline mutations and 74% for those with such alleles in BRCA2.2,3 Similarly, the penetrance for ovarian cancer by the age of 80 is about 22% for carriers of deleterious germline variants in either BRCA1 or BRCA2.3 The development of therapy-related myeloid neoplasms (tMN), such as myelodysplastic syndrome and acute myeloid leukemia, as well as other hematopoietic malignancies (HM) is a well-recognized complication of the treatment of solid tumors, and they arise in approximately 2% of patients treated with cisplatin or olaparib.4 Although some recent clinical trials of olaparib have reported lower rates of associated myelodysplastic syndrome/acute myeloid leukemia, data from the Federal Drug Administration pharmacovigilance program continue to suggest an increased risk.5 Germline carriers of deleterious BRCA1/2 variants treated with these agents are at a higher risk of developing a t-MN, and our group and others observe that about 20% of women who have been treated for breast cancer and subsequently develop a t-MN have deleterious germline variants in BRCA1/2 or related genes.6-8 Double-strand break DNA repair pathways are also important to the pathogenesis of de novo myelodysplastic syndrome/acute myeloid leukemia and other HM. However, because most studies on germline risk for HM from germline BRCA1/2 mutations have focused on t-MN, it is not known whether HM risk is conferred exclusively in the setting of therapy. Some previous studies did not find increased risks of de novo HM within patients carrying BRCA mutations; however, these cohorts were ascertained from patients and families primarily presenting with breast or ovarian cancers.9 Despite this, there remains little work examining the role that BRCA mutations play in the subpopulation of patients and families who present with her-

editary HM rather than solid organ malignancies. Therefore, we sought to determine the frequency and prevalence of germline BRCA1/2 mutations in patients with de novo HM without a prior diagnosis of a solid organ malignancy and in the absence of prior cytotoxic treatment or radiotherapy. We identified patients at the University of Chicago diagnosed with a HM from February 2012 to October 2020, who were heterozygous for pathogenic (P) or likely pathogenic (LP) germline BRCA1 or BRCA2 variants. All of the patients had signed informed consent to research. We included patients with de novo HM as well as those with a t-MN, who served as a comparator group. Germline BRCA1/2 testing criteria included a personal and family history of a HM as well as the National Comprehensive Cancer Network malignancy-specific guidelines for germline testing.3 Testing for germline BRCA1/2 variants was performed in a CLIA-certified clinical genomics laboratory using standard clinical protocols. Variants were interpreted as per American College of Medical Genetics and Genomics/Association for Molecular Pathology guidelines, with germline status of the BRCA1/2 variant confirmed by DNA derived from cultured skin fibroblasts (Online Supplementary Table S1). The variants for the gnomAD control cohort were obtained through the web portal, for the gnomAD non-cancer population, version 3.1.2.10 For the allele burden calculation, variants were considered damaging if they were known P, defined as having been deposited as P or LP into ClinVar, or were expected to result in clear loss-of-function by disrupting the protein product (e.g., large deletions, frameshifts). Variants with conflicting interpretations in ClinVar were included if the majority (>50%) of deposits were pathogenic. HM were assessed for acquired loss of heterozygosity (LoH) via clinical molecular profiling and next-generation sequencing as well as by clinical karyotyping/fluorescence in situ hybridization analysis for loss of chr17 (BRCA1) or chr13 (BRCA2), with copy number variations confirmed by chromosomal microarrays for those with available samples to avoid missing smaller chromosome rearrangements that could not be seen on clinical cytogenetic/fluorescence in situ hybridization analysis. We identified 25 individuals with a diagnosis of a HM who also had deleterious germline BRCA1 or BRCA2 variants: 14 in BRCA1, and 11 in BRCA2 (Online Supplementary Table S1). The patients’ demographics and clinical findings are presented in Table 1, and the germline variants are detailed in Online Supplementary Table S1. The most common ethnic backgrounds were Western European (BRCA1; 7 [50%], BRCA2; 6 [55%]) and Ashkenazi Jewish (BRCA1; 4 [29%],

Haematologica | 109 January 2024

351

LETTER TO THE EDITOR Table 1. Characteristics of patients carrying BRCA1/2 mutations diagnosed with hematopoietic malignancies. Patients’ characteristics

BRCA1, N=14

BRCA2, N=11

Gender, N (%)† Female Male

10 (71) 4 (29)

7 (64) 4 (36)

Ethnic background, N (%) Western European Ashkenazi Jewish Eastern European Northern European African American Unknown

7 (50) 4 (29) 2 (14) 1 (7) 0 (0) 0 (0)

6 (55) 1 (9) 1 (9) 0 (0) 1 (9) 2 (18)

Age at first malignancy in yrs, median (range)‡

48.7 (20.1-74.6)

66.7 (36.6-74.4)

Age at first malignancy in yrs (HM vs. solid), median (range) HM as first diagnosis Solid tumor as first diagnosis

49.5 (21.3-67.4) 63.4 (52.8-70.3)

47.3 (20.1-52.7) 66.9 (36.6-74.4)

Number of malignancies, N (%) 1 2 3

3 (21) 8 (57) 3 (21)

6 (55) 4 (36) 1 (9)

Order of HM, N (%) HM diagnosed as first cancer HM diagnosed after prior cancer

8 (57) 6 (43)

8 (73) 3 (27)

N=28 total 11 (39) 6 (21) 0 (0) 7 (25) 2 (7) 1 (4) 1 (4) 0 (0) 0 (0) 0 (0)

N=17 total 6 (35) 2 (12) 4 (24) 1 (6) 0 (0) 1 (6) 0 (0) 1 (6) 1 (6) 1 (6)

Treatment received for solid malignancy, N (%) Chemotherapy Radiation Surgery

3 (21) 2 (14) 10 (71)

2 (18) 2 (18) 3 (27)

Treatment received for HM, N (%) Chemotherapy Radiation Autologous hematopoietic stem cell transplant Allogeneic hematopoietic stem cell transplant

14 (100) 1 (7) 0 (0) 3 (21)

11 (100) 1 (9) 4 (36) 4 (27)

Type of malignancy, N (%) Myeloid Lymphoid Multiple myeloma Breast Gynecological Prostate Colon Melanoma Thyroid Bladder

χ P=0.178 (NS) for gender difference between germline BRCA1 and BRCA2 mutation carriers. ‡t-test P=0.014 for difference in age of first malignancy between germline BRCA1 and BRCA2 mutation carriers. yrs: years; HM: hematopoietic malignancy; NS: not statistically significant.

† 2

BRCA2; 1 [9%]). Contrary to the prevailing notion that individuals with P/LP BRCA1/2 variants develop t-MN exclusively, eight (57%) of BRCA1-mutated and eight (73%) of BRCA2-mutated patients developed a HM as their first cancer. Several of these patients developed two (BRCA1: 8 [57%], BRCA2: 4 [36%]) or more than two (BRCA1: 3 [21%], BRCA2: 1 [9%]) independent malignancies, which included a range of diagnoses, both commonly associated (e.g., breast and prostate cancer) and not commonly associated

(e.g., colon cancer) with BRCA1/2 mutations. An allele burden represents a calculation of the relative frequency at which a mutated allele is found in a population with a phenotype of interest (e.g., HM) versus a control population (e.g., healthy gnomAD controls). Comparing the allele burden of probands with P or loss-of-function germline BRCA mutations who had a de novo HM diagnosis within the cohort of all patients with clinical panel-based testing for germline predisposition to a diagnosed HM over the

Haematologica | 109 January 2024

352

LETTER TO THE EDITOR same time period within the overall University of Chicago cohort versus a non-cancer control population (gnomAD v.3.1.2, non-cancer), there is enrichment for P or loss-offunction BRCA1 (OR=8.61, 95% CI: 4.00-18.51, P<0.0001) and BRCA2 (OR=4.65, 95% CI: 2.29-9.43, P<0.0001) alleles (Table 2). In the germline BRCA1-mutated patient cohort (n=14), eight developed a HM, either myeloid or lymphoid, as their first cancer diagnosis (Figure 1A). Subtypes included: chronic myeloid leukemia (n=1), myelodysplastic syndrome (n=1), systemic mastocytosis (n=1), Hodgkin lymphoma (n=1), acute lymphoblastic leukemia (n=1), diffuse large B-cell lymphoma (n=1), chronic lymphocytic leukemia (n=1), and follicular lymphoma (n=1). The average age at diagnosis of the first HM was 49.5 years (range, 21.3-67.4 years), compared to 63.4 years (range, 52.8-70.3 years) for the patients with deleterious germline BRCA1 variants who developed a solid tumor malignancy first. Among the eight who developed a HM first, five went on to develop secondary solid malignancies, including breast (n=3), prostate (n=1), and colon (n=1) cancer, with two of these patients then developing a tertiary t-MN. The other six developed a solid tumor as their first cancer (Figure 1B). Subtypes included breast (n=4), ovarian (n=1), and Fallopian tube (n=1). All six of these patients went on to develop a secondary t-MN, with one patient developing a tertiary lymphoid neoplasm. Treatment received included: chemotherapy (17/33 treatment events, 52%), radiation (3/33 treatment events, 9%), allogeneic hematopoietic stem cell transplantation (3/33 treatment events, 9%), and surgery (3/33 treatment events, 9%). Within the overall University of Chicago cohort with clinical cancer predisposition testing and a HM, patients with deleterious germline BRCA1 variants constituted 1.6% (17/1074) of overall HM diagnoses, 2% (11/544) of myeloid diagnoses, 1.6% (6/371) of lymphoid diagnoses, and no multiple mye-

loma diagnoses. In the BRCA2 patient cohort (n=11), eight individuals developed a HM, either myeloid or lymphoid/multiple myeloma, as their first cancer diagnosis (Figure 1C). Subtypes included multiple myeloma (n=4), acute myeloid leukemia (n=3), and chronic lymphoytic leukemia (n=1). The average age of a HM diagnosis as the first cancer was 47.3 years (range, 20.1-52.7 years) versus 66.9 years (range, 36.6-74.4 years) for patients who developed a solid organ malignancy first. Among the eight patients who developed a HM as a first diagnosis, one subsequently developed prostate cancer and one developed therapy-related acute lymphocytic leukemia. The other three BRCA2 patients developed a solid organ malignancy as their first cancer, including breast cancer (n=1), melanoma (n=1), and bladder cancer (n=1). The one patient with melanoma subsequently developed thyroid cancer, which was treated with radiation. All three of these patients eventually developed a t-MN. The treatments received in the BRCA2 cohort included chemotherapy (13/26 treatment events, 50%), radiation (3/26 treatment events, 12%), autologous stem cell transplantation (4/26 treatment events, 15%), allogeneic stem cell transplantation (3/26 treatment events, 12%) and surgery (7/26 treatment events, 27%). Within the overall University of Chicago cohort with clinical cancer predisposition testing and a HM, patients with deleterious germline BRCA2 variants accounted for 1.1% (12/1074) of overall HM, 1.1% (6/544) of myeloid diagnoses, 0.5% (2/371) of lymphoid diagnoses, and 1.8% (4/218) of multiple myeloma diagnoses. Analysis of BRCA1/2 within solid tumors arising in individuals with deleterious germline variants show LoH in 90% of BRCA1-associated breast cancer and 54% of BRCA2-associated breast tumors, functionally implicating the importance of the germline BRCA gene in tumor development.11 We therefore tested for BRCA1/2 LoH within leukemic cells

Table 2. Allele burden calculation for loss-of-function BRCA1/2 variants and occurrence of primary de novo hematopoietic malignancies. Probands with P/LoF† germline BRCA mutations with clinical panel testing for a first HM diagnosis BRCA1

BRCA2

Total

P/LoF† germline BRCA variants in a non-cancer gnomAD control population (v3.1.2)

7 total variants‡

112 total variants

1,074 total tests

147,895 total alleles*

8 total variants

237 total variants¶

1,074 total tests

147,870 total alleles*

15 total variants

1,074 total tests

Allele burden calculation OR

95% CI

8.61

4.00-18.51

<0.0001

4.65

2.29-9.43

<0.0001

Pathogenicity was assessed as per ClinVar deposit, conflicting reports included if majority deposits were pathogenic or likely pathogenic; if no ClinVar deposition, clear loss-of-function protein damaging insertions or deletions were included. ‡One patient excluded due to non-proband status/research initiated testing. *Represents the average of total allele numbers reported per variant. ¶Excludes the BRCA2 p.Lys3326Ter variant reported as benign in ClinVar. P: pathogenic; LoF: loss-of-function; HM: hematopoietic malignancy; gnomAD: Genome Aggregation Database; OR: odds ratio; CI: confidence interval. †

Haematologica | 109 January 2024

353

LETTER TO THE EDITOR

Figure 1. Time course of malignancies in patients with loss-of-function germline BRCA1/2 variants. The swimmer plot depicts each patient in a distinct row, with the age of diagnosis of the malignancy indicated at the end of each block and the color denoting the type of malignancy. Symbols between intervals denote treatments: chemotherapy, autologous (auto) or allogeneic (allo) hematopoietic stem cell transplant (SCT), surgery, and radiation (XRT). (A) Half of germline BRCA1 mutation carriers had a first diagnosis of hematopoietic malignancies, comprising myeloid (chronic myeloid leukemia [N=1], myelodysplastic syndrome [N=1], and systemic mastocytosis [N=1]) and lymphoid (Hodgkin lymphoma [N=1], acute lymphocytic leukemia [N=1], and nonHodgkin lymphoma [N=3]) malignancies. (B) Among the six patients with a germline BRCA1 mutation who developed a solid organ malignancy first, all developed a subsequent therapy-related myeloid neoplasm, with one developing a tertiary lymphoid neoplasm. (C) Among the germline BRCA2-mutated patients, eight developed a hematopoietic malignancy first: multiple myeloma (N=4), acute myeloid leukemia (N=3), and chronic lymphocytic leukemia (N=1). One patient subsequently developed prostate cancer and one a therapy-related acute lymphocytic leukemia. (D) Three of the BRCA2-mutated patients developed solid organ malignancies first: melanoma (N=1), bladder (N=1), and breast (N=1), all of whom subsequently developed therapy-related myeloid neoplasms.

(Online Supplementary Table S1). Among patients with germline BRCA1 mutations and material evaluable for LoH, 13% (1/8) had evidence of LoH, with a chr17 deletion. Among similar patients with germline BRCA2 mutations, 33% (2/6) had evidence of LoH, with one showing loss of chr13 and one having a somatic BRCA2 mutation with a variant allele frequency of 27%. Overall, among germline BRCA1/2 mutation carriers, we observed a 23% rate of LoH. These findings suggest that deleterious germline BRCA1/2 variants are seen frequently in patients with de novo HM, and it is possible that they may confer risk for de novo HM, outside of the context of previous exposure to cytotoxic chemotherapy or radiation. As with other germline mutations, it is important to remember that deleterious germline BRCA1/2 alleles are present in all cells of the body, and consequently, homologous recombination activity is defective globally. Hematopoietic cells may be particularly vul-

nerable to defective DNA repair mechanisms as they must divide continually throughout an individual’s lifetime, and, thus, are more vulnerable to genotoxic stress. Further, overactivity of non-homologous end joining relative to homologous recombination promotes the development of leukemogenic hematopoietic stem and progenitor cells. In keeping with this, mice lacking Brca1 in the hematopoietic compartment develop bone marrow failure and HM after stress erythropoiesis.12 Mutations in other components of BRCA adjacent pathways, particularly the FANC gene family known to be defective in Fanconi anemia, are also well described as conferring risk for the development of spontaneous myeloid malignancies and other HM.13 One novel finding in our study was the unexpected prevalence of lymphoid malignancies, particularly multiple myeloma in patients with deleterious germline BRCA2 variants, given that mutations in FANC and other genes related to homologous

Haematologica | 109 January 2024

354

LETTER TO THE EDITOR recombination tend to have hematopoietic manifestations related to myeloid malignancies and bone marrow failure. However, accumulation of chromosomal abnormalities and gene mutations is also fundamental to the pathogenesis of lymphoid neoplasms, and one report linked germline BRCA2 mutations to a risk for non-Hodgkin lymphoma in adolescents and young adults, and a mouse model of truncating Brca2 mutations was shown to develop thymic lymphomas.14,15 Our study represents a retrospective, single-center patient cohort, and the attendant possibility of ascertainment or referral bias in our patient database must be considered when interpreting these findings. In addition, differences in population composition between our cohort and the gnomAD comparison cohort (e.g., a higher proportion of Ashkenazi Jewish patients in our cohort) may also represent a confounding factor. Differences in upstream variant calling and curation between our local testing and gnomAD could be another confounder. However, we would note that similar allele burden testing approaches using public data such as gnomAD are an accepted approach in the literature and have been applied to other disorders. Among a handful of previous smaller studies that have addressed the question of de novo HM development in BRCA mutation carriers, some have demonstrated an increased risk.15-17 An appreciable number, though not all, of the patients in our cohort demonstrated LoH at the second BRCA1/2 allele through either chromosomal loss or a damaging somatic mutation. However, only about half of BRCA2 germline carriers with breast cancer will have LoH, suggesting that LoH may not be found in the tumor tissue of all patients with BRCA1/2-associated malignancies. Although it is possible that in patients with HM in the absence of LoH, the germline BRCA1/2 mutation is merely a bystander, we note that other solid organ malignancies (e.g., BRCA-associated pancreatic cancer) also display variable degrees of LoH, making the exact role of LoH in the pathogenesis of BRCA-associated malignancies controversial.18 Further, HM are often epigenetically-driven diseases, and there are multiple studies demonstrating promoter hypermethylation and epigenetic silencing of BRCA1 and BRCA2 in HM, particularly in myeloid malignancies.19,20 We were unable to assess for epigenetic loss of BRCA1/2 in our cohort, but future studies should assess whether this is a prominent mechanism of LoH in patients with HM and germline BRCA1/2 mutations. Although some larger cohorts have suggested that there is no increased risk for HM in carriers of BRCA mutations,9 these studies primarily ascertained patients being tested for a known personal or family history of breast or ovarian cancer, possibly pre-selecting for a specific disease phenotype, utilized self-reported family histories, and are susceptible to misclassification bias. Although our work suggests the possibility of an association between germline BRCA mutations and de novo HM development, larger fu-

ture studies focused on populations specifically ascertained with a question of a familial predisposition to HM will be needed to establish definitively the causality and magnitude of risk of developing de novo HM with these alleles.

Authors Ryan J. Stubbins,1,2,3* Anase S. Asom,1* Peng Wang,4 Angela M. Lager,4 Alexander Gary4 and Lucy A. Godley1,5 Section of Hematology/Oncology, Department of Medicine, The

University of Chicago, Chicago, IL, USA; 2Division of Hematology, Department of Medicine, University of British Columbia, Vancouver, British Columbia, Canada; 3Leukemia/BMT Program of BC, BC Cancer, Vancouver, British Columbia, Canada; 4Department of Pathology, The University of Chicago, Chicago, IL, USA; and 5Department of Human Genetics, The University of Chicago, Chicago, IL, USA *RJS and ASA contributed equally as first authors. Correspondence: L.A. GODLEY - lucy.godley@northwestern.edu https://doi.org/10.3324/haematol.2022.281580 Received: June 12, 2022. Accepted: April 24, 2023. Early view: May 4, 2023. ©2024 Ferrata Storti Foundation Published under a CC BY-NC license Disclosures RJS has received honoraria from AbbVie. The other authors have no conflicts of interest to disclose. Contributions RJS, ASA, and LAG reviewed and analyzed the clinical data and wrote and edited the manuscript. PW analyzed the molecular profiling data of all patients. AML and AG reviewed clinical cytogenetics and FISH analysis and also performed and analyzed chromosomal microarrays on all available patients’ material. RJS and ASA generated the Tables and Figure. Acknowledgments We are deeply indebted to our patients and their families for their continued participation in our research. We also thank Stephen Arnovitz and Michael Drazer for their comments on this work Data-sharing statement Clinical sequencing data are available upon request.

Haematologica | 109 January 2024

355

LETTER TO THE EDITOR

References 1. Verma P, Greenberg RA. Communication between chromatin and homologous recombination. Curr Opin Genet Dev. 2021;71:1-9. 2. Tung NM, Boughey JC, Pierce LJ, et al. Management of hereditary breast cancer: American Society of Clinical Oncology, American Society for Radiation Oncology, and Society of Surgical Oncology Guideline. J Clin Oncol. 2020;38(18):2080-2106. 3. Daly MB, Pilarski R, Yurgelun MB, et al. NCCN guidelines insights: genetic/familial high-risk assessment: breast, ovarian, and pancreatic, version 1.2020. J Natl Compr Canc Netw. 2020;18(4):380-391. 4. Liang F, Zhang S, Xue H, Chen Q. Risk of second primary cancers in cancer patients treated with cisplatin: a systematic review and meta-analysis of randomized studies. BMC Cancer. 2017;17(1):871. 5. Zhao Q, Ma P, Fu P, et al. Myelodysplastic syndrome/acute myeloid leukemia following the use of poly-ADP ribose polymerase (PARP) inhibitors: a real-world analysis of postmarketing surveillance data. Front Pharmacol. 2022;13:912256. 6. Churpek JE, Marquez R, Neistadt B, et al. Inherited mutations in cancer susceptibility genes are common among survivors of breast cancer who develop therapy-related leukemia. Cancer. 2016;122(2):304-311. 7. Singhal D, Hahn CN, Feurstein S, et al. Targeted gene panels identify a high frequency of pathogenic germline variants in patients diagnosed with a hematological malignancy and at least one other independent cancer. Leukemia. 2021;35(11):3245-3256. 8. Schulz E, Valentin A, Ulz P, et al. Germline mutations in the DNA damage response genes BRCA1, BRCA2, BARD1 and TP53 in patients with therapy related myeloid neoplasms. J Med Genet. 2012l;49(7):422-428. 9. Li S, Silvestri V, Leslie G, et al. Cancer risks associated with BRCA1 and BRCA2 pathogenic variants. J Clin Oncol. 2022;40(14):1529-1541.

10. Lek M, Karczewski KJ, Minikel EV, et al. Analysis of proteincoding genetic variation in 60,706 humans. Nature. 2016;536(7616):285-291. 11. Maxwell KN, Wubbenhorst B, Wenz BM, et al. BRCA locusspecific loss of heterozygosity in germline BRCA1 and BRCA2 carriers. Nat Commun. 2017;8(1):319. 12. Vasanthakumar A, Arnovitz S, Marquez R, et al. Brca1 deficiency causes bone marrow failure and spontaneous hematologic malignancies in mice. Blood. 2016;127(3):310-313. 13. Przychodzen B, Makishima H, Sekeres MA, et al. Fanconi anemia germline variants as susceptibility factors in aplastic anemia, MDS and AML. Oncotarget. 2018;9(2):2050-2057. 14. Friedman LS, Thistlethwaite FC, Patel KJ, et al. Thymic lymphomas in mice with a truncating mutation in Brca2. Cancer Res. 1998;58(7):1338-1343. 15. Wang Z, Wilson CL, Armstrong GT, et al. Association of germline BRCA2 mutations with the risk of pediatric or adolescent nonHodgkin lymphoma. JAMA Oncol. 2019;5(9):1362-1364. 16. Iqbal J, Nussenzweig A, Lubinski J, et al. The incidence of leukaemia in women with BRCA1 and BRCA2 mutations: an international prospective cohort study. Br J Cancer. 2016;114(10):1160-1164. 17. Pouliot GP, Degar J, Hinze L, et al. Fanconi-BRCA pathway mutations in childhood T-cell acute lymphoblastic leukemia. PLoS One. 2019;14(11):e0221288. 18. Mekonnen N, Yang H, Shin YK. Homologous recombination deficiency in ovarian, breast, colorectal, pancreatic, non-small cell lung and prostate cancers, and the mechanisms of resistance to PARP inhibitors. Front Oncol. 2022;12:880643. 19. Poh W, Dilley RL, Moliterno AR, et al. BRCA1 promoter methylation is linked to defective homologous recombination repair and elevated miR-155 to disrupt myeloid differentiation in myeloid malignancies. Clin Cancer Res. 2019;25(8):2513-2522. 20. Yang C, Zhang Q, Tang X, et al. BRCA2 promoter hypermethylation as a biomarker for the leukemic transformation of myeloproliferative neoplasms. Epigenomics. 2022;14(7):391-403.

Haematologica | 109 January 2024

356

CASE REPORT

Immune-related adverse events with bispecific T-cell engager therapy targeting B-cell maturation antigen Bispecific antibodies (BispAb) targeting BCMAxCD3 have been recently approved for the treatment of patients with relapsed and refractory myeloma previously exposed to immunomodulatory drug, proteasome inhibitor, and antiCD38 antibodies. For instance, the phase I/II MajesTEC-1 study demonstrated high efficacy of teclistamab in advanced (median of 5 prior lines), triple-class exposed and refractory myeloma, with an overall response rate of 63% and median progression-free survival of 11.3 months.1 In BispAb studies, most common treatment-emergent adverse events (AE) were cytokine release syndrome (CRS), hematological toxicities and immune effector cell associated neurotoxicity syndrome. Dysimmune events were neither observed nor classified as such but were exclusion criteria for protocol enrollment.2 Immune-related adverse events (irAE) are well described for the immune checkpoint inhibitor therapy such as anti-PD1 or antiCTLA-4. This type of AE could affect different organs, such as gastrointestinal, endocrinological, skin or rheumatological systems.3 The explanation of these irAE under checkpoints inhibitors seems straightforward with the activity enhancement of T cells.3 Thus, through two thoroughly documented clinical cases, we report here how dysimmunity could be induced by T-cell engaging BispAb, with very different patterns of symptoms. Case 1 Case 1 is a 77-year-old woman without a significant medical history, who has been treated for IgG κ multiple myeloma. Her ninth line consisted of a CD3xBCMA BispAb, a stringent complete remission (sCR) from the 13th cycle was obtained. A grade 1 CRS was noted at first cycle and was resolutive. Shortly after sCR, the patient presented with recurring oligoarthritis. There was a biological inflammatory syndrome (C reactive protein [CRP] increased and thrombocytosis). Blood cultures were positive twice for Neisseria cinerea, and while echocardiography showed no signs of infective endocarditis, she was treated with ceftriaxone (2 g/day) for 14 days and the oligoarthritis was resolved (Figure 1). The positron emission tomography-comuted tomography (PET-CT) scan that had been performed for the disease’s extension showed grade 3 (the highest grade) parietal hypermetabolism of the thoracic aorta, brachiocephalic artery trunk, right subclavian artery, and left common iliac artery (Figure 2A-D, bottom row with arrows). On examination, no recent or previous cephalic, visual, mandibular, or rhizome manifestations, no chest, back, abdominal, or lumbar pain with an inflam-

matory appearance were observed. The temporal and peripheral pulses were perceived without vascular murmur. Except moderate inflammatory syndrome with a CRP of 8 mg/L, blood results were normal. Blood cultures and infectious serologies were negative. Systematic etiological exams for inflammatory vasculitis were negative. The temporal artery biopsy didn’t demonstrate Horton's disease. Ultrasound-Doppler of the encephalic and temporal arteries demonstrated a significant diffuse hyperplasia of the intima, at the origin of the right subclavian artery. A segmental inflammatory aspect of the trunk of the left temporal artery with halo sign was also noticed (UltraSound findings, data not shown). Given the context of severe immunosuppression and recent infection, this arthritis was presumed to be related to an infectious cause, relying on ceftriaxone treatment to recover. No improvement of PETCT findings at 8 weeks from artritis diagnosis were noted. It was decided to start a second line therapy with cotrimoxazole for 8 weeks. The patient underwent a new PETCT after 2 months, with no improvement (vasuclar PET scan [PETVasc] global 27/48). A PETVasc is an additional score with a maximum score at 48 (Figure 2A, C, top and bottom rows). A systemic corticosteroid therapy at a dose of 0.7 mg/kg/day in progressive decrease over 12 months was introduced (Figure 1). Several PET-CT showed a clear decrease in segmental hyperfixation of the wall of large and medium-sized vessels (PETVasc global 23/48) (Figure 2A, C, D). End-of-treatment PET-CT, performed just after steroid withdrawal, showed an iconographic relapse. Referring to the teclistamab-induced vasculitis, internists introduced tocilizumab subcutaneous weekly, with maintening corticotherapy. teclistamab was stopped during the first month (Figure 1). Case 2 The second case reported here concerns a 66-year-old woman with a notable history of immune thrombocytopenia (ITP) refractory to corticosteroid (splenectomy in 1995). She was diagnosed with a symptomatic multiple myeloma, treated with five lines. After the first cycle of CD3xBCMA BispAb, a grade 1 CRS with ITP decompensation (resolved with intravenous immunglobulin) was observed. At the beginning of cycle 11, the patient presented with an abrupt polyarticular presentation with distal interphalangeal joint and proximal interphalangeal joint (PIJ) swelling of the third and second fingers of the left hand, as well as the carpometacarpal joint of the second finger of the left hand. Pain of the left knee and hip, with morning rusting were also

Haematologica | 109 January 2024

357

CASE REPORT

Figure 1. Clinical, radiological, and therapeutical chronology of the two cases. M: month; sCR: stringent complete remission; VGPR: very good partial response; TEP: tomography by emission of positrons.

notified. At the beginning of the 12th cycle, the inflammatory rhythmed polyarthralgias were still present, but migratory also with swelling of the ankles. No family history of inflammatory rheumatism, no uveitis, no dactylitis, no transit abnormalities were reported. Blood testing was normal for uric acid. The immune checkup came back as negative (antinuclear antibodies, rheumatoid factors and anti-CCP) as well as infectious investigations (blood culture, hepatitis polymearse chain reaction). On X-rays, there were some erosions suspected as consequences of inflammation without argument for chondrocalcinosis or osteoarthritis. Ultrasound-Doppler images (Figure 2E-H), there was a grade 2 synovitis of wrists and PIJ. Furthermore, there was a tenosynovitis of the common extensors of the fingers and there were bilateral crural synovitis and fibular tenosynovitis of the ankles. The patient has been recovering gradually under corticosteroid therapy introduced as 10 mg per day with a continuous tapering. From the first dose decrement, complete regression of the morning rusting in the lower limbs was observed, followed by disappearance of hands and wrist synovitis symptoms. Teclistamab was discontinued 3 months (Figure 1).

These two cases illustrate two presumed irAE due to BispAb during multiple myeloma treatment. Indeed, in the first case, it seems particularly challenging to decipher whether it is primary inflammatory vasculitis or potentially induced by BispAb. However, the germ involved here, which usually damages the small vessel walls directly, has been controlled by prolonged antibiotic therapy and is, therefore, unlikely to be the cause of persistent vasculitis.4 Relating those events to drug exposure may be contra-intuitive as most AE related to BispAb are on the side of immunosuppression. Indeed, it may seem paradoxical that an antibody targeting BCMA can induce autoimmune reactions while some drugs targeting its soluble receptor/ligand (BAFF) is used to treat autoimmune disease with belimumab for instance,5 this situation has been extensively described for anti-TNFα therapy in genesis lupus.6 Biological rationale may enlighten this particular course of events. Important to remember is that BCMA is a transmembrane receptor naturally expressed on plasma cells and is consistently expressed at high levels on malignant plasma cells. BCMA binds to BAFF, a cytokine expressed by B-lineage cells, and leads to the activa-

Haematologica | 109 January 2024

358

CASE REPORT

Figure 2. Positron emission tomography and ultrasound pictures of the two cases. (A-D) Positron emission tomography (PET)computed tomography (CT) scans showing grade 3 parietal hypermetabolism of the thoracic aorta, brachiocephalic artery trunk, right subclavian artery, and left common iliac artery. Case 1 : (A) PET coronals, (B) CT coronals, (C) fused coronals, (D) whole body PET. Top row: after corticotherapy; bottom row: before corticotherapy. Red arrow: right subclavian artery hypermetabolism; orange arrow: right brachiocephalic trunk hypermetabolism. Case 2 (E-H): (E) right radiocarpal synovitis, longitudinal section (B mode), (F) right tenosynovitis, extensor digitorum communis (B mode), (G) right tenosynovitis, flexor digitorum communis (B mode), (H): right tenosynovitis, extensor digitorum brevis and extensor digitorum longus with Doppler hyperemia. Triangle stands for radial and medial carpal synovitis (E only). Numbers 1, 2, and 3, respectively stand for radius, lunatum, and capitatum; fiveand six-pointed stars respectively stand for extensor digitorum communis, and flexor digitorum communis (G). Diamond in (H) stands for the extensor digitorum brevis and the extensor digitorum longus. Large arrow in (F, G) highlights synovitis effusion around tendon sheaths. Thin arrow in (H) demonstrates doppler vascularization.

Haematologica | 109 January 2024

359

CASE REPORT Table 1. Effects of the lack of BAFF or BCMA in murine lupus model. Deficiency

Mouse strain

Result

BAFF

NZM2328

Delay in autoantibodies

BCMA

Nba2.Yaa

Enhances autoantibodies and exacerbates disease

NZM2328

Increased spleen size No effect on autoantibodies, plasma cells or mortality

Table adapted with permission from Jackson and Davidson. Immunol Rev. 2019.9

tion of kinases (MAPK8/JNK) by NFκB transcription that stimulate plasma cells to maintain humoral immunity. BAFF is required for normal B-cell homeostasis, but it also promotes the survival of malignant B cells. BCMA can also bind to a related protein, APRIL involved in B-cell development and autoimmune response. The complex interactions between BCMA and its ligands BAFF and APRIL explain BCMA implication in the maintenance of plasma cells, but also in several cancers, autoimmune and infectious diseases.7 This effect could be explained by the increase of the BCMA’s ligand levels along with the inhibition of its receptor via bispecific therapies in multiple myeloma (CD3-BCMA). BCMA may give rise to dysimmune disease via its targeting. Supporting this hypothesis, BCMA deficiency in mice exacerbated phenotypes with dysimmune diseases8 (Table 1). Different studies have shown that serum levels of BAFF were associated with a major disease activity.10 When BCMA is allosterically occupied, other BCMA ligands could increase, and would likely bind to its targets (BAFF-R and APRIL-R) (Table 1). Indeed, this explains why some immunotherapy treatments for autoimmune diseases (such as lupus or vasculitis11), take leverage of this effect, such as atacicept12 which targets both BAFF and APRIL-R; or tabalumab13 which also targets both BAFF and BCMA. The second hypothesis for the development of post-BCMA autoimmunity could be explained by the expansion of follicular T-helper cells. Indeed, the uncontrolled expansion of follicular T-helper cells (TFH) activates autoreactive B cells to produce antibodies inducing autoimmunity. TFH cells express BCMA and BR3 receptors and accumulate in the spleen when BCMA is absent. BCMA deficiency in T cells can promote TFH cell expansion, autoantibody production, and IFN production by TFH cells via BR3.3. In an experimental study,14 blocking BAFF or IFNγ in vivo reduced TFH cell accumulation and improved autoimmunity in BCMA-deficient animals but not in BR3-deficient ones. BCMA acts intrinsically to T cells to limit the number of autoreactive TFH, and thereby regulates autoreactive B-cell responses.15 BAFF blocking agents could be evaluated to reduce autoimmune reactions related to BCMAxC3 BispAb. However, we cannot neglect that there were no strong reports of dysimmunity with other anti-BCMA therapies as

cytotoxic coupled-agents (e.g., belantamab mafodotin) or chimeric antigen receptor T cells (e.g., Ide-cel or Cilta-cel). Thus, we can hypothesize that bispecific agents with T-cell engagement may increase unspecifc T-cell response or reverse T-cell exhaustion,16 leading to dysimmunity. Against this assumption, dysimmune events with using other BispAb (targeting GPRC5D or FcRH5) were not reported. Thus, incidence and physiopathology of these dysimmune complications need to be explored in prospective studies with more patients and longer follow-up.

Authors Bénédicte Piron,1 Mathilde Bastien,1 Chloé Antier,1 Romain DallaTorre,2 Bastien Jamet,3 Thomas Gastinne,1 Viviane Dubruille,1 Philippe Moreau,1,4,5 Jérôme Martin,6 Antoine Bénichou,7 Cyrille Touzeau1,4,5 and Benoît Tessoulin1,4 1

Department of Hematology, University Hospital; 2Department of

Rheumatology, University Hospital; 3Department of Nuclear Medicine, University Hospital; 4Centre de Recherche en Cancérologie et Immunologie Intégrée Nantes Angers, INSERM UMR 1307, CNRS UMR 6075; 5Site de Recherche Intégrée sur le Cancer (SIRIC), Imaging and Longitudinal Investigations to Ameliorate Decision-Making (ILIAD), INCA-DGOS-INSERM 12558, Immunology Laboratory, CIMNA; 6Immunology Laboratory, CIMNA and 7

Department of Internal and Vascular Medicine, University Hospital,

Nantes, France Correspondence: B. TESSOULIN - benoit.tessoulin@chu-nantes.fr https://doi.org/10.3324/haematol.2023.282919 Received: March 15, 2023. Accepted: July 11, 2023. Early view: July 20, 2023. ©2024 Ferrata Storti Foundation Published under a CC BY-NC license Disclosures No conflicts of interest to disclose. Contributions BP, CT and BT wrote the manuscript. BP, MB, CA, TG, VD, CT, RDT, AB and BT treated the patients. RDT and BJ provided imaging data. All authors critically reviewed the manuscript. Data-sharing statement Data supporting this article are available from the corresponding author.

Haematologica | 109 January 2024

360

CASE REPORT

References 1. Moreau P, Girgis S, Goldberg JD. Teclistamab in relapsed or refractory multiple myeloma. Reply. N Engl J Med. 2022;387(18):1722-1723. 2. Usmani SZ, Garfall AL, van de Donk NWCJ, et al. Teclistamab, a B-cell maturation antigen × CD3 bispecific antibody, in patients with relapsed or refractory multiple myeloma (MajesTEC-1): a multicentre, open-label, single-arm, phase 1 study. Lancet. 2021;398(10301):665-674. 3. Sophia C. Weinmann and David S. Pisetsky. Mechanisms of immune-related adverse events during the treatment of cancer with immune checkpoint inhibitors. Rheumatology (Oxford). 2019;58(Suppl 7):vii59-vii67. 4. Teng GG, Chatham WW. Vasculitis related to viral and other microbial agents. Best Pract Res Clin Rheumatol. 2015;29(2):226-243. 5. Jin X, Ding C. Belimumab–an anti-BLyS human monoclonal antibody for rheumatoid arthritis. Expert Opin Biol Ther. 2013;13(2):315-322. 6. De Bandt M. Anti-TNF-alpha-induced lupus. Arthritis Res Ther. 2019;21(1):235. 7. Khattar P, Pichardo J, Jungbluth A, et al. B-cell maturation antigen is exclusively expressed in a wide range of B-cell and plasma cell neoplasm and in a potential therapeutic target for Bcma directed therapies. Blood. 2017;130(Suppl 1):S2755. 8. Jacob CO, Yu N, Sindhava V, et al. Differential development of systemic lupus erythematosus in NZM 2328 mice deficient in discrete pairs of BAFF receptors. Arthritis Rheumatol. 2015;67(9):2523-2535.

9. Jackson SW, Davidson A. BAFF inhibition in SLE-Is tolerance restored? Immunol Rev. 2019;292(1):102-119. 10. Fang Wei, Yan Chang, Wei Wei. The role of BAFF in the progression of rheumatoid arthritis. Cytokine. 2015;76(2):537-544. 11. Alesaeidi S, Darvishi M, Dabiri S, et al. Current understanding and unknown aspects of the treatment of granulomatosis with polyangiitis (Wegener's granulomatosis): opportunities for future studies. Curr Rheumatol Rev. 2020;16(4):257-266. 12. Merrill JT, Wallace DJ, Wax S, et al. Efficacy and safety of atacicept in patients with systemic lupus erythematosus: results of a twenty-four-week, multicenter, randomized, double-blind, placebo-controlled, parallel-arm, phase IIb study. Arthritis Rheumatol. 2018;70(2):266-276. 13. Shaker OG, Tawfic SO, El-Tawdy AM et al. Expression of TNF-α, APRIL and BCMA in Behcet's disease. J Immunol Res. 2014;2014:380405. 14. Coquery CM, Loo WM, Wade NS, et al. BAFF regulates follicular helper t cells and affects their accumulation and interferon-γ production in autoimmunity. Arthritis Rheumatol. 2015;67(3):773-784. 15. Jiang C, Loo WM, Greenley EJ, et al. B cell maturation antigen deficiency exacerbates lymphoproliferation and autoimmunity in murine lupus. J Immunol. 2011;186(11):6136-6147. 16. Gao Z, Feng Y, Xu J, Liang J. T-cell exhaustion in immunemediated inflammatory diseases: new implications for immunotherapy. Front Immunol. 2022;13:977394.

Haematologica | 109 January 2024

361

COMMENT

CD49d in chronic lymphocytic leukemia: a molecule with multiple regulation layers. Comment to “Sialylation regulates migration in chronic lymphocytic leukemia” We read with interest the study by Natoni et al. on the regulation of migration in chronic lymphocytic leukemia (CLL) cells by sialylation, recently published in Haematologica.1 In the study, the authors investigate the sialylation status of CLL cells, evaluating specific expression of α2-3 and α2-6 sialic acids, demonstrating that this particular modification is highly represented on CLL cells.1 Furthermore, they show that sialylation can affect the migratory capacity of CLL cells toward VCAM-1 and fibronectin (FN), which, in turn, is effectively suppressed upon treatment with neuraminidase (a glycoside hydrolase specific for sialic acids). Notably, VCAM-1 and FN are two known cognate ligands for the CD49d integrin, which is identified in this work as highly sialylated. Sialylation, the addition of negatively-charged sialic acids onto N-linked and O-linked protein glycans is a frequent post-translational modification for cell membrane proteins involved in cell-cell and cell-matrix interactions. It occurs within the Golgi network during protein maturation and is regulated by a wide variety of substrate-specific sialyltransferases, which are frequently found dysregulated in cancer.2 In the B-cell context, known sugar modifications include mannosylation of the B-cell receptor (BCR)-influencing antigen engagement and signaling,3 and, as reported by Natoni’s group, sialylation of the CD49d integrin in multiple myeloma4 and in CLL.1 The latter observation is particularly relevant, since CD49d represents, along with mutations of IGHV genes, TP53 and NOTCH1, one of the major biological prognosticators in CLL, and patients affected by a CD49d-expressing CLL (about 40% of all cases) usually experience a worse clinical outcome.5 Functionally, CD49d is the α4 chain of the α4/β1 complex (CD49d/CD29) composing the very late antigen-4 (VLA-4) integrin. VLA-4 operates as a cell-cell and cell-matrix receptor by interacting with its specific ligands: VCAM-1 involved in cell-cell interactions, and FN and EMILIN1 both involved in cell-matrix interactions. All three ligands are present in CLLinvolved tissues and are available for VLA-4 engagement.6 As one of the major actors regulating CLL microenvironmental interactions, CD49d expression and function are regulated at multiple levels, both genetic and epigenetic, as well as through protein-protein cooperation and/or conformational changes. In this complex scenario, sialylation, as described by Natoni and colleagues,1 may represent the last regulation identified of the many layers of biochemical modifications and/or physical interactions of CD49d with other molecules, which can eventually contribute to fine-tune CD49d-me-

diated activities such as migration and adhesion. Along this line, particularly illustrative is the case of trisomy 12 CLL. This subset is characterized by peculiar features of microenvironmental dependency, lymph node involvement and frequent Richter transformation, which includes NOTCH1 mutations, expression of (BCR) with stereotyped features (subset 8) and elevated expression of several integrins.7 Among these, CD49d stands out as almost universally expressed in this trisomy 12 CLL, mainly due to a methylationdependent epigenetic regulation mechanism of the ITGA4 gene (encoding for CD49d).7 As another layer of expression control, upregulation of CD49d expression in CLL also occurs outside the trisomy 12 subset, involving NF-κB pathway activation via NOTCH1 mutations. In fact, NOTCH1 mutation or in vitro activation of the NOTCH1 pathway leads to nuclear translocation of the NF-κB RelA subunit and upregulation of CD49d levels, whereas inhibition of the NF-κB pathway results in a downregulation of the integrin expression.8 A functional layer of regulation of the activity of the expressed CD49d molecule may come from its interaction with CD38, another microenvironmental receptor known to be expressed by CLL cells especially in cases with a bad prognosis.9 CD38, often found simultaneously expressed with CD49d, has a propensity to form with CD49d supramolecular complexes on the cell surface that potentiate CD49d-mediated adhesion and spreading.10 CD49d and CD38 are also interconnected by a sort of “humoral” link, as CD38 engagement by its counter receptor CD31 results in the production by CLL cells of two chemokines (CCL3 and CCL4) which in turn recruit monocyte/macrophages that, by secreting TNFα, contribute to upregulate the stromal/endothelial expression of VCAM-1, thus favoring CD49d/VCAM-1 adhesion and CLL cell survival.11 Another known layer regulating CD49d activities is represented by the activation of VLA-4 via “inside-out” through BCR stimulation, a functional mechanism present in CLL cells as a remnant of an identical mechanism occurring in secondary lymphoid organs during the antigen-dependent maturation process of normal B lymphocytes.12 In this context, stimuli originating from a fit BCR/antigen interaction can activate VLA-4 to promote cell adhesion to VCAM-1-expressing follicular dendritic cells, such a tonic interaction providing rescue of B cells from apoptosis and promote their differentiation.12 The interaction between BCR and VLA-4 in CLL could be relevant, given the increasing use of BCR signaling inhibitors

Haematologica | 109 January 2024

362

COMMENT such as ibrutinib or acalabrutinib for most CLL patients.6 These drugs have the common feature of favoring lymph node shrinkage and the release of lymphocytes in the blood stream, allegedly by hampering integrin-dependent adhesion mechanisms. However, CD49d-expressing CLL show reduced redistribution lymphocytosis and nodal shrinkage, a clinical manifestation of the residual capability of VLA-4 integrin to be activated at high affinity via BCR-mediated inside-out mechanism, thus promoting CLL cell retention.6 This also becomes evident by the higher expression of CD49d within the bone marrow and proliferative compartments, as defined by the relative expression of CD5 and CXCR4 markers.13 These (epi)genetic- and protein-regulated mechanisms, tightly intertwine with the recently described sialylation of CD49d in CLL cells, as reported by Natoni et al.1 which provides a novel, post-translational layer of regulation of integrin dynamics in CLL. Integrin sialylation has been described in a variety of tumors, and associated with increased migration potential, metastasis, immune evasion and cell survival. As an example, sialylation of the VLA-5 integrin (α5/β1) preserved colon adenocarcinoma cells from anoikis (apoptosis in response to lack of external stimuli).2 Although most of the research on sialylated integrins has been conducted within solid tumors, the studies by Natoni et al. point also at a significant role for this modification in hematological malignancies, with the reports of CD49d sialylation in CLL and in multiple myeloma.1,4 Much research is still needed to fully understand the functional dynamics of integrins in CLL, where VLA-4 expression represents one of the major actors of microenvironmental

interactions, thus mediating proliferation, survival and drug resistance.

Authors Federico Pozzo, Erika Tissino, Antonella Zucchetto and Valter Gattei Clinical and Experimental Onco-Hematology Unit, Centro di Riferimento Oncologico di Aviano (CRO) IRCCS, Aviano, Italy Correspondence: F. POZZO - federico.pozzo@cro.it https://doi.org/10.3324/haematol.2023.283237 Received: March 29, 2023. Accepted: May 10, 2023. Early view: May 18, 2023. ©2023 Ferrata Storti Foundation Published under a CC BY-NC license Disclosures No conflicts of interest to disclose. Contributions FP and VG wrote and revised the manuscript. ET and AZ contributed to writing and revising the manuscript.

References 1. Natoni A, Cerreto M, De Propris MS, et al. Sialylation regulates migration in chronic lymphocytic leukemia. Haematologica. 2023;108(7):1851-1860. 2. Seales EC, Jurado GA, Brunson BA, Wakefield JK, Frost AR, Bellis SL. Hypersialylation of beta1 integrins, observed in colon adenocarcinoma, may contribute to cancer progression by upregulating cell motility. Cancer Res. 2005;65(11):4645-4652. 3. Krysov S, Potter KN, Mockridge CI, et al. Surface IgM of CLL cells displays unusual glycans indicative of engagement of antigen in vivo. Blood. 2010;115(21):4198-4205. 4. Natoni A, Farrell ML, Harris S, et al. Sialyltransferase inhibition leads to inhibition of tumor cell interactions with E-selectin, VCAM1, and MADCAM1, and improves survival in a human multiple myeloma mouse model. Haematologica. 2020;105(2):457-467. 5. Dal Bo M, Bulian P, Bomben R, et al. CD49d prevails over the novel recurrent mutations as independent prognosticator of overall survival in chronic lymphocytic leukemia. Leukemia. 2016;30(10):2011-2018. 6. Tissino E, Benedetti D, Herman SEM, et al. Functional and clinical relevance of VLA-4 (CD49d/CD29) in ibrutinib-treated chronic lymphocytic leukemia. J Exp Med. 2018;215(2):681-697. 7. Zucchetto A, Caldana C, Benedetti D, et al. CD49d is overexpressed by trisomy 12 chronic lymphocytic leukemia cells: evidence for a

methylation-dependent regulation mechanism. Blood. 2013;122(19):3317-3321. 8. Benedetti D, Tissino E, Pozzo F, et al. NOTCH1 mutations are associated with high CD49d expression in chronic lymphocytic leukemia: link between the NOTCH1 and the NF-kappaB pathways. Leukemia. 2018;32(3):654-662. 9. Malavasi F, Deaglio S, Damle R, Cutrona G, Ferrarini M, Chiorazzi N. CD38 and chronic lymphocytic leukemia: a decade later. Blood. 2011;118(13):3470-3478. 10. Zucchetto A, Vaisitti T, Benedetti D, et al. The CD49d/CD29 complex is physically and functionally associated with CD38 in B-cell chronic lymphocytic leukemia cells. Leukemia. 2012;26(6):1301-1312. 11. Zucchetto A, Benedetti D, Tripodo C, et al. CD38/CD31, the CCL3 and CCL4 chemokines, and CD49d/vascular cell adhesion molecule-1 are interchained by sequential events sustaining chronic lymphocytic leukemia cell survival. Cancer Res. 2009;69(9):4001-4009. 12. Spaargaren M, Beuling EA, Rurup ML, et al. The B cell antigen receptor controls integrin activity through Btk and PLCgamma2. J Exp Med. 2003;198(10):1539-1550. 13. Tissino E, Pozzo F, Benedetti D, et al. CD49d promotes disease progression in chronic lymphocytic leukemia: new insights from CD49d bimodal expression. Blood. 2020;135(15):1244-1254.

Haematologica | 109 January 2024

363

ERRATUM

Erratum to: Comparing the two leading erythroid lines BEL-A and HUDEP-2 Deborah E. Daniels,1,2 Damien J. Downes,3 Ivan Ferrer-Vicens,1 Daniel C. J. Ferguson,1 Belinda K. Singleton,2,4 Marieangela C. Wilson,1 Kongtana Trakarnsanga,5 Ryo Kurita,6 Yukio Nakamura,7 David J. Anstee2,4 and Jan Frayne1,2 1

School of Biochemistry, University of Bristol, Bristol, UK; 2NIHR Blood and Transplant Research Unit, University of Bristol, Bristol, UK; 3MRC Molecular Haematology Unit, MRC Weatherall Institute of Molecular Medicine, Radcliffe Department of Medicine, University of Oxford, Oxford, UK; 4Bristol Institute for Transfusion Sciences, National Health Service Blood and Transplant (NHSBT), Bristol, UK; 5Department of Biochemistry, Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand; 6Department of Research and Development, Central Blood Institute, Blood Service Headquarters, Japanese Red Cross Society, Tokyo, Japan and 7Cell Engineering Division, RIKEN BioResource Research Center, Ibaraki, Japan

Correspondence: J. Frayne jan.frayne@bristol.ac.uk Received: Accepted:

October 19, 2023. October 19, 2023.

With reference to our paper published in the August issue 2020 of Haematologica,1 during the final processing of the manuscript the labeling of Figure 2E got accidentally inverted. On both the BEL-A and HUDEP-2 graphs, γ and β globin labels are the wrong way around. The corrected panel E of Figure 2 is shown here. A

Figure 2. Globin subunit expression by BEL-A and HUDEP-2. (A). Western blots of lysates obtained from late stage (day 10) HUDEP-2 in EDM media, HUDEP-2 and BEL-A in Bristol media and adult peripheral blood (PB; day 17) erythroid cells incubated with antibodies to α-, β- and γ-globin (sc-514378; sc-21757; sc-21756; all Santa Cruz). (B) Analysis of α-, β- and γ-globin mRNA levels in BEL-A and HUDEP-2 in Bristol media on day 6 of culture by quantitative PCR (qPCR). Target gene expression was normalized to PABPC1 reference gene (using the 2-ΔΔCt method) and compared to a BEL-A control sample. Results are means ± standard deviation (SD), N=4. TaqMan gene expression assays used were: HBA (Hs00361191_g1), HBB (Hs00747223_g1), HBG (Hs00361131_g1) and PABPC1 (Hs00743792_s1) (Fisher Scientific). (C) H3K27ac chromatin immunoprecipitation (ChIP) was carried out in 5x106 day 4 cells by fixing in 1% formaldehyde and performing immunoprecipitation (IP) with the ChIP Assay Kit (Merck Millipore) using 0.3 μg anti-H3K27ac (ab4729; Abcam). Enrichment at the HBG1/2 promoters (i) and control regions (ii) was determined using qPCR by comparing to a background control (hg19, chr16:6,002,736-6,002,834). Bars depict mean and SD for IP from 4 separate differentiations (circles). *Mann-Whitney rank sum P=0.0286. (D-E) Response of BEL-A and HUDEP-2 cells to hydroxyurea treatment. Hydroxyurea (H8627; Sigma) was added to cells at 50 μM every 2 days from four days prior to the start of differentiation until harvesting of cells on day 10 of differentiation. (D) Representative Western blot of cell lysates from ± hydroxyurea (HU) cultures incubated with antibodies to α-, β- and γ-globin and (E) densitometric analysis of the bands obtained from at least 3 independent experiments normalized to α-globin expression (means ± SD). *P<0.05 two-tailed Student t test. (F-G) Expression of BCL11A, GATA1 and KLF1 by BEL-A and HUDEP-2. β-actin was used as a protein loading control. (F) Representative western blots of lysates obtained from cells at day 0, 2, 4 and 6 of BEL-A and HUDEP-2 cultures incubated with BCL11A (ab19487; Abcam), GATA1 (sc-266; Santa Cruz), KLF1 (sc-14034; Santa Cruz) and β-actin (A1978; Sigma) antibodies, and (G) densitometric analysis of the bands obtained from 2 independent experiments normalized to β-actin expression and then compared to BEL-A day 0 (means ± SD).

References 1. Daniels DE, Downes DJ, Ferrer-Vicens I, et al. Comparing the two leading erythroid lines BEL-A and HUDEP-2. Haematologica. 2020:105(8):e389-e394. Haematologica | 109 - January 2024

364

ERRATUM

Erratum to: DUSP22-rearranged ALK-negative anaplastic large cell lymphoma is a pathogenetically distinct disease but can have variable clinical outcome Kerry J. Savage1,2 and Graham W. Slack1,3

Correspondence: K.J. Savage ksavage@bccancer.bc.ca 2

Centre for Lymphoid Cancer, BC Cancer; University of British Columbia and Division of Medical Oncology, BC Cancer and 3University of British Columbia and Department of Pathology, BC Cancer, Vancouver, Canada 1

Received: Accepted:

September 13, 2023. September 13, 2023.

We made non-substantive typographical corrections and the following substantive changes to the original figure of our editorial “DUSP22-rearranged ALK-negative anaplastic large cell lymphoma is a pathogenetically distinct disease but can have variable clinical outcome”1: WHO 2nd edition 2001 is changed to WHO 3rd edition 2001; WHO 3rd edition 2008 is changed to WHO 4th edition 2008; and WHO 4th edition 2016 is changed to WHO Revised 4th edition 2016.

Figure 1. Current classification of systemic anaplastic large cell lymphoma. The most common t(2;5) (ALK-NPM) is shown. Rare variant rearrangements involving the ALK gene on 2p23 and different partner genes on other chromosomes can occur in 15-25% of cases of ALK-positive anaplastic large cell lymphoma. ALCL: anaplastic large cell lymphoma; WHO: World Health Organization; ALK: anaplastic lymphoma kinase.

References 1. Savage KJ, Slack GW. DUSP22-rearranged ALK-negative anaplastic large cell lymphoma is a pathogenetically distinct disease but can have variable clinical outcome. Haematologica. 2023;108(6):1463-1467. Haematologica | 109 - January 2024

365

Haematologica | 109 January 2024

366