Haematologica. Volume 107, Issue 4 by Haematologica

haematologica Journal of the Ferrata Storti Foundation

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

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

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

Statistical Consultant Catherine Klersy (Pavia)

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

Managing Director Antonio Majocchi (Pavia)

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

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

haematologica Journal of the Ferrata Storti Foundation

Brief information on Haematologica Haematologica (print edition, pISSN 0390-6078, eISSN 1592-8721) publishes peer-reviewed papers on all areas of experimental and clinical hematology. The journal is owned by a non-profit organization, the Ferrata Storti Foundation, and serves the scientific community following the recommendations of the World Association of Medical Editors (www.wame.org) and the International Committee of Medical Journal Editors (www.icmje.org). Haematologica publishes Editorials, Original articles, Review articles, Perspective articles, Editorials, Guideline articles, Letters to the Editor, Case reports & Case series and Comments. Manuscripts should be prepared according to our guidelines (www.haematologica.org/information-for-authors), and the Uniform Requirements for Manuscripts Submitted to Biomedical Journals, prepared by the International Committee of Medical Journal Editors (www.icmje.org). Manuscripts should be submitted online at http://www.haematologica.org/. Conflict of interests. According to the International Committee of Medical Journal Editors (http://www.icmje.org/#conflicts), “Public trust in the peer review process and the credibility of published articles depend in part on how well conflict of interest is handled during writing, peer review, and editorial decision making”. The ad hoc journal’s policy is reported in detail at www.haematologica.org/content/policies. Transfer of Copyright and Permission to Reproduce Parts of Published Papers. Authors will grant copyright of their articles to the Ferrata Storti Foundation. No formal permission will be required to reproduce parts (tables or illustrations) of published papers, provided the source is quoted appropriately and reproduction has no commercial intent. Reproductions with commercial intent will require written permission and payment of royalties. Subscription. Detailed information about subscriptions is available at www.haematologica.org. Haematologica is an open access journal and access to the online journal is free. For subscriptions to the printed issue of the journal, please contact: Haematologica Office, via Giuseppe Belli 4, 27100 Pavia, Italy (phone +39.0382.27129, fax +39.0382.394705, E-mail: info@haematologica.org). Rates of the printed edition for the year 2022 are as following: Institutional: Euro 700 Personal: Euro 170 Advertisements. Contact the Advertising Manager, Haematologica Office, via Giuseppe Belli 4, 27100 Pavia, Italy (phone +39.0382.27129, fax +39.0382.394705, e-mail: marketing@haematologica.org). Disclaimer. Whilst every effort is made by the publishers and the editorial board to see that no inaccurate or misleading data, opinion or statement appears in this journal, they wish to make it clear that the data and opinions appearing in the articles or advertisements herein are the responsibility of the contributor or advisor concerned. Accordingly, the publisher, the editorial board and their respective employees, officers and agents accept no liability whatsoever for the consequences of any inaccurate or misleading data, opinion or statement. Whilst all due care is taken to ensure that drug doses and other quantities are presented accurately, readers are advised that new methods and techniques involving drug usage, and described within this journal, should only be followed in conjunction with the drug manufacturer’s own published literature.

Direttore responsabile: Prof. Carlo Balduini; Autorizzazione del Tribunale di Pavia n. 63 del 5 marzo 1955. Printing: Press Up, zona Via Cassia Km 36, 300 Zona Ind.le Settevene - 01036 Nepi (VT)

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

haematologica Journal of the Ferrata Storti Foundation

Table of Contents Volume 107, Issue 4: April 2022 About the Cover 781

Images from the Haematologica Atlas of Hematologic Cytology: precursor lymphoid neoplasms, cytochemistry and immunocytochemistry Rosangela Invernizzi

https://doi.org/10.3324/haematol.2022.280759

Landmark Paper in Hematology 782

The first achievement of complete remission in childhood leukemia by treatment with the folic acid antagonist aminopterin Shai Izraeli

https://doi.org/10.3324/haematol.2022.280670

Editorials 783

FLT3-ITD signals bad news for core binding factor acute myeloid leukemia unless trisomy 22 comes to the rescue Sun Loo and Andrew H. Wei

https://doi.org/10.3324/haematol.2021.279409

785

Thrombotic thrombocytopenic purpura and other immune-mediated blood disorders following vaccination against SARS-CoV-2 Pier Mannuccio Mannucci

https://doi.org/10.3324/haematol.2021.279649

787

Not all mismatches are equal: importance of alloreactivity direction Jacinta Perram and Nada Hamad

https://doi.org/10.3324/haematol.2021.279587

Review Article 790

The mitochondrial anti-apoptotic dependencies of hematologic malignancies: from disease biology to advances in precision medicine Isacco Ferrarini et al.

https://doi.org/10.3324/haematol.2021.280201

Articles Acute Lymphoblastic Leukemia 803 LAMP-5 is an essential inflammatory-signaling regulator and novel immunotherapy target for mixed lineage leukemia-rearranged acute leukemia Gabriel Gracia-Maldonado et al.

https://doi.org/10.3324/haematol.2020.257451

Acute Myeloid Leukemia 816 Interleukin 4 promotes phagocytosis of murine leukemia cells counteracted by CD47 upregulation Pablo Peña-Martínez et al. https://doi.org/10.3324/haematol.2020.270421

825

Pevonedistat and azacitidine upregulate NOXA (PMAIP1) to increase sensitivity to venetoclax in preclinical models of acute myeloid leukemia Dan Cojocari et al. https://doi.org/10.3324/haematol.2020.272609

836

Characteristics and outcome of patients with core-binding factor acute myeloid leukemia and FLT3-ITD: results from an international collaborative study Sabine Kayser et al. https://doi.org/10.3324/haematol.2021.278645

Haematologica 2022; vol. 107 no. 4 - April 2022 http://www.haematologica.org/

haematologica Journal of the Ferrata Storti Foundation

Bone Marrow Transplantation 844 Refined HLA-DPB1 mismatch with molecular algorithms predicts outcomes in hematopoietic stem cell transplantation Jun Zou et al. https://doi.org/10.3324/haematol.2021.278993

Cell Therapy and Immunotherapy 857 Comparison of immune reconstitution between anti-T-lymphocyte globulin and posttransplant cyclophosphamide as acute graft-versus-host disease prophylaxis in allogeneic myeloablative peripheral blood stem cell transplantation Radwan Massoud et al. https://doi.org/10.3324/haematol.2020.271445

Chronic Lymphocytic Leukemia 868 The complex karyotype landscape in chronic lymphocytic leukemia allows the refinement of the risk of Richter syndrome transformation Andrea Visentin et al. https://doi.org/10.3324/haematol.2021.278304

877

IGHV-associated methylation signatures more accurately predict clinical outcomes of chronic lymphocytic leukemia patients than IGHV mutation load Dianna Hussmann et al. https://doi.org/10.3324/haematol.2021.278477

Hematopoiesis 887 Perturbed hematopoiesis in individuals with germline DNMT3A overgrowth Tatton-Brown-Rahman syndrome Ayala Tovy et al. https://doi.org/10.3324/haematol.2021.278990

Hodgkin Lymphoma 899 Improved outcomes of high-risk relapsed Hodgkin lymphoma patients after high-dose chemotherapy: a 15-year analysis Yago Nieto et al. https://doi.org/10.3324/haematol.2021.278311

909

Inhibitors of ADAM10 reduce Hodgkin lymphoma cell growth in 3D microenvironments and enhance brentuximab-vedotin effect Roberta Pece et al. https://doi.org/10.3324/haematol.2021.278469

Plasma Cell Disorders 921 DIS3 mutations in multiple myeloma impact the transcriptional signature and clinical outcome Katia Todoerti et al. https://doi.org/10.3324/haematol.2021.278342

Platelet Biology & its Disorders 933 The GPIbα intracellular tail - role in transducing VWF- and collagen/GPVI-mediated signaling Adela Constantinescu-Bercu et al. https://doi.org/10.3324/haematol.2020.278242

947

Comparative analysis of ChAdOx1 nCoV-19 and Ad26.COV2.S SARS-CoV-2 vector vaccines Stephan Michalik et al. https://doi.org/10.3324/haematol.2021.280154

Red Cell Biology & its Disorders 958 Brain injury pathophysiology study by a multimodal approach in children with sickle cell anemia with no intra or extra cranial arteriopathy Valentine Brousse et al. https://doi.org/10.3324/haematol.2020.278226

Haematologica 2022; vol. 107 no. 4 - April 2022 http://www.haematologica.org/

haematologica Journal of the Ferrata Storti Foundation

Letters to the Editor 966

Loss of 5-hydroxymethylcytosine expression is near-universal in B-cell lymphomas with variable mutations in epigenetic regulators Kevin S. Tanager et al.

https://doi.org/10.3324/haematol.2021.279648

970

https://doi.org/10.3324/haematol.2021.279677

975

Early testicular maturation is sensitive to depletion of spermatogonial pool in sickle cell disease Klara M. Benninghoven-Frey et al.

https://doi.org/10.3324/haematol.2021.279253

980

Targeting B-cell maturation antigen increases sensitivity of multiple myeloma cells to MCL-1 inhibition Marta Cuenca et al.

https://doi.org/10.3324/haematol.2021.279517

984

Phase Ib dose-escalation study of the selective, non-covalent, reversible Bruton’s tyrosine kinase inhibitor vecabrutinib in B-cell malignancies John N. Allan et al.

https://doi.org/10.3324/haematol.2021.280061

988

In vitro and in vivo effects of short-term cold storage of platelets in PAS-C S. Lawrence Bailey et al.

https://doi.org/10.3324/haematol.2021.279865

991

Conventional interferon-α 2b versus hydroxyurea for newly-diagnosed patients with polycythemia vera in a real world setting: a retrospective study based on 286 patients from a single center Dan Liu et al.

https://doi.org/10.3324/haematol.2021.280080

996

Daratumumab with or without chemotherapy in relapsed and refractory acute lymphoblastic leukemia. A retrospective observational Campus ALL study Marco Cerrano et al.

https://doi.org/10.3324/haematol.2021.279851

1000

https://doi.org/10.3324/haematol.2021.280300

1004

Acute lymphoblastic leukemia cells are able to infiltrate the brain subventricular zone stem cell niche and impair neurogenesis Lidia M. Fernández-Sevilla et al.

https://doi.org/10.3324/haematol.2021.279383

Case Reports 1008

Immune-mediated thrombotic thrombocytopenic purpura following administration of Pfizer-BioNTech COVID-19 vaccine Gaetano Giuffrida et al.

https://doi.org/10.3324/haematol.2021.279535

1011

VEXAS syndrome in a female patient with constitutional 45,X (Turner syndrome) Ryan J. Stubbins et al.

https://doi.org/10.3324/haematol.2021.280238

1014

A case series of primary cutaneous B-cell lymphomas with atypical presentations: diagnostic and therapeutic challenges Emily Correia et al.

https://doi.org/10.3324/haematol.2021.279992

Haematologica 2022; vol. 107 no. 4 - April 2022 http://www.haematologica.org/

haematologica Journal of the Ferrata Storti Foundation

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

Ancient Greek

Scientific Latin

Modern English

haematologicus (adjective) = related to blood haematologica (adjective, plural and neuter, used as a noun) = hematological subjects The oldest hematology journal, publishing the newest research results. 2020 JCR impact factor = 9.94

ABOUT THE COVER Images from the Haematologica Atlas of Hematologic Cytology: precursor lymphoid neoplasms, cytochemistry and immunocytochemistry Rosangela Invernizzi University of Pavia, Pavia, Italy E-mail: ROSANGELA INVERNIZZI - rosangela.invernizzi@unipv.it doi:10.3324/haematol.2022.280759

n the absence of signs of morphological differentiation, the lineage of acute leukemia blasts can be assessed by cytochemistry, flow-cytometry or immunocytochemistry. The cytochemical features of acute lymphoblastic leukemia are shown in the Figure. Most lymphoblasts reveal strong periodic acid Schiff (PAS) staining with granules arranged in perinuclear rings (A) or large, sometimes single, cytoplasmic blocks of glycogen (B). Differently from myeloblasts, lymphoblasts are negative for the peroxidase (C, left) and Sudan black (C, right) reactions; note also the strong cytoplasmic positivity of neutrophils for both reactions in panel (C). T-lymphoblasts, differently from B-lymphoblasts, are characterized by focal acid phosphatase reactivity, due to the enzyme localization in the Golgi zone (D), and also by strong, localized, paranuclear dipeptidyl aminopeptidase IV (DAP IV) activity (E). Nuclear terminal deoxynucleotidyl transferase (TdT) may be detected by immunocytochemistry in both B-lineage (F) and T-lineage lymphoblasts.1 Disclosures No conflicts of interest to disclose.

Reference 1. Invernizzi R. Precursor lymphoid neoplasms. Haematologica. 2020;105(Suppl. 1):127-138.

haematologica | 2022; 107(4)

781

LANDMARK PAPER IN HEMATOLOGY The first achievement of complete remission in childhood leukemia by treatment with the folic acid antagonist aminopterin Shai Izraeli Schneider Children’s Medical Center of Israel, Tel Aviv University, Tel Aviv, Israel E-mail: sizraeli@gmail.com doi:10.3324/haematol.2022.280670 TITLE

Temporary remissions in acute leukemia in children produced by folic acid antagonist, 4-aminopteroyl-glutamic acid (aminopterin).

AUTHORS

Farber S, Diamond LK. Mercer RD, Slyvester RF Jr, Wolff JA.

JOURNAL

New England Journal of Medicine. 1948;238(23):787-793.

hese days almost 90% of children with acute lymphoblastic leukemia and 70% of those with acute myelogenous leukemia are cured. The first significant step towards these results was published by Dr. Sidney Farber and colleagues 74 years ago.1 Their achievement was preceded by a devastating failure. Following the great success of treatment of folate deficiency by conjugates of folic acid, and given the morphological similarity between megaloblastic anemia and leukemia, Dr. Farber attempted to treat children with leukemia with folate conjugates. Remarkably, he was perceptive enough to note that treatment with folates had the opposite outcome from that desired: it markedly accelerated the leukemias, as he observed both clinically and in postmortem examinations.2 These observations led him to conduct a clinical trial with the newly synthetised anti-folate, amiopterin;3 the results were published on June 3, 1948 in the New England Journal of Medicine.1 Sixteen children with leukemia were treated with aminopterin, ten entered transient remissions characterized by a reduction or even complete disappearance of blasts from peripheral blood and bone marrow with recovery of normal hematopoiesis (Figure 1, from the original article). These hematologic findings were accompanied by resolution of the clinical symptoms, regression of hepatosplenomegaly and, in

one case, regression of subcutaneous nodules that were presumed to be leukemic. The major toxicity was “severe stomatitis” and, in one case, pancytopenia and empty bone marrow, for which therapy with liver extracts was attempted. The longest complete remission off therapy was 47 days. Fast-forwarding to our times, methotrexate is one of the cornerstones of treatment of acute lymphoblastic leukemia. Severe mucositis is indeed one of the major toxicities and folate, in the form of leucovorin, is the main rescue treatment after therapy with high-dose methotrexate. The remarkable observation by Dr. Farber that folate conjugates accelerate leukemia growth serves as an important reminder that vitamins and nutritional supplements, given with the intention to “strengthen the patient”, may fuel a tumor. Cancer and normal cells compete for the same resources with the former being more dependent on the essential fuel. Treatment with L-asparaginase followed the introduction of anti-folates as an effective means to starve tumor cells. Interestingly, controlled calorie restriction has recently been shown to further improve the rate of molecular remission of children with acute lymphoblastic leukemia and is now being tested in a large clinical trial funded by the National Institutes of Health.4 Figure 1. A reproduction of the original figure from the paper in the New England Journal of Medicine1 demonstrating the morphology of the bone marrow before (A) and after 1 month of treatment with aminopterin and crude liver extract (B).

References 1. Farber S, Diamond LK. Mercer RD, Slyvester RF Jr, Wolff JA.Temporary remissions in acute leukemia in children produced by folic acid antagonist, 4aminopteroyl-glutamic acid (aminopterin). N Engl J Med. 1948;238(23):787-793. 2. Farber S, Cutler EC, Hawkins JW, Harrison JH, Peirce EC 2nd, Lenz GG. The action of pteroylglutamic conjugates on man. Science. 1947;106(2764):619621. 3. Seeger DR, Smith JM Jr, Hultquist ME. Antagonist for pteroylglutamic acid. J Am Chem Soc. 1947;69(10):2567. 4. Orgel E, Framson C, Buxton R, et al. Caloric and nutrient restriction to augment chemotherapy efficacy for acute lymphoblastic leukemia: the IDEAL trial. Blood Adv. 2021;5(7):1853-1861.

782

haematologica | 2022; 107(4)

EDITORIALS FLT3-ITD signals bad news for core binding factor acute myeloid leukemia unless trisomy 22 comes to the rescue Sun Loo1 and Andrew H. Wei1,2 1

Department of Clinical Haematology, Alfred Hospital and 2Australian Centre for Blood Diseases, Monash University, Melbourne, Victoria, Australia E-mail: ANDREW H. WEI - andrew.wei@monash.edu doi:10.3324/haematol.2021.279409

tructural rearrangements resulting in either t(8;21)(q22;q22) [RUNX1-RUNX1T1] or inv(16)(p13q22)/t(16;16)(p13.1;q22) [CBFB-MYH11] are pathognomonic for core binding factor (CBF) acute myeloid leukemia (AML). Prognostic classifications have consistently positioned CBF AML as a favorable entity, particularly if the patient can tolerate conventional induction and consolidation chemotherapy. Optimal outcomes for patients with CBF disease are achieved through incorporation of gemtuzumab ozogamicin into 7+3 based induction and high-dose cytarabine into the consolidation phase of therapy.1,2 Recent molecular studies have highlighted striking differences in the genomic landscape between the two forms of CBF AML. Although kinase activating mutations are observed frequently in both groups, RUNX1-RUNX1T1 more commonly harbors mutations in ASXL1 (14%), ASXL2 (14%), TET2 (11%), RAD21 (11%) and ZBTB7A (19%), whereas CBFB-MYH11 AML is more frequently associated with WT1 mutation (10%). At the cytogenetic level, t(8;21) is more closely linked to del(9q) or loss of a sex chromosome, whereas inv(16) may occur in the company of del(7q) and trisomy 22 abnormalities.3-5 In terms of prognosis, although there is general agreement that additional cytogenetic abnormalities do not consistently increase the risk of relapse in CBF AML, the role of kinase activating mutations has been more controversial.6 The predominant kinase activating mutations in CBF AML involve RAS (27%), KIT (26%) and FLT3 (17%).5 The presence of mutant RAS is generally associated with a favorable prognosis in CBF AML.5 In contrast, several series suggest that KIT mutations, in particular exon 17 mutations, are associated with increased relapse risk among patients with RUNX1RUNX1T1, whereas prognostic concordance is lacking for CBFB-MYH11 AML.7,8 The paper published by Kayser and colleagues9 in this issue of Haematologica is a multi-institutional retrospective cohort analysis addressing the role of FLT3-internal tandem duplication (ITD) co-mutation in CBF AML. The study included 97 patients with similar proportions of t(8;21)(q22;q22) and inv(16)(p13q22)/t(16;16)(p13.1;q22). Most were treated intensively, resulting in a very high complete remission rate of 98%, despite the presence of FLT3-ITD, with only three patients receiving concomitant FLT3 inhibitor. Allogeneic hematopoietic cell transplant (HCT) was performed in 14% of the patient population in first complete remission. Among patients not transplanted in first complete remission, almost 40% relapsed with subsequent allogeneic HCT performed in ~39% of this group. In this analysis of patients with FLT3ITD CBF AML, the authors found that allogeneic HCT was only beneficial for patients at relapse, whereas outcomes were not improved by allogeneic HCT in first complete remission. If allogeneic HCT was not performed at relapse, there were no long-term survivors. Long-term survival was

haematologica | 2022; 107(4)

also absent for the small group of patients treated non-intensively. These results prompted the authors to conclude that patients with FLT3-ITD CBF AML should be given intensive induction and consolidation therapy, when possible, and to reserve allogeneic HCT as a strategy in second complete remission in the event of relapse after first-line therapy. A major caveat is the retrospective nature of the study, which introduces the risk of potential bias. Only 39% of relapsing patients were transplanted, suggesting that the opportunity for cure was lost for the majority of those in whom primary therapy failed. The failure to observe enhanced outcomes for those treated in first complete remission, however, suggests that not all patients with FLT3-ITD CBF AML have a poor prognosis and that heterogeneity in survival must exist. In search of genetic factors differentiating prognosis in CBF AML, Kayser et al. identified an association between inv(16) and trisomy 22 in 23% of cases. Although prior studies have already reported favorable outcome for this chromosomal duet,10 the current study extends this finding to patients with trisomy 22, inv(16) and FLT3-ITD mutation. For patients with this molecular triad, relapse-free survival at 4 years was 80%, compared to only 38% for other patients. The authors conclude that patients with CBF and FLT3-ITD with inv(16) and trisomy 22 should be classified as favorable risk, the remainder as poor risk. It remains uncertain, however, whether outcomes would be improved by upfront allogeneic HCT in first complete remission or whether transplant at relapse would suffice for this poor-risk CBF subgroup with FLT3-ITD. Another intriguing question is what candidate genes are carried on chromosome 22, which when amplified by just one copy, can result in dramatic enhancement of prognosis in patients with FLT3-ITD CBF AML. A major limitation of the study was the absence of flow or molecular measurable residual disease (MRD) correlation with these prognostic observations. Favorable prognosis in CBF AML is strengthened by multi-log reduction or eradication of MRD after commencing treatment. Despite an admirable effort to refine prognostic outcomes in FLT3-ITD CBF AML, a recurring question is whether the importance of baseline prognostic risk stratification is diminished by dynamic assessment of post-treatment MRD. Although current European LeukemiaNET guidance recommends posttreatment MRD monitoring every 3 months, several studies suggest that the window of opportunity to intervene between initial detection of MRD progression by reversetranscriptase quantitative polymerase chain reaction (RTqPCR ) and clinical relapse is too narrow, making it logistically difficult to orchestrate a meaningful therapeutic intervention.12-14 Increasing the intensity of MRD monitoring with more frequent peripheral blood surveillance e.g., monthly for the first 12 months when relapse risk is highest, could enable earlier detection of rising MRD. It remains to be proven whether overall survival would be enhanced by earlier, pre-

783

Editorials

emptive intervention, as opposed to salvage at the time of morphological progression. With allogeneic HCT in second complete remission the main priority for patients with relapsing disease, it is likely that early detection and treatment to suppress rising MRD could increase the proportion of patients bridged to transplant in remission and negative for MRD. Alternatively, it remains an open question whether outcomes will be improved by a pre-transplant MRD reduction strategy, or whether equivalent outcomes could be achieved by proceeding directly to transplantation, especially if myeloablative conditioning is planned. The median time to relapse from detection of MRD failure to clinical relapse is only about 3-4 months.12 Therefore, a pre-emptive MRD suppression strategy could buy the treating team more time, keeping the patient in remission and free from relapse until the allogeneic HCT can be organized and carried out. In terms of targeting FLT3 to improve clinical outcome in FLT3-ITD CBF AML, treatment could be introduced at the induction/consolidation stage, during maintenance, preemptively at the time of MRD progression, at morphological relapse, or as maintenance therapy in the post-allogeneic HCT setting. Unfortunately, robust data to answer any of these questions are lacking, with patients harboring FLT3-ITD CBF accounting for only ~2% of the AML population, making randomized trial data with any new or future agent or combinations within this orphan sub-population an unlikely prospect. The RATIFY trial, which examined the role of midostaurin during induction, consolidation and maintenance in patients with FLT3-mutant AML, only enrolled 16 patients (4%) with CBF AML to the midostaurin arm.15 In the SORAML trial, the FLT3 inhibitor sorafenib was combined with standard induction and consolidation therapy and as maintenance for 12 months.16 In the favorable cytogenetic risk group, which formed only 10% of the study population, sorafenib was associated with improved event-free, relapse-free and overall survival in a post-hoc subgroup analysis. The outcomes of patients with FLT3-ITD within this CBF subgroup were, however, not defined. In summary, as the genomic age continues to reveal further prognostic heterogeneity within conventional AML subgroups, we will increasingly be challenged with when to pull the trigger on the use of allogeneic HCT and when to use a growing number of newly approved AML drugs, such as FLT3 inhibitors and so forth, for uncommon clinical scenarios for which definitive randomized evidence may never become available. The current work by Kayser et al.9 adds to the growing list of AML scenarios in which the presence of FLT3-ITD represents bad news, including among patients with CBF AML. Physicians are likely to formulate a logic circuit that suggests that: (i) it makes sense to use an FLT3 inhibitor to target FLT3-ITD when detected in CBF AML; (ii) patients with concurrent trisomy 22 should not be candidates for allogeneic HCT in first complete remission; (iii) close monitoring of MRD, potentially with RT-qPCR performed monthly on blood for at least the first 12 months, is warranted; and (iv) allogeneic HCT should be ready to action early if MRD progression is confirmed.

784

Disclosures AW has served on advisory boards for Novartis, Janssen, Amgen, Roche, Pfizer, Abbvie, Servier, Celgene-BMS, Macrogenics, Agios, and Gilead; receives research funding to the Institution from Novartis, Abbvie, Servier, Celgene-BMS, Astra Zeneca, and Amgen; serves on speakers bureaus for Abbvie, Novartis, Celgene; and receives royalty payments from the Walter and Eliza Hall Institute of Medical Research related to venetoclax Contributions Both authors wrote and reviewed the paper.

References 1. Magina KN, Pregartner G, Zebisch A, et al. Cytarabine dose in the consolidation treatment of AML: a systematic review and metaanalysis. Blood. 2017;130(7):946-948. 2. Hills RK, Castaigne S, Appelbaum FR, et al. Addition of gemtuzumab ozogamicin to induction chemotherapy in adult patients with acute myeloid leukaemia: a meta-analysis of individual patient data from randomised controlled trials. Lancet Oncol. 2014;15(9):986996. 3. Opatz S, Bamopoulos SA, Metzeler KH, et al. The clinical mutatome of core binding factor leukemia. Leukemia. 2020;34(6):1553-1562. 4. Faber ZJ, Chen X, Gedman AL, et al. The genomic landscape of corebinding factor acute myeloid leukemias. Nat Genet. 2016;48(12): 1551-1556. 5. Jahn N, Terzer T, Strang E, et al. Genomic heterogeneity in corebinding factor acute myeloid leukemia and its clinical implication. Blood Adv. 2020;4(24):6342-6352. 6. Han SY, Mrozek K, Voutsinas J, et al. Secondary cytogenetic abnormalities in core-binding factor AML harboring inv(16) vs t(8;21). Blood Adv. 2021;5(10):2481-2489. 7. Ishikawa Y, Kawashima N, Atsuta Y, et al. Prospective evaluation of prognostic impact of KIT mutations on acute myeloid leukemia with RUNX1-RUNX1T1 and CBFB-MYH11. Blood Adv. 2020;4(1):66-75. 8. Paschka P, Marcucci G, Ruppert AS, et al. Adverse prognostic significance of KIT mutations in adult acute myeloid leukemia with inv(16) and t(8;21): a Cancer and Leukemia Group B study. J Clin Oncol. 2006;24(24):3904-3911. 9. Kayser S, Kramer M, Martínez-Cuadrón D, et al. Characteristics and outcome of patients with core binding factor acute myeloid leukemia and FLT3-ITD: results from an international collaborative study. Haematologica. 2022;107(4):836-843. 10. Marcucci G, Mrozek K, Ruppert AS, et al. Prognostic factors and outcome of core binding factor acute myeloid leukemia patients with t(8;21) differ from those of patients with inv(16): a Cancer and Leukemia Group B study. J Clin Oncol. 2005;23(24):5705-5717. 11. 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. 12. Willekens C, Blanchet O, Renneville A, et al. Prospective long-term minimal residual disease monitoring using RQ-PCR in RUNX1RUNX1T1-positive acute myeloid leukemia: results of the French CBF-2006 trial. Haematologica. 2016;101(3):328-335. 13. Yin JA, O'Brien MA, Hills RK, et al. Minimal residual disease monitoring by quantitative RT-PCR in core binding factor AML allows risk stratification and predicts relapse: results of the United Kingdom MRC AML-15 trial. Blood. 2012;120(14):2826-2835. 14. Puckrin R, Atenafu EG, Claudio JO, et al. Measurable residual disease monitoring provides insufficient lead-time to prevent morphologic relapse in the majority of patients with core-binding factor acute myeloid leukemia. Haematologica. 2021;106(1):56-63. 15. Stone RM, Mandrekar SJ, Sanford BL, et al. Midostaurin plus chemotherapy for acute myeloid leukemia with a FLT3 mutation. N Engl J Med. 2017;377(5):454-464. 16. Rollig C, Serve H, Huttmann A, et al. Addition of sorafenib versus placebo to standard therapy in patients aged 60 years or younger with newly diagnosed acute myeloid leukaemia (SORAML): a multicentre, phase 2, randomised controlled trial. Lancet Oncol. 2015;16(16):1691-1699.

haematologica | 2022; 107(4)

Editorials

Thrombotic thrombocytopenic purpura and other immune-mediated blood disorders following vaccination against SARS-CoV-2 Pier Mannuccio Mannucci Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico, Angelo Bianchi Bonomi Hemophilia and Thrombosis Center, Milan, Italy E-mail: PIER MANNUCCIO MANNUCCI - piermannuccio.mannucci@policlinico.mi.it doi:10.3324/haematol.2021.279649

n this issue of Haematologica, Giuffrida et al.1 report two cases of new-onset, immune-mediated thrombotic thrombocytopenic purpura (TTP) in 81-year-old and 30-year-old women diagnosed with this very rare disease 14 and 18 days after the first dose of the mRNA-based vaccine against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) manufactured by PfizerBioNTech. The older woman (case 1) had a history of diabetes and connective tissue disease positive for antinuclear antibodies, whereas the younger (case 2) was negative regarding clinical history and laboratory markers of potential triggers of TTP such as autoimmune disorders, tumors and infections. Both women were promptly treated with glucocorticoids and daily sessions of plasma exchange, each followed by the nanobody caplacizumab. This state-of-the-art therapeutic approach based upon plasma therapy, immunomodulation and anti-von Willebrand factor medicines was successful in the younger woman, who had rapid normalization of a very low platelet count, even though plasma ADAMTS13 was still unmeasurable on days 14 and 30 after eight plasma exchanges and anti-ADAMTS13 were still present. The older woman with comorbidities had only a modest improvement of platelet count and she died suddenly after the second plasma exchange as the result of an illdefined cardiac event, thus once again emphasizing that TTP is still associated with a significant mortality toll notwithstanding prompt and impeccable management. The main interest of these two cases lies in the fact that new-onset autoimmune TTP occurred within 2 to 3 weeks after the first dose of a vaccine to protect against coronavirus disease 2019 (COVID-19). Administration of the vaccine within this short time window prior to the TTP episode as well as no evidence for other causes (at least in the younger woman) are consistent with causality according to the World Health Organization criteria for post-vaccination adverse events.2 Until now, new-onset TTP had been reported as a single case after the Johnson

& Johnson vaccine, which is based on a human adenovirus vector,3 and a relapse of recurrent TTP which occurred 6 days after the second dose of the PfizerBioNTech vaccine.4 The new-onset cases described by Giuffrida et al.1 of such a rare immune-mediated blood disease associated with a bleeding tendency follow the report of a mRNA-vaccine (Pfizer-BioNTech)-associated case of autoimmune hemophilia due to anti-factor VIII antibodies5 and multiple cases of immune thrombocytopenic purpura (ITP) due to platelet autoantibodies occurring after either of the two mRNA-based vaccines produced by Pfizer and Moderna.6 Common features of these cases are that the majority of them occurred in women, at young but also at older ages, thus reproducing the two typical age peaks of occurrence of autoimmune diseases. At variance with the recent reports of vaccineinduced immune thrombotic thrombocytopenia (VITT),7 these cases were not associated with thrombosis in cerebral or abdominal veins but only with hemorrhagic symptoms compatible with the degree of thrombocytopenia in ITP and TTP and of factor VIII deficiency in autoimmune acquired hemophilia. Another feature that distinguishes these cases from VITT is that they were not accompanied by serological positivity for autoantibodies directed against platelet factor 4. Table 1 summarizes the main clinical symptoms and laboratory findings in the different thrombocytopenias that did occur after vaccination against COVID-19. What are the general messages that may be drawn from these reports of immune-mediated hematological diseases associated with a bleeding tendency in persons recently vaccinated to prevent COVID-19? It is well established that a number of diseases due to the formation of autoantibodies against autologous cells and/or proteins may occur after vaccination against various infectious agents:8-10 common examples are measlesmumps-rubella and diphtheria-tetanus-pertussis vaccines, but also vaccines against polio, rabies, influenza

Table 1. Main features of vaccine-induced, immune mediated thrombocytopenias.

Disease (and acronym)

Severe thrombocytopenia (<10x109/L)

Mucocutaneous bleeding symptoms

Intracerebral hemorrhage

Associated thrombosis

Thrombosis sites

Laboratory diagnosis

Immune thrombocytopenic purpura (ITP) Thrombotic thrombocytopenic purpura (TTP)

Frequent

Rare

Anti-platelet antibodies

Frequent

Rare

Frequent, microvascular

Vaccine-induced immune thrombotic thrombocytopenia (VITT)

Frequent

Rare

Frequent

Frequent, macrovascular

Microcirculation of heart, brain and GI tract Cerebral and abdominal veins

ADAMTS-13 deficiency and ADAMTS-13 antibody Anti-PF4 ELISA positivity

GI: gastrointestinal; ADAMTS13: a disintegrin and metalloproteinase with a thrombospondin type 1 motif, member 13; PF4: platelet factor 4; ELISA: enzyme-linked immunosorbent assay.

haematologica | 2022; 107(4)

785

Editorials

and bacterial pneumonia, especially in children but also in adults. There is no evidence that the innovative technologies recently developed for anti-COVID vaccine production have a particular role in the dysregulation of the immune system that led to the production of antibodies other than those towards the spike SARS-CoV-2 protein, because autoimmune diseases have occurred after all types of vaccines, spanning from those traditionally based upon inactivated virions to those newly employing viral DNA vectors or mRNA technology.8-10 Only VITT appears to be peculiar, because this complication has so far been described with convincing documentation only in patients receiving the vaccines based on adenoviral vectors, such as the AstraZeneca and the Johnson & Johnson products. In VITT the very rare but catastrophic thrombohemorrhagic complications are due to the formation of highly pathogenic autoantibodies against a complex between platelet factor 4 and a still poorly defined polyanion that triggers platelet activation, consumptive thrombocytopenia and a hypercoagulable state perhaps amplified by antibody-induced NETosis.7 However, it is not yet fully understood why venous thrombi occur in unusual sites, and the source and composition of the polyanion are still unlcear. Moreover, it remains uncertain whether or not these rare post-vaccination disorders are more frequent than expected in the population at large, because epidemiologically-based studies evaluating their incidences in vaccinated versus non-vaccinated persons are scanty or absent. The reported prevalences in vaccinated people, usually affected by limited sample size, range from 1 in 50.000-100.000 for VITT depending on the age and gender of the vaccine recipients to a lower prevalence (1 in 1,000,000) for ITP.6,11,12 An array of innate or adaptive immunological mechanisms may be responsible for these adverse events, but vaccine-induced danger signals accompanied by inflammation, as well as antigenic mimicry with activation of quiescent autoreacting B and T cells, are the most plausible.8,10 It is unlikely that adjuvants, frequently employed in some vaccines in order to boost antibody production towards the target antigen, play a mechanistic role, because the currently licensed anti-COVID vaccines do not need nor contain typical adjuvants such as squalene and aluminum, because their RNA and DNA components offer intrinsic ‘adjuvanticity’. On the whole, these exceptional cases of immunemediated hematological diseases associated with bleeding and/or thrombosis that have occurred in the current frame of global vaccination of more than 400 million people should not put in doubt nor jeopardize, in general

786

and in the specific instance of COVID-19, the effectiveness of vaccines, which are the only weapon currently available to control this pandemic. The majority of ITP and TTP cases seem to be less severe than VITT and are usually not life-threatening, except in older individuals with multiple comorbid conditions, such as case 2. In addition, it appears that, albeit with the limited amount of available knowledge given the recent onset and short follow-up of these complications, responses to state-ofthe-art therapies, as well as tendencies to recur or become chronic, are not overtly different from those of cases that occur irrespective of vaccination. By the same token, no prophylactic measure is warranted before or after vaccination, because useless and potentially dangerous. Disclosures No conflicts of interest to disclose.

References 1. Giuffrida G, Condorelli A, Di Giorgio MA, et al. Immune-mediated thrombotic thrombocytopenic purpura following administration of the Pfizer-BioNTech COVID-19 vaccine. Haematologica. 2022; 107(4):1008-1010. 2. World Health Organization. Global manual on surveillance of adverse effect following immunization. 2015 Update. World Health Organization. 2014. 3. Yocum A, Simon EL. Thrombotic thrombocytopenic purpura after Ad26.COV2-S vaccination. Am J Emerg Med. 2021 Nov;49:441.e3441.e4 4. Sissa C, Al-Khaffaf A, Frattini F, et al. Relapse of thrombotic thrombocytopenic purpura after COVID-19 vaccine. Transfus Apher Sci. 2021;60(4):103145. 5. Radwi M, Farsi S. A case report of acquired hemophilia following COVID-19 vaccine. J Thromb Haemost. 2021;19(6):1515-1518. 6. Lee EJ, Cines DB, Gernsheimer T, et al. Thrombocytopenia following Pfizer and Moderna SARS-CoV-2 vaccination. Am J Hematol. 2021;96(5):534-537. 7. Cines DB, Bussel JB. SARS-CoV-2 vaccine-induced immune thrombotic thrombocytopenia. N Engl J Med. 2021;384(23):2254-2256. 8. Guimarães LE, Baker B, Perricone C, Shoenfeld Y. Vaccines, adjuvants and autoimmunity. Pharmacol Res. 2015;100:190-209. 9. Perricone C, Ceccarelli F, Nesher G, et al. Immune thrombocytopenic purpura (ITP) associated with vaccinations: a review of reported cases. Immunol Res. 2014;60(2-3):226-235. 10. Watad A, De Marco G, Mahajna H, et al. Immune-mediated disease flares or new-onset disease in 27 subjects following mRNA/DNA SARS-CoV-2 vaccination. Vaccines (Basel). 2021;9(5):435. 11. Simpson CR, Shi T, Vasileiou E, et al. First-dose ChAdOx1 and BNT162b2 COVID-19 vaccines and thrombocytopenic, thromboembolic and hemorrhagic events in Scotland. Nat Med. 2021;27(7):1290-1297. 12. Pottegård A, Lund LC, Karlstad Ø, et al. Arterial events, venous thromboembolism, thrombocytopenia, and bleeding after vaccination with Oxford-AstraZeneca ChAdOx1-S in Denmark and Norway: population based cohort study. BMJ. 20215;373:n1114.

haematologica | 2022; 107(4)

Editorials

Not all mismatches are equal: importance of alloreactivity direction Jacinta Perram1,2 and Nada Hamad1,2,3 1

Department of Haematology, St Vincent's Hospital Sydney; 2St Vincent's Clinical School Sydney, University of New South Wales and School of Medicine Sydney, University of Notre Dame Australia, Sydney, Australia

E-mail: NADA HAMAD - nada.hamad@svha.org.au doi:10.3324/haematol.2021.279587

llogeneic hematopoietic stem cell transplantation (HSCT) remains the only curative therapy for numerous hematologic malignant and benign conditions. Transplantation does, however, come with significant risks of morbidity and mortality related to the transplant, graft-versus-host disease (GVHD) and relapse. The art of medicine in allogeneic HSCT lies in walking the tightrope between relapse risk and GVHD. We have come a long way in minimizing the risks of GVHD with rates of clinically significant (grade 2-4) acute GVHD and moderate to severe chronic GVHD reported to be as low as 50% and 30%, respectively.1,2 This is, no doubt, related to more widespread use of T-cell depletion methods and better HLA typing.3 Optimal donor selection in the absence of a matched relative relies on an assessment of the relative risk, and selection of the donor most genetically suitable based on HLA matching at HLA-A, -B, -C, DR and DQ. Human-leukocyte-antigen DPB1 (HLA-DPB1) mismatch is known to be broadly associated with decreased relapse at the cost of increased rates of acute GVHD.4 HLA-DPB1 mismatch in otherwise matched donors is common, yet our ability to predict GVHD severity based on this mismatch is limited. In this issue, Zou et al.5 present data supporting HLA-DPB1 mismatch associated risks of acute GVHD while using clinical correlation to investigate the clinical impact of HLA-DPB1 molecular mismatch. In the last two decades, donor selection algorithms have classified HLA-DPB1 mismatches as permissive or nonpermissive, based on functional toxicity assays and T-cell

epitope (TCE) analysis.6,7 While these methods assess the qualitative character of a mismatch, they do not evaluate direction or anticipate immunogenicity of a given mismatch. This poorly characterized potential risk represents a limitation of current donor selection algorithms. By contrast, molecular matching techniques assess structural components of epitopes, called eplets, allowing quantification of donor and recipient mismatched eplets (ME). This quantification, when combined with the PIRCHE score (PS), a predictor of TCE alloreactivity, has been shown to predict immunogenicity and clinical outcomes in haploidentical transplant recipients.8 Zou et al.5 present novel data on the use of molecular algorithms for HLA-DPB1 mismatch in a cohort of more than 1,500 patients who received an unrelated donor transplant between 2005-2018 at The University of Texas MD Anderson Cancer Center. The primary question in this study is whether molecular matching offers superior prognostic guidance than the traditional TCE model. The group reports concordance testing of bidirectional ME and PS, as well as the TCE model, with acute GVHD outcomes. The central finding is that high levels of ME in the graft-versus-host (GVH) direction are the strongest single predictor of acute GVHD. The authors propose use of molecular algorithms to guide the choice of or augmentation of acute GVHD prophylaxis. Another crucial finding is the importance of direction of alloreactivity. Bidirectional high ME or PS mismatch is universally associated with high rates of acute GVHD and

Figure 1. Competing risks regression for acute graftversus-host disease grade 2-4. GVHD: graft-versus-host disease; HR: hazard ratio; ME: mismatched eplets; GVH: graft-versus-host; HVG: host-versus-graft.

haematologica | 2022; 107(4)

787

Editorials

Figure 2. Competing risks regression for non-relapse mortality. HR: hazard ratio; ME: mismatched eplets; GVH: graft-versus-host.

Figure 3. Competing risks regression for relapse. HR: hazard ratio; ME: mismatched eplets; GVH: graft-versus-host; HVG: host-versus-graft.

relapse, suggesting a synergistic effect. The reduced relapse risk purported to arise from HLA-DPB1 permissive mismatch was only observed among those with high ME or PS in the GVH direction, and not those with isolated high PE or MS in the host-versus-graft (HVG) direction, who in fact had an increased rate of relapse. This is a clinically important outcome as it forces re-evaluation of the rationale for tolerating increased GVHD in recipients of transplants with HLA-DPB1 permissive mismatch. Building on the existing TCE model for HLA-DPB1 mismatch classification, there are several findings offering refined stratification. Among the permissive mismatch group, high HVG ME and PS are associated with a high risk of GVHD, yet among non-permissive mismatched

788

cases, an isolated high HVG ME is associated with low GVHD risk. The empirical inconsistency of these results highlights the persisting incomplete understanding of HVG pathophysiology. While the outcomes reported here have potential to alter practice in the future, the authors acknowledge the need for significant further study. Involvement of only a single center is a significant limitation. The retrospective nature of this research is not problematic in itself given the correlative nature of the analyses performed. However, one of the great challenges in allogeneic HSCT is the rapidly evolving landscape across which research is performed. Since the commencement of this study, the adoption of T-cell-depleting therapies has rapidly expanded.

haematologica | 2022; 107(4)

Editorials

Across the trial period, the use of anti-thymocyte globulin increased significantly (23.7% to 41.2%). It is unknown whether, and if so how, this has augmented results. Confirmatory investigations will be important to validate current findings and confirm them as enduring in the setting of routine T-cell depletion. Another important development has been the use of post-transplant cyclophosphamide in haploidentical allogeneic HSCT. The associated rapid improvement in clinical outcomes has shifted the donor selection paradigm. If anything, the uptake of haploidentical transplants reinforces the importance of this trial. With increased availability of alternative donors, the imperative to refine outcome prediction in HLA-DPB1 mismatch is all the more relevant. A further potentially significant development is related to the evolution of therapies for GVHD.2,9 More efficacious treatment and prevention strategies for GVHD may redefine donor selection algorithms, permitting mismatches that were previously prohibited. Zou et al.5 present an important novel approach to assessment of HLA-DPB1 mismatch permissibility. The authors acknowledge that confirmation of their findings with multi-center data is needed before refinement of algorithms can be considered. One of the great challenges for the field of HSCT moving forward is access to progressively more specialized molecular testing. There is also the need to embrace international research collaborations to allow realtime outcome reporting in a rapidly progressing field.

haematologica | 2022; 107(4)

Disclosures No conflicts of interest to disclose. Contributions JP and NH wrote and edited the editorial.

References 1. Jagasia M, Perales MA, Schroeder M, et al. Ruxolitinib for the treatment of steroid-refractory acute GVHD (REACH1): a multicenter, open-label phase 2 trial. Blood. 2020;135(20):1739-1749. 2. Zeiser R, Polverelli N, Ram R, et al. Ruxolitinib for glucocorticoidrefractory chronic graft-versus-host disease. N Engl J Med. 2021;385(3):228-238. 3. Bacigalupo A. ATG in allogeneic stem cell transplantation: standard of care in 2017? Point. Blood Adv. 2017;1(9):569-572. 4. Fleischhauer K, Shaw BE, Gooley T, et al. Effect of T-cell-epitope matching at HLA-DPB1 in recipients of unrelated-donor haemopoieticcell transplantation: a retrospective study. Lancet Oncol. 2012;13(4):366-374. 5. Zou J, Kongtim P, Oran B, et al. Refined HLA-DPB1 mismatch with molecular algorithms predicts outcomes in hematopoietic stem cell transplantation. Haematologica. 2022;107(4):844-866. 6. Zino E, Frumento G, Marktel S, et al. A T-cell epitope encoded by a subset of HLA-DPB1 alleles determines nonpermissive mismatches for hematologic stem cell transplantation. Blood, 2004;103(4):1417-1424. 7. Crivello P, Zito L, Sizzano F, et al. The impact of amino acid variability on alloreactivity defines a functional distance predictive of permissive HLA-DPB1 mismatches in hematopoietic stem cell transplantation. Biol Blood Marrow Transplant. 2015;21(2):233-241. 8. Zou J, Ciurea S, Kongtim P, et al. Molecular dispartiy in human leukocyte antigens is associated with outcomes in haploidentical stem cell transplantation. Blood Adv. 2020;4(15):3474-3485. 9. Zeiser R, von Bubnoff N, Butler J, et al. Ruxolitinib for glucocorticoirrefractory acute graft-versus-host disease. N Engl J Med. 2020;382(19):1800-1810.

789

REVIEW ARTICLE Ferrata Storti Foundation

The mitochondrial anti-apoptotic dependencies of hematologic malignancies: from disease biology to advances in precision medicine Isacco Ferrarini, Antonella Rigo and Carlo Visco Department of Medicine, Section of Hematology, University of Verona, Verona, Italy

Haematologica 2022 Volume 107(4):790-802

ABSTRACT

itochondria are critical organelles in the regulation of intrinsic apoptosis. As a general feature of blood cancers, different antiapoptotic members of the BCL-2 protein family localize at the outer mitochondrial membrane to sequester variable amounts of proapoptotic activators, and hence protect cancer cells from death induction. However, the impact of distinct anti-apoptotic members on apoptosis prevention, a concept termed anti-apoptotic dependence, differs remarkably across disease entities. Over the last two decades, several genetic and functional methodologies have been established to uncover the anti-apoptotic dependencies of the majority of blood cancers, inspiring the development of a new class of small molecules called BH3 mimetics. In this review, we highlight the rationale of targeting mitochondrial apoptosis in hematology, and provide a comprehensive map of the anti-apoptotic dependencies that are currently guiding novel therapeutic strategies. Cell-extrinsic and -intrinsic mechanisms conferring resistance to BH3 mimetics are also examined, with insights on potential strategies to overcome them. Finally, we discuss how the field of mitochondrial apoptosis might be complemented with other dimensions of precision medicine for more successful treatment of ‘highly complex’ hematologic malignancies.

Correspondence: ISACCO FERRARINI isacco.ferrarini@univr.it CARLO VISCO carlo.visco@univr.it Received: October 18, 2021. Accepted: January 7, 2022. Prepublished: January 20, 2022. https://doi.org/10.3324/haematol.2021.280201

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

790

Introduction Prevention of programmed cell death is a hallmark of cancer cells and efforts to re-establish pro-death pathways have been the mainstay of research in the field of anti-cancer therapeutics.1 Among modalities of programmed cell death, apoptosis is the best characterized in terms of triggering stimuli, sequencing of biochemical events, intracellular organelles involved, and morphological changes.2 Two interconnected forms of apoptosis have been described: the extrinsic and the intrinsic (i.e., mitochondrial) pathways. The former is triggered on the cell surface by the engagement of death receptors, such as tumor necrosis factor receptor and tumor necrosis factor-related apoptosis-inducing ligand receptor, and proceeds through caspase-8/10 activation and BID cleavage.3 The latter is induced by oncogenic signaling, nutrient deprivation, genotoxic drugs and other cellular stressors, and is regulated at the level of the outer mitochondrial membrane by pro- and anti-apoptotic BCL-2 family members.4

Overview of the BCL-2 family The BCL-2 family members are distinguished into three main categories: antiapoptotic, pro-apoptotic BH3-only, and pro-apoptotic effectors4 (Figure 1). BCL-2, MCL-1, BCL-w, BCL-B, BCL-xL, and BFL-1 are the main pro-survival relatives and contain all four BH domains.4,5 BID, BIM, PUMA, NOXA, BAD, HRK, and BMF are the pro-apoptotic BH3-only proteins, and only share the BH3 domain with the rest of the family. Among these, BID and BIM can directly activate the pro-apoptotic effectors, while the others act as sensitizers by antagonizing the pro-survival

haematologica | 2022; 107(4)

The anti-apoptotic landscape of blood cancers

members.6 BAX, BAK, and BOK contain three out of four BH domains and represent the pro-apoptotic effector proteins that form homodimers and heterodimers through the outer mitochondrial membrane.7 Overall, the mitochondrial cascade of apoptosis starts when a death stimulus triggers the translocation of BH3-only proteins to the mitochondrial surface. This leads to subsequent displacement of further pro-apoptotic activators and effectors from the pro-survival members, promoting BAX/BAK assembly through the outer mitochondrial membrane. These pore-like structures provoke the cytoplasmic leakage of cytochrome c and other apoptogenic factors, which in turn activate the executioner caspases, and disrupt the mitochondrial transmembrane potential that is necessary for oxidative phosphorylation (OxPHOS).8 BH3-only proteins have different interaction modalities for each anti-apoptotic member. BAD, HRK and NOXA bind preferentially to selected anti-apoptotic members, whereas BIM, BID, PUMA and BMF interact indiscrimi-

nately with all of them. BAD binds BCL-2, BCL-w and BCL-xL, HRK binds BCL-xL, while NOXA preferentially binds MCL-1 and BFL-1.6 In hematologic malignancies, there is selective pressure for upregulating the pro-survival members via genetic and non-genetic mechanisms. For example, in both follicular lymphoma (FL) and double-hit diffuse large-B cell lymphoma (DLBCL), t(14;18) juxtaposes BCL2 to the immunoglobulin heavy chain (IGH) locus, increasing BCL-2 protein levels.9,10 In chronic lymphocytic leukemia (CLL) with del(13q), the lack of microRNA 15/16 derepresses BCL-2 expression.11 Multiple myeloma (MM) and selected subtypes of DLBCL harbor the 1q21 amplification, which leads to MCL-1 overexpression.12,13 On the other hand, pro-apoptotic BH3-only members are sometimes downregulated. BIM is epigenetically silenced in a subset of patients with acute lymphoblastic leukemia (ALL), leading to glucocorticoid resistance and poorer outcomes.14 Similarly, a subgroup of patients with

Figure 1. The BCL-2 protein family. The three subgroups of the BCL-2 protein family and their recurrent genetic alterations are shown in the colored boxes. A schematic of the interactions occurring among the BCL-2 family members is represented in the lower right box, based on the “indirect activation” model. Overall, the antiapoptotic proteins sequester the pro-apoptotic BH3-only activators (A) and effectors, preventing the initiation of the apoptotic cascade. By contrast, the pro-apoptotic sensitizers (S) antagonize the anti-apoptotic members thus freeing the activators, which in turn trigger the polymerization of the effectors. This creates pore-like structures at the outer mitochondrial membrane which favor the release of cytochrome c (cyt c) and other apoptogenic factors. While the model shown here has been widely adopted to define the concept of apoptotic priming, recent evidence suggests that, at least in selected cancer types, BAX and BAK activation may only require that these members are freed from the anti-apoptotics, with no need of direct interaction with pro-apoptotic members. DLBCL: diffuse large B-cell lymphoma; FL: follicular lymphoma; ABC: activated B-cell; T-NHL: T-cell non-Hodgkin lymphoma; MM: multiple myeloma; HL: Hodgkin lymphoma; MCL: mantle cell lymphoma; GC-R: ALL glucocorticoid-resistant acute lymphoblastic leukemia; BL: Burkitt lymphoma.

haematologica | 2022; 107(4)

791

I. Ferrarini et al.

mantle cell lymphoma expresses low levels of BIM and is less likely to achieve complete response to standard treatments.15 All of these mechanisms converge towards apoptosis evasion, a common denominator for cancer cells.

Evasion from mitochondrial apoptosis Solid tumors and hematologic malignancies evade mitochondrial apoptosis in markedly different ways (Figure 2), perhaps reminiscent of the biology of the normal tissue counterparts.16 Diseases such as kidney, colorectal, and cervical cancer show an extremely poor proclivity to activate intrinsic apoptosis, even when strong pro-apoptotic stressors are directly applied on cancer cell mitochondria. Such low apoptotic priming, defined by the near absence of BH3-only activators on the mitochondrial surface, is a major factor contributing to the chemoresistance of solid tumors, and suggests targeting solely this pathway might not be successful against these diseases.16 By contrast, mitochondria of blood cancer cells struggle to maintain the integrity of the outer mitochondrial membrane due to its occupation by several proapoptotic activators (i.e., high mitochondrial priming). In this type of cancers, apoptotic evasion is based on the activity of several anti-apoptotic proteins aiming at buffering large amounts of pro-apoptotics that dynamically shuttle between the cytoplasm and the mitochondria.16 The need for an efficient anti-apoptotic arsenal creates a specific vulnerability that is being successfully targeted by venetoclax and other BH3 mimetics.17 This class of small molecules targets the interaction interface between the anti- and pro-apoptotic members thus allowing the latter to initiate the apoptotic cascade. While in some cases a single anti-apoptotic protein is the only barrier to apoptotic triggering, in others multiple antiapoptotic members act in concert to oppose outer mitochondrial membrane permeabilization18-20 (Figure 2).

Genetic and functional approaches to detect anti-apoptotic dependencies Over the last 15 years, several genetic and functional methodologies have been set up to derive anti-apoptotic dependencies in cancer. Genetic knock-out of selected anti-apoptotic proteins using CRISPR-Cas9 or related gene-editing techniques have pointed out the anti-apoptotic role of MCL-1 and BCL-w in myc-driven lymphomas.21-23 Moreover, doxycycline-inducible silencing RNA (siRNA) targeting specific pro-survival proteins has been successfully transfected in human lymphoma cell lines to evaluate which member has the greatest impact on in vitro cell survival.24 While the strength of these approaches lies in their ability to accurately inform about the role of a selected BCL-2 family gene, they are poorly applicable to primary cells from cancer patients. As their use is mostly limited to cancer cell lines, genetic approaches are precious to infer general principles of apoptotic regulation in a given cancer type, but lack scalability to large numbers of patient-derived samples. This is of relevance because some of the most common hematologic malignancies such as acute myeloid leukemia (AML) and DLBCL show heterogeneity of anti-apoptotic dependencies across patients, and perhaps even within patients over the course of their disease.19,25,26 Functional approaches provide higher scalability but less molecular insight. They are mainly based on in vitro and ex vivo pharmacological targeting of pro-survival BH3 proteins, and on BH3 profiling.27-30 Side-by-side comparisons of different BH3 mimetics targeting distinct antiapoptotic proteins have been performed on cancer cell lines and primary samples, and several readouts such as intracellular ATP content, annexin V externalization, and caspase 3/7 activation have been utilized to detect differences in cell viability.27,28 These pharmacological assays inform about the anti-apoptotic dependencies of tumor samples, working at the same time as apoptosis-tailored drug sensitivity screens with potential impact on preci-

Figure 2. Different forms of evasion from mitochondrial apoptosis. Three distinct scenarios are depicted. In cancers such as chronic lymphocytic leukemia, large amounts of pro-apoptotic activators are sequestered (i.e., high priming) by a single anti-apoptotic relative. In other hematologic cancers (e.g., B-cell acute lymphoblastic leukemia), apoptotic priming is high, but the pro-apoptotics are concurrently sequestered by multiple anti-apoptotic relatives (e.g., BCL-2, MCL-1 and BCL-xL). In solid tumors, pro-apoptotic members are mostly not bound to the anti-apoptotic proteins, and hence sensitivity to BH3 mimetics is generally low. CLL: chronic lymphocytic leukemia; B-ALL: B-cell acute lymphoblastic leukemia.

792

haematologica | 2022; 107(4)

The anti-apoptotic landscape of blood cancers

sion strategies. Long BH3-mimetic incubation times (i.e., more than 18-24 hours), which may be needed to detect many of the cell-death readouts, is a common shortcoming of these assays as primary cells do not often survive long-term in ex vivo cultures.31 Moreover, it has been reported that in vitro-cultured primary cancer cells lose similarities with the original tumor over time,32 thus weakening the reliability of the results. Off-target effects of some BH3 mimetics, for example BCL-2-independent inhibition of OxPHOS by venetoclax,33 might be an additional limitation of these assays when the primary aim is to study the biology of the disease rather than the drug efficacy. A clinically applicable declination of such approaches, called the BH3-mimetic toolkit, has been employed on MM samples using CD138 loss as a flow cytometry readout of cancer cell death. Venetoclax (a BCL-2 inhibitor), A1155463, A1331852 (both BCL-xL inhibitors), and A1210477 (a MCL-1 inhibitor) were tested at multiple concentrations, and three dependency groups were derived in an unbiased way using analytical tools.30 BH3 profiling is a different functional technique based on exposing cancer cell mitochondria to an array of pro-apoptotic peptides with distinct binding modalities for different anti-apoptotic members. In this assay, the anti-apoptotic addiction of cancer cells can be inferred by the pattern of cytochrome c release upon peptide incubation.34 Due to the short incubation time, BH3 profiling averts the risk of artifactual genetic selections or functional modifications that could potentially occur during prolonged ex vivo culture. In addition, the specificity of treating peptides that act directly on the mitochondrial surface renders this assay particularly focused on the BCL-2 fam-

ily dynamics, without off-target effects that might instead be encountered using small molecules. A potential downside is that BH3 profiling is an organelle-centered method that does not take into account how other cellular components might react to peptide-induced cytochrome c release. For instance, defective activation of downstream cytosolic caspases, which sometimes occurs in solid tumors due to somatic mutations,35 may blunt the apoptotic response triggered by cytochrome c leakage. BH3 profiling also functions as a platform to set up an additional assay, named dynamic BH3 profiling, which measures changes in apoptotic priming and anti-apoptotic dependencies triggered by drug candidates. In this case, the incubation with pro-apoptotic peptides is preceded by ex vivo treatment with a panel of drugs, with the aim of identifying those that most efficiently lower the threshold for cytochrome c release.36,37 Dynamic BH3 profiling has proven useful to assess the impact of Bruton tyrosine kinase (BTK) inhibitors on BCL-2 dependence of CLL cells,38 and as a functional precision medicine strategy for T-cell prolymphocytic leukemia.39

The anti-apoptotic map of hematologic malignancies Extensive preclinical research and clinical observations are building our knowledge on how cancer cells evade apoptosis. The anti-apoptotic map of hematologic malignancies (Figure 3) has recently guided highly effective treatments for CLL and AML,40,41 and is currently inspiring novel antineoplastic regimens for other types of blood

Figure 3. The anti-apoptotic map of hematologic malignancies. The colored map shows the antiapoptotic dependencies of several hematologic malignancies, based on preclinical data and clinical results (see references). With the exception of follicular lymphoma, in which the BCL-xL dependence is driven by microenvironmental stimuli, all the other anti-apoptotic dependencies depicted here refer to cell-intrinsic dependencies. AML: acute myeloid leukemia; MLL-r B-ALL: MLL-rearranged B-cell acute lymphoblastic leukemia; ETP T-ALL: early T-cell acute lymphoblastic leukemia; CLL: chronic lymphocytic leukemia; DLBCL: diffuse large B-cell lymphoma; FL: follicular lymphoma; MM multiple myeloma; ALCL: anaplastic large T-cell lymphoma; T-NHL NOS: T-cell nonHodgkin lymphoma not otherwise specified; CTCL: cutaneous T-cell lymphoma; T-PLL: T-cell prolymphocytic leukemia; HL: Hodgkin lymphoma; BPDCN: blastic plasmacytoid dendritic cell neoplasia.

haematologica | 2022; 107(4)

793

I. Ferrarini et al.

disorders. This section focuses on key preclinical findings relevant to apoptosis avoidance. Clinical results of BH3 mimetics in hematology will only be touched on, as they have been recently reviewed by Roberts and colleagues.42

Acute myeloid leukemia and myelodysplastic syndromes Regulation of intrinsic apoptosis in AML shows both intratumor and interpatient heterogeneity. In 2014, Pan and co-workers identified a BCL-2 dependence for roughly 80% of primary AML cases, with rapid apoptotic triggering upon ex vivo exposure to venetoclax. BCL-2 protein expression correlated with sensitivity to venetoclax, whereas BCL-xL and, to a lesser extent MCL-1, showed anti-correlation with susceptibility to BCL-2 inhibition.43 A phase II trial evaluating the activity of venetoclax as a single agent in high-risk AML demonstrated an overall response rate of 19%, with particularly favorable responses among patients carrying isocitrate dehydrogenase 1/2 (IDH1/2) mutations.41 BH3 profiling identified patients who were more likely to stay on venetoclax therapy longer than 30 days, working as a predictive functional assay with potential clinical applications.41 Bone marrow cells from patients with high-risk myelodysplastic syndromes are also sensitive to BCL-2 antagonism. In vitro treatment with ABT-737 or venetoclax depletes the myelodysplastic syndrome progenitor compartment, and decreases the colony-forming capacity and the percentage of CD34+ cells.44 The combination of venetoclax plus intensive chemotherapy led to complete remission in 82% of patients with newly-diagnosed AML and highrisk myelodysplastic syndromes.45 Despite the meaningful clinical activity of venetoclax-based regimens in these settings, about 20% of patients are primarily refractory. Among AML patients achieving complete remission with venetoclax plus azacytidine, the median duration of response was only 11.3 months.46 This suggests that a subset of AML cases is not BCL-2 dependent, and that additional groups of patients may harbor subclonal dependencies to different anti-apoptotic proteins. A recent study addressed this point and found that BCL-2 dependence decreases through stages of AML morphological maturation.47 Indeed, lower BCL-2 expression, at both mRNA and protein levels, was observed in FrenchAmerican-British (FAB) M5 AML compared to FABM0/M1/M2 leukemia. A combination of venetoclax plus azacitidine failed to inhibit OxPHOS in monocytic AML, with a modest impact on cell viability.47 By contrast, the MCL-1 inhibitor VU661013 combined with azacitidine significantly suppressed OxPHOS in monocytic cases and was more effective than venetoclax-based treatments in inducing cell death. Genetic knockdown of MCL-1 was sufficient to trigger apoptosis in primary monocytic AML specimens, further indicating their reliance on MCL-1 to maintain survival.47 Accordingly, 62% of patients with FAB-M5 AML were refractory to venetoclax-azacitidine, whereas only 8% of non-FAB-M5 cases did not respond to this regimen.47 A side-by-side comparison of BH3 mimetics in mediating AML cell killing confirmed that a subset of immortalized and primary AML cells is highly sensitive (low micromolar/nanomolar range) to the MCL1 inhibitor S63845, but not to BCL-xL or BCL-2 inhibitors.27

Acute lymphoblastic leukemia B-cell ALL shows concurrent dependence on BCL-2 and 794

BCL-xL, with MCL-1 being identified as a further antiapoptotic member able to confer resistance to single or dual BCL-2/BCL-xL antagonism.48-50 Indeed, sensitivity to venetoclax, which was predicted by BCL2 gene expression level as well as BH3 profiling, was highly heterogeneous in a panel of B-ALL cell lines and patient-derived xenograft, with EC50 values ranging from 1.8 nM to 5.5 mM.49 Accordingly, venetoclax was effective in vivo in only a minority of B-ALL xenografts, whereas combined inhibition of BCL-2 and BCL-xL resulted in synergistic killing of most B-ALL in vivo models.51 A recently reported phase I trial of the combination of venetoclax with low-dose navitoclax (dual BCL-2/BCL-xL inhibitor) plus chemotherapy in relapsed/refractory (R/R) B-ALL has shown encouraging results, with a complete remission rate of 60%.52 Despite the heterogeneity and the high degree of anti-apoptotic co-dependencies in B-ALL, the subgroup carrying the t(4;11) translocation proved uniformly BCL-2-dependent.51,53 The fusion protein MLL/AF4 activates BCL2 transcription via DOT1L-mediated H3K79me2/3, without altering the expression level of other anti-apoptotic members. This renders MLL-rearranged cells susceptible to venetoclax-induced apoptosis in vitro, and sensitive to venetoclax-based combinations in vivo.53 The anti-apoptotic dependency of T-ALL reflects the maturation stage of lymphoblasts. Early T-cell progenitor (ETP) ALL cells most closely resemble early thymic, CD4– /CD8– T cells, and are dependent upon BCL-2.54 Instead, non-ETP T-ALL cells have a gene expression and phenotypic profile resembling that of more mature, CD4+/CD8+ T cells, expressing abundant BCL-xL protein levels, and having functional dependency upon BCL-xL.54 While ETP-ALL patient-derived xenograft models are very sensitive to venetoclax in vivo, the non-ETP counterpart is relatively resistant.54 Moreover, sensitivity to long-term BCL-2 inhibition in ETP-ALL might be compromised by microenvironment-derived signals. More specifically, the spleen has been identified as a sanctuary site for residual ETP lymphoblasts following venetoclax treatment. Such surviving cells display decreased BCL-2 expression and reduced BCL-2 dependence, with requirement of concomitant MCL-1 inhibition to evoke robust cell death.55

Chronic lymphocytic leukemia In the light of its broad heterogeneity in terms of chromosomal aberrations, gene mutations, clinical characteristics, and drug response profiles,56-58 CLL shows surprisingly homogeneous anti-apoptotic regulation.18 More than three decades of basic discoveries have pointed out that BCL-2 is overexpressed in CLL cells compared to normal B lymphocytes due to gene promoter hypomethylation,59 miR15/16 downregulation,11 and, more rarely, BCL2 translocation.60 Such genetic bases for BCL-2 addiction have been functionally confirmed by BH3 profiling, which revealed a clear-cut BCL-2 dependence of CLL cells regardless of TP53 status and previous lines of therapy.18,61 This set the stage for the clinical introduction of venetoclax, the first approved selective BCL-2 inhibitor, which achieved an overall response rate of 79% with 20% of complete remissions in the R/R CLL setting.40 While cell-intrinsic genetic programs seem to be responsible for the exceedingly high BCL-2 dependency of CLL cells, signals from the microenvironment can add further layers of anti-apoptotic protection. Indeed, antihaematologica | 2022; 107(4)

The anti-apoptotic landscape of blood cancers

gen stimulation and CXCL12, both converging on BTK, decrease apoptotic priming and BCL-2 dependence.62 As a consequence, ibrutinib and other inhibitors of the B-cell receptor signaling pathway increase BCL-2 dependence by impeding extrinsic signals to upmodulate MCL-1 and BCL-xL which would ultimately favor additional pro-survival forces.38 The CLARITY study, exploring the combination of ibrutinib and venetoclax in R/R CLL, found 51% complete remissions with a high rate of minimal residual disease eradication, which enabled treatment cessation in a subset of patients.63

B-cell non-Hodgkin lymphoma DLBCL, the most common type of B-cell nonHodgkin lymphoma (NHL), is remarkably heterogeneous in terms of mutational landscape and clinical pictures.64 Anti-apoptotic dependencies show heterogeneity as well, and do not correlate with cell of origin.19,26,65 Loss of the anti-apoptotic BCL-w was reported to delay MYC-driven lymphoma development in Em-Myc transgenic mice by augmenting MYC-induced apoptosis.22 Moreover, BCL-w is overexpressed in a subset of DLBCL characterized by shorter overall survival, suggesting a primary role for apoptosis evasion.22 Despite this, a recent study found that CRISPR/Cas9-mediated loss of BCL-w did not trigger apoptosis in DLBCL cell lines, nor did it increase the sensitivity to BH3 mimetics targeting additional pro-survival proteins, casting doubts about the role of BCL-w in human DLBCL.28 MYC-driven lymphomagenesis is also sustained by MCL-1, which is highly expressed in activated B-cell DLBCL. 66 MCL1 copy number abnormalities were detectable in 25.7% of activated B-cell -DLBCL versus 12.5% of germinal-center DLBCL, and constitutive STAT3 signaling further contributes to MCL-1 upregulation in selected cases. 66 The MCL-1 antagonist AZD5991, currently in clinical testing, curtails tumor growth and disrupts mitochondrial metabolism in MCL-1-addicted DLBCL cells via TP53- and BAXdependent mechanisms. 13 BCL2 is also frequently deregulated in DLBCL due to the t(14;18) chromosomal translocation, gene mutations, copy number alterations and amplifications.67 Such genetic events confer poor prognosis especially when combined with MYC amplification.10 While in some cases BCL2 dysregulation is associated with functional dependency on BCL-2 and high sensitivity to venetoclax, in others targeting additional pro-survival members is needed to achieve apoptosis. Indeed, several B-cell receptor-dependent DLBCL lines with genetic bases for BCL-2 dysregulation require dual targeting of PI3Ka/d, which in turns decreases MCL-1 abundance, and BCL-2 to commit cell death.68 Anti-apoptotic heterogeneity and frequent co-dependencies account for the disappointing results of singleagent venetoclax against DLBCL, with an overall response rate of only 18% in R/R cases.69 Less common subtypes of B-NHL, including follicular lymphoma and mantle cell lymphoma, display better clinical responses to BCL-2 inhibition.69,70 However, durable remissions are not frequent and may require combination therapies. At least in follicular lymphoma, in which high BCL-2 expression caused by the t(14;18) translocation has ever been considered the pathogenic hallmark, concurrent antagonism of microenvironment-induced BCLxL is needed to efficiently trigger apoptosis.71 haematologica | 2022; 107(4)

Multiple myeloma At least one-third of MM cases are predominantly MCL-1 dependent, whereas the others are either BCL-2 dependent or characterized by BCL-2/MCL-1 co-dependence.12,20,30,72 Moreover, a minority of MM cell lines and primary samples show some degree of BCL-xL dependence.30,72 MCL-1 addiction can be driven by genetic alteration and microenvironmental cues. The MCL1 gene is located on 1q21, which is amplified in 43-72% of MM patients depending on disease status.73 1q21 copy gain directly correlates with MCL1 transcript abundance and MCL-1 protein expression.12 Importantly, plasma cells from 1q21-amplified cases are especially sensitive to MCL-1 inhibition, while proving relatively resistant to venetoclax.12 Interleukin-6 released by bone marrow stromal cells can further amplify MCL-1 dependence through MCL-1 transcriptional upregulation and BIM phosphorylation. Such events switch the BIM binding partner from BCL-2 and BCL-xL to MCL-1.74,75 Blocking the interleukin6 signaling pathway decreases MCL-1 dependence and enhances sensitivity to venetoclax.76,77 On the other hand, BCL-2-dependent cases are enriched among MM with the t(11;14)(q13;q32) translocation, which is detected in 15% to 20% of cases.78 In the phase I study of venetoclax monotherapy in R/R MM, the overall response rate was 21% in the whole population, but reached up to 40% in the subgroup harboring t(11;14).79 High BCL2:MCL1 and BCL2:BCL2L1 mRNA expression ratios correlated with venetoclax sensitivity.79 Given the anti-apoptotic heterogeneity and the frequent co-dependencies of MM cells, dual targeting of BCL-2 and MCL-1 has been assessed in preclinical studies, showing a profound synergism potentially translatable to the clinic.77 Moreover, the BH3mimetic toolkit revealed an increase in MCL-1 addiction from 33% at diagnosis to 69% at relapse, indicating temporal remodeling of cellular dependencies.30 This assay also identified a subset of newly diagnosed MM patients not sensitive to any of the BH3 mimetics, highlighting that apoptosis targeting might not always be a suitable therapeutic option.30

T-cell non-Hodgkin lymphoma Protein expression analyses identified MCL-1 as the major anti-apoptotic member in cell lines and primary samples of systemic T-cell NHL.80,81 Copy number gains involving the MCL1 locus were found in ten out of 21 TNHL cell lines.80 Accordingly, loss of a single MCL1 allele delayed tumor formation in T-NHL mouse models and compromised the viability of neoplastic T cells,81 BH3 profiling confirmed that most T-NHL cell lines, especially those from anaplastic T-cell lymphomas and other peripheral T-cell lymphomas, are MCL-1 dependent.80 The development of several MCL-1 antagonists has recently allowed the translation of these biological findings into pharmacological approaches. Indeed, AZD5991 reduced tumor volumes in vivo, and synergized with cyclophosphamide, vincristine, doxorubicin and prednisone to improve survival of mice with T-NHL.80 By contrast, cutaneous T-NHL cases are primarily BCL-xL dependent, and some of them carry copy number alteration of the BCL2L1 gene.80 A proteolysis targeting chimera that targets BCL-xL for degradation has been developed to effectively kill cutaneous T-NHL cells in vitro and in vivo, without causing significant thrombocytopenia as previously reported for navitoclax.82 795

I. Ferrarini et al.

T-cell prolymphocytic leukemia As compared to CLL, T-cell prolymphocytic leukemia is less primed for apoptosis, less BCL-2 dependent, and often BCL-2/MCL-1 co-dependent.39,83 BCL-xL dependence is usually more pronounced in T-cell prolymphocytic leukemia than in CLL, but remains of secondary importance when compared with BCL-2 and MCL-1 dependence.39 Clinical reports suggest that venetoclax monotherapy is often inadequate to get durable disease control.84 As demonstrated by dynamic BH3 profiling, inhibitors of histone deacetylase and the JAK/STAT pathway increase apoptotic priming and BCL-2 dependence, strengthening the pro-apoptotic effect of venetoclax in vitro and in vivo.39

Hodgkin lymphoma As compared to normal B lymphocytes, Hodgkin lymphoma cells express higher levels of BCL-w and BCL-xL mRNA.85 BCL-w expression further increases in advanced stages and in the R/R setting.85 Fluorescence in situ hybridization analyses on Hodgkin lymphoma samples showed that copy number gains of chromosome 14 and amplifications of the BCL-w containing region were common events in the pathogenesis of Hodgkin lymphoma and led to almost invariably high BCL-w protein expression as detected by immunohistochemistry. Genetic knockdown of BCL-w or pharmacological antagonism of BCL-xL significantly reduced Hodgkin lymphoma cell viability, pointing to a primary role for these

anti-apoptotic proteins in sustaining Hodgkin lymphoma growth.85

Blastic plasmacytoid dendritic cell neoplasm Preclinical data indicate that blastic plasmacytoid dendritic cell neoplasm is primarily BCL-2 dependent. Blastic plasmacytoid dendritic cell neoplasm cells collected from patient-derived xenografts or directly from patients’ skin and bone marrow showed higher BCL-2 dependency compared with randomly selected AML cases.86 In vivo experiments confirmed these findings and, to date, several reports have been published about patients with relapsed blastic plasmacytoid dendritic cell neoplasm successfully treated with single-agent venetoclax.87,88 Such a therapeutic approach is currently being investigated in a phase I clinical trial (NCT03485547).

Resistance to BH3 mimetics: a tale of cellular interactions, mitochondrial biology and mutational pressure With the growing use of venetoclax in clinical practice, several resistance mechanisms have been outlined, especially in the context of CLL and AML which currently represent the major indications for BCL-2 inhibition. Three broad concepts are emerging. First, clinical progressions on venetoclax are mostly underpinned by “polyclonal patterns”, whereby different leukemic clones within the same

Figure 4. Mechanisms of venetoclax resistance. The figure depicts the four major modalities of resistance to venetoclax: cell-extrinsic mechanisms, outer and inner mitochondrial adaptation, genomic alterations, and expansion of intrinsically resistant subclones. Mechanisms highlighted in purple have been described in acute myeloid leukemia. Those highlighted in blue have been reported in chronic lymphocytic leukemia. See text for pathway details. SCF: stem cell factors; OxPHOS: oxidative phosphorylation; FAO: fatty acid oxidation.

796

haematologica | 2022; 107(4)

The anti-apoptotic landscape of blood cancers

Table 1. Treatment strategies that may prevent/overcome escape from venetoclax based on preclinical mechanistic predictions.

Treatment strategy

Indication

Mode of action of partner drug

Escape mechanism: extrinsic interactions and downstream signaling pathways cirmtuzumab + venetoclax CLL anti-ROR1 duvelisib + venetoclax CLL PI3Kgd inhibitor copanlisib + venetoclax DLBCL PI3Kad inhibitor ruxolitinib + venetoclax AML JAK inhibitor BP1001 + venetoclax AML L-Grb2 antisense oligonucleotide ibrutinib + venetoclax CLL covalent BTK inhibitor loxo305 + venetoclax + rituximab CLL non-covalent BTK inhibitor plerixafor + venetoclax AML anti-CXCR4 Escape mechanism: anti-apoptotic and metabolic mitochondrial adaptation AMG 176 + venetoclax R/R heme malignancies MCL-1 inhibitor S64315 + venetoclax AML MCL-1 inhibitor AZD5991 + venetoclax AML MCL-1 inhibitor Navitoclax + venetoclax + chemo B-ALL BCL-xL inhibitor adi-peg 20 + venetoclax + azacytidine AML arginine depleting enzyme Omacetaxine + venetoclax AML protein translation inhibitor IACS-010759 + venetoclax AML OxPHOS inhibitor ONC201 + venetoclax AML OxPHOS inhibitor Escape mechanism: gene alterations BGB-11417 B-cell malignancies BCL-2 inhibitor (active against G101V) eprenetapopt + venetoclax MCL p53 reactivator

Study phase

NCT number

I I/II I/II I II II III /

NCT03797261 NCT03534323 NCT04572763 NCT03874052 NCT02781883 NCT02756897 NCT04965493 /

I I I/II I I I / /

NCT03797261 NCT03672695 NCT03218683 NCT03181126 NCT05001828 NCT04874194 / /

I II

NCT04883957 NCT04990778

NCT: National Clinical Trials; CLL: chronic lymphocytic leukemia; DLBCL: diffuse large B-cell lymphoma; AML: acute myeloid leukemia; R/R: relapsed or refractory; ALL: acute lymphoblastic leukemia; MCL: mantle cell lymphoma.

patient take distinct paths to survive BCL-2 inhibition independent of each other.89 Secondly, the general biological principles driving resistance to one specific BH3 mimetic are shared with other BCL-2 family antagonists. Although this section is mainly focused on the mechanisms of resistance to venetoclax, the only Food and Drug Administration-approved BH3 mimetic so far, early data are piling up about similar escape trajectories occurring in cells treated with MCL-1 or BCL-xL antagonists.25 Thirdly, resistance to BH3 mimetics is highly “multimodal”, as multiple cellular components can be rewired to undermine the efficacy of BCL-2 family antagonism. The four major modalities of acquisition of resistance to venetoclax are based on cell-extrinsic interactions,90,91 mitochondrial adaptation,92 genomic alterations,89 and, as previously highlighted for the monocytic escape of AML, the emergence of intrinsically resistant clones47 (Figure 4). While such variety of resistance mechanisms contrasts with the simpler mutational evolution often observed in patients treated with kinase inhibitors,93 it offers several clues to plan at best subsequent therapies, and to design strategic drug combinations as well (Table 1).

Cell-extrinsic interactions In AML, CXCL12 released by surrounding stromal cells protects the leukemic clone from BCL-2 inhibition by activating the CD44/CXCR4 complex, which in turn induces the transcription of several pro-survival embryonic stem-cell core transcription factors.90 Likewise, in CLL CD40 ligation enhances non-BCL-2 anti-apoptotic dependencies by inducing BCL-xL transcription via the canonical as well as the non-canonical NF-kB pathway.91 Induction of pro-survival members by cell-extrinsic factors may account for the relatively less durable response to venetoclax observed among CLL patients with >5 cm haematologica | 2022; 107(4)

lymph nodes, where most of the cell-to-cell and paracrine interactions take place.40

Mitochondrial adaptations Mitochondria are active players in the initiation of venetoclax resistance due to anti-apoptotic remodeling, occurring at the outer mitochondrial membrane, and metabolic reprogramming, occurring at the inner mitochondrial membrane. Functional shifts towards alternative antiapoptotic defenses render leukemic mitochondria progressively less vulnerable to BCL-2 antagonism, with dual BCL-2 and MCL-1 antagonism outperforming the individual targeting.25,92,94 Moreover, a global reduction of mitochondrial apoptotic priming is quite common in AML acquiring resistance to different types of BH3 mimetics.25 As apoptotic regulation is tightly connected with bioenergetic processes, the upmodulation of mitochondrial metabolic pathways confers resistance to BH3 mimetics. While in venetoclax-sensitive AML stem cells OxPHOS is suppressed through the inhibition of amino acid metabolism, in resistant clones it is restored through the enhancement of fatty acid oxidation.95 In this context, the role of BCL-2 remains unclear because venetoclax is reported to inhibit OxPHOS independently of BCL-2 expression.96 Thus, resistant cells might have evolved the ability to restore OxPHOS, bypassing the inhibitory effect of venetoclax. In addition, resistant AML stem cells show elevated nicotinamide metabolism which promotes the uptake and catabolism of amino acids, as well as the conversion of fatty acids into the tricarboxylic acid cycle intermediates 2-oxoglutarate and malate. Indeed, inhibitors of nicotinamide phosphoribosyltransferase specifically target leukemic cells with acquired resistance to venetoclax.97 The modulation of mitochondrial cristae ultrastructure also impacts on bioenergetics and BH3 mimetic sensitivi797

I. Ferrarini et al.

ty. The mitochondrial chaperonin CLPB has been recently identified as a crucial interactor of the cristae-shaping protein OPA1, and its downregulation alters cristae architecture and renders cytochrome c more prone for cytosolic release.98 An independent genome-wide CRISPR screen identified genes involved in mitochondrial translation as an additional circuit to bypass BCL-2 antagonism, and antibiotics targeting the mitochondrial ribosomes effectively overcome venetoclax resistance.99

Genomic alterations Emergence of subclones harboring BCL2 mutations is increasingly described in CLL patients progressing on venetoclax, with the most frequent being Gly101Val and substitutions at Asp103.89 These BCL-2 structural variants decrease venetoclax binding affinity by several folds. Interestingly, BCL-2 mutants maintain the ability to bind and sequester BIM, thus preserving their fundamental role in the regulation of apoptotic balance.89 In other

Figure 5. Integrated precision medicine for ‘highly complex’ hematologic malignancies. (A) Classification of hematologic malignancies based on their molecular/functional complexity and treatment requirements. In chronic myeloid leukemia, the prototype of ‘purely genomic’ malignancies, a single genomic aberration drives disease biology and treatment modalities. In chronic lymphocytic leukemia, a model for ‘mostly functional’ malignancies, a few functional pathways (Bcell receptor signaling, BCL-2 dependence) sustain leukemic growth and have proven to be the best drug targets so far. In acute myeloid leukemia and other ‘highly complex’ malignancies, multiple driver mutations and functional oncogenic pathways co-occur to promote cancer growth and escape treatments. This category might be approached with an integrated precision medicine strategy. (B) Integrated precision medicine is based on interrogation of different static (i.e., microenvironment, surfaceome, genomics) and functional (i.e., anti-apoptotic and metabolic dependencies, signaling pathways, drug sensitivity) domains through dedicated assays (italics). Each of these assays will provide information about different tumor-specific vulnerabilities (e.g., BH3 profiling might highlight BCL-2 dependence, extracellular flux analysis might indicate oxidative phosphorylation utilization, microenvironment analysis might demonstrate CTLA-4-based interactions). Eventually, a combination of drugs targeting vulnerabilities from different domains is suggested, with potential benefits against ‘highly complex’ malignancies. CML: chronic myeloid leukemia; APL: acute promyelocytic leukemia; HCL: hairy cell leukemia; CNL: chronic neutrophilic leukemia; AML: acute myeloid leukemia; DLBCL: diffuse large B-cell lymphoma; ALL: acute lymphoblastic leukemia; MM: multiple myeloma; T-NHL: T-cell non-Hodgkin lymphoma; OxPHOS: oxidative phosphorylation.

798

haematologica | 2022; 107(4)

The anti-apoptotic landscape of blood cancers

cases, focal amplification of MCL1 or larger chromosomal gains at 1q drive venetoclax resistance. In this context, MCL-1 overexpression takes control of the anti-apoptotic dependencies at the expense of BCL-2.92 A recent work indicates that also BAX variants, including missense, nonsense, frameshift, and splice site mutations, emerge in AML progressing on venetoclax.100 In AML cell lines, BAX but not BAK1 loss is associated with resistance to BCL-2 and MCL-1 antagonists.100 A mutation at BAX c.370 was detected in a case of mantle cell lymphoma that relapsed on venetoclax.101 In contrast to BCL2 mutations, BAX alterations function as a more downstream resistance mechanism potentially affecting sensitivity to a wider range of BH3 mimetics, with important implications when MCL-1 and BCL-xL inhibitors become available for clinical practice. Mutations in genes that do not encode BCL-2 family members have also been implicated in resistance mechanisms. Although TP53-disrupted AML and CLL cells are sensitive to acute venetoclax treatment, they are capable of escaping chronic BCL-2 inhibition, partly because of an increased threshold for BAK/BAX activation.102 Moreover, lack of TP53 activity reduces the transcription of the pro-apoptotic genes PUMA and NOXA, potentially heightening the apoptotic threshold and impairing the efficacy of individual BH3 mimetics when used at suboptimal doses or over long periods of time.102 In AML, KRAS and PTPN11 mutations also decrease sensitivity to venetoclax. Mutant KRAS downregulates BCL-2 and BAX, while it upregulates MCL-1 and BCL2A1, possibly through the activation of the NFkB pathway. Similarly, mutant PTPN11 increases the expression of BCL-xL, MCL-1 and its phosphorylated form, potentially targetable by the correspondent inhibitors.103

Anti-apoptotic profiles at the forefront of integrated precision medicine Until very recently, the paradigm of precision medicine in clinical oncology consisted in matching the right drug with the right patient based on a tumor’s genomic signature. Results from the NCI-MATCH trial indicated, however, that only 37.6% of enrolled patients had actionable alterations and only 17.8% were assigned to a treatment arm.104 Furthermore, resistance-conferring tumor mutations were found in 71.3% of specimens, thus lowering the chance of meaningful responses.104 Overall, the cooccurrence of multiple driver mutations, the poor ability to predict which mutation is a driver and which is a mere bystander, the complexity of resistance mechanisms, and the paucity of available drugs compared to the variety of genomic alterations limit the success rate of genomicdriven precision medicine. The emergence of mitochondrial apoptosis as a cancer vulnerability highlights that successful targets can be found outside genomic alterations, and that genomic alterations not always predict response to mitochondrial targeting. Indeed, BCL-2 rearrangements or expression level do not always correlate with venetoclax response. By contrast, functional assays were able to identify tumors that were more likely to respond to BCL-2 inhibition in the clinic.20,49 Moreover, in the context of CLL, results of BH3 profiling correlated with lymphocyte count reduction upon venetoclax initiation in vivo.61 Although the reliability of these approaches haematologica | 2022; 107(4)

in predicting clinical endpoints is still under investigation (e.g., NCT03943342, NCT03214562, NCT03709758), it looks clear that targeting functional cell biology domains (e.g., mitochondrial apoptosis, B-cell receptor signaling, immune interactions), in addition to the static mutational repertoire, provides therapeutic benefits in hematology. Nevertheless, there should be awareness that mitochondrial apoptosis is only one of the functional domains of cancer cell biology and, for most hematologic malignancies, mitochondrial targeting alone does not provide deep and durable remissions. Blood cancers might be currently classified into three categories based on their underpinnings and treatment requirements. The rare ‘purely genomic’ malignancies have a single genetic abnormality that almost entirely sustains the neoplastic growth. In this case, genomic-driven precision medicine is a highly effective therapeutic strategy and has already gained success. CML is the prototype of these diseases that are effectively treated with molecules targeting their genetic hallmarks.105 The ‘mostly functional’ malignancies, such as CLL, are particularly vulnerable to the targeting of selected oncogenic pathways that have been discovered through cell biology experiments rather than mutational analyses.106 While CLL cells have several mutations across their genome, the targets of the most effective drugs, such as ibrutinib and venetoclax, are never mutated.106 In this category, knowledge built on perturbation of live cells proved perhaps more useful than DNA sequencing in terms of therapeutic applications, and targeting only one functional domain at a time yields considerable results.40 The ‘highly complex’ malignancies, such as AML and DLBCL, are those in which different genetic drivers frequently co-occur, and multiple functional domains simultaneously sustain cancer cell survival.107 In this context, single-agent treatments rarely provide impressive results due to intrinsic and adaptive resistance. An integrated precision medicine approach encompassing genetic and functional testing might be needed to improve the outcome of this category, especially in the R/R setting in which tumor heterogeneity is further amplified.107 As illustrated in Figure 5, analysis on an individual basis of anti-apoptotic dependencies together with complementary static and functional measurements might help to reach this goal in the future. A prerequisite for this approach will be to design clinically applicable ex vivo assays, each with the ability to interrogate one specific domain of cancer cell biology. Because cancer cells usually maintain BAX and BAK expression, and hence are susceptible to the restoration of mitochondrial apoptosis, rational targeting of extra-mitochondrial vulnerabilities through integrated precision medicine might eventually increase apoptotic priming and enhance the effectiveness of concomitant BCL-2 family antagonism.

Concluding remarks The clinical success of venetoclax in CLL and AML generated considerable enthusiasm on targeting apoptosis in hematology. Basic discoveries in the field will be further rewarded with the possible clinical introduction of MCL1 and BCL-xL inhibitors, which are currently under investigation. Different methods have been established to derive tumor-specific anti-apoptotic dependencies, with 799

I. Ferrarini et al.

potential clinical applications. In particular, functional precision medicine platforms such as BH3 profiling and the BH3-mimetic toolkit have proven promising to prioritize apoptosis-targeting agents in a clinically appropriate timeframe. In several studies, they predicted clinical results more accurately than the expression of the BCL-2 family proteins or other static measurements. Moreover, these approaches are somehow breaking the paradigm of precision medicine as an omics-based concept, and pave the way for novel companion diagnostic assays based on ex vivo perturbation of live cells. To pursue this path, clinical and technical efforts will be needed to obtain adequate amounts of live cancer cells from patients, and to standardize ex vivo protocols aiming at minimizing interlaboratory variability. Despite the undoubted advantages that BH3 domain-related pharmacology has been provid-

References 1. Carneiro BA, El-Deiry WS. Targeting apoptosis in cancer therapy. Nat Rev Clin Oncol. 2020;17(7):395-417. 2. Galluzzi L, Vitale I, Aaronson SA, et al. Molecular mechanisms of cell death: recommendations of the Nomenclature Committee on Cell Death 2018. Cell Death Differ. 2018;25(3):486-541. 3. Tummers B, Green DR. Caspase-8: regulating life and death. Immunol Rev. 2017;277(1):76-89. 4. Kale J, Osterlund EJ, Andrews DW. BCL-2 family proteins: changing partners in the dance towards death. Cell Death Differ. 2018;25(1):65-80. 5. Hamouda MA, Jacquel A, Robert G, et al. BCL-B (BCL2L10) is overexpressed in patients suffering from multiple myeloma (MM) and drives an MM-like disease in transgenic mice. J Exp Med. 2016;213(9): 1705-1722. 6. Letai A, Bassik MC, Walensky LD, Sorcinelli MD, Weiler S, Korsmeyer SJ. Distinct BH3 domains either sensitize or activate mitochondrial apoptosis, serving as prototype cancer therapeutics. Cancer Cell. 2002;2(3): 183-192. 7. Moldoveanu T, Czabotar PE. BAX, BAK, and BOK: a coming of age for the BCL-2 family effector proteins. Cold Spring Harb Perspect Biol. 2020;12(4):a036319. 8. Kim H, Tu HC, Ren D, et al. Stepwise activation of BAX and BAK by tBID, BIM, and PUMA initiates mitochondrial apoptosis. Mol Cell. 2009;36(3):487-499. 9. Devan J, Janikova A, Mraz M. New concepts in follicular lymphoma biology: from BCL2 to epigenetic regulators and non-coding RNAs. Semin Oncol. 2018;45(5-6):291-302. 10. Riedell PA, Smith SM. Double hit and double expressors in lymphoma: definition and treatment. Cancer. 2018;124(24):4622-4632. 11. Pekarsky Y, Balatti V, Croce CM. BCL2 and miR-15/16: from gene discovery to treatment. Cell Death Differ. 2018;25(1):21-26. 12. Slomp A, Moesbergen LM, Gong JN, et al. Multiple myeloma with 1q21 amplification is highly sensitive to MCL-1 targeting. Blood Adv. 2019;3(24):4202-4214. 13. Liu T, Lam V, Thieme E, et al. Pharmacologic targeting of Mcl-1 induces mitochondrial dysfunction and apoptosis in B-cell lymphoma cells in a TP53- and BAX-dependent manner. Clin Cancer Res. 2021;27(17):49104922.

800

ing for some hematologic malignancies, it is clear that a multitude of other genetic and functional dependencies exists in complex cancers. In such cases, different angles of cell biology need to be explored simultaneously to instruct more effective combination strategies. Disclosures No conflicts of interest to disclose. Contributions IF, AR, and CV conceived and wrote the manuscript, and reviewed the literature. Acknowledgments This research received no external funding. The figures were created with Biorender.com.

14. Bachmann PS, Piazza RG, Janes ME, et al. Epigenetic silencing of BIM in glucocorticoid poor-responsive pediatric acute lymphoblastic leukemia, and its reversal by histone deacetylase inhibition. Blood. 2010;116(16): 3013-3022. 15. Wang JD, Katz SG, Morgan EA, Yang DT, Pan X, Xu ML. Proapoptotic protein BIM as a novel prognostic marker in mantle cell lymphoma. Hum Pathol. 2019;93:54-64. 16. Ni Chonghaile T, Sarosiek KA, Vo TT, et al. Pretreatment mitochondrial priming correlates with clinical response to cytotoxic chemotherapy. Science. 2011;334(6059): 1129-1133. 17. Cerella C, Dicato M, Diederich M. BH3 mimetics in AML therapy: death and beyond? Trends Pharmacol Sci. 2020;41(11): 793-814. 18. Del Gaizo Moore V, Brown JR, Certo M, Love TM, Novina CD, Letai A. Chronic lymphocytic leukemia requires BCL2 to sequester prodeath BIM, explaining sensitivity to BCL2 antagonist ABT-737. J Clin Invest. 2007;117(1):112-121. 19. de Jong MRW, Langendonk M, Reitsma B, et al. Heterogeneous pattern of dependence on anti-apoptotic BCL-2 family proteins upon CHOP treatment in diffuse large B-cell lymphoma. Int J Mol Sci. 2019;20(23):6036. 20. Touzeau C, Ryan J, Guerriero J, et al. BH3 profiling identifies heterogeneous dependency on Bcl-2 family members in multiple myeloma and predicts sensitivity to BH3 mimetics. Leukemia. 2016;30(3):761-764. 21. Grabow S, Delbridge AR, Aubrey BJ, Vandenberg CJ, Strasser A. Loss of a single Mcl-1 allele inhibits MYC-driven lymphomagenesis by sensitizing pro-B cells to apoptosis. Cell Rep. 2016;14(10):2337-2347. 22. Adams CM, Kim AS, Mitra R, Choi JK, Gong JZ, Eischen CM. BCL-W has a fundamental role in B cell survival and lymphomagenesis. J Clin Invest. 2017;127(2):635-650. 23. Dengler MA, Teh CE, Thijssen R, et al. Potent efficacy of MCL-1 inhibitor-based therapies in preclinical models of mantle cell lymphoma. Oncogene. 2020;39(9):20092023. 24. Kelly GL, Grabow S, Glaser SP, et al. Targeting of MCL-1 kills MYC-driven mouse and human lymphomas even when they bear mutations in p53. Genes Dev. 2014;28(1):58-70. 25. Bhatt S, Pioso MS, Olesinski EA, et al. Reduced mitochondrial apoptotic priming drives resistance to BH3 mimetics in acute myeloid leukemia. Cancer Cell. 2020;38(6):

872-890. 26. 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. 27. Ewald L, Dittmann J, Vogler M, Fulda S. Side-by-side comparison of BH3-mimetics identifies MCL-1 as a key therapeutic target in AML. Cell Death Dis. 2019;10(12):917. 28. Diepstraten ST, Chang C, Tai L, et al. BCLW is dispensable for the sustained survival of select Burkitt lymphoma and diffuse large B-cell lymphoma cell lines. Blood Adv. 2020;4(2):356-366. 29. Ryan J, Montero J, Rocco J, Letai A. iBH3: simple, fixable BH3 profiling to determine apoptotic priming in primary tissue by flow cytometry. Biol Chem. 2016;397(7):671-678. 30. Gomez-Bougie P, Maiga S, Tessoulin B, et al. BH3-mimetic toolkit guides the respective use of BCL2 and MCL1 BH3-mimetics in myeloma treatment. Blood. 2018;132(25): 2656-2669. 31. Samson DJ, Seidenfeld J, Ziegler K, Aronson N. Chemotherapy sensitivity and resistance assays: a systematic review. J Clin Oncol. 2004;22(17):3618-3630. 32. Shah N, Oseth L, Tran H, Hirsch B, LeBien TW. Clonal variation in the B-lineage acute lymphoblastic leukemia response to multiple cytokines and bone marrow stromal cells. Cancer Res. 2001;61(13):5268-5274. 33. 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. 34. Del Gaizo Moore V, Letai A. BH3 profiling-measuring integrated function of the mitochondrial apoptotic pathway to predict cell fate decisions. Cancer Lett. 2013;332(2):202205. 35. Ghavami S, Hashemi M, Ande SR, et al. Apoptosis and cancer: mutations within caspase genes. J Med Genet. 2009;46(8):497510. 36. Montero J, Sarosiek KA, DeAngelo JD, et al. Drug-induced death signaling strategy rapidly predicts cancer response to chemotherapy. Cell. 2015;160(5):977-989. 37. Montero J, Letai A. Dynamic BH3 profilingpoking cancer cells with a stick. Mol Cell Oncol. 2016;3(3):e1040144. 38. Deng J, Isik E, Fernandes SM, Brown JR, Letai A, Davids MS. Bruton's tyrosine kinase inhibition increases BCL-2 dependence and enhances sensitivity to venetoclax in chronic

haematologica | 2022; 107(4)

The anti-apoptotic landscape of blood cancers

lymphocytic leukemia. Leukemia. 2017;31 (10):2075-2084. 39. Herbaux C, Kornauth C, Poulain S, et al. BH3 profiling identifies ruxolitinib as a promising partner for venetoclax to treat Tcell prolymphocytic leukemia. Blood. 2021;137(25):3495-3506. 40. 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. 41. Konopleva M, Pollyea DA, Potluri J, et al. Efficacy and biological correlates of response in a phase II study of venetoclax monotherapy in patients with acute myelogenous leukemia. Cancer Discov. 2016;6(10):11061117. 42. Roberts AW, Wei AH, Huang DCS. BCL2 and MCL1 inhibitors for hematologic malignancies. Blood. 2021;138(13):1120-1136. 43. Pan R, Hogdal LJ, Benito JM, et al. Selective BCL-2 inhibition by ABT-199 causes on-target cell death in acute myeloid leukemia. Cancer Discov. 2014;4(3):362-375. 44. Jilg S, Reidel V, Müller-Thomas C, et al. Blockade of BCL-2 proteins efficiently induces apoptosis in progenitor cells of highrisk myelodysplastic syndromes patients. Leukemia. 2016;30(1):112-123. 45. Kadia TM, Reville PK, Borthakur G, et al. Venetoclax plus intensive chemotherapy with cladribine, idarubicin, and cytarabine in patients with newly diagnosed acute myeloid leukaemia or high-risk myelodysplastic syndrome: a cohort from a singlecentre, single-arm, phase 2 trial. Lancet Haematol. 2021;8(8):e552-e561. 46. 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. 47. 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. 48. Del Gaizo Moore V, Schlis KD, Sallan SE, Armstrong SA, Letai A. BCL-2 dependence and ABT-737 sensitivity in acute lymphoblastic leukemia. Blood. 2008;111(4): 2300-2309. 49. Seyfried F, Demir S, Horl RL, et al. Prediction of venetoclax activity in precursor B-ALL by functional assessment of apoptosis signaling. Cell Death Dis. 2019;10(8):571. 50. Alford SE, Kothari A, Loeff FC, et al. BH3 inhibitor sensitivity and Bcl-2 dependence in primary acute lymphoblastic leukemia cells. Cancer Res. 2015;75(7):1366-1375. 51. Khaw SL, Suryani S, Evans K, et al. Venetoclax responses of pediatric ALL xenografts reveal sensitivity of MLLrearranged leukemia. Blood. 2016;128(10): 1382-1395. 52. Pullarkat VA, Lacayo NJ, Jabbour E, et al. Venetoclax and navitoclax in combination with chemotherapy in patients with relapsed or refractory acute lymphoblastic leukemia and lymphoblastic lymphoma. Cancer Discov. 2021;11(6):1440-1453. 53. Benito JM, Godfrey L, Kojima K, et al. MLLrearranged acute lymphoblastic leukemias activate BCL-2 through H3K79 methylation and are sensitive to the BCL-2-specific antagonist ABT-199. Cell Rep. 2015;13(12): 2715-2727. 54. Chonghaile TN, Roderick JE, Glenfield C, et al. Maturation stage of T-cell acute lymphoblastic leukemia determines BCL-2 versus BCL-XL dependence and sensitivity to

haematologica | 2022; 107(4)

ABT-199. Cancer Discov. 2014;4(9):10741087. 55. Di Grande A, Peirs S, Donovan PD, et al. The spleen as a sanctuary site for residual leukemic cells following ABT-199 monotherapy in ETP-ALL. Blood Adv. 2021;5(7):1963-1976. 56. Moia R, Patriarca A, Schipani M, Gaidano G. The biology of chronic lymphocytic leukemia: diagnostic and prognostic implications. Cancer J. 2021;27(4):266-274. 57. Gohil SH, Wu CJ. Dissecting CLL through high-dimensional single-cell technologies. Blood. 2019;133(13):1446-1456. 58. Mavridou D, Psatha K, Aivaliotis M. Proteomics and drug repurposing in CLL towards precision medicine. Cancers (Basel). 2021;13(14):3391. 59. Hanada M, Delia D, Aiello A, Stadtmauer E, Reed JC. Bcl-2 gene hypomethylation and high-level expression in B-cell chronic lymphocytic leukemia. Blood. 1993;82(6):18201828. 60. Fang H, Reichard KK, Rabe KG, et al. IGH translocations in chronic lymphocytic leukemia: clinicopathologic features and clinical outcomes. Am J Hematol. 2019;94(3):338-345. 61. Anderson MA, Deng J, Seymour JF, et al. The BCL2 selective inhibitor venetoclax induces rapid onset apoptosis of CLL cells in patients via a TP53-independent mechanism. Blood. 2016;127(25):3215-3224. 62. Davids MS, Deng J, Wiestner A, et al. Decreased mitochondrial apoptotic priming underlies stroma-mediated treatment resistance in chronic lymphocytic leukemia. Blood. 2012;120(17):3501-3509. 63. Hillmen P, Rawstron AC, Brock K, et al. Ibrutinib plus venetoclax in relapsed/refractory chronic lymphocytic leukemia: the CLARITY study. J Clin Oncol. 2019;37(30): 2722-2729. 64. Morin RD, Arthur SE, Hodson DJ. Molecular profiling in diffuse large B-cell lymphoma: why so many types of subtypes? Br J Haematol. 2022;196(4):814829. 65. Rys RN, Wever CM, Geoffrion D, et al. Apoptotic blocks in primary non-Hodgkin B cell lymphomas identified by BH3 profiling. Cancers (Basel). 2021;13(5):1002. 66. Wenzel SS, Grau M, Mavis C, et al. MCL1 is deregulated in subgroups of diffuse large Bcell lymphoma. Leukemia. 2013;27(6):13811390. 67. Klanova M, Klener P. BCL-2 Proteins in pathogenesis and therapy of B-cell nonHodgkin lymphomas. Cancers (Basel). 2020;12(4):938. 68. Bojarczuk K, Wienand K, Ryan JA, et al. Targeted inhibition of PI3Kalpha/delta is synergistic with BCL-2 blockade in genetically defined subtypes of DLBCL. Blood. 2019;133(1):70-80. 69. Davids MS, Roberts AW, Seymour JF, et al. Phase I first-in-human study of venetoclax in patients with relapsed or refractory nonHodgkin lymphoma. J Clin Oncol. 2017;35(8):826-833. 70. Davids MS, Roberts AW, Kenkre VP, et al. Long-term follow-up of patients with relapsed or refractory non-Hodgkin lymphoma treated with venetoclax in a phase I, first-in-human study. Clin Cancer Res. 2021;27(17):4690-4695. 71. Serrat N, Guerrero-Hernandez M, MatasCespedes A, et al. PI3Kdelta inhibition reshapes follicular lymphoma-immune microenvironment cross talk and unleashes the activity of venetoclax. Blood Adv.

2020;4(17):4217-4231. 72. Punnoose EA, Leverson JD, Peale F, et al. Expression profile of BCL-2, BCL-XL, and MCL-1 predicts pharmacological response to the BCL-2 selective antagonist venetoclax in multiple myeloma models. Mol Cancer Ther. 2016;15(5):1132-1144. 73. Hanamura I, Stewart JP, Huang Y, et al. Frequent gain of chromosome band 1q21 in plasma-cell dyscrasias detected by fluorescence in situ hybridization: incidence increases from MGUS to relapsed myeloma and is related to prognosis and disease progression following tandem stem-cell transplantation. Blood. 2006;108(5):1724-1732. 74. Wuilleme-Toumi S, Robillard N, Gomez P, et al. Mcl-1 is overexpressed in multiple myeloma and associated with relapse and shorter survival. Leukemia. 2005;19(7):12481252. 75. Gupta VA, Matulis SM, Conage-Pough JE, et al. Bone marrow microenvironment-derived signals induce Mcl-1 dependence in multiple myeloma. Blood. 2017;129(14):1969-1979. 76. Wang M, Wu D, Liu P, Deng J. Silence of MCL-1 upstream signaling by shRNA abrogates multiple myeloma growth. Exp Hematol Oncol. 2014;3(1):27. 77. Algarin EM, Diaz-Tejedor A, Mogollon P, et al. Preclinical evaluation of the simultaneous inhibition of MCL-1 and BCL-2 with the combination of S63845 and venetoclax in multiple myeloma. Haematologica. 2020; 105(3):e116-e120. 78. Touzeau C, Maciag P, Amiot M, Moreau P. Targeting Bcl-2 for the treatment of multiple myeloma. Leukemia. 2018;32(9):1899-1907. 79. Kumar S, Kaufman JL, Gasparetto C, et al. Efficacy of venetoclax as targeted therapy for relapsed/refractory t(11;14) multiple myeloma. Blood. 2017;130(22):2401-2409. 80. Koch R, Christie AL, Crombie JL, et al. Biomarker-driven strategy for MCL1 inhibition in T-cell lymphomas. Blood. 2019;133 (6):566-575. 81. Spinner S, Crispatzu G, Yi JH, et al. Re-activation of mitochondrial apoptosis inhibits T-cell lymphoma survival and treatment resistance. Leukemia. 2016;30(7):1520-1530. 82. He Y, Koch R, Budamagunta V, et al. DT2216-a Bcl-xL-specific degrader is highly active against Bcl-xL-dependent T cell lymphomas. J Hematol Oncol. 2020;13(1):95. 83. Smith VM, Lomas O, Constantine D, et al. Dual dependence on BCL2 and MCL1 in Tcell prolymphocytic leukemia. Blood Adv. 2020;4(3):525-529. 84. Kornauth C, Herbaux C, Boidol B, et al. Rationale for the combination of venetoclax and ibrutinib in T-prolymphocytic leukemia. Haematologica. 2021;106(8):2251-2256. 85. Adams CM, Mitra R, Vogel AN, Liu J, Gong JZ, Eischen CM. Targeting BCL-W and BCLXL as a therapeutic strategy for Hodgkin lymphoma. Leukemia. 2020;34(3):947-952. 86. Montero J, Stephansky J, Cai T, et al. Blastic plasmacytoid dendritic cell neoplasm is dependent on BCL2 and sensitive to venetoclax. Cancer Discov. 2017;7(2):156-164. 87. Beziat G, Ysebaert L, Gaudin C, Steinmeyer Z, Balardy L. Venetoclax to treat relapsed blastic plasmacytoid dendritic cell neoplasm: a case-report and review of literature. Leuk Res. 2019;85:106199. 88. Agha ME, Monaghan SA, Swerdlow SH. Venetoclax in a patient with a blastic plasmacytoid dendritic-cell neoplasm. N Engl J Med. 2018;379(15):1479-1481. 89. Blombery P, Thompson ER, Nguyen T, et al. Multiple BCL2 mutations cooccurring with Gly101Val emerge in chronic lymphocytic

801

I. Ferrarini et al. leukemia progression on venetoclax. Blood. 2020;135(10):773-777. 90. Yu X, Munoz-Sagredo L, Streule K, et al. CD44 loss of function sensitizes AML cells to the BCL-2 inhibitor venetoclax by decreasing CXCL12-driven survival cues. Blood. 2021;138(12):1067-1080. 91. Haselager M, Thijssen R, West C, et al. Regulation of Bcl-XL by non-canonical NFkappaB in the context of CD40-induced drug resistance in CLL. Cell Death Differ. 2021;28(5):1658-1668. 92. Guieze R, Liu VM, Rosebrock D, et al. Mitochondrial reprogramming underlies resistance to BCL-2 inhibition in lymphoid malignancies. Cancer Cell. 2019;36(4):369384. 93. Bixby D, Talpaz M. Mechanisms of resistance to tyrosine kinase inhibitors in chronic myeloid leukemia and recent therapeutic strategies to overcome resistance. Hematology Am Soc Hematol Educ Program. 2009;461-476. 94. 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. 95. Stevens BM, Jones CL, Pollyea DA, et al.

802

Fatty acid metabolism underlies venetoclax resistance in acute myeloid leukemia stem cells. Nat Cancer. 2020;1(12):1176-1187. 96. 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. 97. Jones CL, Stevens BM, Pollyea DA, et al. Nicotinamide metabolism mediates resistance to venetoclax in relapsed acute myeloid leukemia stem cells. Cell Stem Cell. 2020;27(5):748-764. 98. Chen X, Glytsou C, Zhou H, et al. Targeting mitochondrial structure sensitizes acute myeloid leukemia to venetoclax treatment. Cancer Discov. 2019;9(7):890-909. 99. Sharon D, Cathelin S, Mirali S, et al. Inhibition of mitochondrial translation overcomes venetoclax resistance in AML through activation of the integrated stress response. Sci Transl Med. 2019;11(516): eaax2863. 100. Moujalled DM, Brown FC, Pomilio G, et al. Acquired mutations in BAX confer resistance to BH3 mimetics in acute myeloid leukemia. Blood. 2020;136(Suppl 1):7-8. 101. Thompson ER, Nguyen T, Kankanige Y, et al. High clonal complexity of resistance mechanisms occurring at progression after

single-agent targeted therapy strategies in chronic lymphocytic leukemia. Blood. 2020;136(Suppl 1):15-16. 102. Thijssen R, Diepstraten ST, Moujalled D, et al. Intact TP-53 function is essential for sustaining durable responses to BH3-mimetic drugs in leukemias. Blood. 2021;137(20): 2721-2735. 103. Zhang H, Nakauchi Y, Kohnke T, et al. Integrated analysis of patient samples identifies biomarkers for venetoclax efficacy and combination strategies in acute myeloid leukemia. Nat Cancer. 2020;1(8):826-839. 104. Flaherty KT, Gray RJ, Chen AP, et al. Molecular landscape and actionable alterations in a genomically guided cancer clinical trial: National Cancer Institute Molecular Analysis for Therapy Choice (NCIMATCH). J Clin Oncol. 2020;38(33):38833894. 105. Hehlmann R. Chronic myeloid leukemia in 2020. Hemasphere. 2020;4(5):e468. 106. Letai A. Functional precision cancer medicine-moving beyond pure genomics. Nat Med. 2017;23(9):1028-1035. 107. Adashek JJ, Subbiah V, Kurzrock R. From tissue-agnostic to N-of-one therapies: (r)evolution of the precision paradigm. Trends Cancer. 2021;7(1):15-28.

haematologica | 2022; 107(4)

ARTICLE

Acute Lymphoblastic Leukemia

LAMP-5 is an essential inflammatory-signaling regulator and novel immunotherapy target for mixed lineage leukemia-rearranged acute leukemia

Ferrata Storti Foundation

Gabriel Gracia-Maldonado,1,2 Jason Clark,1,2 Matthew Burwinkel,1,2 Brenay Greenslade,1,3 Mark Wunderlich,1,4 Nathan Salomonis,5,7 Dario Leone,6 Evelina Gatti,6 Philippe Pierre,6 Ashish R. Kumar1,2,7 and Lynn H. Lee1,3,7 1 Cancer and Blood Diseases Institute, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH, USA; 2Division of Bone Marrow Transplantation and Immune Deficiency, Cincinnati, Children’s Hospital Medical Center, Cincinnati, OH, USA; 3Division of Oncology, Cincinnati Children's Hospital Medical Center, Cincinnati, OH, USA; 4Division of Experimental Hematology and Cancer Biology, Cincinnati, Children’s Hospital Medical Center, Cincinnati, OH, USA; 5Division of Biomedical Informatics, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH, USA; 6Aix Marseille Université, CNRS, INSERM, Centre d'Immunologie de Marseille-Luminy (CIML), Marseille, France and 7Department of Pediatrics, University of Cincinnati School of Medicine, Cincinnati, OH, USA

Haematologica 2022 Volume 107(3):803-815

ABSTRACT

lthough great advances have been made in understanding the pathobiology of mixed lineage leukemia-rearranged (MLL-r) leukemias, therapies for this leukemia have remained limited, and clinical outcomes remain bleak. In order to identify novel targets for immunotherapy treatments, we compiled a lineage-independent MLL-r leukemia gene signature using publicly available data sets. Data from large leukemia repositories were filtered through the in silico human surfaceome, providing a list of highly predicted cell surface proteins overexpressed in MLL-r leukemias. LAMP5, a lysosomal associated membrane protein, is expressed highly and specifically in MLL-r leukemia. We found that LAMP5 is a direct target of the oncogenic MLL-fusion protein. LAMP5 depletion significantly inhibited leukemia cell growth in vitro and in vivo. Functional studies showed that LAMP-5 is a novel modulator of innate-immune pathways in MLL-r leukemias. Downregulation of LAMP5 led to inhibition of NF-kB signaling and increased activation of type-1 interferon signaling downstream of Toll-like receptor/interleukin 1 receptor activation. These effects were attributable to the critical role of LAMP-5 in transferring the signal flux from interferon signaling endosomes to pro-inflammatory signaling endosomes. Depletion of IRF7 was able to partially rescue the cell growth inhibition upon LAMP5 downregulation. Lastly, LAMP-5 was readily detected on the surface of MLL-r leukemia cells. Targeting surface LAMP-5 using an antibody-drug conjugate leads to significant cell viability decrease specifically in MLL-r leukemias. Overall, based on the limited expression throughout human tissues, we postulate that LAMP-5 could potentially serve as an immunotherapeutic target with a wide therapeutic window to treat MLL-r leukemias.

Introduction Translocations in the mixed lineage leukemia (MLL) gene account for 10% of all human leukemias and are associated with pediatric, adult, and therapy-related cases. In infants, around 80% of acute lymphoid leukemia (ALL) and 35%-50% of acute myeloid leukemia (AML) cases carry a translocation in the MLL gene.1 However, despite improvements in conventional chemotherapy treatments for leukemia, patients with MLL-rearranged leukemia (MLL-r) have a poor response to treatment

haematologica | 2022; 107(4)

Correspondence: LYNN H. LEE lynn.lee@cchmc.org ASHISH R. KUMAR ashish.kumar@cchmc.org Received: April 30, 2020. Accepted: April 2, 2021. Pre-published: April 29, 2021. https://doi.org/10.3324/haematol.2020.257451

803

G. Gracia-Maldonado et al.

and poor prognosis.2,3 Immunotherapy strategies have proven effective in multiple blood cancers, mainly targeting lineage-specific proteins like CD19 (blinatumomab, tisagenlecleucel) and CD33 (gemtuzumab), abundantly expressed in ALL and AML patients, respectively.4 However, mounting evidence in recent clinical trials and case reports have shown that patients with MLL-rearrangements frequently relapse after treatment with CD19 immunotherapies, arising as AML or mixed phenotype acute leukemia (MPAL).5-12 The exact mechanism of lineage switch induced by CD19 immunotherapies is still unclear. One approach to overcome the lineage switching is to develop MLL-r specific immunotherapies targeting cell surface proteins essential for the survival of MLL-r leukemias. Recently, NG2/CSPG4 and CD133/PROM1 have been shown to be promising MLL-r specific immunotherapy targets, however, these targets are restricted to lymphoid lineage, increasing the potential for lineage switching within the leukemic population.13-15 Gene-expression profiling based on underlying cytogenetic mutations is one way to identify proteins that are overexpressed and thus might be essential for the propagation of the specific leukemia.16,17 Both AML and ALL with MLL-rearrangements share a common gene signature that is distinct from that of MLL-germline (MLL-G) leukemias.18 Most of the well-studied and validated MLLr gene targets however are DNA binding proteins like the HOXA gene cluster and its co-factor MEIS1,19,20 which are not suitable targets for immunotherapy. In several of the published gene-expression studies, we found LAMP5 significantly and specifically overexpressed in MLL-r leukemias.18,21–23 LAMP-5 is a member of the lysosome-associated membrane protein (LAMP) family. In contrast to other LAMP proteins which show widespread expression, Lamp5 expression in mice is confined to several regions of the postnatal brain. In neurons, the protein was found to recycle between the plasma membrane and a non-classical endosomal vesicle.24–26 In humans, aside from its conserved expression in the brain, LAMP5 is specifically expressed in plasmacytoid dendritic cells (pDC).27 Upon activation of pDC, LAMP-5 aids in the transport of Toll-like receptor 9 (TLR9) from early endosomal to lysosomal signaling vesicles, thereby regulating type 1 interferon (IFN-1) and pro-inflammatory signaling respectively, downstream of TLR9 activation.28 Importantly, results of in silico modeling predict LAMP-5 as a cell surface protein.29 In this report, we demonstrate LAMP5 as being highly expressed and essential for MLLr leukemias through the regulation of innate immune signaling and describe its potential as a target for MLL-r specific immunotherapy.

Animal experiments All animal experiments were carried out in accordance with the guidelines of the Institutional Animal Care and Use Committee (IACUC). For xenograft experiments with MV4;11 and MLL-AF10 primary patient cells, immunocompromised NOD-Rag1null IL2rgnull (NRG) (Jackson Laboratories, stock no. 007799) recipient mice were conditioned with busulfan and transplanted 24 hours later. In xenograft experiments bone marrow samples were collected 4 weeks after transplantation as well as when signs of leukemia were present; bone marrow aspirates were analyzed via flow cytometry for the presence of human CD45+ cells and the presence of short hairpin RNA (shRNA)-transduced Venus+ cells

Retroviral and lentiviral transductions Retroviral and lentiviral supernatants were generated by transfection of HEK293T cells using the FuGENE 6 reagent (Promega) or Lipofectamine 3000 (Thermo Fisher Scientific) according to the manufacturer’s recommendations. All lentiviral shRNA constructs were purchased from Millipore Sigma. Cells transduced with constructs containing a fluorescent marker (Venus) were sorted 4-5 days after transduction by using MoFlo XDP (Beckman Coulter), FACSAria (BD Biosciences), or a SONY SH800S (Sony Biotechnology).

Flow cytometry For apoptosis assays, cells were incubated with allophycocyanin (APC)-conjugated annexin V (BD Biosciences) for 15 minutes at room temperature in 1X annexin V binding buffer (BD Biosciences) followed by staining with 7-aminoactinomycin D (7AAD) (eBioscience). For surface LAMP-5 detection, cells were incubated with 3 µg anti-human LAMP-5 therapeutic antibody overnight and then stained with anti-mouse IgG1-PE. Data were acquired on a BD FACSCanto analyzer and results were analyzed using FlowJo Version 10 (BD Biosciences).

Colony-forming unit assays Transduced human cells were sorted 4-5 days after transduction and were cultured in methylcellulose-containing media (StemCell Technologies, H4434). Colonies were scored 10-14 days after plating.

Real-time- and quantitative real-time polymerase chain reaction Total RNA was extracted from human puromycin-selected or sorted Venus+ cells using the RNeasy Mini kit (QIAGEN). For quantitative real-time polymerase chain reaction (RT-PCR) 5-10 ng of cDNA was analyzed using iTaq Universal SYBR Green Supermix (Bio-Rad) or PowerUP SYBR Green (Thermo Fisher Scientific) in a StepOnePlus RT-PCR machine (Applied Biosystems).

Cell viability Methods Cell lines and primary patient-derived xenograft cells Human leukemia cell lines were maintained in Iscove’s Modified Dulbecco Medium (IMDM) or Roswell Park Memorial Institute (RPMI) 1640 medium supplemented with 10% fetal bovine serum (FBS), 1% penicillin, and 1% streptomycin. Fully de-identified primary cells were obtained from the Cincinnati Children’s Hospital Medical Center Biorepository. Cells were cultured in IMDM supplemented with 20% FBS and 10 ng/mL human cytokines including SCF, FLT3-ligand, thrombopoietin, IL-3, and IL-6.

804

MOLM-13, RS4;11, THP-1, and Kasumi-1 cells were plated at 10,000 cells per well in a 96-well plate. Cells were incubated with LAMP-5 therapeutic antibody (Creative Biolabs) and anti-mouse immunoglobulin G (IgG) Fc-DM1 antibody with non-cleavable linker (Moradec) at 5 ng/uL and 1 ng/uL final concentrations, respectively. In order to measure cell viability CellTiter-Glo® 2.0 Cell Viability Assay (Promega) was used following the manufacturer’s protocol.

Immunofluorescence Cells seeded on alcian blue-treated coverslips were fixed with 3.5% paraformaldehyde and permeabilized with 0.05% saponin. Cells were stained overnight with primary antibodies against

haematologica | 2022; 107(4)

LAMP5 as an essential target of MLL-r leukemias

LAMP-5, MYD88, and LAMP-1. Immunofluorescence and confocal microscopy were performed with a Zeiss LSM580 63x objective and accompanying imaging software.

Statistics The statistical methodology used, and sample sizes are described in the individual Figure legends. t-tests were two-tailed unless otherwise stated. Results are presented as mean ± standard error of the mean (SEM) unless otherwise stated. A two-sided time-stratified Cochran-Mantel-Haenszel was used for the Kaplan-Meier Survival analysis. ROC curves were used to determine the diagnostic utility of LAMP5 mRNA. The sensitivity and specificity were identified at the optimal cutoff point that was chosen at which Youden's index was maximal. A significance level cutoff of 0.05 was used unless otherwise stated. Statistical analysis was performed using GraphPad Prism. More detailed information on the materials and methods used can be found in the Online Supplementary Appendix.

Results LAMP5 is highly expressed in mixed lineage leukemia-rearranged leukemias and is a direct target of the mixed lineage leukemia-fusion protein In order to determine genes that are highly expressed in AML and ALL with MLL-rearrangements, we compared recently published RNA sequencing (RNA-seq) studies that identified differentially expressed genes between MLL-r and MLL-G leukemias in both AML and B-ALL samples30,31(Figure 1A). Twenty-seven genes were commonly overexpressed in MLL-r ALL and AML (Online Supplementary Table S1). Using the in silico human surfaceome tool (http://wlab.ethz.ch/surfaceome/) five of these 27 genes were predicted to be expressed on the cell surface29 (Online Supplementary Table S1). Of the five predicted proteins, LAMP-5 stood out for being present in multiple previous MLL-r leukemia gene expression studies17,18,21–23 (Online Supplementary Figure S1A and B). We further validated the specificity of LAMP5 expression in MLL-r leukemias by analyzing the 1,109 pediatric leukemia patient samples from the St.Jude PeCan Portal which revealed LAMP5 as significantly overexpressed in 92% of ALL and 72% of AML with MLL-rearrangements16 (Figure 1B). In order to determine if LAMP5 expression could discriminate between MLL-r leukemia and MLL-G leukemia patients, we performed a receiving operating curve (ROC) analysis. LAMP5 achieved a statistically significant area under the curve (AUc) score in both the microarray innovations in both the (MILE) (GSE13159) and the St. Jude PeCan datasets, with high sensitivity and specificity at the optimal cutoff points (Figure 1C). Further, a Kaplan-Meier survival analysis of B-ALL and AML patients correlated higher expression of LAMP5 with poor survival (Online Supplementary Figure S2A and B). At the protein level, patient-derived xenograft (PDX) pediatric AML and ALL samples show high expression of LAMP-5 only in the MLL-r samples as compared to MLL-G (Figure 1D). Similar results were seen in human MLL-r AML and ALL cells lines (MOLM13, MV4;11, THP-1, and RS4;11) at the mRNA (Figure 1E) and protein (Figure 1F) levels as compared to MLL-G leukemia cell lines (HL-60, Kasumi-1, K562, REH, and RCH-ACV) and normal human CD34-enriched cord blood cells (CB-CD34+ cells). Translocations of the MLL haematologica | 2022; 107(4)

locus generate MLL fusion proteins (MLL-FP) which activate transcription of downstream target genes.32,33 In order to determine if LAMP5 expression was dependent on the MLL-FP, we transformed CB-CD34+ cells with a retrovirus carrying a tetracycline-repressible MLL-AF9 construct. Treatment of transformed cells with doxycycline led to a simultaneous reduction in the levels of both MLL-AF9 and LAMP5 (Figure 1G). in order to determine if the MLL-FP directly activates the LAMP5 gene locus, we interrogated previously published MLL-FP chromatin immunoprecipitation sequencing (ChIP-seq) datasets derived from the SEM, RS4;11, MV4;11, THP-1, and ML2 cell lines, CD34+ cells transformed with FLAG-MLL-Af4, and primary patient sample.32–36 Almost all cell lines exhibited peaks within the LAMP5 promoter region, suggesting direct binding of the MLL-FP (as evidenced by coincident signal in both MLL and fusion partner ChIP-seq tracks). Additionally, there was accompanying significant enrichment of H3K79me2 and H3K79me3 along the gene body, further supporting our hypothesis that LAMP5 undergoes transcriptional activation in MLL-r leukemia via direct targeting by the MLL-FP complex (Figure 1H). In mice, Lamp5 does not show any expression in blood, as it does in humans (Online Supplemental Figure S3A and B). Furthermore, we did not detect upregulation of Lamp5 in mouse models of MLL-AF9, E2A-HLF, and AML1-ETO leukemia, hence we focused our studies exclusively on human cells (Online Supplemental Figure S3C).

LAMP-5 is required for in vitro and in vivo leukemia cell survival The ideal immunotherapy target should be essential for the survival of MLL-r leukemias. In order to test the functional role of LAMP-5 in MLL-r leukemia, we transduced both MLL-r leukemia (MOLM-13, MV4;11, RS4;11, THP-1) and MLL-G leukemia cells (Kasumi-1 and REH) with lentiviral shRNA vectors targeting LAMP5. We obtained efficient knockdown of LAMP5 with two independent hairpins as compared to non-targeting control (NT) (Online Supplementary Figure S4A). Upon LAMP5 depletion, we observed a significant reduction of cell growth in MLL-r leukemia cell lines (Figure 2A), while no effect was seen in Kasumi-1 and REH (Figure 2B). Additionally, LAMP5 knockdown led to a significant decrease in colony-forming units (CFU) in the MLL-r leukemia cell lines (Figure 2C) suggesting an effect on the clonogenicity of these cells. Furthermore, LAMP5 knockdown led to apoptosis in MLL-r leukemia cells, as evident by a significant increase in annexin V and 7-AAD double-positive staining (Figure 2D; Online Supplementalry Figure S4B). We next sought to determine the role of LAMP-5 in leukemia propagation in vivo. MV4;11 cells were transduced with shLAMP5-2 or NT control followed by transplantation into immunocompromised NOD-Rag1null IL2rgnull (NRG) mice (Figure 2E). In the bone marrow, both groups showed similar human cell engraftment based on human CD45 expression. On the other hand, the transduced Venus+ fraction was significantly reduced in shLAMP5-2 compared to shNT mice 4 weeks after transplantation (Figure 2F, left panels). We repeated this experiment using cells from an AML PDX with MLL-r (MLL-AF10) leukemia. We again observed a significant reduction in the proportion of Venus+ cells with LAMP5 knockdown compared to NT control (Figure 2F, right panels). Overall, these data underscore a critical role for LAMP5 in the growth of MLL-r leukemia cells. 805

G. Gracia-Maldonado et al.

LAMP-5 is required for activation of Toll-like receptor/interleukin 1 receptor signaling in leukemia Acute leukemias exhibiting constitutive activation of innate immune signaling pathways have been characterized as having a pro-inflammatory profile which is required for their survival.37 These physiologic cellular systems involve TLR/IL-1R signaling and culminate in the release of pro-inflammatory cytokines via NF-kB and/or of type I interferons (IFN-1).38 Recent studies reveal heightened activation of NF-kB signaling in MLL-r leukemia compared to other leukemias.39 Furthermore, MLL-r leukemias have been shown to require the TLR/IL1R signaling pathway to survive, through degradation of the wild-type MLL protein, allowing the MLL-FP to bind to its target genes without restriction.40 Recently, Combes

et al. showed that LAMP-5 plays an important role in controlling the subcellular location of TLR9 after activation in human pDC. Upon activation of TLR9, LAMP-5 shuttles TLR9 from the VAMP3+-interferon response factor signaling endosome (IRF-SE), to the LAMP-1+ pro-inflammatory-signaling endosome (PI-SE). This transition of TLR localization in turn acts as a negative regulator of IFN-1 signaling.28 Based on the known role of LAMP-5 in TLR9 localization in pDC, we first examined the localization of intracellular LAMP-5 in MOLM-13 cells. We performed co-staining of MOLM-13 cells with antibodies against LAMP-5, LAMP-1, and myeloid differentiation primary response 88 (MYD88), a scaffold protein that is required for TLR and IL-1R signaling. Confocal microscopy showed that in MOLM-13 leukemia cells, LAMP-5 local-

Figure 1. Continued on following page.

806

haematologica | 2022; 107(4)

LAMP5 as an essential target of MLL-r leukemias

Figure 1. LAMP5 is highly expressed in mixed lineage leukemia-rearranged leukemias and is a direct target of the mixed lineage leukemia-fusion protein. (A) The intersection of published gene expression signatures composed of genes overexpressed in mixed lineage leukemia-rearranged (MLL-r) acute myeloid leukemia (AML) and acute lymphoid leukemia (ALL) when compared to MLL-germline (MLL-G) leukemias. (B) Log2 FPKM expression of LAMP5 in AML and ALL pediatric patients with MLL-rearrangement (AML MLL-r, n=36 and ALL MLL-r, n=76) compared to MLL-G (AML MLL-G, n=270) (ALL MLL-G, n=727) patients. Data obtained from the St. Jude PeCan Portal and presented as median value with quartiles (t-test, ***P<0.0001). (C) Receiving operating curve (ROC) analysis showing the capacity of LAMP5 to discriminate acute leukemia patients with MLL-G or MLL-r leukemias. Data obtained from GSE13159 and St. Jude Pecan Portal. (D) Western blot analysis of LAMP5 expression in pediatric primary ALL and AML samples. Actin or vinculin was used as a loading control. (E) Relative expression of LAMP5 in MLL-r leukemia (MV4;11, MOLM-13, THP-1, and RS4;11) and MLL-G leukemia (K562, HL-60, Kasumi-1, REH, RCH-ACV) cell lines. The graph represents the relative expression of LAMP5 normalized to b-ACTIN. Data are from three biological replicates. LAMP5 expression in cord blood cells was set as 1.0. Bars show mean ± standard error of the mean (SEM). (F) Western blot analysis of the LAMP-5 levels in MLL-r leukemia and MLL-G leukemia human cell lines. CD34+ cord blood cells were used as control. Actin was used as a loading control. (G) Quantitative real-time polymerase chain reaction (RT-PCR) analysis of LAMP5 and MLL-AF9 gene expression in CD34+ cord blood cells transformed with a tetracycline-repressible MLL-AF9 construct. Gene expression was analyzed 24 hours after doxycycline incubation. Relative expression of LAMP5 and MLL-AF9 was normalized to b-ACTIN. (H) Representative chromatin immunoprecipitation sequencing (ChIP-seq) tracks at the LAMP5 locus from different MLL-r cell lines. ChIP-seq data were obtained from GSE95511 for ML-2, GSE79899 for MV4;11 and THP-1, GSE38403 for RS4;11, GSE38338 for SEM, GSE84116 for CB CD34+ MLL-Af4, and GSE83671 for primary patient MLL-AF4.

ized to LAMP-1+ vesicles. As suspected, we found MYD88 accumulating highly in the periphery of LAMP1+ vesicles in MLL-r leukemia, suggestive of TLR/IL-1R activation (Figure 3A). Conversely, in Kasumi-1 cells, MYD88 does not co-localize with LAMP-1+ vesicles. However, overexpression of wild-type LAMP-5 in this cell line led to the relocation of MYD88 around LAMP-1+ vesicles (Figure 3B). We subsequently hypothesized that LAMP-5 loss may dampen TLR/IL-1R signaling in MLL-r leukemias. We thus analyzed known effector proteins downstream of TLR/IL1R activation by western blot. Upon LAMP5 knockdown, we observed a reduction in phosphorylated IRAK1, NFkB, p38, and JNK, key players in the signal transduction downstream of TLR/IL-1R (Figure 3C). In order to further determine the impact of LAMP5 depletion in TLR-mediated NF-kB activation, we measured NF-kB activity using the THP-1 NF-kB-SEAP cell line, which contains an NF-kB inducible secreted embryonic alkaline phosphatase (SEAP) haematologica | 2022; 107(4)

reporter. Robust activation of NF-kB was evident in control cells upon incubation with PAM3CSK4 (TLR2 agonist) or LPS (TLR4 agonist). Knockdown of LAMP5 led to a near-complete blockade of this activation, suggesting that TLR-induced NF-kB signaling is disrupted upon LAMP5 depletion (Figure 3D). Correspondingly, in Kasumi-1 cells, overexpression of LAMP5 led to increased phosphorylation of p38, JNK, and NF-kB along with increased cell growth (Figure 3E and F). A previous study showed that NF-kB plays a critical role in MLL-r leukemias.39 We thus hypothesized that NFkB activation would rescue the cell growth defect seen by LAMP5 depletion. We induced persistent activation of NF-kB in leukemia cells by overexpressing a constitutively active version of inhibitor of nuclear factor kB kinase subunit b (IKBKB-EE) in these cells.41 Despite sustained NF-kB activation, knockdown of LAMP5 in MOLM-13 and RS4;11 cells still led to growth inhibition, suggesting that loss of NF-kB is not the only signaling event being 807

G. Gracia-Maldonado et al.

affected by LAMP5 depletion (Online Supplementary Figure S5A and B). A potential mechanism underlying this essentiality was proposed by Wang et al., where they suggested that loss of LAMP5 in MLL-r leukemia led to degradation of the MLL-FP due to increased autophagy.42 However, in our experiments, we did not observe any change in the

levels of the MLL-FP or LC3 A/B in THP-1 and MOLM-13 cells upon LAMP5 depletion (Online Supplemental Figure 6A and B). Overall, these results underscore a critical role for LAMP-5 in the activation of TLR/IL-1R signaling in MLL-r leukemia, while also indicating the presence of additional attributes that are also essential.

Figure 2. LAMP5 expression is required for mixed lineage leukemia-rearranged leukemia survival in vitro and in vivo. (A and B) In vitro growth of MLL-r and MLL-G leukemia cell lines (A) (MOLM-13, MV4;11, RS4;11 and THP-1) and (B) (Kasumi-1 and REH) respectively upon short hairpin RNA (shRNA) knockdown of LAMP5. Data are from three independent experiments, t-test, **P<0.01, ***P<0.001. (C) Colony-forming units (CFU) of MV4;11, MOLM-13, THP-1, RS4;11 cells upon LAMP5 shRNA knockdown. Data are from three biological replicates, represented as mean and SEM, t-test, **P<0.01, ***P<0.001. (D) Percentage of annexin V+/7-AAD+ cells after transduction with shNT or shLAMP5-2. Data are from three biological replicates, represented as mean and standard error of the mean (SEM) of at least three experiments. t-test, *P<0.05, **P<0.01 (E) Schematic of in vivo xenograft transplantation. MV4;11 or MLL-AF10 PDX cells were transduced with short hairpin non-targeting control (shNT) or shLAMP5-2 (shLAMP5) lentivirus and the mixed population of Venus+ and Venus- cells were transplanted into mice. (F) Plots show the percentage of human CD45+ (upper) and Venus+ cells in the CD45+ fraction (lower) in MV4;11 (left) and MLL-AF10 patient-derived xenograft (PDX) sample (right). Data are from eight biological replicates for MV4;11 and five biological replicates for the MLL-AF10 PDX, represented as mean and SEM, t-test, **P<0.01, ***P<0.001.

808

haematologica | 2022; 107(4)

LAMP5 as an essential target of MLL-r leukemias

LAMP-5 is a negative regulator of interferon-1 signaling in mixed lineage leukemia-rearranged leukemias Since activation of NF-kB was not sufficient to rescue the cell growth inhibition seen upon LAMP5 depletion, we next sought to understand the mechanistic significance of the inflammatory-signal-regulation function of LAMP-5 in MLL-r leukemia. In pDC, the carboxy-terminal YKHM domain of LAMP-5 was found to be required for normal localization of LAMP-5 and transportation of TLR9 from the early endosome vesicle to the pro-inflammatory vesicle.24,27,28 We thus overexpressed wild-type LAMP5 (LAMP5-WT), a Y276A mutant LAMP5 (LAMP5mut), or control vector (EV) in MV4;11 and THP-1 cells, followed by selective knockdown of endogenous LAMP5 using an shRNA targeting the 3’UTR region of LAMP5 (Online Supplementary Figure S7A). Overexpression of LAMP5-WT completely prevented cell growth inhibition and apoptosis upon knockdown of endogenous LAMP5, validating LAMP5 as the main target of the shRNA. In contrast, LAMP5-mut was unable to rescue cell growth or apoptosis in MV4;11 (Figure 4A and B). In pDC, LAMP5 knockdown or overexpression of LAMP5-mut induced IFN-1 activation upon TLR9-stimulation, due to retention of TLR9 in the IRF-SE. In order to determine the effect of

LAMP-5 on IFN-1 signaling in MLL-r leukemia, we turned to THP-1-ISG-SEAP cells containing an interferon-stimulated gene (ISG) inducible-SEAP reporter. Upon TLR activation by PAM3CSK4, IFN-1 signaling activation was evident only in the LAMP5-depleted cells but not in the control condition (Figure 4C). Furthermore, gene set enrichment analysis of RNA-seq from MOLM-13 cells transduced with shNT or shLAMP5-2 showed enrichment of IFN gene signatures (Online Supplementary Figure S8). Additionally, we validated the increase in IFN-1 signaling in several MLL-r cell lines by demonstrably increased expression of interferon a2 (IFNA2) and interferon b (IFNB) upon depletion of LAMP5 (Figure 4D). In order to assess the role of LAMP5-depletion mediated IFN-1 activation on cell growth, we performed knockdown of interferon regulatory factor 7 (IRF7), a known regulator of IFN signaling downstream of TLR/IL1R activation, along with LAMP5 in MV4;11 and THP-1 cells. We found that loss of IRF7 alone had no significant effect on MLL-r leukemia cell growth but importantly, its depletion prevented the growth inhibition observed upon LAMP5 knockdown (Figures 4E and F; Online Supplementary Figure S9A and B). Collectively, these results demonstrate that a critical function of LAMP-5 in MLL-r leukemias is to promote the transfer of TLR/IL-1R from the IFN-1–activating signaling

Figure 3. LAMP-5 is required for activation of Toll-like receptor/interleukin 1 receptor signaling. (A) Representative confocal microscopy images showing MOLM-13 cells stained with LAMP-5 (red), LAMP-1 (blue), and MYD88 (green); scale bar =1 mm. (B) Confocal microscopy image showing Kasumi-1 cells overexpressing empty vector (EV) or wild-type LAMP5 (LAMP5-WT) stained with antibodies against LAMP-5, LAMP-1, and MYD88; scale bar =1 mm. (C) Western blot analysis showing that LAMP5 depletion (shL5) led to a decrease of p-IRAK1, p-p38, p-JNK, and p-NF-kB, known downstream targets of Toll-like receptor (TLR) signaling. (D) THP-1-Blue-NFκB reporter cell line was treated with PAM3CSK4 10 ng/mL or LPS 100 ng/mL in the presence or absence of LAMP-5. Data are from three independent experiments. t-test, ***, P<0.001. (E) Western blot analysis of Kasumi-1 cells with overexpression of empty vector (EV) or LAMP5 showing increased activation of p-NF-kB, p-p38, and p-JNK. (F) In vitro cell growth of Kasumi-1 cells overexpressing EV or LAMP5-WT. Data are from three individual experiments. t-test, ***P<0.001.

haematologica | 2022; 107(4)

809

G. Gracia-Maldonado et al.

Figure 4. Continued on following page.

810

haematologica | 2022; 107(4)

LAMP5 as an essential target of MLL-r leukemias

Figure 4. LAMP-5 is a negative regulator of interferon-1 signaling in mixed lineage leukemia-rearranged leukemias. (A) In vitro growth of MV4;11 and THP-1 cells overexpressing empty vector control (EV), wildtype LAMP5 (LAMP5-WT), or mutated LAMP5 (LAMP5-mut) upon shRNA knockdown of LAMP5. Data are from three independent experiments. t-test ***, P<0.001. (B) Fold change of % apoptotic cells in MV4;11 cell line overexpressing empty vector (EV), wild-type LAMP5 (LAMP5WT), or mutant LAMP5 (LAMP5-mut) upon short hairpin RNA (shRNA) knockdown of LAMP5. Data are from three independent experiments. t-test, *P<0.05, **P<0.01. (C) THP-1 ISG blue reporter cell line was untreated (UT) or treated with 10 ng/mL PAM3CSK4 in the presence or absence of LAMP-5. Data are from three independent experiments. Bars show mean ± standard error of the mean (SEM). t-test, ***P<0.001. (D) Relative expression of LAMP5, IFNA2, and IFNB upon knockdown of LAMP5 in MV4;11, MOLM-13, THP-1 and RS4;11 cells. The graph represents the relative expression of LAMP5, IFNA2, and IFNB normalized to b-actin. Data are from three biological replicates. Bars show mean ± SEM. t-test *, P<0.05. (E) In vitro growth of MV4;11 after LAMP5, or IRF7 or LAMP5+IRF7 shRNA knockdown. Data are from three independent experiments, represented as mean and ± standard deviation. ***P<0.001. (F) Colony-forming units (CFU) of THP-1 cells upon shRNA knockdown of LAMP5, IRF7, or LAMP5+IRF7 together. Data are from three independent experiments, represented as mean and ± SEM. t-test **, P<0.01.

cascade to the pro-inflammatory signaling cascade. Depletion of LAMP5 thus leads not only to loss of NF-kB activation but also to activation of IFN-1-signaling, the latter inducing cell death.

Surface LAMP-5 can be detected and targeted with antibody drug conjugate therapy LAMP-5 has been found to briefly localize in the plasma membrane of cortical neurons in mice and is highly predicted to reach the cell membrane based on the human surfaceome.24,29 We thus sought to confirm if LAMP-5 was expressed on the surface of MLL-r leukemia cells. Using an antibody targeting the N-terminus of LAMP-5, we were able to detect LAMP-5 on the surface of MLL-r leukemia cell lines, while none was detected in the MLL-G leukemias (Figure 5A and B). In order to validate the specificity of the antibody, we overexpressed LAMP5 or control empty vector (EV) in Kasumi-1 cells. We detected surface LAMP-5 only in the cells that express high levels of LAMP5 (Figure 5C). As a proof-of-concept for potential therapeutic use, we used a secondary antibody conjugated to the tubulin-toxin Mertansine, targeting the surface-LAMP-5 antibody. We observed that a 72-hour treatment with this antibody-sandwich comprised of the surface LAMP-5 antibody along with the secondary antibody drug conjugate (ADC) antibody is sufficient to reduce cell viability in MLL-r leukemia cell lines MOLM-13, RS4;11 and THP-1, while no effect was seen in Kasumi-1 cells (Figure 5D). These results suggest that LAMP-5 could be exploited as an MLL-r specific biomarker and could potentially be used as a target for immunotherapy.

Discussion Our findings further reaffirm LAMP5 as a novel and essential core gene in MLL-r leukemias, directly upregulated by the MLL-FP. Additionally, we found that one of the critical functions of LAMP-5 is to regulate innatehaematologica | 2022; 107(4)

immune signaling in MLL-r leukemias, specifically directing the flux of activity away from IRF-SE towards the PISE, leading to constant activation of NF-kB (Figure 6). Recent discoveries have highlighted how the specific subcellular location and timing of TLR activation affect signaling outcomes in normal immune cells.43 Combes et al. showed that LAMP-5 is a negative regulator of IFN-1 signaling in pDC wherein it transports activated TLR9 from the IRF-SE to the PI-SE. Although dispensable for pDC cell survival, LAMP5 depletion led to unrestricted activation of IFN-1 signaling. Furthermore, aberrant expression of LAMP-5 can lead to diminished activation of pDC in tumors and contribute to their immunomodulatory phenotype by decreasing the IFN-1 production capacity.28 However, how these mechanisms function in leukemia is still poorly understood. Innate immune signaling and inflammation have been shown to play a crucial role in acute leukemias.37 MLL-r leukemias rely on activation of NF-kB downstream of TLR/IL-1R to maintain the MLL-FP gene signature and block cell differentiation.39,40 Furthermore, it has been shown that treatment with IFN-1 or activation of IFN-1 signaling is deleterious for MLL-r leukemias.44 In our study, we describe a novel role for LAMP-5 in maintaining NF-kB activation and blocking IFN-1 signaling downstream of TLR/IL-1R in MLL-r leukemias. We show that LAMP-5 acts as a molecular switch to maintain active TLR/IL-1R signaling in the pro-inflammatory endosome leading to NF-kB activation, whereas LAMP5 depletion leads to activation of IFN-1 signaling and cell death. This suggests that both the LAMP-5-mediated induction of pro-inflammatory signaling and inhibition of IFN-1 signaling contribute to the pathogenesis of MLL-r leukemias. We confirmed that activation of IFN-1 signaling upon LAMP-5 depletion was deleterious for leukemia propagation, and that by depleting IRF7, cell growth and clonogenicity were rescued in LAMP5-depleted cells. This suggests that increased IFN-1 signaling is at least partly responsible for inducing cell death upon LAMP5 depletion. Additionally, overexpression of LAMP5 in MLL-G leukemia led to increased acti811

G. Gracia-Maldonado et al.

vation of NF-kB, p38, and JNK, and increased cell growth, which suggests that this signaling pathway might be contributing to the therapy-resistant phenotype of MLL-r leukemias. In humans, LAMP5 expression is generally restricted to the brain and blood. In blood, LAMP5 is exclusively expressed in nonactivated pDC,27 wherein LAMP-5 resides in the ERGIC compartment and is transported to endo-lysosomal vesicles upon TLR9 activation.27,28 We found that the aberrant increased expression of LAMP-5 in MLL-r leukemia leads to its accumulation in the plasma membrane, as demonstrated by a novel LAMP-5 antibody targeting the N-terminus of the protein. The detec-

tion of LAMP-5 on the surface of MLL leukemias provides the opportunity to potentially use it as a target for immunotherapy in this treatment-refractory malignancy. Furthermore, LAMP-5 is highly expressed in other cancers such as multiple myeloma (MM) and blastic plasmacytoid dendritic cell neoplasm (BPDCN).45,46 Therefore, LAMP-5 immunotherapies could benefit other blood diseases. Finally, total loss of Lamp5 had no major effects on the health or lifespan of mice, only causing minor behavioral effects like deficits in olfactory discrimination and increased startle response to auditory and tactile stimuli,25,26 suggesting that there could be a wide therapeutic window for LAMP-5-directed therapies in humans.

Figure 5. Surface LAMP-5 can be detected and targeted with antibody drug conjugate therapy. (A) Representative histogram plots showing LAMP-5 surface expression in mixed lineage leukemia-rearranged (MLL-r) leukemia (MOLM-13, RS4;11, MV4;11, and THP-1) and MLL-germline (MLL-G) leukemia (Kasumi-1, HL-60, and NB4) cell lines. (B) Graph showing mean fluorescence intensity (MFI) of LAMP-5 surface staining in MLL-r leukemias vs. MLL-G leukemias represented as mean and ± standard deviation (SD). t-test,**P<0.01. (C) Representative histogram of LAMP-5 staining in Kasumi-1 expressing empty vector (EV) or LAMP5, confirming the specificity of the antibody. (D) MOLM-13, RS4;11, THP-1, and Kasumi-1 cells were incubated with surface LAMP-5 antibody clone D1 and aMFc-NC-DM1 antibody drug conjugate (ADC) antibody for 72 hours. Bar graph represents cell viability from three biological replicates presented as mean and ± SEM. t-test, *P<0.05.

812

haematologica | 2022; 107(4)

LAMP5 as an essential target of MLL-r leukemias

Figure 6. Proposed model to illustrate the mechanism of action of LAMP-5 in mixed lineage leukemia-rearranged leukemias and potential immunotherapy usage. Left panel: 1. The mixed lineage leukemia (MLL)-FP induces expression of LAMP-5. 2. LAMP-5 gets internalized from the cell surface to the interferon signaling endosome (IFN-SE), 3. and 4. LAMP-5 is quickly shuttled to the LAMP-1+ pro-inflammatory signaling endosome (PI-SE), activating NF-kΒ signaling. 5. NF-kΒ activates proinflammatory signaling. Right panel: 6. depletion of LAMP-5 leads to blockage of transport of TLR to PI-SE, with retention in and activation of the IFN-SE and 7. induction of interferon related genes and cell death. 8. Surface-LAMP-5 can be targeted in mixed lineage leukemia-rearranged (MLL-r) leukemias with immunotherapies.

Similar to our observations, Wang et al. recently showed that LAMP-5 is essential for the survival of MLLr leukemias in vitro and in vivo using shRNA knockdown. However, they propose that LAMP-5 is a negative regulator of autophagy leading to MLL-FP stabilization. They show that LAMP-5 and ATG5 co-localize in MLL-r leukemia cells and that blockade of autophagy is sufficient to rescue the increased levels of apoptosis after LAMP5 knockdown.42 We were unable to detect any significant change in the levels of the MLL-FP or LC3A/B upon LAMP5 knockdown. Since TLR-mediated innate immune signaling can regulate autophagy, the function of LAMP-5 in regulating autophagy as described by Wang et al. may be downstream of its impact on endosome-lysosome trafficking.47 On the other hand, it is also possible that these effects are not directly linked, and that LAMP5 might exert its growth-promoting effects in MLL-r leukemia by multiple mechanisms. It is notable; however, that the role of autophagy in leukemia is controversial. In murine MLL leukemia models, heterozygous loss of Atg5 leads to increased leukemia cell proliferation in vitro and more aggressive leukemia in vivo, while homozygous loss is lethal to these cells.48 Additionally, while some studies suggest that Atg5-dependent autophagy may contribute to the development of MLL-AF9 driven leukemia but dispensable for propagation and chemosensitivity, others suggest that Atg5-dependent autophagy is dispensable haematologica | 2022; 107(4)

altogether.49,50 Overall, our results show that LAMP-5 localizes both on the surface and in LAMP-1+ endosomes in leukemia, leading to constitutive activation of proinflammatory signaling, and dampening of interferon-signaling and that it can be used as a target for immunotherapy. Disclosures No conflicts of interest to disclose. Contributions GGM, LHL and ARK contributed to study conception and design; GGM, JC, MB and BG acquired data; GGM, JC, MW, DL, PP, EG and LHL analyzed and interpreted data; NS and LHL analyzed and interpreted RNA-seq data; GGM, JC, LHL and ARK wrote and revised the manuscript; GGM, JC, DL, PPE and ARK reviewed the manuscript; MW, DL and JC provided administrative, technical, or material support. Acknowledgments We would like to thank Daniel Starczcynowski, Ph.D., for his intellectual input. We thank J. Bailey and V. Summey for assistance with transplantations (Comprehensive Mouse and Cancer Core at CCHMC). We would like to acknowledge the assistance of the Research Flow Cytometry Core in the Division of Rheumatology at Cincinnati Children’s Hospital Medical Center. All flow cytometric data were acquired using equipment 813

G. Gracia-Maldonado et al.

maintained by the Research Flow Cytometry Core in the Division of Rheumatology at Cincinnati Children’s Hospital Medical Center. Funding The SH800S is supported by an NIH Shared Instrumentation Grant (S10OD023410). MW was supported by an NIH grant (R50 CA211404). NS was supported by an

References 1. Meyer C, Burmeister T, Gröger D, et al. The MLL recombinome of acute leukemias in 2017. Leukemia. 2018;32(2):273-284. 2. Hilden JM, Dinndorf PA, Meerbaum SO, et al. Analysis of prognostic factors of acute lymphoblastic leukemia in infants: report on CCG 1953 from the Children’s Oncology Group. Blood. 2006;108(2):441451. 3. Pieters R, Schrappe M, de Lorenzo P, et al. A treatment protocol for infants younger than 1 year with acute lymphoblastic leukaemia (Interfant-99): an observational study and a multicentre randomised trial. Lancet. 2007;370(9583):240-250. 4. Bauer J, Nelde A, Bilich T, Walz JS. Antigen targets for the development of immunotherapies in leukemia. Int J Mol Sci. 2019;20(6):1397. 5. Jacoby E, Nguyen SM, Fountaine TJ, et al. CD19 CAR immune pressure induces Bprecursor acute lymphoblastic leukaemia lineage switch exposing inherent leukaemic plasticity. Nat Commun. 2016;7(1):12320. 6. Rayes A, McMasters RL, O’Brien MM. Lineage switch in MLL-rearranged infant leukemia following CD19-directed therapy. Pediatr Blood Cancer. 2016;63(6):11131115. 7. Haddox CL, Mangaonkar AA, Chen D, et al. Blinatumomab-induced lineage switch of B-ALL with t(4:11)(q21;q23) KMT2A/AFF1 into an aggressive AML: pre- and post-switch phenotypic, cytogenetic and molecular analysis. Blood Cancer J. 2017;7(9):e607. 8. Balducci E, Nivaggioni V, Boudjarane J, et al. Lineage switch from B acute lymphoblastic leukemia to acute monocytic leukemia with persistent t(4;11)(q21;q23) and cytogenetic evolution under CD19-targeted therapy. Ann Hematol. 2017;96(9):1579-1581. 9. Wölfl M, Rasche M, Eyrich M, Schmid R, Reinhardt D, Schlegel PG. Spontaneous reversion of a lineage switch following an initial blinatumomab-induced ALL-toAML switch in MLL-rearranged infant ALL. Blood Adv. 2018;2(12):1382-1385. 10. Aldoss I, Song JY. Extramedullary relapse of KMT2A(MLL)-rearranged acute lymphoblastic leukemia with lineage switch following blinatumomab. Blood. 2018; 131(22):2507. 11. He RR, Nayer Z, Hogan M, et al. Immunotherapy- (Blinatumomab-) related lineage switch of KMT2A/AFF1 rearranged B-lymphoblastic leukemia into acute myeloid leukemia/myeloid sarcoma and subsequently into B/myeloid mixed phenotype acute leukemia. Case Rep Hematol. 2019;2019:7394619. 12. Fournier E, Inchiappa L, Delattre C, et al. Increased risk of adverse acute myeloid leukemia after anti-CD19-targeted

814

NIH grant (R01 CA226802). PP and EV were supported by the Institut National du Cancer (INCA) (PLBIO17-187), Canceropole Paca GEFLUC (RAK18024AAA), and Fondation ARC PJA (20131200330). GGM was supported by the Chateaubriand Fellowship. ARK was supported by a Hyundai Hope on Wheels grant. LHL is a St. Baldrick’s Foundation Scholar and is supported by grants from CancerFree KIDS and the NIH (L40 HL143713-01).

immunotherapies in KMT2A -rearranged acute lymphoblastic leukemia: a case report and review of the literature. Leuk Lymphoma. 2019;60(7):1827-1830. 13. Godfrey L, Crump NT, O’Byrne S, et al. H3K79me2/3 controls enhancer–promoter interactions and activation of the pan-cancer stem cell marker PROM1/CD133 in MLL-AF4 leukemia cells. Leukemia. 2020;35(1):90-106. 14. Li D, Hu Y, Jin Z, et al. TanCAR T cells targeting CD19 and CD133 efficiently eliminate MLL leukemic cells. Leukemia. 2018;32(9):2012-2016. 15. Lopez-Millan B, Sanchéz-Martínez D, Roca-Ho H, et al. NG2 antigen is a therapeutic target for MLL-rearranged B-cell acute lymphoblastic leukemia. Leukemia. 2019;33(7):1557-1569. 16. Ma X, Liu Y, Liu Y, et al. Pan-cancer genome and transcriptome analyses of 1,699 paediatric leukaemias and solid tumours. Nature. 2018;555(7696):371-376. 17. Haferlach T, Kohlmann A, Wieczorek L, et al. Clinical utility of microarray-based gene expression profiling in the diagnosis and subclassification of leukemia: report from the International Microarray Innovations in Leukemia Study Group. J Clin Oncol. 2010;28(15):2529-2537. 18. Zangrando A, Dell’orto MC, te Kronnie G, Basso G. MLL rearrangements in pediatric acute lymphoblastic and myeloblastic leukemias: MLL specific and lineage specific signatures. BMC Med Genomics. 2009;2(1):36. 19. Roychoudhury J, Clark JP, GraciaMaldonado G, et al. MEIS1 regulates an HLF-oxidative stress axis in MLL-fusion gene leukemia. Blood. 2015;125(16):25442552. 20. 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. 21. Valk PJ, Verhaak RG, Beijen MA, et al. Prognostically useful gene-expression profiles in acute myeloid leukemia. N Engl J Med. 2004;350(16):1617-1628. 22. Ross ME, Mahfouz R, Onciu M, et al. Gene expression profiling of pediatric acute myelogenous leukemia. Blood. 2004; 104(12):3679-3687. 23. Stam RW, Schneider P, Hagelstein JAP, et al. Gene expression profiling–based dissection of MLL translocated and MLL germline acute lymphoblastic leukemia in infants. Blood. 2010;115(14):2835-2844. 24. David A, Tiveron M-C, Defays A, et al. BAD-LAMP defines a subset of early endocytic organelles in subpopulations of cortical projection neurons. J Cell Sci. 2007; 120(2):353-365. 25. Tiveron M-C, Beurrier C, Céni C, et al. LAMP5 fine-tunes GABAergic synaptic transmission in defined circuits of the mouse brain. PLoS One. 2016;11(6): e0157052.

26. Koebis M, Urata S, Shinoda Y, et al. LAMP5 in presynaptic inhibitory terminals in the hindbrain and spinal cord: a role in startle response and auditory processing. Mol Brain. 2019;12(1):20. 27. Defays A, David A, de Gassart A, et al. BAD-LAMP is a novel biomarker of nonactivated human plasmacytoid dendritic cells. Blood. 2011;118(3):609-617. 28. Combes A, Camosseto V, N’Guessan P, et al. BAD-LAMP controls TLR9 trafficking and signalling in human plasmacytoid dendritic cells. Nat Commun. 2017;8(1):913. 29. Bausch-Fluck D, Goldmann U, Müller S, et al. The in silico human surfaceome. Proc Natl Acad Sci U S A. 2018;115(46):E10988E10997. 30. Lavallée V-P, Baccelli I, Krosl J, et al. The transcriptomic landscape and directed chemical interrogation of MLL-rearranged acute myeloid leukemias. Nat Genet. 2015; 47(9):1030-1037. 31. Gu Z, Churchman ML, Roberts KG, et al. PAX5-driven subtypes of B-progenitor acute lymphoblastic leukemia. Nat Genet. 2019;51(2):296-307. 32. Benito JM, Godfrey L, Kojima K, et al. MLL-rearranged acute lymphoblastic leukemias activate BCL-2 through H3K79 methylation and are sensitive to the BCL2-specific antagonist ABT-199. Cell Rep. 2015;13(12):2715-2727. 33. Lin S, Luo RT, Ptasinska A, et al. Instructive role of MLL-fusion proteins revealed by a model of t(4;11) pro-B acute lymphoblastic leukemia. Cancer Cell. 2016;30(5):737-749. 34. Prange K, Mandoli A, Kuznetsova T, et al. MLL-AF9 and MLL-AF4 oncofusion proteins bind a distinct enhancer repertoire and target the RUNX1 program in 11q23 acute myeloid leukemia. Oncogene. 2017; 36(23):3346-3356. 35. Numata A, Kwok HS, Kawasaki A, et al. The basic helix-loop-helix transcription factor SHARP1 is an oncogenic driver in MLL-AF6 acute myelogenous leukemia. Nat Commun. 2018;9(1):1-16. 36. Kerry J, Godfrey L, Repapi E, et al. MLLAF4 spreading identifies binding sites that are distinct from super-enhancers and that govern sensitivity to DOT1L inhibition in leukemia. Cell Rep. 2017;18(2):482-495. 37. Hemmati S, Haque T, Gritsman K. Inflammatory signaling pathways in preleukemic and leukemic stem cells. Front Oncol. 2017;7:265. 38. Cohen P. The TLR and IL-1 signalling network at a glance. J Cell Sci. 2014; 127(11):2383-2390. 39. Kuo H-P, Wang Z, Lee D-F, et al. Epigenetic roles of MLL oncoproteins are dependent on NF-κB. Cancer Cell. 2013;24(4):423437. 40. Liang K, Volk AG, Haug JS, et al. Therapeutic targeting of MLL degradation pathways in MLL-rearranged leukemia. Cell. 2017;168(1-2):59-72. 41. Mercurio F, Zhu H, Murray BW, et al. IKK-

haematologica | 2022; 107(4)

LAMP5 as an essential target of MLL-r leukemias

1 and IKK-2: cytokine-activated IkB kinases essential for NF-kappaB activation. Science. 1997;278(5339):860-866. 42. Wang W-T, Han C, Sun Y-M, et al. Activation of the lysosome-associated membrane protein LAMP5 by DOT1L serves as a bodyguard for MLL fusion oncoproteins to evade degradation in leukemia. Clin Cancer Res. 2019;25(9):2795-2808. 43. Oosenbrug T, van de Graaff MJ, Ressing ME, van Kasteren SI. Chemical tools for studying TLR signaling dynamics. Cell Chem Biol. 2017;24(7):801-812. 44. Tracey L, Streck CJ, Du Z, et al. NF-kB activation mediates resistance to IFN in MLL-

haematologica | 2022; 107(4)

rearranged acute lymphoblastic leukemia. Leukemia. 2010;24(4):806-812. 45. Beird HC, Khan M, Wang F, et al. Features of non-activation dendritic state and immune deficiency in blastic plasmacytoid dendritic cell neoplasm (BPDCN). Blood Cancer J. 2019;9(12):99. 46. Ledergor G, Weiner A, Zada M, et al. Single cell dissection of plasma cell heterogeneity in symptomatic and asymptomatic myeloma. Nat Med. 2018;24(12):18671876. 47. Into T, Inomata M, Takayama E, Takigawa T. Autophagy in regulation of Toll-like receptor signaling. Cell Signal. 2012;24(6): 1150-1162.

48. Watson A, Riffelmacher T, Stranks A, et al. Autophagy limits proliferation and glycolytic metabolism in acute myeloid leukemia. Cell Death Discov. 2015;1:15008. 49. Liu Q, Chen L, Atkinson JM, Claxton DF, Wang H-G. Atg5-dependent autophagy contributes to the development of acute myeloid leukemia in an MLL-AF9-driven mouse model. Cell Death Dis. 2016;7(9): e2361. 50. Chen X, Clark J, Wunderlich M, et al. Autophagy is dispensable for Kmt2a/MllMllt3/Af9 AML maintenance and antileukemic effect of chloroquine. Autophagy. 2017;13(5):955-966.

815

ARTICLE Ferrata Storti Foundation

Acute Myeloid Leukemia

Interleukin 4 promotes phagocytosis of murine leukemia cells counteracted by CD47 upregulation Pablo Peña-Martínez, Ramprasad Ramakrishnan, Carl Högberg, Caroline Jansson, David Gisselsson Nord and Marcus Järås Division of Clinical Genetics, Department of Laboratory Medicine, Lund University, Lund, Sweden

Haematologica 2022 Volume 107(4):816-824

ABSTRACT

Correspondence: MARCUS JÄRÅS marcus.jaras@med.lu.se Received: August 25, 2020. Accepted: April 27, 2021. Pre-published: May 6, 2021.

ytokines are key regulators of tumor immune surveillance by controlling immune cell activity. Here, we investigated whether interleukin 4 (IL4) has antileukemic activity via immune-mediated mechanisms in an in vivo murine model of acute myeloid leukemia driven by the MLL–AF9 fusion gene. Although IL4 strongly inhibited leukemia development in immunocompetent mice, the effect was diminished in immune-deficient recipient mice, demonstrating that the antileukemic effect of IL4 in vivo is dependent on the host immune system. Using flow cytometric analysis and immunohistochemistry, we revealed that the antileukemic effect of IL4 coincided with an expansion of F4/80+ macrophages in the bone marrow and spleen. To elucidate whether this macrophage expansion was responsible for the antileukemic effect, we depleted macrophages in vivo with clodronate liposomes. Macrophage depletion eliminated the antileukemic effect of IL4, showing that macrophages mediated the IL4-induced killing of leukemia cells. In addition, IL4 enhanced murine macrophage-mediated phagocytosis of leukemia cells in vitro. Global transcriptomic analysis of macrophages revealed an enrichment of signatures associated with alternatively activated macrophages and increased phagocytosis upon IL4 stimulation. Notably, IL4 concurrently induced Stat6-dependent upregulation of CD47 on leukemia cells, which suppressed macrophage activity. Consistent with this finding, combining CD47 blockade with IL4 stimulation enhanced macrophage-mediated phagocytosis of leukemia cells. Thus, IL4 has two counteracting roles in regulating phagocytosis in mice; enhancing macrophage-mediated killing of leukemia cells, but also inducing CD47 expression that protects target cells from excessive phagocytosis. Taken together, our data suggest that combined strategies that activate macrophages and block CD47 have therapeutic potential in acute myeloid leukemia.

Introduction https://doi.org/10.3324/haematol.2020.270421

816

Acute myeloid leukemia (AML) is a fatal disease characterized by an accumulation of myeloid blasts in the bone marrow. For AML to develop, the malignant cells must escape tumor immune surveillance. Several evasion mechanisms have been described in AML, mainly associated with suppression of natural killer (NK) cells and macrophages.1-3 Suppression of NK cells is mediated by secretion of ligands from the leukemic blasts and through direct cell–cell interactions with leukemic cells.4 An absence of NKG2D ligands on leukemia stem cells mediates their immune evasion.5 The main inhibitory signal to macrophages is CD47, which is upregulated on AML cells and protects them from phagocytosis.2 Paradoxically, tumor-associated macrophages in AML also contribute to immune suppression.6,7 Whereas interleukin (IL)2 and IL15 promote restoration of NK cell function in AML,8 anti-CD47 blocking antibodies can rescue macrophage function.9 Whether cytokine treatment can restore and boost macrophage-mediated antileukemic activity is currently unclear. In a syngeneic murine AML model, we previously found that IL4 exerts

haematologica | 2022; 107(4)

IL4 enhances phagocytosis of murine leukemia cells

antileukemic activity by inducing Stat6-dependent apoptosis of AML cells.10 Elevated IL4 levels in mice eradicate AML cells in both the spleen and bone marrow, resulting in increased survival. Under physiological conditions, IL4 is a pleiotropic cytokine that regulates several immunological processes, such as B-cell class switching, T helper cell maturation, alternative activation of macrophages, and activation of NK cells.11,12 IL4 can bind to the IL4 receptor (IL4R) type I receptor complex, a heterodimer of the IL4R alpha (IL4RA) and IL2 receptor subunit gamma (IL2RG) chains, or to the IL4R type II receptor complex, a dimer of IL4RA and IL13RA1.13 Whether immune cells also mediate the antileukemic activity of IL4 has not been previously explored. In this study, we show that IL4 regulates phagocytosis by enhancing macrophage-mediated killing of AML cells and increasing CD47 expression on leukemia cells, which inhibits macrophages. Combined blockade of CD47 and IL4 stimulation enhanced macrophage-mediated killing of AML cells. Hence, our data suggest that combined strategies that activate macrophages and block CD47 have therapeutic potential in AML.

For human phagocytosis assays, we labeled human leukemia cell lines with the PKH67 green fluorescent cell dye according to the manufacturer’s instructions (Sigma-Aldrich, Darmstadt, Germany) and stained macrophages with the PKH26 red fluorescent cell dye (Sigma-Aldrich). AML cells were mixed with human macrophages in a 2:1 ratio and incubated for either 2 h (Mono Mac 6 cells) or 18 h (MA9:16 cells). The percentage of PKH26+ PKH67+ macrophages was determined by FACS.

RNA sequencing analysis Global gene expression profiling was performed on sorted F4/80+ spleen cells from mice transplanted with IL4-overexpressing leukemia cells and non-transplanted irradiated controls. Cells were collected 12 days after irradiation. In addition, RNA sequencing was performed on macrophages produced in vitro by stimulating murine monocytes for 7 days with murine (m)CSF1 (25 ng/mL) and mIL4 (20 ng/mL) or only mCSF1. Raw data and normalized gene expression data are available in the Gene Expression Omnibus database under accession number GSE155048.

Results Methods

The antileukemic activity of interleukin 4 in vivo is predominantly mediated via immune cells

The murine leukemia model

To determine whether immune cells contribute to the previously described antileukemic effects of IL4 in vivo,10 we used a murine AML model driven by the MLL–AF9 (KMT2A-MLLT3) fusion gene.14 The leukemia cells were generated in a dsRed+ transgenic background, allowing for convenient tracking of leukemia cells upon serial transplantations.16,17 Serial passaging of leukemia cells in mice did not alter IL4RA expression on AML blasts (Online Supplementary Figure S1A). Consistent with previous results,10 we confirmed that elevated IL4 levels mediated by retroviral expression in c-Kit+ AML cells transplanted into mice (IL4 group) resulted in strong in vivo antileukemic activity. The IL4 group showed prolonged survival compared to controls and had almost no leukemia cells in the bone marrow or spleen at the time of sacrifice (Figure 1A, B, Online Supplementary Figure S1B). To address whether the antileukemic activity of IL4 in vivo was immune-mediated, we used two strains of immunodeficient recipient mice: NOD/SCID mice, which lack T and B cells and have decreased activity of both NK cells and macrophages,18 and NSG mice, which additionally lack NK cells.19 In NOD/SCID animals, the antileukemic effect of IL4 was reduced, and we observed increased levels of leukemia cells in the bone marrow and spleens compared to the levels in immunocompetent mice (Figure 1C, Online Supplementary Figure S1C). These findings suggest that immune cells at least partially mediate the antileukemic effect of IL4. To further characterize the antileukemic effect of IL4, we used the NSG mouse strain, which lacks a functional IL4 receptor type I complex because of deficiency in the Il2rg gene. Of note, in NSG mice, the antileukemic effect of IL4 was abolished, and survival was even shorter than in controls, with high levels of leukemia cells in both the bone marrow and spleens at the time of sacrifice (Figure 1D, Online Supplementary Figure S1D). These findings suggest that the antileukemic effect of IL4 in vivo depends on immune cells expressing the IL4 receptor type I complex.

All animal experiments were conducted according to the protocol approved by the Animal Care and Use Committee of the Lund/Malmö Ethical Committee. MLL–AF9 leukemias were generated in a dsRed C57BL/6 transgenic background (6051; Jackson Laboratory, Bar Harbour, NY, USA), as previously described.14 The MLL–AF9 leukemia was serially propagated in sublethally irradiated (600 cGy) C57BL/6 recipient mice and leukemia stem cells were enriched as previously described.15 All experiments involving murine leukemia cells were performed using tertiary or quaternary transplanted leukemia cells. As immunodeficient murine recipients, sublethally irradiated (250 cGy) NOD/SCID and NOD.CgPrkdcscidIl2rgtm1Wjl/SzJl (NSG) mice were used (in-house breeding). All mice used in experiments were age- and sex-matched.

In vivo depletion of macrophages To deplete macrophages in mice transplanted with retrovirally transduced leukemia cells, we used intraperitoneal (i.p.) injection of 200 mL of clodronate liposomes (5 mg/mL; Liposoma B.V., Amsterdam, the Netherlands). Controls were injected with phosphate-buffered saline. We administered the first injection of clodronate liposomes 1 day before injections of leukemia cells and repeated the procedure every tenth day. All mice in the survival experiments were sacrificed based on at least one of the following criteria: immobility, hunched back, hind leg paralysis, or dehydration.

Phagocytosis assay For mouse phagocytosis assays, c-Kit+ dsRed+ murine MLL–AF9 leukemia cells were added to macrophage cultures in a 2:1 ratio. After 18 h, the cells were stained with a BV421–F4/80 antibody (BioLegend, San Diego, CA, USA), and the percentage of F4/80+dsRed+ cells was determined by FACS analysis. For the CD47 blocking experiments, we incubated c-Kit+ dsRed+ murine MLL–AF9 leukemia cells for 30 min with an anti-CD47 antibody or rat IgG2a isotype control (30 mg/mL; BioXCell, Lebanon, NH, USA), before co-culture with macrophages for 1 h at 37°C. The percentage of F4/80+dsRed+ cells was determined by flow cytometry as described above.

haematologica | 2022; 107(4)

817

P. Peña-Martinez et al.

Figure 1. Interleukin-4 has antileukemic activity in a microenvironment-dependent manner. (A) dsRed+ c-Kit+ MLL–AF9 acute myeloid leukemia (AML) cells were transduced with retroviral vectors coexpressing green fluorescent protein (GFP) and a murine interleukin 4 cDNA (MIG–IL4) or an empty control vector (MIG). Two days later, sorted GFP+ AML cells were transplanted into sublethally irradiated mice. (B) Transplantation of 10,000 leukemia cells into C57BL/6 mice. Kaplan-Meier survival curves (9 mice per group, pooled from 2 independent experiments), and percentage of leukemia (dsRed+) cells in the bone marrow (BM) of mice at the time of sacrifice. (C) Transplantation of 30,000 leukemia cells into NOD/SCID mice. Kaplan-Meier survival curves (6 mice per group) and percentage of leukemia cells in the BM of mice at the time of sacrifice. (D) Transplantation of 30,000 leukemia cells into NSG mice. Kaplan-Meier survival curves (14 mice per group, pooled from 2 independent experiments), and percentage of leukemia cells in the BM of mice at the time of sacrifice. ***P<0.001; ****P<0.0001.

Interleukin 4 expands macrophages in vivo To identify the type of immune cell that mediates the IL4-induced antileukemic effects, we analyzed the hematopoietic compartment in mice receiving IL4-secreting AML cells. At day 19 after transplantation, we detected no IL4-induced alterations in blood cell lineages by flow cytometry (Figure 2A). Moreover, at this time-point, we detected no circulating leukemia cells in the blood of mice in the IL4 group (Figure 2B). In contrast, at day 27 after transplantation, the white blood cell, red blood cell, and platelet counts in the IL4 group were reduced compared to those in controls that had not been injected with leukemia cells (Figure 2C, Online Supplementary Figure S2A, B). Of note, at the time of sacrifice, when the mice had succumbed to disease (Figure 1B), there was significant expansion of F4/80+ macrophages in the bone marrow (on average, 2.4% vs. 1%; P<0.001) and spleens (on average, 7.9% vs. 1.3%; P<0.0001) of IL4 mice (Figure 2D, Online Supplementary Figure S2C, D). We confirmed this IL4-induced increase in the proportion of macrophages by immunohistochemistry (Figure 2E, Online Supplementary Figure S2E). We also confirmed IL4RA expression on the F4/80+ cells from both groups of mice, supporting that IL4 receptor signaling may directly stimulate macrophages in this model (Online Supplementary Figure S2F). Hematoxylin staining of sections revealed extramedullary hematopoiesis in the spleens of the IL4 mice, as indicated by a marked increase in megakaryocytes and altered spleen architecture with increased red pulp and decreased white pulp (Online Supplementary Figure S2G). In addition to a reduction in leukemia cells, the decrease in circulating white blood cells, increased extramedullary hematopoiesis, and hypocellular bone marrow indicated 818

that the elevated IL4 levels resulted in bone marrow failure in these animals. By contrast, in NSG mice, the IL4 group exhibited high levels of leukemia cells in the bone marrow, similar to the levels in the MIG control group (Online Supplementary Figure S2H).

Interleukin 4 stimulation increases murine macrophage-mediated phagocytosis of leukemia cells To assess whether the IL4-induced expansion of macrophages in vivo was responsible for the antileukemic activity of IL4, we depleted macrophages by intraperitoneal injections of clodronate liposomes,20,21 followed by injection of IL4-secreting AML cells (Figure 3A). Efficient depletion of macrophages was observed in the spleen but not in the bone marrow (Figure 3B). Consistent with the macrophage depletion, we found a proportional increase of leukemia cells in the spleen of these mice (on average, 33% vs. 6%; P<0.05), but not in the bone marrow (Figure 3C). In contrast, depletion of macrophages had no effect on the level of leukemia cells in the MIG control group (Online Supplementary Figure S3A, B). These findings suggest that macrophages mediate the IL4-induced killing of leukemia cells. Because macrophages kill cells by phagocytosis, we next assessed whether IL4 stimulation results in increased macrophage-mediated phagocytosis of leukemia cells in culture. Murine monocytes isolated from bone marrow were differentiated into macrophages for 7 days by supplementation of the culture medium with CSF1 (Figure 3D). The addition of IL4 to the medium resulted in increased phagocytosis of leukemia cells, as evident by macrophage acquisition of dsRed fluorescence (Figure 3E, F). Consistent with a more activated state, the IL4-stimuhaematologica | 2022; 107(4)

IL4 enhances phagocytosis of murine leukemia cells

Figure 2. Interleukin 4 stimulation increases the frequency of macrophages in vivo. C57BL/6 mice were transplanted with 30,000 sorted green fluorescent protein (GFP)+ MLL-AF9 acute myeloid leukemia (AML) cells 2 days after transduction with retroviral vectors co-expressing GFP and a murine interleukin 4 cDNA (MIG–IL4) or a control vector (MIG). (A) Percentages of blood cell populations within dsRed– cells 19 days after transplantation (n=3). (B) Percentage of leukemia (dsRed+ ) cells in the peripheral blood on day 19 after transplantation (n=3). (C) White blood cell counts at days 12 and 27 for MIG–IL4 and non-transplanted irradiated control mice (IL4 group, n=4; controls, n=3). (D) Percentage of F4/80+ cells within dsRed– cells in bone marrow and spleens of mice at the time of sacrifice (controls, n=4; IL4 group, n=5). (E) Representative immunohistochemistry staining of F4/80+ cells in bone marrow (40×; scale bar, 20 mm) and spleens (10x; scale bar, 100 mm). BM: bone marrow; N.D.: not detected; PB: peripheral blood; WBC: white blood cell; IHC: immunohistochemistry. **P<0.01; ***P<0.001; ****P<0.0001.

haematologica | 2022; 107(4)

819

P. Peña-Martinez et al.

Figure 3. Interleukin 4 stimulation causes macrophage-mediated depletion of leukemia cells in vivo. (A) C57BL/6 mice were transplanted with 30,000 sorted green fluorescent protein (GFP)+ MLL-AF9 acute myeloid leukemia (AML) cells transduced with retroviral vectors expressing a murine interleukin 4 cDNA (MIG–IL4) or GFP only (MIG; data presented in Online Supplementary Figure S2). One day prior to transplantation, mice received intraperitoneal (i.p.) injections of clodronate liposomes (MΦdep group; n=4) or phosphate-buffered saline as control (n=5). Every tenth day, new i.p. injections were performed. (B) Percentage of F4/80+ cells and (C) leukemia cells in bone marrow (BM) and spleens at the time of sacrifice in the IL4 group. (D) Monocytes were isolated from mouse BM and differentiated into macrophages in culture with mCSF1 (25 ng/mL) and mIL4 (20 ng/mL) for 7 days, and then MLL-AF9 dsRed+ AML cells were co-cultured with the macrophages. (E) Representative flow cytometry contour plots showing dsRed+ cells within F4/80+ cells in freshly mixed cultures (0 h) and after 18 h of co-culture with macrophages and dsRed+ leukemia cells. (F) Phagocytosis assay with dsRed+ AML cells and murine macrophages (n=3). The percentage of dsRed+ cells within F4/80+ cells is presented. (G) CD14+ cells were isolated from human blood and differentiated into macrophages in culture with human (h)CSF1 (25 ng/mL) and hIL4 (20 ng/mL) for 7 days and then co-cultured with membrane-stained AML cell lines. (H) Phagocytosis assay with PKH67+ MA9:16 cells and PKH26+ human macrophages (n=4). The percentage of PKH67+ cells within PKH26+ cells is presented. (I) Phagocytosis assay with PKH67+ Mono Mac 6 cells and PKH26+ human macrophages (n=5). BM, bone marrow; MM6, Mono Mac 6; MΦ, macrophage. **P<0.01; ***P<0.001; ****P<0.0001.

lated macrophages had an increased volume and were less irregular than unstimulated cells, as evaluated using phase holograph imaging (Online Supplementary Figure S4). In contrast to its effect on murine macrophages, human IL4 is well known to differentiate human monocytes into anti-inflammatory macrophages.22 To assess how human IL4 affects phagocytosis of leukemia cells, human macrophages were stimulated with IL4 before mixing with AML cell lines. In line with a differential role of IL4 in mice and humans, IL4 suppressed human macrophagemediated phagocytosis of the AML cells (Figure 3G-I).

Interleukin 4 induces polarization of macrophages To investigate how IL4 affects the global gene expression of macrophages, we performed RNA sequencing of murine macrophages generated in vitro with or without IL4 stimulation. In addition, we performed RNA sequencing on sorted dsRed– F4/80+ macrophages from mice transplanted with IL4-expressing leukemia cells and macrophages from leukemic control mice. In agreement with a described role for IL4 in promoting macrophage 820

polarization, IL4 induced the expression of several genes associated with alternative activation of macrophages, including Arg1, Chil3, and Retnla (Figure 4A, Online Supplementary Figure S5A, B),22,23 which were among the most differentially upregulated genes (Online Supplementary Tables S1 and S2). Of note, IL4 also induced strong upregulation in vivo of the chemokine Ccl24, a biomarker for macrophages that originate from monocytes rather than tissue-resident macrophages (Figure 4A).24 Moreover, the IL4-induced macrophages showed downregulation of genes such as Cd68, which is associated with tumor-associated macrophages (Figure 4B),25 indicating that IL4 differentiates macrophages into a phenotype that is distinct from tumor-associated macrophages. We next performed gene set enrichment analysis to identify gene expression signatures enriched in the IL4-induced macrophages in vivo. In accordance with increased phagocytosis of macrophages stimulated with IL4 in vitro, we found an enrichment of phagocytosis signatures in macrophages harvested from mice in the IL4 group (Figure 4C). Moreover, IL4 stimulation resulted in enrichment of genes haematologica | 2022; 107(4)

IL4 enhances phagocytosis of murine leukemia cells

Figure 4. Interleukin 4 expands macrophages enriched for gene expression signatures associated with alternative activation of macrophages and phagocytosis. RNA sequencing was performed on murine macrophages generated from monocytes in vitro, and on sorted dsRed-F4/80+ macrophages from mice in the interleukin 4 (IL4) and control groups. (A) Volcano plots displaying differential gene expression between IL4-stimulated macrophages and control macrophages in vitro (left plot), and macrophages from mice in the IL4 or control group (right plot). The y-axis corresponds to the –log10(q-value) and the x-axis to the log2 of the gene expression fold change. Green dots represent significantly differentially expressed genes with a q-value <0.05 and fold change >2.0. (B) Heatmap showing expression of genes associated with upregulation in tumor-associated macrophages. IL4-stimulated macrophages and control macrophages were harvested from mice. (C) Gene set enrichment analysis revealed enrichment of phagocytosis and MHC protein complex signatures in macrophages harvested from mice. FDR, false discovery rate; GO: gene ontology; MΦ, macrophage; NES, normalized enrichment score; TAM: tumor-associated macrophage.

associated with major histocompatibility complex (MHC) proteins (Figure 4C). To determine the influence of the in vivo microenvironment, we compared the gene expression profiles of IL4-stimulated macrophages generated in vitro and those generated in vivo (Online Supplementary Table S3). Macrophages generated in vivo exhibited a preferential upregulation of several markers associated with inflammation and immune activation (Online Supplementary Figure S5C, D). Altogether, the gene expression data suggest that IL4 stimulation leads to an expansion of monocyte-derived macrophages with increased phagocytic activity.

Interleukin 4 upregulates CD47 in a Stat6-dependent manner We next searched for IL4-induced mechanisms in leukemia cells that might affect their interactions with macrophages. Interestingly, the macrophage-inhibitory protein CD47 was upregulated on leukemia cells in the IL4 group compared to controls at the time of sacrifice (Figure 5A). Consistent with this finding, IL4 induced the expression of CD47 in leukemia cells in a dose-dependent manner, showing that IL4 activates signaling that induces CD47 expression (Figure 5B). Moreover, according to RNA sequencing data that we had previously generated,10 Cd47 haematologica | 2022; 107(4)

was upregulated in both c-Kit+ AML cells and normal c-Kit+ bone marrow cells stimulated with IL4 (Figure 5C). We next explored the mechanistic basis of the IL4induced upregulation of CD47. Because STAT6 is a critical downstream mediator of IL4R signaling, we used CRISPR/Cas9 genetic engineering to knock out Stat6 in Cas9-expressing MLL–AF9 AML cells using Stat6 sgRNA that we had previously characterized.10 Stat6 disruption hindered the IL4-induced upregulation of CD47 (Figure 5D), demonstrating that IL4 upregulates CD47 in a STAT6-dependent manner. Thus, in addition to activating murine macrophages, we identified a previously unknown role of IL4 in protecting cells from phagocytosis via CD47 upregulation.

Combined interleukin 4 treatment and CD47 blockade results in enhanced macrophage-mediated phagocytosis of acute myeloid leukemia cells Because CD47 protects cells from phagocytosis, we next evaluated whether the IL4-induced upregulation of CD47 on AML cells counteracts enhanced phagocytosis by IL4-stimulated macrophages. Consistent with this hypothesis, AML cells pre-treated for 24 h with IL4 and washed before co-culture with macrophages were partial821

P. Peña-Martinez et al.

ly resistant to phagocytosis (Figure 5E). To overcome the inhibitory signal provided by increased CD47 expression, we used an a-CD47 blocking antibody. Combined blocking of CD47 on AML cells and IL4 stimulation of macrophages resulted in enhanced phagocytosis of AML cells (Figure 5F). These findings show that IL4 has a dual role in murine phagocytosis by directly activating macrophages and enhancing their phagocytic activity, while also inducing CD47 expression that counteracts phagocytosis in target cells.

Discussion Distinct types of macrophages control tumor development. Whereas tumor-associated macrophages promote tumor development by suppressing the immune system, other types of macrophages achieve tumor immune surveillance through phagocytosis of malignant cells.26-29 We found that IL4 has antileukemic effects in mice, predominantly mediated by alternatively activated macrophages that normally play a key role in tissue repair and immune

Figure 5. Combined interleukin 4 stimulation and CD47 blockade result in enhanced macrophage-mediated phagocytosis of acute myeloid leukemia cells. (A) Representative histograms showing CD47 expression on acute myeloid leukemia (AML) cells in bone marrow (BM) and spleens of mice transplanted with dsRed+ leukemia cells transduced with the MIG–interleukin 4 (MIG-IL4) or control (MIG) vectors. (B) CD47 expression on AML cells following IL4 stimulation for 24 h. (C) Cd47 expression shown as FPKM values of normalized reads from RNA sequencing data of c-Kit+ dsRed+ leukemia cells and c-Kit+ normal BM cells stimulated with IL4 for 18 h. Data are presented as box and whiskers diagrams; the line indicates median, box limits are first and third quartiles, and bars indicate maximum and minimum values. (D) CD47 expression measured by flow cytometry after 24 h of stimulation with murine (m)IL4 (100 ng/mL) in cells transduced with lentiviral vectors expressing Stat6 or control sgRNA. (E) Phagocytosis assay with macrophages derived from murine BM monocytes stimulated with mCSF1 (25 ng/mL) and mIL4 (20 ng/mL) for 7 days. The AML cells were treated with mIL4 (100 ng/mL) or no IL4 (control) for 24 h prior to co-culture (n=3). Phagocytosis is presented as the percentage of dsRed+ cells within F4/80+ cells. (F) Phagocytosis assay with mouse BM monocyte-derived macrophages stimulated for 7 days with mCSF1 (25 ng/mL) and mIL4 (20 ng/mL) or mCSF1 only (n=3). AML cells were cultured for 1 h with a blocking anti-CD47 antibody or corresponding isotype control and then mixed with the macrophages. FPKM, fragments per kilobase million; gMFI, geometric mean fluorescence intensity; NBM, normal bone marrow. *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001.

822

haematologica | 2022; 107(4)

IL4 enhances phagocytosis of murine leukemia cells

regulation.30,31 The observed expansion of alternatively activated macrophages is consistent with findings showing that IL4, via the IL4 receptor type I complex, promotes the outgrowth of macrophages beyond homeostatic levels in the setting of nematode infections.32 However, nematode infections trigger the expansion of tissue resident macrophages.32 In contrast, the IL4-induced macrophages with antileukemic activity showed higher expression of Ccl24, Mrc1, and Pdcd1lg2, suggesting that they are of monocytic origin, from either the bone marrow or peripheral blood.24 Among hematopoietic cells, only macrophages showed increased numbers following enforced expression of IL4 in vivo. IL4 also boosted the phagocytic activity of murine monocyte-derived macrophages in vitro, suggesting that IL4 acts directly on the monocytes/macrophages that mediate the antileukemic effect. Moreover, consistent with their increased phagocytic activity, the IL4-induced macrophages were functionally and molecularly distinct from tumor-associated macrophages, which are classically associated with an alternatively activated phenotype.25 Furthermore, the IL4-induced macrophages were functionally distinct from AML-associated macrophages, which polarize into a leukemia-supportive state that accelerates disease development.3 The reason why IL4 induced stronger macrophage activation in vivo than in vitro could be related to interactions with other immune cells or the AML blasts, resulting in enhanced phagocytic activity. Of note, the macrophages were dependent on IL4 for their anti-leukemic activity as depletion of macrophages in the MIG control group did not affect the leukemia burden. Constitutive expression of IL4 in mice has not been linked previously to anti-cancer activity, but it has been associated with excessive phagocytosis resulting in decreased blood cell counts, extramedullary hematopoiesis, and increased mortality.33,34 We found that IL4 induced potent antileukemic activity, with some mice surviving long-term without signs of disease or problems of tolerability, while other mice eventually had to be sacrificed despite very low levels of leukemia cells in their bone marrow and spleen. The low blood cell counts and expansion of megakaryocytes in the spleen indicated extramedullary hematopoiesis and suggests that elevated IL4 levels induced macrophage activation with excessive phagocytosis. This pattern resembles that of hemophagocytic lymphohistiocytosis (HLH), a disease characterized by aberrantly activated macrophages.35 Hence, we speculate that the cause of death of non-leukemic mice in the IL4 group was due to the HLH-like symptoms. Of note, the leukemic cells were selectively depleted, indicating that the IL4-induced macrophages preferentially attacked them. The reason is unclear but could be related to altered expression of genes by leukemia cells that regulate macrophages, such as MHC class I molecules or calreticulin.36,37 In addition to IL4 boosting macrophage-mediated phagocytosis, stimulation of AML cells with IL4 induced STAT6-dependent upregulation of CD47, revealing a pre-

References 1. Costello RT, Sivori S, Marcenaro E, et al. Defective expression and function of natural killer cell-triggering receptors in patients with acute myeloid leukemia. Blood.

haematologica | 2022; 107(4)

viously unrecognized mechanism that regulates CD47 expression and thereby protects cells from phagocytosis. This mechanism could possibly have evolved to protect endogenous cells from phagocytosis in areas in which high IL4 levels activate macrophages to fight invading pathogens. Consistent with these findings, a superenhancer region with binding sites for STAT6 has been shown to regulate CD47 expression,38 providing a putative mechanistic basis for how CD47 is upregulated via the IL4/STAT6 pathway. Given that combined IL4 stimulation and CD47 inhibition enhanced macrophage-mediated phagocytosis of AML cells, our data suggest therapeutic potential for strategies that combine direct activation of macrophages with blocking of inhibitory signals to macrophages. Because IL4 has opposing effects in murine and human macrophages, we speculate that other cytokines that activate human macrophages may also upregulate CD47 or other ‘don’t eat me’ signals on target cells. Identifying these mechanisms may translate into new therapeutic opportunities in AML and possibly other types of cancer. In summary, here we show that IL4 has a potent in vivo antileukemic effect in mice by promoting macrophagemediated phagocytosis of AML cells. IL4 stimulation induced CD47 upregulation in a STAT6–dependent manner, and combined IL4 stimulation with CD47 blockade further enhanced macrophage-mediated phagocytosis of AML cells. These findings deepen our understanding of how IL4 regulates murine macrophages and suggest that strategies to combine macrophage activation with CD47 inhibition should be explored further as a therapeutic approach in cancer. Disclosures No conflicts of interest to disclose. Contributions PPM, RR, CH and CJ performed research, PPM and MJ analyzed data and wrote the manuscript, and all other authors contributed with valuable comments. Acknowledgments The authors thank Dr Benjamin Ebert (Brigham and Women’s Hospital, Boston, MA, USA) for sharing the dsRed+ MLL-AF9 leukemia cells. We also thank Dr James Mulloy, (University of Cincinnati, Cincinnati, OH, USA) for sharing the MA9:16 cells. Funding We thank the following granting agencies for their support: the Swedish Cancer Society, the Swedish Childhood Cancer Foundation, the Swedish Research Council, the Crafoord Foundation, the Royal Physiographic Society in Lund, and the Medical Faculty of Lund University. Data-sharing statement Raw data and normalized gene expression data are available in the Gene Expression Omnibus database under accession number GSE155048.

2002;99(10):3661-3667. 2. Jaiswal S, Jamieson CH, Pang WW, et al. CD47 is upregulated on circulating hematopoietic stem cells and leukemia cells to avoid phagocytosis. Cell. 2009; 138(2):271-285.

3. Al-Matary YS, Botezatu L, Opalka B, et al. Acute myeloid leukemia cells polarize macrophages towards a leukemia supporting state in a growth factor independence 1 dependent manner. Haematologica. 2016; 101(10):1216-1227.

823

P. Peña-Martinez et al. 4. 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. 5. Paczulla AM, Rothfelder K, Raffel S, et al. Absence of NKG2D ligands defines leukaemia stem cells and mediates their immune evasion. Nature. 2019; 572(7768): 254-259. 6. Petty AJ, Yang Y. Tumor-associated macrophages in hematologic malignancies: new insights and targeted therapies. Cells. 2019;8(12):1526. 7. Li Y, You MJ, Yang Y, Hu D, Tian C. The role of tumor-associated macrophages in leukemia. Acta Haematol. 2020;143(2):112117. 8. Wrangle JM, Patterson A, Johnson CB, et al. IL-2 and beyond in cancer immunotherapy. J Interferon Cytokine Res. 2018;38(2):45-68. 9. Majeti R, Chao MP, Alizadeh AA, et al. CD47 is an adverse prognostic factor and therapeutic antibody target on human acute myeloid leukemia stem cells. Cell. 2009;138(2):286-299. 10. Peña-Martínez P, Eriksson M, Ramakrishnan R, et al. Interleukin 4 induces apoptosis of acute myeloid leukemia cells in a Stat6-dependent manner. Leukemia. 2018;32(3):588-596. 11. Li Z, Chen L, Qin Z. Paradoxical roles of IL4 in tumor immunity. Cell Mol Immunol. 2009;6(6):415-422. 12. Kiniwa T, Enomoto Y, Terazawa N, et al. NK cells activated by Interleukin-4 in cooperation with Interleukin-15 exhibit distinctive characteristics. Proc Natl Acad Sci U S A. 2016;113(36):10139-10144. 13. Paul WE. History of interleukin-4. Cytokine. 2015;75(1):3-7. 14. 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. 15. Eriksson M, Peña-Martínez P, Ramakrishnan R, et al. Agonistic targeting of TLR1/TLR2 induces p38 MAPK-dependent apoptosis and NFkB-dependent differentiation of AML cells. Blood Adv. 2017;1(23):2046-2057.

824

16. Miller PG, Al-Shahrour F, Hartwell KA, et al. In vivo RNAi screening identifies a leukemia-specific dependence on integrin beta 3 signaling. Cancer Cell. 2013; 24(1):45-58. 17. Järås M, Miller PG, Chu LP, et al. Csnk1a1 inhibition has p53-dependent therapeutic efficacy in acute myeloid leukemia. J Exp Med. 2014;211(4):605-612. 18. Piganelli JD, Martin T, Haskins K. Splenic macrophages from the NOD mouse are defective in the ability to present antigen. Diabetes. 1998;47(8):1212-1218. 19. Ito M, Hiramatsu H, Kobayashi K, et al. NOD/SCID/gcnull mouse: an excellent recipient mouse model for engraftment of human cells. Blood. 2002;100(9):3175. 20. Claassen I, Van Rooijen N, Claassen E. A new method for removal of mononuclear phagocytes from heterogeneous cell populations in vitro, using the liposome-mediated macrophage 'suicide' technique. J Immunol Methods. 1990;134(2):153-161. 21. Qian Q, Jutila MA, Van Rooijen N, Cutler JE. Elimination of mouse splenic macrophages correlates with increased susceptibility to experimental disseminated candidiasis. J Immunol. 1994;152(10):50005008. 22. Loke Pn, Nair MG, Parkinson J, Guiliano D, Blaxter M, Allen JE. IL-4 dependent alternatively-activated macrophages have a distinctive in vivo gene expression phenotype. BMC Immunol. 2002;3:7. 23. Martinez FO, Helming L, Milde R, et al. Genetic programs expressed in resting and IL-4 alternatively activated mouse and human macrophages: similarities and differences. Blood. 2013;121(9):e57-e69. 24. Gundra UM, Girgis NM, Ruckerl D, et al. Alternatively activated macrophages derived from monocytes and tissue macrophages are phenotypically and functionally distinct. Blood. 2014;123(20):e110. 25. Haas L, Obenauf AC. Allies or enemies-the multifaceted role of myeloid cells in the tumor microenvironment. Front Immunol. 2019;10:2746. 26. Chen Y, Zhang X. Pivotal regulators of tissue homeostasis and cancer: macrophages. Exp Hematol Oncol. 2017;6:23. 27. Feng M, Chen JY, Weissman-Tsukamoto R,

et al. Macrophages eat cancer cells using their own calreticulin as a guide: roles of TLR and Btk. Proc Natl Acad Sci U S A. 2015;112(7):2145-2150. 28. Jaiswal S, Chao MP, Majeti R, Weissman IL. Macrophages as mediators of tumor immunosurveillance. Trends Immunol. 2010;31(6):212-219. 29. Loyher P-L, Hamon P, Laviron M, et al. Macrophages of distinct origins contribute to tumor development in the lung. J Exp Med. 2018;215(10):2536-2553. 30. Gordon S, Martinez FO. Alternative activation of macrophages: mechanism and functions. Immunity. 2010;32(5):593-604. 31. Sica A, Mantovani A. Macrophage plasticity and polarization: in vivo veritas. J Clin Invest. 2012;122(3):787-795. 32. Jenkins SJ, Ruckerl D, Thomas GD, et al. IL-4 directly signals tissue-resident macrophages to proliferate beyond homeostatic levels controlled by CSF-1. J Exp Med. 2013;210(11):2477. 33. Erb KJ, Rüger B, von Brevern M, Ryffel B, Schimpl A, Rivett K. Constitutive expression of interleukin (IL)-4 in vivo causes autoimmune-type disorders in mice. J Exp Med. 1997;185(2):329-339. 34. Milner JD, Orekov T, Ward JM, et al. Sustained IL-4 exposure leads to a novel pathway for hemophagocytosis, inflammation, and tissue macrophage accumulation. Blood. 2010;116(14):2476-2483. 35. La Rosée P, Horne A, Hines M, et al. Recommendations for the management of hemophagocytic lymphohistiocytosis in adults. Blood. 2019;133(23):2465-2477. 36. Barkal AA, Weiskopf K, Kao KS, et al. Engagement of MHC class I by the inhibitory receptor LILRB1 suppresses macrophages and is a target of cancer immunotherapy. Nat Immunol. 2018; 19(1):76-84. 37. Feng M, Marjon KD, Zhu F, et al. Programmed cell removal by calreticulin in tissue homeostasis and cancer. Nat Commun. 2018;9(1):3194. 38. Betancur PA, Abraham BJ, Yiu YY, et al. A CD47-associated super-enhancer links proinflammatory signalling to CD47 upregulation in breast cancer. Nat Commun. 2017; 8:14802.

haematologica | 2022; 107(4)

ARTICLE

Acute Myeloid Leukemia

Pevonedistat and azacitidine upregulate NOXA (PMAIP1) to increase sensitivity to venetoclax in preclinical models of acute myeloid leukemia Dan Cojocari,1* Brianna N Smith,2-4*Julie J. Purkal,1 Maria P. Arrate,3 Jason D. Huska,1 Yu Xiao,1 Agnieszka Gorska,3 Leah J. Hogdal,5 Haley E. Ramsey,3,4 Erwin R. Boghaert,1 Darren C. Phillips1# and Michael R. Savona3,4,6,7# 1

Oncology Discovery, AbbVie Inc., North Chicago, IL; 2Department of Pediatrics, 3 Medicine, and 4Program in Cancer Biology, Vanderbilt University School of Medicine, Nashville, TN; 5Precision Medicine, AbbVie Inc., North Chicago, IL; 6Vanderbilt-Ingram Cancer Center, Nashville, TN and 7Center for Immunobiology, Vanderbilt University School of Medicine, Nashville, TN, USA

Ferrata Storti Foundation

Haematologica 2022 Volume 107(4):825-835

*DC and BNS contributed equally as co-first authors. # DCP and MRS contributed equally as senior authors.

ABSTRACT

ysregulation of apoptotic machinery is one mechanism by which acute myeloid leukemia (AML) acquires a clonal survival advantage. B-cell lymphoma protein-2 (BCL2) overexpression is a common feature in hematologic malignancies. The selective BCL2 inhibitor, venetoclax (VEN) is used in combination with azacitidine (AZA), a DNAmethyltransferase inhibitor (DNMTi), to treat patients with AML. Despite promising response rates to VEN/AZA, resistance to the agent is common. One identified mechanism of resistance is the upregulation of myeloid cell leukemia-1 protein (MCL1). Pevonedistat (PEV), a novel agent that inhibits NEDD8-activating enzyme, and AZA both upregulate NOXA (PMAIP1), a BCL2 family protein that competes with effector molecules at the BH3 binding site of MCL1. We demonstrate that PEV/AZA combination induces NOXA to a greater degree than either PEV or AZA alone, which enhances VEN-mediated apoptosis. Herein, using AML cell lines and primary AML patient samples ex vivo, including in cells with genetic alterations linked to treatment resistance, we demonstrate robust activity of the PEV/VEN/AZA triplet. These findings were corroborated in preclinical systemic engrafted models of AML. Collectively, these results provide rational for combining PEV/VEN/AZA as a novel therapeutic approach in overcoming AML resistance in current therapies.

Introduction Acute myeloid leukemia (AML) is a heterogeneous hematopoietic neoplasm characterized by the arrest of differentiation and clonal proliferation of myeloid precursor cells. Despite recent advances, only 24% of patients survive their disease beyond 5 years of diagnosis.1–6 One mechanism by which AML clones acquire a survival advantage is the dysregulation of apoptotic machinery that in normal cells regulates homeostasis in the bone marrow. Apoptosis is controlled at the mitochondrial level via mechanisms regulated by the B-cell lymphoma protein-2 (BCL2) family of proteins.7 This group of proteins is divided into three sub-groups which include pro-apoptotic “BH3-only” proteins BIM, BID, PUMA, NOXA, BAD, BIK, BMF and HRK, and the effector proteins BAX and BAK that are induced by various cell death stimuli to trigger mitochondrial outer membrane permeabilization (MOMP) and apoptosis.8,9 A separate group of anti-apoptotic family members, including BCL2, BCL-XL, MCL1, BCL-W, and BCL2-A1, function to bind and inhibit the pro-apoptotic family members, preventing MOMP and subsequently apoptosis.8,10,11 BCL2 overexpression in human hematologic malignancies is associated with poor responses to conventional chemotherapy.12,13 A deep understanding

haematologica | 2022; 107(4)

Correspondence: MICHAEL R. SAVONA michael.savona@vanderbilt.edu Received: September 17, 2020. Accepted: April 7, 2021. Pre-published: April 15, 2021. https://doi.org/10.3324/haematol.2020.272609

825

D. Cojocari et al.

of mitochondrial protein control of apoptosis has fueled a desire to discover small molecules specifically intended to occupy the hydrophobic BH3 binding site of antiapoptotic proteins and allow initiation of apoptosis by effector molecules.14,15 Venetoclax (VEN), a selective BCL2 inhibitor, binds BCL2 to directly inhibit sequestration of pro-apoptotic proteins such as the activator BIM. Free BIM can bind to and activate BAX/BAK, inducing conformational changes that result in BAX/BAK homo-oligomerization and mitochondrial outer membrane permeabilization (MOMP), initiating apoptosis.16,17 VEN has been effective in the clinic for patients with chronic lymphocytic leukemia but has limited efficacy in relapsed-refractory AML as a single agent.18,19 In recent clinical trials in newly-diagnosed AML patients ineligible for intensive chemotherapy, VEN demonstrated an overall response rate of 67% or 48% in combination with DNA methyltransferase inhibitors (DNMTi; AZA or decitabine) or low-dose cytosine arabinoside (LDAC), respectively.20,21 Despite this progress, many patients treated with VEN+DNMTi/LDAC ultimately relapse, and a subset of patients never respond.20–23 One proposed mechanism of resistance to VEN is cellular upregulation of the anti-apoptotic protein myeloid cell leukemia-1 protein (MCL1). VEN in combination with selective MCL1 inhibitors has demonstrated enhanced cytotoxic activity over either agent alone in AML cells in vitro and in xenograft models pre-clinically, but efficacy of this combination in the clinic has yet to be reported.24 Both MCL1-dependent and MCL1-independent mechanisms of VEN resistance are emerging and new approaches aimed at addressing these are moving toward the clinic.25-29 Pevonedistat (PEV) was developed as a targeted inhibitor of NEDD-8 activating enzyme (NAE), which activated the Cullin-RING E3 ubiquitin ligases (CRL) in a process called neddylation. Thus, PEV disrupts the proteasomal-mediated degradation of proteins targeted by the CRL, leading to their accumulation.30 Neddylation is upregulated in AML, and pevonedistat has been tested as a single agent and in combination with the DNMTi, AZA.31-33 In a phase Ib clinical trial of untreated AML patients ≥60 years of age, an intention to treat analysis revealed PEV in combination with AZA induced an overall response rate of 50%.31 Interestingly, PEV and AZA both upregulate NOXA (PMAIP1), a pro-apoptotic BCL2 family protein known to compete with effector molecules at the BH3 binding site of MCL1 to inhibit its anti-apoptotic function.10,34,35 Given the emerging critical role of MCL1 in VEN resistance, we postulated that the combination of PEV/AZA will synergize with VEN in a triple combination with efficacy superior to VEN/AZA or VEN/PEV alone. Herein, we demonstrate that the triple combination of VEN/PEV/AZA induces robust activity in preclinical models of AML that is superior to either agent alone or as combination doublets. Leveraging AML cell lines and primary AML patient samples ex vivo, we demonstrate that the PEV/AZA combination induces NOXA to a greater extent than PEV or AZA alone to further enhance VEN-mediated apoptosis. Apoptosis induced by the VEN/PEV/AZA combination required PMAIP1, the gene encoding NOXA, since its deletion abrogated the activity of this triplet. Importantly, the VEN/PEV/AZA combination enhanced the kinetics of apoptosis compared to other treatment variations in AML cell lines and patient samples that manifest in vivo to drive durable anti-leukemic activity in cell 826

line-derived and patient-derived xenograft models of AML. Our work provides important mechanistic insight to support the use of these agents in combination in ongoing clinical trials.

Methods Cell culture and reagents AML cell lines and patient cells were cultured as previously described, and were short tantem repeats (STR) validated and mycoplasma tested.34 See the Online Supplementary Materials and Methods for reagent details.

Patient samples Experiments were conducted on primary patient samples provided by the Vanderbilt-Ingram Cancer Center Hematopoietic Malignancies Repository, after obtaining informed consent, and approval of the Vanderbilt University Medical Center Institutional Review Board.

Knockout cell line generation Genomic deletion of BBC3 and PMAIP1 were previously described.34 The BAK1/BAX-deficient OCI-AML5 cells were generated using the same protocol except by combining the two CRISPR RNA (crRNA) targeting BAK1/BAX. ATF4-deficient cells were generated by combining the two crRNA (Online Supplementary Table S1).

Cell proliferation assay For combinatorial studies, cells were treated for 24 hours as previously described.34 For primary cells, compounds were dispensed into a 384-well plate using the Echo 555 liquid handler (Labcyte) and 2,000-8,000 cells/well were incubated for 24 hours or 72 hours for VEN/AZA resistant patient samples and cell viability was measured using the Cell TiterGlo (Promega). Percent viability was defined as the relative luminescence units (RLU) of each well divided by the RLU of cells in dimethyl sulfoxide (DMSO) control and the effective concentration (EC50) to induce 50% cell death were determined by nonlinear regression algorithms using Prism 8.0 (GraphPad).

Western blot Western blotting was performed as previously described.34 Membranes were incubated with the respective primary antibodies (Online Supplementary Table S2).

Patient-derived xenografts 2×106 primary AML mononuclear cells were engrafted in 7-9week-old female NSGS (NOD-scid IL2Rgnull3Tg(hSCF/hGMCSF/hIL3)) mice (The Jackson Laboratory) as described in the Online Supplementary Materials and Methods. Chimerism was assessed weekly in the peripheral blood, and, at time of tissue harvest, in the bone marrow and spleen. Animal experiments were conducted in accordance with guidelines approved by the IACUC at VUMC.

Cell line-derived xenografts Female fox chase SCID beige mice were employed for the flank xenograft model; female NOD SCID gamma mice were employed in the systemic engraftment of OCI-AML2-Red-Fluc cell line (Charles River Laboratories) and performed as described in the Online Supplementary Materials and Methods. Animal studies were conducted in accordance with the guidelines established by the AbbVie Institutional Animal Care and Use Committee.

haematologica | 2022; 107(4)

PEV/VEN/AZA therapy in preclinical AML

Flow cytometry Red blood cells were lysed with EL buffer on ice (Qiagen), with remaining cells washed and resuspended in 1X phosphate buffered saline (PBS) with 1% bovine serum albumin (BSA) and stained for 15 minutes with the conjugated antibodies listed in the Online Supplementary Table S2. Cells were washed and submitted for flow cytometric analysis using a 3-laser LSRII (Becton Dickinson).

Immunohistochemistry Tissues were fixed in 4% paraformaldehyde for 48 hours and stored in 70% ethanol, embedded in paraffin and sectioned at 5 µm after bone tissue was decalcified. Sections were dewaxed in Xylene and rehydrated in successive ethanol baths. Standard Mayer’s hematoxylin and eosin (H&E) staining was performed. Antigen retrieval using a standard pH 6 sodium citrate buffer (BioGenex) was performed and sections were stained with monoclonal mouse anti-human CD45 (Dako, M0701, dilution 1:200) using M.O.M. Kit (Vector).

Kinetics of caspase-3/-7 activation and cell death Activated caspase-3/-7 (aCasp-3/-7+) and cell death (DRAQ7+) was evaluated using an IncuCyte S3 (Sartorius) as described in detail previously.34 Area under the curve (AUC) was calculated over the 48 hours period using Prism 8.0 (GraphPad).

Gene expression 100,000 AML cells (OCI-AML5, MV4-11, THP-1) were cultured in 96-well plates three biological replicates were treated with pevonedistat for 24 hours. QuantiGene Plex assay (SigmaAldrich ) was performed as instructed by the manufacturer.

Statistical analysis An unpaired 2-tailed Student t-test was performed unless otherwise indicated. All statistical analyses were completed using Prism 8.0 software (GraphPad). The effects of VEN/PEV combinatorial activity were calculated using the zero interaction potency (ZIP) synergy model, which compares observed and expected combination effects.36–38

Results Deletion of PMAIP1 abrogates the synergistic activity of pevonedistat and venetoclax in acute myeloid leukemia cell lines We tested a panel of AML cell lines for the combinatorial activity of VEN and PEV using a 10 by 3 concentration matrix, respectively (Online Supplementary Figure S1A). The two agents were synergistic in 15 of 18 and highly synergistic (d>5%) in nine of 18 AML cell lines (Figure 1A) as assessed by the ZIP synergy model.38 In order to further validate this in vitro observation, a subcutaneous xenograft model of the MV4-11 cell line was established in the flank of immunocompromised mice. The mice were divided into four treatment groups and treated daily for two weeks with either PEV, VEN, or the combination of PEV and VEN. Both VEN and PEV each moderately inhibited MV4-11 tumor growth to similar degrees. However, the two agents together displayed combinatorial activity resulting in pronounced tumor regression and a tumor growth delay (Figure 1B). In order to explore the molecular drivers behind the VEN/PEV synergistic activity, we measured changes in

haematologica | 2022; 107(4)

the BCL2 family proteins and gene expression in AML cell lines. Treatment with PEV caused a robust upregulation in NOXA protein levels in all five AML cell lines (Figure 1C), and increased gene expression of its transcript, PMAIP1, in the three of the AML cell lines assessed (Figure 1D). The BH3-only protein PUMA was inconsistently induced across these cell lines at the protein or transcript level (Figure 1C and D). In contrast to the BH3-only pro-apoptotic members, MCL1 was downregulated in all cell lines treated with PEV (Figure 1C). In order to establish the importance of NOXA for either PEV or VEN/PEV activity, we utilized three PMAIP1-deficient AML cell lines.34 Deletion of PMAIP1 reduced PEVinduced death in OCI-AML5, Kasumi-1, and MV4-11 cells (Figure 1E; Online Supplementary Figure S1B) and abrogated the synergistic activity of VEN/PEV in all three cell lines (Figure 1F and G). These findings illuminate the importance of NOXA for the synergistic activity between VEN and PEV.

Venetoclax/pevonedistat/azacitidine induces BAX/BAK-dependent apoptosis in acute myeloid leukemia cell lines Previously, we demonstrated that NOXA is essential for the combinatorial activity between VEN and AZA in AML cell lines, providing an insight into the mechanism of VEN/AZA clinical activity in AML patients.34 We explored the impact of adding PEV to the VEN/AZA combination to AML cell lines in vitro. A panel of seven AML cell lines were treated with ten doses of VEN and three doses of AZA. Upon the addition of PEV at 100 nM or 370 nM, the overall cell viability decreased in the VEN/AZA combination matrix, indicating an added benefit of PEV to the VEN/AZA cell killing in all seven cell lines (Figure 2A), including in THP-1 cells where VEN/AZA was not synergistic but VEN/PEV was (Figure 1A).34 Similarly, the addition of AZA to VEN/PEV increased cell death in SET-2 and U937 cells (Figure 2A). Thus, VEN/PEV/AZA triple combination improves the activity observed with either the VEN/PEV or VEN/AZA treatments, overcoming the inherent resistance of some AML cells to these treatment doublets. In order to explore the kinetics of cell death with the VEN/AZA/PEV combination, caspase-3/-7 activation and cell death (DRAQ7 uptake) were measured over time following the addition of individual agents, the double combinations, or the triple combination. The triple VEN/AZA/PEV treatment induced a more rapid activation of caspase-3/-7 and cell death in MV4-11 (Figure 2B) and Kasumi-1 (Online Supplementary Figure 2A) compared to other variations of this treatment combination. Similarly, the overall levels of caspase-3/-7 and cell death over the 48-hour period, as measured by the area under the kinetic curve (AUCk), was significantly higher than the singlet or doublet treatments (Figure 2C). We asked whether the cell death induced by the triplet is dependent upon the intrinsic apoptosis signaling pathway. OCI-AML5 cells deficient in both BAX and BAK1 were generated using CRISPRCas9 clonal gene editing and the absence of BAX and BAK protein expression confirmed (Figure 2D). The triplet combination treatment in parental OCI-AML5 cells induced significant caspase-3/-7 activation and cell death, that was completely abrogated by BAX/BAK1 deletion (Figure 2E; Online Supplementary Figure S2B).

827

D. Cojocari et al. A

Figure 1. Deletion of PMAIP1 abrogates the synergistic activity of pevonedistat and venetoclax in acute myeloid leukemia cell lines. (A) Synergistic activity between venetoclax (VEN) and pevonedistat (PEV) as determined by the zero interaction potency (ZIP) model following 24 hours of treatment of a panel of acute myeloid leukemia (AML) cell lines. (B) Tumor growth over time in the MV4-11 subcutaneous xenograft model treated (black bar) with vehicle control, VEN (50 mg/kg; 14-day daily [QDx14]; orally [PO]), PEV (60 mg/kg; QDx14; intraperitoneal [IP]), or the two agents in combination (QD×14). Data are presented as the mean tumor volume ± standard error of the mean (SEM) from eight mice per treatment group. (C) Western blot analysis of NOXA, PUMA, BCL2, BCL-XL, and MCL1 proteins in a panel of AML cell lines treated with PEV (OCI-AML5: 1 mM; Kasumi-1: 1.5 mM; MV4-11: 0.4 mM; OCI-AML2: 0.15 mM; U937: 2 mM) for 24 hours. (D) Gene expression analysis of PMAIP1, BBC3, BCL2L1, MCL1, and BCL2 in OCI-AML5, MV4-11, and THP-1 cells following treatment with PEV at the indicated concentrations (OCI-AML5: 0.4 mM and 1 mM; MV4-11: 0.2 mM and 0.4 mM) after 24 hour of treatment relative to the dimethyl sulfoxide (DMSO) control cells (t-test, *P<0.05, n=3). (E) Cell viability of three parental or PMAIP1-/- AML cell lines after 24 hours of treatment with VEN and PEV in combination (n=3). (F) Synergy measured by ZIP model metric for the VEN and PEV combination in three parental or PMAIP1-/- AML cell lines after 24 hours of treatment (n=3) (lower panel). (G) Visualization of the calculated 2D ZIP synergy map for the parental and PMAIP1-/- MV4/11 from (F).

828

haematologica | 2022; 107(4)

PEV/VEN/AZA therapy in preclinical AML

Azacitidine- and pevonedistat-induced NOXA contributes to the combinatorial activity with venetoclax NOXA is important for efficient VEN, PEV or AZA single-agent activity, as well as for VEN/AZA34 or VEN/PEV synergy (Figure 1F).35 However, PMAIP1 is induced by AZA or PEV through distinct mechanisms. AZA treatment induces the integrated stress response (ISR) pathway to upregulate PMAIP1, whereas PEV stabilizes the transcrip-

tion factor ATF4 to enhance PMAIP1 expression.34,39 In order to understand the effects of combining AZA and PEV on the ISR’s ATF4-NOXA axis, we treated the parental and PMAIP1 deficient OCI-AML5 cell lines for 24 hours with either agent alone or in combination. In parental cells, AZA or PEV alone or in combination induced ATF4 or its transcriptional targets CHOP and NOXA and led to apoptosis, as measured by levels of cleaved PARP (Figure 3A). Similarly, AZA/PEV combination induced DDIT3 and

Figure 2. Adding azacitidine to venetoclax/pevonedistat treatment significantly decreases acute myeloid leukemia cell line viability in vitro that is dependent on BAX/BAK-mediated apoptosis. (A) Cell viability matrix measure in a panel of acute myeloid leukemia (AML) cell lines following 24 hours of treatment with venetoclax (VEN) (10 mM top dose 1:3 dilution, except OCI-AML2, MV4-11 top dose of 300 nM) and azacitidine (AZA) (0.3 and 1 mM) and pevonedistat (PEV) at the indicated doses. (B) Activated caspase-3/-7 positive cells (aCasp-3/-7+) or dead cells (DRAQ7+) were counted over time following treatment with the VEN (1 nM), AZA (1 mM), PEV (100 nM), or indicated combinations of these compounds in MV4-11 cells (n=5). (C) Area under the kinetic curves (AUCk) from (B) was calculated over a 48-hour period and plotted as total positive cells (n=5). (D) Western blot analysis of BAX and BAK expression in the parental and BAX-/-/BAK1-/- OCI-AML5 cell lines. (E) OCIAML5 cells were treated with VEN (10 nM), AZA (3 mM), PEV (100 nM) or indicated combinations, and the AUC of the total aCasp-3/-7+ or DRAQ7+ cells calculated over a 48-hour time period were plotted as total aCasp-3/-7+ (left) or DRAQ7+ (right) cells positive cells (n=5).

haematologica | 2022; 107(4)

829

D. Cojocari et al.

Figure 3. Azacitidine- and pevonedistat-induced NOXA contributes to the combinatorial activity with venetoclax. (A) Western blot analysis of total PARP, ATF4, CHOP, and NOXA protein in the parental and the PMAIP1-/- OCI-AML5 cell line treated with azacitidine (AZA) (1 mM), pevonedistat (PEV) (0.37 mM) or both for 24 hours. b-actin was used a protein loading control. (B) DDIT3 and PMAIP1 gene expression was measured in cell lines from (A) under the same conditions. (C) Western blot analysis of ATF4, CHOP, eIF2a, phosphor eIF2a Ser51 and NOXA protein in the parental and the ATF4-/- OCI-AML5 cell line treated as indicated in (A). (D) Cell viability of OCI-AML5 cell line following 24 hours of treatment with venetoclax (VEN) (10 mM top dose 1:3 dilution) and AZA (0.3 mM, 1 mM and 3 mM) and PEV (370 nM) (n=3). (E) Cell viability as measured by area under the curve (AUC) of the VEN dose response curve (from C and Online Supplementary Figures S3A and B) of three parental or PMAIP1-/- acute myeloid leukemia (AML) cell lines following 24 hours of treatment with VEN (10 mM top dose 1:3 dilution, MV4-11 top dose of 0.3 mM) and AZA (0.3 mM, 1 mM and 3 mM) and PEV 100 nM and 370 nM (n=3). (F) Western blot analysis of total PARP, BAX, BAK, MCL1, NOXA protein in the parental and the BAX-/-/BAK1-/- OCI-AML5 cell line treated as indicated in (A). DMSO: dimethyl sulfoxide.

830

haematologica | 2022; 107(4)

PEV/VEN/AZA therapy in preclinical AML

PMAIP1 transcripts to a greater level than the single agent treatments in OCI-AML5 and Kasumi-1 cells (Figure 3B). In contrast, the PMAIP1-deficient OCI-AML5 cells displayed reduced levels of cleaved PARP, while the ATF4 and CHOP induction was not affected by the gene deletion (Figure 3A). Upon ATF4 ablation in OCI-AML5, either agent alone, or the combination, was unable to induce CHOP or NOXA (Figure 3C). Furthermore, unlike AZA, PEV did not induce significant phosphorylation of eIF2a at Ser 51, a marker of ISR activation, while still inducing ATF4 and NOXA. This suggests PEV stabilizes ATF4 in the absence

of ISR. Thus, ATF4 is required for the induction of NOXA by AZA/PEV. In order to understand if NOXA is critical for the PEV/VEN-mediated sensitization of AML cell lines to VEN, we treated the parental and PMAIP1-deficient cell lines with the VEN/PEV/AZA triple combination and measured cell viability after 24 hours. Relative to the parental cell lines, PMAIP1-deficiency reduced the potency of VEN-mediated cell death alone or in combination with AZA, PEV, or AZA/PEV in OCI-AML5 (Figure 3D and 3E), Kasumi-1 (Online Supplementary Figure 3A) and MV4-11 (Online Supplementary Figure 3B) cell lines. NOXA can pro-

Figure 4. The venetoclax/pevonedistat/azacitidine triple-combination treatment induces durable responses in a systemic xenograft model of acute myeloid leukemia. (A) Tumor growth from whole-body ROI bioluminescent signal (total photons/second) of systemically engrafted OCI-AML2-Red-Fluc tumor cells measured over time and treated (black bar) with vehicle control, venetoclax (VEN) (50 mg/kg, 14-day daily [QDx14], orally [PO]), AZA (8 mg/kg, every 7 days for three treatments [Q7D×3], intravenously [IV]), pevonedistat (PEV) (60 mg/kg, QD×14, intraperitoneal [IP]) or the combinations of the two or all three agents (n=6-8 mice per treatment group). (B) Representative images of animals’ bioluminescent signal of the treatment cohorts from panel (A) showing significant delay in in vivo acute myeloid leukemia (AML) growth with either single- or dual-treatment combinations, while no growth (BLI signal) was detected in the triple-combination treatment.

haematologica | 2022; 107(4)

831

D. Cojocari et al.

Figure 5. The venetoclax/pevonedistat/azacitidine triple-combination treatment is efficacious in preclinical primary acute myeloid leukemia models. (A) Primary acute myeloid leukemia (AML) samples from seven different patients with different mutational profiles treated ex vivo for 24 or 72 hours (venetoclax/azacitidine [VEN/AZA] resistant patients) with pevonedistat (PEV) (0.3 mM), AZA (0.3 mM) and VEN (0.01 mM) alone, and combinations. (B) Western blot analysis of a primary AML patient samples after 12 hours (AML008) or 24 hours (AML004, AML005, AML006 and AML009) of treatment with dimethyl sulfoxide (DMSO) control, AZA alone, PEV alone, and PEV/AZA in combination. Protein levels of BCL2 family members MCL1, BCL2, NOXA, cleaved PARP, and b-actin. (C) Mice were injected with 2x106 patient primary AML 004 cells via tail vein on day 1, 24 hours after cesium irradiation and were treated (black bar) with vehicle, PEV (30 mg/kg; every other day for 28 days [QOD×28]), AZA (1.5 mg/kg; once every day for 7 days [QD×7]), VEN (15 mg/kg; once every day for 28 days [QDx28]), VEN/AZA or triple combination VEN/AZA/PEV. Human CD45 positive (hCD45+) cells were measured weekly by flow cytometry in peripheral blood (PB) from week 2 through week 12 post-transplant (t-test, *P<0.05). (D) Percent of hCD45+ cells in bone marrow and spleen tissue were measured via flow cytometry on day of tissue harvest week 12 post-transplant. (E) Immunohistochemistry of bone marrow (femur) and spleen (20X), stained with monoclonal antibody for hCD45 in experimental mice.

mote the degradation of MCL1 by the proteosome and executioner caspases can also cleave MCL1 during apoptosis.40 In order to understand the mechanism of MCL1 loss upon treatment with either AZA or PEV, the OCI-AML5 cell line deficient in BAX and BAK1 were treated with these agents (Figure 3F). MCL1 protein degradation induced by either AZA or PEV was not observed in the OCI-AML5 cell lines deficient in either PMAIP1 (Figure 3A), or BAX/BAK1 (Figure 3E). In the BAX-/- BAK1-/- cells, MCL1 was not degraded despite NOXA upregulation (Figure 3F), thus MCL1 degradation by either AZA or PEV requires both sensitization by NOXA and activation of apoptosis by BAX/BAK. Taken together, these data demonstrate that NOXA can be induced by both AZA and PEV and that it has a central role in driving their combinatorial activity with VEN. 832

The venetoclax/pevonedistat/azacitidine combination treatment is highly active in preclinical models of acute myeloid leukemia The in vivo efficacy of the triple combination of PEV, VEN and AZA was tested in a systemic murine model that allowed monitoring the tumor burden of xenografted bioluminescent OCI-AML2-Red-Fluc cells. Animals were distributed into eight treatment cohorts and treated with either vehicle alone, VEN (50 mg/kg, QD×14, orally [PO]), 5-Aza (8 mg/kg, Q7D×3, intravenous [IV]), PEV (60 mg/kg, QD×14, intraperitoneal [IP]) or the combinations of the two or all three agents. Animals were imaged at regular intervals until they reached endpoint (1×1010 photons/second). The single-agent treatments were active in this model, and the doublets were more efficacious than the single-agents (Figure 4A). Most notably, the triplet combihaematologica | 2022; 107(4)

PEV/VEN/AZA therapy in preclinical AML

nation of PEV, VEN and AZA demonstrated enhanced tumor growth inhibition compared to the singlet or doublet treatment cohorts, driving durable responses with no evidence of tumor growth (zero of eight mice) by day 117, 93 days after treatment had ceased (Figure 4A and B). Importantly, none of the cohorts experienced a significant decrease in body weight (Online Supplementary Figure 4A). In order to further explore the utility of the VEN/AZA/PEV triple combination, we treated a panel of primary AML patient samples with different mutational, karyotype profiles, and treatment failure (Online Supplementary Table S3) with PEV, AZA and VEN alone, and combinations of these agents (Figure 5A). Improved combinatorial activity was noted in the triplet combination for all patient samples when compared to VEN/AZA. In order to test whether PEV can also improve efficacy in patient samples with significant resistance to VEN/AZA in vitro, we tested four patients with in vitro resistance to the doublet. In this subset of patients, the triplet combination also demonstrated superior efficacy compared to VEN/AZA alone (Figure 5A). Western blot analysis of five primary AML patient samples treated with PEV/AZA in combination for 12 or 24 hours increased NOXA protein expression further than either drug alone, and coincided with a decrease in MCL1 protein without impacting BCL2 expression (Figure 5B). In order to validate our ex vivo findings in a patient-derived xenograft model, NSGS mice were transplanted with de novo cells from patient AML 004. Next generation sequencing of this patient’s AML revealed a complex mutational profile including mutations in FLT3ITD, NPM1, IDH2, and DNMT3A. Chimerism was established, and 5 weeks post-transplant we began treating the mice with PEV, AZA, VEN, combination VEN/AZA (the current standard of care) or the triplet combination at subtherapeutic doses of 30 mg/kg of PEV, 1.5 mg/kg of AZA, and 15 mg/kg of VEN for 28 days. Vehicle-treated mice rapidly succumbed to leukemia by week 12 post-transplant (Figure 5C). At this point, the experiment was concluded, and tissue was harvested for evaluation of human chimerism in the bone marrow and spleen (Figure 5D). The triplet combination resulted in decreases of tumor burden in the bone marrow (0.6±0.3%) beyond any single agent treatment (P<0.05; VEN: 79.0±7.8%; PEV: 40.8±7.5%; AZA: 69.9±12.6%) or VEN/AZA alone (P<0.05; 8.5±5.8%) (Figure 5D). Staining with anti-hCD45 antibody revealed lowest AML chimerism in the group treated with the triplet combination (Figure 5E). Normal tissues were unaffected in this experiment, and combination of the three agents neither led to any significant effects on the weight of the mice (Online Supplementary Figure S4B), nor exhibited signs of stress. Ex vivo study of the triplet combination did not lead to a significant reduction in total colony formation in normal human CD34+ cells (Online Supplementary Figure S5).

Discussion Novel drug combinations of VEN with DNMTi or LDAC, have led to dramatic improvements in responses in patients with AML.20,22 Despite these advances, some patients do not respond or will relapse on this therapy emphasizing a need to develop novel treatment strategies. We explored the impact of adding the NAE inhibitor, PEV, to the clinically relevant treatment regihaematologica | 2022; 107(4)

men of VEN/AZA in experimental models of AML. We demonstrate that the VEN/PEV/AZA triple combination requires the induction of the BH3-only protein NOXA to enhance the kinetics and depth of apoptosis in AML cell lines in vitro when compared to other treatment variations of these drug components. These observations are reflected in vivo, where the VEN/PEV/AZA combination triplet induces durable anti-leukemic responses in systemic cell line-derived and patient-derived xenograft models of AML. Although the addition of VEN to AZA in newly diagnosed AML patients ineligible for intensive chemotherapy improved overall survival (14.7 months VEN/AZA vs. 9.6 months AZA), a subset of AML patients present with shorter durations of response or never respond to treatment.23,41 A previous report suggested the potential of PEV to synergize with VEN in cell lines in vitro.35 In order to build upon these observations and to explore the mechanism of this interaction, we extensively explored the potential combinatorial activity extensively with, and without AZA. VEN and AZA are effective in most AML cell lines in vitro and reflect clinical observations of diminished responses in cases of mutant TP53.34,41,42 Adding PEV to the VEN/AZA treatment paradigm enhanced cell death in AML cell lines and patient samples, consistent with our previous finding of NOXA induction in AML exposed to AZA,34 and consistent with synergistic NOXA induction with the combination of PEV and AZA. This coincided with diminished MCL1 protein which the primary target of NOXA in the mitochondria. To that end, we saw enhanced killing of AML in cell lines and patient samples in which VEN/PEV or then VEN/AZA combination activity was limited. In order to capture the response heterogeneity of AML we tested the triple combination in a panel of primary AML samples from patients found ultimately to be refractory to conventional chemotherapy, and/or VEN/DNMTi treatment (Online Supplementary Table S3). The triplet combination was found to effectively kill the malignant AML cells in ex vivo assays and in a disseminated AML patient-derived xenograft (PDX) model (AML004). Of note, we observed ex vivo activity of all three agents in the AML003 and AML005 patient cells even though in the clinic these patients eventually failed VEN/DNMTi treatment. Considering the heterogeneity that characterizes AML and the clonal selection that occurs under existing treatment strategies that can eventually contribute to disease relapse,43,44 we speculate that the breadth of activity observed with VEN/PEV/AZA across AML cell lines and patient samples harboring differing genetic mutations may enable more durable responses in AML patients. An additional feature of the combinatorial activity of the VEN/PEV/AZA treatment was the increased rate of caspase-3/-7 activation and the resulting AML cell death over either agent as a monotherapy or as a combination doublet. These in vitro observations translated in vivo where the VEN/PEV/AZA treatment drove durable antileukemic activity in a systemic model of OCI-AML2, with no evidence of tumor growth observed at 93 days posttreatment cessation. Utilizing a PDX model of AML, we demonstrated that the VEN/PEV/AZA combination reduced the leukemic burden within the spleen and bone marrow to a greater extent than VEN, AZA or VEN/AZA in combination. In efforts to minimize potential toxicity, we lowered the doses of each agent used in this study, 833

D. Cojocari et al.

which if administered alone may be sub-therapeutic. However, even at these low dosages, the VEN/PEV/AZA treatment proved safe and effective, offering an opportunity to potentially mitigate hematologic toxicity seen with azanucleosides or LDAC, and in recent combination trials with VEN.23,45,46 PMAIP1 induction is critical to the activity of AZA and PEV, and their respective synergy with VEN;34,35,39 however; the mechanism by which AZA and PEV induce PMAIP1 (NOXA) expression are distinct. PEV inhibition of NAE and the subsequent inactivation of CRL, leads to an accumulation of the CRL substrates.47 Importantly, CRL substrates include the transcription factors MYC and ATF4, which transcriptionally induce the expression of PMAIP1.35,39 Recently, we showed that AZA induced cellular stress and ATF4/NOXA through the upstream activation of the ISR pathway and eIF2a phosphorylation.34 Here, we observed that ATF4, and its transcriptional targets CHOP and NOXA were all induced by either AZA or PEV, which when combined, enhanced NOXA expression further in AML cells and was associated with elevated PARP cleavage and cell death. Deletion of PMAIP1 in this combination decreased the magnitude of apoptosis, measured by PARP cleavage, but not the ATF4/CHOP induction. This implies that NOXA, and not ATF4 or its other targets, is responsible for the ensuing apoptosis. Furthermore, deleting PMAIP1 resulted in significant decrease in cell death induced by the VEN/AZA/PEV triple combination, indicating a critical role for NOXA in the apoptosis-induction mechanism of this combination. NOXA has its greatest affinity for MCL1 over BCL-XL and BCL2, and when upregulated, serves to reduce the anti-apoptotic function of MCL1,48,49 subsequently priming cells to venetoclax-mediated apoptosis. When AZA and PEV are combined, their additive effects on NOXA induction increases the kinetics and depth of venetoclaxmediated caspase-3/-7 activation and cell death that proceeds in a BAX/BAK-dependent manner.

References 1. Ravandi F, Ritchie EK, Sayar H, et al. Vosaroxin plus cytarabine versus placebo plus cytarabine in patients with first relapsed or refractory acute myeloid leukaemia (VALOR): a randomised, controlled, double-blind, multinational, phase 3 study. Lancet Oncol. 2015;16(9):10251036. 2. 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. 3. Cortes JE, Goldberg SL, Feldman EJ, et al. Phase II, multicenter, randomized trial of CPX-351 (cytarabine:Daunorubicin) liposome injection versus intensive salvage therapy in adults with first relapse AML. Cancer. 2015;121(2):234-242. 4. Lancet JE, Cortes JE, Hogge DE, et al. Phase II, multicenter, randomized, open label trial of CPX-351 (cytarabine:daunorubicin) liposome injection versus cytarabine and daunorubicin in patients with untreated AML 60-75 years of age. Blood. 2014; 123(21):3239-3246. 5. Shah A, Andersson TML, Rachet B, Björkholm M, Lambert PC. Survival and

834

While treatment with VEN/AZA in patients with AML has been successful, additional therapies are needed to prevent or rescue patients from relapse of their disease. Our studies provide mechanistic insight as to how addition of PEV to VEN/AZA is synergistic and suggest that this triple combination will be a promising treatment strategy. Clinical trials, including NCT03863257 and NCT04172844, are ongoing to assess the safety and efficacy of the triplet combination in patients with AML. The question of whether sequential dosing with PEV/AZA in cases of VEN resistance, or in cases of de novo MCL1 dependence, is tenable and effective in the clinic remains, and should be explored in future studies. Disclosures MRS receives research funding from Astex, Incyte, Takeda, TG Therapeutics; has equity in Karyopharm; serves on advisory boards and DSMBs for AbbVie, BMS, Celgene, Karyopharm, Novartis, Ryvu, Sierra Oncology, Taiho, Takeda, TG Therapeutics; and consults for Karyopharm and Ryvu. DC, JP, JH, YX, LJH, ERB and DCP are employees of AbbVie. DC, JP, YX, EB & DCP are stockholders of AbbVie Inc. Contributions DC, BNS, DP and MS designed the study, performed experiments and analyzed the data; JP, MA, JH, YZ, AG, LH, HR, EB performed experiments and analyzed data; DP and MS supervised the study. All authors contributed to the writing and critical review of the manuscript and agreed on its submission. Funding BNS receives support from the Department of Health and Human Services National Institutes of Health National Cancer Institute under grant number 5K12CA090625. MRS is a Leukemia and Lymphoma Society Clinical Scholar and is supported by the E.P. Evans Foundation, Adventure Allie Fund, the Biff Ruttenberg Foundation, the Beverly and George Rawlings Directorship, and NIHP30 CA068485-19.

cure of acute myeloid leukaemia in England, 1971-2006: a population-based study. Br J Haematol. 2013;162(4):509-516. 6. Shallis RM, Wang R, Davidoff A, Ma X, Zeidan AM. Epidemiology of acute myeloid leukemia: Recent progress and enduring challenges. Blood Rev. 2019; 36:70-87. 7. Yang E, Korsmeyer SJ. Molecular thanatopsis: A discourse on the BCL2 family and cell death. Blood. 1996;88(2):386-401. 8. Merino D, Kelly GL, Lessene G, Wei AH, Roberts AW, Strasser A. BH3-mimetic drugs: blazing the trail for new cancer medicines. Cancer Cell. 2018;34(6):879-891. 9. Adams JM, Cory S. The BCL-2 arbiters of apoptosis and their growing role as cancer targets. Cell Death Differ. 2018;25(1):27-36. 10. Brunelle JK, Letai A. Control of mitochondrial apoptosis by the Bcl-2 family. J Cell Sci. 2009;122(4):437-441. 11. Del Gaizo Moore V, Letai A. BH3 profilingmeasuring integrated function of the mitochondrial apoptotic pathway to predict cell fate decisions. Cancer Lett. 2013; 332(2):202-205. 12. Campos L, Rouault JP, Sabido O, et al. High expression of bcl-2 protein in acute myeloid leukemia cells is associated with poor response to chemotherapy. Blood.

1993;81(11):3091-3096. 13. Robertson LE, Plunkett W, McConnell K, Keating MJ, McDonnell TJ. Bcl-2 expression in chronic lymphocytic leukemia and its correlation with the induction of apoptosis and clinical outcome. Leukemia. 1996; 10(3):456-459. 14. Pan R, Hogdal LJ, Benito JM, et al. Selective BCL-2 inhibition by ABT-199 causes on-target cell death in acute myeloid Leukemia. Cancer Discov. 2014;4(3):362-675. 15. Niu X, Wang G, Wang Y, et al. Acute myeloid leukemia cells harboring MLL fusion genes or with the acute promyelocytic leukemia phenotype are sensitive to the Bcl-2-selective inhibitor ABT-199. Leukemia. 2014;28(7):1557-1560. 16. Sarosiek KA, Chi X, Bachman JA, et al. BID preferentially activates BAK while BIM preferentially activates BAX, affecting chemotherapy response. Mol Cell. 2013;51(6):751-765. 17. Chonghaile TN, Letai A. Mimicking the BH3 domain to kill cancer cells. Oncogene. 2008;27(0 1):S149-S157. 18. 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. 19. Konopleva M, Pollyea DA, Potluri J, et al.

haematologica | 2022; 107(4)

PEV/VEN/AZA therapy in preclinical AML

Efficacy and biological correlates of response in a phase II study of venetoclax monotherapy in patients with acute myelogenous leukemia. Cancer Discov. 2016; 6(10):1106-1117. 20. 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. 21. Wei AH, Montesinos P, Ivanov V, et al. Venetoclax plus LDAC for patients with untreated AML ineligible for intensive chemotherapy: phase 3 randomized placebo-controlled trial. Blood. 2020; 135(24):2137-2145. 22. Wei AH, Strickland SA, Hou JZ, et al. Venetoclax combined with low-dose cytarabine for previously untreated patients with acute myeloid leukemia: Results from a phase Ib/II study. J Clin Oncol. 2019;37(15):1277-1284. 23. 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. 24. Ramsey HE, Fischer MA, Lee T, et al. A novel MCL1 inhibitor combined with venetoclax rescues venetoclax-resistant acute myelogenous Leukemia. Cancer Discov. 2018;8(12):1566-1581. 25. Chen X, Glytsou C, Zhou H, et al. Targeting mitochondrial structure sensitizes acute myeloid Leukemia to venetoclax treatment. Cancer Discov. 2019; 9(7):890-909. 26. Nechiporuk T, Kurtz SE, Nikolova O, et al. The TP53 apoptotic network is a primary mediator of resistance to BCL2 inhibition in AML cells. Cancer Discov. 2019;9(7):910925. 27. Savona MR, Rathmell JC. Mitochondrial homeostasis in AML and gasping for response in resistance to BCL2 blockade. Cancer Discov. 2019;9(7):831-833. 28. Fischer MA, Friedlander SY, Arrate MP, et al. Venetoclax response is enhanced by selective inhibitor of nuclear export compounds in hematologic malignancies. Blood Adv. 2020;4(3):586-598. 29. Chen CC, Yang CF, Yang MH, et al.

haematologica | 2022; 107(4)

Pretreatment prognostic factors and treatment outcome in elderly patients with de novo acute myeloid leukemia. Ann Oncol. 2005;16(8):1366-1373. 30. Soucy TA, Smith PG, Milhollen MA, et al. An inhibitor of NEDD8-activating enzyme as a new approach to treat cancer. Nature. 2009;458(7239):732-736. 31. Swords RT, Coutre S, Maris MB, et al. Pevonedistat, a first-in-class NEDD8-activating enzyme inhibitor, combined with azacitidine in patients with AML. Blood. 2018;131(13):1415-1424. 32. Nawrocki ST, Griffin P, Kelly KR, Carew JS. MLN4924: a novel first-in-class inhibitor of NEDD8-activating enzyme for cancer therapy. Expert Opin Investig Drugs. 2012;21(10):1563-1573. 33. Khalife J, Radomska HS, Santhanam R, et al. Pharmacological targeting of miR-155 via the NEDD8-activating enzyme inhibitor MLN4924 (Pevonedistat) in FLT3ITD acute myeloid leukemia. Leukemia. 2015;29(10):1981-1992. 34. Jin S, Cojocari D, Purkal JJ, et al. 5Azacitidine induces NOXA to prime AML cells for venetoclax-mediated apoptosis. Clin Cancer Res. 2020;26(13):3371-3383. 35. Knorr KL, Schneider PA, Meng XW, et al. MLN4924 induces Noxa upregulation in acute myelogenous leukemia and synergizes with Bcl-2 inhibitors. Cell Death Differ. 2015;22(12):2133-2142. 36. Yadav B, Pemovska T, Szwajda A, et al. Quantitative scoring of differential drug sensitivity for individually optimized anticancer therapies. Sci Rep. 2014;4:5193. 37. Ianevski A, Giri AK, Aittokallio T. SynergyFinder 2.0: visual analytics of multi-drug combination synergies. Nucleic Acids Res. 2020;48(W1):W488-W493. 38. Yadav B, Wennerberg K, Aittokallio T, Tang J. Searching for drug synergy in complex dose-response landscapes using an interaction potency model. Comput Struct Biotechnol J. 2015;13:504-513. 39. Liu X, Jiang Y, Wu J, et al. NEDD8-activating enzyme inhibitor, MLN4924 (Pevonedistat) induces NOXA-dependent apoptosis through up-regulation of ATF-4. Biochem Biophys Res Commun. 2017;

488(1):1-5. 40. Senichkin V V., Streletskaia AY, Gorbunova AS, Zhivotovsky B, Kopeina GS. Saga of Mcl-1: regulation from transcription to degradation. Cell Death Differ. 2020; 27(2):405-419. 41. Chyla BJ, Harb J, Mantis C, et al. Response to venetoclax in combination with low intensity therapy (LDAC or HMA) in untreated patients with acute myeloid leukemia patients with IDH, FLT3 and other mutations and correlations with BCL2 family expression. Blood. 2019; 134(Suppl 1):S546. 42. Pei S, Pollyea DA, Gustafson A, et al. Monocytic subclones confer resistance to venetoclax-based therapy in patients with acute myeloid leukemia. AACR Journals.org Cancer Discov. 2020;10(4): 101-116. 43. Döhner H, Weisdorf DJ, Bloomfield CD. Acute myeloid leukemia. N Engl J Med. 2015;373(12):1136-1152. 44. Papaemmanuil E, Gerstung M, Bullinger L, et al. Genomic classification and prognosis in acute myeloid leukemia. N Engl J Med. 2016;374(23):2209-2221. 45. DiNardo CD, Pratz KW, Letai A, et al. Safety and preliminary efficacy of venetoclax with decitabine or azacitidine in elderly patients with previously untreated acute myeloid leukaemia: a non-randomised, open-label, phase 1b study. Lancet Oncol. 2018;19(2):216-228. 46. Silverman LR, Demakos EP, Peterson BL, et al. Randomized controlled trial of azacitidine in patients with the myelodysplastic syndrome: A study of the cancer and leukemia group B. J Clin Oncol. 2002; 20(10):2429-2440. 47. Emanuele MJ, Elia AE, Xu Q, et al. Global identification of modular cullin-RING ligase substrates. Cell 2011;147(2):459-474. 48. Smith AJ, Dai H, Correia C, et al. Noxa/Bcl2 protein interactions contribute to bortezomib resistance in human lymphoid cells. J Biol Chem. 2011;286(20):17682-17692. 49. Gomez-Bougie P, Wuilleme-Toumi S, Menoret E, et al. Noxa up-regulation and Mcl-1 cleavage are associated to apoptosis induction by bortezomib in multiple myeloma. Cancer Res. 2007;67(11):5418-5424.

835

ARTICLE Ferrata Storti Foundation

Haematologica 2022 Volume 107(4):836-843

Acute Myeloid Leukemia

Characteristics and outcome of patients with core-binding factor acute myeloid leukemia and FLT3-ITD: results from an international collaborative study Sabine Kayser,1,2,3 Michael Kramer,4 David Martínez-Cuadrón,5,6 Justin Grenet,7 Klaus H. Metzeler,1,8 Zuzana Sustkova,9 Marlise R. Luskin,10 Andrew M. Brunner,11 Michelle A. Elliott,12 Cristina Gil,13 Sandra Casal Marini,14 Zdeněk Ráčil,9,15 Petr Cetkovsky,16 Jan Novak,15 Alexander E. Perl,7 Uwe Platzbecker,1 Friedrich Stölzel,4 Anthony D. Ho,3 Christian Thiede,4 Richard M. Stone,10 Christoph Röllig,4 Pau Montesinos,5,6 Richard F. Schlenk2,3,17# and Mark J. Levis18# Medical Clinic and Policlinic I, Hematology and Cellular Therapy, University Hospital Leipzig, Leipzig, Germany; 2NCT Trial Center, National Center of Tumor Diseases, German Cancer Research Center (DKFZ), Heidelberg, Germany; 3Department of Internal Medicine V, Heidelberg University Hospital, Heidelberg, Germany; 4Department of Medicine I, University Hospital Carl-Gustav-Carus, Dresden, Germany; 5Hematology Department, Hospital Universitari i Politècnic, La Fe, València, Spain; 6CIBERONC, Instituto Carlos III, Madrid, Spain; 7Division of Hematology & Oncology, Abramson Cancer Center, Perelman School of Medicine at the University of Pennsylvania, Philadelphia, PA, USA; 8Laboratory for Leukemia Diagnostics, Department of Medicine III, University Hospital, LMU Munich, Munich, Germany; 9Department of Internal Medicine, Hematology and Oncology, Masaryk University and University Hospital Brno, Brno, Czech Republic; 10Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA, USA; 11Massachusetts General Hospital, Boston, MA, USA; 12Division of Hematology, Department of Internal Medicine, Mayo Clinic, Rochester, MN, USA; 13 Hospital General, Alicante, Spain; 14Department of Clinical Haematology, Centro Hospitalar e Universitário de Coimbra, Coimbra, Portugal; 15Institute of Hematology and Blood Transfusion, Prague, Czech Republic; 16Department of Internal Medicine and Haematology, 3rd Faculty of Medicine, Charles University and Faculty Hospital Kralovske Vinohrady, Prague, Czech Republic; 17Department of Medical Oncology, National Center for Tumor Diseases (NCT), Heidelberg University Hospital, Heidelberg, Germany and 18 Sidney Kimmel Comprehensive Cancer Center, Johns Hopkins University, Baltimore, MD, USA. 1

Presented in part at the 60th Annual Meeting of the American Society of Hematology in Orlando (FL, USA) December 8, 2019.

RFS and MJL contributed equally as co-senior authors.

Correspondence: SABINE KAYSER s.kayser@dkfz-heidelberg.de Received: February 23, 2021. Accepted: June 16, 2021. Pre-published: August 5, 2021. https://doi.org/10.3324/haematol.2021.278645

836

ABSTRACT

he aim of this study was to evaluate the prognostic impact of FLT3-ITD in core-binding factor acute myeloid leukemia (CBFAML) in an international, multicenter survey of 97 patients of whom 52% had t(8;21)(q22;q22) and 48% had inv(16)(p13q22)/t(16;16)(p13;q22). The median age of the patients was 53 years (range, 19-81). Complete remission after anthracycline-based induction (n=86) and non-intensive therapy (n=11) was achieved in 97% and 36% of the patients, respectively. The median follow-up was 4.43 years (95% confidence interval [95% CI]: 3.35-7.39 years). The median survival after intensive and non-intensive treatment was not reached and 0.96 years, respectively. Among intensively treated patients, inv(16) with trisomy 22 (n=11) was associated with a favorable 4-year relapse-free survival rate of 80% (95% CI: 59-100%) as compared to 38% (95% CI: 27-54%; P=0.02) in all other patients with CBFAML/FLT3-ITD (n=75). Overall, 24 patients underwent allogeneic hematopoietic cell transplantation (HCT), 12 in first complete remission and 12 after relapse. Allogeneic HCT in first complete remission was not beneficial (P=0.60); however, allogeneic HCT seemed to improve median survival in relapsed patients compared to that of patients treated with chemotherapy (not reached vs. 0.6 years, respectively; P=0.002). Excluding patients with inv(16) with trisomy 22, our data indicate that haematologica | 2022; 107(4)

Outcome of CBF-AML with FLT3-ITD

the outcome of CBF-AML patients with FLT3-ITD may be inferior to that of patients without FLT3-ITD (based on previously published data), suggesting that prognostically CBF-AML patients with FLT3-ITD should not be classified as favorable-risk. FLT3-inhibitors may improve the outcome of these patients.

Introduction Core binding factor acute myeloid leukemia (CBFAML) is defined by the presence of either t(8;21)(q22;q22)/ RUNX1/RUNX1T1 or inv(16) (p13.1q22)/t(16;16) (p13.1;q22)/ MYH11-CBFB and is recognized by the World Health Organization classification as a separate entity within the category “AML with recurrent genetic abnormalities”.1 Both aberrations result in formation of novel chimeric fusions involving genes of the CBF complex, a master regulator of definitive hematopoiesis.2 CBF-AML is associated with a favorable outcome, particularly if treated with repeated cycles of high-dose cytarabine as post-remission therapy.3-6 The 10year overall survival (OS) rate was reported to be 58% in FLT3 internal tandem duplication (FLT3-ITD)-negative patients.7 Thus, CBF-AML is categorized into the favorable-risk group according to the National Comprehensive Cancer Network guidelines,8 as well as the European LeukemiaNet recommendations,9 regardless of the FLT3ITD mutational status. Nevertheless, 30-40% of patients with CBF-AML experience relapse.3,7,10 FLT3-ITD mutations occur in roughly 5-10% of adults with CBF-AML.11 In a murine transplantation model the co-transduction of FLT3-ITD with RUNX1-RUNX1T112 or CBFB-MYH1113 promoted progression to AML, indicating the cooperative nature of FLT3 aberrations. However, the prognostic relevance of such aberrations in CBF-AML is still controversial. In a study of 176 patients with inv(16) AML, those with FLT3 mutations (n=30) had an inferior OS compared to those with wild-type FLT3.14 Of note, in the same analysis trisomy 22 was a favorable prognostic marker irrespective of concomitant FLT3 mutations,14 confirming the findings of previous studies,3,10 in which trisomy 22 was associated with a lower cumulative incidence of relapse (CIR) than that in patients with only inv(16) (estimated long-term CIR rates of 42% and 66%, respectively; P = 0.02)3 and an excellent relapse-free survival (RFS) of 82%.10 The results of two other studies were also in line with these findings.15,16 Currently, it is unclear whether patients with CBF-AML and KIT or FLT3 mutations may benefit from allogeneic hematopoietic stem cell transplantation (HCT) in first complete remission (CR1). A meta-analysis of several prospective trials evaluating the impact of allogeneic HCT for AML in CR1 did not show a benefit of allogeneic HCT on RFS or OS for favorable-risk AML (n=547) as compared to non-allogeneic HCT therapies including post-remission chemotherapy, autologous HCT, or both.17 In line with these findings, Burnett et al. reported that CBF-AML patients who underwent allogeneic HCT in CR1 had no survival benefit compared to patients not receiving HCT.18 Another retrospective analysis of younger (<60 years) AML patients with t(8;21) compared the outcomes of 118 patients who received allogeneic HCT from a matched-related donor with the outcomes of 132 patients treated with cytarabine-based chemotherapy.19 After allogeneic HCT, the risk of relapse was significantly lower (hazard ratio=0.47; P=0.014), but the treathaematologica | 2022; 107(4)

ment-related mortality was significantly higher (hazard ratio=6.76; P<0.001) than after chemotherapy.19 No benefit regarding RFS and OS was found for allogeneic HCT in the entire study cohort. Two other studies compared the results of allogeneic versus autologous HCT in CBFAML.20,21 Gorin et al. reported a significantly higher relapse risk after autologous HCT than after allogeneic HCT in patients with t(8;21) (28% vs. 15%, P=0.03) but not in patients with inv(16) AML.20 Again, treatmentrelated mortality was significantly higher after allogeneic HCT than afer autologous HCT in both t(8;21) patients (24% vs. 6%, P=0.003) and in those with inv(16) (14% vs. 2%, P=0.003), but the type of transplant did not affect leukemia-free survival.20 A Japanese study also showed comparable results for OS after allogeneic and autologous HCT in CR1 for both t(8;21) and inv(16) AML.21 Nevertheless, the mutational status of FLT3-ITD was not taken into account in any of these reports.17-21 The prognostic impact of FLT3-ITD in CBF-AML is still a matter of debate. The objectives of our study were to characterize CBF-AML patients with FLT3-ITD within an international, multicenter cohort study and compare outcomes according to treatment strategies, with a specific focus on the impact of allogeneic HCT, as compared to conventional chemotherapy, on survival.

Methods Patients and treatment Information on 97 adult patients with CBF-AML diagnosed between 1996 and 2019 (prior to 2000, n=7; 2000-2010, n=39; after 2010, n=51) was collected from eight study groups/institutions in the USA and Europe. Participating centers were chosen upon network relationships of the first and last authors. Detailed case report forms (including information on baseline characteristics, chemotherapy, allogeneic HCT, response, and survival) were collected from all participating centers. Inclusion criteria were adult CBF-AML patients with FLT3-ITD and all patients who fulfilled these criteria were included by the participating groups/institutions. The diagnosis of AML was based on FrenchAmerican-British Cooperative Group criteria,22 and, after 2003, on revised International Working Group criteria.23 Chromosome banding was performed using standard techniques, and karyotypes were described according to the International System for Human Cytogenetic Nomenclature.24 FLT3 mutation screening for ITD and point mutations within the tyrosine kinase domain (TKD) was carried out at each institution as previously described.25,26 Data collection and analysis were approved by the institutional review boards of the participating centers.

Treatment Eighty-six (89%) of the 97 patients received intensive induction treatment either within clinical trials (n=30) or according to local institutional standards (n=56). Treatment protocols included the Study Alliance Leukemia AML60+ (n=2),27 AML96 (n=18)28 and AML2003 (n=9)29 as well as the CALGB/Ratify trial (n=1).30 Induction therapy of the 86 patients consisted of the anthracycline/cytarabine based “7+3” regimen (n=62) or compa-

837

S. Kayser et al.

rable intensive treatment (n=24); additionally, five of the intensively treated patients received midostaurin and four patients received gemtuzumab ozogamicin. Eleven (11%) of the 97 patients were treated non-intensively. Of these, five received azacitidine, either alone (n=2) or in combination with venetoclax (n=2) or sorafenib (n=1); four patients received fludarabine and low-dose cytarabine, one patient was treated with tipifarnib and etoposide within a clinical trial and one patient was treated with hydroxyurea only. Response was assessed according to International Working Group recommendations.23 All clinical studies were approved by the institutional review boards of the participating centers. All patients provided written informed consent to participation in one of the treatment trials or to therapy according to local standards.

Statistical analyses Survival endpoints including OS, RFS, CIR and cumulative incidence of death in CR (CID) were defined according to the revised recommendations of the International Working Group.23 Patients’ characteristics were compared with the Kruskal-Wallis rank sum test for continuous variables and Fisher exact test for categorical variables. The median follow-up time was computed using the reverse Kaplan-Meier estimate.31 The Kaplan-Meier method was used to estimate the distribution of RFS and OS.32 Confidence interval (CI) estimates for survival curves were based on the cumulative hazard function using the Greenwood formula for variance estimation. Log-rank tests were employed to compare survival curves between groups. A Cox proportional hazards regression model was used to identify prognostic variables for RFS.33 CIR and CID and their standard errors were computed according to the method described by Gray34 and included only patients attaining CR. The effect of allogeneic HCT on OS as a time-dependent intervening event was tested using the Mantel-Byar method.35 The method of Simon and Makuch was applied to estimate survival distributions with respect to timedependent interventions.36 The individuals at risk were initially all represented in the chemotherapy group. If patients received an allogeneic HCT, they were censored at this time point in the chemotherapy group and further followed up within the allogeneic HCT group. All statistical analyses were performed with the statistical software environment R, version 3.3.1, using the R packages prodlim, version 1.5.7, and survival, version 2.39-5.37

Results Study cohort Demographic and clinical data were collected from 97 patients (Study Alliance Leukemia, n=46; Spanish Programa Español de Tratamientos en Hematología [PETHEMA], n=20; Johns Hopkins University, Baltimore, n=8; Perelman School of Medicine at the University of Pennsylvania, n=6; University of Munich, n=6; Czech Leukemia Centers, n=6; Dana-Faber Cancer Institute and Massachusetts General Hospital, Boston, n=3; and Mayo Clinic Rochester, n=2) diagnosed with CBF-AML between 1996 and 2019. The median age of the patients was 53 years (range, 19-81 years) and 45 patients (46%) were female. The patients’ baseline characteristics are summarized in Table 1. Median white blood cell (WBC): count was higher in patients with inv(16)/t(16;16) than in patients with t(8;21). In addition, median WBC count in patients with inv(16)/t(16;16) was lower in those with tri838

somy 22 (n=11; median WBC 28.8x109/L; range, 3.9186.7x109/L) than in those without trisomy 22 (n=36; median WBC 54.8x109/L; range, 2.7-298x109/L; P=0.18).

Cytogenetic and molecular analyses The balanced translocation t(8;21)(q22;q22) was present in 50 (52%) of the 97 patients. It occurred as a sole abnormality in 15 (30%) patients, while additional cytogenetic abnormalities were present in 35 (70%) patients, most frequently loss of a sex chromosome (n=26; loss of X or Y n=13, each), three or more abnormalities (n=10) and deletion of the long arm of chromosome 9 (del(9q), n=6). Of the del(9q) cases, all but one co-occurred within a karyotype with three or more abnormalities. An inv(16)(p13q22) (n=46) or t(16;16)(p13;q22) (n=1) was detected in 47 (48%) patients. It was the sole abnormality in 25 (53%) patients, while concurrent cytogenetic abnormalities were present in the other 22 (47%) patients, most frequently trisomy 22 (n=11), three or more abnormalities (n=7), trisomy 8 (n=8; all except one within a karyotype with ≥3 abnormalities) as well as monosomy 7 or deletion of the long arm of chromosome 7 (del(7q); n=4; all within a karyotype with ≥3 abnormalities).

Table 1. Baseline characteristics of patients with acute myeloid leukemia and core-binding factor leukemia.

All patients (n=97) Median age, years 53 (range) (19-81) Female, n. (%) 45 (46) Median WBC, 109/L 20.5 (range) (1.8-298) Missing 2 Median Hb, g/dL 8.6 (range) (4.6-14.3) Missing 5 Median platelets, 109/L 33 (range) (7-372) Missing 5 Median BM blasts, % 60 (range) (0-98) Missing 8 Cytogenetics, n. (%) CBF as sole abn 41 (42) CBF + other abn 56 (58) Trisomy 22 12 (12) Trisomy 8 7 (7) Disease type, n. (%) De novo AML 87 (90) Secondary AML 2 (2) Therapy-related AML 8 (8) Median FLT3-ITD allelic ratio 0.35 (range) 0.003-50 Missing 20 FLT3-TKD N. (%) 10 (21) Missing 49

Inv(16) (n=47)

t(8;21) (n=50)

P-value

50 (19-81) 28 (56) 43.4 (2.7-298) 8.7 (5.6-14.3) 2 33 (7-261) 2 61 (0-98) 2

53.5 (22-77) 17(36) 14.3 (1.8-153.4) 2 8.4 (4.6-12.0) 3 33 (7-373) 3 58 (17-96) 6

0.95

26 (55) 21 (45) 11 (23) 5 (11)

15 (30) 35 (70) 1 (2) 2 (4)

42 (89) 1 (2) 4 (9)

45 (90) 1 (2) 4 (8)

0.99

0.32 0.003-50 9

0.35 0.005-34 11

0.99

8 (29) 19

2 (10) 30

0.07 <0.001

0.1

0.92

0.49

0.01 0.002 0.26

0.16

abn: aberration; allo: allogeneic; AML: acute myeloid leukemia; BM: bone marrow; CBF: core-binding factor; FLT3: fms-related tyrosine kinase 3; Hb: hemoglobin; TKD: tyrosine kinase domain; WBC: white blood cell count. Results may not add-up to 100 due to rounding.

haematologica | 2022; 107(4)

Outcome of CBF-AML with FLT3-ITD

The FLT3-ITD allelic ratio was available in 77 (79%) patients and the median allelic ratio was 0.35 (range, 0.003-50). Median WBC count was higher in patients with a high allelic ratio than in those with a low allelic ratio (31.7x109/L vs. 16x109/L, P=0.02). The median FLT3-ITD size and number of ITD clones were available for 29 (30%) patients. The median FLT3ITD size was 39 (range, 3-120) base-pairs and most of the patients harbored one clone (1 clone, n=24; 2 clones, n=4, 3 clones, n=1). Besides the FLT3-ITD, ten (21%) of 48 patients with available data also harbored a FLT3-TKD (Table 1).

Response to induction therapy Data on response to induction therapy were available for all 97 patients. Of the intensively treated patients (n=86), CR after induction therapy was achieved in 84 (98%), including one patient who achieved CR after salvage therapy with 1 g cytarabine every 12 h on 4 days as well as mitoxantrone 12 mg/day on 3 days. Early death occurred in two (2%) patients; none of the intensively treated patients had refractory disease. All patients, who received “7+3” treatment, either with midostaurin (n=5) or gemtuzumab ozogamicin (n=4), achieved CR. Eleven patients were treated less intensively because of their older age (median: 72 years; range, 40-81 years) or comorbidities. Of these 11 less intensively treated patients, four (36%) achieved CR, one had a partial remission, four were refractory and two patients died early. The four patients who achieved CR had been treated with a hypomethylating agent (n=1), venetoclax + azacitidine (n=1) and fludarabine + low-dose cytarabine (n=2).

Further therapy including intensive consolidation and allogeneic hematopoietic cell transplantation Seventy-two (86%) of 84 intensively treated patients in CR1 received intensive chemotherapy consolidation consisting of high-dose cytarabine with or without additional chemotherapy (mitoxantrone and/or amsacrine, n=25). Precise information on applied consolidation cycles was available for 54 patients. Of those, ten patients received four cycles of consolidation, 14 patients received three cycles, seven received two cycles and 23 patients received one cycle. For analysis, we compared patients who received one or two consolidation cycles with those who received more than two cycles of consolidation. There was no difference in CIR between patients who received one or two cycles and those who received more than two cycles (P=0.97). One of the transplanted patients received maintenance with gilteritinib for 2 years after allogeneic HCT within a randomized trial. In addition, one patient with intensive chemotherapy consolidation received maintenance with midostaurin. Twelve (14%) patients proceeded to allogeneic HCT in CR1 with five of the transplanted patients receiving some consolidation chemotherapy prior to their transplant. There was no difference in baseline characteristics, such as median WBC count, median age and median FLT3-ITD allelic ratio, between patients proceeding to allogeneic HCT in CR1 and those given consolidation chemotherapy, (data not shown). Among the patients consolidated with chemotherapy, relapses occurred in 31 and there were six treatmentrelated deaths after consolidation. In patients consolidated with allogeneic HCT in CR1 three patients relapsed haematologica | 2022; 107(4)

and there was one treatment-related death. In those relapsing after chemotherapy, allogeneic HCT was performed in 12 patients: eight in second CR (CR2) and four with active disease. FLT3 mutational status was available in 17 (50%) of 34 relapsed patients who received intensive treatment. Of these, eight (47%) were still FLT3-ITD-positive. Interestingly, one of the ITD-positive patients developed a new FLT3-TKD mutation.

Characteristics of patients undergoing allogeneic hematopoietic cell transplantation Overall, an allogeneic HCT was performed in 24 (25%) of the 97 patients, either in CR1 (n=12; inv(16), n=4; t(8;21), n=8) or CR2 (n=8; inv(16), n=5; t(8;21), n=3), or with active disease (n=4; inv(16), n=2; t(8;21), n=2). Thirteen patients received myeloablative conditioning, including total body irradiation in eight patients; additionally eight patients received reduced-intensity conditioning (missing, n=3). The stem cell source was a matched related donor in 11 cases, a matched unrelated donor in ten cases, a haplo-identical donor in two, and unknown in one of the 24 patients.

Cumulative incidences of relapse and death in complete remission, and survival The median follow-up of the entire cohort was 4.43 years (95% CI: 3.35-7.39 years). The median and 4-year OS of the entire cohort were 4.48 years (95% CI: 2.48-not reached) and 51% (95% CI: 41-64%), respectively. In intensively treated patients RFS and OS were not different between patients with inv(16) and those with t(8;21) (P=0.70 and P=0.80, respectively) (Figure 1). Furthermore, CIR (P=0.26) (Figure 2A) and CID (P=0.96) (Figure 2B) were comparable in patients proceeding to allogeneic HCT in CR1 or not. However, in relapsed patients survival was dismal without allogeneic HCT (n=22) irrespective of CBF-AML type, with a median survival of 0.6 years after relapse (95% CI: 0.31-1.11 years), and none of the patients survived beyond 2 years. In contrast, in relapsed patients proceeding to allogeneic HCT either in CR2 or with active disease the median survival was not reached and survival at 4 years was 53% (95% CI: 30-94% Figure 3). In a Mantel-Byar analysis including allogeneic HCT performed after relapse as a time-dependent event, survival after relapse was significantly improved by allogeneic HCT (P=0.002). Since supportive care might have had an influence on outcome, we performed a Cox regression analysis. This analysis revealed no impact of date of diagnosis either as a continuous variable (P=0.92) or as a dichotomized variable (on the year 2010; P=0.23). The median survival of non-intensively treated patients was 0.96 years (95% CI: 0.24-not reached) and none of these patients survived beyond 3 years. Exploratory subset analysis revealed trisomy 22 in patients with inv(16) as a significant prognostic factor for RFS (n=11; P=0.02) (Figure 4A); the outcome of these patients was favorable with a 4-year RFS rate of 80% (95% CI: 59-100%), whereas all other CBF patients had a high relapse rate resulting in a 4-year RFS rate of 38% (95% CI: 27-54%, P=0.02) (Figure 4A). In addition, OS tended to be higher in patients with inv(16) and trisomy 22 (P=0.10) than in those with all other CBF-AML (Figure 4B). Other relevant prognostic factors, such as type of CBF839

S. Kayser et al.

AML, older age (≥60 years), WBC count, platelet count, trisomy 8, complex karyotype, and high FLT3-ITD allelic ratio (≥0.5) were not identified as significant variables either for RFS or OS (Table 2). In addition, loss of the Y chromosome in patients with t(8;21) had no impact on outcome (RFS, P=0.7; OS, P=0.3). Trisomy 22 was the only variable with a significant effect on the RFS endpoint (hazard ratio=0.22; P=0.04) (Table 2).

The focus of our study was to characterize adult CBFAML patients with FLT3-ITD in an international, multicenter cohort study and compare outcomes according to treat-

ment strategies, with a specific focus on the impact of allogeneic HCT, as compared to conventional chemotherapy, on survival. Secondary chromosome aberrations can be detected in more than 60% of AML cases with t(8;21) and in 35% to 40% of those with inv(16). In line with published data,10,38 the most frequent secondary chromosome aberration in our cohort of t(8;21) AML patients was loss of a sex chromosome, whereas the most frequent secondary chromosome aberration in inv(16) AML was trisomy 22. In addition, we found a higher WBC count in patients with inv(16) than in those with t(8;21). In contrast to previous reports there was no impact of WBC count, older age (≥60 years) or loss of the sex chromosome10,19,39 on outcome.

Discussion

Figure 1. Kaplan-Meier plots of survival in intensively treated patients according to type of core-binding factor acute myeloid leukemia. (A) Relapse-free survival. (B) Overall survival.

Figure 2. Plots of cumulative incidence of relapse and cumulative incidence of death according to treatment strategy in first complete remission. (A) Cumulative incidence of relapse. (B) Cumulative incidence of death. Only patients attaining complete remission are included. Treatment strategy is divided into consolidation chemotherapy or allogeneic hematopoietic stem cell transplantation (allo-HCT) in first complete remission (CR1).

840

haematologica | 2022; 107(4)

Outcome of CBF-AML with FLT3-ITD

In our cohort, remission rate after intensive treatment was very high, as was reported in CBF-AML without FLT3ITD,3,10 suggesting that CBF-AML is highly chemosensitive regardless of a concurrent FLT3-ITD. We confirmed the excellent prognosis of patients with inv(16) and trisomy 22,3,10,14 despite the additional presence of a FLT3-ITD. To date, it is unclear why patients with an inv(16) and trisomy 22 so rarely relapse after intensive induction and consolidation therapy. Obviously, leukemic clones harboring both abnormalities are very chemosensitive. Our study adds to previous knowledge that despite the proliferative signal induced by a FLT3-ITD40 and the chemoresistance induced by high FLT3-ITD allelic ratios41 patients with inv(16) and trisomy 22 remain extremely chemosensitive. The underlying pathogenetic mechanism by which trisomy 22 exerts its prognostic impact remains elusive. Regarding the outcome of intensively treated CBF patients exhibiting a FLT3-ITD without trisomy 22, results were dismal with a RFS rate of 38% after 4 years. The relapse rate in these patients was high and confirmed findings from previous studies.3,10,14 In comparison, an OS rate of 58% after 10 years was reported in FLT3-ITD-negative patients.7 In our cohort OS was 51% after 4 years in FLT3-

ITD-positive patients as compared to 58% after 10 years in those with wild-type FLT3.7 In addition, the outcome of intensively treated patients was not affected by CBF subtype, inv(16) and t(8;21), or CR1 consolidation approach (chemotherapy or transplantation). These results might argue for the benefit of repeated cycles of intensive chemotherapy as post-remission treatment, i.e., high-dose cytarabine in this subgroup of patients, although we would like to emphasize that this needs to be validated in a larger cohort. Biologically, a FLT3-ITD in CBF-AML seems to impair the favorable prognosis, comparable to its negative impact in acute promyelocytic leukemia, at least in those patients not treated with all-trans retinoic acid and arsenic trioxide.42,43 Despite the limitation that data on measurable residual disease were not available in our cohort, outcome was inferior compared to published data in FLT3-ITD-negative patients.7 Thus, the FLT3 mutational status should be taken into account when classifying CBF-AML; patients with FLT3-ITD should not be classified within the favor-

Figure 3. Simon Makuch plot of overall survival measured from the date of relapse in relapsed patients illustrating the impact of allogeneic hematopoietic stem cell transplantation as a time-dependent event. Allo-HCT: allogeneic hematopoietic stem cell transplantation.

Table 2. Univariable Cox models on relapse-free and overall survival.

HR Type of CBF-AML Older age (≥60 years) Log (WBC) Platelets Complex karyotype High FLT3-ITD allelic ratio (≥0.5) Trisomy 8 Trisomy 22 10

0.89 0.61 0.42 1.00 0.68 1.19 0.94 0.22

RFS P-value 0.72 0.14 0.22 0.54 0.32 0.63 0.92 0.04

HR 0.91 0.63 1.43 1.00 0.96 1.83 1.53 0.35

OS P-value 0.80 0.20 0.34 0.50 0.91 0.13 0.48 0.15

AML: acute myeloid leukemia; CBF: core-binding factor; FLT3: fms-related tyrosine kinase 3; HR: hazard ratio; ITD: internal tandem duplication; OS: overall survival; RFS: relapse-free survival, WBC: white blood cell count..

haematologica | 2022; 107(4)

Figure 4. Kaplan-Meier plots of survival of patients with inversion 16 and trisomy 22 as compared to all other core-binding factor acute myeloid leukemia patients. (A) Relapse-free survival. (B) Overall survival. CBF-AML: core-binding factor acute myeloid leukemia.

841

S. Kayser et al.

able-risk category.8,9 Rather, these patients might be candidates for targeted treatment with tyrosine kinase inhibitors as well as intensive chemotherapy.30,44 In addition, there is evidence that gemtuzumab ozogamicin in combination with chemotherapy particularly benefits patients with FLT3-ITD mutations45 as well as patients with CBF-AML.46 However, in our cohort only a few patients were treated with either midostaurin or gemtuzumab ozogamicin, so the effect of these drugs on outcome could not be evaluated. The impact of midostaurin in combination with gemtuzumab ozogamicin on outcome is currently being evaluated within a phase I/II trial (ClinicalTrials.gov Identifier: NCT04385290). The survival of relapsed patients was dismal without allogeneic HCT, irrespective of CBF-AML type, with a median survival of 0.6 years after relapse and none of the patients survived beyond 2 years. In contrast, in patients proceeding to allogeneic HCT after relapse, either in CR2 or with active disease, the median survival had not been reached and survival at 4 years was 53%, arguing that allogeneic HCT should be the preferred approach in relapsed patients.3,10,21 However, we would like to emphasize that retrospectively collected data have serious limitations since the factors for allocating patients to allogeneic HCT, such as co-morbidities, individual assessment of the treating physician, choice of conditioning, and availability of a donor, remain unknown and these factors need to be taken into account when evaluating the value of allogeneic HCT in our series. In conclusion, despite a high remission rate patients with FLT3-ITD had an inferior outcome compared to those without FLT3-ITD, based on previously published data on CBF-AML. Thus, CBF-AML with FLT3-ITD

References 1. Swerdlow SH, Campo E, Harris NL, et al. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues, Revised 4th Edition. WHO Press, Geneva, Switzerland, 2017. 2. Speck NA, Gilliland DG. Core-binding factors in haematopoiesis and leukaemia. Nat Rev Cancer. 2002;2(7):502-513. 3. Marcucci G, Mrózek K, Ruppert AS, et al. Prognostic factors and outcome of core binding factor acute myeloid leukemia patients with t(8;21) differ from those of patients with inv(16): a Cancer and Leukemia Group B study. J Clin Oncol. 2005;23(24):5705-5717. 4. Bloomfield CD, Lawrence D, Byrd JC, et al. Frequency of prolonged remission duration after high-dose cytarabine intensification in acute myeloid leukemia varies by cytogenetic subtype. Cancer Res. 1998;58(18): 4173-4179 5. Byrd JC, Ruppert AS, Mrózek K, et al. Repetitive cycles of high-dose cytarabine benefit patients with acute myeloid leukemia and inv(16)(p13q22) or t(16;16)(p13;q22): results from CALGB 8461. J Clin Oncol. 2004;22(6):1087-1094. 6. Miyawaki S, Ohtake S, Fujisawa S, et al. A randomized comparison of 4 courses of standard-dose multiagent chemotherapy versus 3 courses of high-dose cytarabine alone in postremission therapy for acute myeloid leukemia in adults: the JALSG AML201 study. Blood. 2011;117(8):2366-2372.

842

should not be classified within the favorable-risk category. Our data suggest that allogeneic HCT should be the preferred approach in relapsed patients. CBF-AML with FLT3-ITD represents a further therapeutic target for tyrosine kinase inhibitors as well as gemtuzumab ozogamicin and should be included in combined FLT3inhibitor/CD33-antibody trials (ClinicalTrials.gov Identifier: NCT04385290). Disclosures No conflicts of interest to disclose. Contributions SK and RFS were responsible for the concept of this paper, contributed to the literature search data collection, analyzed and interpreted data, and wrote the manuscript. MJL was responsible for the concept of this paper, contributed to the literature search data collection, contributed patients, analyzed and interpreted data, and critically revised the manuscript. CTh performed research and critically revised the manuscript. MK, DM-C, JG, KHM, ZS, MRL, AMB, MAE, CG, SCM, ZR, PC, JN, AEP, FS, ADH, UP, RMS, CR, and PM contributed patients and critically revised the manuscript. All authors reviewed and approved the final manuscript. Funding SK was supported by the Olympia-Morata fellowship program from the Medical Faculty of the Heidelberg University. MJL is supported by a grant from the NCI (NCI Leukemia SPORE P50 CA100632). ZS, ZR, PC and JN were supported by the Ministry of the Czech Republic, grant n. 15-25809A. The authors also acknowledge support from Leipzig University for Open Access Publishing.

7. Allen C, Hills RK, Lamb K, et al. The importance of relative mutant level for evaluating impact on outcome of KIT, FLT3 and CBL mutations in core-binding factor acute myeloid leukemia. Leukemia. 2013;27(9): 1891-1901. 8. O'Donnell MR, Tallman MS, Abboud CN, et al. Acute myeloid leukemia, version 3.2017, NCCN clinical practice guidelines in oncology. J Natl Compr Canc Netw. 2017;15(7):926-957. 9. 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. 10. Schlenk RF, Benner A, Krauter J, et al. Individual patient data-based meta-analysis of patients aged 16 to 60 years with core binding factor acute myeloid leukemia: a survey of the German Acute Myeloid Leukemia Intergroup. J Clin Oncol. 2004;22(18):3741-3750. 11. Faber ZJ, Chen X, Gedman AL, et al. The genomic landscape of core-binding factor acute myeloid leukemias. Nat Genet. 2016;48(12):1551-1556. 12. Schessl C, Rawat VP, Cusan M, et al. The AML1-ETO fusion gene and the FLT3 length mutation collaborate in inducing acute leukemia in mice. J Clin Invest. 2005;115(8): 2159-2168. 13. Kim HG, Kojima K, Swindle CS, et al. FLT3ITD cooperates with inv(16) to promote progression to acute myeloid leukemia. Blood. 2008;111(3):1567-1574. 14. Paschka P, Du J, Schlenk RF, et al. Secondary

genetic lesions in acute myeloid leukemia with inv(16) or t(16;16): a study of the German-Austrian AML Study Group (AMLSG). Blood. 2013;121(1):170-177. 15. Boissel N, Leroy H, Brethon B, et al. Incidence and prognostic impact of c-Kit, FLT3, and Ras gene mutations in core binding factor acute myeloid leukemia (CBFAML). Leukemia. 2006;20(6):965-970. 16. Jones D, Yao H, Romans A, et al. Modeling interactions between leukemia-specific chromosomal changes, somatic mutations, and gene expression patterns during progression of core-binding factor leukemias. Genes Chromosomes Cancer. 2010;49(2): 182-191. 17. Koreth J, Schlenk R, Kopecky KJ, et al. Allogeneic stem cell transplantation for acute myeloid leukemia in first complete remission: systematic review and metaanalysis of prospective clinical trials. JAMA. 2009;301(22):2349-2361. 18. 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. 19. Schlenk RF, Pasquini MC, Perez WS, et al. HLA-identical sibling allogeneic transplants versus chemotherapy in acute myelogenous leukemia with t(8;21) in first complete remission: collaborative study between the German AML Intergroup and CIBMTR. Biol Blood Marrow Transplant. 2008;14(2):187196. 20. Gorin NC, Labopin M, Frassoni F, et al. Identical outcome after autologous or allo-

haematologica | 2022; 107(4)

Outcome of CBF-AML with FLT3-ITD

geneic genoidentical hematopoietic stemcell transplantation in first remission of acute myelocytic leukemia carrying inversion 16 or t(8;21): a retrospective study from the European Cooperative Group for Blood and Marrow Transplantation. J Clin Oncol. 2008;26(19):3183-3188. 21. Kuwatsuka Y, Miyamura K, Suzuki R, et al. Hematopoietic stem cell transplantation for core binding factor acute myeloid leukemia: t(8;21) and inv(16) represent different clinical outcomes. Blood. 2009;113(9):2096-2103. 22. Bennett JM, Catovsky D, Daniel MT, et al. Proposed revised criteria for the classification of acute myeloid leukemia. A report of the French-American-British Cooperative Group. Ann Intern Med. 1985;103(4):620625. 23. 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. 24. Mitelman F: ISCN: an International System for Human Cytogenetic Nomenclature. Basel, Switzerland: S. Karger. 1995. 25. Yokota S, Kiyoi H, Nakao M, et al. Internal tandem duplication of the FLT3 gene is preferentially seen in acute myeloid leukemia and myelodysplastic syndrome among various hematological malignancies. A study on a large series of patients and cell lines. Leukemia. 1997;11(10):1605-1609. 26. Thiede C, Steudel C, Mohr B, et al. Analysis of FLT3-activating mutations in 979 patients with acute myelogenous leukemia: association with FAB subtypes and identification of subgroups with poor prognosis. Blood. 2002;99(12):4326-4335. 27. Röllig C, Kramer M, Gabrecht M, et al. Intermediate-dose cytarabine plus mitoxantrone versus standard-dose cytarabine plus daunorubicin for acute myeloid

haematologica | 2022; 107(4)

leukemia in elderly patients. Ann Oncol. 2018;29(4):973-978. 28. Röllig C, Thiede C, Gramatzki M, et al. A novel prognostic model in elderly patients with acute myeloid leukemia: results of 909 patients entered into the prospective AML96 trial. Blood. 2010;116(6):971-997 29. Schaich M, Parmentier S, Kramer M, et al. High-dose cytarabine consolidation with or without additional amsacrine and mitoxantrone in acute myeloid leukemia: results of the prospective randomized AML2003 trial. J Clin Oncol. 2013;31:2094-2102. 30. Stone RM, Mandrekar SJ, Sanford BL, et al. Midostaurin plus chemotherapy for acute myeloid leukemia with a FLT3 mutation. N Engl J Med. 2017;377(5):454-464. 31. Schemper M, Smith TL. A note on quantifying follow-up in studies of failure time. Control Clin Trials. 1996;17(4):343-346. 32. Kaplan E, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53(282):457-481. 33. Cox DR. Regression models and life tables (with discussion). J R Stat Soc. 1972;34(2): 187-220. 34. Gray RJ. A class of k-sample tests for comparing the cumulative incidence of a competing risk. Ann Stat. 1988;16(3):1141-1154. 35. Mantel N, Byar D. Evaluation of responsetime data involving transient states: an illustration using heart transplant data. J Am Stat Assoc. 1974;69(345):81-86. 36. Simon R, Makuch RW. A non-parametric graphical representation of the relationship between survival and the occurrence of an event: application to responder versus nonresponder bias. Stat Med. 1984;3(1):35-44. 37. R Development Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing. Vienna, Austria, 2014. 38. Paschka P. Core binding factor acute myeloid leukemia. Semin Oncol. 2008;35(4): 410-417. 39. Opatz S, Bamopoulos SA, Metzeler KH, et al.

The clinical mutatome of core binding factor leukemia. Leukemia. 2020;34(6):1553-1562. 40. Brandts CH, Sargin B, Rode M, et al. Constitutive activation of Akt by Flt3 internal tandem duplications is necessary for increased survival, proliferation, and myeloid transformation. Cancer Res. 2005;65:9643-9650. 41. Schlenk RF, Kayser S, Bullinger L, et al. Differential impact of allelic ratio and insertion site in FLT3-ITD-positive AML with respect to allogeneic transplantation. Blood. 2014;124(23):3441-3449. 42. Lucena-Araujo AR, Kim HT, Jacomo RH, et al. Internal tandem duplication of the FLT3 gene confers poor overall survival in patients with acute promyelocytic leukemia treated with all-trans retinoic acid and anthracycline-based chemotherapy: an International Consortium on Acute Promyelocytic Leukemia study. Ann Hematol. 2014;93(12): 2001-2010. 43. Cicconi L, Divona M, Ciardi C, et al. PMLRARa kinetics and impact of FLT3-ITD mutations in newly diagnosed acute promyelocytic leukaemia treated with ATRA and ATO or ATRA and chemotherapy. Leukemia 2016;30(10):1987-1992. 44. Paschka P, Schlenk RF, Weber D, et al. Adding dasatinib to intensive treatment in core-binding factor acute myeloid leukemiaresults of the AMLSG 11-08 trial. Leukemia. 2018;32(7):1621-1630. 45. Lambert J, Pautas C, Terré C, et al. Gemtuzumab ozogamicin for de novo acute myeloid leukemia: final efficacy and safety updates from the open-label, phase III ALFA-0701 trial. Haematologica. 2019;104 (1):113-119. 46. Hills RK, Castaigne S, Appelbaum FR, et al. Addition of gemtuzumab ozogamicin to induction chemotherapy in adult patients with acute myeloid leukaemia: a metaanalysis of individual patient data from randomised controlled trials. Lancet Oncol. 2014;15(9):986-996.

843

ARTICLE Ferrata Storti Foundation

Bone Marrow Transplantation

Refined HLA-DPB1 mismatch with molecular algorithms predicts outcomes in hematopoietic stem cell transplantation Jun Zou,1* Piyanuch Kongtim,2* Betül Oran,3 Vasilis Kosmoliaptsis,4 Yudith Carmazzi,1 Junsheng Ma,5 Liang Li,5 Gabriela Rondon,3 Samer Srour,3 Hannah C. Copley,4 David Partlow,1 Stefan O. Ciurea,2 Uri Greenbaum,3 Qing Ma,3 Elizabeth J. Shpall,3 Richard E. Champlin3# and Kai Cao1#

Haematologica 2022 Volume 107(4):844-856

Department of Laboratory Medicine, The University of Texas MD Anderson Cancer Center, Houston, TX, USA; 2Division of Hematology/Oncology, Department of Medicine, Chao Family Comprehensive Cancer Center, University of California, Irvine, CA, USA; 3 Department of Stem Cell Transplantation and Cellular Therapy, The University of Texas MD Anderson Cancer Center, Houston, TX, USA; 4Department of Surgery, University of Cambridge, and NIHR Cambridge Biomedical Research Centre, Cambridge, UK and 5 Department of Biostatistics, The University of Texas MD Anderson Cancer Center, Houston, TX, USA 1

*JZ and PK contributed equally as co-first authors. REC and KC contributed equally as co-senior authors.

ABSTRACT

Correspondence: JUN ZOU jzou@mdanderson.org Received: April 14, 2021. Accepted: July 9, 2021. Pre-published: August 26, 2021. https://doi.org/10.3324/haematol.2021.278993

844

LA-DPB1 mismatches between donor and recipient are commonly seen in allogeneic hematopoietic stem cell transplantation from an unrelated donor. HLA-DPB1 mismatch, conventionally determined by the similarity of the T-cell epitope (TCE), is associated with an increased risk of acute graft-versus-host disease (GVHD) and a decreased risk of disease relapse. We investigated the clinical impact of HLA-DPB1 molecular mismatch quantified by mismatched eplets (ME) and the Predicted Indirectly Recognizable HLA Epitopes Score (PS) in a cohort of 1,514 patients receiving hematopoietic stem cell transplants from unrelated donors matched at HLA-A, -B, -C, -DRB1/3/4/5, and DQB1 loci. HLA-DPB1 alloimmunity in the graft-versus-host direction, determined by high graft-versus-host ME/PS, was associated with a reduced risk of relapse (hazard ratio [HR]=0.83, P=0.05 for ME) and increased risk of grade 2-4 acute GVHD (HR=1.44, P<0.001 for ME), whereas high host-versus-graft ME/PS was only associated with an increased risk of grade 2-4 acute GVHD (HR=1.26, P=0.004 for ME). Notably, in the permissive mismatch subgroup classified by TCE grouping, high host-versus-graft ME/PS was associated with an increased risk of relapse (HR=1.36, P=0.026 for ME) and grade 2-4 acute GVHD (HR=1.43, P=0.003 for PS-II). Decision curve analysis showed that graftversus-host ME outperformed other models and provided the best clinical net benefit for the modification of acute GVHD prophylaxis regimens in patients with a high risk of developing clinically significant acute GVHD. In conclusion, molecular assessment of HLA-DPB1 mismatch enables separate prediction of host-versus-graft or graft-versus-host alloresponse quantitatively and allows further refinement of HLA-DPB1 permissiveness as defined by conventional TCE grouping.

Introduction Currently, allogeneic hematopoietic stem cell transplantation (HSCT) is the only curative therapy for many hematologic malignancies. Although modern immunosuppressive therapy and transplant interventions have significantly improved non-relapse mortality (NRM) over years,1 as a major complication after HSCT, acute graft-versus-host disease (GVHD) occurs in 20 to 80% of recipients with 15% mortality.2 It is well established that patients who undergo allogeneic HSCT

haematologica | 2022; 107(4)

Refined HLA-DPB1 mismatch with molecular algorithms

from an HLA-mismatched unrelated donor are more likely to have a higher incidence of acute GVHD and suboptimal clinical outcomes.3-5 Among patients who have received HLA-A, -B, -C, -DRB1, and -DQB1 matched (10/10) grafts from unrelated donors, the disparity between the donor and recipient at the HLA-DPB1 locus is associated with an increased risk of GVHD but is counterbalanced by a reduced risk of relapse.6,7 Given the weak linkage disequilibrium between the DP locus and DR/DQ loci, mismatching at the HLA-DPB1 locus is observed in about 75-90% of transplants from unrelated donors regardless of matching at other HLA loci.7-10 Pioneering studies have classified HLA-DPB1 mismatches as permissive or nonpermissive using the functional toxicity assay and by analyzing the similarity of Tcell epitopes (TCE).11,12 The initial experimental hypothesis has been confirmed clinically and translated into a donor selection algorithm; permissive HLA-DPB1 mismatches are associated with elicited alloreactivity resulting in a beneficial graft-versus-leukemia (GVL) effect with clinically tolerable GVHD.13,14 This approach has significantly expanded the likelihood of finding suitable unrelated donors and reduced the risks of mortality by avoiding donors with nonpermissive mismatches.7,15,16 Although the TCE model assigns permissiveness based on T-cell alloreactivity within the same or from different immunogenicity groups,11 another partially overlapping model predicts HLA-DPB1 immunogenicity with similar success by analyzing expression levels of the specific HLA-DPB1 allele.17,18 Modern HLA molecular matching methods may open new avenues for alloimmune risk assessment and help to quantitatively refine the traditional TCE grouping. Additionally, the different direction of HLA-DPB1 nonpermissive mismatches defined by the TCE model, i.e., either in the host-versus-graft (HVG) or graft-versus-host (GVH) direction, appears to have a similar impact on the risk of GVHD and mortality in HSCT from unrelated donors.7,8,16 Although the underlying mechanism of nonpermissive mismatch in the HVG direction remains unclear, recent compelling evidence showed that peripheral host T cells present in the skin and gut are primed by donor-derived antigen-presenting cells and contribute to the development of GVHD.19-21 Computational prediction methods could separately assess immunogenicity from a donor’s or recipient’s perspective in a quantitative manner, which might shed light on the alloreactive mechanisms that mediate GVHD risk and the GVL effect in HSCT from HLA-DPB1 mismatched donors. HLAMatchmaker, one of the best-studied molecular matching strategies, compares eplets, which are the key structural component of epitopes, between the donor and recipient. The amount of mismatched eplets (ME) between donor and recipient has been shown to correlate with the level of immune response and is associated with clinical outcomes in patients who have undergone haploidentical HSCT.22 As HLAMatchmaker focuses mainly on surface-accessible positions, TCE that are derived from polymorphisms on the non-exposed region of HLA molecules could be overlooked.23,24 Alloreactivity in transplantation is critically dependent on T-cell responses via the indirect recognition pathway in which polymorphic HLA-derived peptides are presented to T cells. Although various approaches have been described to predict TCE

haematologica | 2022; 107(4)

through the indirect recognition pathway, Predicted Indirectly Recognizable HLA Epitopes (PIRCHE), with PIRCHE score (PS)-I representing CD8+ T-cell alloreactivity and PS-II representing CD4+ T-cell alloreactivity, is widely and successfully used for this purpose.25 In the present study, we sought to comprehensively validate the molecular mismatch algorithms in predicting the risks associated with HLA-DPB1 mismatches in a relatively large cohort of patients with malignant disease who underwent HSCT from unrelated donors. Furthermore, we hypothesized that in silico quantification could refine the current definition of TCE grouping, especially in the permissive or nonpermissive mismatch subgroups, given that significantly different T-cell cross-reactivities are seen in various HLA-DPB1 alleles within the same subgroup.11

Methods Patients and transplant characteristics Our cohort included consecutively treated patients with hematologic malignancies who were 18 years of age or older and underwent allogeneic HSCT at The University of Texas MD Anderson Cancer Center (UTMDACC) between June 2005 and December 2018. All patients in our analysis received HSCT from an HLA-A, -B, -C, DRB1, -DQB1, -DRB3/4/5 matched unrelated donor to minimize the confounding alloreactivity caused by HLA mismatch from other loci. Clinical and laboratory data were collected from electronic medical records. All patients provided written informed consent for HSCT in accordance with the Declaration of Helsinki. A retrospective data review protocol and a waiver of informed consent were approved by the UTMDACC Institutional Review Board.

HLA typing and ME and PS analyses Patients included in the study had donor and recipient HLA typing performed at the HLA-A, -B, -C, DRB1, -DRB3/4/5, DQB1, and -DPB1 loci using sequence-based typing methods at high resolution.26 ME load at the HLA-DPB1 locus was measured using the HLAMatchmaker module incorporated in HLA Fusion software v4.3, which identifies theoretically predicted eplets based on crystallized HLA molecule models27 and identifies ME by comparing donor and recipient eplets. The analyses were performed separately in both the GVH and HVG directions.22 Eplet repertoires are listed in the HLA Epitope Registry (http://www.epitopes.net/downloads.html). The PS for mismatched HLA-DPB1 in the GVH direction was calculated using the HSCT module from the PIRCHE online matching service (http://www.pirche.com/pirche/#/). The PS for mismatched HLADPB1 in the HVG direction was calculated by inverting the patient and donor in the input fields using the same HSCT module.

HLA-DPB1 permissiveness defined by the TCE model HLA-DPB1 mismatches between the donor and the recipient were classified into permissive and nonpermissive mismatches according to TCE algorithms (version 2.0) on the IPD-IMGT/HLA website (https://www.ebi.ac.uk/ipd/imgt/hla/dpb.html).28 As previously described,26 the direction of HLA-DPB1 mismatch, either in the GVH or HVG direction, was assigned. Transplants were therefore classified into four groups: (i) HLA-DPB1 matched, (ii) permissive mismatched, (iii) nonpermissive mismatched in the HVG direction, and (iv) nonpermissive mismatched in the GVH direction.

845

J. Zou et al.

Statistical analysis The primary outcome was acute GVHD and secondary outcomes were overall survival, progression-free survival, relapse, NRM, and neutrophil engraftment. Univariate and multivariable Cox proportional hazards regression was used to determine the impact of baseline characteristics, PS, ME, and HLA-DPB1 matching on survival outcomes, while univariate and multivariable sub-distributional hazards regression was used to analyze cumulative incidence outcomes, including relapse, NRM, acute GVHD, and engraftment. All regression models were tested for proportional hazards assumption and interaction terms. Each PS, ME, and HLA-DPB1 match group with a P value <0.1 in the univariate analysis was analyzed in separate multivariable regression models adjusted for significant baseline characteristics. PS and ME were analyzed as both continuous variables and categorical variables (low versus high), and they were analyzed only as categorical variables in multivariable analyses. To determine the optimal cutoff for low versus high PS and ME groups, the concordance probabilities of PS and ME for acute GVHD prediction were tested at the 25th, 50th, and 75th percentile cutoffs. The cutoffs at the 50th percentile were selected for the analysis to maximize the concordance probability. The discrimination power of the TCE, ME, and PS models on acute GVHD was compared using the Harrell C-concordance index. A decision-curve analysis29,30 was performed to assess the net clinical benefit of all models in deciding on GVHD regimen modification. Outcome definitions and details of the statistical analysis are described in the Online Supplementary Material.

Results Patients’ characteristics and HLA-DPB1 matching status defined by TCE and in silico methods The analysis included 1,514 patients with a median age of 56 years (range, 18-79). The characteristics of the patients and their transplants are listed in Table 1. The majority of patients received a peripheral blood graft (62%) and GVHD prophylaxis with tacrolimus and mycophenolate (83%). Seventy-four percent of patients received anti-thymocyte globulin as a part of GVHD prophylaxis. The variables that were significantly different between subgroups were bone marrow stem cell source (with 29% in the GVH nonpermissive group versus 37.6% in the whole group) and the year of HSCT. The number of transplants with nonpermissive mismatch was significantly reduced in recent years (2014-2018) compared to the previous years (27.6% versus 36.5%, respectively), likely due to the awareness of the adverse effect of nonpermissive mismatch. HLA-DPB1 permissive mismatch was present in 43.0% of patients, and nonpermissive HLA-DPB1 mismatches in the GVH and HVG directions were noted in 17.7% and 15.1% of patients, respectively. The median follow-up duration in 695 surviving patients was 57.1 months (range, 3.4-148.4). ME, PS-I, and PS-II were quantified in both HVG and GVH directions (Table 1). High concordances between the functional TCE grouping and in silico methods were noted. The median ME, PS-I, and PS-II values in the GVH direction in the GVH nonpermissive mismatch group were significantly higher than the corresponding values in the HVG nonpermissive mismatch group and in the 846

permissive mismatch group. Likewise, the median ME, PS-I, and PS-II values in the HVG direction were considerably higher in the HVG nonpermissive mismatch group than in the GVH nonpermissive mismatch and permissive mismatch groups. No or weakly positive correlations were seen between GVH and HVG ME, PS-I, and PS-II values, indicating that ME/PS from the donor perspective were different from ME/PS from the recipient perspective, whereas positive correlations were observed between PS-I and PS-II values and between ME and PS values in the same direction (GVH or HVG) (Online Supplementary Figure S1). The number of patients in the low and high PS and ME groups and TCE model are summarized in Online Supplementary Tables S1 and S2.

Impact of HLA-DPB1 matching status defined by TCE, ME, and PS on post-transplant outcomes In the entire cohort, molecular mismatches in the GVH direction were associated with a reduced risk of relapse and increased risk of GVHD and NRM, whereas mismatch in the HVG direction was associated only with increased risk of GVHD without relapse protection. Results from multivariable analyses showed that HLADPB1 mismatches by TCE grouping, ME, PS-I, and PS-II in both the GVH and HVG directions were strongly associated with an increased risk of clinically significant acute GVHD after adjustment for significant baseline characteristics (Figure 1A). Using conventional TCE grouping, compared with the HLA-DPB1 matched group, those with permissive mismatch, GVH nonpermissive mismatch, and HVG nonpermissive mismatch had an increased risk of grade 2-4 acute GVHD (permissive: hazard ratio [HR]=1.42, 95% confidence interval [95% CI]: 1.15-1.76, P=0.001; GVH nonpermissive: HR=1.99, 95% CI: 1.55-2.55, P<0.001; HVG nonpermissive: HR=1.80, 95% CI: 1.38-2.35, P<0.001). Using the median cutoff of ME, the risk of grade 2-4 acute GVHD was 1.44 (95% CI: 1.23-1.68, P<0.001) and 1.26 (95% CI: 1.08-1.48, P=0.004) times higher in those with high ME in the GVH and HVG direction, respectively, than in those with low ME in the same direction. Similarly, having a high PS in the GVH direction was associated with an increased risk of grade 2-4 acute GVHD (PS-I: HR=1.39, 95% CI: 1.19-1.63, P<0.001; PS-II: HR=1.40, 95% CI: 1.19-1.64, P<0.001). Having a high PS in the HVG direction was also associated with an increased risk of grade 2-4 acute GVHD (PS-I: HR=1.32, 95% CI: 1.12-1.54, P=0.001; PS-II: HR=1.24, 95% CI: 1.05-1.45, P=0.009). The associations of ME, PS-I, and PS-II in the GVH direction with grade 2-4 acute GVHD risk were independent of the associations of ME, PS-I, and PS-II in the HVG direction with grade 2-4 acute GVHD. However, higher risks of grade 2-4 acute GVHD were seen in patients who had high ME, PS-I, or PS-II in both the GVH and HVG directions than in those with low ME, PS-I, or PS-II in both directions (Figure 1A, Online Supplementary Figure S2A). For NRM, HLA-DPB1 nonpermissive mismatch in either the GVH direction (HR=1.67, 95% CI: 1.24-2.27, P=0.001) or HVG direction (HR=1.46, 95% CI: 1.05-2.03, P=0.025) was associated with a significantly increased risk of NRM compared with that in the matched group, whereas no association was seen between NRM and permissive mismatch status. The strong association of high haematologica | 2022; 107(4)

Refined HLA-DPB1 mismatch with molecular algorithms

Table 1. Clinical characteristics of patients who underwent hematopoietic stem cell transplantation from unrelated donors.

Characteristic

Entire cohort, n=1514

Median age in years (range) 56 (18-79) Age >50 years, n (%) 991 (65.5) Donor age in years (range) 30 (18-71) Donor age >40 years, n (%) 288 (19.0) Female, n (%) 614 (40.6) Donor-recipient sex combination, n (%) Female to female 178 (11.8) Female to male 211 (13.9) Male to female 436 (28.8) Male to male 689 (45.5) ABO matching, n (%) Match 724 (47.8) Minor mismatch 351 (23.2) Major mismatch 333 (22.0) Bidirectional mismatch 106 (7.0) Donor-recipient CMV serostatus (n=1510), n (%) NR-NR 192 (12.7) NR-R 734 (48.6) R-NR 99 (6.6) R-R 485 (32.1) Diagnosis, n (%) AML/MDS 673 (44.5) Other hematologic malignancies 841 (55.5) DRI, n (%) Low 228 (15.1) Intermediate 600 (39.6) High 518 (34.2) Very high 168 (11.1) HCT-CI, median (range) 3 (0-11) HCT-CI ≥3, n (%) 766 (50.6) Prior AlloHSCT, n (%) 36 (2.4) Prior AutoHSCT, n (%) 120 (7.9) HSCT protocol, n (%) Clinical trial protocol 962 (63.5) Standard of care 552 (36.5) Conditioning regimen intensity, n (%) MA 1024 (67.6) RIC/NMA 490 (32.4) Stem cell source, n (%) PB 945 (62.4) BM 569 (37.6) GVHD regimen (n=1513), n (%) Tacrolimus/methotrexate 1268 (83.8) PTCY 185 (12.2) Others 60 (4.0) ATG, n (%) 1121 (74.0) Year of HSCT, n (%) 2005-2009 359 (23.7) 2010-2013 531 (35.1) 2014-2018 624 (41.2) Quantified ME, PS-I, and PS-II, median (range) GVH DP ME 4 (0-22) GVH PS-I 0 (0-14) GVH PS-II 2 (0-28) HVG DP ME 4 (0-20) HVG PS-I 0 (0-17) HVG PS-II 1 (0-34)

Match, n=366

HLA-DPB1 match by TCE grouping Permissive GVH nonpermissive HVG nonpermissive mismatch, mismatch, mismatch, n=651 n=269 n=228

55 (18-76 ) 237 (64.8) 30 (18-63) 59 (16.1) 141 (38.5)

56 (18-76) 437 (67.1) 30 (18-58) 122 (18.8) 259 (39.8)

56 (20-77) 172 (63.9) 30 (18-59) 61 (22.7) 120 (44.6)

57 (20-79) 145 (63.6) 29 (19-71) 46 (20.2) 94 (41.2)

42 (11.5) 48 (13.1) 99 (27.0) 177 (48.4)

74 (11.4) 97 (14.9) 185 (28.4) 295 (45.3)

35 (13.0) 42 (15.6) 85 (31.6) 107 (39.8)

27 (11.8) 24 (10.5) 67 (29.4) 110 (48.3)

177 (48.4) 77 (21.0) 81 (22.1) 31 (8.5)

313 (48.1) 159 (24.4) 147 (22.6) 32 (4.9)

129 (48.0) 56 (20.8) 62 (23.0) 22 (8.2)

105 (46.1) 59 (25.9) 43 (18.9) 21 (9.2)

52 (14.2) 182 (49.9) 23 (6.3) 108 (29.6)

83 (12.8) 303 (46.8) 47 (7.3) 215 (33.2)

33 (12.3) 137 (50.9) 18 (6.7) 81 (30.1)

24 (10.5) 112 (49.1) 11 (4.8) 81 (35.5)

170 (46.5) 196 (53.6)

293 (45.0) 358 (55.0)

107 (39.8) 162 (60.2)

103 (45.2) 125 (54.8)

63 (17.2) 139 (38.0) 130 (35.5) 34 (9.3) 3 (0-11) 187 (51.1) 8 (2.2) 30 (8.2)

89 (13.7) 269 (41.3) 219 (33.6) 74 (11.4) 2 years (0-11) 325 (49.9) 14 (2.2) 52 (8.0)

42 (15.6) 97 (36.1) 96 (35.7) 34 (12.6) 3 years (0-10) 150 (55.8) 8 (3.0) 21 (7.8)

34 (14.9) 95 (41.7) 73 (32.0) 26 (11.4) 2 years (0-11) 104 (45.6) 6 (2.6) 17 (7.5)

234 (63.9) 132 (36.1)

407 (62.5) 244 (37.5)

164 (61.0) 105 (39.0)

157 (68.9) 71 (31.1)

245 (66.9) 121 (33.1)

454 (69.7) 197 (30.3)

183 (68.0) 86 (32.0)

142 (62.3) 86 (37.7)

202 (55.2) 164 (44.8)

411 (63.1) 240 (36.9)

191 (71) 78 (29)

141 (61.8) 87 (38.2)

295 (80.6) 55 15.0 16 (4.4) 266 (72.7)

547 (84.2) 79 (12.2) 24 (3.7) 477 (73.3)

229 (85.1) 29 (10.8) 11 (4.0) 205 (76.2)

197 (86.4) 22 (9.7) 9 (4.0) 173 (75.9)

68 (18.6) 105 (28.7) 193 (52.7)

163 (25.0) 229 (35.2) 259 (39.8)

72 (26.8) 103 (38.3) 94 (34.9)

56 (24.6) 94 (41.2) 78 (34.2)

0 (0-0) 0 (0-0) 0 (0-0) 0 (0-0) 0 (0-0) 0 (0-0)

5 (0-22) 1 (0-13) 3 (0-22) 5 (0-19) 1 (0-14) 3 (0-34)

9 (0-19) 3 (0-9) 8 (0-28) 5 (0-20) 0 (0-17) 1 (0-22)

5 (0-21) 1 (0-14) 2 (0-27) 9 (1-19) 3 (0-10) 8(0-25)

0.972 0.673 0.387 0.207 0.447 0.566

0.259

0.725

0.376

0.710

0.261 0.152 0.878 0.990 0.274

0.223

0.001

0.540

0.657 <0.001

<0.001 <0.001 <0.001 <0.001 <0.001 <0.001

Notes and abbreviations on following page.

haematologica | 2022; 107(4)

847

J. Zou et al.

Note: Percentages may not add up to 100 because of rounding. P values of categorical variables were from the Fisher exact or c2 test. P values of continuous variables were from analysis of variance or the Kruskal-Wallis test. There were four missing data points for donor-recipient cytomegalovirus serostatus and one missing data point for the graft-versus-host disease regimen. HSCT: hematopoietic stem cell transplantation; AlloHSCT: allogeneic hematopoietic stem cell transplantation; AutoHSCT: autologous hematopoietic stem cell transplantation; TCE: T-cell epitope; GVH: graft-versus-host; HVG: host-versus-graft; CMV: cytomegalovirus; NR: nonreactive; R: reactive; AML: acute myeloid leukemia; MDS: myelodysplastic syndrome; DRI: Disease Risk Index; HCT-CI: Hematopoietic Cell Transplant-Comorbidity Index; MA: myeloablative; RIC: reduced-intensity conditioning; NMA: nonmyeloablative; PB: peripheral blood; BM: bone marrow; GVHD: graft-versus-host disease; PTCY: post-transplant cyclophosphamide; ATG: antithymocyte globulin; DP ME: HLA-DPB1 mismatched eplets; PS-I, Predicted Indirectly Recognizable HLA Epitopes score I; PS-II, Predicted Indirectly Recognizable HLA Epitopes score II.

GVH PS-I and GVH PS-II with grade 2-4 acute GVHD risk resulted in an increased risk of NRM (GVH PS-I: HR=1.31, 95% CI: 1.07-1.60, P=0.008; GVH PS-II: HR=1.34, 95% CI: 1.10-1.63, P=0.004), whereas HVG PSI (HR=1.22, 95% CI: 1.01-1.49, P=0.041), but not HVG PS-II, was associated with an increased risk of NRM, and neither GVH nor HVG ME was significantly associated with NRM. In the analysis of combined groups, NRM risk was highest in those with high GVH and high HVG PS-I (HR=1.48, 95% CI: 1.15-1.91, P=0.002) and in those with high GVH and high HVG PS-II (HR=1.50, 95% CI: 1.161.94, P=0.002) (Figure 1B).

HLA-DPB1 nonpermissive mismatch in the GVH direction was associated with not only an increased risk of acute GVHD but also a reduced risk of relapse (HR=0.64, 95% CI: 0.47-0.86, P=0.003), whereas permissive mismatch and HVG nonpermissive mismatch were not significantly associated with risk of relapse. Similar results were seen in patients with high GVH ME, PS-I, and PS-II, which were associated with reduced risk of relapse (ME: HR=0.83, 95% CI: 0.70-0.99, P=0.05; PS-I: HR=0.82, 95% CI: 0.68-0.98, P=0.032; PS-II: HR=0.79, 95% CI: 0.66-0.95, P=0.011), whereas HVG ME, PS-I, and PS-II were not associated with a reduced

Figure 1. Figure continued on following page.

848

haematologica | 2022; 107(4)

Refined HLA-DPB1 mismatch with molecular algorithms

Figure 1. Forest plots showing results from multivariable analyses of the impact of molecular mismatch scores (ME, PS-I, PS-II) and traditional T-cell epitope grouping on outcomes, stratified by the mismatch in the graft-versus-host and host-versus-graft direction. (A) Acute graft-versus-host disease grade 2-4. (B) Non-relapse mortality. (C) Relapse. (D) Overall survival. Dots and bars in the forest plots represent adjusted hazard ratios and 95% confidence intervals. PS and ME were categorized into low and high groups using the median as a cutoff point. ME: mismatched eplets, PS: Predicted Indirectly Recognizable HLA Epitope score; GVH: graftversus-host: HVG: host-versus-graft; GVHD: graft-versus-host disease.

risk of relapse (Figure 1C). Relapse risk was significantly lower in patients with high GVH ME combined with low HVG ME than in patients with low ME in both directions (Figure 1C, Online Supplementary Figure S2B). Neither HLA-DPB1 mismatch permissiveness nor molecular mismatches were found to be associated with overall survival (Figure 1D, Online Supplementary Table S3), progression-free survival (Online Supplementary Table S4), or engraftment in the present study cohort. In the permissive mismatch group, GVH alloimmunity determined by ME and PS was associated with an increased risk of GVHD, whereas HVG alloimmunity determined by ME and PS was associated with an increased risk of relapse and GVHD. Consistent with the previous report,26 permissive mismatch represented the largest subgroup in our cohort of patients who underwent HSCT from unrelated donors. Results from the multivariable analyses showed that the haematologica | 2022; 107(4)

alloimmunity predicted by ME or PS, in either the HVG or the GVH direction, was associated with a trend of increased risk of grade 2-4 acute GVHD (Figure 2A). In particular, HVG PS-II was associated with a significantly increased risk of grade 2-4 acute GVHD (HR=1.43, 95% CI: 1.13-1.82, P=0.003). This finding was further confirmed by our analysis of combined groups, in which a significantly increased risk of grade 2-4 acute GVHD was observed in the group with high ME (Figure 2B) or PS-II in both directions. However, high GVH ME or PS without concurrent HVG alloimmunity was not associated with an increased risk of acute GVHD. Similar to what we observed in the entire cohort, no anti-leukemia benefit was associated with HVG alloresponse assessed by ME or PS. Moreover, high ME in the HVG direction was associated with an increased risk of relapse in the permissive mismatch group (HR=1.36, 95% CI: 1.02-1.76, P=0.026) (Figure 2C), and this was more 849

J. Zou et al.

pronounced in the group with high HVG ME coupled with low alloimmunity in the GVH direction (Figure 2D). Molecular mismatches assessed by ME or PS were not associated with the risk of NRM, overall survival, or progression-free survival in this permissive mismatch subgroup.

grade 2-4 acute GVHD (HR=2.82, 95% CI: 1.41-5.62, P=0.003) (Figure 3B). No significant association between the molecular mismatch factors and relapse (Online Supplementary Figure S3), NRM, engraftment, overall survival, or progressionfree survival was identified.

In the GVH nonpermissive mismatch group, ME in the GVH direction was associated with a higher incidence of grade 2-4 acute GVHD, and HVG ME could synergistically contribute to this risk. Alloimmunity quantified by ME appeared to be more clinically relevant than alloimmunity quantified by PS in the GVH nonpermissive mismatch group. Results from the multivariable analyses showed that high ME in the GVH direction was associated with an increased risk of grade 2-4 acute GVHD (HR=1.64, 95% CI: 1.16-2.31, P=0.005) (Figure 3A). Although HVG ME itself was not associated with the risk of acute GVHD, those with high ME in both directions had a significantly increased risk of

In the HVG nonpermissive mismatch group, ME and PS-I in the GVH direction were associated with worse NRM without an increased risk of GVHD None of the mismatch factors was associated with the risk of relapse or acute GVHD in the HVG nonpermissive mismatch group with high HVG alloimmunity settings (Online Supplementary Figure S4A, B). Although no association with the risk of acute GVHD was identified, alloimmunity in the GVH direction determined by ME and PS-I was associated with an increased risk of NRM (ME: HR=1.90, 95% CI: 1.18-3.07, P=0.008, Figure 4A; PS-I: HR=1.60, 95% CI:1.04-2.60, P=0.024, Figure 4B), indicat-

Figure 2. Figure continued on following page.

850

haematologica | 2022; 107(4)

Refined HLA-DPB1 mismatch with molecular algorithms

Figure 2. Forest plots showing results from the multivariable analyses of the impact of molecular mismatch scores (ME, PS-I, and PS-II) on outcomes in the permissive mismatch group, stratified by the mismatch in the graft-versus-host and host-versus-graft direction. (A) Acute graft-versus-host disease (GVHD) grade 2-4. (B) Adjusted cumulative incidence of acute GVHD grade 2-4. (C) Relapse. (D) Adjusted cumulative incidence of relapse. Dots and bars in the forest plots represent adjusted hazard ratios and 95% confidence intervals. PS and ME were categorized into low and high groups using the median as a cutoff point. ME: mismatched eplets, PS: Predicted Indirectly Recognizable HLA Epitope score; GVH: graft-versus-host: HVG: host-versus-graft; GVHD: graft-versus-host disease; HR: hazard ratio.

ing that the increased risk of NRM observed here may not be mostly attributed to GVHD. Additionally, a lower incidence of neutrophil engraftment was observed in the group with high ME in the HVG direction, likely attributable to the alloimmunity towards the graft (HR=0.73, 95% CI: 0.56-0.96, P=0.028 for low GVH ME + high HVG ME). Predictive performance of the TCE, ME, and PS models Results from the concordance test showed that the ME in the GVH direction provided better discriminative ability for the prediction of clinically significant acute GVHD with a concordance index of 0.595 compared with other models. The values of the concordance index of GVH PS I, GVH PS II, HVG ME, HVG PS I, HVG PS II, and TCE were haematologica | 2022; 107(4)

0.560, 0.556, 0.545, 0.541, 0.542, and 0.566, respectively. Moreover, decision curve analysis29 was conducted to compare the clinical application of different matching models. We found that ME in the GVH direction outperformed other models, including the conventional TCE model, and provided the best net clinical benefit for the modification of the acute GVHD prophylaxis regimen in patients with a high risk of developing clinically significant acute GVHD (Figure 5).

Discussion Relapse and GVHD remain two major causes of morbidity and mortality in patients with hematologic malig851

J. Zou et al.

nancies undergoing HSCT. It has been accepted that donor T-cell–mediated alloimmune responses are the key mediators of beneficial GVL and adverse GVHD effects. A better understanding of T-cell alloreactivity in patients receiving HSCT would help to minimize the risk of GVHD while still preserving GVL activity. With recent progress in bioinformatics and molecular HLA typing, in silico prediction of immunogenicity has evolved rapidly, and several algorithms with a different focus have been shown to be predictive of outcomes in patients who have undergone HSCT.22,31 In the present comprehensive study in a cohort of patients with hematologic malignancies, we demonstrat-

ed that HLAMatchmaker and PIRCHE can be used to assess histocompatibility in HSCT at the molecular level. Using the decision curve analysis method that incorporates clinical considerations, it was found that ME in the GVH direction has advantages over other predictive models including the conventional TCE model, in aiding the decision whether or not to modify the acute GVHD prophylaxis regimen. In patients with a high risk of developing clinically significant acute GVHD predicted by high ME in both GVH and HVG directions, the addition of therapy based on T-cell depletion to the prophylactic regimen may reduce the incidences and intensity of GVHD. Moreover, ME and PS can quantitatively refine the con-

Figure 3. Forest plots showing results from the multivariable analyses of the impact of molecular mismatch scores (ME, PS-I, PS-II) on outcomes in patients with HLA-DPB1 nonpermissive mismatch in the graft-versus-host (GVH) direction, stratified by ME GVH and host-versus-graft combinations. (A) Acute graft-versus-host disease (GVHD) grade 2-4. (B) Adjusted cumulative incidence of acute GVHD grade 2-4. Dots and bars in the forest plots represent adjusted hazard ratios and 95% confidence intervals. PS and ME were categorized into low and high groups using the median as a cutoff point. ME: mismatched eplets, PS: Predicted Indirectly Recognizable HLA Epitope score; GVH: graft-versus-host: HVG: host-versus-graft; GVHD: graft-versus-host disease; HR: hazard ratio.

852

haematologica | 2022; 107(4)

Refined HLA-DPB1 mismatch with molecular algorithms

ventional TCE grouping, so the finding here will aid prioritization of the donors even within the same TCE group. Using the HLA-DPB1 TCE model, Fleischhauer et al. concluded that mismatches in different directions (HVG versus GVH) did not differ in terms of acute GVHD and mortality risk.32 However, bidirectional mismatches seemed to work synergistically and were associated with an increased risk of GVHD. How to reconcile HVG alloimmunity remains unclear, because host T cells in circulation are believed to be depleted by conditioning regimens during HSCT. Recent studies indicate that peripheral host T cells resident in the skin and gut are stimulated by the mismatched HLA and, as a result, the activated

host T cells secrete higher levels of inflammatory cytokines and contribute to GVHD in addition to graft Tcell immunity.19,21 For the first time, we demonstrate that the direction of alloreactivity may be better reflected by ME or PS in different directions. The elicited GVH alloreactivity defined by PS and ME seems to contribute to GVL along with GVHD, whereas HVG alloreactivity is likely to augment GVHD without the anti-leukemia effect. In the HLA-DPB1 permissive mismatch group, the largest subgroup of patients within our cohort, the elicited HVG alloreactivity appears to counteract the antileukemia effect exerted by GVH alloimmunity, discouraging the use of donors with a high load of HVG ME/PS in patients with HLA-DPB1 permissive mismatch. These

Figure 4. Adjusted cumulative incidence of non-relapse mortality in patients with HLA-DPB1 nonpermissive mismatch in the host-versus-graft direction. (A) Stratified by the number of mismatched eplets (ME) in the graft-versus-host (GVH) direction. (B) Stratified by Predicted Indirectly Recognizable HLA Epitopes score-I (PS-I) in the GVH direction. HR: hazard ratio.

haematologica | 2022; 107(4)

853

J. Zou et al.

Figure 5. The clinical net benefit of the TCE, ME, and PS models in deciding to modify graft-versus-host disease (GVHD) prophylaxis regimen for patients with a high risk of developing clinically significant acute GVHD in comparison with a “treat/modify all” and “treat/modify none” strategy. Y-axis represents the net clinical benefit (positive values) or risk (negative values) of using model-guided GVHD regimen modification in comparison with no GVHD regimen modification (net clinical benefit =0). The X-axis represents threshold probabilities of acute GVHD grade 2-4 at 100 days after transplantation. TCE: T-cell epitopes; ME: mismatched eplets, PS: Predicted Indirectly Recognizable HLA Epitope score; GVH: graft-versus-host: HVG: host-versus-graft; aGVHD: acute graft-versushost disease.

findings not only assist donor selection and risk stratification in HSCT from unrelated donors but also provide valuable insights into the mechanism and process of alloimmunity in this setting. In agreement with recent studies on DP mismatches using the TCE model33 or DP expression model,34 associations of the nonpermissive mismatch and overall survival or transplant-related mortality were not found in our cohort. This is perhaps attributable to a high degree of HLA matching degree in the cohort, recent advances in GVHD prophylaxis and reduced incidence of severe GVHD. The majority of our patients received in-vivo Tcell depletion which may lessen the alloresponse derived from DP mismatch and reduce the severity and incidence of acute GVHD.35 Additionally, several recent studies documented an improved outcome with post-transplant cyclophosphamide in patients receiving not only haploidentical transplants but also in transplants from matched unrelated donors,36 it may be particularly effective for individuals with high ME/PS due to the profound effect of this treatment on GVHD outcomes compared with conventional GVHD prevention regimens.37 However, due to the low number of patients who received post-transplant cyclophosphamide in the current study, future large prospective studies are warranted to confirm our hypothesis. The predictive value of the HLAMatchmaker and PIRCHE algorithms has been demonstrated in HSCT from HLA-mismatched unrelated donors or haploidentical donors.24,31,38 Although HLAMatchmaker mainly focuses on epitopes directly recognized by B cells, alloreactive T-cell clones that are specific to certain eplets identified by HLAMatchmaker have also been found,39-41 suggesting that HLAMatchmaker reveals many polymorphic residues overlapping in both B-cell epitopes and T-cell epitopes. Consistent with a previous study,42 we observed a considerable correlation between ME load and PS. However, the disparity determined by ME load appears to be more clinically relevant in our study. Analysis of the topographic location of immunogenic amino acids identified with both methods demonstrated 854

that a significant number of polymorphic amino acids, especially in the b-sheet and a-3 domain, were not colocalized.42 Therefore, an optimized algorithm that considers both direct and indirect alloresponses would be more predictive of risks or benefits in the context of HSCT with HLA-mismatched donors. Unlike the TCE and the expression model that has been extensively studied and shown to be clinically relevant for HSCT in several high power studies,7,8,17,18 the molecular mismatching algorithms have been primarily studied in the solid organ transplant setting in the assessment of antibody-mediated rejection. The predictive value of ME or PIRCHE was only reported in a few small studies in HSCT settings.38,43-45 and further validation is warranted before routine clinical application. The heterogeneity of the cohort and retrospective nature of the current study may have biased our results. In conclusion, molecular HLA disparity and subsequent alloresponse assessed by in silico methods are useful in the prediction of clinical outcomes. In addition to conventional TCE grouping, additional information provided by ME and PS can be used to refine the permissiveness of HLA-DPB1 mismatches. In the present study, high alloimmunity in both the HVG and GVH directions, revealed by high PS or ME, is associated with an increased risk of GVHD. Nevertheless, only GVH ME or PS was associated with a reduced risk of relapse. An integrated study in which patients’ immune cells are characterized and comprehensively analyzed will provide deeper and better insights into the process of GVH response and the contribution from host T cells. Disclosures No conflicts of interest to disclose. Contributions JZ, PK, REC, and KC designed the study and contributed to collecting and interpreting the data and writing the manuscript; JZ and PK wrote the initial draft of the manuscript; PK, JM, and LL contributed to the statistical analysis and interpretation of statistical results and reviewed and approved the manuscript; haematologica | 2022; 107(4)

Refined HLA-DPB1 mismatch with molecular algorithms

BO, VK, YC, SS, HCC, DP, SOC, and QM contributed to the data collection and analysis and reviewed and approved the manuscript; GR contributed to data collection and reviewed and approved the manuscript; BO, SS, UG, EJS, and REC contributed to the treatment of patients and reviewed, edited, and approved the final version of the manuscript. Acknowledgments The authors would like to thank Kevin Harrell and Dr. JarHow Lee from Thermo Fisher Scientific for their help in eplet analysis for this manuscript. We thank Erica Goodoff, Senior Scientific Editor in the Research Medical Library at The

References 1. Tanaka Y, Kurosawa S, Tajima K, et al. Analysis of non-relapse mortality and causes of death over 15 years following allogeneic hematopoietic stem cell transplantation. Bone Marrow Ttransplant. 2016;51(4):553559. 2. Pasquini MC, Wang Z, Horowitz MM, Gale RP. 2010 report from the Center for International Blood and Marrow Transplant Research (CIBMTR): current uses and outcomes of hematopoietic cell transplants for blood and bone marrow disorders. Clin Transpl. 2010;87-105. 3. Lee SJ, Klein J, Haagenson M, et al. High-resolution donor-recipient HLA matching contributes to the success of unrelated donor marrow transplantation. Blood. 2007;110 (13):4576-4583. 4. Petersdorf EW, Hansen JA, Martin PJ, et al. Major-histocompatibility-complex class I alleles and antigens in hematopoietic-cell transplantation. N Engl J Med. 2001;345 (25):1794-1800. 5. Petersdorf EW, Kollman C, Hurley CK, et al. Effect of HLA class II gene disparity on clinical outcome in unrelated donor hematopoietic cell transplantation for chronic myeloid leukemia: the US National Marrow Donor Program experience. Blood. 2001;98(10): 2922-2929. 6. Petersdorf EW, Gooley T, Malkki M, et al. The biological significance of HLA-DP gene variation in haematopoietic cell transplantation. Br J Haematol. 2001;112(4):988-994. 7. Fleischhauer K, Shaw BE, Gooley T, et al. Effect of T-cell-epitope matching at HLADPB1 in recipients of unrelated-donor haemopoietic-cell transplantation: a retrospective study. Lancet Oncol. 2012;13(4): 366-374. 8. Pidala J, Lee SJ, Ahn KW, et al. Nonpermissive HLA-DPB1 mismatch increases mortality after myeloablative unrelated allogeneic hematopoietic cell transplantation. Blood. 2014;124(16):25962606. 9. Varney MD, Lester S, McCluskey J, Gao X, Tait BD. Matching for HLA DPA1 and DPB1 alleles in unrelated bone marrow transplantation. Hum iImmunol. 1999;60(6):532-538. 10. Hurley CK, Baxter-Lowe LA, Begovich AB, et al. The extent of HLA class II allele level disparity in unrelated bone marrow transplantation: analysis of 1259 National Marrow Donor Program donor-recipient pairs. Bone Marrow Transplant. 2000;25 (4):385-393. 11. Zino E, Frumento G, Marktel S, et al. A Tcell epitope encoded by a subset of HLADPB1 alleles determines nonpermissive mismatches for hematologic stem cell trans-

haematologica | 2022; 107(4)

University of Texas MD Anderson Cancer Center, for editing this article. Funding VK acknowledges funding from an NIHR Fellowship (PDF2016-09-065). This research was partially supported by the Cancer Center Support Grant of MD Anderson (NIH: P30CA016672 to L.L.). Data-sharing statement For data sharing, contact the corresponding author: jzou@mdanderson.org

plantation. Blood. 2004;103(4):1417-1424. 12. Fleischhauer K, Locatelli F, Zecca M, et al. Graft rejection after unrelated donor hematopoietic stem cell transplantation for thalassemia is associated with nonpermissive HLA-DPB1 disparity in host-versusgraft direction. Blood. 2006;107(7):29842992. 13. Fleischhauer K, Shaw BE. HLA-DP in unrelated hematopoietic cell transplantation revisited: challenges and opportunities. Blood. 2017;130(9):1089-1096. 14. Fleischhauer K, Beelen DW. HLA mismatching as a strategy to reduce relapse after alternative donor transplantation. Semin Hematol. 2016;53(2):57-64. 15. Shaw BE, Robinson J, Fleischhauer K, Madrigal JA, Marsh SG. Translating the HLA-DPB1 T-cell epitope-matching algorithm into clinical practice. Bone Marrow Transplant. 2013;48(12):1510-1512. 16. Crocchiolo R, Zino E, Vago L, et al. Nonpermissive HLA-DPB1 disparity is a significant independent risk factor for mortality after unrelated hematopoietic stem cell transplantation. Blood. 2009;114(7):14371444. 17. Petersdorf EW, Malkki M, O'Huigin C, et al. High HLA-DP expression and graft-versushost disease. N Engl J Med. 2015;373(7):599609. 18. Morishima S, Shiina T, Suzuki S, et al. Evolutionary basis of HLA-DPB1 alleles affects acute GVHD in unrelated donor stem cell transplantation. Blood. 2018;131(7):808817. 19. Divito SJ, Aasebo AT, Matos TR, et al. Peripheral host T cells survive hematopoietic stem cell transplantation and promote graft-versus-host disease. J Clin Invest. 2020;130(9):4624-4636. 20. Young JW. Alternative mechanisms that mediate graft-versus-host disease in allogeneic hematopoietic cell transplants. J Clin Invest. 2020;130(9):4532-4535. 21. Jardine L, Cytlak U, Gunawan M, et al. Donor monocyte-derived macrophages promote human acute graft-versus-host disease. J Clin Invest. 2020;130(9):4574-4586. 22. Zou J, Ciurea SO, Kongtim P, et al. Molecular disparity in human leukocyte antigens is associated with outcomes in haploidentical stem cell transplantation. Blood Adv. 2020;4(15):3474-3485. 23. Kramer CSM, Israeli M, Mulder A, et al. The long and winding road towards epitope matching in clinical transplantation. Transpl Int. 2019;32(1):16-24. 24. Rimando J, Slade M, DiPersio JF, et al. HLA epitope mismatch in haploidentical transplantation is associated with decreased relapse and delayed engraftment. Blood Adv. 2018;2(24):3590-3601.

25. Geneugelijk K, Thus KA, Spierings E. Predicting alloreactivity in transplantation. J Immunol Res. 2014;2014:159479. 26. Oran B, Saliba RM, Carmazzi Y, et al. Effect of nonpermissive HLA-DPB1 mismatches after unrelated allogeneic transplantation with in vivo T-cell depletion. Blood. 2018;131(11):1248-1257. 27. Duquesnoy RJ. Reflections on HLA epitopebased matching for transplantation. Front Immunol. 2016;7:469. 28. Crivello P, Zito L, Sizzano F, et al. The impact of amino acid variability on alloreactivity defines a functional distance predictive of permissive HLA-DPB1 mismatches in hematopoietic stem cell transplantation. Biol Blood Marrow Transplant. 2015;21(2):233241. 29. Vickers AJ, Elkin EB. Decision curve analysis: a novel method for evaluating prediction models. Med Decis Making. 2006;26(6):565574. 30. Vickers AJ, van Calster B, Steyerberg EW. A simple, step-by-step guide to interpreting decision curve analysis. Diagn Progn Res. 2019;3:18. 31. Rimando J, Slade M, DiPersio JF, et al. The Predicted Indirectly Recognizable HLA Epitopes (PIRCHE) score for HLA class I graft-versus-host disparity is associated with increased acute graft-versus-host disease in haploidentical transplantation with posttransplantation cyclophosphamide. Biol Blood Marrow Transplant. 2020;26(1):123131. 32. Fleischhauer K, Ahn KW, Wang HL, et al. Directionality of non-permissive HLA-DPB1 T-cell epitope group mismatches does not improve clinical risk stratification in 8/8 matched unrelated donor hematopoietic cell transplantation. Bone Marrow Ttransplant. 2017;52(9):1280-1287. 33. Gagne K, Loiseau P, Dubois V, et al. Is there any impact of HLA-DPB1 disparity in 10/10 HLA-matched unrelated hematopoietic SCT? Results of a French multicentric retrospective study. Bone Marrow Ttransplant. 2015;50(2):232-236. 34. Petersdorf EW, Bengtsson M, De Santis D, et al. Role of HLA-DP expression in graft-versus-host disease after unrelated donor transplantation. J Clin Oncol. 2020;38(24):27122718. 35. Arai Y, Jo T, Matsui H, Kondo T, TakaoriKondo A. Efficacy of antithymocyte globulin for allogeneic hematopoietic cell transplantation: a systematic review and metaanalysis. Leuk Lymphoma. 2017;58(8):18401848. 36. Kanakry CG, O'Donnell PV, Furlong T, et al. Multi-institutional study of post-transplantation cyclophosphamide as single-agent graft-versus-host disease prophylaxis after

855

J. Zou et al. allogeneic bone marrow transplantation using myeloablative busulfan and fludarabine conditioning. J Clin Oncol. 2014;32(31): 3497-3505. 37. Ruggeri A, Sun Y, Labopin M, et al. Posttransplant cyclophosphamide versus antithymocyte globulin as graft- versus-host disease prophylaxis in haploidentical transplant. Haematologica. 2017;102(2):401-410. 38. Geneugelijk K, Thus KA, van Deutekom HWM, et al. Exploratory study of predicted indirectly recognizable HLA epitopes in mismatched hematopoietic cell transplantations. Front Immunol. 2019;10:880. 39. van Seventer GA, Huis B, Melief CJ, Ivanyi P. Fine specificity of human HLA-B7-specific cytotoxic T-lymphocyte clones. I. Identification of HLA-B7 subtypes and his-

856

totopes of the HLA-B7 cross-reacting group. Hum Immunol. 1986;16(4):375-389. 40. Smith KD, Epperson DF, Lutz CT. Alloreactive cytotoxic T-lymphocytedefined HLA-B7 subtypes differ in peptide antigen presentation. Immunogenetics. 1996;43(1-2):27-37. 41. Hiraiwa M, Yamamoto J, Matsumoto K, et al. T cell can recognize the allospecificities formed by the substitution of amino acids associated with HLA-Bw4/Bw6 public epitopes. Hum Immunol. 1991;32(1):41-45. 42. Otten HG, Calis JJ, Kesmir C, van Zuilen AD, Spierings E. Predicted indirectly recognizable HLA epitopes presented by HLADR correlate with the de novo development of donor-specific HLA IgG antibodies after kidney transplantation. Hum Immunol.

2013;74(3):290-296. 43. Thus KA, de Hoop TA, de Weger RA, Bierings MB, Boelens JJ, Spierings E. Predicted indirectly recognizable HLA epitopes class I promote antileukemia responses after cord blood transplantation: indications for a potential novel donor selection tool. Biol Blood Marrow Transplant. 2016;22(1):170-173. 44. Stenger W, Kunkele A, Niemann M, et al. Donor selection in a pediatric stem cell transplantation cohort using PIRCHE and HLA-DPB1 typing. Pediatr Blood Cancer. 2020;67(3):e28127. 45. Thus KA, Te Boome L, Kuball J, Spierings E. Indirectly recognized HLA-C mismatches and their potential role in transplant outcome. Front Immunol. 2014;5:210.

haematologica | 2022; 107(4)

ARTICLE

Cell Therapy & Immunotherapy

Comparison of immune reconstitution between anti-T-lymphocyte globulin and posttransplant cyclophosphamide as acute graft-versus-host disease prophylaxis in allogeneic myeloablative peripheral blood stem cell transplantation Radwan Massoud, Nico Gagelmann, Ulrike Fritzsche-Friedland, Gaby Zeck, Silke Heidenreich, Christine Wolschke, Francis Ayuk, Maximilian Christopeit and Nicolaus Kröger

Ferrata Storti Foundation

Haematologica 2022 Volume 107(4):857-867

Department of Stem Cell Transplantation, University Medical Center HamburgEppendorf, Hamburg, Germany

ABSTRACT

nti-T-cell lymphocyte globulin (ATLG) and posttransplant cyclophosphamide (PTCy) are now widely used strategies to prevent graft-versus-host disease (GVHD) after allogeneic stem cell transplantation. Data comparing immune reconstitution (IR) between ATLG and PTCy is scarce. This retrospective study conducted at the University Medical Center Hamburg-Eppendorf (UKE) compares PTCy (n=123) and ATLG (n=476) after myeloablative allogeneic peripheral blood stem cell transplant. Detailed phenotypes of T, B natural killer (NK), natural killer T (NKT) cells were analyzed by multicolor flow at day 30, 100 and 180 posttransplant. Incidence of infections, viral reactivations, GVHD and relapse were collected. Neutrophil engraftment was significantly delayed in the PTCy group (median day 12 vs. day 10, P<0.001) with a high incidence of infection before day+100 in the PTCy arm but a higher Epstein-Barr virus reactivation in the ATLG arm and comparable cytomegalovirus reactivation. Overall incidence of acute GVHD was similar but moderate/severe chronic GVHD was seen more often after PTCy (44% vs. 38%, P=0.005). ATLG resulted in a faster reconstitution of CD8+ T, NK, NKT and gdT cells while CD4 T cells and B cells reconstituted faster after PTCy. Similar reconstitution was observed for T-regulatory cells and B cells. Non-relapse mortality relapse incidence, disease-free survival, and overall survival did not differ significantly between both arms. Even though differences in IR were related to a decreased incidence of infection and moderate/severe cGVHD in the ATLG group they had no impact on any of the other long-term outcomes. However, it remains undetermined which regimen is better as GVHD prophylaxis.

Correspondence: NICOLAUS KRÖGER nkroeger@uke.uni-hamburg.de Received: September 10, 2020. Accepted: March 31, 2021. Pre-published: April 8, 2021. https://doi.org/10.3324/haematol.2020.271445

Introduction Allogeneic stem cell transplantation (allo-SCT) is a potentially curative treatment strategy for hematological diseases.1 This is attributed mainly to the graftversus-tumor effect derived from transferring the donor’s immune system to the recipient.2 However, the benefit of allo-SCT may be offset by increased transplantrelated mortality (TRM) especially due to graft- versus-host disease (GVHD).3 In an attempt to decrease the incidence of GVHD physicians employ a multitude of strategies, including in vivo T-cell depletion (TCD) with pretransplant anti-T-lymphocyte globulin (ATLG)4,5 and/or posttransplant cyclophosphamide (PTCy).6-8 GVHD prophylaxis with PTCy decreases the incidence of graft rejection and GVHD by impairing the function of alloreactive T cells.9 However, data is scarce on the effect of PTCy effect on immune reconstitution (IR) post-allo-SCT, especially when compared to the standard use of ATLG as a TCD tool.4,5,10,11 In this

haematologica | 2022; 107(4)

857

R. Massoud et al.

study we aim to compare the IR kinetics and transplant outcomes between ATLG and PTCy as TCD strategies in patients undergoing allo-SCT with myeloablative conditioning (MAC) and peripheral blood stem cells (PBSC).

Methods This retrospective study conducted at the University Medical Center Hamburg-Eppendorf (UKE) with a primary outcome to compare IR between PTCy versus ATLG in vivo TCD in adult patients who received MAC PBSC allo-SCT. Secondary outcomes included incidence of viral reactivations, engraftment, infections, acute GVHD (aGHVD), chronic GVHD (cGVHD), non-relapse mortality (NRM), progression-free survival (PFS), overall survival (OS). All patients signed written informed consents for treatment, and the study was approved by the Institutional Review Board of UKE. MAC regimens were defined according to working group definition.12 ATLG (Grafalon®, Neovii, Switzerland) was given at a dose of 30 mg/kg for related donor or 60 mg/kg for unrelated donor with a trend in recent years to give the lower dose for both groups. All ATLG doses were fractionated between days 4 to -1. PTCy was administered as 50 mg/kg/day. Posttransplant immunosuppression was given on days +3 and +4 combined with calcineurin inhibitor (tacrolimus for unrelated donor or haplo-identical donor, and cyclosporine for related donor) and mycophenolate mofetil for mismatched transplants. Similar supportive care was used for all patients per institutional guidelines including antimicrobial prophylaxis consisting of fluoroquinelone for bacterial infections, trimethorpin-sulfamethoxazole or pentamidine for Pnemocystis jiroveci, micafungin for fungal infections and acyclovir for viral infections. Patients were screened weekly for cytomegalovirus (CMV) and Epstein-Barr virus (EBV) by blood polymerase chain reaction (PCR). Neutrophil engraftment was defined as the first 3 consecutive days with a measure of an absolute neutrophil count >0.5x109/L. Platelet engraftment was defined as the first consecutive days with a platelet count >20x109/L without transfusion support. Acute GVHD was graded according to standard criteria.13 Chronic GVHD was graded according to National Institute of Health (NIH) criteria routinely at every visit after transplantation.14

Infections were defined as any microbial testing with a positive result and requiring therapy. As per institution guidelines, blood samples were collected for each patient on days 30, 100 and 180 post-allo-SCT. Samples were used directly after red blood cell lysis following 10 minutes of incubation with erythrocyte lysing reagent without fixative. Immunophenotypes were assessed using four color cytometry using mouse anti-human antibodies for the following cells: T lymphocytes (CD3+), activated T lymphocytes (CD3+HLADR+), T-helper cells (CD3+/CD4+), cytotoxic T cells (CD3+/CD8+), B lymphocytes (CD19+), B-lymphocyte subpopulations (CD19+CD5+CD1d+)(CD19+CD27+), naïve B cells (CD19+CD27-CD10+), natural killer (NK) cells (CD56+CD3-), natural killer T (NKT) cells (CD56+CD3+), naïve T-helper cells (CD4+CD45RA+), memory T-helper cells (CD4+CD45R0+), naïve cytotoxic T cells (CD8+CD45RA+), memory cytotoxic T cells (CD8+CD45R0+), gdT cells (gdTCR+,CD3+), regulatory T cells (CD4+CD25+CD127low-neg).

Statistical analyses All data was retrospectively collected, and was summarized

858

by standard descriptive statistical methods. c2 test was used to compare categorical variables, whereas continuous variables were compared using Student’s t-test. We defined disease-free survival (DFS) as survival without relapse or progression of hematological disease; we censored patients without disease or progression at the time of the last follow-up. We defined OS and NRM as death from any cause, and without evidence of relapse respectively. We used the Kaplan-Meier method to calculate the probabilities of moderate/severe cGVHD relapse-free survival, DFS and OS; and the cumulative incidence functions were used to estimate RI and NRM in a competing risk setting. All analysis was performed using SPSS version 26.0 and ACCorD.

Results Patients and transplant characteristics In order to have comparable groups we selected only patients receiving myeloablative conditioning for their first allo-SCT with only PBSC as a stem cell source. From 2005 to 2019, 599 patients were included in the study. Four hundred and seventy-six patients received ATLG, with 34% and 66% receiving 30 mg/kg and 60 mg/kg ATLG respectively. One hundred and twenty-three patients received PTCy. The median age at transplant was 53 years (range, 18-75 years) in both groups. Seventy nine percent and 72% were transplanted from a full match donor (HLA10/10) in the ATLG and PTCy group, respectively. All patients, donor and transplant characteristics are listed in Table 1.

Transplant outcomes All transplant outcomes are summarized in Table 2.

Engraftment Platelet and neutrophil engraftment were significantly delayed in the PTCy group when compared to the ATLG group; with a median of 12 days (range, 8-36 days) to neutrophil in the ATLG versus 16 days (range, 12-27 days) in PTCy group (P<0.001); and a median of 15 days (range, 8-99 days) to platelet engraftment in the ATLG versus 21 days (range, 9-99 days) in the PTCy group (P=0.024).

Cytomegalovirus infections and Epstein-Barr virus reactivation We observed no significant differences in incidence of CMV reactivation before day 100 (ATLG 46%, PTCy 50%). The overall incidence of infection before day 100 was significantly higher in the PTCy group (91%) when compared to the ATLG group (75%), P<0.001. The incidence of EBV reactivation before days 100 in the ATLG group was higher than the PTCy group (33% vs. 16%, P<0.001).

Graft-versus-host disease The cumulative incidences of aGVHD grade 2-4 and 34 were similar between the two groups with 36% and 15% in the ATLG group and 40% and 12 % in the PTCy group. The cumulative incidence of all grade cGVHD were similar between the two groups, 15% and 27% in the ATLG and PTCy groups respectively, however we observed a higher cumulative incidence of moderate and severe cGVHD in the PTCy group, 38% ATLG versus 44% PTCy (P=0.005). haematologica | 2022; 107(4)

Immune reconstitution with ATLG vs. PTCy

continued from previous coloum

Table1. Patients and transplant characteristics.

Patients

ATLG N (%)

Total patients 476 (100) ATLG dose 30 mg/kg 162 (34) 60 mg/kg 314 (66) Mean age (Standard Deviation) 50 (14) Disease ALL 27 (10) AML 206 (43) CML 16 (3) MDS 43 (9) MDS-MPN 6 (1) HL 4 (1) NHL 75 (16) MM 64 (13) PMF 12 (3) Other AL 3 (1) ECOG 0 115 (30) 1 241 (63) 2 23 (6) 3 3 (1) Mean KI at SCT (Standard Deviation)86 (12) Mean donor age (Standard Deviation)36 (12) Donor/ recipient CMV serology D-/R151 (32) D+/R+ 194 (41) D-/R+ 62 (13) D+/R68 (14) Donor-Recipient sex No mismatch 322 (67) Male-Male 241 (51) Female-Female 81 (17) Mismatch 154 (32) Male - Female 101 (21) Female - Male 53 (11) ABO incompatibility Isogroup 182 (39) Minor 127 (27) Major 111 (24) Bidirectional 48 (10) Median year of transplant (range) 2013 (2005-2019) Type of transplant Related 77 (16) Unrelated 399 (84) Full match (HLA10/10) 377 (79) Mismatch (HLA <10/10) 99 (21) MRD 74 (16) MMRD 3 (1) MUD 303 (64) MMUD 96 (20) 6 Mean CD34 x10 /kg (SD) infused 11.55 (64) Conditioning details Busulfan based 256 (54) TBI based 130 (27) Other 90 (19) TBI 146 (31) TBI dose ≤10 Gy 44 (9) >10 Gy 102 (21)

PTCy N(%)

123 (100)

50 (13) 35 (29) 23 (19) 1 (1) 2 (2) 4 (3) 2 (2) 13 (11) 38 (31) 2 (2) 3 (2) 23 (20) 81 (72) 9 (8) 0 (0) 83 (11) 37 (14) 41 (34) 58 (47) 9 (7) 15 (12) 75 (61) 59 (48) 16 (13) 48 (39) 32 (26) 16 (13) 57 (47) 23 (19) 33 (27) 8 (7) 2015 (2005-2019) 45 (37) 78 (63) 88 (72) 35 (29) 31 (25) 14 (11) 57 (46) 21 (17) 7.18 (2) 29 (24) 55 (45) 39 (32) 55 (45) 21 (17) 34 (28)

NS <0.001

Immune suppression CNI+MMF CNI+MTX CNI other

413 (87) 57 (12) 1 (0.2) 5 (1)

<0.001

77 (62) 0 (0) 23 (19) 23 (19)

ATLG: anti T-cell lymphocyte globulin; PTCy: and post-transplant cyclophosphamide; NS: statistically not significant; ALL: acute lymphoblastic leukemia; AML: acute myeloid leukemia; CML: chronid myeloid leukemia; MDS: myelodysplastic syndrome; MDS-MPN: myelodysplastic syndrome - myeloproliferative neoplasm; HL: Hodgkin lymphoma; NHL: non Hodgkin lymphoma; MM: multiple myeloma; PMF: primary myelofibrosis; Other AL: other acute leukemia; ECOG: Eastern Cooperative Oncology Group performance status; KI: Karnofsy index: CMV: cytomegalovirus; D-: donor with negative CMV serology; D+: donor with positive CMV serology; R-: recipient with negative CMV serology; R+: recipient with positive CMV serology; MRD: matched related door; MMRD: mismatched related donor; MUD: matched unrelated donor; MMUD: mismatched unrelated donor; SD: standard deviation; TBI: total body irradiation; CNI: calcineurin inhibitor; MMF: mycophenolate mofetil; age in years.

Table 2. Transplant outcomes.

Transplant outcomes ATLG N (%) NS

0.011 NS NS

<0.001

<0.001 NS NS <0.001

NS <0.001

0.003 0.001 NS

Patients Leukocytes engraftment median days (range) Platelet engraftment median days (range) CMV reactivation [15 in ATLG, 54 in PTCy missing] EBV reactivation [34 in ATG missing, 4 in PTCy missing] Overall incidence of infection by day 100 NRM at 3 years Relapse incidence DFS at 3 years OS at 3 years Moderate/severe cGVHD relapse-free survival

PTCy N(%)

476 (100) 12 (8-36)

123 (100) 16 (12-27) <0.001

15 (8-99)

21 (9-99)

0.024

214 (46)

60 (50)

131 (33)

19 (16)

<0.001

344 (75)

109 (91) <0.001

16% 34% 50% 65% 40%

30% 29% 42% 58% 27%

0.006 NS NS NS NS

NS: statistically non-significant; NRM: non-relapse mortality; OS: overall survival; DFS: disease-free survival; cGVHD: chronic graft-versus-host disease; EBV: Epstein-Barr virus; CMV: cytomegalovirus; PTCy: posttransplant cyclophosphamide; ATLG: anti Tcell lymphocyte globulin; SD: standard deviation.

Non-relapse mortality PTCy was associated with a higher NRM when compared to ATLG on univariate analysis, in addition a positive patient CMV serology, patient age >52 years, donor age >34 years and female donor were also associated with higher NRM. On multivariate analysis, NRM was not affected by ATLG and PTCy; only donor sex, patient age and CMV serology had a significant impact on NRM (Table3).

Relapse, disease-free survival and overall survival After a median follow-up of 16 months (range, 1-169 months) we observed no significant differences in terms of DFS (at 3 years ATLG 51%, PTCy 42%, P=0.3), relapse incidence (ATLG 34% vs. PTCy 29%, P=0.261), OS (65% vs. 58%, P=0.663) or moderate/severe cGVHD relapsefree survival at 3 years was (ATLG 40% vs. PTCy 27%, P=0.068) between the two groups.

Immune reconstitution We observed a faster reconstitution of CD3 T lympho-

haematologica | 2022; 107(4)

859

R. Massoud et al.

cytes (CD3+) (P<0.05) and activated T cells (CD3+/HLADR+) after ATLG than PTCy (P<0.05) (Figure 1A and B). In addition, as shown in Figure 2 and the Online Supplementary Table S1 the reconstitution of cytotoxic T cells (CD3+/CD8+) and also of naïve cytotoxic T cells (CD3+/CD8+/CD45RA+) (P=0.017) was significantly faster in the ATLG group (on day 90, P=0.002), while the reconstitution of memory cytotoxic T cells (CD3+/CD8+/CD45R0+) was comparable (Figure 2). In contrast to cytotoxic CD8+ positive cells, CD3 helper cells (CD3+/CD4+) had a trend for faster reconstitution in the PTCy group (Figure 3), which was sustained for naïve (CD4+/CD45RA+) (P=0.002) as well as for memory helper cells (CD4+/CD45R0+) (P<0.001). The reconstitution of B cells (CD19+) was similar in the ATLG and PTCy group (Figure 4A) and a trend for faster reconstitution of naïve B-cells (CD19+/CD27-/CD10+) was observed in the PTCy group (Figure 4B). NK-cell reconstitution was faster on day 30 in the ATLG group (P<0.001), however the values on day 100 and 180 were similar after ATLG and PTCy (Figure 5A). Furthermore, NKT cells and gdT cells reconstitution was faster in the ATLG group (P<0.05) at all time points (Figure 5B and C), while there were no significant differences in regulatory T-cell immune reconstitution between the two groups. All our data is summarized in the Online Supplementary Table S1. All our findings were confirmed in a donor subgroup analysis (Online Supplementary Table S2).

Discussion Although ATLG and PTCy are widely used for GVHD prevention in allo-SCT, data is scarce on their impact on immune reconstitution and only one small prospective study using RIC PBSC has compared immune reconstitution PTCy to ATLG so far.15 In this retrospective study, we compared the influence of ATLG to PTCY on immune reconstitution and transplant outcomes after myeloablative allogeneic PBSC transplant. In our study, we observed some strong differences in terms of cell counts, immune reconstitution, infections moderate/severe cGVHD and EBV reactivation between the two groups. Since NK and gdT cells have a protective role against many bacterial and viral infections including CMV16-32 the longer period of aplasia and the decreased numbers of NK and NKT cells in the PTCy group can explain the higher incidence of infections before day 100 in this group. One of the most recent studies of gdT-cell recovery and their association with transplant outcomes was conducted on 102 pediatric patients with acute leukemia.33 They reported significantly improved PFS and OS in patients with elevated gdT cells, these findings have also been reported in adults.34,35 In addition they reported a significantly lower incidence of infections with a total absence of bacterial infections in the high gdT-cell group.33 Our findings fall in line with the literature. We observed an early recovery of the gdT-cell population in both groups independent of the donor subtype. In addition, gdT cell were consistently higher in the ATLG group in all evaluations when compared to the PTCy group, which may explain the decreased overall incidence of infections in this group. Even though our study was not designed for long term outcomes, we observed no significant difference in DFS or OS between the two groups. 860

Table 3. Multivariate non-relapse mortality

Multivariate

NRM HR (95% CI) P-value

ATLG vs. PTCy

1.6 (0.98-2.48) 0.061 Patient CMV serology 1.73 (1.09-2.75) negative vs. positive 0.02 Patient age 1.69 (1.05-2.73) <52 vs. ≥ 52 0.03 Donor Age 1.32 (0.81-2.13) <34 vs. ≥ 34 0.26 Donor sex 1.61 (1.01-2.59) Male vs. Female 0.048 CD34 x106/kg 0.748 (0.46-1.2) <7.2 vs. ≥ 7.2 0.24 ECOG 1.4 (0.87-2.14) 3 vs. 0-2 0.17 NRM: non-relapse mortality; HR: hazard ratio; CI: confidence interval; ATLG: anti T-cell lymphocyte globulin; PTCy: posttransplant cyclophosphamide; ECOG: Eastern Cooperative Oncology group.

Retiere et al. observed a rapid NK recovery within day 30 after allo-HSCT in the ATLG and PTCy group, however while they reported an increase in NK-cell counts in the ATLG group, they did not observe any effect of the donor type on these values and concluded the recovery rate was a direct effect of the difference in GVHD prophylaxis between the two groups.15 Our results fall in line with Retiere’s, we observed rapid reconstitution of NK cells at day 30, and we observed a significantly higher percentage and absolute count of NK cells at day 30 in the ATLG group. This was validated by our donor subgroup analysis which allows us to conclude that this was a direct effect of the difference in the TCD strategy. This supports the hypothesis that ATLG spares NK cells while PTCy targets them.15 Rubio et al. reported that early recovery of NKT cells post T-cell-repleted allo-SCT, and a high NKT-cell dose in the graft are associated with protection from aGVHD.36,37 In addition Tae et al. report an increase in the incidence of aGVHD and of relapse in patients with lower NKT-cell counts post-allo-SCT.38 Retiere et al. did not observe any significant differences in the NKT-cell population between ATLG and PTCy.15 However, in our study, we observed a higher number and percentage of NKT cells in the ATLG group when compared to PTCy and this was confirmed by our subgroup donor analysis. Nonetheless, we did not observe any significant differences in the incidence of relapse or aGVHD, while we observed a significantly lower incidence of moderate and severe cGVHD in the ATLG group when compared to the PTCy group. Servais et al. studied the impact of ATLG on IR post MAC PBSC allo-SCT.39 They looked precisely at memory and naïve T cells and they observed that ATLG selectively depletes naïve CD4+ T and naïve CD8+ T cells whereas it does not significantly impact memory Tcells. 39 Our results fall in line with Servais et al., as we observed a progressive increase in the naïve to memory ratio both in the CD4+ T cells and CD8+ T cells, indicating that the effect of ATLG effect is more pronounced on naïve T cells than on memory T cells. However, in our study the effect of ATLG on naïve and memory CD4+ T cells was haematologica | 2022; 107(4)

Immune reconstitution with ATLG vs. PTCy

Figure 1. Comparison between ATLG and PTCy regarding immune reconstitution of (A) activated T cells CD3+/HLADR+) and (B) all T cells (CD3+) P*=P-value at day 30; P**=P-value at day 100; P***=P-value at day 180; %: percentage of cells; Absolut: absolute number of cells. TLG: anti Tcell lymphocyte globulin; PTCy: and post-transplant cyclophosphamide.

more pronounced than PTCy. In our study the reconstitution of CD8+ T cells was faster than that of CD4+ Tcells with CD8+ T cells recovering by day 100, while the CD4+ T-cell reconstitution was not observed at the last evaluation of the immune profile on day 180. In addition, the CD4+/CD8+ ratio did not return to normal in either of the two groups, which indicates incomplete recovery of the CD4+ T-cell compartment. These findings were confirmed by subgroup analysis according to the donor. It is well established that reconstitution of the T-cell compartment after allo-HSCT arises from both homeohaematologica | 2022; 107(4)

static peripheral expansion (HPE) of donor T cells transferred with the graft and from the novel production of naïve T cells in the thymus.40,41 In patients receiving MAC most of the T cells originate from HPE, and given that ATLG persist for several weeks in circulation,42,43 it can be hypothesized that ATLG selectively depletes donor naïve T cells while it spares other T-cell populations. This differential cytotoxic activity of ATLG has been demonstrated in vitro.44 In addition, HPE occurs more asymmetrically between T cells, with CD8+ T cells having higher proliferating potential by HPE when compared to CD4+ T cells.39 This may explain the decreased 861

R. Massoud et al.

Figure 2. Comparison between ATLG and PTCY regarding immune reconstitution of CD8+ cells. (A) Total CD8+ T cells; (B) naïve CD8+ T cells; (C) memory CD8+ T cells. P*=Pvalue at day 30; P**=P-value at day 100; P***=P-value at day 180; %: percentage of cells; Absolut: absolute number of cells. ATLG: anti T-cell lymphocyte globulin; PTCy: and post-transplant cyclophosphamide.

862

haematologica | 2022; 107(4)

Immune reconstitution with ATLG vs. PTCy

A Figure 3. Immune reconstitution for CD4+ Tcell ATLG versus PTCy. (A) Total CD4+ T cells; (B) naïve CD4+ T cells; (C) memory CD4+ T cells. P*=P-value at day 30; P**=P-value at day 100; P***=P-value at day 180; %: percentage of cells; Absolut: absolute number of cells. ATLG: anti T-cell lymphocyte globulin; PTCy: and post-transplant cyclophosphamide.

haematologica | 2022; 107(4)

863

R. Massoud et al.

A Figure 4. Comparison between ATLG and PTCy regarding immune reconstitution of B cells(A) Total B cells; (B) naïve B cells. P*=Pvalue at day 30; P**=P-value at day 100; P***=P-value at day 180; %: percentage of cells; Absolut: absolute number of cells. ATLG: anti T-cell lymphocyte globulin; PTCy: and post-transplant cyclophosphamide.

CD4+ T-cell population and increased CD8+ T-cell numbers in the ATLG group. Another explanation for the decreased number of CD8+ T cells in the PTCy group is the ability of PTCy to selectively target proliferating NK and CD8+ T cells more than CD4+ T cells.15 These findings were all validated in the donor subgroup analyses which makes it safe to assume that the observed discrepancies between the ATLG and PTCy groups can be attributed to the difference in the TCD between the groups. From our findings, we can hypothesize that PTCy has less impact on all the CD4+ T cells, while it has increased activity against CD8+ T cells, which was expressed by higher proliferation of CD4+ T cells in the PTCy and CD8+ T cells in the ATLG group. In addition, we can hypothesize that ATLG has a more pronounced effect on memory CD8+ and CD4+ T cells than PTCy. 864

The higher percentage of T lymphocytes and activated lymphocytes in the ATLG group may be explained by the higher CD8+ T-cell and NTK-cell reconstitution observed in this group. It has been proven in animal models that Tregs suppress GVHD without decreasing GVL,45 and that they accelerate post-transplant T-cell immune reconstitution in murine models.46 In our study Tregs persisted after transplant and we observed no significant differences in Treg reconstitution post-allo-HSCT between the two groups. This may be explained by the cyclophosphamide resistance of Tregs conferred by their increase in the expression of aldehyde dehydrogenase for cyclophosphamide detoxification which allow them to persist posttransplant in the PTCy setting,10,15 and by the selective sparing of Tregs by ATLG.39 After allo-HSCT the numbers of total B cells normalize haematologica | 2022; 107(4)

Immune reconstitution with ATLG vs. PTCy

Figure 5. Comparison between ATLG and PTCy regarding immune reconstitution of innate immune system (A) B natural killer (NK) cells; (B) natural killer T (NKT) cells; (C) gdT cells. P*=P-value at day 30; P**= P-value at day 100; P***=P-value at day 180; %: percentage of cells; Absolut: absolute number of cells. ATLG: anti T-cell lymphocyte globulin; PTCy: and post-transplant cyclophosphamide.

haematologica | 2022; 107(4)

865

R. Massoud et al.

within 3 months to 1 year.47-49 However, the reconstitution in the first year compromise mainly transitional and naïve subsets and memory B cells occurs much later. Our results show normalization of the total B-lymphocyte count at day 60, but no complete recovery of naïve and memory Bcells at last follow-up. In addition, we observed a higher percentage and count of naïve B cells in the PTCy group at day 100. Even though we observed a higher incidence of NRM in the PTCY group on univariate analysis, this did not persist on multivariate analysis (hazard ratio 1.6; 95% Confidence Interval: 0.98-2.48; P=0.061). This difference can be explained by a higher proportion of high risk patients (higher proportion of mismatch transplants and ALL) in the PTCY group when compared to ATLG. Given the retrospective nature of our study and the heterogeneity of our patient population especially the differences in donor types, ATLG dosing, immune suppression regiments and conditioning regimens, it should be noted that our study is prone to bias especially with univariate analysis. In an attempt to reduce bias and render the population more homogenous we selected consecutive patients undergoing allo-SCT only with MAC regimens and PBSC as a stem cell source. In addition, we conducted subgroup analysis in which we found anecdotal differences in clinical outcomes and IR between the 30 mg/kg and 60 mg/kg ATLG dose, and we confirmed our findings

References 1. Copelan EA. Hematopoietic stem-cell transplantation. N Engl J Med. 2006;354 (17):1813-1826. 2. Horowitz MM, Gale RP, Sondel PM, et al. Graft-versus-leukemia reactions after bone marrow transplantation. Blood. 1990;75(3): 555-562. 3. Glucksberg H, Storb R, Fefer A, et al. Clinical manifestations of graft-versus-host disease in human recipients of marrow from HL-A-matched sibling donors. Transplantation. 1974;18(4):295-304. 4. Chang YJ, Zhao XY, Huang XJ. Immune reconstitution after haploidentical hematopoietic stem cell transplantation. Biol Blood Marrow Transplant. 2014;20 (4):440-449. 5. Roberto A, Castagna L, Zanon V, et al. Role of naive-derived T memory stem cells in Tcell reconstitution following allogeneic transplantation. Blood. 2015;125(18):28552864. 6. Kanakry CG, O'Donnell PV, Furlong T, et al. Multi-institutional study of post-transplantation cyclophosphamide as singleagent graft-versus-host disease prophylaxis after allogeneic bone marrow transplantation using myeloablative busulfan and fludarabine conditioning. J Clin Oncol. 2014;32(31):3497-3505. 7. Kanakry CG, Tsai HL, Bolanos-Meade J, et al. Single-agent GVHD prophylaxis with posttransplantation cyclophosphamide after myeloablative, HLA-matched BMT for AML, ALL, and MDS. Blood. 2014;124 (25):3817-3827. 8. Jacoby E, Chen A, Loeb DM, et al. Singleagent post-transplantation cyclophosphamide as graft-versus-host disease prophylaxis after human leukocyte antigenmatched related bone marrow transplanta-

866

in a subgroup analysis according to donor type. However, our findings should be confirmed in more homogenous prospective studies.

Conclusion Acknowledging the bias associated with our study especially its retrospective nature, while taking into consideration the large sample size, it is safe to conclude that a better CD8+ T-cell, NK-cell, NKT-cell and gdT-cell reconstitution is observed in the ATLG group while improved CD4+ recovery is a hallmark of the PTCy group. Even though these findings have been translated into a decreased incidence of infection and moderate/severe cGVHD in the ATLG group they had no impact on any of the other long-term outcomes. So, it remains undetermined which TCD strategy is better to consider and results from well-designed randomized studies are needed. Disclosures No conflicts of interest to disclose. Contributions RM and NK designed the study, analyzed data, interpreted results, and wrote the manuscript; NG, UFF, GZ, SH, CW, FA, and MC collected and analyzed data. All authors approved the final version of the manuscript.

tion for pediatric and young adult patients with hematologic malignancies. Biol Blood Marrow Transplant. 2016;22(1):112-118. 9. Wachsmuth LP, Patterson MT, Eckhaus MA, Venzon DJ, Gress RE, Kanakry CG. Post-transplantation cyclophosphamide prevents graft-versus-host disease by inducing alloreactive T cell dysfunction and suppression. J Clin Invest. 2019;129(6): 2357-2373. 10. Kanakry CG, Coffey DG, Towlerton AM, et al. Origin and evolution of the T cell repertoire after posttransplantation cyclophosphamide. JCI Insight. 2016;1(5): e86252. 11. Rubio MT, Labopin M, Blaise D, et al. The impact of graft-versus-host disease prophylaxis in reduced-intensity conditioning allogeneic stem cell transplant in acute myeloid leukemia: a study from the Acute Leukemia Working Party of the European Group for Blood and Marrow Transplantation. Haematologica. 2015;100 (5):683-689. 12. Bacigalupo A, Ballen K, Rizzo D, et al. Defining the intensity of conditioning regimens: working definitions. Biol Blood Marrow Transplant. 2009;15(12):16281633. 13. Przepiorka D, Weisdorf D, Martin P, et al. 1994 Consensus Conference on Acute GVHD Grading. Bone Marrow Transplant. 1995;15(6):825-828. 14. 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. 15. Retiere C, Willem C, Guillaume T, et al. Impact on early outcomes and immune reconstitution of high-dose post-transplant cyclophosphamide vs anti-thymocyte glob-

ulin after reduced intensity conditioning peripheral blood stem cell allogeneic transplantation. Oncotarget. 2018;9(14):1145111464. 16. Horowitz A, Stegmann KA, Riley EM. Activation of natural killer cells during microbial infections. Front Immunol. 2011;2:88. 17. Adib-Conquy M, Scott-Algara D, Cavaillon JM, Souza-Fonseca-Guimaraes F. TLRmediated activation of NK cells and their role in bacterial/viral immune responses in mammals. Immunol Cell Biol. 2014;92(3): 256-262. 18. Lunemann S, Malone DF, Hengst J, et al. Compromised function of natural killer cells in acute and chronic viral hepatitis. J Infect Dis. 2014;209(9):1362-1373. 19. Hall LJ, Murphy CT, Hurley G, et al. Natural killer cells protect against mucosal and systemic infection with the enteric pathogen Citrobacter rodentium. Infect Immun. 2013;81(2):460-469. 20. Han X, Fan Y, Wang S, Jiao L, Qiu H, Yang X. NK cells contribute to intracellular bacterial infection-mediated inhibition of allergic responses. J Immunol. 2008;180(7):46214628. 21. Nogusa S, Ritz BW, Kassim SH, Jennings SR, Gardner EM. Characterization of agerelated changes in natural killer cells during primary influenza infection in mice. Mech Ageing Dev. 2008;129(4):223-230. 22. Alter G, Martin MP, Teigen N, et al. Differential natural killer cell-mediated inhibition of HIV-1 replication based on distinct KIR/HLA subtypes. J Exp Med. 2007;204(12):3027-3036. 23. Harshan KV, Gangadharam PR. In vivo depletion of natural killer cell activity leads to enhanced multiplication of Mycobacterium avium complex in mice. Infect Immun. 1991;59(8):2818-2821. 24. Katz P, Yeager H, Jr., Whalen G, Evans M,

haematologica | 2022; 107(4)

Immune reconstitution with ATLG vs. PTCy

Swartz RP, Roecklein J. Natural killer cellmediated lysis of Mycobacterium-avium complex-infected monocytes. J Clin Immunol. 1990;10(1):71-77. 25. Blanchard DK, Stewart WE 2nd, Klein TW, Friedman H, Djeu JY. Cytolytic activity of human peripheral blood leukocytes against Legionella pneumophila-infected monocytes: characterization of the effector cell and augmentation by interleukin 2. J Immunol. 1987;139(2):551-556. 26. Klimpel GR, Niesel DW, Klimpel KD. Natural cytotoxic effector cell activity against Shigella flexneri-infected HeLa cells. J Immunol. 1986;136(3):1081-1086. 27. Vantourout P, Hayday A. Six-of-the-best: unique contributions of gd T cells to immunology. Nat Rev Immunol. 2013;13 (2):88-100. 28. Kalyan S, Kabelitz D. Defining the nature of human gd T cells: a biographical sketch of the highly empathetic. Cell Mol Immunol. 2013;10(1):21-29. 29. Zheng J, Liu Y, Lau Y-L, Tu W. gd-T cells: an unpolished sword in human anti-infection immunity. Cell Mol Immunol. 2013;10(1): 50-57. 30. Scheper W, van Dorp S, Kersting S, et al. gdT cells elicited by CMV reactivation after allo-SCT cross-recognize CMV and leukemia. Leukemia. 2013;27(6):1328-1338. 31. Knight A, Madrigal AJ, Grace S, et al. The role of Vd2-negative gd T cells during cytomegalovirus reactivation in recipients of allogeneic stem cell transplantation. Blood. 2010;116(12):2164-2172. 32. Elmaagacli AH, Steckel NK, Koldehoff M, et al. Early human cytomegalovirus replication after transplantation is associated with a decreased relapse risk: evidence for a putative virus-versus-leukemia effect in acute myeloid leukemia patients. Blood. 2011;118(5):1402-1412. 33. Perko R, Kang G, Sunkara A, Leung W, Thomas PG, Dallas MH. Gamma delta T

haematologica | 2022; 107(4)

cell reconstitution is associated with fewer infections and improved event-free survival after hematopoietic stem cell transplantation for pediatric leukemia. Biol Blood Marrow Transplant. 2015;21(1):130-136. 34. Lamb Jr L, Gee AP, Hazlett LJ, et al. Influence of T cell depletion method on circulating gd T cell reconstitution and potential role in the graft-versus-leukemia effect. Cytotherapy. 1999;1(1):7-19. 35. Godder K, Henslee-Downey P, Mehta J, et al. Long term disease-free survival in acute leukemia patients recovering with increased gd T cells after partially mismatched related donor bone marrow transplantation. Bone Marrow Transplant. 2007;39(12):751-757. 36. Salio M, Silk JD, Yvonne Jones E, Cerundolo V. Biology of CD1-and MR1restricted T cells. Annu Rev immunol. 2014;32:323-366. 37. Bendelac A, Savage PB, Teyton L. The biology of NKT cells. Annu Rev Immunol. 2007;25:297-336. 38. Kim TW, Park S-S, Lim J-Y, et al. Predictive role of circulating immune cell subtypes early after allogeneic hematopoietic stem cell transplantation in patients with acute leukemia. Int J Stem Cell 2019;12(1):73-83. 39. Servais S, Menten-Dedoyart C, Beguin Y, et al. Impact of pre-transplant anti-T cell globulin (ATG) on immune recovery after myeloablative allogeneic peripheral blood stem cell transplantation. PLoS One. 2015;10(6): e0130026. 40. Bosch M, Dhadda M, Hoegh-Petersen M, et al. Immune reconstitution after anti-thymocyte globulin-conditioned hematopoietic cell transplantation. Cytotherapy. 2012;14(10):1258-1275. 41. Toubert A, Glauzy S, Douay C, Clave E. Thymus and immune reconstitution after allogeneic hematopoietic stem cell transplantation in humans: never say never again. Tissue Antigens. 2012;79(2):83-89.

42. Chawla S, Dharmani-Khan P, Liu Y, et al. High serum level of antithymocyte globulin immediately before graft infusion is associated with a low likelihood of chronic, but not acute, graft-versus-host disease. Biol Blood Marrow Transplant. 2014;20(8): 1156-1162. 43. Remberger M, Sundberg B. Rabbitimmunoglobulin G levels in patients receiving thymoglobulin as part of conditioning before unrelated donor stem cell transplantation. Haematologica. 2005;90(7):931-938. 44. Pearl JP, Parris J, Hale DA, et al. Immunocompetent T-cells with a memorylike phenotype are the dominant cell type following antibody-mediated T-cell depletion. Am J Transplant. 2005;5(3):465-474. 45. Edinger M, Hoffmann P, Ermann J, et al. CD4+ CD25+ regulatory T cells preserve graft-versus-tumor activity while inhibiting graft-versus-host disease after bone marrow transplantation. Nat Med. 2003;9(9): 1144-1150. 46. Nguyen VH, Shashidhar S, Chang DS, et al. The impact of regulatory T cells on T-cell immunity following hematopoietic cell transplantation. Blood. 2008;111(2):945953. 47. Storek J, Ferrara S, Ku N, Giorgi JV, Champlin RE, Saxon A. B cell reconstitution after human bone marrow transplantation: recapitulation of ontogeny? Bone Marrow Transplant. 1993;12(4):387-398. 48. Storek J, Dawson MA, Storer B, et al. Immune reconstitution after allogeneic marrow transplantation compared with blood stem cell transplantation. Blood. 2001;97(11):3380-3389. 49. Small TN, Keever CA, Weiner-Fedus S, Heller G, O'Reilly RJ, Flomenberg N. B-cell differentiation following autologous, conventional, or T-cell depleted bone marrow transplantation: a recapitulation of normal B-cell ontogeny. Blood. 1990;76(8):16471656.

867

ARTICLE Ferrata Storti Foundation

Haematologica 2022 Volume 107(4):868-876

Chronic Lymphocytic Leukemia

The complex karyotype landscape in chronic lymphocytic leukemia allows the refinement of the risk of Richter syndrome transformation Andrea Visentin,1,2 Laura Bonaldi,3 Gian Matteo Rigolin,4 Francesca Romana Mauro,5 Annalisa Martines,3 Federica Frezzato,1,2 Stefano Pravato,1,2 Leila Romano Gargarella,1,2 Maria Antonella Bardi,4 Maurizio Cavallari4, Eleonora Volta,4 Francesco Cavazzini,4 Mauro Nanni,5 Monica Facco,1,2 Francesco Piazza,1,2 Anna Guarini,5 Robin Foà,5 Gianpietro Semenzato,1,2 Antonio Cuneo4 and Livio Trentin1,2 1 Hematology and Clinical Immunology Unit, Department of Medicine, University of Padua, Padua; 2Veneto Institute of Molecular Medicine, Padua; 3Immunology and Molecular Oncology Unit, Veneto Institute of Oncology IOV-IRCSS, Padua; 4Hematology section, Department of Medical Sciences, Azienda Ospedaliera-Universitaria, Arcispedale S. Anna, University of Ferrara, Ferrara; 5Hematology division, Department of Precision and Translational Medicine, "Sapienza" University, Rome, Italy

ABSTRACT

Correspondence: LIVIO TRENTIN livio.trentin@unipd.it ANDREA VISENTIN andrea.visentin@unipd.it Received: January 5, 2021. Accepted: May 21, 2021. Pre-published: June 3, 2021. https://doi.org/10.3324/haematol.2021.278304

868

omplex karyotype (CK) at chronic lymphocytic leukemia (CLL) diagnosis is a negative biomarker of adverse outcome. Since the impact of CK and its subtypes, namely type-2 CK (CK with major structural abnormalities) or high-CK (CK with ≥5 chromosome abnormalities), on the risk of developing Richter syndrome (RS) is unknown, we carried out a multicenter real-life retrospective study to test its prognostic impact. Among 540 CLL patients, 107 harbored a CK at CLL diagnosis, 78 were classified as CK2 and 52 as high-CK. Twenty-eight patients developed RS during a median follow-up of 6.7 years. At the time of CLL diagnosis, CK2 and high-CK were more common and predicted the highest risk of RS transformation, together with advanced Binet stage, unmutated (U)-IGHV, 11q-, and TP53 abnormalities. We integrated these variables into a hierarchical model: high-CK and/or CK2 patients showed a 10-year time to RS (TTRS) of 31%; U-IGHV/11q/TP53 abnormalities/Binet stage B-C patients had a 10-year TTRS of 12%; mutated (M)-IGHV without CK and TP53 disruption a 10-year TTRS of 3% (P<0.0001). We herein demonstrate that CK landscape at CLL diagnosis allows the risk of RS transformation to be refined and we recapitulated clinico-biological variables into a prognostic model.

Introduction Chronic lymphocytic leukemia (CLL), the most common leukemia in western countries, is a remarkably heterogeneous disease, with some patients never requiring treatments and others with a highly aggressive and/or rapidly progressive clinical course.1,2 Richter syndrome (RS) is the transformation of CLL into an aggressive lymphoma, most commonly resembling diffuse large B-cell lymphoma (DLBCL) or Hodgkin lymphoma (HL) variants.3,4 It is characterized by fast growing lymphadenopathies, 18-fluorodeoxyglucose (FDG) positron emission tomography computerized tomography (PET-CT)-avid masses, B symptoms, worsening performance status and increased lactate dehydrogenase levels.5 It is a challenging task to distinguish RS from progressive CLL and it is even more difficult to study prognostic markers since the frequency of RS transformation affects 2%-10% of CLL patients.5 Several studies have proved that chromosome banding analysis is able to refine the prognostic stratification of CLL compared to fluorescence in situ hybridization (FISH) analysis. In fact, 22%-36% of CLL cases with ‘normal’ FISH carry a chromosomal aberration following stimulated karyotypic analyses. In particular, complex karyotype (CK), defined by the presence of at least three chromosome lesions

haematologica | 2022; 107(4)

Complex karyotype subtypes and RS

in the same clone, is detectable in 14%-34% of CLL cases.6-9 The presence of a CK is both a negative prognostic and predictive biomarker associated with an adverse outcome6,10 and worse response to chemoimmunotherapy,7, 11 as well as to novel agents,12,13 regardless of the CLLIPI score or IGHV mutational status.8 However, the CK itself is a heterogeneous quantitative and qualitative cytogenetic category that includes numerical (i.e., monosomies and trisomies) and structural abnormalities (i.e., balanced and unbalanced translocations, marker chromosomes, isochromosomes, deletions, insertions and additions). Recently, collaborative studies have demonstrated that among CK cases assessed at CLL diagnosis, those harboring five or more chromosome abnormalities (high-CK)14 or those with major structural abnormalities, also called type-2 CK (CK2),15,16 identify highly aggressive disease subsets with a poor outcome; the latter is also characterized by a peculiar mRNA expression profile.15,17 Indeed, most of the patients included in these retrospective studies were managed with chemoimmunotherapy.14-16 However, the presence of CK has rarely been associated to the development of RS18 and, to date, it is unknown whether CK subtypes, namely high-CK or CK2, could help to identify patients at a higher risk of developing an RS at CLL diagnosis. In this multicenter retrospective study, we documented, for the first time, that the presence of a CK at CLL diagnosis is associated with an increased risk of developing an RS. In particular, patients with CK2 and high-CK had the highest likelihood of RS transformation. Finally, we recapitulated clinico-biological variables associated with RS into a prognostic model defining three statistically distinct classes of risk of developing RS, the lowest risk for IGHV gene mutated (M-IGHV) being patients without any CK subtypes and an absence of TP53 abnormalities, and the highest risk for patients harboring highCK and/or CK2 subtypes.

Methods Study design Inclusion criteria for this study were diagnosis of CLL according to the 2008 iwCLL guidelines,19 histologically confirmed diagnosis of RS (diffuse large B-cell lymphoma or high-grade B-cell lymphoma), age >18 years and chromosome banding analysis performed within one year of diagnosis in patients without features of disease progression. Data included in the comparative analysis were gender, age, Binet stage,19 CLL treatment prior to RS, 11q2223 deletion by FISH,20 IGHV gene mutational analysis21 and TP53 abnormalities including gene deletions (deletion 17p13) or mutations,22 and b2-microglobulin level >3.5 mg/L. The primary endpoint was the impact of overall CK, CK2 and highCK on the time to Richter syndrome (TTRS) transformation. The correlation of RS with clinical and biologic variables and their impact on TTRS were secondary endpoints. This study was approved by the local research ethics committee and informed consent was obtained from all patients.

Chromosome banding analysis Cytogenetic analysis was performed on peripheral blood after a 72h exposure to 500 mM CpG ODN DSP30 (Roche, Risch, CH) mitogen + 20U/mL IL2 (Roche). Cultures were exposed overnight to 0.1 mg/mL colcemid (Gibco® Karyomax Colcemid, ThermoFisher, Waltham, MA, USA) to obtain metaphases and

haematologica | 2022; 107(4)

then harvested following standard procedures. Karyotype was described after the analysis of at least 20 G-banded metaphases using IKAROS software (MetasYstems, Altlhusseim, Germany), according to international guidelines (ISCN 2016). The definition of a complex karyotype (CK) was defined by the presence of three or more chromosome abnormalities in the same clone.6,8,23,24 According to the literature, CK2 is represented by CK cases with major structural rearrangements that are unbalanced translocations, chromosome additions, insertion, duplications, ring, dicentric and marker chromosomes, whereas complex karyotypes with balanced translocations, deletions, monosomies or trisomies is defined as type-1 (CK1).16 High-CK cases were those presenting at least five chromosome abnormalities.14 Chromosome abnormalities found in only one metaphase were not considered as clonal, and were excluded. Karyotypes were reported by local cytogeneticists (AM, MAB and MN) and reviewed by LB and AV. Detail descriptions of IGHV mutational status,25-29 an assessment of stereotyped B-cell receptor (BCR),30,31 cytogenetics by fluorescence in situ hybridization (FISH),26,32 TP53 gene mutation,22 and NOTCH1 c.7544_7545delCT analysis33 are available in the Online Supplementary Methods.

Statistical analysis Categorical variables were compared by the c2 test (for Binet stages and FISH) or the Fisher exact test (gender, treatment, TP53 and IGHV), when appropriate. Continuous variables (median age) were compared using the Mann-Whitney test. TTRS was calculated starting from the date of CLL diagnosis to RS transformation (event) or last known follow-up (censored).19,34 Overall Survival (OS) was calculated starting from the date of CLL or RS diagnosis, when specified, to death for any cause, or to last known followup. Survival analyses were performed by the Kaplan-Meier method and the Log-rank test was used to compare survival curves between groups. The Cox regression model was employed to estimate hazard ratios (HR). The Cox proportional hazard assumption was assessed based on the scaled Schoenfeld residuals. The stability of our model was internally validated by the bootstrap 0.632 method with B=540. The Harrell concordance index (c-index; 1.0 indicates a perfect discrimination, while a value of 0.5 indicates equivalence to chance) was used to compare our prognostic model.35 The prediction error was calculated as 1 - cindex, corrected for optimism and estimated using the 0.632 bootstrap method.36 Akaike information criterium (AIC) was calculated using the AIC function with R (an open-source statistical package downloadable from http://www.r-project.org).37 A P value >0.05 was considered as not significant.

Results Patients’ characteristics We gathered data from 540 treatment-naive CLL patients with chromosome banding analysis assessed within 12 months from diagnosis at three Italian centers (Table 1). The median age at diagnosis of the whole case series was 63±12 years, 61% were male, 75% showed Binet A stage, the median b2-microglobulin was 2.93 mg/L, 57% of patients were U-IGHV, 11% harbored TP53 abnormalities (8% 17p13 deletion and 3% only TP53 mutation) and 20% a CK (Figure S1A). NOTCH1 mutation was assessed in 47 patients at CLL diagnosis and it was found in two subjects who further developed RS. Two hundred and fifty-two patients subsequently received at least one line of therapy - 31% FCR (fludarabine, cyclophosphamide, rituximab), 17% BR (bendamus869

A. Visentin et al.

Table 1. Clinical and biological features of patients.

Age (years) median±sd Age at RS (years) median±sd Gender Female Male Binet stage A B-C b2-microglobulin (mg/L)* median±sd IGHV status M-IGHV U-IGHV FISH 13q or Normal +12 11qTP53 abn Normal Disrupted KARYOTYPE no CK CK QUALITATIVE no CK CK1 CK2 QUANTITATIVE 0 1-2 3-4 ≥5 RS SCORE low-risk Int.-risk high-risk

Population (n=540)

RS (n= 28)

no RS (n=512)

63±12

63±9.8

63±12

0.8793

68±12

n.a.

211 (39%) 329 (61%)

11 (39%) 17 (61%)

200 (39%) 312 (61%)

>0.9999

407 (75%) 133 (25%)

15 (54%) 13 (46%)

392 (77%) 120 (23%)

0.0113

2.9±1.5

3.2±0.98

2.9±1.6

0.1216

232 (43%) 309 (57%)

6 (21%) 22 (79%)

225 (44%) 287 (56%)

0.019

404 (75%) 84 (15%) 52 (10%)

19 (68%) 3 (11%) 6 (21%)

385 (75%) 81 (16%) 46 (9%)

0.0861 0.0415+

482 (89%) 58 (11%)

19 (68%) 9 (32%)

463 (90%) 49 (10%)

0.0043

433 (80%) 107 (20%)

14 (50%) 14 (50%)

419 (82%) 93 (18%)

0.0002

433 (80%) 29 (5%) 78 (14%)

14 (50%) 1 (4%) 13 (46%)

419 (82%) 28 (5%) 65 (13%)

<0.0001

165 (30%) 269 (50%) 54 (10%) 52 (10%)

3 (11%) 11 (39%) 3 (11%) 11 (39%)

162 (32%) 258 (50%) 52 (10%) 41 (8%)

<0.0001

212 (39%) 247 (46%) 81 (15%)

3 (11%) 12 (43%) 13 (46%)

209 (41%) 235 (46%) 68 (13%)

<0.0001

RS: Richter syndrome; sd: standard deviation; M-IGHV: mutated IGHV gene; U-IGHV: unmutated IGHV gene; 11q-: del11q22-23 by interphase FISH analysis; TP53 abn: TP53 abnormalities include deletions and/or mutations; CK: complex karyotype; CK1: type-1 CK; CK2: type-2 CK; highCK: ≥5 chromosome abnormalities; high-risk: CK2 and/or highCK; int.risk: U-IGHV/11q-/TP53abn/Binet B-C; low-risk: M-IGHV without CK and TP53 wild type; n.a.: not applicable; * data available from 520 (96%) patients, 26 (93%) who developed an RS and 494 (96%) who did not transform; + Analysis between subgroups with 11q- and others.

tine, rituximab), 10% ibrutinib, 5% chlorambucil plus an anti-CD20 monoclonal antibody, 2% venetoclax, 35% other treatments such as FC or chlorambucil single agent as first line therapy - and 90 died during the follow-up. According to the qualitative CK subtype, 29 of 107 (27%) patients displayed a CK1 and 78 (73%) a CK2 (Online Supplementary Figure S1A, Table 1) whereas, according to the number of chromosome lesions, 165 (30%) patients had a normal karyotype (i.e., 46,(XX) or 46,(XY) for females and males, respectively), 268 (50%) had one or two lesions, and 54 (10%) three or four abnor870

malities and 52 (10%) were classified as high-CK (i.e., ≥5 chromosome lesions) (Online Supplementary Figure S1A and S2A, Table 1). In particular, a high-CK was more common in CK2 than in CK1 patients, being present in 63% of patients harboring a CK2 subtype but in only 10% of CK1 patients (P<0.0001, Online Supplementary Figure S2A). As a preliminary step for our further analysis, we confirmed the established prognostic role of overall CK, CK with major unbalanced abnormalities (i.e., CK2) and highCK in our dataset (Figure S2B-D). The 10-year OS was 54% and 79% for CK and no-CK patients, respectivehaematologica | 2022; 107(4)

Complex karyotype subtypes and RS

Table 2. Hazard ratios (HR) for the time to Richter syndrome.

TTRS ≥65 years Age+ Male b2MG high* Binet B-C U-IGHV +12 11qTP53 abn CK CK2 High-CK RS MODEL Low-risk Int.-risk High-risk

Univariate analysis 95% C.I.

Multivariate analysis 95% C.I.

1.4 1.02 1.0 1.8 3.9 4.0 0.8 4.6 9.5 7.4 8.8 9.9

0.6-2.9 0.9-1.1 0.5-2.1 0.6-5.6 1.6-9.6 1.9-8.6 0.3-2.4 1.3-16.7 2.9-31.4 3.0-18.3 4.9-19.8 6.5-22.9

0.4289 0.2120 0.9379 0.2925 0.0024 0.0004 0.6675 0.0215 0.0002 <0.0001 <0.0001 <0.0001

2.9 4.5 2.8 3.9 4.7 5.6 6.9

1.4-6.3 1.8-11.3 1.1-6.9 1.8-8.7 2.2-9.9 2.7-11.8 3.3-14.9

0.0039 0.0011 0.0285 0.0008 <0.0001 <0.0001 <0.0001

1.0 4.0 13.6

1.4-11.4 7.1-20.9

0.0101 <0.0001

1.00 3.4 9.2

1.5-7.5 4.7-17.3

0.0023 <0.0001

TTRS: time to Richter syndrome; b2MG high: beta2-microglobulin >3.5mg/L; U-IGHV: unmutated IGHV gene; TP53 abn: TP53 abnormalities include deletions and/or mutations; CK: complex karyotype; CK1: type-1 CK; CK2: type-2 CK; highCK: ≥5 chromosome abnormalities; High-risk: CK2 and/or highCK; Int.-risk: U-IGHV/11q-/TP53abn/Binet B-C; Lowrisk: M-IGHV without CK and TP53 wild type; n.a.: not applicable; + : age considered as continuous variable; * : data available from 520 (96%) patients, 26 (93%) who developed an RS and 494 (96%) who did not transform.

ly (P<0.0001, Online Supplementary Figure S2B); 48% vs. 72% vs. 79% for CK2, CK1 and no-CK (P<0.0001, Online Supplementary Figure S2C), respectively; 44% vs. 64% vs. 70% vs. 90% for patients with ≥5 (i.e., high-CK), 4-3, 2-1 and without chromosome abnormalities (P<0.0001, Online Supplementary Figure S2D), respectively.

Clinico-biological features of patients who developed a RS transformation Twenty-eight (5.2%) patients developed a histologically confirmed RS over a median follow-up of 6.7 years (Figure S1B). The median age at RS diagnosis was 68 years (range 38-84), 61% were male, 75% had received a CLL treatment in the past, 79% were U-IGHV, 32% presented TP53 abnormalities, 50% harbored a CK at CLL diagnosis, which included 46% and 39% of CK2 and high-CK subtypes, respectively. Eight cases showed deletion of 9p21.3, i.e., the locus of CDKN2A gene, all with a CK. In particular, 8 of 8 were classified as CK2 and 6 of 8 as high-CK subtype. Only one patient, who developed RS, received ibrutinib frontline. We also observed that more patients who developed an RS displayed an advanced Binet stage at CLL diagnosis (P=0.0113) and were enriched in U-IGHV (P=0.0191), TP53 abnormalities (P=0.0043), CK overall (P=0.0002), CK2 (P<0.0001) and high-CK (P<0.0001) cases as compared to patients who did not develop an RS (Table 1, Figure S1C). Age at CLL diagnosis (median age 63.5 and 63.3 years), gender distribution (both 61%), trisomy of chromosome 12 (11% and 16%), b2-microglobulin (median levels 3.2mg/L and 2.9mg/L) and stereotyped BCR (10.7% vs. 9.8%) had a superimposable distribution among patients with and without an RS transformation (Table 1).

Prognosticators of Richter Syndrome The cumulative incidence of RS slowly increases over haematologica | 2022; 107(4)

time. As shown in Figure 1A, 2.6%, 12% and 13% of patients developed an RS within five, ten and 15 years after CLL diagnosis, respectively. We observed that patients with a CK, overall (Figure 1B) and its subtypes (Figure 1C-D), had a very high risk of developing an RS. The estimated ten-year TTRS to be 25% vs. 8% (P<0.0001), 38% vs. 8% (P<0.0001) and 41% vs. 8% (P<0.0001) for patients with CK vs. no-CK, CK2 vs. other patients (i.e., CK1 or noCK), highCK vs. other patients (i.e., 3-4 or 1-2 or, 0 chromosome abnormalities) respectively (Figure 1C-D and S3G-H). Multivariate analysis revealed that CK overall was associated with a more than four-fold higher risk of developing an RS (HR 4.7, 95% CI 2.2-9.9, P<0.0001). This risk was even higher for CK subtypes, being more than five-fold (HR 5.6, 95% CI 2.7-11.8, P<0.0001) and seven-fold (HR 6.9, 95% CI 3.3-14.9, P<0.0001) higher for patients harboring CK2 and high-CK subtypes, respectively (Table 2). Other variables associated with TTRS at univariate and multivariate analysis were Binet stage B-C, U-IGHV, 11q-, TP53 abnormalities (Table 2, Online Supplementary Figure S3A-E). Among CK2 and/or high-CK patients (n=81), 32 (39%) patients carried TP53 abnormalities and 52 (63%) an U-IGHV status. We found that TP53 abnormalities and IGHV status mildly impact on the risk of developing RS among CK2 and/or high-CK subgroup, but the difference was not statistically significant (Figure S4A-B, P=0.1150 and P=0.1405, respectively). These data suggest that CK subtypes per se represent a stronger prognosticator of RS transformation than conventional biologic markers such as TP53 disruption and U-IGHV conformation. The median OS from CLL diagnosis for the whole population was not reached and the estimated ten-year OS was 73% (Figure S5A). Patients who developed an RS had a shorter OS (Figure 2A). The median OS was seven years vs. not reached and the estimated ten-year OS was 16% 871

A. Visentin et al.

Figure 1. Kaplan Meyer curves of time to Richter syndrome. The upper-left (A) panel shows the time to Richter syndrome (RS) transformation for the whole population. Patients with a CK overall (B), CK2 (C) or high-CK (D) have a significantly increased a risk of developing an RS compared to the other patients (Log-rank test, P<0.0001).

vs. 79% for patients who developed an RS vs. those who did not transform (Figure 2E, P<0.0001), respectively. Variables that were associated with a higher risk of death in multivariate analysis are summarized in Table S1. The median time from CLL diagnosis to RS transformation was 5.3 years. ranging from 0.10 years to 10.8 years. In only one patient was RS was diagnosed within six months of CLL diagnosis. The median OS from RS transformation was 5.3 months and the two-year OS was only 20% (Figure 2B). The OS from RS was not affected by the presence of a CK at CLL diagnosis nor its subtypes (Figure S5B-C). The two-year OS from RS was 28% vs. 10% for CK2 cases and other patients (i.e., CK1 and no-CK) (P=0.3317) and 24% vs 16% for high-CK cases and other patients, respectively (i.e., <5 chromosome abnormalities) (P=0.9864) (Figure S5C-D). No traditional prognostic markers could foresee the risk of death after RS diagnosis in our population (Table S2).

A Richter syndrome prognostic model By integrating CK subtypes, TP53 abnormalities, 11q deletion, IGHV mutational status and Binet stages based on HR values, we developed a hierarchical model leading 872

to the identification of three statistically different groups. These were ranked from the shortest to the longest TTRS, as follows: 81 (15%) patients were classified as high-CK and/or CK2, and had a five-year and ten-year TTRS of 13% and 31%; 247 (46%) patients displayed a U-IGHV status or 11q- or TP53 disruption or Binet stage B-C, and showed the five-year and ten-year TTRS of 0.9% and 12%; 212 (39%) patients were M-IGHV without CK and TP53 abnormalities, and had a five-year and ten-year TTRS of 0.7% and 3% (Figure 3, P<0.0001). Multivariate analysis confirmed that the former subgroup (i.e., high-CK and/or CK2) was associated with the highest risk of RS transformation (HR 9.2, 95% CI 3.8-46, P<0.0001) compared to the low-risk group, one which is characterized by the presence of M-IGHV without CK and TP53 abnormalities. Patients with U-IGHV or 11q- or TP53 abnormalities or Binet stage B-C had an intermediate risk, with a threefold higher risk of RS compared to low-risk patients (HR 3.4, 95% CI 1.5-7.5, P=0.0023) (Table 2). Our model was also internally validated using the bootstrap 0.632 method showing a prediction error of 0.26. Finally, the c-index for our proposed model was 0.81 for TTRS and the Akaike information criterium was 286. These results indicate that haematologica | 2022; 107(4)

Complex karyotype subtypes and RS

Figure 2. Kaplan Meyer curves of overall survival. The left panel (E) shows the overall survival analysis for patients with RS transformation and those who did not develop an RS (no RS). Patients who developed an RS had a shorter overall survival, calculated from CLL diagnosis (Log-rank test, P<0.0001). The right (F) panel shows the overall survival after RS transformation, confirming these patients’ very poor prognosis.

our model had a good prediction accuracy for the risk of developing an RS; higher than of the CLL-IPI38 (c-index 0.69, prediction error 0.28, Akaike information criterium 301) and the Barcelona-Brno39 (c-index 0.74, prediction error 0.25, Akaike information criterium 292) scores accuracies applied to our population (Figure S4C-D). Based on the lower Akaike score, our RS prognostic model better predicts the risk of developing an RS than the available comparators.

Discussion In this multicenter retrospective study, we demonstrated that patients harboring a CK at CLL diagnosis, in particular those with CK2 and/or high-CK, are characterized by the highest risk of developing an RS transformation. Subsequently, by integrating data of CK subtypes with other clinical and biologic variables associated with the risk of RS, we were able to define an RS prognostic model. To minimize selection and attrition biases, as well as imprecise reporting of data inherent to observational studies, we asked the clinicians to report all patients who performed stimulated cytogenetic analysis within the first year of diagnosis. We analyzed the reported data and performed computerized and manual consistency checks on each case report form. RS is a rare and an aggressive complication of CLL patients, affecting between 2% and 10% of CLLs.34 Most RS patients are elderly, have a poor performance status and suffer from several comorbidities which limit the use of intensive chemoimmunotherapy.40 Since the majority of patients are primary refractory to first-line treatment and only a few are able to undergo allogenic stem cell transplantation procedures, the reported estimated survival after a diagnosis of RS is usually less than one year, even with the introduction of targeted-therapy41,42 and immune checkpoint inhibitors.43 For these reasons, the standard of care of patients with RS remains a primary unmet need. Known biologic risk factors for the development of RS are TP53 and CDKN2A aberrations, NOTCH1 haematologica | 2022; 107(4)

mutation and a stereotype BCR subset #8.44,45 To date, the impact of CK at CLL diagnosis on the risk of developing RS has been investigated in only a few studies.46,47 The German CLL study group has recently reviewed the clinical features of RS patients as part of their clinical trials.34 In this study, 3.5% of CLL developed an RS transformation after a median observation time of 4.4 years. The median age at RS was 65 years and the median OS after RS was 9.4 months, which was significantly longer for HL compared to the DLBCL variant (median OS 83 months vs. 8.7 months, respectively). Adverse risk factors at trial enrollment, such as 17p13 deletion by FISH, high b2-microglobulin and CLL-IPI scores were more common in patients who developed an RS34 while NOTCH1 mutations and stereotype #8 were not recurrent in RS cases.34 Conversely, among the 204 RS from the Mayo clinic, the median OS after RS diagnosis was 12 months.48 In a multivariate Cox regression analysis, prior CLL treatment and older age, but not TP53 disruption, were associated with a shorter OS.48 The results of our real-life study are in line with the GCLLSG and Mayo clinic reports, even though our patients were slightly older; this could explain the shorter OS in our RS cohort (median survival after RS is 5.3 months). Comparable survival rates, between six and 12 months, have been observed in other retrospective analyses.34,44,48 In addition, advanced Binet stage, U-IGHV and 11q- were also significantly associated with an RS risk in our patients. Chromosome banding analysis in CLL is capable of identifying chromosomal abnormalities that are missed by FISH analysis, sometimes fulfilling CK criteria.6,24,49,50 Genomic microarrays have also emerged as a valuable tool for genome-wide studies in CLL. However, in a recent study, no significant differences emerged in patients’ classification, time to first treatment, OS and prediction accuracy between chromosome banding analysis and genomic microarrays.51 The prognostic and predictive role of CK, defined by the presence of at least three chromosomal lesions, is evident at diagnosis,6,8 as well as at disease progression7 and in relapsed/refractory patients treated with ibrutinib13,52 or venetoclax.12 Of note, CK was not a prog873

A. Visentin et al.

Figure 3. The Richter syndrome scoring system. Kaplan-Meier curve of time to Richter syndrome transformation according to the Richter syndrome scoring system. Patients were classified at high-risk if they were high-CK and/or CK2 at CLL diagnosis (blue curve); at intermediate-risk if they displayed unmutated IGHV status (UIGHV), 11q22-23 deletion (11q), TP53 abnormalities (including deletions or mutations, TP53 abn) or Binet stage B-C (grey curve); at low-risk if they were IGHV mutated (M-IGHV) patients without CK and wild-type TP53 gene (TP53 not deleted non mutated) (orange curve).

nostic marker of survival on multivariate analysis for patients treated frontline with ibrutinib±rituximab53 while treatment with idelalisib plus rituximab seems to have a comparable efficacy in R/R patients with and without CK.54-56 CK has been found in 14%-35% of CLL depending on the study in question,6,10 and identifies a heterogeneous cytogenetic category in terms of quantitative and qualitative characteristics. Data from the literature has documented that the presence of at least five chromosomal aberrations is associated with a very aggressive clinical course independent of the IGHV status and TP53 lesions.14 Our collaborative group has previously demonstrated that almost 70% of CK cases harbor major structural aberrations such as unbalanced translocations and ring or marker chromosomes.15 This subset, called CK2, was associated with a higher incidence of TP53 aberrations, chemo-refractoriness, early relapse after chemoimmunotherapy, and a shorter OS at multivariate analysis.15 In addition, the prognostic and predictive accuracy of CK subtypes is enhanced when it is combined with IGHV mutational status.16 Interestingly, a recent analysis of the international CLL14 clinical trial suggests that the fixedduration combination of obinutuzumab plus venetoclax seems to overcome the negative predictive impact of CK, both in terms of undetectable minimal residual disease rates and progression-free survival.57 The presence of CK has been sporadically linked to the development of RS.18 In a retrospective study on CLL patients treated with FCR, one of four cases with RS had a CK.9 Anderson et al.12 found a CK in 48% of the 25 patients who progressed on venetoclax, including eight of 17 patients with RS. Rogers et al.,47 reported a CK in 67% patients who developed an RS and found that a CK had an 874

adverse impact on the R-EPOCH regimen. A recent study from Ohio State University found that six of nine patients with a near-tetraploidy (four copies of most chromosomes) karyotype developed an RS.46 At multivariate analysis, near-tetraploidy and CK predicted ibrutinib discontinuation due to transformation.46 Although the exact mechanism that favors the development of an RS in patients with CK is unknown, the strong association between CK and TP53 abnormalities, short telomere length and, consequently, increased chromosome instability could play a relevant role.18,58 Thanks to stimulated chromosome banding analysis, we were able to identify a CK in 20% of 540 CLL patients and could demonstrate that patients harboring a CK2 or a high-CK had a six and seven-fold increased risk of developing RS. We therefore suggest that the integration of CK subtypes, together with IGHV mutational status, TP53 abnormalities, 11q22-23 deletion and Binet stage, may allow the prognostic risk of RS transformation (Figure 3) to be refined. Indeed, we could show that M-IGHV patients without any CK subtypes and a wild-type TP53 gene are characterized by a very low risk of developing RS, at only 0.7% five years from CLL diagnosis. On the other hand, patients with CK subtypes, both CK2 and/or high-CK, are characterized by the highest risk of developing RS, with 31% of them experiencing a disease transformation within ten years of diagnosis. In addition, our model seems to better predict the risk of RS transformation than the available scoring systems. Our results, like most data found in the literature, derived from a cohort of patients treated mainly with chemoimmunotherapy, a choice made partly due to their longer follow-up. Although the cumulative incidence of RS among patients haematologica | 2022; 107(4)

Complex karyotype subtypes and RS

treated with chemo/chemoimmunotherapy seems to be higher than patients treated with BTK or BCL2 inhibitors, this difference was not statistically significant (P=0.3337, Online Supplementary Figure S5E). In addition, after validation by an independent cohort of patients treated frontline with targeted drugs and in a prospective study, our prognostic model might be used in the follow-up management of patients with CLL. In particular, patients with a CK2 and/or high-CK should be carefully monitored for the development of an RS during their follow-up. Disclosures AV received honoraria from Janssen, Abbvie, Italfarmaco. LT received research funding by Gilead, Roche, Janssen and Takeda, advisory board for Roche, Shire and Abbvie, Astrazeneca. GMR received research funding by Gilead. FRM advisory board for Janssen, Shire and Abbvie. AC advisory board and speaker bureau for Roche, Abbvie, Gilead and Janssen. GS board member of Abbvie, Roche, Janssen and Celgene. RF advisory board or speaker bureau for Roche, Abbvie, Celgene, Incyte, Amgen, Janssen, Gilead and Novartis. Contributions AV designed the study, performed statistical analysis, visited patients and wrote the article; SP, LRG, MC, EV and FC and provided intellectual inputs and visited patients; LB, AM, MAB

References 1. Scarfo L, Ferreri AJ, Ghia P. Chronic lymphocytic leukaemia. Crit Rev Oncol Hematol. 2016;104:169-182. 2. Visentin A, Facco M, Frezzato F, et al. Integrated CLL scoring system, a new and simple index to predict time to treatment and overall survival in patients with chronic lymphocytic leukemia. Clin Lymphoma Myeloma Leuk. 2015;15(10):612-620 3. Mauro FR, Galieni P, Tedeschi A, et al. Factors predicting survival in chronic lymphocytic leukemia patients developing Richter syndrome transformation into Hodgkin lymphoma. Am J Hematol. 2017; 92(6):529-535. 4. Visentin A, Imbergamo S, Gurrieri C, et al. Major infections, secondary cancers and autoimmune diseases occur in different clinical subsets of chronic lymphocytic leukaemia patients. Eur J Cancer. 2017; 72:103-111. 5. Vitale C, Ferrajoli A. Richter syndrome in chronic lymphocytic leukemia. Curr hematol Malig Rep. 2016;11(1):43-51. 6. Baliakas P, Iskas M, Gardiner A, et al. Chromosomal translocations and karyotype complexity in chronic lymphocytic leukemia: a systematic reappraisal of classic cytogenetic data. Am J Hematol. 2014; 89(3):249-255. 7. Herling CD, Klaumunzer M, Rocha CK, et al. Complex karyotypes and KRAS and POT1 mutations impact outcome in CLL after chlorambucil-based chemotherapy or chemoimmunotherapy. Blood. 2016; 128(3):395-404. 8. Rigolin GM, Cavallari M, Quaglia FM, et al. In CLL, comorbidities and the complex karyotype are associated with an inferior outcome independently of CLL-IPI. Blood. 2017;129(26):3495-2498. 9. Le Bris Y, Struski S, Guieze R, et al. Major prognostic value of complex karyotype in

haematologica | 2022; 107(4)

and MN performed cytogenetic tests; FF, MF and AG performed cytofluorimetric and IGHV analysis; FRM, GMR, PF, GS, RF, AC and LT visited patients, provided intellectual inputs and reviewed the article. Funding This work was supported by funds from Associazione Italiana per la Ricerca sul Cancro (A.I.R.C.) projects to LT (IG-25024), Gilead fellowship program 2018 to LT, Special Program ‘Metastatic disease: the key unmet need in oncology’, AIRC 5x1000 (No. 21198) to RF, Fondo di Ateneo per la Ricerca 2016, 2017 of the University of Ferrara to GMR and FC, Fondo di Incentivazione alla Ricerca 2017 of the University of Ferrara to GMR, Ministero dell’Istruzione, dell’Università e della Ricerca PRIN 2015 to AC (2015ZMRFEA). AV received a research fellowship from the University of Padua supported by ONLUS Ricerca per Credere nella Vita (RCV) odv, Padua, Italy. This study was approved by the local research ethics committee of Padua hospital and informed consent was obtained from all patients. Data sharing statement The datasets generated and analyzed during the current study are not publicly available due to the data protection and lack of consent from the patients. Access to data is strictly limited to the researchers who have obtained permission for data processing.

addition to TP53 and IGHV mutational status in first-line chronic lymphocytic leukemia. Hematol Oncol. 2017;35(4):664670. 10. Rigolin GM, del Giudice I, Formigaro L, et al. Chromosome aberrations detected by conventional karyotyping using novel mitogens in chronic lymphocytic leukemia: Clinical and biologic correlations. Genes Chromosomes Cancer. 2015;54(12):818826. 11. Badoux XC, Keating MJ, Wang X, et al. Fludarabine, cyclophosphamide, and rituximab chemoimmunotherapy is highly effective treatment for relapsed patients with CLL. Blood. 2011;117(11):3016-3024. 12. Anderson MA, Tam C, Lew TE, et al. Clinicopathological features and outcomes of progression of CLL on the BCL2 inhibitor venetoclax. Blood. 2017; 129(25):3362-3370. 13. Thompson PA, O'Brien SM, Wierda WG, et al. Complex karyotype is a stronger predictor than del(17p) for an inferior outcome in relapsed or refractory chronic lymphocytic leukemia patients treated with ibrutinibbased regimens. Cancer. 2015; 121(20): 3612-3621. 14. Baliakas P, Jeromin S, Iskas M, F, et al. Cytogenetic complexity in chronic lymphocytic leukemia: definitions, associations, and clinical impact. Blood. 2019; 133(11):1205-1216. 15. Rigolin GM, Saccenti E, Guardalben E, et al. In chronic lymphocytic leukaemia with complex karyotype, major structural abnormalities identify a subset of patients with inferior outcome and distinct biological characteristics. Br J Haematol. 2018; 181(2):229-233. 16. Visentin A, Bonaldi L, Rigolin GM, et al. The combination of complex karyotype subtypes and IGHV mutational status identifies new prognostic and predictive groups in chronic lymphocytic leukaemia. Br J Cancer. 2019;121(2):150-156.

17. Rigolin GM, Saccenti E, Melandri A, et al. In chronic lymphocytic leukaemia, SLAMF1 deregulation is associated with genomic complexity and independently predicts a worse outcome. Br J Haematol. 2021;192(6):1068-1072. 18. Cavallari M, Cavazzini F, Bardi A, et al. Biological significance and prognostic/predictive impact of complex karyotype in chronic lymphocytic leukemia. Oncotarget. 2018;9(76):34398-34412. 19. Hallek M, Cheson BD, Catovsky D, et al. Guidelines for the diagnosis and treatment of chronic lymphocytic leukemia: a report from the International Workshop on Chronic Lymphocytic Leukemia updating the National Cancer Institute-Working Group 1996 guidelines. Blood. 2008; 111(12):5446-5456. 20. Hallek M. Chronic lymphocytic leukemia: 2015 Update on diagnosis, risk stratification, and treatment. Am J Hematol. 2015;90(5):446-460. 21. Langerak AW, Davi F, Ghia P, et al. Immunoglobulin sequence analysis and prognostication in CLL: guidelines from the ERIC review board for reliable interpretation of problematic cases. Leukemia. 2011; 25(6):979-984. 22. Malcikova J, Tausch E, Rossi D, et al. ERIC recommendations for TP53 mutation analysis in chronic lymphocytic leukemiaupdate on methodological approaches and results interpretation. Leukemia. 2018; 32(5):1070-1080. 23. Blanco G, Puiggros A, Baliakas P, et al. Karyotypic complexity rather than chromosome 8 abnormalities aggravates the outcome of chronic lymphocytic leukemia patients with TP53 aberrations. Oncotarget. 2016;7(49):80916-80924. 24. Kreinitz N, Polliack A, Tadmor T. Chronic lymphocytic leukemia is becoming more complex: how to define complex karyotype? Leuk Lymphoma. 2018;59(3):521522.

875

A. Visentin et al. 25. Terrin L, Trentin L, Degan M, et al. Telomerase expression in B-cell chronic lymphocytic leukemia predicts survival and delineates subgroups of patients with the same igVH mutation status and different outcome. Leukemia. 2007;21(5):965-972. 26. Raponi S, Del Giudice I, Marinelli M, et al. Genetic landscape of ultra-stable chronic lymphocytic leukemia patients. Ann Oncol. 2018;29(4):966-972. 27. Brochet X, Lefranc MP, Giudicelli V. IMGT/V-QUEST: the highly customized and integrated system for IG and TR standardized V-J and V-D-J sequence analysis. Nucl Acids Res. 2008;36(Web Server issue):W503-508. 28. Hamblin TJ, Davis Z, Gardiner A, Oscier DG, Stevenson FK. Unmutated Ig V(H) genes are associated with a more aggressive form of chronic lymphocytic leukemia. Blood. 1999;94(6):1848-1854. 29. Visentin A, Facco M, Gurrieri C, et al. Prognostic and predictive effect of IGHV mutational status and load in chronic lymphocytic leukemia: focus on FCR and BR treatments. Clin Lymphoma Myeloma Leuk. 2019;19(10):678-685. 30. Bystry V, Agathangelidis A, Bikos V, Set al. ARResT/assignsubsets: a novel application for robust subclassification of chronic lymphocytic leukemia based on B cell receptor IG stereotypy. Bioinformatics. 2015; 31(23): 3844-3846. 31. Agathangelidis A, Chatzidimitriou A, Gemenetzi K, et al. Higher-order connections between stereotyped subsets: implications for improved patient classification in CLL. Blood. 2021;137(10):1365-1376. 32. Dohner H, Stilgenbauer S, Benner A, et al. Genomic aberrations and survival in chronic lymphocytic leukemia. N Engl J Med. 2000;343(26):1910-1916. 33. Rossi D, Rasi S, Fabbri G, et al. Mutations of NOTCH1 are an independent predictor of survival in chronic lymphocytic leukemia. Blood. 2012;119(2):521-529. 34. Al-Sawaf O, Robrecht S, Bahlo J, et al. Richter transformation in chronic lymphocytic leukemia (CLL)-a pooled analysis of German CLL Study Group (GCLLSG) front line treatment trials. Leukemia. 2021; 35(1): 169-176. 35. Harrell FE Jr, Lee KL, Mark DB. Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med. 1996;15(4):361-387. 36. Iba K, Shinozaki T, Maruo K, Noma H. Reevaluation of the comparative effectiveness

876

of bootstrap-based optimism correction methods in the development of multivariable clinical prediction models. BMC Med Res Methodol. 2021;21(1):9. 37. Cohen JA, Rossi FM, Zucchetto A, et al. A laboratory-based scoring system predicts early treatment in Rai 0 chronic lymphocytic leukemia. Haematologica. 2020; 105(6):1613-1620. 38. International CLLIPIwg. An international prognostic index for patients with chronic lymphocytic leukaemia (CLL-IPI): a metaanalysis of individual patient data. Lancet Oncol. 2016;17(6):779-790. 39. Delgado J, Doubek M, Baumann T, et al. Chronic lymphocytic leukemia: a prognostic model comprising only two biomarkers (IGHV mutational status and FISH cytogenetics) separates patients with different outcome and simplifies the CLL-IPI. Am J Hematol. 2017;92(4):375-380. 40. Condoluci A, Rossi D. Richter syndrome. Curr Oncol Rep. 2021;23(3):26. 41. Ayers EC, Mato AR. Richter's Transformation in the era of kinase inhibitor therapy: a review. Clin Lymphoma Myeloma Leuk. 2017;17(1):1-6. 42. Visentin A, Imbergamo S, Scomazzon E, et al. BCR kinase inhibitors, idelalisib and ibrutinib, are active and effective in Richter syndrome. Br J Haematol. 2019;185(1):193197. 43. Ding W, LaPlant BR, Call TG, et al. Pembrolizumab in patients with CLL and Richter transformation or with relapsed CLL. Blood. 2017;129(26):3419-3427. 44. Rossi D, Spina V, Gaidano G. Biology and treatment of Richter syndrome. Blood. 2018;131(25):2761-2772. 45. Fabbri G, Khiabanian H, Holmes AB, et al. Genetic lesions associated with chronic lymphocytic leukemia transformation to Richter syndrome. J Exp Med. 2013; 210(11):2273-2288. 46. Miller CR, Ruppert AS, Heerema NA, et al. Near-tetraploidy is associated with Richter transformation in chronic lymphocytic leukemia patients receiving ibrutinib. Blood Adv. 2017;1(19):1584-1588. 47. Rogers KA, Huang Y, Ruppert AS, et al. A single-institution retrospective cohort study of first-line R-EPOCH chemoimmunotherapy for Richter syndrome demonstrating complex chronic lymphocytic leukaemia karyotype as an adverse prognostic factor. Br J Haematol. 2018; 180(2):259-266. 48. Wang Y, Tschautscher MA, Rabe KG, et al. Clinical characteristics and outcomes of

Richter transformation: experience of 204 patients from a single center. Haematologica. 2020;105(3):765-773. 49. Baliakas P, Puiggros A, Xochelli A, et al. Additional trisomies amongst patients with chronic lymphocytic leukemia carrying trisomy 12: the accompanying chromosome makes a difference. Haematologica. 2016; 101(7):e299-302. 50. Puiggros A, Collado R, Calasanz MJ, et al. Patients with chronic lymphocytic leukemia and complex karyotype show an adverse outcome even in absence of TP53/ATM FISH deletions. Oncotarget. 2017;8(33):54297-54303. 51. Ramos-Campoy S, Puiggros A, Bea S, et al. Chromosome banding analysis and genomic microarrays are both useful but not equivalent methods for genomic complexity risk stratification in chronic lymphocytic leukemia patients. Haematologica. 2022; 107(3):593-603. 52. Morabito F, Del Poeta G, Mauro FR, et al. TP53 disruption as a risk factor in the era of targeted therapies: a multicenter retrospective study of 525 chronic lymphocytic leukemia cases. Am J Hematol. 2021;96(8): E306-E310. 53. Woyach JA, Ruppert AS, Heerema NA, et al. Ibrutinib regimens versus chemoimmunotherapy in older patients with untreated CLL. N Engl J Med. 2018;379(26):25172528. 54. Kreuzer KA, Furman RR, Stilgenbauer S, et al. The impact of complex karyotype on the overall survival of patients with relapsed chronic lymphocytic leukemia treated with idelalisib plus rituximab. Leukemia. 2020;34(1):296-300. 55. Rigolin GM, Cavazzini F, Piciocchi A, et al. Efficacy of idelalisib and rituximab in relapsed/refractory chronic lymphocytic leukemia treated outside of clinical trials. a report of the gimema working group. Hematol Oncol. 2021;39(3):326-335. 56. Visentin A, Frezzato F, Severin F, et al. Lights and shade of next-generation Pi3k inhibitors in chronic lymphocytic leukemia. Onco Targets Ther. 2020; 13:9679-9688. 57. Al-Sawaf O, Lilienweiss E, Bahlo J, et al. High efficacy of venetoclax plus obinutuzumab in patients with complex karyotype and chronic lymphocytic leukemia. Blood. 2020;135(11):866-870. 58. Jebaraj BMC, Tausch E, Landau DA, et al. Short telomeres are associated with inferior outcome, genomic complexity, and clonal evolution in chronic lymphocytic leukemia. Leukemia. 2019;33(9):2183-2194.

haematologica | 2022; 107(4)

ARTICLE

Chronic Lymphocytic Leukemia

IGHV-associated methylation signatures more accurately predict clinical outcomes of chronic lymphocytic leukemia patients than IGHV mutation load Dianna Hussmann,1 Anna Starnawska,1,2,3,4 Louise Kristensen,5 Iben Daugaard,1,6 Astrid Thomsen,1 Tina E. Kjeldsen,1 Christine Søholm Hansen,2,7 Jonas Bybjerg-Grauholm,2,7 Karina Dalsgaard Johansen,8 Maja Ludvigsen,8,9 Thomas Kristensen,5 Thomas Stauffer Larsen,10 Michael Boe Møller,5 Charlotte Guldborg Nyvold,11 Lise Lotte Hansen1 and Tomasz K. Wojdacz1,12,13 Department of Biomedicine, Aarhus University, Aarhus, Denmark; 2The Lundbeck Foundation Initiative for Integrative Psychiatric Research, iPSYCH, Aarhus, Denmark; 3 Center for Integrative Sequencing, iSEQ, Aarhus University, Aarhus, Denmark; 4Center for Genomics and Personalized Medicine, CGPM, Aarhus University, Aarhus, Denmark; 5 Department of Pathology, Odense University Hospital, Odense, Denmark; 6Department of Pathology, Aarhus University Hospital, Aarhus, Denmark; 7Department for Congenital Disorders, Statens Serum Institut, Copenhagen, Denmark; 8Department of Hematology, Aarhus University Hospital, Aarhus, Denmark; 9Department of Clinical Medicine, Aarhus University, Aarhus, Denmark; 10Department of Haematology, Odense University Hospital, Odense, Denmark; 11Haematology-Pathology Research Laboratory, Research Unit for Haematology and Research Unit for Pathology, University of Southern Denmark and Odense University Hospital, Odense, Denmark; 12Aarhus Institute of Advanced Studies, Aarhus University, Aarhus, Denmark and 13Independent Clinical Epigenetics Laboratory, Pomeranian Medical University, Szczecin, Poland 1

Ferrata Storti Foundation

Haematologica 2022 Volume 107(4):877-886

ABSTRACT

urrently, no molecular biomarker indices are used in standard care to make treatment decisions at diagnosis of chronic lymphocytic leukemia (CLL). We used Infinium MethylationEPIC array data from diagnostic blood samples of 114 CLL patients and developed a procedure to stratify patients based on methylation signatures associated with mutation load of the IGHV gene. This procedure allowed us to predict the time to treatment with a hazard ratio (HR) of 8.34 (95% confidence interval [CI]: 4.54-15.30), as opposed to a HR of 4.35 (95% CI: 2.60-7.28) using IGHV mutation status. Detailed evaluation of 17 cases for which the two classification procedures gave discrepant results showed that these cases were incorrectly classified using IGHV status. Moreover, methylation-based classification stratified patients with different overall survival (HR=1.82; 95% CI: 1.07-3.09), which was not possible using IGHV status. Furthermore, we assessed the performance of the developed classification procedure using published HumanMethylation450 array data for 159 patients for whom information on time to treatment, overall survival and relapse was available. Despite 450K array methylation data not containing all the biomarkers used in our classification procedure, methylation signatures again stratified patients with significantly better accuracy than did IGHV mutation load regarding all available clinical outcomes. Thus, stratification using IGHV-associated methylation signatures may provide better prognostic power than IGHV mutation status.

Introduction Most patients diagnosed with chronic lymphocytic leukemia (CLL) have asymptomatic, early-stage disease at the time of diagnosis but the subsequent disease course is highly variable, with some patients experiencing early progression and others living for many years with indolent disease.1 Immediate treatment after diagnosis does not seem to improve patients’ survival.2-5 Consequently, to reduce

haematologica | 2022; 107(4)

Correspondence: TOMASZ K. WOJDACZ tomasz.wojdacz@pum.edu.pl Received: January 31, 2021. Accepted: May 21, 2021. Pre-published: June 3, 2021. https://doi.org/10.3324/haematol.2021.278477

877

D. Hussmann et al.

unnecessary harmful complications following therapy, the majority of CLL patients are managed with a “watch and wait” strategy,6 and treatment is only initiated at disease progression. This is assessed according to clinical symptoms defined by the Rai and Binet staging systems.79 However, with the advent of new therapies it is wellrecognized that some patients can potentially benefit from earlier intervention.9 Molecular biomarker-based indices, as opposed to clinical staging, are likely to reflect the complex biology of CLL and, therefore, predict patients’ outcomes more accurately.10 However, the development of biomarker-based indices in CLL is still ongoing. Recent large multicenter studies, investigating the prognostic power of various known molecular and clinical biomarkers, have proposed two new biomarker indices: the International Prognostic Index for Chronic Lymphocytic Leukemia (CLL-IPI) and the International Prognostic Score for Early-stage CLL (IPS-E).11,12 The CLL-IPI index is based on TP53 aberrations, IGHV mutation status, b2-microglobulin concentration, clinical Rai/Binet stage, and age, with TP53 aberrations predicting overall survival (OS) most accurately in multivariable modeling.11 However, lesions affecting the TP53 locus are rather rare and other studies have shown that IGHV-mutated patients with TP53 locus aberrations experience a rather indolent disease course.13-15 The IPS-E index was developed for early-stage patients with asymptomatic disease and time to first treatment (TTFT) as a primary outcome.12 This index includes IGHV status, absolute lymphocyte count and palpable lymph nodes. TP53 status did not show independent prognostic power in this index, which indicates that this biomarker may provide no clinical relevance for predicting TTFT for early-stage patients. In both of the above indices, stratification of patients into mutated (M-CLL) or unmutated (U-CLL), according to IGHV mutation load, plays a central role.16 It is well-established that the CLL methylome reflects, to a large extent, the natural history of the B cell.17-20 Recent studies have also shown that the CLL methylome can guide the stratification of patients experiencing different clinical outcomes both at diagnosis18,19,21 and in clinical trials.20 Specifically, Kulis et al. and, subsequently, Queirós et al. have shown that methylation signatures can stratify CLL patients into three groups experiencing different clinical outcomes: the n-CLL (naïve B-cell-like CLL), i-CLL (intermediate CLL), and m-CLL (memory B-cell-like CLL) subgroups.18,21 The identified methylation signatures were closely related to IGHV mutation status, with the n-CLL and m-CLL subgroups consisting mainly of U-CLL patients and M-CLL patients, respectively. The new iCLL subgroup included borderline M-CLL and U-CLL patients, as they were found to display both an intermediate load of mutations in the IGHV gene and intermediate clinical outcomes.18,21 Further studies of i-CLL patients have shown that certain molecular features are enriched in this group of patients, such as poor-prognostic subset #2 characteristics.21,22 The subset #2 i-CLL cases seem to constitute an aggressive subgroup of i-CLL with clinical prognosis resembling the prognosis of n-CLL patients.22 Thus, the diagnostic utility of this classification needs to be studied further. The above findings clearly indicate that methylation signatures of CLL cells are largely associated with the mutation load of the IGHV gene and that they have prog878

nostic significance. In this study, we developed a procedure for classifying patients based on methylation changes associated with IGHV mutation load, comparing the prognostic power of this classification procedure to predict clinical outcomes with that of patients’ stratification based on IGHV mutation load alone.

Methods Clinical material Our cohort of patients has already been described;23,24 the patients’ clinicobiological characteristics are summarized in Table 1 (see Online Supplementary File, Patient Cohort Section). The Ethics Committee of the Region of Southern Denmark approved the study (approval number: S-20100128).

Genome-wide DNA methylation analysis To assess genome-wide DNA methylation, we analyzed 400 ng of DNA with the Illumina Infinium MethylationEPIC Beadchip (EPIC) array. Raw data were processed in R using the RnBeads package25 with default filtering settings including the removal of probes, which were: (i) outside CpG context; (ii) overlapping single-nucleotide polymorphisms; (iii) targeting sex chromosomes; (iv) missing b-values; (v) showing a standard deviation of b-values <0.005; and (vi) cross-reactive probes.26 bvalues were normalized using the BMIQ method27 followed by noob background correction.28 We assessed the sample purity using the methylomic data,19 and included only patients’ samples with at least 85% B cells (n=114) to limit the impact of celltype composition.

Bioinformatic and statistical analyses Bioinformatic and statistical analyses were performed in R version 3.6.1, Stata/SE 15.0 (StataCorp, TX, USA), and Qlucore Omics Explorer 3.4 (Qlucore, Lund, Sweden). We used linear regression to test the association between methylation levels at individual CpG loci (b-values) and mutation load of IGHV (as percentage identity to germline sequence to avoid specific cutoff29) for a total of 671,684 CpG, using P<10-8 as recommended for methylomic studies.30 Only CpG with qualitative methylation changes defined as an interquartile range of minimum 0.80 were included in subsequent analysis (Online Supplementary File, Section 1). The primary clinical endpoint used to develop the classification procedure was time to treatment (TTT). CpG with methylation levels associated with TTT were selected using Cox regression with the significance threshold of P<10-7; this was chosen to identify the most associated CpG and to control for false-positive results. CpG independently associated with TTT were identified in a multivariable Cox regression model using a backward elimination procedure with P<0.05. Classification of IGHV mutation load (IGHV status) into mutated (M-CLL) and unmutated (U-CLL) was based on 98% identity cutoff to the germline sequence.16 The strength of association between two classification methods was quantified by the odds ratio (OR) using Woolf approximation to calculate 95% confidence intervals (CI). Secondary clinical endpoints were OS and relapse.31 The prognostic accuracy of a classification method in predicting the clinical outcomes was evaluated using hazard ratios (HR) from univariate and multivariable Cox regression models, and by KaplanMeier plots combined with log-rank tests and estimation of median time to event. The Cox regression model assumptions were tested using Schoenfeld residuals, and P values <0.05 were considered as statistically significant results.

haematologica | 2022; 107(4)

IGHV-associated methylation in CLL prognostics

Validation of EPIC microarray data with methylation-sensitive high resolution melting The microarray data were validated using methylation-sensitive high-resolution melting.32 The details of the assay design can be found in Online Supplementary Methods, Section 2.

Stratification of patients using IGHV-associated methylation signatures from 450K data We used data from an independent CLL cohort (n=159)33 previously published by Kulis et al.18 and Queirós et al.21 to test whether HumanMethylation450 BeadChip (450K) data are sufficient to stratify patients using our procedure.

This analysis also showed that methylation status of the individual CpG sites predicted the clinical outcomes of patients with very similar accuracy (Online Supplementary File, Figures S2 and S3), and that none of the CpG sites was uniformly informative to predict short TTT (Online Supplementary Figure S1). Then, to combine the information from all nine CpG sites, we counted the number of CpG that predicted a short TTT for each patient and compared the HR between groups of patients with a different number of the CpG sites predicting short TTT. We performed this analysis for a series of different b-value cutoffs for individual CpG sites to allow us to establish a b-value cutoff at which the final stratification of patients was most accurate (Online Supplementary Figures S4 and S5).

Results Identification of methylation signatures that independently predict short time to treatment To investigate whether IGHV-associated methylation signatures can more accurately classify patients with aggressive disease at diagnosis than IGHV mutation status, we first used linear regression and identified 4,518 sites (CpG) in the EPIC array dataset at which the methylation levels (b-values) were associated with the IGHV mutation load (Figure 1A). Due to both technical and biological limitations of quantitative methylation measurements in clinical material (for a detailed description, see Online Supplementary File, Section 1), we focused our analysis on 147 sites of the 4518 CpG at which we also observed qualitative methylation changes (defined as an interquartile range of b-values >0.8) (Figure 1B). As TTT was the primary clinical indicator of aggressive disease in our study, we then used Cox regression to identify 44 CpG among these 147 sites at which the level of methylation (b-values) were associated with an increased hazard of short TTT (Figure 1C). Moreover, as biomarkers that independently predict clinical outcomes are most useful in clinical practice, we applied multivariable Cox regression analysis, performed as a backward elimination model, to select CpG sites at which the methylation levels independently predicted TTT (Figure 1D). This analysis resulted in a final set of nine CpG sites with six CpG located in gene bodies of REPS1 (cg21740960), RRM2B (cg00395579), SMYD3 (cg07395110), IL1B (cg07250315), UBE2R2 (cg02198280), and ATP9B (cg21394039); two CpG did not annotate to any known gene (cg03282117 and cg00185137) and one CpG was located in the S-shelf of a CpG island in the LMBR1 promoter (cg12032915).

Development of a methylation-based classification procedure Next, we assessed whether the methylation status of one of the nine selected CpG sites is sufficient to stratify the patients accurately into two groups with different TTT, or whether combining the information from all CpG sites stratifies patients more accurately. A detailed description of these analyses is provided in the Online Supplementary File, Section 3. Briefly, we used TTT as the primary outcome and estimated the power of the methylation changes at each CpG site to predict TTT using the HR from the Cox regression analysis. These analyses showed that hypomethylation predicted short TTT for one CpG site (cg07395110), while hypermethylation was associated with short TTT for the remaining CpG (Online Supplementary Figure S1). We then compared the HR of the individual CpG sites. haematologica | 2022; 107(4)

Table 1. Clinicobiological characteristics of the patients with chronic lymphocytic leukemia.

Variable Age Median [range], years Age ≤65 years Age >65 years Sex Male Female Binet stage A B+C ZAP70 expression* Low High CD38 expression† Low High Trisomy 12 Absent Present Del(11q) Absent Present Del(13q) Absent Present NOTCH1 mutation Absent Present TP53 aberration‡ Absent Present IGHV status** M-CLL U-CLL Time to treatment Median [range], months Number of treated patients Overall survival Median [range], months Median follow-up time [range], months Number of deceased patients

N (%) 114 71 [49-92] 37 (32 %) 77 (68 %) 114 72 (63 %) 42 (37 %) 114 77 (68 %) 37 (32 %) 114 67 (59 %) 47 (41 %) 114 84 (74 %) 30 (26 %) 112 101 (90 %) 11 (10 %) 114 101 (89 %) 13 (11 %) 113 55 (49 %) 58 (51 %) 114 110 (96 %) 4 (4 %) 114 103 (90 %) 11 (10 %) 114 72 (63 %) 42 (37 %) 114 51.3 [0.1-126] 64 (56 %) 114 98.2 [0.4-144] 98.9 (0.4-144) 57 (50 %)

Del: deletion; M-CLL: chonic lymphocytic leukemia with mutated IGHV; U-CLL: chonic lymphocytic leukemia with unmutated IGHV; * ZAP70 expression is considered to be high when >20% cells are positive; †CD38 expression is considered to be high when >30% cells are positive; ‡including TP53 mutation and del(17p); **germline homology >98 % is considered U-CLL.

879

D. Hussmann et al.

Figure. 1. Identification of methylation changes that independently predict short time to treatment. (A) Test of the association between the methylation level (as bvalue) with the IGHV mutation load (as percentage identity to germline sequence) using a linear regression model for 114 patients and 671 684 CpG sites. Two scatterplots display a non-significant association for cg17698174 (left) and a significant association for cg08090385 (right) between the b-values on the y-axis and IGHV mutation load on the x-axis. A total of 4 518 CpG showed significant association between b-values and IGHV mutation load using a significance threshold of P<10-8. (B) Selection of CpG with qualitative methylation changes. The boxplots display the distribution of methylation at cg08090385 (left) and cg00029031 (right), where each black dot represents a patient (n=114) with the b-value for the specific CpG indicated on the y-axis. The box displays the 25-, 50- and 75-percentiles. An interquartile range (IQR) >0.80 was defined as a qualitative methylation change (Online Supplementary Methods, Section 1), and CpG with an IQR >0.80 were selected for further analyses (n=147). (C) Univariate Cox regression analysis of the association between methylation level (as a continuous b-value) and time to treatment (TTT) in 114 patients for 147 CpG. The Manhattan plot displays the significance to predict TTT for each of the 147 CpG (dots), with the -log10(P value) from Cox regression analysis on the y-axis against the chromosomal location of the CpG on the x-axis. A total of 44 CpG (red dots) showed a significance level below the threshold of P<10-7. (D) Results from the multivariable Cox regression of 44 CpG to identify CpG sites that independently predict TTT performed using backward elimination. A final set of nine CpG sites was identified with a statistical significance of P<0.05.

880

haematologica | 2022; 107(4)

IGHV-associated methylation in CLL prognostics

Overall, this data modeling showed that the combination of the information from all nine CpG had a considerably stronger prognostic power to predict TTT than had information from individual CpG sites. Specifically, the stratification for patients displaying two or more CpG sites with methylation status indicating short TTT (poor prognosis) versus patients with none or one CpG site (favorable prognosis) identified patients experiencing short TTT with a HR of 8.34 (95% CI: 4.54-15.30; P<0.001) (Online Supplementary Figure S5b-f). This HR was a clear improvement, as the power to identify patients with short TTT for the individual CpG sites stratified patients with HR ranging from 4.10 (95% CI: 2.46-6.85; P<0.001) to 6.60 (95% CI: 3.76-11.58; P<0.001) (Online Supplementary Figure S3A-I). The overview of the developed classification procedure is shown in Figure 2.

Methylation-based classification predicts time to treatment with significantly higher accuracy than does IGHV mutation status Next, we compared the power to predict TTT of the methylation-based classification with stratification using IGHV mutation status (using the most frequent cutoff at 98% germline identity16). In our cohort, the methylationbased classification identified 53 patients with a poor prognosis and median TTT of 13.1 months (95% CI: 4.120.1), and 61 patients with a favorable prognosis for whom the median TTT was not reached. At the same time, stratification based on IGHV status identified 42 UCLL patients with a median TTT of 10.1 months (95% CI: 3.4-21.1) and 72 M-CLL patients for whom the median TTT was not reached. Cox regression analyses showed that the methylation-based classification was significantly more accurate in predicting the need for treatment as described by a HR of 8.34 (95% CI: 4.54-15.30; P<0.001), compared to a HR of 4.35 (95% CI: 2.60-7.28; P<0.001) for IGHV status. This was further corroborated by the Kaplan-Meier analyses shown in Figure 3A, B. The two stratification methods provided discrepant classifications for 17 patients (Online Supplementary Figure S6). The methylation-based classification predicted a

poor prognosis for 14 M-CLL cases. Those patients, however, experienced a significantly shorter median TTT of 16.2 months (95% CI: 3.9-37.9), than the median TTT of the remaining M-CLL patients (n=58) who did not reach the median TTT (P<0.0001) (Figure 3C, dotted curves). Similarly, the median TTT for the three U-CLL patients predicted to have a favorable prognosis according to the methylation-based classification was 70.0 months (95% CI: 64.7-not reached), and significantly longer than the median TTT of the remaining U-CLL patients (n=39), which was 8.0 months (95% CI: 1.9-20.1; P=0.0188) (Figure 3C, dashed curves). These Kaplan-Meier curves clearly indicate that the methylation-based classification predicted TTT more accurately for the discrepantly classified patients. We further analyzed the IGHV mutation load of the discrepant cases, and found that they displayed an intermediate level of IGHV mutations; this was significantly different and closer to the 98% cutoff than that of the remaining patients with similar IGHV status (Online Supplementary Figure S7). This may indicate a limitation of the IGHV mutation-based stratification of these cases.

Accuracy of methylation-based classification to predict overall survival We then compared the accuracy of the two classifications to predict OS. Overall, 57 out of 114 patients in our study cohort experienced events and the median followup time was 98.9 months (95% CI: 94.4-117.6). Cox regression and Kaplan-Meier analyses showed that patients stratified using the methylation-based classification had significantly different OS (Cox regression: HR=1.82; 95% CI: 1.07-3.09; P=0.027; Kaplan-Meier: P=0.0246) (Figure 3D). At the same time, IGHV statusbased stratification was not able to identify patients with different OS in our cohort (Cox regression: HR=1.35; 95% CI: 0.80-2.28; P=0.263; Kaplan-Meier: P=0.2608) (Figure 3E). We did not find significant differences in OS for the 17 patients with discrepant classification between the IGHV status- and methylation-based classifications (Figure 3F and Online Supplementary Figure S6). However,

Figure. 2. The methylation classification procedure based on the nine selected CpG sites. The final procedure for methylation-based classification of patients into having a favorable or poor prognosis. (A) Methylation levels (b-values) at nine selected CpG sites obtained from the EPIC array for two random patients. (B) Classification of the b-value according to the cutoff as 1 if the b-values predict short time to treatment (TTT). (C) The number of CpG predicting short TTT is counted for each patient, and the patient is classified as having a favorable prognosis if 0-1 CpG predicts a short TTT, or as having a poor prognosis if 2-9 CpG predict a short TTT.

haematologica | 2022; 107(4)

881

D. Hussmann et al.

the follow-up time in our cohort was relatively short and an increased number of events is likely needed to increase the power of this analysis.

Methylation-based stratification of patients from 450K array data The cohort size available in this study did not allow us to divide patients into discovery and validation cohorts, which would be the most accurate way of assessing the prognostic power of a proposed procedure for stratifying CLL patients. Furthermore, we were not able to identify a publicly-available EPIC array dataset from a similar CLL cohort that could be used to validate our findings. The majority of genome-wide methylation profiling studies in CLL have, so far, been performed using the 450K array; a previous generation of the methylomic microarray. We assessed whether limited data obtained using the 450K BeadChip, which contained only three of the nine CpG sites we used to classify patients (cg00395579, cg12032915, and cg21394039), allow for the accurate stratification of patients according to the classification procedure we developed. The data we used here have been previously published and came from 159 CLL patients with TTT data available for 138 patients (34 events), and OS data for 139 patients (33 events).18,21 Relapse data were available for a subset of the patients in this cohort (74 patients/74 events), allowing us to make a

preliminary assessment of the power of methylationbased classification to predict relapse. Even with the limited data available for this cohort, our methylation-based classification procedure was able to stratify patients into two groups with different TTT (Online Supplementary Figure S8A) with a similar strength to that observed in our cohort: HR=8.41 (95% CI: 3.74-18.89; P<0.001). The methylation classification also stratified patients with different OS with HR=6.03 (95% CI: 2.65-13.73; P<0.001), and a different likelihood of relapse with HR=2.38 (95% CI: 1.33-4.25; P=0.003) (Online Supplementary Figure S8B, C). In this cohort, we also compared the performance of the methylation-based classification with that of stratification using IGHV status. The methylation-based classification stratified patients (96 patients/13 events) with different TTT with HR=5.20 (95% CI: 1.53-17.71; P=0.008) (Figure 4A; P=0.0038), as opposed to IGHV status which only stratified patients with borderline statistical significance: HR=2.97 (95% CI: 0.96-9.17; P=0.059) (Figure 4B; P<0.0001). The analysis of OS for patients in this cohort (97 patients/14 events) showed similar results to those observed in our CLL cohort, among whom only the methylation-based classification was able to stratify patients with different OS (HR=5.18; 95% CI: 1.62-16.53; P=0.006) (Figure 4D; P=0.0022), and IGHV status was not informative (HR) 2.46; 95% CI: 0.84-7.24; P=0.102)

Figure 3. Kaplan-Meier analyses of time to treatment and overall survival for the methylation-based classification and IGHV status stratifications. (A-C) Kaplan Meier curves describing time to treatment , and (D-F) Kaplan-Meier curves describing overall survival for stratification of patients using methylation-based classification (A and D), IGHV status (B and E), or both stratification methods (C and F) in our cohort of patients. In (C and F), the curves represent patients classified according to both stratification procedures: patients with mutated IGHV (dotted line) and unmutated IGHV (dashed lines) are represented by different colors according to the prediction by the methylation classification into favorable (green) or poor (red) prognosis. The log-rank test for equality was performed between all groups in (C and F), and all P values are listed below: green/dotted curve versus red/dotted curve (C: P<0.0001; F: P=0.1601); green/dotted curve versus green/dashed curve (C: P=0.3698; F: P=0.2236); green/dotted curve versus red/dashed curve (C: P<0.0001; F: P=0.0675); red/dotted curve versus green/dashed curve (C: P=0.0719; F: P=0.0945); red/dotted curve versus red/dashed curve (C: P=0.5284; F: P=0.9315); green/dashed curve versus red/dashed curve (C: P=0.0188; F: P=0.1048). M-CLL: patients with chronic lymphocytic leukemia with mutated IGHV; U-CLL: patients with chronic lymphocytic leukemia with unmutated IGHV.

882

haematologica | 2022; 107(4)

IGHV-associated methylation in CLL prognostics

(Figure 4E; P=0.0916). Moreover, despite a limited number of patients with available IGHV status and relapse data (39 patients/39 events), the methylation-based classification still stratified patients experiencing different times to relapse with a HR=3.55 (95% CI: 1.54-8.18; P=0.003) (Figure 4G; P=0.0018), whereas IGHV status was not informative (HR=1.05; 95% CI: 0.55-2.01; P=0.872) (Figure 4B; P=0.8721).

Ten patients were classified discrepantly by the two classification procedures. The statistical analyses of data for those patients were of very limited power. However, three U-CLL patients classified as likely to have a favorable prognosis according to methylation signature did not experience an event but participated long enough in the study to speculate that they did indeed have both a favorable TTT and OS, as indicated by the dashed green

Figure 4. Kaplan-Meier analyses of time to treatment and overall survival for methylation-based classification and IGHV status stratifications in the cohort for which 450K data were available. (A-C) Kaplan-Meier curves describing time to treatment, (D-F) Kaplan-Meier curves describing overall survival, and (G-I) Kaplan-Meier curves describing relapse for patients stratified using the methylation-based classification (A, D, and G), IGHV status (B, E, and H), or both stratification methods (C, F, and I) in an independent cohort of patients (450K data). In (C, F, and I) the curves represent patients classified according to both stratification procedures: chronic lymphocytic leukemia patients with mutated IGHV (dotted) and unmutated IGHV (dashed) were colored according to the prediction by the methylation classification into favorable (green) or poor (red) prognosis. In (C), the case with mutated IGHV with a favorable prognosis (dotted/red curve) is visually difficult to spot behind the other curves in the top left of the Kaplan-Meier plot. The log-rank test for equality was performed between all groups in (C, F, and I), and all P are listed below: green/dotted curve versus red/dotted curve (C: P=0.7598; F: P=0.0323; I: P=0.0070); green/dotted curve versus green/dashed curve (C: P=0.4285; F: P=0.5530; I: P=0.0880); green/dotted curve versus red/dashed curve (C: P=0.0033; F: P=0.0134; I: P=0.0306); red/dotted curve versus green/dashed curve (C: P=not available because there were no events; F: P=0.1415; I: P=0.0405); red/dotted curve versus red/dashed curve (C: P=0.5550; F: P=0.3584; I: P=0.0251); green/dashed curve versus red/dashed curve (C: P=0.1034; F: P=0.1854; I: P=0.0111). M-CLL: patients with chronic lymphocytic leukemia with mutated IGHV; U-CLL: patients with chronic lymphocytic leukemia with unmutated IGHV.

haematologica | 2022; 107(4)

883

D. Hussmann et al.

Table 2. Multivariable Cox regression analysis for time to treatment and overall survival according to the methylation-based classification adjusted for clinicobiological biomarkers.

Variable

Time to treatment HR (95% CI)

Methylation-based classification Binet stage

Favorable Poor A

Del(13q) Del(11q)

B+C Absent Present Absent Present

8.33 (4.28-16.19)

Overall survival P <0.001

4.87 (2.68-8.85)

<0.001

0.52 (0.30-0.91)

0.021

0.30 (0.14-0.65)

0.002

Variable Age at diagnosis Methylation-based classification

HR (95% CI)

≤65 years >65 years Favorable

3.11 (1.59-6.09)

0.001

Poor

1.96 (1.15-3.35)

0.013

The table displays the final model with variables showing independent prognostic potential with hazard ratios indicating the likelihood of an event given the biomarker status in the specific row. The time to treatment (TTT) model was built on data from 113 patients with chronic lymphocytic leukemia (CLL) among whom 63 started treatment and the overall survival model was built on data from 114 CLL patients of whom 57 died. Adjustment for confounding variables was performed using backward elimination, and P values <0.05 were considered statistically significant. The models were built using all standard clinicobiological biomarkers available, including: age at diagnosis (≤65 vs. >65 years), Binet stage (A vs. B+C), sex, ZAP70 expression, CD38 expression, trisomy 12, del(11q), del(13q), NOTCH1 mutation, and TP53 aberrations. HR: hazard ratio; CI: confidence interval. OS, overall survival; TTT, time to treatment.

Kaplan-Meier curves in Figure 4C and 4F, respectively. Similarly, the dotted red Kaplan-Meier curves in Figure 4C and 4F for seven M-CLL patients classified by methylation signatures as likely to have a poor prognosis suggest short TTT and OS. Furthermore, the relapse data for seven discrepant patients confirmed that the two U-CLL patients classified as having a favorable prognosis had a significantly different time to relapse than that of the remaining U-CLL patients (Figure 4I, dashed curves; P=0.0111), and likewise, the five M-CLL patients classified as having a poor prognosis had a significantly different time to relapse compared to that of the remaining MCLL patients (Figure 4I, dotted curves; P=0.0070). The IGHV mutation load was not available for this cohort and we were not able to assess whether the mutation loads of the discrepantly classified patients were close to the IGHV mutation cutoff, suggesting a difficulty in classifying those patients similar to those in our cohort.

Association of IGHV status and methylation-based classification with standard clinicobiological biomarkers of chronic lymphocytic leukemia In our cohort, we also analyzed the association of standard clinicobiological biomarkers used in CLL prognostication with both methylation-based classification and IGHV status. The analysis was based on all variables available for this cohort, including: sex, age, Binet stage, ZAP70 expression, CD38 expression, del(11q), del(13q), trisomy 12, NOTCH1 mutation, and TP53 locus aberrations (Online Supplementary Table S1, Online Supplementary Figure S6). Advanced Binet stage, ZAP70 expression, CD38 expression, del(11q), del(13q), and NOTCH1 mutation were significantly associated with both U-CLL patients (for IGHV status stratification) and with poor prognosis patients, according to the methylation-based classification. Furthermore, classification of patients as UCLL was significantly associated with sex; other biomarkers did not show statistically significant associations with any of the subgroups of patients. The frequency of the biomarkers in the discrepantly stratified patients were too low for definite conclusions to be drawn (Online Supplementary Table S2); however, most of the discrepantly stratified patients had early-stage disease (Binet stage A: 13/17). In univariate Cox regression analyses, the methylation-based classification predicted TTT most 884

accurately among all standard clinicobiological CLL biomarkers, and only age predicted OS more accurately than did the methylation-based classification (Online Supplementary Table 3). The multivariable models that included all the above biomarkers and were developed using the backward elimination procedure confirmed an independent power of methylation-based classification to predict TTT with a HR=8.33 (95% CI: 4.28-16.19; P<0.001) along with Binet stage, del(13q) and del(11q), and OS with a HR=1.96 (95% CI: 1.15-3.35; P=0.013) along with age (Table 2). In an identical modeling procedure, IGHV mutation status independently predicted TTT with HR=2.35 (95% CI: 1.34-4.15; P=0.003) along with Binet stage, NOTCH1 mutation, and ZAP70 expression, but was not informative regarding OS (Table 3). As the CpG sites in our stratification procedure were selected based on the association between the methylation levels and IGHV mutation load (Figure 1A), the multivariable modeling was performed separately for those variables due to the expected intercorrelation. We also compared the prognosis of patients classified with our procedure with that of the biological subgroups identified by classification procedure recently described by Duran-Ferrer et al.34 This procedure identified: 35 nCLL, 20 i-CLL and 59 m-CLL with distinct TTT in our cohort (Online Supplementary Figure S9A). All 35 n-CLL predicted to experience poor prognosis were also classified as likely to have a poor prognosis with our classifier. However, our classification procedure stratified 59 mCLL cases into six with a poor prognosis and 53 with a favorable prognosis. Similarly, 20 i-CLL patients were stratified into 12 with a poor prognosis and eight with a favorable prognosis. The groups identified by our procedure did indeed experience, statistically, significantly different outcomes, as illustrated by the Kaplan-Meier curves in Online Supplementary Figure S9B, C. In the cohort of patients for whom 450k data were available, all 66 nCLL cases were predicted to experience a poor prognosis according to our classification, and out of 64 m-CLL cases, one was predicted to have a poor prognosis. Of the 29 i-CLL cases, 14 were predicted to have a favorable prognosis while 15 were predicted to have a poor prognosis. The comparison of clinical data for the discrepant cases was not possible as most of their time data were censored. haematologica | 2022; 107(4)

IGHV-associated methylation in CLL prognostics

Table 3. Multivariable Cox regression analysis for time to treatment and overall survival according to IGHV status adjusted for clinicobiological biomarkers.

Variable Binet stage NOTCH1 mutation IGHV status ZAP70 expression

Time to treatment HR (95% CI) A B+C Absent Present M-CLL U-CLL Low High

4.49 (2.55-7.91)

Overall survival P <0.001

3.57 (1.20-10.66)

0.022

2.35 (1.34-4.15)

0.003

2.07 (1.19-3.59)

0.009

Variable Age at diagnosis ZAP70 expression

≤65 years >65 years Low High

HR (95% CI)

3.28 (1.66-6.45)

0.001

2.04 (1.20-3.47)

0.008

The table displays the final model with variables showing independent prognostic potential with hazard ratios indicating the likelihood of an event given the biomarker status in the specific row.The model was built on data from 114 patients with chronic lymphocytic leukemia (CLL) with 64 events. Adjustment for confounding variables was performed using backward elimination, and P values <0.05 were considered statistically significant. The models were built using all standard clinicobiological biomarkers available, including: age at diagnosis (≤65 vs. >65 years), Binet stage (A vs. B+C), sex, ZAP70 expression, CD38 expression, trisomy 12, del(11q), del(13q), NOTCH1 mutation, and TP53 aberrations. HR: hazard ratio; CI: confidence interval. M-CLL: chronic lymphocytic leukemia with mutated IGHV; U-CLL: chronic lymphocytic leukemia with unmutated IGHV.

Polymerase chain reaction validation of microarray data To follow good laboratory practice, we performed a technical validation of methylation measurements obtained from the microarray analysis in our cohort with methylation-sensitive high-resolution melting. The results obtained corroborated the microarray data (Online Supplementary Figure S10).

Discussion The initiation of treatment of CLL patients is still based on progression according to clinical symptoms. However, as a substantial group of CLL patients progresses shortly after diagnosis, or rapidly experiences relapse, it is generally acknowledged that some patients may benefit from earlier intervention. IGHV mutation load, together with TP53 aberrations, are currently the most widely adopted prognostic markers in CLL diagnostics; however, molecular biomarkers are not considered in the decision to treat, and the most clinically relevant cutoff for IGHV status is still debated.29 Moreover, some studies have indicated that TP53 locus aberrations may not to be informative for patients with early-stage disease.11,12 The prognostic value of methylation changes in CLL has been described;18,21 however, the clinical utility of methylation signatures directly associated with IGHV mutation load has not yet been studied. Here we investigated the prognostic power of the IGHV-associated methylation changes in CLL and developed a procedure for classifying patients based on those signatures. We then evaluated the prognostic accuracy of the developed classification procedure and found that it provided a significantly more accurate prediction of TTT and OS than the stratification based on IGHV status alone. Furthermore, we assessed the prognostic validity of classification in an independent cohort of patients;18,21 despite the fact that the methylation data for this cohort were limited, compared to IGHV status the methylation signatures in the independent CLL cohort also displayed significantly higher prognostic accuracy to predict TTT, OS and relapse. Moreover, the analysis of clinical outcomes is this cohort indicated that considerably longer follow-up (138.0 months vs. 98.9 months in our cohort) further improved the accuracy of the methylation classification (HR=6.03; 95% CI: 2.65-13.74) compared to haematologica | 2022; 107(4)

that in our cohort (HR=1.82; 95% CI: 1.07-3.09). At the same time, we did not see an improvement of the prognostic value of IGHV status regarding OS with the longer follow-up, which was not informative in either cohort. However, the fact that OS was not informative may be attributed to the specificity of the patients in these two cohorts because IGHV status predicted OS in other studies.35 Due to limited data availability, we were not able to evaluate our findings in the context of an already proposed methylation signature-based stratification21 and CLL-classification indices (such as the CLL-IPI and IPSE).11,12 Nevertheless, our data indicate that it is plausible that the performance of biomarker indices that use IGHV mutation status will improve with the implementation of the proposed patients’ classification procedure based on methylation changes. It is also important to note that the genome-wide methylation screening technology used in this project has already been proposed for diagnostic use in glioblastoma,36,37 indicating that, despite its still high cost, this technology is worth considering given the data quality and the amount of data obtained from single experiments. In summary, our results show that IGHV-associated methylation signatures may be more accurate than IGHV mutation status in predicting CLL patients’ outcomes, including the identification of patients with aggressive disease at diagnosis as well as treatment outcomes. Our results also indicate that the prognostic power of biomarker indices including IGHV mutation status can, potentially, be improved with the addition of methylation markers, but this needs to be addressed in further studies. Disclosures No conflicts of interest to disclose. Contributions DH and TKW designed the study; TEK, ML, and KDJ managed the preparation of patients’ samples and quality control; JBG performed bisulfite treatment and microarray analysis; DH performed the bioinformatic and statistical analyses assisted by TKW and AS; DH, TKW, AS, and LLH interpreted results; DH designed the methylation-sensitive high-resolution melting assays, and AT performed them; DH, TKW, and LLH wrote the manuscript draft; TKW and LLH supervised the study; LK, ID, TK, TSL, MBM and CGN participated in the collection of clin885

D. Hussmann et al.

ical samples and clinical data, and standard biomarker analyses. All authors critically reviewed the manuscript and approved the final version. Acknowledgements We thank the clinical contributors and the data producers for the publicly available dataset through ICGC used for independent validation. Funding This work was supported by Harboefonden, Knud og Edith

References 1. Fabbri G, Dalla-Favera R. The molecular pathogenesis of chronic lymphocytic leukaemia. Nat Rev Cancer. 2016;16(3):145162. 2. Spanish Cooperative Group P. Treatment of chronic lymphocytic leukemia: a preliminary report of Spanish (Pethema) trials. Leuk Lymphoma. 1991;5(Suppl 1):89-91. 3. Shustik C, Mick R, Silver R, et al. Treatment of early chronic lymphocytic leukemia: intermittent chlorambucil versus observation. Hematol Oncol. 1988;6(1):7-12. 4. Dighiero G, Maloum K, Desablens B, et al. Chlorambucil in indolent chronic lymphocytic leukemia. French Cooperative Group on Chronic Lymphocytic Leukemia. N Engl J Med. 1998;338(21):1506-1514. 5. Geisler C, Hansen MM, Yeap BY, et al. Chemotherapeutic options in chronic lymphocytic leukemia: a meta-analysis of the randomized trials. J Natl Cancer Inst. 1999;91(10):861-868. 6. Eichhorst B, Robak T, Montserrat E, et al. Chronic lymphocytic leukaemia: ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2015;26(Suppl 5):v78-84. 7. Rai KR, Sawitsky A, Cronkite EP, et al. Clinical staging of chronic lymphocytic leukemia. Blood. 1975;46(2):219-234. 8. Binet JL, Auquier A, Dighiero G, et al. A new prognostic classification of chronic lymphocytic leukemia derived from a multivariate survival analysis. Cancer. 1981;48(1):198206. 9. Hallek M, Shanafelt TD, Eichhorst B. Chronic lymphocytic leukaemia. Lancet. 2018;391(10129):1524-1537. 10. Hallek M, Cheson BD, Catovsky D, et al. Guidelines for the diagnosis and treatment of chronic lymphocytic leukemia: a report from the International Workshop on Chronic Lymphocytic Leukemia updating the National Cancer Institute-Working Group 1996 guidelines. Blood. 2008;111(12):5446-5456. 11. International CLL-IPI working group. An international prognostic index for patients with chronic lymphocytic leukaemia (CLLIPI): a meta-analysis of individual patient data. Lancet Oncol. 2016;17(6):779-790. 12. Condoluci A, Terzi di Bergamo L, Langerbeins P, et al. International prognostic score for asymptomatic early-stage chronic lymphocytic leukemia. Blood. 2020;135(21): 1859-1869. 13. Hu B, Patel KP, Chen HC, et al. Association

886

Eriksens Mindefond and Arvid Nilsson Fonden. The Graduate School of Health, Aarhus University, Denmark, funded the PhD fellowship for DH. TKW was supported by the ‘Polish Returns’ program from the Polish National Agency for Academic Exchange and the individual Maria Skłodowska-Curie fellowship CONFUND at Aarhus Institute of Advanced Studies. AS was supported by the Lundbeck Foundation. Data-sharing statement The raw data upon which we built the classification procedure are available in the Online Supplementary File.

of gene mutations with time-to-first treatment in 384 treatment-naive chronic lymphocytic leukaemia patients. Br J Haematol. 2019;187(3):307-318. 14. Tam CS, Shanafelt TD, Wierda WG, et al. De novo deletion 17p13.1 chronic lymphocytic leukemia shows significant clinical heterogeneity: the M. D. Anderson and Mayo Clinic experience. Blood. 2009;114(5):957964. 15. Best OG, Gardiner AC, Davis ZA, et al. A subset of Binet stage A CLL patients with TP53 abnormalities and mutated IGHV genes have stable disease. Leukemia. 2009;23(1):212-214. 16. Rosenquist R, Ghia P, Hadzidimitriou A, et al. Immunoglobulin gene sequence analysis in chronic lymphocytic leukemia: updated ERIC recommendations. Leukemia. 2017;31 (7):1477-1481. 17. Kulis M, Merkel A, Heath S, et al. Wholegenome fingerprint of the DNA methylome during human B cell differentiation. Nat Genet. 2015;47(7):746-756. 18. Kulis M, Heath S, Bibikova M, et al. Epigenomic analysis detects widespread gene-body DNA hypomethylation in chronic lymphocytic leukemia. Nat Genet. 2012;44(11):1236-1242. 19. Oakes CC, Seifert M, Assenovl Y, et al. DNA methylation dynamics during B cell maturation underlie a continuum of disease phenotypes in chronic lymphocytic leukemia. Nat Genet. 2016;48(3):253-264. 20. Wojdacz TK, Amarasinghe HE, Kadalayil L, et al. Clinical significance of DNA methylation in chronic lymphocytic leukemia patients: results from 3 UK clinical trials. Blood Adv. 2019;3(16):2474-2481. 21. Queiros AC, Villamor N, Clot G, et al. A Bcell epigenetic signature defines three biologic subgroups of chronic lymphocytic leukemia with clinical impact. Leukemia. 2015;29(3):598-605. 22. Bhoi S, Ljungstrom V, Baliakas P, et al. Prognostic impact of epigenetic classification in chronic lymphocytic leukemia: the case of subset #2. Epigenetics. 2016;11(6): 449-455. 23. Kristensen L, Kristensen T, Abildgaard N, et al. LPL gene expression is associated with poor prognosis in CLL and closely related to NOTCH1 mutations. Eur J Haematol. 2016;97(2):175-182. 24. Kristensen L, Kristensen T, Abildgaard N, et al. High expression of PI3K core complex genes is associated with poor prognosis in chronic lymphocytic leukemia. Leuk Res. 2015;39(6):555-560.

25. Assenov Y, Muller F, Lutsik P, et al. Comprehensive analysis of DNA methylation data with RnBeads. Nat Methods. 2014;11(11):1138-1140. 26. McCartney DL, Walker RM, Morris SW, et al. Identification of polymorphic and off-target probe binding sites on the Illumina Infinium MethylationEPIC BeadChip. Genom Data. 2016;9:22-24. 27. Teschendorff AE, Marabita F, Lechner M, et al. A beta-mixture quantile normalization method for correcting probe design bias in Illumina Infinium 450 k DNA methylation data. Bioinformatics. 2013;29(2):189-196. 28. Triche TJ, Weisenberger DJ, Van Den Berg D, et al. Low-level processing of Illumina Infinium DNA Methylation BeadArrays. Nucleic Acids Res. 2013;41(7):e90. 29. Davis Z, Forconi F, Parker A, et al. The outcome of chronic lymphocytic leukaemia patients with 97% IGHV gene identity to germline is distinct from cases with < 97% identity and similar to those with 98% identity. Br J Haematol. 2016;173(1):127-136. 30. Saffari A, Silver MJ, Zavattari P, et al. Estimation of a significance threshold for epigenome-wide association studies. Genet Epidemiol. 2018;42(1):20-33. 31. Hallek M, Cheson BD, Catovsky D, et al. iwCLL guidelines for diagnosis, indications for treatment, response assessment, and supportive management of CLL. Blood. 2018;131(25):2745-2760. 32. Wojdacz TK, Dobrovic A, Hansen LL. Methylation-sensitive high-resolution melting. Nat Protoc. 2008;3(12):1903-1908. 33. https://dcc.icgc.org, project code: CLLE-ES. [last accessed January 30, 2022] 34. Duran-Ferrer M, Clot G, Nadeu F, et al. The proliferative history shapes the DNA methylome of B-cell tumors and predicts clinical outcome. Nat Cancer. 2020;1(11): 1066-1081. 35. Rotbain EC, Frederiksen H, Hjalgrim H, et al. IGHV mutational status and outcome for patients with chronic lymphocytic leukemia upon treatment: a Danish nationwide population-based study. Haematologica. 2020;105(6):1621-1629. 36. Karimi S, Zuccato JA, Mamatjan Y, et al. The central nervous system tumor methylation classifier changes neuro-oncology practice for challenging brain tumor diagnoses and directly impacts patient care. Clin Epigenetics. 2019;11(1):185. 37. Capper D, Jones DTW, Sill M, et al. DNA methylation-based classification of central nervous system tumours. Nature. 2018;555 (7697):469-474.

haematologica | 2022; 107(4)

ARTICLE

Hematopoiesis

Perturbed hematopoiesis in individuals with germline DNMT3A overgrowth Tatton-Brown-Rahman syndrome Ayala Tovy,1,2 Carina Rosas,1,2 Amos S. Gaikwad,3 Geraldo Medrano,3 Linda Zhang,1,2,4 Jaime M. Reyes,1,2,5 Yung-Hsin Huang,1 Tatsuhiko Arakawa,1,2 Kristen Kurtz,3 Shannon E. Conneely,3 Anna G. Guzman,1,2 Rogelio Aguilar,3 Anne Gao,3 Chun-Wei Chen,1,2,4 Jean J. Kim,1,2,6 Melissa T. Carter,7 Amaia Lasa-Aranzasti,8 Irene Valenzuela,8 Lionel Van Maldergem,9 Lorenzo Brunetti,310 M. John Hicks,11 Andrea N. Marcogliese,3 Margaret A. Goodell1,2,3,4,5,6# and Rachel E. Rau1,3#

Ferrata Storti Foundation

Haematologica 2022 Volume 107(4):887-898

1 Stem Cells and Regenerative Medicine Center, Baylor College of Medicine, Houston, TX, USA; 2Department of Molecular and Cellular Biology, Baylor College of Medicine, Houston, TX, USA; 3Department of Pediatrics, Baylor College of Medicine and Texas Children's Hospital, Houston, TX, USA; 4Graduate Program in Translational Biology and Molecular Medicine, Baylor College of Medicine, Houston, TX, USA; 5Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, TX, USA; 6 Department of Education, Innovation and Technology, Baylor College of Medicine, Houston, TX, USA; 7Department of Genetics, Children’s Hospital of Eastern Ontario, Ottawa, Ontario, Canada; 8Department of Clinical and Molecular Genetics, Vall d´Hebron University Hospital and Medicine Genetics Group, Vall d´Hebron Research Institute, Barcelona, Spain; 9Centre de Génétique Humaine and Integrative and Cognitive Neuroscience Research Unit EA481, University of Franche-Comté, Besancon, France; 10 Department of Medicine and Surgery, University of Perugia, Perugia, Italy and 11 Department of Pathology Texas Children’s Hospital and Department of Pathology and Immunology, Baylor College of Medicine, Houston, TX, USA #

MAG and RER contributed equally as co-senior authors.

ABSTRACT

atton-Brown-Rahman syndrome (TBRS) is an overgrowth disorder caused by germline heterozygous mutations in the DNA methyltransferase DNMT3A. DNMT3A is a critical regulator of hematopoietic stem cell (HSC) differentiation and somatic DNMT3A mutations are frequent in hematologic malignancies and clonal hematopoiesis. Yet, the impact of constitutive DNMT3A mutation on hematopoiesis in TBRS is undefined. In order to establish how constitutive mutation of DNMT3A impacts blood development in TBRS we gathered clinical data and analyzed blood parameters in 18 individuals with TBRS. We also determined the distribution of major peripheral blood cell lineages by flow cytometric analyses. Our analyses revealed non-anemic macrocytosis, a relative decrease in lymphocytes and increase in neutrophils in TBRS individuals compared to unaffected controls. We were able to recapitulate these hematologic phenotypes in multiple murine models of TBRS and identified rare hematological and non-hematological malignancies associated with constitutive Dnmt3a mutation. We further show that loss of DNMT3A in TBRS is associated with an altered DNA methylation landscape in hematopoietic cells affecting regions critical to stem cell function and tumorigenesis. Overall, our data identify key hematopoietic effects driven by DNMT3A mutation with clinical implications for individuals with TBRS and DNMT3A-associated clonal hematopoiesis or malignancies.

Introduction Tatton-Brown-Rahman syndrome (TBRS; OMIM: 615879) is a germline dominant disorder due to constitutive heterozygous mutations in the de novo DNA methyltransferase DNMT3A. Individuals with TBRS are typically tall, obese, macrocephalic and have varying degrees of intellectual disability.1,2 Thus far, approximately 60 TBRS individuals have been described in the literature. However, the number of diagnosed indi-

haematologica | 2022; 107(4)

Correspondence: RACHEL E. RAU rachel.rau@bcm.edu MARGARET A. GOODELL goodell@bcm.edu Received: April 14, 2021. Accepted: May 21, 2021. Pre-published: June 3, 2021. https://doi.org/10.3324/haematol.2021.278990

887

A. Tovy et al.

viduals is steadily increasing with growing awareness of DNMT3A mutations as a cause of overgrowth and developmental disorders. As the full extent of TBRS clinical features is not yet defined, ongoing efforts to characterize the effects of constitutive DNMT3A loss will be critical to facilitate identification of affected individuals and improve their care and quality of life. DNMT3A is perhaps best known as a regulator of blood development governing the balance between hematopoietic stem cell (HSC) self-renewal and differentiation.3 Loss-offunction DNMT3A mutations are highly prevalent across a spectrum of adult hematologic malignancies, further confirming its indispensable role in hematopoiesis.4-6 In addition, DNMT3A is the most commonly mutated gene in clonal hematopoiesis (CH),7,8 a phenomenon of aging associated with mutant hematopoietic stem and progenitor cell (HSPC) expansion.9 Individuals with CH have a significantly increased risk for hematologic malignancy development. This increased risk is likely due to a competitive growth advantage that leads to an accumulation of DNMT3Amutant HSPC in the bone marrow over time,3 with a proportional increase in the likelihood of acquiring collaborating leukemogenic mutations. Consistent with this timedependent model of clonal expansion and malignant transformation is the fact that, while common in adult hematologic malignancies, DNMT3A mutations are exceedingly rare in pediatric leukemias.10-13 The mutation spectrum reported for TBRS individuals covers all the functional domains of DNMT3A and overlaps with CH and hematologic malignancies.1 However, in contrast to the somatic mutations in CH and cancer, in TBRS individuals, the DNMT3A mutation is constitutive, arising before embryogenesis. This timing raises the concern that the negative hematopoietic sequelae associated with somatic DNMT3A mutations in older individuals could have an early onset in TBRS. Reports of recurrent infections, bleeding tendencies, and leukemias in individuals with TBRS suggest that there may be a correlation between constitutive DNMT3A loss and abnormal hematopoiesis.1,14,15 Despite these concerns and anecdotal observations, how constitutive loss of DNMT3A impacts blood formation and leukemia predisposition in TBRS remains unknown. In order to define basal hematopoiesis in TBRS, we extensively characterized the hematological parameters and peripheral blood cellular composition of TBRS individuals with a spectrum of DNMT3A mutations and deletions. We found that germline DNMT3A mutations impact multilineage blood development, resulting in clinically relevant phenotypes. We further identified DNA methylation changes as one mechanism through which constitutive loss of DNMT3A alters blood development in TBRS. In order to confirm the observed hematopoietic effects, we extensively investigated hematopoiesis in multiple mouse models of TBRS. These murine models recapitulated the major hematologic phenotypes identified in humans. Longitudinal studies of the TBRS mouse models demonstrated that while the risk of hematologic malignancies is increased in TBRS, disease penetrance overall is low.

assent where appropriate were obtained from each participant and/or his or her guardian in accordance with the declaration of Helsinki prior to any study procedures. Peripheral blood samples from individuals with germline DNMT3A mutations or deletions and unaffected controls were collected in EDTA-coated tubes. Available medical records for enrolled patients were reviewed for growth parameters, laboratory studies and hematologic diagnoses.

Mouse models All mice were housed in AAALAC-accredited, specificpathogen-free animal care facilities at Baylor College of Medicine, and all procedures were approved by the Institutional Animal Care and Use Committee. C57BL/6 mice of both sexes were used unless stated otherwise, and experimental mice were separated by sex and housed with four mice per cage. All mice were immune-competent and healthy prior to the experiments described. Mice were bred and maintained at regular housing temperatures (23°C) and 12-hour (h) light/12h dark cycle. In order to generate germline Dnmt3a heterozygous haploinsufficient mice we crossed Dnmt3+/fl mice with Ellacre transgenic mice obtained from The Jackson Laboratories (Bar Harbor, ME). The DNMT3A p.293 deletion and DNMT3A p.W577R mutation were generated in our laboratory utilizing CRISPR strategies with a single guide RNA and single strand DNA template injected into mouse blastocysts.16 Guide sequence and DNA template are described in the Online Supplementary Table S3. For all assays mouse data was obtained at 12 or 15 months of age.

Lymphoblastoid cells generation Peripheral blood mononuclear cells were isolated from whole blood by density gradient centrifugation using Lymphoprep density gradient medium (STEMCELL Technologies, Vancouver, Canada) and then incubated with concentrated Epstein–Barr virus in a total of 200 mL of RPMI medium for 30 minutes. The cells were then plated in a flat-bottomed 96-well plate with 1 mg/mL cyclosporin A (Sandoz Pharmaceuticals, Washington, DC). Cells were fed bi-weekly until lymphoblastoid cell lines (LCL) were established.

Whole genome bisulfite sequencing DNA was extracted from LCL using the DNeasy kit (Qiagen, Hilden, Germany). DNA was fragmented prior to bisulfite conversation using the NEBNext fragmentase kit, following the manufacturer’s protocol (New England BioLabs, Inc., Ipswich, MA). We used 200 ng of DNA to prepare whole genome bisulfite sequencing (WGBS) libraries with the Swift-whole genome bisulfite kit according to the manufacturer’s instruction (Swift Biosciences, Ann Arbor, MI). Libraries were sequenced and processed as previously described.17 The Wildcard ALignment tool (WALT) pipeline was used for duplicate removal and hypomethylated regions (HMR) calling.18 A minimum of five reads per CpG was used to calculate methylation values. Coverage per base for mutant or control cells was >7x. For identification of differentially methylated regions, we used the WALT pipeline default settings.

Statistical analyses Methods Human samples This study was approved by the Institutional Review Board of Baylor College of Medicine. Written informed consent and

888

Statistical analyses of the clinical data were performed with the GraphPad Prism 8 software (GraphPad Software, San Diego, CA). P-values were interpreted as statistically significant if less than 0.05, unless otherwise stated. Throughout the manuscript data are expressed as the mean +/- standard error of the mean, unless otherwise stated. The statistical significance of the

haematologica | 2022; 107(4)

DNMT3A mutations in TBRS

Figure 1. DNMT3A mutations and deletions identified in patients with Tatton-Brown-Rahman syndrome. Map of DNMT3A variants identified in individuals with Tatton-Brown-Rahman syndrome (TBRS) based on clinical sequencing data from our patient cohort, published data, and information obtained from the TBRS community. AA: amino acid; PWWP: Pro-Trp-Trp-Pro motif domain; ADD: ATRX-DNMT3L-DNMT3L domain; Mtase: methyltransferase domain.

differences between two groups was calculated using unpaired Student’s t-test (two-sided, without assuming equal standard deviations), Fisher’s exact test or the nonparametric MannWhitney test where appropriate. Statistical details are also described in the figure legends, including the number of replicates, animals or human samples per group (denoted by “n”).

Results Characteristics of germline DNMT3A mutations in Tatton-Brown-Rahman syndrome In order to identify the current spectrum of TBRS variants, we gathered mutational information on 48 individuals with TBRS utilizing the TBRS organization database. These individuals harbored mutations (missense, frameshift, deletions) in the PWWP, ADD and methyltransferase domains, including ~30 not previously reported in TBRS (Figure 1, Table 1; Online Supplementary Table S1). We also identified two TBRS individuals with deletion of the entire DNMT3A gene. These data show that, as previously suggested, DNMT3A mutations in TBRS can occur across all the DNMT3A functional domains and are shared between TBRS, CH and leukemia.

White blood cell differential of Tatton-Brown-Rahman syndrome individuals is characterized by increased neutrophils and decreased lymphocytes In order to assess the composition of the blood of individuals with TBRS, we enrolled 18 individuals with TBRS, seven unaffected siblings, and four unrelated unaffected individuals (Table 1; Online Supplementary Table S2). The clinical and genetic information for the included TBRS individuals is shown in Table 1. One individual in the TBRS cohort had severe iron deficiency anemia and one had hypogammaglobulinemia requiring intravenous immunoglobulin (IVIG) replacement. No other blood or immune defects were reported in either cohort. We performed complete blood cell counts (CBC) with white blood cell (WBC) differential on the enrolled individuals, excluding from the analysis the TBRS patient with severe iron deficiency anemia. An additional four TBRS individuals and one control did not have a CBC performed, thus 13 TBRS individuals and nine controls were included in our analysis. The age of TBRS individuals (median 15.2 years, range, 3.5-35.8 years) with CBC was not significantly different than controls (median 10.0 years, range, 3.5-40 years). We found that while the total WBC count was not difhaematologica | 2022; 107(4)

ferent between the groups, in TBRS individuals the percentage of neutrophils was significantly increased, and percentage of monocytes and lymphocytes significantly decreased relative to controls (Figure 2A to D). Also, while hemoglobin and red cell counts (RBC) were not significantly different (Figure 2E and F) compared to unaffected individuals, those with TBRS had a significantly increased mean corpuscular volumes (MCV) and mean corpuscular hemoglobin (MCH) levels, with no difference in mean corpuscular hemoglobin concentration (MCHC) (Figure 2G; Online Supplementary Figure 1; Figure 2H, respectively). As MCV varies by age and sex, we plotted the MCV of our enrolled TBRS individuals by age/sex, including past CBC from medical records where available. We found that overall, the vast majority of TBRS individuals had MCV above the 50th percentile at all ages examined including ten of 14 with an MCV ≥90th percentile and eight of 14 >97th percentile for age/sex at one or more timepoints (Figure 2I and J). All other CBC parameters were similar between TBRS and unaffected individuals (Online Supplementary Figure 1). These finding indicate that heterozygous germline DNMT3A lesions impact multilineage hematopoiesis in the absence of overt clinical blood disorders.

Analytical characterization of blood populations in Tatton-Brown-Rahman syndrome individuals We next conducted immunophenotypic analyses of the peripheral blood from 15 TBRS individuals and ten controls. The age of TBRS versus control individuals with immunophenotypic data was not significantly different (median 13.5 years and 10 years, respectively). Our findings confirmed neutrophil expansion in the blood of TBRS individuals (Figure 3A). Given reports of recurrent infections and hypogammaglobulinemia in individuals with TBRS as well as mouse model data suggesting a role for DNMT3A in T-cell development,19 we conducted a detailed immunophenotypic analysis of T- and B-cell subsets. We found a relative reduction in CD19+, CD20+, and CD22+ B cells in TBRS individuals compared to unaffected controls (Figure 3B; Online Supplementary Figure S2A and B). We also identified a trend towards reduced total CD3+ T cells that did not reach statistical significance but found significant differences in the relative populations of CD4+ and CD8+ expressing T cell subsets with a higher CD4 to CD8 ratio in TBRS compared to unaffected individuals (Figure 3C; Online Supplementary Figure 2C). These results confirm a relative increase in neutrophils in individuals with TBRS and identify significant changes in B- and T-cell populations. 889

A. Tovy et al.

Table 1. Table of Tatton-Brown-Rahman syndrome individuals’ clinical and mutation information.

Sex

Age Height Height Weight Weight (years) (cm) percentile (kg) percentile

RR01 RR04 RR07 RR16 RR17 RR18 RR19 RR20 RR22 RR24 RR25 RR26 RR28 RR34 RR38 RR40

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

3.8 21.8 2.6 17.6 34.5 20.8 35.8 16* 3.5 5.7 13.5 16.0 18* 5.9 7.0 15.2

110.4 183 104.8 183 196 196 194 172 99 130 168 180.3 168 n/k 137 193

96 83 >99 84 >99 >99 >99 93 47 >99 78 >99 77 n/k >99 >99

19.9 89 26.2 79.5 79.4 74.4 77.5 68.5 17.7 33.1 83.9 115.5 64 n/k 30 74.5

99 55 >99 84 35 23 30 88 87 >99 99 99 75 n/k 91 89

RR43 RR45

F M

2.0 3.7

93 107

99 92

15 21

97 98

Mutation type

DNMT3A coding DNMT3A sequence variant variant protein

Missense c.901C>T Frameshift deletion c.297del Gene deletion 2p23.3 del Missense c.1748 G>A Frameshift duplication c.1238dupG Missense c.2309 C>T Missense c.2207 G>A Missense c.2063 G>A Missense c.1978 T>C Gene deletion 2p23.3 del Missense c.2645 G>A Missense c.929 T>A Missense c.2114T>C Missense c.2129G Missense c.919C>T Frameshift deletion/ c.2432_2434delinsC insertion Missense c.1627 G>T Missense c.2645 G>A

p.R301W p.M99fs p.C583Y p.F414fs*7 p.S770L p.R736H p.A688H p.Y660H p.R882H p.I310N p.V704A p.C710Y p.P307S

p.G543C p.R882H

Flow Y/N

CBC

Y N Ex Y Y Y Y Y Y Y Y Y N Y Y Y

Y Y

N N

Blood and medical records we collected from 18 Tatton-Brown-Rahman syndrome (TBRS) individuals. Presented is mutation data and the type of analyses that were done on each sample. *: approximate; n/k: not known; Flow: flow cytometric analysis of peripheral blood; CBC: complete blood cell count with white blood cell differential; Y: yes; N: no; Ex: excluded due to severe iron deficiency anemia.

Myeloid cells expand in the peripheral blood of mice with germline Dnmt3a mutations In order to validate the findings of our human cohort, we examined hematopoiesis in three mouse models with constitutive Dnmt3a lesions. We first evaluated mice with a heterozygous germline deletion affecting amino acid 293 (HET293) (mouse equivalent of human amino acid 297) in the PWWP domain of DNMT3A, a lesion previously reported in TBRS.1,2 Consistent with our human data, CBC revealed significantly increased percentage of neutrophils and decreased lymphocytes compared to wildtype littermate controls (WT) without a significant difference in the overall WBC count (Figure 4A to C; Online Supplementary Figure S3A). In order to evaluate for potential phenotype-modifying effects of differing Dnmt3a mutations on hematopoiesis, we also analyzed the peripheral blood of mice with heterozygous germline mutations affecting the DNMT3A ADD domain amino acid W577R (HET577) (mouse equivalent of human W581R), a mutation which has been reported in CH,20 and mice with heterozygous deletion of Dnmt3a (HET), recapitulating a complete loss of function or deletion of one DNMT3A copy. At 1 year of age, we compared the peripheral blood of HET577 and HET mice to their WT littermates and found that the percentage of neutrophils was significantly increased whereas the percentage of lymphocytes was significantly decreased (Online Supplementary Figure 3B and C). HET577 mice also had significantly decreased total WBC and decreased platelets, unlike the HET293 and HET mice and the TBRS cohort. Flow cytometric immunophenotypic analysis of peripheral blood leukocytes at 12 and 15 months demonstrated significant myeloid expansion and B-cell reduction in the HET293 mice and HET577 (Figure 4D to F; Online 890

Supplementary Figure S4A and B, respectively) consistent with findings in our TBRS cohort. Like TBRS individuals, the percentage of Ly6G+ neutrophils in HET293 mice was significantly increased relative to littermate controls with no significant difference in the percentage of Ly6C+ monocytes (Figure 4G). Together, these results demonstrate that multiple constitutive Dnmt3a mutant mouse models recapitulate the hematopoietic phenotypes observed in TBRS. Previous publications have shown that the key inflammatory cytokine, interleukin 6 (IL6), is abnormally elevated in DNMT3A-mediated CH.21 We therefore postulated that the neutrophil expansion identified in TBRS and our murine models may be in response to inflammation driven by mutation or deletion of Dnmt3a. In order to explore this possibility, we measured IL6 levels in the serum of HET293 mice and WT littermates at 12 months and in the TBRS individuals. Although in the young TBRS individuals IL6 was not altered, in HET293 mice we measured significantly elevated levels of IL6 (Online Supplementary Figure 4C and D). Our findings suggest that germline Dnmt3a mutations are consistently associated with neutrophil expansion, potentially in response to heightened inflammation.

Stem and myeloid cells expand in bone marrow of aged germline Dnmt3a-mutant mice Knockout (KO) of Dnmt3a in hematopoietic stem cells has been widely investigated in bone marrow transplantation experiments, in which transplant of Dnmt3a-KO HSC leads to expansion of the stem cell pool and increased repopulation capability.22 The impact of DNMT3A mutation or loss on hematopoietic stem and progenitor cell populations in an unperturbed model such as TBRS has not been fully investigated. In order to address this we next analyzed the composition of hematopoietic cells in haematologica | 2022; 107(4)

DNMT3A mutations in TBRS

Figure 2. Blood of Tatton-Brown-Rahman syndrome individuals is characterized by relative increase in neutrophils, decrease in lymphocytes and non-anemic macrocytosis. Complete blood cell counts were performed on peripheral blood from Tatton-Brown-Rahman syndrome (TBRS) individuals (n=13) and unaffected controls (n=9) including (A) total white blood cell (WBC) count and WBC differential with percent of (B) neutrophils, (C) monocytes, and (D) lymphocytes. Red blood cell (RBC) indices were also compared including (E) hemoglobin, (F) total RBC number, (G) mean corpuscular volume (MCV), and (H) mean corpuscular hemoglobin concentration (MCHC). One TBRS individual had no available RBC count. MCV plotted by age for (I) TBRS males and (J) TBRS females including data from study CBC and from CBC data extracted from medical records for some individuals over multiple timepoints.

the bone marrow of HET293 mice compared to WT littermate controls. Like the blood, the bone marrow of HET293 mice showed relative myeloid expansion and reduced frequency of B-cells (Figure 4 H to J). Comparison of the stem/progenitor cell compartment (gating scheme in the Online Supplementary Figure S4E and F) in 15-monthold mice without any overt hematologic malignancies, showed a moderate but significant expansion of hematopoietic stem cells and multipotent progenitor cells (Figure 4K and L). The relative frequency of other stem/progenitor populations, including common myeloid progenitors, common lymphoid progenitors, granulocytes/monocyte progenitors, megakaryocyte/erythroid progenitors did not differ significantly (Online Supplementary Figure 4G and H). haematologica | 2022; 107(4)

These finding support that the germline mutation of Dnmt3a, like somatic loss of Dnmt3a in the hematopoietic compartment, results in stem cell expansion with aging and leads to myeloid expansion in the bone marrow.3

Germline Dnmt3a-mutant mice have defects in lymphocytes production Evaluation of T- and B-cell subsets in the blood of TBRS individuals revealed significantly reduced CD19+ B cells and an altered ratio of CD4/CD8 T cells in those with TBRS. We therefore performed flow cytometric analysis of B- and T-cell subsets in the peripheral blood of Dnmt3a-mutant mice. We found that unlike TBRS individuals, in HET293 and HET577 mice the CD4/CD8 T-cell ratio was significantly lower than WT controls (Online 891

A. Tovy et al. A

Figure 3. Immunophenotypic analysis of Tatton-Brown-Rahman syndrome individuals identifies neutrophil expansion and deficiencies in specific T- and B- cell subsets. Flow cytometry analysis was performed on peripheral blood from Tatton-Brown-Rahman syndrome (TBRS) individuals (n=15) and unaffected controls (n=10). Percent of leukocytes categorized by immunophenotype as (A) neutrophils, (B) B cells expressing CD10, CD19, CD20 and/or CD22, (C) total CD3+ T cells and the percentage of CD3+ T cells that are CD4+ and CD8+. (D) Quantification of the CD4/CD8 ratio of controls versus TBRS individuals.

Supplementary Figure S5A). Quantification of B-cell subsets by flow cytometry demonstrated that although HET293 mice had a relative decrease in total B220+ B cells, the proportion of different B-cell subsets was not altered when compared to WT mice (Online Supplementary Figure S5B). Only mature B lymphocytes can effectively contribute to the immune response, and maturation of B-cells that occurs in the spleen. Therefore, in order to determine if the B-cell reduction reflects defects in splenic B-cell maturation we analyzed the frequency of transitional immature B cells in the spleen by flow cytometry (T1 and T2 populations, respectively).23 T1 B cells are bone marrow derived immature B cells that migrate to the spleen where they mature into T2 B cells which are the progenitors for mature B cells.24 We identified a significant decrease in the frequency of T1 B cells but not of T2 B cells in the spleen of HET293 mice compared to WT controls (Figure 5A and B). These data suggest that DNMT3A loss leads to a reduction in B-cell frequency via decreased B-cell progenitors in the bone marrow. While reduced in numbers, these progenitors are capable of normal maturation.

Germline Dnmt3a-mutant mice have defects in erythropoiesis Evaluation of RBC indices in our TBRS cohort revealed increased average MCV and MCH relative to unaffected individuals. This finding was recapitulated in the HET293, HET577, and HET murine models (Figure 5D and E; Online Supplementary Figure S3), although unlike the TBRS individuals, HET293 mice additionally had significantly 892

decreased RBC counts compared to WT controls (Figure 5C to E). In order to further characterize erythropoiesis in TBRS, we performed flow cytometric examination of erythroid development in the bone marrow of HET293 mice.25 Staining of bone marrow cells with CD71 and TER119 is used to identify developmental stages of erythropoiesis, labeled ProE (proerythroblasts), Erythroblasts A-C (EryA, EryB and EryC), which correspond to sequential steps in erythroid development.26 Although the total percentage of TER119 positive cells was not significantly different in HET293 mice compared to WT littermates, HET293 mice had significantly fewer large, immature erythroblasts (EryA) and more small mature erythroblasts (EryC) (Figure 5F and G). Thus, loss of DNMT3A in our murine model leads to a subtle, but significant effect on early erythroid development and differentiation.

Dnmt3a-mutant mice develop hematologic malignancies at higher rates than controls Given the frequency of somatic DNMT3A mutations in adult hematologic malignancies, as well as case reports documenting hematologic malignancies in individuals with TBRS,1,14 we looked for malignancies in cohorts of HET293 mice at the age of 15 months. At this age, eight of 36 (22%) HET293 mice had malignancies compared to one of 35 (3%) WT littermate controls (P=0.028). Of the eight malignancies in the HET293 mice, seven were hematologic including both myeloid and lymphoid diseases (Online Supplementary Table S7; Online Supplementary Figure 6 A and B) consistent with results from mice with heterozygous haematologica | 2022; 107(4)

DNMT3A mutations in TBRS

Figure 4. Tatton-Brown-Rahman syndrome mouse model characterized by myeloid expansion and increased frequency of hematopoietic stem and progenitor cells in the marrow. Complete blood cell counts performed on the blood of a representative cohort of mice with heterozygous in frame deletion of amino acid 293 of DNMT3A (HET293) (n=23) and wild-type (WT) littermate controls (n=31). All mice included in the analyses did not display hematologic malignancies. Displayed comparisons of (A) total white blood cell (WBC) count, (B) percentage of neutrophils and (C) percentage of lymphocytes. Flow cytometric analysis of peripheral blood CD45+ leukocytes depicting (D) relative distribution of myeloid (defined as cells expressing CD11b and/or Ly6G), T cells (defined as CD3 and CD4+ and/or CD8+ cells) and B cells (defined as B220+ cells) in the HET293 mice compared to WT. Quantification of the percentage of (E) myeloid and (F) B cells in the HET293 mice and WT mice as determined by flow cytometry from (D). (G) Analysis of the different subtypes of CD11b+ myeloid cells into neutrophils (Ly6G expressing cells) or monocytes (Ly6C expressing cells). Flow cytometric analysis of bone marrow CD45+ leukocytes depicting H) relative distribution of myeloid, T cells and B cells in the HET293 mice compared to WT. Quantification of the percentage of I) myeloid and J) B cells in the HET293 mice and WT mice as determined by flow cytometry. Bone marrow flow cytometry assessment of HET293 and WT mice showing the percent of (K) hematopoietic stem/progenitor cells defined by SLAM markers and (L) multipotent progenitor (MPP) cells.

haematologica | 2022; 107(4)

893

A. Tovy et al.

germline deletion of Dnmt3a.27 The affected mice were moribund, with weight loss, pallor, adenopathy, tumors and splenomegaly (Figure 6A). Two mice had malignant myeloid neoplasms: one acute myeloid leukemia and one myeloproliferative neoplasm with infiltration of malignant cell into blood, bone marrow, liver and spleen. Four mice

had lymphomas including B- (n=3) and T-cell (n=1) malignancies (Figure 6B). We also identified histolytic sarcoma in two mice, both of which had an additional malignancy (one with lymphoma and one with angiosarcoma). Two mice had angiosarcomas, a non-hematological malignancy (Online Supplementaty Figure S7).

Figure 5. Differences in lymphoid and erythroid compartments in Tatton-Brown-Rahman syndrome mouse model. Quantification of the proportion of B220+ splenic B cells that are (A) T1 B cells expressing immunglobulin (Ig) M and intermediate levels of IgD and (B) T2 B cells expressing both IgM and IgD in HET293 mice (n=19) and wild-type (WT) littermates (n=20). From complete blood cell counts performed on peripheral blood, comparison of (C) red blood cell (RBC) number, (D) mean corpuscular volume (MCV), and (E) mean corpuscular hemoglobin concentration (MCHC) of WT mice (n=29) and HET293 mice (n=24). (F) Flow cytometric gating strategy for assessment of the erythroid development for representative WT and HET293 mice. Viable cells were gated based on their TER119 expression and then TER119+ cells were plotted by forward scatter and CD71 expression levels to identify erythroblasts in different developmental stages (Ery A-C). (G) Top: the total proportion of TER119 and proerythroblasts (ProE) cells. Bottom: the proportion of the indicated populations within the TER119+ fraction defined by CD71 and FSC (n=18) and HET293 mice (n=16).

894

haematologica | 2022; 107(4)

DNMT3A mutations in TBRS

These data, combined with case reports of individuals with TBRS and hematologic cancers, suggest that while the presence of a germline variant of DNMT3A increases the relative risk for the development of hematologic malignancies, a majority will not develop a blood cancer.

Constitutive loss of DNMT3A in Tatton-Brown-Rahman syndrome leads to significant hypomethylation in hematopoietic cells In TBRS individuals only one functional allele of DNMT3A remains. In order to examine if constitutive loss of DNMT3A in TBRS impacts the DNA methylation landscape, we performed whole genome bisulfite sequencing on LCL derived from B cells of a TBRS individual with a DNMT3A-297deletion (297del) and from a sibling of a

TBRS individual with WT DNMT3A. We measured a 6% decrease in global DNA methylation in the 297del LCL compared to WT LCL (60.34% and 66.77%, respectively) (Figure 7A). When we analyzed the distribution of DNA methylation in the WT LCL compared to the 297del LCL, we observed a significant decrease in DNA methylation in enhancer regions. This observation suggests that, similar to leukemia,28 in TBRS hematopoietic cells DNMT3A loss impacts methylation at regulatory enhancer regions (Figure 7B). We identified 1,068 differentially methylated regions (DMR) in 297del compared to WT LCL (Online Supplementary Table S8). Interestingly, some of these DMR include key loci with known importance in blood development such as the HOXA cluster, similar to the hypomethylation patterns we previously reported for

Figure 6. Development of hematologic malignancies in Tatton-Brown-Rahman syndrome mouse model. A subset of the HET293 mice had hematologic malignancies at 15 months of age. These mice were noted to have (A) enlarged spleens relative to wild-type (WT) and HET293 mice without malignancies. (B) Pathologic evaluation of the bone marrow, spleen and liver of a WT mouse (WT #1, top row), a HET293 mouse with acute myeloid leukemia noted in the bone marrow, spleen and liver (HET293 2b, second row), and a HET293 with T-cell lymphoma in the bone marrow, spleen and liver (HET293 #40, third row) and focal histiocytic sarcoma in the bone marrow (HET293 #40, bottom image).

haematologica | 2022; 107(4)

895

A. Tovy et al.

LCLs generated from an individual with a constitutive DNMT3A 771Q mutation17 (Figure 7C). These results support that the heterozygous loss of DNMT3A leads to hematopoietic defects likely through altered DNA methylation.

Discussion In order to examine the impact of constitutive loss of one copy of DNMT3A on blood production, we analyzed primary human specimens from patients with TBRS and murine models recapitulating pathogenic Dnmt3a variants identified in TBRS, CH and hematologic malignancies. Our data reveal a shift in the distribution of leukocytes with an overall increase in the myeloid compartment, specifically neutrophils, in individuals with TBRS compared to controls. Further, we identified a significant reduction in the percentage of lymphocytes and specific B- and T-cell subsets. We also noted erythropoiesis defects in TBRS, manifested as increased MCV. We further found that our TBRS murine models developed hematologic malignancies of low penetrance after a long latency. We also identified differences in the blood parameters of mice with differing Dnmt3a lesions (PWWP domain, ADD domain, or complete deletion of one allele), raising the possibility that mutation-specific effects on DNMT3A function may modify blood phenotypes. This work demonstrates that constitutive heterozygous loss of one DNMT3A allele leads to significant multilineage perturbations of hematopoiesis. Our findings have potential clinical implications for individuals with TBRS. In our TBRS cohort and all examined TBRS mouse models, we found significantly reduced B cells relative to controls. Additionally, one of our enrolled patients reported hypogammaglobulinemia requiring IVIG supplementation. Therefore, our observations suggest that evaluation of immunoglobulin levels may be warranted for individuals with TBRS, particularly those with recurrent infections. We also identified significant differences in T-cell subsets between TBRS indi896

Figure 7. Altered DNA methylation in enhancer regions of hematopoietic cells of a Tatton-Brown-Rahman syndrome individual. Whole genome bisulfite sequencing was performed on a lymphoblastoid cell line (LCL) derived from a Tatton-Brown-Rahman syndrome (TBRS) individual with a heterozygous deletion of amino acid 297 in the PWWP domain of DNMT3A (297 del) and an LCL derived from the unaffected sibling of an individual with TBRS (control). (A) Overall % CpG methylation of control and 297 del LCL. (B) Plot represents density of DNA CpG methylation of enhancer regions in control and 297 del LCL. (C) Genome browser tracks of CpG methylation at the HOXA locus of the control and 297 del LCL compared to a previously published DNMT3A p.R771Q mutant LCL.

viduals and controls. In our TBRS cohort, the CD4/CD8 ratio was increased, a finding that has been associated with obesity which is noted in most of our TBRS cohort.29 Conversely, in our murine models, we found a decreased CD4/CD8 ratio in mice with Dnmt3a mutation or deletion. There are known differences in lymphoid development between human and mouse that may explain this discrepancy, but further investigations will be needed to fully understand the impact of DNMT3A loss on T-cell development.30 Lower CD4/CD8 ratios have been associated with altered immune responses and inflammation; perhaps this relative change might explain the increased level of inflammatory cytokine IL6 in our aged mutant mice.31 Overall, the clinical consequences of these immune cell phenotypes will require additional evaluation and longitudinal follow-up of affected individuals with a focus on how these B- and Tcell deficits impact susceptibility to infections and response to vaccinations. In our cohort, we found increased MCV in TBRS individuals relative to controls. This finding is similar to the nonanemic macrocytosis characteristic of Down syndrome (DS) and Williams-Beuren syndrome (WBS).32,33 It is likely that in TBRS, like in DS and WBS, isolated macrocytosis is a benign variant of normal, warranting conservative observation only. Indeed, multiple individuals in our cohort had MCV consistently at or above the 90th percentile for sex/age over many years yet have not developed any hematologic disorders. It is notable that increased red cell distribution width (RDW), reflecting variability in RBC size, is a consistent finding in CH.9,21 Our data suggest that DNMT3A-mutant HSPC generate larger RBC than RBC derived from non-mutant HSPC. Therefore, in DNMT3Amutant CH, increased RDW may be attributable to a mixed population of WT and DNMT3A-mutant cells in the peripheral blood. An expansion of the DNMT3A-mutant HSPC would therefore be expected to result in an increased MCV due to increased representation of DNMT3A-mutant RBC in the blood. Indeed, increased MCV in individuals with CH has been associated with increased risk of hematologic malignancy development.34 haematologica | 2022; 107(4)

DNMT3A mutations in TBRS

Unlike the TBRS patients, the aged HET293 mice were noted to have a modest, but significant decrease in RBC count compared to littermate controls. This may reflect a feature associated with aging and warrants further longitudinal studies on TBRS individuals to definitively determine changes in erythropoiesis over time. In our TBRS cohort and murine models, we also identified neutrophil expansion, potentially in response to inflammation. CH is associated with increased risk not only for hematological malignancies, but also for cardiovascular disease, largely due to a pathologic inflammatory state.35 Previous studies have shown that loss in hematopoietic cells of Dnmt3a or Tet2, another commonly mutated gene in CH, leads to increased expression of inflammatory chemokines and cytokines and accelerated atherosclerosis.21,36,37 Our cohort of TBRS individuals relative to unaffected controls did not show significantly increased IL6. However, it should be noted that our TBRS cohort included only children and young adults and it is possible an increase in inflammatory cytokines detectable by the relatively insensitive enzyme-linked immunosorbant assay may only be apparent with aging. If confirmed, increased inflammation with aging in TBRS individuals may elevate the risk of atherosclerotic cardiovascular disease, warranting lifetime monitoring. This monitoring may be of particular importance given that TBRS is associated with congenital cardiac defects.1,37 Given the association between somatic DNMT3A mutations and leukemia, malignancy risk is a major concern for TBRS individuals. While there are case reports in the literature of TBRS individuals developing hematologic malignancies, the absolute risk is currently undefined. In this study we found that unperturbed HET293 mice developed spontaneous hematologic malignancies at higher rates than WT mice, consistent with constitutive Dnmt3a-HET and previous studies using complete somatic Dnmt3a deletion.27,38,39 We also identified malignancies which are not often associated with DNMT3A-mutations and not reported to date in TBRS; histolytic sarcoma and angiosarcoma. Interestingly, pathologic examination of non-moribund aged HET293 mice revealed extramedullary hematopoiesis (EMH) involving spleen, liver and kidney in a subset of mice (Online Supplementary Figure S7). EMH can be a feature of myeloproliferative neoplasm.40 However, in our TBRS models this finding is of uncertain significance as the affected HET293 mice were healthy without significant blood count abnormalities. Overall, our findings support a possible increased risk for hematologic malignancies in individuals with TBRS. Thus, TBRS individuals and their families should be educated about the signs and symptoms of leukemia that would warrant medical evaluation. However, the relatively low penetrance of malignancy in our mouse model, even at advanced ages, indicates that leukemia is

References 1. Tatton-Brown K, Zachariou A, Loveday C, et al. The Tatton-Brown-Rahman syndrome: a clinical study of 55 individuals with de novo constitutive DNMT3A variants. Wellcome Open Res. 2018;3:46. 2. Tatton-Brown K, Seal S, Ruark E, et al. Mutations in the DNA methyltransferase gene DNMT3A cause an overgrowth syn-

haematologica | 2022; 107(4)

not an inevitable consequence of having a constitutive DNMT3A mutation. Indeed, our previous report demonstrated that, although loss of DNMT3A leads to a significant competitive advantage in the blood, this advantage is not necessarily associated with leukemogenesis.17 Natural history studies of TBRS individuals are needed to define the incidence of hematologic malignancy, the disease spectrum, and risk predictors for malignancy development to enable tailored surveillance guidelines for prospective monitoring. In conclusion, we show that constitutive loss of DNMT3A significantly impacts multilineage blood development and leads to phenotypic changes with clinical implications. While our findings offer important insights into blood development in individuals with TBRS, these results may also have broader implications. The DNMT3A mutations in TBRS and CH largely overlap and, therefore, TBRS may offer a unique opportunity to address the impact of DNMT3A mutations on hematopoiesis in a state mimicking accelerated CH. Disclosures No conflicts of interest to disclose. Contributions AT, MAG and RER conceived the project, discussed and designed experiments; AT performed experiments and analysis with assistance of CR, GM, LZ, KK, SEC, AGG, RA and CWC; AGS, AG and ANM performed flow cytometry analysis on TBRS individuals; ANM and MJH performed all the pathological analysis in this manuscript; YHH, TA and AT generated the mouse models for all studies; JJK, MTC, ALA, IV, RS, LVM and LB assisted with collection of blood and medical data from TBRS patients; JMR conducted the bioinformatic analyses in the manuscript; AT, MAG and RER wrote the manuscript; all authors interpreted the results and edited the manuscript. Acknowledgments We thank our research participants and their families for their contributions and active involvement. We thank the TBRS Community for their kind help in supporting this project. We especially wish to thank the TBRS Community board of directors, particularly Jill Kiernan and Kerry Grens. We also thank C. Gillespie for critical review. Funding This project was supported by the Baylor College of Medicine’s Human Stem Cell Core and Baylor College of Medicine’s Cytometry and Cell Sorting Core, which are funded in part by the institution and the NIH CA125123, OD028591, AI036211, A125123, RR024574, DK092833, CA183252, K08CA201611 and the HHMI James H. Gilliam Fellowships. Publication costs were generously supported by the Texas Children's Hospital Young Investigators Endowed Fund.

drome with intellectual disability. Nat Genet. 2014;46(4):385-388. 3. Challen GA, Sun D, Jeong M, et al. Dnmt3a is essential for hematopoietic stem cell differentiation. Nat Genet. 2011;44(1):23-31. 4. Yang L, Rau R, Goodell MA. DNMT3A in haematological malignancies. Nat Rev Cancer. 2015;15(3):152-165. 5. Ley TJ, Ding L, Walter MJ, et al. DNMT3A mutations in acute myeloid leukemia. N Engl J Med. 2010;363(25):2424-2433.

6. Brunetti L, Gundry MC, Goodell MA. DNMT3A in leukemia. Cold Spring Harb Perspect Med. 2017;7(2):a030320. 7. Genovese G, Kahler AK, Handsaker RE, et al. Clonal hematopoiesis and blood-cancer risk inferred from blood DNA sequence. N Engl J Med. 2014;371(26):2477-2487. 8. 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.

897

A. Tovy et al. 9. Abelson S, Collord G, Ng SWK, et al. Prediction of acute myeloid leukaemia risk in healthy individuals. Nature. 2018; 559(7714):400-404. 10. Bolouri H, Farrar JE, Triche T, Jr., et al. The molecular landscape of pediatric acute myeloid leukemia reveals recurrent structural alterations and age-specific mutational interactions. Nat Med. 2018;24(1):103112. 11. Ho PA, Kutny MA, Alonzo TA, et al. Leukemic mutations in the methylationassociated genes DNMT3A and IDH2 are rare events in pediatric AML: a report from the Children's Oncology Group. Pediatr Blood Cancer. 2011;57(2):204-209. 12. Liu Y, Easton J, Shao Y, et al. The genomic landscape of pediatric and young adult T-lineage acute lymphoblastic leukemia. Nat Genet. 2017;49(8):1211-1218. 13. Zaliova M, Stuchly J, Winkowska L, et al. Genomic landscape of pediatric B-other acute lymphoblastic leukemia in a consecutive European cohort. Haematologica. 2019;104(7):1396-1406. 14. Hollink I, van den Ouweland AMW, Beverloo HB, Arentsen-Peters S, Zwaan CM, Wagner A. Acute myeloid leukaemia in a case with Tatton-Brown-Rahman syndrome: the peculiar DNMT3A R882 mutation. J Med Genet. 2017;54(12):805-808. 15. Community T. 2020 [cited; Available from: https://tbrsyndrome.org/ 16. Huang Y, Tovy A, Sundaramurthy V, et al. Nearly a third of clonal hematopoiesisassociated DNMT3A mutations reduce protein stability and may be associated with poorer prognosis. Blood. 2018; 132(Suppl 1):S1315. 17. Tovy A, Reyes JM, Gundry MC, et al. Tissue-biased expansion of DNMT3Amutant clones in a mosaic individual is associated with conserved epigenetic erosion. Cell Stem Cell. 2020;27(2):326-335. 18. Chen H, Smith AD, Chen T. WALT: fast and accurate read mapping for bisulfite sequencing. Bioinformatics. 2016; 32(22):3507-3509. 19. Ladle BH, Li KP, Phillips MJ, et al. De novo DNA methylation by DNA methyltrans-

898

ferase 3a controls early effector CD8+ Tcell fate decisions following activation. Proc Natl Acad Sci U S A. 2016;113(38):1063110636. 20. Watson CJ, Papula AL, Poon GYP, et al. The evolutionary dynamics and fitness landscape of clonal hematopoiesis. Science. 2020;367(6485):1449-1454. 21. Bick AG, Weinstock JS, Nandakumar SK, et al. Inherited causes of clonal haematopoiesis in 97,691 whole genomes. Nature. 2020;586(7831):763-768. 22. Jeong M, Park HJ, Celik H, et al. Loss of Dnmt3a immortalizes hematopoietic stem cells in vivo. Cell Rep. 2018;23(1):1-10. 23. Petro JB, Gerstein RM, Lowe J, Carter RS, Shinners N, Khan WN. Transitional type 1 and 2 B lymphocyte subsets are differentially responsive to antigen receptor signaling. J Biol Chem. 2002;277(50):4800948019. 24. Loder F, Mutschler B, Ray RJ, et al. B cell development in the spleen takes place in discrete steps and is determined by the quality of B cell receptor-derived signals. J Exp Med. 1999;190(1):75-89. 25. Socolovsky M, Nam H, Fleming MD, Haase VH, Brugnara C, Lodish HF. Ineffective erythropoiesis in Stat5a(-/-)5b (-/-) mice due to decreased survival of early erythroblasts. Blood. 2001;98(12):32613273. 26. Koulnis M, Pop R, Porpiglia E, Shearstone JR, Hidalgo D, Socolovsky M. Identification and analysis of mouse erythroid progenitors using the CD71/TER119 flow-cytometric assay. J Vis Exp. 2011;(54):2809. 27. Cole CB, Russler-Germain DA, Ketkar S, et al. Haploinsufficiency for DNA methyltransferase 3A predisposes hematopoietic cells to myeloid malignancies. J Clin Invest. 2017;127(10):3657-3674. 28. Spencer DH, Russler-Germain DA, Ketkar S, et al. CpG island hypermethylation mediated by DNMT3A is a consequence of AML progression. Cell. 2017;168(5):801816. 29. van der Weerd K, Dik WA, Schrijver B, et al. Morbidly obese human subjects have

increased peripheral blood CD4+ T cells with skewing toward a Treg- and Th2dominated phenotype. Diabetes. 2012;61(2):401-408. 30. Kumar BV, Connors TJ, Farber DL. Human T cell development, localization, and function throughout life. Immunity. 2018; 48(2):202-213. 31. McBride JA, Striker R. Imbalance in the game of T cells: what can the CD4/CD8 Tcell ratio tell us about HIV and health? PLoS Pathog. 2017;13(11):e1006624. 32. Wachtel TJ, Pueschel SM. Macrocytosis in Down syndrome. Am J Ment Retard. 1991;95(4):417-420. 33. Yu E, Feinn R, Bona R, et al. Mild macrocytosis in Williams-Beuren syndrome. Eur J Med Genet. 2020;63(3):103740. 34. Sperling AS, Gibson CJ, Ebert BL. The genetics of myelodysplastic syndrome: from clonal haematopoiesis to secondary leukaemia. Nat Rev Cancer. 2017;17(1):519. 35. Steensma DP, Ebert BL. Clonal hematopoiesis as a model for premalignant changes during aging. Exp Hematol. 2020; 83:48-56. 36. Fuster JJ, MacLauchlan S, Zuriaga MA, et al. Clonal hematopoiesis associated with TET2 deficiency accelerates atherosclerosis development in mice. Science. 2017;355 (6327):842-847. 37. Jaiswal S, Natarajan P, Ebert BL. Clonal hematopoiesis and atherosclerosis. N Engl J Med. 2017;377(14):1401-1402. 38. Celik H, Mallaney C, Kothari A, et al. Enforced differentiation of Dnmt3a-null bone marrow leads to failure with c-Kit mutations driving leukemic transformation. Blood. 2015;125(4):619-628. 39. Mayle A, Yang L, Rodriguez B, et al. Dnmt3a loss predisposes murine hematopoietic stem cells to malignant transformation. Blood. 2015;125(4):629638. 40. Fan N, Lavu S, Hanson CA, Tefferi A. Extramedullary hematopoiesis in the absence of myeloproliferative neoplasm: Mayo Clinic case series of 309 patients. Blood Cancer J. 2018;8(12):119.

haematologica | 2022; 107(4)

ARTICLE

Hodgkin Lymphoma

Improved outcomes of high-risk relapsed Hodgkin lymphoma patients after high-dose chemotherapy: a 15-year analysis Yago Nieto,1 Stephen Gruschkus,2 Benigno C. Valdez,1 Roy B. Jones,1 Paolo Anderlini,1 Chitra Hosing,1 Uday Popat,1 Muzaffar Qazilbash,1 Partow Kebriaei,1 Amin Alousi,1 Neeraj Saini,1 Samer Srour,1 Katayoun Rezvani,1 Jeremy Ramdial,1 Melissa Barnett,1 Alison Gulbis,3 Terri Lynn Shigle,3 Sairah Ahmed,4 Swaminathan Iyer,4 Hun Lee,4 Ranjit Nair,4 Simrit Parmar,4 Raphael Steiner,4 Bouthaina Dabaja,5 Chelsea Pinnix,5 Jillian Gunther,5 Branko Cuglievan,6 Kris Mahadeo,6 Sajad Khazal,6 Hubert Chuang,7 Richard Champlin,1 Elizabeth J. Shpall1 and Borje S. Andersson1

Ferrata Storti Foundation

Haematologica 2022 Volume 107(4):899-908

1 Department of Stem Cell Transplantation and Cellular Therapy, University of Texas MD Anderson Cancer Center; 2Biostatistics, University of Texas MD Anderson Cancer Center; 3Pharmacy, University of Texas MD Anderson Cancer Center; 4Lymphoma and Myeloma, University of Texas MD Anderson Cancer Center; 5Radiation Oncology, University of Texas MD Anderson Cancer Center; 6Pediatrics, University of Texas MD Anderson Cancer Center and 7Nuclear Medicine, University of Texas MD Anderson Cancer Center, Houston, TX, USA

ABSTRACT

igh-dose chemotherapy and autologous stem-cell transplant (HDC/ASCT) is standard treatment for chemosensitive relapsed classical Hodgkin lymphoma, although outcomes of high-risk relapse (HRR) patients remain suboptimal. We retrospectively analyzed all HRR classical Hodgkin lymphoma patients treated with HDC/ASCT at our institution between 01/01/2005 and 12/31/2019. HRR criteria included primary refractory disease/relapse within 1 year, extranodal extension, B symptoms, requiring more than one salvage line, or positron emission tomography (PET)-positive disease at ASCT. All patients met the same ASCT eligibility criteria. We treated 501 patients with BEAM (n=146), busulphan/melphalan (BuMel) (n=38), gemcitabine(Gem)/BuMel (n=189) and vorinostat/Gem/BuMel (n=128). The Gem/BuMel and vorinostat/Gem/BuMel cohorts had more HRR criteria and more patients with PET-positive disease at ASCT. Treatment with brentuximab vedotin (BV) or anti-PD1 prior to ASCT, PET-negative disease at ASCT, and maintenance BV increased over time. BEAM and BuMel predominated in earlier years (2005-2007), GemBuMel and BEAM in middle years (2008-2015), and vorinostat/GemBuMel and BEAM in later years (2016-2019). The median follow-up is 50 months (range, 6-186). Outcomes improved over time, with 2-year progressionfree survival (PFS)/overall survival (OS) rates of 58%/82% (2005-2007), 59%/83% (2008-2011), 71%/94% (2012-2015) and 86%/99% (20162019) (P<0.0001). Five-year PFS/OS rates were 72%/87% after vorinostat/GemBuMel, 55%/75% after GemBuMel, 45%/61% after BEAM, and 39%/57% after BuMel (PFS: P=0.0003; OS: P<0.0001). These differences persisted within the PET-negative and PET-positive subgroups. Prior BV and vorinostat/GemBuMel were independent predictors of more favorable outcome, whereas primary refractory disease, ≥2 salvage lines, bulky relapse, B symptoms and PET-positivity at ASCT correlated independently with unfavorable outcomes. In conclusion, post-HDC/ASCT outcomes of patients with HRR classic Hodgkin lymphoma have improved over the last 15 years. Pre-ASCT BV treatment and optimized synergistic HDC (vorinostat/GemBuMel) were associated with this improvement.

haematologica | 2022; 107(4)

Correspondence: YAGO NIETO ynieto@mdanderson.org Received: January 5, 2021. Accepted: April 22, 2021. Pre-published: May 6, 2021. https://doi.org/10.3324/haematol.2021.278311

899

Y. Nieto et al.

Introduction High-dose chemotherapy (HDC) with autologous stem-cell transplant (ASCT) is standard treatment of relapsed classical Hodgkin lymphoma (cHL).1,2 Adverse predictors of post-ASCT outcome include primary refractory disease, short first complete remissions, extranodal extension, bulky lesions or B symptoms at the time of relapse, performance status ≥1 at relapse, relapse within a prior radiation field, requirement for more than one line of salvage chemotherapy and, particularly, the persistence of metabolically active tumor on pre-HDC positron emission tomography (PET).3-5 The BEAM regimen (carmustine/etoposide/cytarabine/ melphalan) has long been the standard HDC combination for cHL despite its suboptimal results in patients with highrisk relapses (HRR), whose long-term progression-free survival (PFS) rate is around 50%.3,5,6 Efforts to improve ASCT outcomes have focused mainly in the pretransplant and posttransplant settings. Pretransplant PET-guided use of non-cross-resistant chemotherapy and incorporation of novel drugs, such as brentuximab vedotin (BV), seem to improve results.7,8 In the post-ASCT setting, a randomized trial of maintenance BV after BEAM for HRR cHL showed improvement of 5-year PFS from 41% to 59% as compared to that with placebo.9,10 In contrast, despite clearly suboptimal results obtained with BEAM, little effort has gone into developing a more efficacious HDC program, save for the notable exceptions of attempts to deliver HDC in a tandem fashion.11-13 We have systematically sought to develop more effective HDC regimens based on investigations of synergistic interactions between its components. We started with a combination of pharmacokinetically-guided intravenous busulfan with melphalan (BuMel), which was as safe as BEAM but appeared not to be more effective.14 Following the demonstration of marked preclinical synergy between gemcitabine and BuMel, we next investigated clinically the GemBuMel combination.15,16 Finally, our preclinical work on epigenetic modulation, the synergistic interactions between nucleoside analogs and bifunctional DNA-alkylating agents used in HDC17 led us to clinically test vorinostat/GemBuMel18 and azacytidine/vorinostat/ GemBuMel.19 We herein report our experience with HDC with ASCT for HRR cHL over the last 15 years, analyzing patient-, tumor-, and treatment-related factors (pre-ASCT, HDC regimens, and post-ASCT) associated with outcome.

Methods We retrospectively analyzed all patients with HRR cHL treated at MD Anderson Cancer Center with HDC and ASCT between 01/01/2005 and 12/31/2019. This analysis was approved by the Institutional Review Board. As in our sequential GemBuMel trials, HRR was defined for this analysis by one or more of the following criteria: relapse within 1 year or refractoriness to frontline therapy, extranodal extension at relapse, B symptoms at relapse, failure to achieve a complete remission in response to the most recent therapy, or requiring two or more lines of salvage therapy. Lines of salvage chemotherapy were defined as different regimens used to treat persistent/progressive disease and did not include a different regimen given to mobilize peripheral blood progenitor cells following a complete remission. Patients not meeting any HRR criteria were excluded from this analysis. Bulky lesions at relapse were

900

defined as those ≥5 cm. All demographic and tumor-related variables were captured prospectively in our departmental database. During this 15-year period we studied new HDC regimens for HRR cHL in sequential Institutional Review Board-approved clinical trials: a phase II trial of BuMel (NCT00427765),14 a phase I trial of GemBuMel (NCT00410982),15 a phase II study of GemBuMel (NCT01200329),16 a phase I/II trial of vorinostat/GemBuMel (NCI2011-02891),18 and a phase I/II trial of azacytidine/ vorinostat/GemBuMel (NCT01983969).19 The upper age limit of participants in these trials was 65 years and the patients had to have had a performance status 0-2 and normal renal, pulmonary, cardiac and hepatic function (Online Supplementary Material). BuMel,14 GemBuMel,15 vorinostat/GemBuMel,18 and azacytidine/vorinostat/GemBuMel19 were administered as previously described (Online Supplementary Material). Since azacytidine did not improve the activity of vorinostat/GemBuMel, we included those patients with the vorinostat/GemBuMel group in this analysis. Patients with HRR cHL who were eligible for those trials but who instead received standard BEAM were prospectively registered in our database. In addition, both GemBuMel and vorinostat/Gem/Bu/Mel were adopted as standard regimens at our institution upon publication of their trials. Reasons for treating patients off study included their declining trial participation, patients’ lack of clinical trial insurance benefits or periods when no trial was open to enrollment. The choice of HDC regimen off study was at the treating physician’s discretion. Institutional guidelines for supportive care and follow-up visits were followed (Online Supplementary Material). All patients in this analysis had undergone pre-ASCT PET/computed tomography scans, prospectively interpreted as positive (active tumor) or negative (no active tumor) using the International Harmonization Project in Lymphoma (IHPL) criteria up to 2013 (with mediastinal blood pool activity as the reference background),20 and the Deauville score from 2014 thereafter (with a score 1-3 considered a complete remission).21 Post-transplant radiotherapy, delivered at 30-41.4 Gy, was considered for bulky relapses and/or PET-positive lesions at ASCT. Maintenance BV was considered for all patients after the results of the AETHERA study became available.9

Statistical analyses Differences in variables by cohort were assessed with Wilcoxon rank-sum and c2 tests for continuous and categorical variables, respectively.22 PFS and overall survival (OS) were measured from the initiation of HDC to relapse or death, respectively, or last follow-up visit. The Kaplan-Meier method estimated 12-, 24-, and 60-month PFS and OS,23 and differences in outcomes were assessed using the log-rank test.24 Univariable and multivariable Cox regression analyses identified factors associated with PFS and OS. Statistical significance was defined by a=0.05 and all analyses used SAS v.9.4 (Cary, NC, USA).

Results A total of 501 patients with HRR cHL were treated with HDC and ASCT between 01/01/2005 and 12/31/2019 and all are included in this analysis: 189 received GemBuMel (159 on two clinical trials and 30 off study), 128 received vorinostat-GemBuMel (± azacytidine) (41 on trial, 87 off trial), 146 received BEAM (all standard of care), and 38 BuMel (all on study). Thirty-seven patients (7.3%) received maintenance BV. Over a median follow-up of 50 months (range, 6-186), a haematologica | 2022; 107(4)

ASCT for poor-risk relapsed HL in the last 15 years

Figure 1. Progression-free survival and overall survival of all patients. PFS: progression-free survival; OS: overall survival; ASCT: autologous stem-cell transplantation.

total of 205 patients (40.9%) experienced relapse and 130 patients (25.9%) died. Treatment-related mortality from HDC/ASCT was the cause of death of two patients, aged 40 and 45, who died from infectious complications, both around 3 months after BEAM therapy. Other causes of death were progressive disease (n=96), second primary malignancies (n=8), toxicity from post-ASCT salvage therapies (n=21, 19 after allogeneic stem cell transplantation, 2 after BV), unrelated late events (n=2), and unknown (n=1). The causes of death did not vary across the different time periods (Online Supplementary Table S1). The overall 1-year, 2-year and 5-year PFS rates for the whole population were 67% (95% confidence interval [95% CI]: 63-71%), 60% (95% CI: 56-64%), and 55% (95% CI: 50-59%), respectively. The 1-year, 2-year and 5year OS rates were 92% (95% CI: 89-94%), 84% (95% CI: 81-87%), and 73%, respectively (Figure 1). There was a gradual improvement in PFS and OS over time (Figures 2 and 3). The 2-year PFS rates were 48% for those transplanted between 2005-2007, 50.6% for those transplanted between 2008-2010, 64.3% for those transplanted between 2011-2015 and 78.7% for those transplanted between 20162019 (P<0.0001) (Figure 2A). Their respective 2-year OS rates were 74.6%, 76.8%, 89.7% and 96.2% (P<0.0001) (Figure 2B). Seven BEAM patients were ineligible for the clinical trials due to age older than 65 years. Three of them are alive in complete remission at 15 months, 3 years and 5 years after ASCT. The other four relapsed at a median of 17 months after ASCT (range, 4-39 months), and three died from tumor progression. We excluded these seven patients from the cohort and prognostic analyses described below, for which all patients met the same eligibility criteria.

Cohort analyses There were significant differences in disease characteristics among the four cohorts (Table 1). The GemBuMel and vorinostat/GemBuMel groups included more patients with primary refractory disease (P=0.001), bulky relapse (P<0.0001) and more patients with three or more haematologica | 2022; 107(4)

high-risk criteria (P=0.0006), as well as more patients with PET-positive disease at ASCT (P=0.0002), as compared to patients treated with BEAM or BuMel. Patient- and tumor-related variables did not change substantially over time but there was an increase in the use of pre-ASCT BV (P<0.0001) and anti-PD1 (P<0.0001), a decrease in PET-positive disease at ASCT (P=0.0008), and an increase in post-transplant BV (P<0.0001) (Table 2). BEAM and BuMel predominated in earlier years (2005-2007), GemBuMel and BEAM in middle years (2008-2015), and vorinostat/GemBuMel and BEAM in the last 4 years (2016-2019) (P<0.0001). Consequently, the use of post-ASCT maintenance BV was largely restricted to the vorinostat/GemBuMel and BEAM cohorts (P<0.0001). These two cohorts, in particular the one treated with vorinostat/GemBuMel, also received more prior BV (P<0.0001) and anti-PD1 (P=0.0001). To discern a possible confounding effect of having followed two different sets of criteria for interpretation of PET scans (IHPL from 2005-2013 and the Deauville score from 2014-2019) we retrospectively reviewed all patients from the earlier period who had a positive PET at ASCT by IHPL criteria. Thus, those whose PET showed uptake greater than mediastinum but not than liver were reassigned as negative (Deauville score 3). Of 115 patients with a positive PET by IHPL, 23 were reassigned as PETnegative: 15 in the GemBuMel cohort (21.7% of PETpositive patients by IHPL in this cohort), two in the vorinostat/GemBuMel cohort (18.1%), four in the BEAM cohort (16.6%), and two in the BuMel cohort (18.1%). There were significant differences among the four cohorts in PFS (P=0.0003) (Figure 3A) and OS (P<0.0001) (Figure 3B), with patients receiving vorinostat/ GemBuMel having the best outcomes, followed by those treated with GemBuMel, BEAM and BuMel. The respective 2-year and 5-year PFS rates were 73.2% and 71.9% (vorinostat/GemBuMel), 57.3% and 55% (GemBuMel), 56.3% and 45% (BEAM), and 47.4% and 38.9% (BuMel) (Figure 4). Likewise, the respective 2year and 5-year OS rates were 93.8% and 87.3% 901

Y. Nieto et al.

(vorinostat/GemBuMel), 85.5% and 75.5% (GemBuMel), 75.2% and 60.8% (BEAM), and 78.9% and 57.2% (BuMel) (Figure 5). The differences among regimens persisted within the subgroups with PET-negative (PFS: P=0.0002; OS: P<0.0001) (Figure 4A) and PET-positive disease at ASCT (PFS: P=0.002; OS: P<0.0001) (Figure 4B). Likewise, these differences were also seen when patients were analyzed by number of risk factors (Online Supplementary Figures S1-3). Overall responses to HDC, determined around day +30 after ASCT in patients with measurable active disease at the time of transplantation, did not vary among the cohorts: BEAM 76.9%; BuMel 72.7%, GemBuMel

88.3%, and vorinostat/GemBuMel 88.6% (P=0.48). However, complete remission rates were higher after vorinostat/GemBuMel (82.8%) than after GemBuMel (70.1%), BuMel (63.6%) or BEAM (50%) (P=0.03). The median follow-up times of the four cohorts were 97 months (range, 3-189) for those transplanted between 2005-2007, 93 months (range, 6-138) for those transplanted between 2008-2011), 57 months (range, 6-91) for those transplanted between 2012-2015, and 26 months (range, 3-55) for those transplanted between 2016-2019.

Prognostic analyses Univariate analyses of PFS showed that primary refracto-

Table 1. Patient and clinical features of the matched cohorts of patients (n=494).

Age in years, median (range) Gender, male/female, % ASCT year interval 2005-2007 2008-2011 2012-2015 2016-2019 N. of modified AETHERA criteria Median (range) 1 2 3 4 Primary refractory disease Prior disease-free interval# Median (range) <12 months ≥12 months PS ≥1 at relapse Extranodal extension at relapse B symptoms at relapse Prior radiotherapy Relapse within prior RT field Bulky relapse Prior BV Prior anti-PD1 N. of prior lines of therapy Median (range) >2 N. of prior relapses Median (range) >1 Positive PET at ASCT Progressive disease at ASCT Post-ASCT radiotherapy Post-ASCT BV

All (N=494)

BEAM (N=139)

BuMel (N=38)

GemBuMel (N=189)

Vorinostat/GemBuMel (N=128)

34.5(8-65) 57.5 / 42.5

36 (10-65) 58.2 / 41.8

36 (20-63) 65.8 / 34.2

34 (13-65) 56.1 / 43.9

33 (8-62) 56.3 / 43.8

0.31 0.72

97 (19.6%) 138 (27.9%) 157 (31.7%) 102 (20.6%)

57 (41%) 37 (27.2%) 25 (17.9%) 20 (14.7%)

38 (100%) 0 (0%) 0 (0%) 0 (0%)

2 (1.1%) 97 (51.3%) 88 (46.6%) 2 (1.1%)

0 (0%) 4 (3.1%) 44 (34.4%) 80 (62.5%)

<0.0001

2 (1-4) 198 (40%) 196 (39.6%) 84 (17%) 16 (3.2%) 218 (44.1%)

1 (1-4) 72 (51.8%) 48 (34.5%) 16 (11.5%) 3 (2.1%) 45 (34.5%)

2 (1-4) 13 (34.2%) 19 (50%) 5 (13.1%) 1 (2.6%) 12 (31.5%)

2 (1-4) 72 (38%) 77 (40.7%) 35 (18.5%) 5 (2.6%) 100 (53.4%)

2 (1-4) 41 (32%) 52 (40.6%) 28 (21.9%) 7 (5.5%) 61 (50%)

0.001 0.0006

2 (3-242) 186 (67.4%) 90 (32.6%) 179 (36.2%) 205 (41.4%) 79 (16%) 133 (26.9%) 55 (11.1%) 150 (30.3%) 120 (24.3%) 19 (3.8%)

4 (3-115) 65 (69.1%) 29 (30.9%) 55 (37.6%) 44 (31.6%) 21 (15.1%) 42 (30.2%) 12 (8.6%) 22 (15.8%) 25 (18%) 3 (2.1%)

1 (3-98) 24 (92.3%) 2 (76.9%) 11 (28.9%) 13 (34.2%) 5 (13.1%) 11 (28.9%) 4 (10.5%) 7 (18.4%) 0 (0%) 0 (0%)

0 (3-166) 59 (66.3%) 30 (33.7%) 64 (33.8%) 84 (44.4%) 26 (13.7%) 47 (24.9%) 19 (10%) 74 (39.1%) 24 (12.7%) 2 (1%)

0 (3-242) 38 (56.7%) 29 (43.2%) 49 (38.2%) 57 (44.5%) 27 (21.1%) 33 (25.8%) 21 (16.4%) 47 (36.7%) 71 (55.4%) 14 (10.9%)

0.02 0.01 0.55 0.21 0.29 0.85 0.16 <0.0001 <0.0001 0.0001

2 (2-8) 206 (41.7%)

2 (2-8) 44 (31.6%)

2 (2-6) 16 (42.1%)

2 (2-6) 72 (38.1%)

3 (2-7) 74 (57.8%)

<0.0001 <0.0001

1 (1-7) 165 (33.4%) 141 (28.5%) * 118 (23.9%) ** 38 (7.6%) 69 (14%) 37 (7.4%)

1 (1-7) 41 (29.4%) 25 (18%) * 21 (15.1%) ** 4 (2.8%) 11 (7.9%) 14 (10.3%)

1 (1-4) 10 (26.3%) 11 (28.9%) * 9 (23.6%) ** 1 (2.6%) 4 (10.5%) 0 (0%)

1 (1-5) 63 (33.3%) 75 (39.7%) * 60 (31.7%) ** 24 (12.7%) 28 (14.8%) 1 (0.5%)

1 (1-6) 51 (39.8%) 37 (28.9%) * 35 (27.3%) ** 9 (7%) 27 (21.1%) 20 (15.6%)

0.10 0.19 0.0002 * 0.007 ** 0.004 0.02 <0.001

0.001

Values are numbers (percentages) unless otherwise stated. #Disease-free interval excludes patients with primary refractory disease. *PET interpreted per International Harmonization Project in Lymphoma (2005-2013) and Deauville criteria (2014-2019). **All PET interpreted per Deauville criteria. BEAM: carmustine/ etoposide/cytarabine/melphalan; BuMel: busulphan/melphalan; GemBuMel: gemcitabine/busulphan/melphalan; ASCT: autologous stem-cell transplant; PS: performance status; RT: radiotherapy; BV: brentuximab vedotin; PET: positron emission tomography.

902

haematologica | 2022; 107(4)

ASCT for poor-risk relapsed HL in the last 15 years

ry disease (P=0.005), B symptoms at relapse (P=0.009), performance status ≥1 at relapse (P=0.01), more than one prior relapse (P=0.0001), more than two prior lines of therapy (P=0.0004), positive PET at ASCT (P<0.0001), and progressive disease at ASCT (P<0.0001) correlated with an adverse

PFS. In contrast, prior BV treatment (P=0.01), prior anti-PD1 treatment (P=0.03), the use of vorinostat/GemBuMel (P=0.0004), and post-ASCT maintenance therapy with BV (P=0.01) were associated with a favorable PFS (Table 3). In multivariable analyses of PFS, primary refractory dis-

Table 2. Patient and clinical characteristics by treatment year interval (entire population, n=501).

2005-2007 (N=98) Age in years, median (range) Gender, male/female (%) HDC regimen BEAM BuMel GemBuMel Vorinostat/GemBuMel Primary refractory disease Prior disease-free interval * Median (range) <12 months PS ≥1 at relapse Bulky relapse Extranodal extension at relapse B symptoms at relapse N. of prior relapses Median (range) >1 N. of prior lines of therapy Median (range) >2 Prior BV Prior anti-PD1 Positive PET at ASCT Progressive disease at ASCT Post-ASCT radiotherapy Post-ASCT BV

Treatment year interval 2008-2011 2012-2015 (N=138) (N=158)

2016-2019 (N=107)

30 (18-72) 62 (63%) / 36 (37%)

35 (10-69) 80 (58%) / 58 (42%)

32 (13-70) 87 (55%) /71 (45%)

34 (8-71) 59 (55%) /48 (45%)

0.58 0.57

58 (59.2%) 38 (100%) 2 (2%) 0 (0%) 41 (41.8%)

37 (26.8%) 0 (0%) 97 (70.3%) 4 (2.9%) 69 (50%)

26 (16.5%) 0 (0%) 88 (55.7%) 44 (27.8%) 74 (46.8%)

25 (23.4%) 0 (0%) 2 (1.9%) 80 (74.8%) 43 (40.2%)

<0.0001

1 (3-108) 43 (75.4%) 37 (37.7%) 19 (19.4%) 33 (33.7%) 14 (14.3%)

0 (3-166) 45 (65.2%) 48 (34.4%) 45 (32.6%) 64 (46.4%) 19 (13.8%)

3 (3-145) 45 (53.5%) 61 (38.6%) 60 (38%) 68 (43%) 24 (15.2%)

3 (3-242) 83 (62.5%) 33 (30.8%) 27 (25.2%) 40 (37.4%) 22 (20.6%)

0.11 0.16 0.59 0.008 0.19 0.48

1 (1-4) 27 (27.6%)

1 (1-6) 54 (39.1%)

1 (1-7) 51 (32.3%)

1 (1-5) 34 (31.8%)

0.24 0.28

2 (2-6) 34 (34.7%) 0 (0%) 0 (0%) 29 (29.6%) 5 (5.1%) 7 (7.1%) 0 (0%)

2 (2-7) 54 (39.1%) 4 (2.9%) 0 (0%) 58 (42%) 23 (16.7%) 28 (20.3%) 0 (0%)

2 (2-8) 72 (45.6%) 60 (38%) 2 (1.3%) 40 (25.3%) 7 (4.4%) 16 (10.1%) 2 (1.3%)

2 (2-10) 47 (43.9%) 58 (54.2%) 19 (17.8%) 21 (19.6%) 3 (2.8%) 19 (17.8%) 35 (32.7%)

0.29 0.31 <0.0001 <0.0001 0.0008 <0.0001 0.0089 <0.0001

0.39

Values are numbers (percentages) unless otherwise stated. *Disease-free interval excludes patients with primary refractory disease. HDC: high-dose chemotherapy; BEAM: carmustine/etoposide/cytarabine/melphalan; BuMel: busulphan/melphalan; GemBuMel: gemcitabine/busulphan/melphalan; PS: performance status; BV: brentuximab vedotin; PET: positron emission tomography; ASCT: autologous stem-cell transplant.

Figure 2. Outcomes by treatment year. (A) Progression-free survival, (B) overall survival.

haematologica | 2022; 107(4)

903

Y. Nieto et al. Table 3. Cox regression univariable and multivariable analyses of progression-free survival of the matched cohorts of patients.

Age >35 years Female gender Treatment year 2005-2007 2008-2011 2012-2015 2016-2019 Primary refractory disease Prior disease-free interval <1 year N. of prior relapses >1 N. of prior lines of therapy >2 PS ≥1 at relapse Bulky relapse Extranodal extension at relapse B symptoms at relapse Positive PET at ASCT Progressive disease at ASCT Prior BV Prior anti-PD1 HDC regimen Vorinostat/GemBuMel GemBuMel BEAM BuMel Post-ASCT BV Post-ASCT radiotherapy

Univariable 95% CI LL UL

Multivariable 95% CI LL UL

1.17 0.92

0.89 0.70

1.54 1.22

0.26 0.57

1.13 1.02

0.85 0.76

1.50 1.36

0.39 0.9

3.85 3.14 2.07 1 (ref) 1.49 0.76 1.72 1.65 1.28 1.33 0.97 1.58 2.90 2.64 0.65 0.23

2.27 1.87 1.22

6.54 5.28 3.51

<0.0001 <0.0001 0.006

1.14 0.53 1.30 1.26 1.19 1.00 0.73 1.13 2.20 1.77 0.45 0.06

1.96 1.09 2.27 2.17 1.92 1.77 1.28 2.22 3.83 3.94 0.93 0.94

0.003 0.14 0.0001 0.0003 0.01 0.047 0.81 0.008 <0.0001 <0.0001 0.01 0.04 global P=0.0006

1.41 0.92 1.19 1.60 1.15 1.56 0.98 1.68 2.60 1.06 0.58 0.35

1.01 0.60 0.79 1.08 0.78 1.15 0.73 1.19 1.83 0.64 0.36 0.08

1.97 1.40 1.78 2.12 1.52 2.12 1.31 2.37 3.69 1.76 0.93 1.48

0.04 0.69 0.4 0.01 0.7 0.004 0.88 0.003 <0.0001 0.83 0.02 0.15 global P=0.001

1.16 1.38 1.63 0.08 0.63

2.61 3.19 4.62 0.73 1.39

0.007 0.0005 0.0002 0.01 0.75

1 (ref) 1.33 2.19 2.29 0.37 0.66

0.83 1.35 1.24 0.12 0.43

2.13 3.55 4.23 1.20 1.02

0.23 0.001 0.007 0.09 0.06

1 (ref) 1.74 2.10 2.74 0.23 0.94

HR: hazard ratio. 95% CI: 95% confidence interval. LL: lower limit. UL: upper limit; ref: reference; PS: performance status; PET: positron emission tomography; ASCT: autologous stemcell transplant. BV: brentuximab vedotin; HDC: high-dose chemotherapy; GemBuMel: gemcitabine/busulphan/melphalan; BEAM: carmustine/etoposide/cytarabine/melphalan; BuMel: busulphan/ melphalan.

ease (hazard ratio [HR]=1.41 [95% CI: 1.01-1.97], P=0.04), more than two prior lines of therapy (HR=1.60 [95% CI: 1.08-2.36], P=0.01), bulky relapse (HR=1.56 [95% CI: 1.152.12], P=0.004), B symptoms (HR=1.68 [95% CI: 1.19-2.37], P=0.003) and a positive PET at ASCT (HR=2.60 [95% CI: 1.83-3.69], P<0.0001) were independent adverse prognostic factors, whereas prior BV (HR=0.58 [95% CI: 0.36-0.93], P=0.02) and vorinostat/GemBuMel (P<0.0001) were independently associated with improved PFS. The hazard ratios for the other three HDC regimens compared to vorinostat/GemBuMel were: GemBuMel: 1.33 (95% CI: 0.83-2.13), BEAM: 2.19 (95% CI: 1.35-3.55), and BuMel 2.29 (95% CI: 1.24-4.23) (Table 3). The following were unfavorably associated with OS in univariate analyses: age >35 years (P=0.006), B symptoms (P=0.006), performance status ≥1 (P=0.002), more than one prior relapse (P=0.0001), more than two prior lines of therapy (P<0.0001), positive PET at ASCT (P<0.0001), and progressive disease at ASCT (P<0.0001). In contrast, prior BV treatment (P=0.01) and vorinostat/GemBuMel (P<0.0001) were associated with a more favorable OS (Table 4). Multivariable OS analyses identified age >35 years (HR=1.80 [95% CI: 1.24-2.60], P=0.002), B symptoms (HR=1.74 [95% CI: 1.13-2.68], P=0.01), more than two 904

prior lines of therapy (HR=2.11 [95% CI: 1.26-3.53], P=0.004), and positive PET at ASCT (HR=1.88 [95% CI: 1.16-3.04], P=0.01) as independent adverse prognostic factors. On the contrary, prior BV treatment (HR=0.46 [95% CI: 0.23-0.90], P=0.02) and vorinostat/GemBuMel (P<0.0001) were independently associated with better OS. The hazard ratios for the other three HDC regimens compared to vorinostat/GemBuMel were: GemBuMel: 1.63 (95% CI: 0.75-3.57), BEAM: 5.06 (95% CI: 2.30-11.10), and BuMel 5.17 (95% CI: 2.13-12.54) (Table 4). Of note, evaluation of all PET according to the Deauville score did not change the prognostic effect of this variate in univariate analyses or the results for any variable in the multivariate analyses.

Treatment for post-ASCT relapse Patients received a median of two (range, 0-12) lines of salvage therapy for post-ASCT recurrence. Salvage therapies included BV (n=85), conventional chemotherapy (n=72), clinical trials of experimental agents (n=67), allogeneic stem cell transplantation (n=64), anti-PD1 (n=37), radiotherapy (n=27), and unknown (n=14). No salvage therapy was administered to 13 patients. Of the 205 patients who relapsed, 53 are currently in a new clinical haematologica | 2022; 107(4)

ASCT for poor-risk relapsed HL in the last 15 years

complete remission after allogeneic stem cell transplantation (n=26), anti-PD1 (n=13), BV (n=9), radiotherapy (n=3) and chemotherapy (n=2).

Second primary malignancies Out of the entire population (n=501) eight patients developed therapy-related myelodysplastic syndrome and

Figure 3. Outcomes by high-dose chemotherapy regimen. (A) Progression-free survival, (B) overall survival. GemBuMel: gemcitabine/busulphan/melphalan; BEAM: carmustine/etoposide /cytarabine/melphalan; BuMel: busulphan/melphalan.

Table 4. Cox regression univariable and multivariable analyses of overall survival of the matched cohorts of patients.

Univariable 95% CI LL UL

Multivariable 95% CI LL UL

1.63 0.99

1.15 0.69

2.32 1.42

0.006 0.96

1.80 1.18

1.24 0.81

2.60 1.73

0.002 0.39

11.87 7.21 3.41 1 (ref) 1.38 0.70 2.03 2.07 1.37 1.22 1.00 1.86 2.80 3.38 0.52 0.34

3.69 2.23 1.03

38.23 23.31 11.31

<0.0001 0.001 0.04

0.97 0.42 1.42 1.45 1.21 0.84 0.70 1.22 1.96 2.14 0.30 0.05

1.97 1.15 2.89 2.96 1.98 1.77 1.42 2.84 4.00 5.32 0.90 2.42

0.07 0.15 0.0001 <0.0001 0.002 0.29 0.98 0.004 <0.0001 <0.0001 0.01 0.28 global

1.27 0.78 1.25 2.11 1.21 1.51 1.08 1.74 1.88 1.92 0.46 0.72

0.82 0.44 0.74 1.26 0.79 1.00 0.74 1.13 1.16 1.02 0.23 0.09

1.96 1.39 2.13 3.53 1.85 2.29 1.57 2.68 3.04 3.58 0.90 5.64

0.28 0.39 0.40 0.004 0.44 0.05 0.69 0.01 0.01 0.04 0.02 0.75 global

P<0.0001 Vorinostat/GemBuMel GemBuMel BEAM BuMel Post-ASCT BV Post-ASCT radiotherapy

1 (ref) 2.27 4.24 5.53 0.21 1.18

1.14 2.14 2.59 0.03 0.74

4.51 8.39 11.80 1.54 1.89

P<0.0001 0.01 <0.0001 <0.0001 0.12 0.48

1 (ref) 1.63 5.06 5.17 0.41 1.06

0.75 2.30 2.13 0.06 0.62

3.57 11.10 12.54 3.05 1.82

0.22 <0.0001 0.0003 0.38 0.83

HR: hazard ratio. 95% CI: 95% confidence interval. LL: lower limit. UL: upper limit; ref: reference; PS: performance status; PET: positron emission tomography; ASCT: autologous stem-cell transplant. BV: brentuximab vedotin; HDC: high-dose chemotherapy; GemBuMel: gemcitabine/busulphan/melphalan; BEAM: carmustine/etoposide/cytarabine/melphalan; BuMel: busulphan/ melphalan.

haematologica | 2022; 107(4)

905

Y. Nieto et al.

Figure 4. Progression-free survival by high-dose chemotherapy regimen according to positron emission tomography status. (A) Progression-free survival in patients with (A) positron emission tomography (PET)-negative disease and with (B) PET-positive disease.

five patients developed therapy-related acute myeloblastic leukemia: seven after BEAM (4.8%), three after GemBuMel (1.5%), two after BuMel (5.2%) and one after vorinostat/GemBuMel (0.07%), at a median 31 months (range, 5-133) after ASCT. The incidence of therapy-related myelodysplastic syndrome/acute myeloid leukemia did not vary significantly among the cohorts (P=0.13). Cytogenetic findings in these patients included complex abnormalities with -7/del(7q) ± -5/del(5q) (n=7), -7 alone (n=3), 11q+ (n=1), and other abnormalities (n=2). These patients were older (median age 54, range 22-72) than all other patients (n=493) who did not develop therapy-related myelodysplastic syndrome/acute myeloid leukemia (median age 32; range, 8-71) (P=0.0005). Other second primary malignancies were renal-cell carcinoma (2 BEAM patients), Müllerian adenocarcinoma and epithelioid hemangioendothelioma (1 GemBuMel patient each), and diffuse large B-cell lymphoma (1 BuMel patient, 1 vorinostat/GemBuMel patient).

Discussion Our analysis of patients with HRR cHL treated with HDC and ASCT shows a gradual and significant improvement of outcomes over the last 15 years. Improved tumor control with BV before ASCT and the use of more active HDC regimens, particularly vorinostat/GemBuMel, emerged independently as favorable prognostic factors in multivariable analysis. The clinical development of vorinostat/GemBuMel was based on two important preclinical observations. The first one was the synergistic inhibition by gemcitabine of the repair of DNA damage caused by busulfan and melphalan.15 The use of ASCT enables the infusion of gemcitabine at its optimal rate of 10 mg/m2/min, previously shown to avoid saturation of its intracellular enzymatic activation,25 which results in greater activity and myelotoxicity than shorter infusions of this drug,26,27 and in turn optimizes the synergy with the bifunctional DNAalkylating agents.28 Our second major in vitro observation was that relaxation of chromatin after increased histone acetylation with vorinostat facilitated access of gemc906

itabine, busulfan and melphalan, to DNA, which increased DNA damage, apoptosis and cytotoxicity in refractory lymphoma cell lines.17 Those preclinical observations, tested in subsequent clinical trials, are confirmed in the present analysis, and notably did not increase the risk of treatment-related mortality. The other major favorable factor was the use of BV before ASCT. BV has revolutionized the treatment of cHL in the last decade. Following its favorable results and approval in 2011 for post-ASCT relapses,29 BV was moved up to the first or second line of salvage therapy before ASCT,8,30 which allows more patients to receive HDC in a PET-negative complete remission. Lastly, BV was successfully tested in the post-ASCT maintenance setting, in which the randomized AETHERA trial compared 16 cycles of BV with placebo in 329 patients with HRR cHL, defined by the same criteria as in our analysis. The use of BV resulted in improved 2-year PFS (63% vs. 51%) and 5-year PFS (59% vs. 41%), albeit with no OS benefit.9,10 In contrast to our population, this trial did not allow prior BV and more than 60% of patients received BEAM. Despite these differences, we also saw a correlation of maintenance BV with improvement of PFS but not OS in our univariate analyses. Maintenance BV was restricted to patients we treated in later years, which likely resulted in a loss of power and significance in the multivariable analysis. We saw that the pre-ASCT use of the anti-PD1 antibodies nivolumab and pembrolizumab was associated with improved PFS, although this did not hold significance in multivariable analysis, probably due to the small number of patients who received them. This class of drugs has produced another breakthrough in the treatment of Hodgkin lymphoma. Besides their efficacy in post-ASCT relapses,31,32 these drugs can serve as a successful bridge to ASCT by inducing responses in refractory relapses.33 In addition, anti-PD1 might chemosensitize refractory tumors and improve results of HDC.34 The strengths of our analysis include the homogeneity of the population of patients and of the treatments administered in the four cohorts and the large sample size, which allowed us to independently dissect the prognostic value of the patient-, tumor-, and treatment-related variables. On the other hand, our study has several limitahaematologica | 2022; 107(4)

ASCT for poor-risk relapsed HL in the last 15 years

tions. First, we only intended to analyze those patients who ultimately received HDC and ASCT after HDC, and our analyses exclude patients who failed to successfully undergo salvage chemotherapy, e.g., due to morbidity or highly refractory disease. Thus, our population does not represent an unselected real-world cohort of HRR cHL patients. Second, the comparison of the different HDC regimens is nonrandomized. While all of our HRR patients met the eligibility criteria of the prospective trials, physician biases in assigning patients with more aggressive tumors who were perceived to be fitter to a clinical trial instead of standard HDC likely played a role, as was reflected in the higher proportion of patients with positive PET at ASCT or other HRR criteria in the cohorts treatment with GemBuMel with or without vorinostat, compared with the BEAM group. Third, while all patients’ data were captured prospectively, this study is retrospective in nature, and thus, fraught with the usual limitations of these analyses, including the possibility of reporting biases or underreporting of second primary malignancies and other long-term events. Fourth, our analysis, which encompasses a 15-year period, is subject to the changes in ASCT supportive care during this time span. However, refinement of supportive measures does not appear to be the cause of the improvement in results over time, as the treatment-related mortality was minimal. Fifth, the weight in our analysis of some major new treatments of Hodgkin lymphoma incorporated more recently, such as post-ASCT maintenance or the pre-ASCT use of anti-PD1, is limited by smaller numbers of patients. Lastly, since patients in the vorinostat/GemBuMel cohort had worse prognostic features and since this was the HDC regimen most used in the last period (2016-2019), this cohort had the highest use of pre-ASCT and post-ASCT BV, which likely contributed to its better results. Nevertheless, this regimen clearly stands out as an independent favorable factor for both PFS and OS after adjusting for all other variables. However, definite proof of superiority of

References 1. Linch DC, Winfield D, Goldstone AH, et al. Dose intensification with autologous bonemarrow transplantation in relapsed and resistant Hodgkin's disease: results of a BNLI randomised trial. Lancet. 1993; 341(8852):1051-1054. 2. Schmitz N, Pfistner B, Sextro M, et al. Aggressive conventional chemotherapy compared with high-dose chemotherapy with autologous haemopoietic stem-cell transplantation for relapsed chemosensitive Hodgkin's disease: a randomised trial. Lancet. 2002;15;359(9323):2065-2071. 3. Sureda A, Constans M, Iriondo A, et al. Prognostic factors affecting long-term outcome after stem cell transplantation in Hodgkin's lymphoma autografted after a first relapse. Ann Oncol. 2005;16(4):625633. 4. Moskowitz A, Yahalom J, Kewalramani T, et al. Pretransplantation functional imaging predicts outcome following autologous stem cell transplantation for relapsed and refractory Hodgkin lymphoma. Blood. 2010;116(23):4934-4937. 5. Bröckelmann PJ, Müller H, Casasnovas O, et al. Risk factors and a prognostic score for

haematologica | 2022; 107(4)

vorinostat/GemBuMel over BEAM will require a randomized trial. Other novel strategies developed to improve the outcome of HRR cHL patients undergoing ASCT include new maintenance therapies, such as anti-PD-1 alone35 or antiPD-1 plus BV,36 which have shown very promising results. Anti-PD-1 can be easily used after ASCT with vorinostat/GemBuMel. Tandem ASCT based on BEAM has been studied,11-13 but it is unclear, in the absence of a prospective randomized trial, how this approach might compare to a single ASCT using vorinostat/GemBuMel. In conclusion, the outcome of HRR cHL patients treated with HDC and ASCT has improved substantially over the last 15 years. The incorporation of BV into pre-ASCT salvage therapy and the use of pharmacologically optimized, more active HDC regimens, particularly vorinostat/GemBuMel, were associated with these improved results. Disclosures YN has provided consultancy services for Affimed and Novo Nordisk; and has received research funding from Affimed, Novartis, Takeda, Astra-Zeneca, and Biosecura. BA holds a patent for intravenous busulfan. None of the other authors has any conflicts of interest to disclose. Contributions YN designed research, collected data, treated patients, and wrote the manuscript; SG and HC: analyzed the data and wrote the manuscript; BCV, RBJ, PA, CH, UP, MQ, PK, AA, NS, SS, KR, JR, MB, AG, TLS, SA, SI, HL, RN, SP, RS, BD, CP, JG, BC, KM, SK, RC, EJS and BSA treated patients and reviewed the manuscript. Data sharing statement All clinical trial protocols (NCT00427765, NCT00410982, NCT01200329, NCI-2011-02891, NCT01983969) and treatment orders will be made available upon request.

survival after autologous stem-cell transplantation for relapsed or refractory Hodgkin lymphoma. Ann Oncol. 2017; 28(6):1352-1358. 6. Chen Y-B, Lane AA, Logan B, et al. Impact of conditioning regimen on outcomes for patients with lymphoma undergoing highdose therapy with autologous hematopoietic cell transplantation. Biol Blood Marrow Transplant. 2015;21(6):1046-1053. 7. Moskowitz CH, Matasar MJ, Zelenetz AD, et al. Normalization of pre-ASCT, FDGPET imaging with second-line, non-crossresistant, chemotherapy programs improves event-free survival in patients with Hodgkin lymphoma. Blood. 2012; 119(7):1665-1670. 8. Moskowitz AJ, Schöder H, Yahalom J, et al. PET-adapted sequential salvage therapy with brentuximab vedotin followed by augmented ifosfamide, carboplatin, and etoposide for patients with relapsed and refractory Hodgkin's lymphoma: a nonrandomised, open-label, single-centre, phase 2 study. Lancet Oncol. 2015; 16(3):284-292. 9. Moskowitz CH, Nademanee A, Masszi T, et al. Brentuximab vedotin as consolidation therapy after autologous stem-cell transplantation in patients with Hodgkin's lym-

phoma at risk of relapse or progression (AETHERA): a randomised, double-blind, placebo-controlled, phase 3 trial. Lancet. 2015;385(9980):1853-1862. 10. Moskowitz CH, Walewski J, Nademanee A, et al. Five-year PFS from the AETHERA trial of brentuximab vedotin for Hodgkin lymphoma at high risk of progression of relapse. Blood. 2018;132(25):2639-2642. 11. Deau B, Amorim S, Perrot A, et al. Tandem haematopoietic stem cell transplantation for high risk relapsed/refractory Hodgkin lymphoma: a LYSA study.Br J Haematol. 2018;181(3):341-349. 12. Sibon D, Morschhauser F, Resche-Rigon M, et al. Single or tandem autologous stem-cell transplantation for first-relapsed or refractory Hodgkin lymphoma: 10-year followup of the prospective H96 trial by the LYSA/SFGM-TC study group. Haematologica. 2016;101(4):474-481. 13. Smith EP, Li H, Friedberg JW, Constine LS, et al. Tandem autologous hematopoietic cell transplantation for patients with primary progressive or recurrent Hodgkin lymphoma: a SWOG and Blood and Marrow Transplant Clinical Trials Network phase II trial (SWOG S0410/BMT CTN 0703). Biol Blood Marrow Transplant. 2018;24(4):700-707.

907

Y. Nieto et al. 14. Kebriaei P, Madden T, Kazerooni R, et al. Intravenous busulfan plus melphalan is a highly effective, well-tolerated preparative regimen for autologous stem cell transplantation in patients with advanced lymphoid malignancies. Biol Blood Marrow Transplant. 2011;17(3):412-420. 15. Nieto Y, Thall P, Valdez B, et al. High-dose infusional gemcitabine combined with busulfan and melphalan with autologous stem-cell transplant in patients with refractory lymphoid malignancies. Biol Blood Marrow Transplant. 2012;18(11):1677-1686. 16. Nieto Y, Thall PF, Ma J, et al. Phase II trial of high-dose gemcitabine/busulfan/melphalan with autologous stem cell transplantation for primary refractory or poor-risk relapsed Hodgkin lymphoma. Biol Blood Marrow Transplant. 2018;24(8):1602-1609. 17. Valdez BC, Nieto Y, Murray D, et al. Epigenetic modifiers enhance the synergistic cytotoxicity of combined nucleoside analog-DNA alkylating agents in lymphoma cell lines. Exp Hematol. 2012;40 (10):800-810. 18. Nieto Y, Valdez BC, Thall PF, et al. Vorinostat combined with high-dose gemcitabine, busulfan and melphalan with autologous stem-cell transplantation in patients with refractory lymphomas. Biol Blood Marrow Transplant 2015; 21(11):1914-1920. 19. Nieto Y, Valdez BC, Thall PF, et al. Double epigenetic modulation of high-dose chemotherapy with azacytidine and vorinostat for patients with refractory or poor-risk relapsed lymphoma. Cancer. 2016;122(17):2680-2688. 20. 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

908

Lymphoma. J Clin Oncol. 2007;25(5):571578. 21. Cheson BD, Fisher RI, Barrington SF, et al. Recommendations for initial evaluation, staging, and response assessment of Hodgkin and non-Hodgkin lymphoma: the Lugano classification. J Clin Oncol. 2014;32(27):3059-3067. 22. Fisher R. On the interpretation of x2 from contingency tables, and the calculation of P. J Royal Stat Soc. 1022;85:87-94. 23. Kaplan EL, Meier P. Nonparametric estimator from incomplete observations. J Amer Stat Assoc. 1958;53:457-481. 24. Peto R, Peto J. Asymptomatically efficient rank invariant test procedures. J R Stat Soc A. 1971;135:185-198. 25. Grunewald R, Kantarjian H, Keating MJ, et al. Pharmacologically directed design of the dose rate and schedule of 2',2'-difluorodeoxycytidine (Gemcitabine) administration in leukemia. Cancer Res. 1990; 50(21):6823-6826. 26. Tempero M, Plunkett W, Ruiz Van Haperen V, et al. Randomized phase II comparison of dose-intense gemcitabine: thirty-minute infusion and fixed dose rate infusion in patients with pancreatic adenocarcinoma. J Clin Oncol. 2003;21(18):3402-3408. 27. Soo RA, Wang LZ, Tham LS, et al. A multicenter randomized phase II study of carboplatin in combination with gemcitabine at standard rate or fixed dose rate infusion in patients with advanced stage non-smallcell lung cancer. Ann Oncol. 2006; 17(7):1128-1133. 28. Valdez BC, Andersson BS. Interstrand crosslink inducing agents in pretransplant conditioning therapy for hematologic malignancies. Environ Mol Mutagen. 2010; 51(6):659-668. 29. Younes A, Gopal AK, Smith SE, et al. Results of a pivotal phase II study of bren-

tuximab vedotin for patients with relapsed or refractory Hodgkin's lymphoma. J Clin Oncol. 2012;30(18):2183-2189. 30. LaCasce A, Bociek G, Sawas A, et al. Brentuximab vedotin plus bendamustine: a highly active first salvage regimen for relapsed or refractory Hodgkin lymphoma. Blood. 2018;132(1):40-48. 31. Younes A, Santoro A, Shipp M, et al. Nivolumab for classical Hodgkin’s lymphoma after failure of both autologous stem-cell transplantation and brentuximab vedotin: a multicenter, multicohort, singlearm phase 2 trial. Lancet Oncol. 2016; 17(9):1283-1294. 32. Armand PA, Shipp MA, Ribrag V, et al. Programmed death-1 blockade with pembrolizumab in patients with classical Hodgkin lymphoma after brentuximab vedotin failure. J Clin Oncol. 2016; 34(31):3733-3739. 33. Herrera AF, Moskowitz AJ, Bartlett NL, et al. Interim results of brentuximab vedotin in combination with nivolumab in patients with relapsed or refractory Hodgkin lymphoma. Blood. 2018;131(11):1183-1194. 34. Merryman RW, Redd RA, Nieto Y, et al. Outcome of autologous stem cell transplantation following PD-(L)1 based salvage therapy for multiply relapsed patients with classic Hodgkin lymphoma. Blood. 2019;134 (Suppl 1):4571. 35. Armand P, Chen YB, Redd RA, et al. PD-1 blockade with pembrolizumab for classical Hodgkin lymphoma after autologous stem cell transplantation. Blood. 2019;134(1):2229. 36. Herrera AF, Chen L, Nieto Y, et al. Consolidation with nivolumab and brentuximab vedotin after autologous hematopoietic cell transplantation in patients with high-risk Hodgkin lymphoma. Blood. 2020;136(Suppl 1):19-20.

haematologica | 2022; 107(4)

ARTICLE

Hodgkin Lymphoma

Inhibitors of ADAM10 reduce Hodgkin lymphoma cell growth in 3D microenvironments and enhance brentuximab-vedotin effect Roberta Pece,1* Sara Tavella,1* Delfina Costa,2* Serena Varesano,2* Caterina Camodeca,3 Doretta Cuffaro,3 Elisa Nuti,3 Armando Rossello,3 Massimo Alfano,4 Cristina D’Arrigo,5 Denise Galante,5° Jean-Louis Ravetti,6 Marco Gobbi,7 Francesca Tosetti,2 Alessandro Poggi2# and Maria Raffaella Zocchi8# Cellular Oncology Unit, IRCCS Ospedale Policlinico San Martino and Department of Experimental Medicine, University of Genoa, Genoa; 2Molecular Oncology and Angiogenesis Unit, IRCCS Ospedale Policlinico San Martino, Genoa; 3Department of Pharmacy, University of Pisa, Pisa; 4Division of Experimental Oncology/Unit of Urology, URI, IRCCS Ospedale San Raffaele, Milan; 5SCITEC-CNR, Genoa; 6Patology Unit, IRCCS Ospedale Policlinico San Martino, Genoa; 7Clinical Hematology, University of Genoa, Genoa and 8Division of Immunology, Transplants and Infectious Diseases, IRCCS San Raffaele Scientific Institute, Milan, Italy. 1

RP ST DC and SV contributed equally as co-first authors.

AP and MRZ contributed equally as co-senior authors.

Ferrata Storti Foundation

Haematologica 2022 Volume 107(4):909-920

°DG current address: ARPAL (Regional Agency for Environmental Protection, Liguria), Genoa, Italy.

ABSTRACT

hedding of ADAM10 substrates, like TNFa or CD30, can affect both anti-tumor immune response and antibody-drug-conjugate (ADC)-based immunotherapy. We have published two new ADAM10 inhibitors, LT4 and MN8 able to prevent such shedding in Hodgkin lymphoma (HL). Since tumor tissue architecture deeply influences the outcome of anti-cancer treatments, we set up a new threedimensional (3D) culture systems to verify whether ADAM10 inhibitors can contribute to, or enhance, the anti-lymphoma effects of the ADC brentuximab-vedotin (BtxVed). In order to recapitulate some aspects of lymphoma structure and architecture, we assembled two 3D culture models: mixed spheroids made of HL lymph node (LN) mesenchymal stromal cells (MSC) and Reed Sternberg/Hodgkin lymphoma cells (HL cells) or collagen scaffolds repopulated with LN-MSC and HL cells. In these 3D systems we found that: i) the ADAM10 inhibitors LT4 and MN8 reduce ATP content or glucose consumption, related to cell proliferation, increasing lactate dehydrogenase release as a cell damage hallmark; ii) these events are paralleled by mixed spheroids size reduction and inhibition of CD30 and TNFa shedding; iii) the effects observed can be reproduced in repopulated HL LN-derived matrix or collagen scaffolds; iv) ADAM10 inhibitors enhance the anti-lymphoma effect of the anti-CD30 ADC BtxVed both in conventional cultures and in repopulated scaffolds. Thus, we provide evidence for a direct and combined antilymphoma effect of ADAM10 inhibitors with BtxVed, leading to the improvement of ADC effects; this is documented in 3D models recapitulating features of the LN microenvironment, that can be proposed as a reliable tool for anti-lymphoma drug testing.

Introduction ADAM (A Disintegrin And Metalloproteinases) are transmembrane proteins with protease activity exerted on several substrates, including growth factors, cytokines, receptors and their ligands, leading to the release of soluble bioactive molecules.1,2 Some of them, such as tumor necrosis factor (TNF)a, are involved in the develop-

haematologica | 2022; 107(4)

Correspondence: MARIA RAFFAELLA ZOCCHI zocchi.maria@hsr.it Received: February 8, 2021. Accepted: May 28, 2021. Pre-published: June 10, 2021. https://doi.org/10.3324/haematol.2021.278469

909

R. Pece et al.

ment of different cancers.3-6 Moreover, since overexpression of ADAM10 relates with parameters of tumor progression,6,7 ADAM have been proposed as both biomarkers and therapeutic targets for cancer,7,8 and ADAM10 inhibitors with anti-tumor effects have been developed.9-11 ADAM10 expression and increased enzymatic activity has been documented in many tumors including chronic lymphocytic leukemia, acute myeloid leukemia, nonHodgkin and Hodgkin lymphomas (HL).12-14 In particular, we described the overexpression of ADAM10 in the lymph node (LN) microenvironment in HL, together with impaired stimulation of T lymphocytes with anti-tumor activity.14 Likewise, CD30 shedding due to ADAM10 activity has been reported to decrease the efficiency of targeted lymphoma cell killing obtained with anti-CD30 monoclonal antibodies (mAb) in vitro and this effect can be prevented by the use of the inhibitor GI254023X (abbreviated to GIX).15 We have developed the inhibitors LT4 and MN8 with high specificity for ADAM10 with the aim to enhance efficiency and selectivity of action and to reduce the shedding of NKG2D ligands (NKG2DL) and CD30 by HL cell lines. The exposure of HL cells to these compounds significantly increased their sensitivity to lymphocyte-mediated killing due to NKG2D/NKG2DL interaction or due to the antibody-dependent cellular cytotoxicity in the presence of the anti-CD30 mAb Iratumumab.16,17 Thus, the ADAM10 inhibitors, maintaining the expression of surface CD30, could further sensitize HL cells to the anti-lymphoma activity of the anti-CD30-drug conjugated (ADC) humanized mAb brentuximab-vedotin (BtxVed) used for HL treatment.18 All the mentioned reported findings rely on in vitro experiments performed with conventional culture systems that do neither consider the importance of tissue architecture nor the complexity of cell populations at the tumor site. The tissue microenvironment deeply contributes to determine both cancer progression and the outcome of anti-cancer treatments;19-22 thus, newly designed models for the definition of drug safety and efficacy are rapidly developing in the field. In turn, there is increasing evidence that tumor development in humans is not always reproducible and predictable in animal models, mostly used in preclinical studies.23-25 Furthermore, they are very expensive and require a long time to be set up and defined for each tumor. As an alternative, several three-dimensional (3D) culture systems, including spheroids and scaffolds, have been validated by the European Union Reference Laboratories for Alternatives to Animal Testing (EURL ECVAM) as preclinical models, to overcome these inconveniences at least in part.26-29 In this work we demonstrate the anti-lymphoma effect of ADAM10 inhibitors in HL in two different 3D culture systems: mixed spheroids made of LN mesenchymal stromal cells (MSC) and Reed Sternberg/Hodgkin lymphoma cells (from now on HL cells) and collagen sponges repopulated with both LN-MSC and HL cells. In these 3D models we found that: i) the ADAM10 inhibitors LT4 and MN8 reduced HL cell ATP content and glucose consumption related to proliferation, while increasing lactate dehydrogenase (LDH) release as a cell damage hallmark; ii) these events are paralleled by mixed spheroids size reduction and inhibition of soluble CD30 and TNFa shedding; iii) the effects due to ADAM10 inhibitors can be reproduced in LN-derived matrix or collagen scaffolds repopulated with LN-MSC and HL cells; iv) ADAM10 inhibitors exerted a direct anti-lymphoma effect and enhanced the effect of BtxVed. 910

Methods ADAM10 inhibitors LT4, MN8, and cyanine 5.5-conjugated MN8 (CAM36) were synthesized as previously published,16,17,30 compared to GI254023X (GIX, Sigma-Aldrich) and used at 10 mM to 1 mM, alone or with BtxVed ((20-2 mg/mL, Pharmacy Unit, IRCCS Policlinico San Martino).

Spheroids Mixed spheroids of LN-derived MSC and HL cells were prepared as previously published.14,31 L428 or L540 cell lines (DSMZ GmbH), RS773, LN-MSC773, LN-MSC16412 or LN-MSC23274 primary cell lines were stabilized and cultured as previously reported.14 Spheroid dimension was analyzed at 48 hours (h), 72 h and 96 h of culture without or with 10 mM LT4 or MN8, by the CellSens 1.12 software (Olympus).31 Conventional co-cultures were performed with HL cell lines and LN-MSC at the ratio of 10:1 in flat bottom 96-multiwell plates.

Scaffolds Extracellular matrix (ECM) from LN of HL patients (Pathology Unit, IRCCS Policlinico San Martino; IRB approvals 0026910/07, 03/2009, 14/09/15) was prepared as described.32LN-MSC were co-cultured on ECM or AviteneTMmicrofibrillar collagen sponge (Davol Inc.) scaffolds with L428 or L540 cells, in the presence or absence of 10 mM LT4 or MN8, either alone or in combination with 202 mg/mL BtxVed, and analyzed at 48 h, 72 h, 72 h, 96 h and 120 h. Culture supernatants (SN) were recovered for TNFa or soluble CD30 detection.

Confocal microscopy Mixed spheroids were incubated with 10 mM CAM36 for 1 h at 37°C, followed by 1 mM Syto16 (ThermoFisher Scientific) and analyzed in sequence mode with a FV500 confocal microscope (Olympus). Z-stack sections were taken every 2 mm and data was analyzed with FluoView 4.3b software.

Scanning electron microscopy After fixation with 4% paraformaldehyde and 1% osmium tetroxide post-fixation, 10 mm sections from paraffinembedded empty scaffolds or 3D cultures, were collected on glass coverslips, mounted on aluminum stubs and sputter-coated with gold-palladium. The ultrastructure was analyzed on a Hitachi TM3000 Benchtop scanning electron microscopy (SEM) instrument operating at 15 kV acceleration voltage.

Immunohistochemistry and immunofluorescence Five mm serial sections from 3D cultures fixed in Histochoice (Amresco) were stained with rabbit anti-Ki67 antiserum (1:100, Ventana Hoffman-La Roche), anti-CD30 Ber-H2 mAb (2 mg/mL,Ventana), rabbit anti-TGII antiserum (1:100, ThermoScientific), rabbit anti-caspase-3 mAb (1:1,000, Cell Signaling) or an isotypic unrelated antibody (Dako Cytomation). For immunohistochemistry (IHC), biotinylated goat anti-mouse (Biot-GAM) or goat anti-rabbit antiserum (Biot-GAR, BioOptica) was added, followed by horseradish peroxidase (HRP)-conjugated avidin (HRPAv, ThermoScientific) and 3,3′-diaminobenzidine (DAB, Sigma). Immunofluorescence (IF) was performed with antiCD30 mAb, anti-Ki67 antiserum and DAPI, using the autohaematologica | 2022; 107(4)

ADAM10 inhibitors in Hodgkin lymphoma 3D models

mated stainer BOND-Rxm. Slides were observed under a Leica DM-MB2 microscope with a CCD camera (Olympus DP70) or the AperioVERSA or AperioAT2 Scanner and data was analyzed with theAperio Cellular IF Algorithm (Leica Biosystems).33 IHC images were analyzed with the Genie software and the nuclear-count V9 macro (Leica Biosystems).33 ATP, LDH, soluble CD30, TNFa and glucose measurement ATP content was tested using the CellTiter-Glo® Luminescent Kit (Promega Italia).34 LDH was determined using the CytoTox96 Kit (Promega). Soluble CD30 and TNFa were measured by the Picokine ELISA kit (Boster Bio) and the specific cytokine detection kit (PeproTech).34 Glucose was evaluated with the D-glucose Assay (Megazyme), referred to a standard curve.

Statistical analysis Data are presented as mean ± standard error of the mean (SEM) or ± standard deviation (SD). Statistical analysis was performed by two-tailed unpaired Student’s t-test, with Welch correction, using the Graph Pad Prism software 5.0.

Results ADAM10 inhibitors reduce ATP content and the size of Hodgkin lymphoma-mesenchymal stromal cell spheroids In experiments performed under conventional culture conditions, the HL cells RS773 or L428 or L540 (4x105) were cultured for 96 h with medium alone or the vehicle dimethyl sulfoxide (DMSO) (1:1,000), or LT4, MN8 or the commercial inhibitor GIX (10-2.5 mM). ATP content at this time point was consistently lower in the presence of the inhibitors than in the absence (0 mM, culture medium alone) or in the solvent (DMSO) and both LT4 and MN8 were more efficient than GIX, even at low (2.5 mM) concentrations (Online Supplementary Figure S1A). The effect on ATP cell content paralleled an impairment in cell proliferation, as documented by the reduced cell number in the cultures containing ADAM10 inhibitors (Online Supplementary Figure S1B). Online Supplementary Table S1 shows the effect of other three sulfonamido-based hydroxamate compounds, synthesized in our lab, on the ATP intracellular content and TNFa shedding by L428 and L540 cell lines. In the same table, the half maximal inhibitory concentration (IC ) (nM) on ADAM10 or ADAM17 for each compound is also shown to better compare the distinct effects of the inhibitors related to their specificity. In particular, FC410, FC143 and FC130, which are more specific for ADAM17, display IC of 0.1-2.5 mM on TNFa shedding, and IC of 1050 mM on intracellular ATP, while LT4 and MN8 IC 5-10 mM both on TNFa shedding and on intracellular ATP. The IC of GIX (that shows an in vitro IC on the purified ADAM similar to LT4 but, at variance with LT4, has also nM activity on some MMP)17 is 5-10 mM on TNFa shedding, superimposable to that of MN8, but higher (15-20 mM) on intracellular ATP modulation. As a first 3D culture model to test ADAM10 inhibitors in HL microenvironment, mixed spheroids of LN-MSC16412 (2x105) and L540 cells (4x105) were prepared as previously described.31 Figure 1A shows confocal microscopy images of an exemplar mixed spheroid in bright field (left) or after staining with anti-CD30 mAb to identify HL cells (right). Figure 1B shows the confocal analysis of a mixed spheroid 50

incubated with 1 µM Syto16 (blue) to stain nuclei and 10 µM CAM36 (Cy5.5-MN8, red) documenting that the inhibitor reaches the inner part of the 3D structure. The zstack images, taken every 2 mm, depicted in the Online Supplementary Figure S2A and B, confirm this result. Thus, 10 µM (Figure 1C and D) or dilutions (Figure 1E) of LT4 or MN8 were added to the mixed spheroids and cultured for 48 h, 72 h or 96 h. Spheroid dimensions (area in Figure 1C and volume in D) were analyzed in each culture well as described in the Online Supplementary Appendix. In D, the volume of spheroids made of LN-MSC16412 alone is indicated as well. Of note, both LT4 and MN8 could significantly reduce the size of mixed spheroids, and this effect was particularly evident after 96 h (Figure 1C, right, and D). No effect was observed on LN-MSC16412 spheroids (Figure 1 D). In parallel samples, mixed spheroids were exposed to serial dilutions (10-0 mM) of LT4, MN8, GIX or DMSO 1:1,000 for 96 h; then supernatants were recovered for LDH detection and cells lysed for ATP measurement. All ADAM10 inhibitors induced a decrease in ATP cellular content (Figure 1E, left) and an increase in LDH release (Figure 1E, right), with LT4 and MN8 displaying a more efficient effect than GIX. Of note, LT4 or MN8 were effective also in autologous mixed spheroids of LN-MSC773 (2x105) and RS773 cells (4x105), isolated from the same LN and prepared as previously described.16 Both inhibitors (10 mM) could reduce the ATP content at 72 h, and this was more evident at 96 h (Figure 2Aa); at this time point, LT4 and MN8 were more effective than GIX also at 5 mM concentration (Figure 2A and B and the decrease in intracellular ATP was paralleled by a rise in LDH release (Figure 2B) and by a reduction in the secretion of TNFa (Figure 2C). Interestingly, spheroid dimension was significantly lower in the cultures exposed to LT4 at 48h (Figure 2D, left), or MN8 at 72 h (central), at variance to GIX (not shown). In order to verify that the effects of ADAM10 inhibitors on spheroid size and ATP content were mainly directed against HL cells, spheroids made of LN-MSC16412 only were prepared.30 ATP intracellular content, LDH release and size were measured at different time points, in the cultures without (Online Supplementary Figure S3A to C) or with 10 mM LT4 or MN8 (Online Supplementary Figure S3D and E). First, LDH release (A), ATP intracellular content (B) and spheroids size (C) were stable over time. Second, the two ADAM10 inhibitors did neither affect LN-MSC16412 spheroid dimension (D) nor ATP (E), nor LDH release (not shown). Altogether, these data support the hypothesis that blocking of ADAM10 with specific inhibitors can interfere with HL cell growth in a 3D microenvironment composed of stromal and lymphoma cells.

haematologica | 2022; 107(4)

ADAM10 inhibitors decrease CD30 and TNFa shedding in 3D cultures of Hodgkin lymphoma-mesenchymal stromal cells on extracellular matrix scaffolds In order to resemble the architecture of HL more closely, decellularized ECM derived from patient LN were repopulated with LN-MSC16412 (2x105) for 2 days, followed by the addition of 4x105 L428 for further 3 days. Figure 3 shows an example of a repopulated ECM scaffold, with L428 cell identified in IHC with anti-CD30 mAb (A) and LN-MSC16412 stained with anti-TGII antiserum (B). SN were harvested 96 h after addition of 10 mM LT4 or MN8 or GIX for TNFa and soluble CD30 measurement. It is of note that the inhibitors reduced the shedding of TNFa (Figure 911

R. Pece et al.

3C) and CD30 (Figure 3D). Superimposable results were obtained with decellularized LN-derived ECM cultured with MSC16412 and L540 (not shown), confirming that testing ADAM10 inhibitors on 3D cultures of stromal and HL cells on matrix scaffolds from human ECM is feasible. However, ECM obtainable from every LN specimen can barely allow a single/double experiment in triplicate (i.e., three 3D replicates with MSC co-cultered with one HL cell line only). Moreover, this 3D system is difficult to standardize in terms of scaffold size, shape and structure. Thus, we introduced commercial sponges made of microfibrillar collagen (AviteneTM Sponges), used as hemostats in surgery.

These sponges, analyzed by SEM display a 3D structure (Figure 4D to F) similar to that of LN-derived ECM (Figure 4A to C), with a network of round shaped niches, of approximately 10-50 mm of diameter, defined by collagen fibre bundles (arrows). AviteneTM sponges were cut in equalsized scaffolds, and LN-MSC16412 were co-cultured with L540 or L428 HL cells as above. These scaffolds were efficiently repopulated by LN-MSC16412 (TGII+, Figure 5A) and L428 cells (CD30+, Figure 5B) or L540 (not shown) cells after 96 h. Scaffold repopulation was also documented by SEM, where both cell types could be distiguished by morphology (Figure 5C), suggesting that the model was feasible

Figure 1. Effects of ADAM10 inhibitors on lymph node-mesenchymal stromal/Hodgkin lymphoma cell spheroids. (A) Confocal microscopy of mixed spheroids, made of LN-MSC16412 (2x105) and L540 cells (4x105), in bright field (left) or after staining with anti-CD30 monoclonal antibody (mAb), to identify Hodgkin lymphoma (HL) cells (arrows), followed by FITC-GAM (right). (B) Confocal analysis of a mixed spheroid incubated with 1 µM Syto16 (blue) to stain nuclei and 10 µM CAM36 (red): single pseudocolor or merged images and bright field as indicated (FV500 confocal LSM, Olympus, PlanApo 40x NA1.00 oil objective). Images were taken in sequence mode to avoid cross-contribution of each fluorochrome, analyzed with the FluoView4.3b software (Olympus) and shown in pseudocolor or bright field. Scale bar: 10 mm. (C to E) 10 mM (C and D) or dilutions (E) of LT4 or MN8 were added to the mixed spheroids and the cultures were kept at 37°C, for further 48 hours (h) (C, left), 72 h (C, central) or 96 h (C, right, D and E). (C and D) Mixed spheroid dimension (C: area, D: volume) analyzed in each culture well as previously described.31 In (D), also the volume of spheroids made of LN-MSC16412 alone is indicated. Mean ± standard error of the mean (SEM) of triplicates analyzed for each culture condition with a minimum of 50 single spheroids for each sample in three independent experiment. Nil: no drug added. *P<0.01 and **P<0.001 vs. nil. (E) Mixed spheroids exposed to serial dilutions (10-0 mM) of LT4, MN8, GIX or the solvent dimethyl sulpfoxide (DMSO) (1:1,000) for 96 h at 37°C. Left graph: intracellular ATP content (luminescence a.u./104 L540 cells); right graph: lactate dehydrogenase (LDH) detection (O.D.490/104 L540 cells) in the supernatant.*P<0.01 and **P<0.001 vs. dimethyl sulfoxide (DMSO).

912

haematologica | 2022; 107(4)

ADAM10 inhibitors in Hodgkin lymphoma 3D models

and reproducible. Also in this 3D system, 10 mM LT4 and MN8 could significantly reduce the shedding of CD30 (Figure 5D) and TNFa (Figure 5E) by both L428 (left) and L540 (right) HL cells. The anti-shedding effect was evident at 72 h for L428 and at 96 h for L540 cells. These results indicate that ADAM10 inhibitors are functional in 3D cultures recapitulating some features of the HL lymphoma microenvironment, such as ECM and MSC. Furthermore, microfibrillar collagen sponges can substitute LN-derived ECM to allow larger sampling.

LT4 and MN8 lower the number of Ki67+ Hodkin lymphoma cells in 3D repopulated scaffolds In order to analyze HL cell proliferation, AviteneTM sponges repopulated with LN-MSC16412 and L428 or L540 cells and exposed to 10 mM LT4 or MN8, for 72 h and 96 h were paraffin embedded and 5 µm sections were prepared for IF with the anti-Ki67 polyclonal antibody that identifies cycling cells.34 Figure 6A shows representative images of repopulated scaffolds (LN-MSC16412 and L428 cells), cultured for 96 h in medium alone: the anti-CD30 mAb identifies HL cells (red membrane staining) and the anti-Ki67 polyclonal antibody (green nuclear staining) identifies cycling cells. Images from untreated and treated samples were automatically analyzed with the Aperio Cellular IF Algorithm (Leica Biosystems) and the number of CD30+/Ki67+ cells was calculated as described in the Online Supplementary Figure S4. The number of CD30+Ki67+L428 cells was significantly lower in the presence of LT4 or MN8 (Figure 6B); the effect of LT4 was evi-

dent at 72 h (left), while MN8 inhibition was maintained also at 96 h (right). The L540 cell line was less sensitive to the inhibition exerted by the two ADAM10 blockers; nevertheless, the reduction of proliferating HL cells by MN8 was significant at 96 h (Figure 6C).

BtxVed, LT4 and MN8 reduce ATP content, glucose consumption and induce caspase-3 in Hodgkin lymphoma-mesenchymal stromal cell co-cultures Given the reduction of CD30 shedding due to ADAM10 inhibitors, and considering the reported antagonistic effect of soluble CD30 on the therapeutic efficacy of the antibody-drug conjugate (ADC) BtxVed,35 we asked whether in the presence of LT4 or MN8, BtxVed could enhance its antilymphoma effect. First, we set up this experiment in conventional 2D co-cultures of HL and LN-MSC cells. To this aim, L428 or L540 cells were added to LN-MSC16412 and co-cultures were performed in the presence of BtxVed (10 or 1 mg/mL), alone or in combination with 10 mM LT4 or MN8. After 96 h, HL cells were harvested (free of LN-MSC that remained adherent) and counted at the MACS Quant Analyzer 10, while parallel samples were analyzed for ATP content. The two ADAM10 inhibitors, reduced L428 and L540 cell growth by about 50%; this effect was similar to that exerted by 1 mg/mL BtxVed; moreover, LT4 and MN8 could enhance the inhibitory effect of BtxVed (10 µg/mL) by 25% (Figure 7A). Accordingly, the combinatory effect of LT4 or MN8 and BtxVed was detectable in decreasing the content of ATP in HL cells, also with BtxVed used at 1 mg/mL (Figure 7B). Indeed, the inhibitory effect of LT4 or MN8 in combination with BtxVed was 2- or 3-fold that of

B i

Figure 2. Effects of ADAM10 inhibitors on ATP content, lactate dehydrogenase or TNFa release and size of autologous lymph node-mesenchymal stromal/Hodgkin lymphoma cell spheroids. Autologous mixed spheroids of LN-MSC773 (2x105) and RS773 cells (4x105) were prepared as previously described.31 (A to C) 10 mM (panel Ai) or dilutions (panel Aii), B and C) of LT4 or MN8 or GIX were added to the mixed spheroids for 48 hours (h) (Ai) and C), 72 h (Ai)) or 96 h (Ai), Aii) and B). At the indicated time points ATP (Ai) and Aii)), lactate dehydrogenase (LDH) (B) or TNFa (C) were measured by specific assays. Results are expressed as luminescence arbitrary units (a.u./104cells, A) or O.D (a.u./10mcells, B) or pg/mL/104cells (C) and are referred to one representative experiment out of three in triplicate. (D) Spheroid area was measured as previously described31 in each culture well at 48 h (left), 72 h (central), 96 h (right). At least triplicates were analyzed for each culture condition and a minimum of 50 single spheroids for each of three independent experiment. Nil: no drug added (medium with dimethyl sulfoxide 1:1,000). Mean ± standard deviation from three experiments is indicated. *P<0.01 and **P<0.001 vs. nil. .490

haematologica | 2022; 107(4)

913

R. Pece et al.

Figure 3. Decellularized lymph node matrices repopulated with lymph nodemesenchymal stromal and Hodgkin lymphoma cells as 3D culture model to test ADAM10 inhibitors. (A and B) Immunohistochemistry (IHC) of 3D cultures performed on decellularized Hodgkin lymphoma (HL) lymph node (LN) matrices (extracellular matrix [ECM], one reprentative experiment, LN I-19032-16) with 2x105 LN-MSC16412 for 3 days, followed by the addition of 4x105 L428 cells to each ECM/LNMSC16412 scaffold for a further 2 days. (A) 4 mm sections stained with the anti-CD30 monoclonal antibody (mAb), to identify HL (L428) cells, followed by Biot-GAM, HRP-Av and developed with DAB; (B) sections stained with the antiTGII rabbit polyclonal antiserum, to identify LN-MSC16412, followed by BiotGAR, HRP-Av and developed with DAB. Inset in (A) negative control with BiotGAM alone. Slides were counterstained with hematoxylin and analyzed under a Leica DM MB2 microscope (40x objective). (C to D) TNFa C) or soluble CD30 (D) content (pg/mL/106 cells) in the supernatant (SN) recovered after 48 hour (h) from the addition of 10 mM LT4 or MN8 or GIX ADAM10 inhibitors to the 3D cultures, measured by specific enzyme-linked immunosorbant assay. Nil: solvent (dimethylsulfoxide 1:1,000). Results are the mean ± standard deviation from three experiments performed in duplicate with ECM from three HL patients.*P<0.005 vs. nil; **P<0.001 vs. nil.

Figure 4. Structural analysis of lymph node matrices and collagen scaffolds for 3D culture models. Scanning electron microscopy (SEM) images of lymph node (LN) matrix (extracellular matrix [ECM]) specimens obtained, as previously described,32 from one Hodgkin lymphoma (HL) patient (I-19032-16, A and B) or a non-neoplastic LN (I-19273-16, C). SEM of AviteneTMUltrafoam Collagen sponges, at the indicated magnifications, embedded and processed as LN-ECM specimens (D and F). Arrows indicate matrix branches surrounding empty spaces. Magnifications and scale bars are reported in each panel.

914

haematologica | 2022; 107(4)

ADAM10 inhibitors in Hodgkin lymphoma 3D models

A D

Figure 5. AviteneTM microfibrillar collagen scaffolds reconstituted with lymph node-mesenchymal stromal and Hodgkin lymphoma cells as 3D culture model to test ADAM10 inhibitors. AviteneTM scaffolds were cultured with 2x105 LN-MSC16412 for 3 days, followed by 4x105 L428 (A to C and D to E left histograms) or L540 cells (D to E right histograms) for further 2 days, before addition of 10 mM LT4 or MN8 for 72 hours (h) or 96 h. (A) 4 mm sections of repopulated AviteneTM scaffolds stained with anti-TGII polyclonal antiserum followed by Biot-GAR, HRP-Av and developed with DAB; (B) sections stained with anti-CD30 monoclonal antibody (mAb) followed by Biot-GAM, HRP-Av and developed with DAB. Inset in the left pictures: images of the whole repopulated scaffold. Inset in the right pictures: negative control (Nil) with Biot-GAR(A) or Biot-GAM (B) alone. Slides were counterstained with hematoxylin and analyzed under a Leica DM MB2 microscope (left: 20x enlargement, right: 40x enlargement). (C) Scanning electron microscopy (SEM) images of AviteneTM scaffolds, repopulated with LN-MSC16412 and L428 cells (arrows). Magnifications and scale bar are reported in each panel. Inset in the left picture: the whole scaffold with a white square indicating the area enlarged. (D and E) Soluble CD30 (D) or TNFa (E) content (pg/mL/106 cells), measured by specific enzyme-linked immunosorbant assay, in the supernatant (SN) recovered after 72 h or 96 h from addition of ADAM10 inhibitors to the scaffolds repopulated with LN-MSC16412 and L428 (left histograms) orL540 (right histograms). Results are the mean the mean ± standard deviation of quadruplicates from three independent experiments.**P<0.005 vs. nil; ***P<0.001 vs. nil.

BtxVed alone in L428 cells (Figure 7B, left); this effect was still detectable in L540 cells despite its higher sensitivity to BtxVed (Figure 7B, right). In order to closely reproduce the cellularity of a LN node microenvironment, we tried to improve scaffold repopulation by simultaneously seeding a mixture of LN-MSC and HL cells onto the scaffold. This system allowed HL cells to fill the niches between the collagen branches of the scaffold (Online Supplementary Figure S5B), compared to the sequential seeding where empty spaces are still evident (Online Supplementary Figure S5A). Cells maintained their metabolic activity as documented by glucose consumption (Online Supplementary Figure S5C). In this 3D setting, we decided to use BtxVed at 20 mg/mL to maximize its effect on HL cells, compared with one tenth of the concentration (2 mg/mL). In order to avoid sudden cell starvation due to total glucose consumption by 48-72 h (Online Supplementary Figure S5C), half of the medium was replaced at 48 h, after recovering the SN for glucose measurement. Glucose consumption by either L428 (Online Supplementary Figure S5D) or L540 (Online Supplementary Figure S5E) cells rapidly decreased over time in the presence of 20 mg/mL BtxVed, while the effect was less evident with the lower dose of the ADC (2 µg/mL). Glucose consumption in the presence of DMSO haematologica | 2022; 107(4)

dilutions is also shown (Online Supplementary Figure S5F and G; Figure 8E and F). In order to test the hypothesis of a synergic or additive effect of ADAM10 inhibitors and BtxVed on HL cells, we chose the IHC analysis of caspase-3 activation, rather than Ki67 expression, since both the ADC and the inhibitors act mainly by inducing programmed cell death.19,35,36 Moreover, as Ki67 is detectable in the nucleus not only in the mitotic phase but also in the interphase,34 it is possible to miscount Ki67+ cells as proliferating, even though they are in other cell cycle phases. Figure 8 shows scanned images of repopulated scaffolds (A: untreated scaffold, B: scaffold exposed to 20 mg/mL BtxVed) stained with the anti-caspase-3 antibody (subpanels b) and automatically analyzed with the Genie software, combined with the nuclear count V9 macro of Image-Scope software (subpanels c: HL cells in blue and caspase-3+ cells in red). As reported in panel C, the percentage of caspase3+L428 HL cells increased significantly upon treatment with 20 mg/mL BtxVed or 10 mM LT4 or MN8, with a slight additional effect using BtxVed at 20 mg/mL and 10 mM LT4 together. In turn, the percentage of L540 HL cells expressing caspase-3 increased especially upon treatment with BtxVed at 2 mg/mL; of note, both LT4 and MN8 could significantly 915

R. Pece et al.

Figure 6. Effects of ADAM10 inhibitors on Hodgkin lymphoma cell growth in 3D scaffolds repopulated with lymph node-mesenchymal stromal and Hodgkin lymphoma cells. (A) A representative AviteneTM scaffold repopulated by LN-MSC16412 (2x105) and L428 cells (4x105). Sections (5 mm) from paraffin-embedded repopulated scaffolds were stained with DAPI for nuclei (blue), anti-CD30 monoclonal antibody (mAb) followed by anti-mouse Alexa Fluor594 (red) for Hodgkin lymphoma (HL) cells and anti-Ki67 polyclonal antibody followed by anti-rabbit Alexa Fluor488 (green) to identify cycling cells. Images were taken with the Aperio VERSA Digital Pathology Scanner (Leica Biosystems) with a 10x objective. Subpanel b: enlargement of the blue rectangle in subpanel a. Scale bars as indicated. (B and C) AviteneTM scaffolds repopulated with 2x105 MSC16412 and 4x105 L428 (B) or L540 cells (C) and exposed to 10 mM LT4 or MN8 ADAM10 inhibitors for 72 hours (h) or 96 h as indicated. Nil: solvent (dimethyl sulfoxide 1:1,000) were subjected to immunofluorescence (IF) as in (A). At least three sections/scaffold, cut at 15 µm distance, were acquired. Image data were analyzed with the Aperio Cellular IF Algorithm (Leica Biosystems) and the percentage of CD30+/Ki67+ cells was calculated as described in the Online Supplementary Figure S1. Results are the mean ± standard error of the mean (SEM) from three independent experiments performed in duplicate (two scaffolds). *P<0.05 vs. nil; **P<0.005 vs. nil.

enhance the number of caspase-3+ cells when used in combination with 2 mg/mL BtxVed (from about 7-10% with BtxVed alone to 12-16%, Figure 8D). In addition, glucose consumption was measured every 24 h in the supernatant of the 3D cultures, referred to the glucose content in fresh culture medium. In these experiments, BtxVed was used at low doses (2 mg/mL) to emphasize any potential additive effect of ADAM10 inhibitors at 10 mM. Glucose depletion in the medium of LN-MSC23274+L428 repopulated scaffolds increased during time; of note, consumption was significantly reduced at 72 h by BtxVed and at 72 h to 96 h by LT4 and MN8, with an additive effect of BtxVed and LT4 used together already evident at 24 h to 48 h (Figure 8E). In the case of 3D cultures of LNMSC23274+L540 cells (Figure 8F), glucose depletion reached a steady state at 48 h, while the pharmacological effects of ADAM10 inhibitors, were evident at 96 h to 120 h. At these time points, LT4 and MN8 were effective in reducing glucose consumption, at variance with BtxVed; of note, additive effects were observed using combinations of LT4 or MN8 and BtxVed (Figure 8E and F). Glucose consumption did not vary in the presence of DMSO at the same dilution used as solvent, supporting a direct anti-lymphoma action of the drugs (Figure 8E and F). These data indicate a direct effect of ADAM10 inhibitors on HL cell growth in 3D microenvironment and, more importantly, an additive effect to the anti-lymphoma action of BtxVed ADC. 916

Discussion In this work we demonstrate the efficiency of ADAM10 inhibitors in HL growth control in two different 3D culture systems: mixed spheroids made of LN-MSC and HL cells and scaffolds repopulated with both LN-MSC and HL cells. In these 3D models we found that: i) in HL cells LT4 and MN8 reduce ATP content, related to proliferation, and glucose consumption, while increasing LDH release as a marker of cell damage; ii) the two inhibitors lead to mixed spheroids size reduction and inhibition of soluble CD30 and TNFa shedding; iii) both effects can be reproduced in 3D culture systems based on patients LN matrix or AviteneTM collagen scaffolds repopulated with LN-MSC and HL cells; iv) LT4/MN8 enhance the anti-lymphoma effect of BtxVed, evaluated as reduction of proliferation and induction of apoptosis, both in conventional co-cultures and in repopulated scaffolds. The mixed spheroid 3D system, like other spheroids approved as animal-free preclinical models,25-27 displays some advantages such as feasibility, low cost, reproducibility analysis of high sample number at a time. On the other hand, the presence of LN-MSC as a core guarantees a rounded shape of the spheroid, allowing their measurement. Indeed, HL cells alone grow in culture as cell suspension forming little clumps without a real spherical spatial organization. It is of note that in this system, LT4 and MN8 could enter the spheroid and proved to be highly effective haematologica | 2022; 107(4)

ADAM10 inhibitors in Hodgkin lymphoma 3D models

not only in their anti-sheddase activity, but also in the interference with HL cell viability, in terms of ATP content, and cell damage documented by LDH release. These effects were accompanied by a reduction of spheroid size, measured with a tailored image analysis procedure reported in a previous paper,30 as a proof of the limited HL cell growth. The inhibition of soluble CD30 and TNFa shedding was considerable as well; indeed, on one side the release of CD30 would impair the effect of anti-CD30 therapeutic mAb,15,33 on the other side released TNFa would function as a growth factor for HL cells.37,38 It is of note that metabolic impairment, shedding inhibition and spheroid size reduction were obtained with all the three HL cell lines tested, i.e., L428 (derived from pleural effusion of a HL patient), L540 (from bone marrow of a different patient) and RS773 (from a LN of a distinct patient), thus proving that the system can be applied to multiple subjects and the pharmacologic effect of ADAM10 inhibitors can be elicited in HL cells regardless the tumor site from which they derive. With regard to the anti-lymphoma action of ADAM10 inhibitors, it has to be noted that exposure to LT4 and MN8 leads to ADAM10 compartmentalization in endolysosomes possibly interfering with ADAM10 stability due to retention in the degradative pathway while decreasing membrane localization.35 Following their intracellular pathways the inhibitors may encounter different substrates involved in tumor cell growth, not only TNFa in HL, but also mediators linked to cell proliferation, such as Notch1 or receptor associated kinases, limiting their function.1,3,5 haematologica | 2022; 107(4)

Figure 7. ADAM10 inhibitors reduce Hodgkin lymphoma cell growth and ATP content and enhance brentuximab-vedotin effect. (A) L428 (left) or L540 (right) cells (104) were added to LN-MSC16412 (103) previously seeded into 96 microwell plates in the presence of brentuximab-vedotin (BtxVed) (10 mg/mL), alone or in combination with 10 mM LT4 or MN8. After 96 hours (h), 200 ml cell suspension were collected and cells counted at the MACS Quant Analyzer 10 (Miltenyi Biotech GmbH). Results are shown as percentage of cell growth inhibition calculated as the number of cells/well in cultures with the indicated drugs referred to cultures in the solvent (dimethyl sufoxide [DMSO] 1:1,000), as described in the Online Supplementary Appendix. Mean ± standard deviation (SD) of quadruplicates from three independent experiments.*P<0.01 and **P<0.001 vs. BtxVed or LT4 or MN8 alone. (B) Cultures were performed as in (A), with BtxVed used at 10 mg/mL and 1 mg/mL. After 96 h, ATP content measured by specific assay. Results are expressed as percent inhibition of luminescence in cultures exposed to the indicated drugs referred to cultures in the solvent (DMSO 1:1000), as described in the Online Supplementary Appendix. Mean ± SD of quadruplicates from three independent experiments.*P<0.001 vs. BtxVed or LT4 or MN8 alone **P<0.0001 vs. BtxVed alone.

The second 3D system is based on LN extracellular matrix and collagen scaffolds repopulated with LN-derived MSC and HL cells. As a first model, matrices obtained from patient LN specimens were used to test ADAM10 inhibitors. This model is fairly physiological, as the decellularization process allows the removal of cells while maintaining the biochemical composition and tridimensional organization of the tissue of origin32,39 and allows MSC and HL cells to repopulate the structure creating a bona fide, lymphoma microenvironment. In these LN-derived and repopulated ECM, LT4 and MN8 displayed the same anti-sheddase activity observed in the spheroid system, i.e., blocking of CD30 and TNFa release. However, the reduced number of replicates and the difficult standardization of the scaffold size, due to the paucity of bioptic samples and to various shapes of LN-derived ECM, represent limitations of this 3D culture system. Therefore, the microfibrillar collagen sponge AviteneTM, used for hemostatic purposes in surgery, was chosen for further studies. This system allows the preparation of a high number of replicates with homogeneous and reproducible sampling. Interestingly, ultrastructure analyses evidenced that the architecture of these sponges was very similar to that of decellularized matrices obtained from LN biopsies, thus representing a good alternative to test anti-lymphoma drugs, including ADAM10 inhibitors. AviteneTM scaffolds could be actively repopulated by LN-MSC and HL cells, that migrate through the collagen branches into the empty spaces, as shown by SEM. We are aware of the complex LN cellular composition in 917

R. Pece et al.

HL, including multiple types of immunocompetent and inflammatory cells that influence anti-tumor and drug response,38 that cannot be fully recapitulated by LN-MSC. In particular, this has been documented by a multi-center phase II trial with immune checkpoint inhibitors in classical HL that failed both autologous stem-cell transplantation and BtxVed therapy.40 Nevertheless, it has been

reported that fibroblasts from HL lymph node suspensions protect lymphoma cells from BtxVed effects,41 supporting that LN-MSC might represent leading actors in HL response to this ADC. In AviteneTM scaffolds repopulated by LN-MSC and HL cells, the anti-sheddase effect of LT4 and MN8 on CD30 and TNFa was documented using two different HL cell

Figure 8. ADAM10 inhibitors and brentuximab-vedotin induce caspase-3 activation in Hodgkin lymphoma cells and reduce glucose consumption in repopulated scaffolds. AviteneTM repopulated scaffolds (LN-MSC23274+L428 cells), were either untreated (A) and nil (C to F) or exposed to brentuximab-vedotin (BtxVed) (20 mg/mL or 2 mg/mL as indicated) (B to D) or 10 mM LT4 or MN8 (C to F) or BtxVed plus one of the inhibitors as indicated (C to F). (A and B) After 96 hours (h), scaffolds were fixed and 5 mm serial sections were stained in immunohistochemistry (IHC) with the rabbit monoclonal anti-caspase-3 antibody. Images were acquired with the Aperio AT2 Digital Pathology Scanner and data analyzed with the Genie software to identify and count caspase-3+cells. Subpanel a: sections of the whole scaffold; subpanel b: enlargements of squares in subpanel a; subpanel c: Hodgkin lymphoma (HL) cells, identified by morphology (blue), and caspase-3+ cells recognized by the Genie software (red) (C and D) percentage of caspase-3+ cells (C: L428, D: L540) counted in serial sections (3 every 15 mm/each scaffold) by the Genie software, are reported as the mean ± standard deviation (SD) of three serial sections analyzed from three different experiments in duplicate (2 scaffolds/experiment). (C) **P<0.0005 and *P<0.05 vs. nil; (D) **P<0.05 vs. BtxVed and *P<0.05 vs. nil. (E and F) Glucose evaluation in the supernatant (SN) recovered from LNMSC23274+L428 (E) or LN-MSC23274+L540 (F) repopulated scaffolds exposed for the indicated time periods to either 2 mg/mL BtxVed, 10 mM LT4 or MN8, or LT4+BtxVed or MN8+BtxVed. Green symbols: medium containing dimethyl sulfoxide (DMSO) 1:1,000. Glucose was measured with the specific kit (Megazyme) in culture SN recovered every 24 h and data are expressed as percentage glucose consumpion referred to the content in fresh culture medium; mean ± SD from three experiments performed in duplicate (2 scaffolds/experiment). (E) *P<0.02 and **P<0.001 vs. nil; ***P<0.0001 vs. nil and vs. BtxVed. (F) *P<0.01, **P<0.001 and ***P<0.0001 vs. nil and vs. BtxVed.

918

haematologica | 2022; 107(4)

ADAM10 inhibitors in Hodgkin lymphoma 3D models

lines, with a slightly different time course. Actually, the response to ADAM10 inhibitors show a time-dependent biphasic kinetics, at least in the time frame we selected for our observations, while the morphological and biological global response led us to conclude they were effective in reshaping our 3D cultures with an overall, persistent antitumor effect. A possible explanation might be based on the different fusion gene expression involving the hyperactivation of different kinases (ELMO1-SCLO3A1 in L428) related to NFkB regulation. In fact biphasic changes in NFkB as well as TNFa signaling, both inducing either damage and repair like other inflammatory regulators, has been reported to possibly influence the observed effects.42 Another possibility is a different location of ADAM10 in intracellular compartments, that may influence the speed of action of the inhibitors in different cell lines and cell types. We reported that ADAM10 intracellular distribution changes after exposure to LT4 or MN8 functionalized with a linker.35 LT4 and more effectively MN8 induced a substantial reorganization of the intracellular vesicular network, with secretion of extracellular vesicles (EV) carrying the inhibitors.35 It is then possible that LT4 and MN8 can act on CD30 early after uptake, while at a later time point the effect is temporary lost due to extracellular release in EV. Since EV-bound inhibitors are then taken up by tumor and bystander cells, we cannot exclude the possibility that extended biological effects occur later than 96 h. Both inhibitors could reduce the number of cycling HL cells, evaluated by automatic cell count and computerized imaging as the number of CD30+ cells co-expressing the Ki67 marker, in repopulated scaffolds. Also, HL cell metabolism evaluated by glucose consumption, was impaired by ADAM10 inhibitors. Epigenetic effects of DMSO, used for the first solubilization step of LT4 and MN8, on genes mainly controlling glucose metabolism have been reported in human cardiac fibroblasts and primary hepatocytes.43 Nevertheless, in our 2D and 3D culture models glucose consumption did not vary in the presence of DMSO at the same dilution used as solvent, supporting a direct anti-lymphoma action of the drugs. Of note, LT4 and MN8 could enhance the anti-tumor action of BtxVed, rescuing the effect of this ADC at low doses, both in 2D cultures and in the 3D scaffold system, and leading to an increase in the number of caspase-3+ apoptotic HL cells. These effects are conceivably due to the double action of the inhibitors as anti-sheddase on CD30, that is the target of BtxVed,15,16 and TNFa that represents a lymphoma growth factor.37,44 On its own, BtxVed targets CD30 mainly expressed on lymphoma cells, although minimally present also on normal cells, predominantly activated B and T lymphocytes.38 The internalization of the ADC in cell lysosomes results in proteolytic cleavage of the microtubule-disrupting agent monomethyl auristatin E (MMAE), with consequent apoptosis. Nevertheless, the first effect of microtubule disruption may be the impairment of cell proliferation before accelerating apoptosis45 and this might explain why certain HL cell lines, such as L540, display a lower caspase-3 activation in response to the ADC. As a possible undesired effect, MMAE released into the surrounding extracellular matrix might exert toxicity on adjacent normal cells.

haematologica | 2022; 107(4)

Moreover, as MMAE is eliminated through liver and kidney, patients with hepatic diseases or renal failure, who develop severe adverse reaction to full-dose BtxVed, require ADC administration at reduced doses. Also, BtxVed can induce peripheral neuropathy, severe anemia and neutropenia in a fraction of patients regardless of renal and hepatic impairment.36 From this viewpoint, ADAM10 inhibitors could help to contain the ADC dosage in suitable ranges to this purpose. Given the overall efficiency at non toxic doses (5-10 mM) of LT4 and MN8 in eliciting an additive effect on BtxVed anti-lymphoma action in 3D cultures, these compounds might be proposed for a combined therapy. Potential toxicity to normal cells should be considered, as ADAM10 expression is not confined to cancer cells, although it is upregulated in tumors compared to healthy tissues.35 Another limitation to overcome is represented by the need of DMSO for the first solubilization step. However, ADC-based anti-HL therapy is usually based on short repeated cycles of drug administration in order to reduce any potential side effect and this should be useful to limit ADAM10 inhibitors undesired effects as well. In any case, at present the potential therapeutic use of these compounds is only a suggestion since their pharmacodynamics is still to be defined and deserves further studies. In conclusion, our data point toward three main pieces of information: first, a direct anti-lymphoma effect exerted by ADAM10 inhibitors, more efficient than the known commercial inhibitor GIX. Second, the enhancement of BtxVed anti-lymphoma effect, due to a combinatory action of the ADC and the inhibitors, detectable at low and ineffective doses of the ADC. Last, the evidence of these effects in 3D systems, very similar to those approved as preclinical models. Repopulated scaffolds may also represent a starting point to reconstitute the whole LN cellular composition, including inflammatory or endothelial cells that can contribute to HL pathogenesis, influencing drug response as well. Nevertheless, our 3D model based on ECM, stromal cells and lymphoma cells, recapitulates the main aspects of the lymphoma microenvironment architecture, representing a reliable tool for anti-lymphoma drug testing. Disclosures No conflicts of interest to disclose. Contributions RP and ST performed scaffold repopulation, IHC and IF assays; DC performed IHC and image analysis; SV performed experiments with mixed spheroids; CC, DC, EN and AR designed and produced ADAM10 inhibitors; MA performed decelluarization matrix experiments; CD’A and DG performed SEM preparation of samples and analysis; JLR provided LN specimens; MG provided LN specimens and clinical patient information; FT performed ELISA experiments for TNF and soluble CD30 quantitation; AP performed LN-MSC isolation and culture, repopulation of scaffolds, confocal microscope analysis and planned some experiments; MRZ planned, designed and scheduled the experiments. All the authors read and revised the manuscript. Funding This study has been supported by the AIRC IG-17074 grant to MRZ.

919

R. Pece et al.

References 1. Edwards DR, Handsle, MM, Pennington CJ. The ADAM metalloproteinases. Mol Aspects Med. 2008;29(5):258-289. 2. Reiss K, Saftig P. The “A Disintegrin And Metalloprotease” (ADAM) family of sheddases: physiological and cellular functions. Semin Cell Dev Biol. 2009;20(2):126-137. 3. Blobel CP. ADAMs: key components in EGFR signalling and development. Nature Rev Cancer. 2005;6(1):32-43. 4. Rocks N, Paulissen G, El Hour M, et al. Emerging roles of ADAM and ADAMTS metalloproteinases in cancer. Biochimie. 2008;90(2)369-379. 5. Duffy MJ, McKiernan E, O’Donovan N, McGowan P. Role of ADAMs in cancer formation and progression. Clin Cancer Res. 2009;15(4):1140-1144. 6. Murphy G. The ADAMs: Signalling scissors in the tumor microenvironment. Nature Rev Cancer. 2008;8(12):929-941. 7. Duffy MJ, Mullooly M, O'Donovan N, et al. The ADAMs family of proteases: new biomarkers and therapeutic targets for cancer? Clin Proteomics. 2011;8(1):9-13. 8. Saftig P, Reiss K. The “A Disintegrin And Metalloproteases” ADAM10 and ADAM17: novel drug targets with therapeutic potential? Eur J Cell Biol. 2011;90(67):527-535. 9. Zhou BB, Peyton M, He B, et al. Targeting ADAM-mediated ligand cleavage to inhibit HER3 and EGFR pathways in non-small cell lung cancer. Cancer Cell. 2006;10(1):3950. 10. Witters L, Scherle P, Friedman S, et al. Synergistic inhibition with a dual epidermal growth factor receptor/HER-2/neu tyrosine kinase inhibitor and a disintegrin and metalloproteinase inhibitor. Cancer Res. 2008;68(17):7082-7089. 11. Moss ML, Stoeck A, Yan W, Dempsey PJ. ADAM10 as a target for anti-cancer therapy. Curr Pharm Biotechnol. 2008;9(1):2-8. 12. Waldhauer I, Steinle A. Proteolytic release of soluble UL16-binding protein 2 from tumor cells. Cancer Res. 2006;66(5):25202526. 13. Waldhauer I, Goehlsdorf D, Gieseke F, et al. Tumor-associated MICA is shed by ADAM proteases. Cancer Res. 2008;68(15):63686376. 14. Zocchi MR, Catellani S, Canevali P, et al. High ERp5/ADAM10 expression in lymphnode microenvironment and impaired NKG2D-ligands recognition in Hodgkin lymphomas. Blood. 2012;119(6):1479-1489. 15. Eichenauer DA, Simhadri VL, von Strandmann EP, et al. ADAM10 inhibition of human CD30shedding increases specificity of targeted immunotherapy in vitro. Cancer Res. 2007;67(1):332-338. 16. Zocchi MR, Camodeca C, Nuti E, et al. ADAM10 new selective inhibitors reduce NKG2D ligand release sensitizing Hodgkin lymphoma cells to NKG2Dmediated killing. Oncoimmunology. 2015;5(5):e1123367. 17. Camodeca C, Nuti E, Tepshi L, et al. Discovery of a new selective inhibitor of A

920

Disintegrin And Metalloprotease 10 (ADAM10) able to reduce the shedding of NKG2D ligands in Hodgkin's lymphoma cell models. Eur J Med Chem. 2016;111:193-201. 18. Francisco JA, Cerveny CG, Meyer DL, et al. cAC10-vcMMAE, an anti-CD30monomethyl auristatin E conjugate with potent and selective antitumor activity. Blood. 2003;102(4):1458-1465. 19. Steidl C, Connors JM, Gascoyne RD. Molecular pathogenesis of Hodgkin’s lymphoma: Increasing evidence of the importance of the microenvironment. J Clin Oncol. 2011;29(14):1-15. 20. Montes-Moreno S. Hodgkin’s Lymphomas: a tumor recognized by its microenvironment. Adv Hematol. 2011;2011:142395. 21. Hirata E, Sahai E. Tumor microenvironment and differential responses to therapy. Cold Spring Harb Perspect Med. 2017;7(7):a026781. 22. Wu T, Dai Y. Tumor microenvironment and therapeutic response. Cancer Lett. 2017; 387:61-68. 23. Akhtar A. The flaws and human harms of animal experimentation. Camb Q Health Ethics. 2015;24(4):407-419. 24. Enna SJ, Williams M. Defining the role of pharmacology in the emerging world of translational research. Adv Pharmacol. 2009;57:1-30. 25. Ellis LM, Fidler IJ. Finding the tumor copycat. Therapy fails, patients don't. Nat Med. 2010;16(9):974-975. 26. Zanoni M, Piccinini F, Arienti C, et al. 3D tumor spheroid models for in vitro therapeutic screening: a systematic approach to enhance the biological relevance of data obtained. Sci Rep. 2016;6:19103. 27. Rodrigues T, Kundu B, Silva-Correia J, Kundu. Emerging tumor spheroids technologies for 3D in vitro cancer modeling. Pharmacol Ther. 2018;184:201-211. 28. Sant S, Johnston PA. The production of 3D tumor spheroids for cancer drug discovery. Drug Discov Today Technol. 2017;23:2736. 29. Verjans ET, Doijen J, Luyten W, Landuyt B, Schoofs L. Three-dimensional cell culture models for anticancer drug screening: Worth the effort? J Cell Physiol. 2018;233(4):2993-3003. 30. Camodeca C, Nuti E, Tosetti F, et al. Synthesis and in vitro evaluation of ADAM10 and ADAM17 highly selective bioimaging probes. Chem Med Chem. 2018;13(19):2119-2131. 31. Varesano S, Zocchi MR, Poggi A. Zoledronate triggers Vδ2 T cells to destroy and kill spheroids of colon carcinoma: quantitative image analysis of three-dimensional cultures. Front Immunol. 2018;9:998. 32. Nebuloni M, Albarello L, Andolfo A, et al. Insight on colorectal carcinoma infiltration by studying perilesional extracellular matrix. Sci Rep. 2016;6:22522. 33. Caitríona L, Darragh L. Aperio Cellular IF AlgorithmValidation. Leica Biosystems White Paper Series. 34. Sun X, Kaufman PD. Ki-67: more than a proliferation marker. Chromosoma.

2018;127(2):175-186. 35. Tosetti F, Venè R, Camodeca C, et al. Specific ADAM10 inhibitors localize in exosome-like vesicles released by Hodgkin lymphoma and stromal cells and prevent sheddase activity carried to bystander cells. Oncoimmunology. 2018;7(5):e1421889. 36. Scott LJ. Brentuximab vedotin: a review in CD30-positive Hodgkin lymphoma. Drugs. 2017;77(4):435-445. 37. Nakayama S, Yokote T, Tsuji M, et al. Expression of tumour necrosis factor-α and its receptors in Hodgkin lymphoma. Br J Haematol. 2014;167(4):574-577. 38. Renner C, Stenner F. Cancer immunotherapy and the immune response in Hodgkin lymphoma. Front Oncol. 2018;8:193. 39. Genovese L, Zawada L, Tosoni A, et al. Cellular localization, invasion, and turnover are differently influenced by healthy and tumor-derived extracellular matrix. Tissue Eng Part A. 2014;20(1314):2005-2018. 40. Younes A, Santoro A, Shipp M, et al. Nivolumab for classical Hodgkin's lymphoma after failure of both autologous stem-cell transplantation and brentuximabvedotin: a multicentre, multicohort, single-arm phase 2 trial. Lancet Oncol. 2016;17(9):1283-1294. 41. Bankov K, Döring C, Ustaszewski A, et al. Fibroblasts in nodular sclerosing classical hodgkin lymphoma are defined by a specific phenotype and protect tumor cells from brentuximab-vedotin induced Injury. Cancers. 2019;11(11):1687. 42. Wang A, Ashton R, Hensel JA, et al. RANKL-targeted combination therapy with osteoprotegerin variant devoid of TRAIL binding exerts biphasic effects on skeletal remodeling and antitumor immunity. Mol Cancer Ther. 2020;19(12):25852597. 43. Verheijen M, Lienhard M, Schrooders Y, et al. DMSO induces drastic changes in human cellular processes and epigenetic landscape in vitro. Sci Rep. 2019;9(1): 4641. 44. Abreu M, Basti A, Genov N, Mazzoccoli G, Relógio A. The reciprocal interplay between TNFa and the circadian clock impacts on cell proliferation and migration in Hodgkin lymphoma cells. Sci Rep. 2018;8(1):11474. 45. Schönberger S, van Beekum C, Götz B, et al. Brentuximab vedotin exerts profound antiproliferative and pro-apoptotic efficacy in CD30-positive as well as cocultured CD30-negative germ cell tumour cell lines. J Cell Mol Med. 2018;22(1):568-575. 46. Nuti E, Casalini F, Santamaria S, et al. Selective arylsulfonamide inhibitors of ADAM-17: hit optimization and activity in ovarian cancer cell models. J Med Chem. 2013;56(20):8089-8103. 47. Nuti E, Casalini F, Avramova SI, et al. Potent arylsulfonamide inhibitors of tumor necrosis factor-alpha converting enzyme able to reduce activated leukocyte cell adhesion molecule shedding in cancer cell models. J Med Chem. 2010;53(6):26222635.

haematologica | 2022; 107(4)

ARTICLE

Plasma Cell Disorders

DIS3 mutations in multiple myeloma impact the transcriptional signature and clinical outcome

Ferrata Storti Foundation

Katia Todoerti,1,2* Domenica Ronchetti,2* Vanessa Favasuli,2 Francesco Maura,3 Fortunato Morabito,4,5 Niccolò Bolli,1,2 Elisa Taiana1,2# and Antonino Neri1,2# 1 Hematology, Fondazione Cà Granda IRCCS Policlinico, Milan, Italy; 2Department of Oncology and Hemato-oncology, University of Milan, Milan, Italy; 3Myeloma Program, Sylvester Comprehensive Cancer Center, University of Miami, Miami, FL, USA; 4 Biotechnology Research Unit, Aprigliano, A.O./ASP, Cosenza, Italy and 5Department of Hematology and Bone Marrow Transplant Unit, Augusta Victoria Hospital, Jerusalem, Israel.

Haematologica 2022 Volume 107(4):921-932

*KT and DR contributed equally as co-first authors. # ET and AN contributed equally as co-senior authors.

ABSTRACT

IS3 gene mutations occur in roughly 10% of patients with multiple myeloma (MM); furthermore, DIS3 expression can be affected by monosomy 13 and del(13q), which occur in approximately 40% of MM cases. Despite several reports on the prevalence of DIS3 mutations, their contribution to the pathobiology of MM remains largely unknown. We took advantage of the large public CoMMpass dataset to investigate the spectrum of DIS3 mutations in MM and its impact on the transcriptome and clinical outcome. We found that the clinical relevance of DIS3 mutations strictly depended on the co-occurrence of del(13q). In particular, bi-allelic DIS3 lesions significantly affected progression-free survival, independently of other predictors of poor clinical outcome, while mono-allelic events mostly affected overall survival. As expected, DIS3 mutations affect the MM transcriptome involving cellular processes and signaling pathways associated with RNA metabolism, and the deregulation of a large number of long non-coding RNA, among which we identified five distinct transcripts as independent predictors of poorer overall survival and nine of worse progression-free survival, with two (AC015982.2 and AL445228.3) predicting both unfavorable outcomes. These findings strongly prompt further studies investigating the relevance of these long non-coding RNA in MM.

Introduction Multiple myeloma (MM) is a hematologic malignancy that is still incurable despite the remarkable improvements in treatment and patients’ care.1 MM is characterized by a profound genomic instability involving ploidy, structural rearrangements, and a wide array of mutations affecting both putative oncogenes and tumor suppressor genes, such as KRAS, NRAS, TP53, BRAF, TRAF3, FAM46C, and DIS3.2-7 These abnormalities are predicted to influence the biological and clinical behavior of the tumor and yet, despite a clear driver role, only a few carry prognostic value.8-10 Among mutated genes, DIS3 deserves great attention; this gene maps to 13q22.1 and encodes for a highly conserved ribonuclease indispensable for survival in vertebrates.11 DIS3 is a multidomain protein with two different catalytic activities: a 3’–5’ exonucleolytic activity via the RNase II/R (RNB) domain and an endonucleolytic activity via the PilT N-terminal (PIN) domain at the N-terminus.11,12 DIS3 provides catalytic activity to the exosome, a multi-subunit complex involved in RNA degradation and metabolism, including mRNA quality control, gene expression regulation, and small RNA processing.12-15 DIS3 mutations and altered expression have been reported in roughly 10% of MM patients.6,16,17 Moreover, DIS3 expression can be affected by monosomy 13 and del(13q), which occur in approximately 40% of MM patients.18,19 DIS3 muta-

haematologica | 2022; 107(4)

Correspondence: ANTONINO NERI antonino.neri@unimi.it Received: January 14, 2021. Accepted: April 22, 2021. Pre-published: May 6, 2021. https://doi.org/10.3324/haematol.2021.278342

921

K. Todoerti et al.

tions are mainly located within the major ribonuclease domains of the protein, probably impairing DIS3 catalytic activity, as has been demonstrated to occur for some mutated residues.11 Recently, the analysis of a large cohort of MM patients enrolled in the Multiple Myeloma Research Foundation (MMRF) CoMMpass study pointed out that DIS3 mutations are almost exclusively missense, with no nonsense or frameshift mutations, copy neutral loss of heterozygosity, or bi-allelic deletions. In addition, a third of the mutations occur in three codons (D479, D488, and R780) within the RNB domain, almost never associated with del(13q); in contrast, the remaining two-thirds of DIS3 mutations are distributed across the various exons, almost always associated with del(13q).20 Despite the detailed overview of DIS3 mutations, their functional consequences on MM pathogenesis remain largely unknown, to the point that it is not clear whether DIS3 acts as an oncogene or a tumor suppressor gene.21,22 A crucial role for the DIS3 ribonuclease has been demonstrated in human MM cell lines, in which it promotes the maturation of the let-7 microRNA tumor suppressor family; indeed, through the reduction of mature let-7, DIS3 inactivation enhances the translation of let-7 targets such as MYC and RAS, leading to enhanced tumorigenesis.23 However, the extent to which this pathway is affected in MM patients harboring DIS3 mutations and how it may contribute to myelomagenesis need to be investigated further. The clinical relevance of DIS3 mutations in MM was initially investigated by Weissbach et al;22 despite the caution due to the small size of the cohort investigated, their data suggested that response to therapy was affected by DIS3 mutations depending on the presence of such mutations in minor rather than major subclones. Recently, the analysis of a larger cohort of MM cases at diagnosis revealed that DIS3 mutations were significantly associated with shorter event-free survival, showing an even worse outcome when in association with 13q deletion. However, overall survival (OS) was not affected.16 Notably, DIS3 mutations retained their significance for event-free survival in a multivariate analysis including t(4;14) and the high-risk “double hit” group, i.e. patients with bi-allelic TP53 inactivation and amplification of 1q21.10 However, it is becoming a well-established notion that, given the complexity of the genetic background in MM, it is mandatory to assess DIS3 mutations in association with other oncogenic events, and their transcriptomic consequences, to establish their impact on MM prognosis.10,22,24 Based on these considerations, our study mined genomic and transcriptomic data from cases included in the MMRF CoMMpass dataset to dissect the impact of DIS3 mutations in the context of the MM genomic landscape in order to improve the definition of relevant clinical subsets. In addition, we investigated the transcriptomic profile related to DIS3 mutations to elucidate its role in myeloma cells and identify relevant pathways in MM pathobiology.

Methods Multi-omics data in the CoMMpass study Multi-omics data regarding bone marrow MM samples at baseline (BM_1) were freely accessible from the MMRF CoMMpass study (https://research.themmrf.org/) and retrieved from the Interim Analysis 12a (MMRF_CoMMpass_IA12a). Details about molecu-

922

lar and clinical data of the CoMMpass cohort selected for the present study are described in the Online Supplementary Methods.

Statistical and survival analyses Fisher exact test was applied to verify the association between genomic alterations in stratified MM cases. The two-tailed P-value was corrected using the Benjamini-Hochberg method, and adjusted P-values <0.05 were considered statistically significant. Survival analyses were performed using survival and survminer packages in R Bioconductor (version 3.5.1) as reported in the Online Supplementary Methods.

Differential expression analysis of the CoMMpass cohort A global dataset from 655 MM cases was stratified on the basis of DIS3 mutation RNA frequency and 56 DIS3-mutated cases with at least 20% RNA mutational load were considered in the differential expression analysis. Further steps of analysis are described in the Online Supplementary Methods. Principal component analysis was applied on differentially expressed (DE) annotated transcripts by means of the prcomp function in R. Volcano plots were used in R to represent significantly up- or down-regulated transcripts. The heatmap of the expression levels of the top 100 DE annotated transcripts, according to the limma B statistics value, was created using dChip software.25

Functional enrichment analysis on differentially expressed protein-coding genes Gene set enrichment analysis (GSEA)26 was performed on the pre-ranked DE protein-coding gene lists based on the fold change values by computing 1,000 permutations and using default analysis conditions. Further details about GSEA are provided in the Online Supplementary Methods.

Long non-coding RNA expression validation We investigated long non-coding (lnc)RNA expression in a proprietary dataset that includes 43 MM patients (Online Supplementary Table S1), upon written informed consent (Ethical Committee approval n. 575, 03/29/2018, Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico). Further details about this analysis are reported in the Online Supplementary Methods.

Results Assessment of DIS3 mutations in myeloma patients We focused on a CoMMpass cohort of 930 bone marrow plasma cell samples from newly diagnosed MM patients for whom non-synonymous somatic mutation data, gained by whole exome sequencing, were available (Online Supplementary Table S2), identifying 103 DIS3 mutations in 94 of the 930 cases (Figure 1A, Online Supplementary Table S3). The variant allelic frequency ranged between 5.3% and 100% (mean: 48%; median: 43%). The majority of mutations were missense variants (100/103) in the coding region (95/103) or within the region of the splice site (2/103), whereas in a minority of cases they consisted in start-lost (3/103). Three mutations were classified as splice acceptor or donor variants involving intronic sequences (3/103), all by means of SnpEff&SnpSift tools (http://snpeff.sourceforge.net/). Strikingly, DIS3 mutations mainly occurred in the active domains of the protein; in detail, 70/103 mutations fell within the RNB domain and ten in the PIN domain. With regard to the main mutational hotspots reported to occur within the RNB domain, R780 haematologica | 2022; 107(4)

Transcriptomic impact of DIS3 mutations in MM

Figure 1. Distribution of DIS3 non-synonymous somatic variants in 94 patients with multiple myeloma of the CoMMpass cohort and the correlation of the variants with other genetic alterations. (A) Frequency of the 100 missense single nucleotide variants in DIS3 protein domains schematized below the histogram; the splice acceptor or donor variants occurring in intronic regions are indicated by red arrows. (B) Plot of co-occurrence of main copy number alterations, IGH translocations and non-synonymous somatic mutations in 651 cases of multiple myeloma in the CoMMpass dataset entirely profiled by whole exome sequencing and RNA-sequencing.

haematologica | 2022; 107(4)

923

K. Todoerti et al.

Table 1. Cox regression univariate analysis in 630 patients with multiple myeloma for whom all data were available.

Variable ISS I ISS II ISS III del(13q) DIS3mut del(13q)/DIS3mut del(13q) + DIS3mut TP53.alterations (del(17p)/TP53 or TP53mt) 1q21 gain/amp TP53.alterations + 1q21 gain/amp N-RAS mut K-RAS mut BRAF mut RAS/BRAF mut TRAF3 mut FAM46C mut del(1p)/CDKN2C HD WHSC1-FGFR3.trx MAF.trx MYC.trx

N (%) 223 (35.4%) 226 (35.9%) 181 (28.7%) 281 (44.6%) 17 (2.7%) 298 (47.3%) 50 (7.9%) 53 (8.4%) 207 (32.9%) 24 (3.8%) 142 (22.5%) 155 (24.6%) 49 (7.8%) 310 (49.2%) 48 (7.6%) 64 (10.2%) 190 (30.2%) 359 (57.0%) 86 (13.7%) 40 (6.3%) 26 (4.1%)

OS Univariate Cox Analysis adj.P-value HR (95% CI) 2.59E-06 7.13E-01 1.92E-06 9.80E-02 9.80E-02 3.71E-02 1.45E-01 9.42E-01 9.80E-02 1.60E-04 9.42E-01 4.82E-01 2.64E-01 3.60E-01 1.10E-01 4.82E-01 7.43E-02 9.80E-02 3.60E-01 1.03E-01 1.10E-01

0.26 (0.15-0.43) 1.09 (0.76-1.57) 2.60 (1.83-3.70) 1.46 (1.03 -2.07) 2.41 (1.06 - 5.49) 1.63 (1.14 -2.32) 1.60 (0.93-2.74) 0.96 (0.50 -1.84) 1.45 (1.02 - 2.07) 3.64 (2.00 -6.62) 1.02 (0.65-1.50) 1.18 (0.80-1.74) 1.49 (0.84-2.65) 1.22 (0.86-1.74) 0.39 (0.14-1.05) 1.27 (0.73-2.21) 1.55 (1.08 -2.23) 0.70 (0.50-0.99) 1.29 (0.82-2.05) 1.78 (1.00-3.16) 1.91 (0.97-3.78)

PFS Univariate Cox Analysis adj.P-value HR (95% CI) 2.00E-07 6.08E-01 2.00E-07 2.77E-01 2.94E-01 1.52E-01 4.03E-02 7.63E-01 5.63E-02 4.80E-03 6.31E-01 5.79E-01 4.49E-01 6.48E-01 6.48E-01 4.49E-01 3.27E-01 6.60E-02 6.60E-02 3.83E-01 2.64E-01

0.42 (0.31-0.57) 1.09 (0.85-1.42) 2.08 (1.62-2.68) 1.21 (0.94 -1.55 ) 1.67 (0.83 -3.39) 1.27 (0.99 -1.63) 1.71 (1.15-2.54) 0.93 (0.58 -1.49) 1.38 (1.07-1.78) 2.56 (1.49-4.4) 0.91 (0.68-1.23) 1.12 (0.84-1.48) 1.26 (0.81-1.96) 1.07 (0.83-1.37) 0.88 (0.54-1.44) 1.22 (0.82-1.83) 1.20 (0.92-1.56) 0.75 (0.58-0.96) 1.47 (1.06-2.03) 1.32 (0.83-2.09) 1.55 (0.90-2.66)

Number (N) and percentage (%) of positive cases are indicated for each variable with the hazard ratio and 95% confidence interval. P values that were statistically significant after Benjamini-Hochberg adjustment are reported in bold. OS: overall survival; PFS: progression-free survival; HR: hazard ratio; 95% CI: 95% confidence interval; ISS: International Staging System; del: deletion; mut: mutation; amp: amplification; HD: hyperdiploid; trx: translocation.

was involved in 12 MM cases, whereas D488 and D479 residues were affected in 11 MM samples each (Figure 1, Online Supplementary Table S3). We evaluated the association of DIS3 mutations with the presence of the main IGH chromosomal translocations in 724 MM cases (Online Supplementary Table S2). A significant co-occurrence was observed both with t(4;14) and MAF (MAF, MAFB, or MAFA) translocations (P=0.0035). The association of DIS3 mutations with copy number alterations commonly found in MM disease was evaluated in 847 MM cases, for which specific data were available by whole exome sequencing (Online Supplementary Table S2). As expected, a very significant association (P=0.0003) was observed between DIS3 mutations and del(13q): specifically, 62 of 86 patients harboring a DIS3 mutation for whom data on copy number alterations were available showed a 13q deletion (Online Supplementary Table S4). Next, we considered the distribution of del(13q) across the different types of DIS3 mutations. Only nine of 29 MM affected by the mutational hotspots also carried del(13q); notably, the D479 hotspot was involved in five of these nine cases. In contrast, 50 of the 54 cases harboring not-hotspot DIS3 mutations were associated with del(13q) (P<0.0001). Concerning the other main copy number alterations, a very significant association (P=0.0020) was observed with 1q21 gain/amplification (1q gain/amp), occurring in 45 of 86 DIS3-mutated patients, 38 of whom also carrying del(13q). In contrast, a highly significant inverse correlation was found with the hyperdiploid condition (P=0.0003) and the occurrence of 1p22/CDKN2C loss (P=0.0205). Moreover, 17p13/TP53 deletions were present in only four 924

of the 86 cases harboring DIS3 mutations (Online Supplementary Table S4). We then looked at the co-occurrence of DIS3 mutations with the most frequently mutated genes in MM, i.e., KRAS, NRAS, BRAF, FAM46C, TP53, and TRAF3 (Online Supplementary Table S2). A statistically significant association with DIS3 mutations was only observed for BRAF mutations (P=0.0126) (Online Supplementary Table S4). Figure 1B displays the global landscape of co-occurrence of DIS3 mutations with other main molecular lesions.

Correlation of DIS3 mutations with clinical parameters We tested the clinical impact of DIS3 mutations in 930 MM cases with available PFS and OS data. At a median follow-up of 889 and 868 days for PFS and OS, respectively, a significantly lower survival rate for both OS (log-rank P=0.039) and PFS (log-rank P=0.021) was observed in 94 DIS3-mutated cases compared to 836 DIS3 wild-type MM cases. Specifically, the median PFS was 800 days in DIS3-mutated cases versus 1176 days in wild-type MM cases (Figure 2A, B), whereas the OS evaluated at 3 years was 65% in DIS3-mutated cases versus 79% in the wildtype group. As described above, DIS3 mutations co-occur significantly with del(13q) and therefore with the loss of the second allele, a finding affecting several other genes in MM. Such a result prompted us to stratify the cases in the CoMMpass series into four groups according to the absence of both mutated DIS3 and del(13q) (381 cases); the presence of only a DIS3 mutation (24 cases); the del(13q) alone (380 cases); or the occurrence of both a DIS3 mutahaematologica | 2022; 107(4)

Transcriptomic impact of DIS3 mutations in MM

Figure 2. Impact of DIS3 mutations on clinical outcomes. (A, B) Kaplan-Meier survival curves of 930 patients with multiple myeloma (MM) stratified according to the occurrence of DIS3 mutations with respect to overall survival (A) and progression-free survival (B). (C, D) Kaplan-Meier survival curves of 847 patients stratified into four molecular groups according to the presence of DIS3 mutation and del(13q) as single or bi-allelic alterations, with respect to overall survival (C) and progression-free survival (D). Log-rank test P-value measuring the global difference between survival curves and numbers of samples at risk in each group across time are reported. Log-rank test P-values of pairwise comparisons are also reported in (C, D). Statistically significant adjusted P-values following Benjamini-Hochberg (P<0.05) are in bold red. The median follow-up time for the overall survival analysis was 862 days (interquartile range, 594-1,092 days) while that for progression-free survival was 828 days (interquartile range, 535-1,094 days). OS: overall survival; PFS: progression-free survival; del: deletion; mt: mutated; WT: wild-type.

tion and del(13q) (62 cases). As depicted in Figure 2C, D, the presence of the bi-allelic alterations was associated with a poor prognosis in comparison to the wild-type condition, with respective 3-year OS rates of 62% versus 82%, and a median PFS of 772 days versus 1215 days. Furthermore, we tested the prognostic impact on both PFS and OS of DIS3 mutations and del(13q), as single or biallelic lesions, along with other parameters, in 630 MM patients for whom all the information was available (Table 1). A significantly increased risk of death or progression was associated with International Staging System (ISS) stage III and the occurrence of 1q gain/amp in association haematologica | 2022; 107(4)

with TP53 alterations. Moreover, a higher risk of having a shorter OS was also associated with DIS3 mutations or del(13q) as single events (mono-allelic condition). In contrast, the bi-allelic condition was associated with a shorter PFS. In contrast, ISS stage I was associated with significantly shortened time to both death and progression (Table 1). Notably, all these features retained their independent prognostic power both for OS and PFS when tested in Cox regression multivariate analysis (Figure 3). Based on such differential effects of mono- or bi-allelic DIS3 lesions, we re-analyzed the association of these events with the known MM oncogenic lesions. Regarding 925

K. Todoerti et al.

Figure 3. Impact of DIS3 mutations and other clinical/molecular variables on survival of patients with multiple myeloma. (A, B) Forest plots of Cox regression multivariate analyses considering all features with adjusted P-value <0.05 in univariate analysis with regards to overall survival (A) and progression-free survival (B), in 630 patients with multiple myeloma from the CoMMpass cohort for whom all considered data were available. The hazard ratio, 95% confidence interval and P-value are reported for each variable. A global log-rank P-value is reported for each analysis. OS: overall survival; PFS: progression-free survival; ISS: International Staging System.

the main IGH translocations, t(4;14) was exclusively related to the bi-allelic condition (P=0.0170). Furthermore, a higher prevalence of 1q gain/amp was observed in patients with bi-allelic DIS3 events (61% vs. 29%). Finally, a larger fraction of FAM46C mutated cases was evidenced in cases carrying only a DIS3 mutation (5.3% vs. 3.1%) (Online Supplementary Table S5). Finally, we attempted to clarify further the prognostic relevance of DIS3 mutations in relation to well-established first-line regimens, as described in the CoMMpass dataset. In this respect, 930 MM patients (94 of whom harboring DIS3 mutations) were grouped according to the type of therapy they had received: bortezomib/ immunomodulatory drug (IMID)-based, bortezomib-based, IMID-based, or carfilzomib-based (i.e., combining IMID/carfilzomib, bortezomib/IMID/carfilzomib, or single-agent carfilzomib schedules). Overall, carfilzomib-based regimens significantly improved both OS and PFS compared to all other types of combination therapies (Online Supplementary Figure S1). However, no significant difference in either OS or PFS was detected among DIS3-mutated cases irrespective of the specific therapy scheme (Online Supplementary Figure S2). These results suggest that DIS3 mutations could negatively affect carfilzomib-based therapies. However, further validation in a larger prospective cohort of patients is warranted.

Transcriptional expression changes associated with DIS3 mutations and del(13q) To define the gene expression signatures and molecular pathways associated with DIS3 mutations, we focused on cases with expression of the mutation at meaningful levels, thus opting for a stringent cut-off of 20% on RNA 926

mutational load (RNA_ALT_FREQ), chosen on the basis of analysis of the receiving operating characteristic curve (area under the curve, 98%) according to DNA mutational level (variant allele frequency) (Online Supplementary Table S3, Online Supplementary Figure S3A). In further support of this cut-off value, we found a significant global correlation (Pearson correlation r=0.94) between DNA and RNA mutational levels in those 56 MM cases with >20% DIS3 RNA mutational load, compared to the remaining 17 MM cases with lower RNA somatic variant frequencies (r=0.30) (Online Supplementary Figure S3B, C). Furthermore, significantly shorter OS (log-rank P=0.013) and PFS (logrank P=0.0013) were observed in the 56 MM cases with a >20% DIS3 RNA mutational load compared to the 836 DIS3 wild-type cases. To note, no significant difference was appreciated between DIS3 sub-clonal mutations (i.e., <20% DIS3 RNA mutational load) and wild-type cases (Online Supplementary Figure S4). For the analysis, we therefore considered a total of 28,346 expressed transcripts and compared the expression profiles in 56 DIS3-mutated cases versus 582 DIS3 wildtype MM cases. A list of 7,167 DE transcripts, 6,564 of which annotated, was obtained at a low stringency cut-off (false discovery rate <10%). Among them, 3,464 proteincoding genes and 2,062 lncRNA resulted mostly upregulated (79%) in DIS3-mutated patients compared to unmutated patients (Online Supplementary Table S6, Online Supplementary Figure S5). Principal component analyses based on the expression levels of the 6,564 DE annotated transcripts stratified according to the del(13q) aberration confirmed the strong association between DIS3 mutations and del(13q) (Figure 4A). Furthermore, the heatmap of the top 100 most significantly upregulated transcripts, almost haematologica | 2022; 107(4)

Transcriptomic impact of DIS3 mutations in MM

C Figure 4. Transcriptional expression changes associated with DIS3 mutations and del(13q). (A) Principal component analyses of 6,564 differentially expressed transcripts in 56 cases of multiple myeloma (MM) with >20% mutated DIS3 (DIS3mt) versus 582 DIS3 wild-type (DIS3 WT) cases. (B) Heatmap of the top 100 differentially expressed transcripts according to B statistics value, in 56 DIS3 >20% mutated versus 582 DIS3 WT MM cases. The colored scaled bar represents standardized rows by subtracting the mean and dividing by the standard deviation. Samples in each group are further stratified according to the occurrence of del(13q) aberration: 5 DIS3mt and not available (nd) for del(13q) MM cases, 38 MM cases with bi-allelic alteration, 13 MM carrying only DIS3mt, 289 MM with del(13q) as a single lesion and 293 WT MM cases. (C) Venn diagram of differentially expressed transcript lists resulting from 13 MM with DIS3 mutation, 289 MM with del(13q) as a single lesion, or 38 MM carrying bi-allelic alteration compared to 293 WT MM cases.

all involving lncRNA (82%), in the stratified MM samples revealed a stronger pattern of positive regulation in the biallelic condition with respect to DIS3 mutations or del(13q) alone, compared to DIS3 wild-type MM cases (Figure 4B). Based on these findings, we compared the global expression profiles associated with each lesion (13 cases with DIS3 mutations, 289 cases with del(13q), and 38 cases with bi-allelic alteration) to the wild-type condition (293 cases). We found a shared set of 430 DE transcripts (of which 405 annotated) in all three comparisons, 305 of which were lncRNA, whereas 56 were classified as protein-coding genes (Figure 4C). Interestingly, 295 out of 405 transcripts showed a cumulative effect of deregulation haematologica | 2022; 107(4)

across patients with mono-allelic and bi-allelic lesions (Online Supplementary Table S7).

Protein-coding genes: molecular pathways and gene sets modulated in association with DIS3 mutations In order to define which molecular pathways could be modulated in relation to the occurrence of DIS3 mutations in MM, GSEA was performed on the list of DE proteincoding genes that were ranked based on fold change values (Online Supplementary Table S6). The enrichment map on the top 100 GSEA gene sets based on gene ontology biological process terms revealed a complex network of connected functional modules 927

K. Todoerti et al. Table 2. Summary information on the 12 long non-coding RNA significant in multivariate analysis.

Gene stable ID Gene name * ENSG00000232519 AL353807.2 ENSG00000235919 ASH1L-AS1 ENSG00000271991 AC013400.1 ENSG00000271387 AL445228.2 ENSG00000272606 AC015982.2 ENSG00000236206 AL356441.1 ENSG00000255647 AC093510.1 ENSG00000259775 AL138976.2 ENSG00000260236 AC099778.1 ENSG00000272426 BX284668.6 ENSG00000272716 AL121658.1 ENSG00000273355 AP000894.4

Description #

Chr.

Start (bp)

End (bp)

Strand

Neighbor Gene Name**

Pearson's correlation

novel transcript, AS to MSTO1 ASH1L AS RNA 1 [Source:HGNC Symbol; Acc:HGNC:44146] novel transcript, AS to TTC32 novel transcript, AS to C1orf21 novel transcript, AS to PP4R3B novel transcript, sense to MGST3 novel transcript, AS to CETN3 novel transcript, AS to EIF5 novel transcript, AS to PTPN23 novel transcript, AS to CROCC novel transcript, AS to SLC30A6 novel transcript, AS to YES1

1q22

155609776

155610380

-1

MSTO1

r = 0.38, P<6.08E-16

1q22

155562042

155563944

ASH1L; AL353807.4

r= 0.34, P< 6.08E-16 r= 0.63, P< 6.08E-16

2p24

19902025

19902569

TTC32

r = 0.40, P<6.08E-16

1q25

184385753

184386704

-1

C1orf21

r= 0.69, P<6.08E-16

2p16

55617909

55618373

PNPT1

r=0.15, P=2.70E-05

1q24

165598356

165624084

MGST3

r = 0.28, P=8.93E-15

5q14

90410000

90410669

CETN3

r= 0.13, P=6.39E-04

14q32

103331674

103332367

-1

EIF5

r= 0.30, P<6.08E-16

3p21

47379089

47380999

-1

PTPN23

r= 0.11, P=4.42E-03

1p36

16904339

16904776

-1

2p22

32165046

32165757

-1

18p11

813274

813756

RNU1-2 CROCC SLC30A6 SPAST YES1

r=0.15, P=3.00E-05 r=0.18, P=1.36E-06 r= 0.07, P=8.94E-02 r=0.15, P=2.71E-05 r=0.16, P=1.93E-05

*The significance in multivariate analysis is indicated for overall survival (underlined Gene stable ID), overall survival and progression-free survival (italicized Gene stable ID), or progression-free survival (Gene stable ID in plain text). The four lncRNA in bold are those validated by quantitative reverse transcriptase polymerase chain reaction. **Neighbor genes belonging to the list of differentially expressed genes (Online Supplementary Table S6) are marked in bold. ID: identity; Chr: chromosome; AS: antisense.

mainly concerning RNA and protein metabolism, cellcycle regulation, nucleosome organization, immune response, cell proliferation and apoptosis, cell adhesion and tissue development (Online Supplementary Figure S6). Furthermore, GSEA revealed a significant enrichment of transcriptional signatures some of which known to be distinctively associated with main IGH translocations (Online Supplementary Table S8), likely in agreement with the cooccurrence of DIS3 mutations and t(4;14) or MAF translocations (Online Supplementary Table S4), or in an opposite manner, with the hyperdiploid condition (Online Supplementary Tables S4 and S8). Therefore, with the aim of identifying DE transcripts more specifically related to DIS3 mutations, we investigated the transcriptional profiles in MM subgroups with a more homogeneous genetic background, each of them stratified according to the occurrence of DIS3 mutations: i.e., patients carrying t(4;14) and 1q gain/amp; MAF translocations and 1q gain/amp; or hyperdiploid cases (Online Supplementary Figure S7A). Whereas no significant DE transcripts were identified in cases with both MAF translocations and 1q gain/amp (likely due to the limited number of cases), almost 90% of the DE transcripts from the other two comparisons overlapped with the original transcriptional signature (Online Supplementary Figure S7A). Overall, we identified 1,542 shared DE transcripts that could be likely considered as distinctively associated with DIS3 muta928

tions; notably, these transcripts corresponded to the most significantly deregulated ones in the global DE list (Online Supplementary Table S6, Online Supplementary Figure S7B). Focusing on the shared 490 coding transcripts (Online Supplementary Table S6, Online Supplementary Figure S7C), GSEA identified a number of gene sets among those previously obtained from the global analysis; in particular, we confirmed downregulation of the oxidative phosphorylation gene set along with genes involved in RNA and amino acid metabolism and translation, and upregulation of the interferon signaling gene set (Online Supplementary Figure S8, Online Supplementary Table S8). Finally, the shared DE transcript list associated with DIS3 mutations was particularly enriched in lncRNA (782/1542: 51% of all DE transcripts) (Online Supplementary Table S6, Online Supplementary Figure S7D), thus further supporting the notion of a distinctive and stronger impact of DIS3 mutations on the ncRNA transcriptome.

Differential expression patterns of long non-coding RNA associated with DIS3 mutations LncRNA transcriptional patterns specifically associated with DIS3 mutations were further investigated. In detail, based on the pronounced heterogeneity of lncRNA and their lower expression levels as compared to coding transcripts, we applied a more stringent analysis on the 782 haematologica | 2022; 107(4)

Transcriptomic impact of DIS3 mutations in MM

shared DE lncRNA, and focused on the 50 most significant ones (false discovery rate <1%). All lncRNA were upregulated in DIS3-mutated cases compared to DIS3 wild-type cases and were mainly represented by lncRNA antisense to coding genes (80%); the remaining cases included three

lncRNA sense to coding-genes, one long intergenic nonprotein coding RNA, two microRNA host genes and two divergent transcripts (Online Supplementary Table S9). Of note, the novel transcript Z93930.2 is located less than 100 bp antisense to the transcription factor XBP1, a well-estab-

Figure 5. Long non-coding RNA with clinical relevance in patients with multiple myeloma. Scheme of the genomic region of the four long non-coding (lnc)RNA validated by means of quantitative reverse transcriptase polymerase chain reaction in 43 patients with newly diagnosed multiple myeloma including 13 with DIS3 wildtype (WT), 14 with DIS3 WT and del(13q), seven with DIS3 mutation (DIS3mt), and nine with DIS3mt and del(13q). Primer positions are indicated in red below each lncRNA. Differential expression was assessed by the Wilcoxon signed-rank test and statistically significant P-values (<0.05) are reported above each boxplot. Dunn test. P-values for pairwise comparisons are reported in tables under each boxplot, with statistically significant P-values in bold red.

haematologica | 2022; 107(4)

929

K. Todoerti et al.

lished regulator of MM, known to be altered during the initiation and progression of MM.27 Based on the recurrent evidence that the transcription of mRNA and lncRNA appears to be closely regulated, leading to a cis-regulatory relationship,28-30 we investigated the levels of expression of overlapping or nearby transcripts localized in close proximity to the 50 lncRNA (in a window up to 65 kb). We considered 81 mRNA-lncRNA pairs and analyzed the correlation between their expression levels across the entire dataset of 767 MM cases profiled by RNA-sequencing in the CoMMpass cohort. A significant Pearson correlation (r>0.5, P<6.08x10-16) was observed for nine lncRNA-gene pairs; among them, AL121672.3 and MIRLET7BHG, both mapping at 22q13, showed a relevant correlation with each other (r=0.65) and with the PRR34 gene (r=0.66 and r=0.73, respectively) (Online Supplementary Table S10).

Clinical relevance of long non-coding RNA Next, all 50 lncRNA were tested for their relationship to OS and PFS using Kaplan-Meier survival analysis on the 767 MM cases with available RNA-sequencing and survival data. Groups with high versus low expression were determined according to the mean cut-off value for each lncRNA expression level across the entire dataset. Interestingly, higher levels of expression were associated with a poorer clinical outcome in terms of PFS for 35 out of all the 50 tested lncRNA. Furthermore, 15 of them showed an unfavorable prognosis in terms of OS (Online Supplementary Figure S9, Online Supplementary Table S11). The clinical impact of the 35 lncRNA with poorer clinical outcome was further investigated by Cox regression univariate analysis. For 21 lncRNA, their higher expression level was associated with a significantly higher risk in PFS, and for five of them (AC015982.2, AL353807.2, AC013400.1, ASH1L-AS1, and AL445228.3) also in OS (Online Supplementary Table S12). Next, for all these 21 significant lncRNA, high and low lncRNA expression levels were tested in 630 cases, together with other clinically relevant characteristics, i.e. ISS stage I and III, the occurrence of 1q gain/amp in association with TP53 alterations, the presence of DIS3 mutations or del(13q) as single events for OS, or in a bi-allelic condition for PFS. Notably, all the five lncRNA associated with a shorter OS retained their clinical impact when tested in Cox regression multivariate analysis. Two of them (AC015982.2 and AL445228.3) retained an independent significant prognostic power also for PFS; besides these two lncRNA, another seven (AC099778.1, AP000894.4, AL121658.1, AL356441.1, BX284668.6, AC093510.1, AL138976.2) were found to be independent predictors of PFS at multivariate analysis (Table 2, Online Supplementary Table S13). Of note, the biallelic condition lost its independent clinical impact in all the multivariate analyses of the nine lncRNA significant for PFS (Online Supplementary Table S13). Overall, from our analysis 12 lncRNA were predicted to have substantial clinical relevance (Table 2). Specifically, 11 of them code for novel transcripts, and four are antisense to known transcripts whose expression level, when detectable, was highly positively correlated (Table 2, Online Supplementary Table S10). Interestingly, we found four couples of lncRNA-coding genes (antisense or nearby) located on chromosome 1q with highly correlated expression (Table 2); these coding genes (MGST3, ASH1L, MSTO1, and C1orf21) were also significantly upregulated 930

in DIS3-mutated samples as compared to unmutated ones (Online Supplementary Table S6). Finally, the expression of some (AC093510.1, AC015982.2, AL138976.2, and AC099778.1) of these lncRNA was validated by quantitative reverse transcriptase polymerase chain reaction in 43 newly diagnosed MM proprietary samples previously characterized for the presence of DIS3 mutations and for which material was available (Online Supplementary Table S1). In detail, we found that AC093510.1, AC015982.2, and AL138976.2 were significantly upregulated in MM with DIS3 mutations without del(13q) as compared to in DIS3 wild-type samples, whereas AC099778.1 was significantly upregulated only in the bi-allelic condition (Figure 5).

Discussion We took advantage of the large publicly available CoMMpass dataset to investigate the type and frequency of DIS3 mutations in MM and their impact on the transcriptional signature and clinical outcome. In agreement with previously reported data,6,16,17 we assessed that the frequency of DIS3 mutations in newly diagnosed MM is approximately 10%; notably, DIS3 mutations were associated with del(13q) as bi-allelic events in 72% of cases. The majority of DIS3 mutations are missense. Together with their clustering at particular codons, the lack of truncating mutations is not typical of a tumorsuppressor gene and may suggest an oncogenic potential for DIS3. DIS3 mutations showed a clear pattern of cooccurrence with other molecular alterations, mainly chromosomal translocations. This phenomenon could be explained through the interaction between the RNA exosome and the protein activation-induced cytidine deaminase (AID) during the process of class switching and hypermutation in B cells;31 indeed, DIS3 mutations could indirectly, through disruption of an interaction with AID, cause mis-targeting of the somatic hypermutation process leading to chromosomal translocations. In particular, DIS3 mutations are associated with t(4;14), this combination defining a poor prognostic MM subgroup.16 However, our analyses pointed out that t(4;14) occurred statistically significantly in patients with bi-allelic lesions, in agreement with the very frequent association between t(4;14) and del(13q) alterations. Notably, the oncogenic events that most frequently co-occurred with DIS3 mutations were del(13q), 1q gains, t(4;14) and MAF translocations. Overall, this spectrum of molecular lesions suggests a functional constraint of cooperating oncogenic events in MM, with a selection of later lesions being restricted by the ones appearing first in the transformed cell. With regard to this, DIS3 mutations were found to be both clonal in some patients and subclonal in others, meaning they might function sometimes as early and sometimes as late hits.7,22 Our study also pointed out that the clinical relevance of DIS3 mutations depended strictly on the co-occurrence of del(13q). In detail, we established that the bi-allelic lesions significantly affected PFS, whereas the mono-allelic condition predicted worse OS. Notably, these alterations remained valuable independent predictors even when tested in combination with the clinical and poor prognosis molecular variables used to foresee clinical outcome. The differential impact of the bi-allelic or mono-allelic lesions on MM outcome could be related to the fact that our data highlighted two patterns of DIS3 lesions: one in which haematologica | 2022; 107(4)

Transcriptomic impact of DIS3 mutations in MM

del(13q) co-exists with non-hotspot DIS3 mutations, and a second in which hotspot DIS3 mutations rarely show loss of heterozygosity through bi-allelic events. Again, this is a rather intriguing pattern that may suggest either haploinsufficiency for some mutations and not others, or a different function with some mutations showing loss and others showing gain of function. This might also explain the different clinical consequences on PFS and OS observed with the two genetic statuses. Indeed, previous reports showed how different types of DIS3 mutations could lead to diverse biological effects either by impairing exosome function through reduced/modified DIS3 activity,11 or through a dominant-negative effect exerted by mutated DIS3 on the other catalytic subunit, Rrp6, acting in the exosome complex.32 DIS3 is a key component of the multisubunit RNA exosome complex in eukaryotic cells involved in the processing, quality control and degradation of virtually all classes of RNA. Our study further supports and extends the notion that DIS3 mutations affect the transcriptome, showing a stronger impact on noncoding RNA species, mainly lncRNA. Indeed, we found that approximately half of the DE transcripts predicted to be specifically related to the presence of DIS3 mutations are represented by novel, largely uncharacterized lncRNA. Among them, we highlighted 12 lncRNA, five of which are independent predictors of poorer OS and nine of worse PFS, with two of them (AC015982.2 and AL445228.3) predicting both. Moreover, the clinical impact of the nine lncRNA predicting inferior PFS at higher expression levels was independent of the genetic status of DIS3. The effect of these 12 lncRNA on the pathobiology of MM disease remains to be fully elucidated; indeed, they are all novel transcripts, none of them currently reported as being associated with cancer. Interestingly, some of them showed a strict correlation in terms of expression levels with the corresponding nearby genes, thus suggesting a cis-regulatory relationship between the paired transcripts. Although these findings need to be investigated further, this could be the case for genes involved in fundamental molecular pathways, such as MGST3, involved in the production of leukotrienes and

References 1. Munshi NC, Anderson KC. New strategies in the treatment of multiple myeloma. Clin Cancer Res. 2013;19(13):3337-3344. 2. Barbieri M, Manzoni M, Fabris S, et al. Compendium of FAM46C gene mutations in plasma cell dyscrasias. Br J Haematol. 2016;174(4):642-645. 3. Bolli N, Avet-Loiseau H, Wedge DC, et al. Heterogeneity of genomic evolution and mutational profiles in multiple myeloma. Nat Commun. 2014;5:2997. 4. Chapman MA, Lawrence MS, Keats JJ, et al. Initial genome sequencing and analysis of multiple myeloma. Nature. 2011; 471(7339):467-472. 5. Lionetti M, Barbieri M, Manzoni M, et al. Molecular spectrum of TP53 mutations in plasma cell dyscrasias by next generation sequencing: an Italian cohort study and overview of the literature. Oncotarget. 2016;7(16):21353-21361. 6. Lionetti M, Barbieri M, Todoerti K, et al. A compendium of DIS3 mutations and associated transcriptional signatures in plasma

haematologica | 2022; 107(4)

prostaglandin E, which are important mediators of inflammation;33 ASH1L, a methyltransferase already known to be involved in cancer;34 and MSTO1, important for mitochondrial fusion and intracellular distribution.35 Along with these findings, the patterns observed in the context of the DE coding transcripts associated with DIS3 mutations are also interesting. Indeed, we found downregulation of gene sets related to oxidative phosphorylation, metabolism of RNA or amino acids, and translation, and contrariwise, upregulation of interferon signaling. These transcriptomic changes are in line with the functional role of DIS3 in RNA metabolisms36,37 and in agreement with what has been described for yeasts in which mutations have been extensively investigated.38 In particular, the predicted enhanced interferon activity may represent a response to the accumulation of RNA substrates in cells following a deficiency of DIS3.39 Overall, our comprehensive evaluation of the clinical and transcriptional consequences of DIS3 mutations/deficiency in MM strongly indicates that they may play an important role in the mechanisms of MM transformation and progression. Our results may provide important insights for functional studies in order to better understand such mechanisms in MM. Disclosures No conflicts of interest to disclose. Contributions KT designed the study, collected and analyzed data and wrote the paper; DR designed the study, analyzed data and wrote the paper; VF, FM and FM analyzed data; NB wrote the paper; ET and AN designed the study, supervised the study and wrote the manuscript Funding This work was financially supported by grants from the Associazione Italiana Ricerca sul Cancro (AIRC) to AN (IG16722 and IG24365). NB is funded by the European Research Council under the European Union’s Horizon 2020 research and innovation program (grant agreement n. 817997).

cell dyscrasias. Oncotarget. 2015;6(28): 26129-26141. 7. Lohr JG, Stojanov P, Carter SL, et al. Widespread genetic heterogeneity in multiple myeloma: implications for targeted therapy. Cancer Cell. 2014;25(1):91-101. 8. 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. 9. D'Agostino M, Zaccaria GM, Ziccheddu B, et al. Early relapse risk in patients with newly diagnosed multiple myeloma characterized by next-generation sequencing. Clin Cancer Res. 2020;26(18):4832-4841. 10. Walker BA, Mavrommatis K, Wardell CP, et al. A high-risk, double-hit, group of newly diagnosed myeloma identified by genomic analysis. Leukemia. 2019;33(1):159-170. 11. Tomecki R, Drazkowska K, Kucinski I, et al. Multiple myeloma-associated 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.

12. 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. 13. Lorentzen E, Basquin J, Tomecki R, Dziembowski A, Conti E. Structure of the active subunit of the yeast exosome core, Rrp44: diverse modes of substrate recruitment in the RNase II nuclease family. Mol Cell. 2008; 29(6):717-728. 14. 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. 15. Tomecki R, Drazkowska K, Dziembowski A. Mechanisms of RNA degradation by the eukaryotic exosome. Chembiochem. 2010; 11(7):938-945. 16. 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.

931

K. Todoerti et al. 17. 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. 18. 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. 19. 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. 20. 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. 21. de Groen FL, 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. 22. 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. 23. Segalla S, Pivetti S, Todoerti K, et al. The ribonuclease DIS3 promotes let-7 miRNA maturation by degrading the pluripotency factor LIN28B mRNA. Nucleic Acids Res. 2015;43(10):5182-5193.

932

24. Walker BA, Mavrommatis K, Wardell CP, et al. Identification of novel mutational drivers reveals oncogene dependencies in multiple myeloma. Blood. 2018;132(6):587-597. 25. Schadt EE, Li C, Ellis B, Wong WH. Feature extraction and normalization algorithms for high-density oligonucleotide gene expression array data. J Cell Biochem Suppl. 2001;Suppl 37:120-125. 26. Subramanian A, Tamayo P, Mootha VK, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005;102(43):1554515550. 27. Chen L, Li Q, She T, et al. IRE1alpha-XBP1 signaling pathway, a potential therapeutic target in multiple myeloma. Leuk Res. 2016;49:7-12. 28. Sigova AA, Mullen AC, Molinie B, et al. Divergent transcription of long noncoding RNA/mRNA gene pairs in embryonic stem cells. Proc Natl Acad Sci U S A. 2013; 110(8):2876-2881. 29. Tan JY, Smith AAT, Ferreira da Silva M, et al. Cis-acting complex-trait-associated lincRNA expression correlates with modulation of chromosomal architecture. Cell Rep. 2017;18(9):2280-2288. 30. Trinklein ND, Aldred SF, Hartman SJ, Schroeder DI, Otillar RP, Myers RM. An abundance of bidirectional promoters in the human genome. Genome Res. 2004; 14(1):62-66. 31. Pefanis E, Basu U. RNA exosome regulates AID DNA mutator activity in the B cell

genome. Adv Immunol. 2015;127:257-308. 32. Chlebowski A, Lubas M, Jensen TH, Dziembowski A. RNA decay machines: the exosome. Biochim Biophys Acta. 2013;1829(6-7):552-560. 33. Higgins LG, Hayes JD. Mechanisms of induction of cytosolic and microsomal glutathione transferase (GST) genes by xenobiotics and pro-inflammatory agents. Drug Metab Rev. 2011;43(2):92-137. 34. Xu B, Qin T, Yu J, Giordano TJ, Sartor MA, Koenig RJ. Novel role of ASH1L histone methyltransferase in anaplastic thyroid carcinoma. J Biol Chem. 2020;295(26):88348845. 35. Chapman J, Ng YS, Nicholls TJ. The maintenance of mitochondrial DNA integrity and dynamics by mitochondrial membranes. Life. 2020;10(9):164. 36. Davidson L, Francis L, Cordiner RA, et al. Rapid depletion of DIS3, EXOSC10, or XRN2 reveals the immediate impact of exoribonucleolysis on nuclear RNA metabolism and transcriptional control. Cell Rep. 2019;26(10):2779-2791.e5. 37. Kiss DL, Andrulis ED. Genome-wide analysis reveals distinct substrate specificities of Rrp6, Dis3, and core exosome subunits. RNA. 2010;16(4):781-791. 38. Milbury KL, Paul B, Lari A, et al. Exonuclease domain mutants of yeast DIS3 display genome instability. Nucleus. 2019;10(1):21-32. 39. Rigby RE, Rehwinkel J. RNA degradation in antiviral immunity and autoimmunity. Trends Immunol. 2015;36(3):179-188.

haematologica | 2022; 107(4)

ARTICLE

Platelet Biology & its Disorders

The GPIba intracellular tail - role in transducing VWF- and collagen/GPVI-mediated signaling

Ferrata Storti Foundation

Adela Constantinescu-Bercu,1 Yuxiao A Wang,1 Kevin J Woollard,2 Pierre Mangin,3 Karen Vanhoorelbeke,4 James TB Crawley1 and Isabelle I Salles-Crawley1 Center for Hematology, Department of Immunology and Inflammation, Imperial College London, London, UK; 2Center for Inflammatory Disease, Department of Immunology and Inflammation, Imperial College London, London, UK; 3Université de Strasbourg, INSERM, EFS Grand-Est, BPPS UMR-S 1255, FMTS, Strasbourg, France and 4Laboratory for Thrombosis Research, KU Leuven, Kortrijk, Belgium. 1

Haematologica 2022 Volume 107(4):933-946

ABSTRACT

he GPIba-VWF A1 domain interaction is essential for platelet tethering under high shear. Synergy between GPIba and GPVI signaling machineries has been suggested previously, however its molecular mechanism remains unclear. We generated a novel GPIba transgenic mouse (GpIbaΔsig/Δsig) by CRISPR-Cas9 technology to delete the last 24 residues of the GPIba intracellular tail that harbors the 14-3-3 and phosphoinositide-3 kinase binding sites. GpIbaΔsig/Δsig platelets bound von Willebrand factor (VWF) normally under flow. However, they formed fewer filopodia on VWF/botrocetin in the presence of a aIIbb3 blocker, demonstrating that despite normal ligand binding, VWF-dependent signaling is diminished. Activation of GpIbaΔsig/Δsig platelets with ADP and thrombin was normal, but GpIbaΔsig/Δsig platelets stimulated with collagenrelated-peptide (CRP) exhibited markedly decreased P-selectin exposure and aIIbb3 activation, suggesting a role for the GpIba intracellular tail in GPVI-mediated signaling. Consistent with this, while hemostasis was normal in GpIbaΔsig/Δsig mice, diminished tyrosine-phosphorylation, (particularly pSYK) was detected in CRP-stimulated GpIbaΔsig/Δsig platelets as well as reduced platelet spreading on CRP. Platelet responses to rhodocytin were also affected in GpIbaΔsig/Δsig platelets but to a lesser extent than those with CRP. GpIbaΔsig/Δsig platelets formed smaller aggregates than wild-type platelets on collagen-coated microchannels at low, medium and high shear. In response to both VWF and collagen binding, flow assays performed with plasma-free blood or in the presence of aIIbb3- or GPVI-blockers suggested reduced aIIbb3 activation contributes to the phenotype of the GpIbaΔsig/Δsig platelets. Together, these results reveal a new role for the intracellular tail of GPIba in transducing both VWF-GPIba and collagen-GPVI signaling events in platelets.

Introduction In order to fulfil their hemostatic function, platelets are recruited to sites of vessel damage by von Willebrand factor (VWF), which interacts with exposed collagen and, thereafter, to glycoprotein (GP) Iba on the platelet via its A1 domain. VWF-mediated platelet tethering facilitates platelet capture.1 Subsequent interaction of platelets with additional ligands (e.g., aIIbb3-fibrinogen, collagen-GPVI, collagen-a2b1) and changes in platelet phenotype are required to stabilize the platelet plug. Although the VWF-GPIba interaction primarily facilitates platelet recruitment, it also transduces a signal that causes intraplatelet Ca2+ release and activation of the platelet integrin, aIIbb3.2-5 These signaling events are highly dependent upon flow as shear forces induce unfolding of the GPIba mechanosensitive juxtamembrane region that translates the mechanical signal into intracellular biochemical events.6,7 Signaling is dependent upon the binding of adaptor and signaling molecules (e.g., Src kinases, Lyn and c-Src, 14-3-3 isoforms and phosphoinositide-3 kinase [PI3K]) that can associate with the GPIba intracellular tail.8-12 Downstream activation of phospholipase Cg2

haematologica | 2022; 107(4)

Correspondence: ISABELLE I SALLES-CRAWLEY i.salles@imperial.ac.uk Received: December 23, 2020. Accepted: June 10, 2021. Pre-published: June 17, 2021. https://doi.org/10.3324/haematol.2020.278242

933

A. Constantinescu-Bercu et al.

(PLCg2), PI3K-Akt, cGMP-PKG, mitogen activated kinase and LIM kinase 1 pathways have also been reported.13-19 By comparison to other platelet agonists (e.g., collagen, thrombin, ADP, thromboxane A2), signaling through GPIba is considered weak. VWF-GPIba signaling, which we term platelet ‘priming’ rather than activation, does not induce appreciable degranulation.5 Therefore, the contribution of platelet ‘priming’ to normal hemostasis remains unclear as the effects of the other platelet agonists have the potential to mask those of GPIba. However, in scenarios where other platelet agonists are either absent or in low abundance (e.g., platelet recruitment to endothelial or bacterial surfaces), the effects/importance of GPIba signaling may become more prominent.5 GPVI is a collagen/fibrin receptor on the platelet surface that non-covalently associates with Fc receptor g-chain (FcRg) and signals via immunoreceptor tyrosine-based activation motifs (ITAM).20-22 Collagen binding to platelets induces clustering of GPVI, which results in the phosphorylation of FcRg by Src family kinases, Lyn and Fyn, that associate with the intracellular domain of GPVI.23,24 This causes the recruitment and phosphorylation of Syk tyrosine kinase, and formation of a LAT-based signaling complex that can activate PLCg2 and lead to release of intraplatelet Ca2+ stores, activation of protein kinase (PK) C, and ultimately aIIbb3 activation and both a- and dense-granule release.21 Previous studies have suggested functional associations between GPIba and GPVI and/or its co-receptor FcRg.13,25,26 For example, VWF-GPIba-mediated platelet responses are reportedly impaired in GPVI/FcRg deficiencies in both mice and humans.13,27 There is also evidence that VWF can potentiate responses after collagen mediated responses in human platelets.28 However, the molecular basis of GPIba and GPVI receptor crosstalk has not been elucidated. Using a novel GPIba transgenic mouse in which the last 24 amino acids (a.a.) of the GPIba intracellular tail were deleted, we demonstrate the importance of this region not only to VWF-dependent signaling in platelets, but also reveal a major contribution in augmenting GPVI-mediated platelet signaling.

Methods Mice All procedures were performed with the United Kingdom Home Office approval in accordance with the Animals (Scientific Procedures) Act of 1986. GpIbaΔsig/Δsig mice were generated in-house by the Medical Research Council transgenic group at Imperial College using CRISPR-Cas9 technology (Figure 1). Briefly, pronuclear injections (CBAB6F1) were performed with Cas9 mRNA (75 ng/mL), guide RNA (gRNA; 25-50 ng/mL) and single-strand oligo donor DNA (25-50 ng/mL). The donor DNA (GGTAAGGCCTAATGGGCGAGTGGGGCCTCTGGTAGCAGGACGGCGACCCTGAGCTCTGAGTCAGGGTCGTGGTCAGGACCTATTGGGCACAGTGGGCATTA) had 50 bp homology arms at the 5’ and 3’ ends (Integrated DNA Technologies). Embryos were transferred to pseudo-pregnant CBAB6F1 female mice. Two founder mice originated from the same gRNA (CGACCCTGACTCAGAGCTGAGGG) were bred with C57BL/6 mice. F1 GpIbaΔsig/+ mice were bred to obtain GpIbaΔsig/Δsig mice, and GpIba+/+ littermates were used as controls. Genotyping was performed by polymerase chain reaction (PCR) amplification of a GpIba fragment (551 bp) using primers: AAGCACTCACACCACAAGCC

934

and AGTATGAATGAGCGGGAGCC and subsequent Sanger sequencing (Genewiz). Experimental procedures were performed as previously described.29,30 Additional details are included in the Online Supplementary Appendix.

Results Generation of GpIbaΔsig/Δsig mice Sequence identity between human and murine GPIba intracellular region is very high, supporting the contention that their functions are well conserved (Figure 1A). In order to evaluate the role of the GPIba intracellular tail upon both VWF- and collagen/GPVI-mediated signaling, we generated a novel transgenic mouse (GpIbaΔsig/Δsig) using CRISPR-Cas9 technology. We introduced a point mutation (Ser695Stop) that resulted in a premature stop codon that deletes the last 24 a.a. of the GPIba intracellular tail (a.a. 695-718) containing the entire 14-3-3 isoform and PI3K binding region,10,12 but maintains the upstream filamin binding site in GPIba (residues 668-681 in murine GPIba)31 (Figure 1A and B). Introduction of the mutation was confirmed by sequencing and by western blotting using an anti-GPIba antibody that recognizes the terminal region of the intracellular tail (Figure 1C to E). GpIbaΔsig/Δsig mice were viable and born with the expected Mendelian frequencies.

GpIbaΔsig/Δsig mice platelet count, platelet size and hemostatic function GpIbaΔsig/Δsig mice had mildly reduced (~20%) platelet counts and slightly larger platelet size (Figure 2A and B), but other hematological parameters were unaffected (Online Supplementary Table S1). This is in contrast to the severe thrombocytopenia and giant platelets observed in complete GPIba deficiency in mice or Bernard-Soulier patients.32,33 Expression of the major platelet receptors, GPVI, aIIbb3, GPIba, and the extracellular region of GPIba was unaltered on GpIbaΔsig/Δsig platelet surfaces (Figure 2C). In order to assess hemostatic function in GpIbaΔsig/Δsig mice, we performed tail bleeding assays. Unlike Vwf-/- mice or mice lacking the extracellular domains of GPIba,32,34,35 GpIbaΔsig/Δsig mice displayed normal blood loss following tail transection (Figure 2D), suggesting that GpIbaΔsig/Δsig platelets can be recruited to sites of vessel damage similar to wildtype mice. There was no difference between GpIbaΔsig/Δsig mice and wild-type littermates in a non-ablative laser-induced thrombus formation, as measured by the kinetics and extent, of both platelet accumulation and fibrin deposition (Figure 2E to G; Online Supplementary Figure S1; Online Supplementary Video S1).29,30,36 These results support the contention that deletion of the GPIba does not appreciably influence either platelet recruitment or their ability to support thrombin generation. In this model, platelet accumulation requires both VWF and thrombin but has less dependency upon collagen exposure or GPVI signaling due to the non-ablative injury.37,38

GpIbaΔsig/Δsig platelets bind von Willebrand factor (VWF) normally, but exhibit decreased VWF-mediated signaling In order to specifically examine the effect of the GPIba intracellular tail truncation upon VWF-dependent platelet capture, we coated microchannels with murine VWF over haematologica | 2022; 107(4)

The role of GPIba in GPVI-signaling

Figure 1. Generation and characterization of GpIbaΔsig/Δsig mice. (A) Sequence alignment of the last 100 amino acids (a.a.) of human and mouse GPIba. Sequence identities are highlighted in red. Filamin binding region: (a.a. 560-573) and (a.a. 668-681) for human and mouse GPIba; PI3K/14-3-3 binding region: (a.a. 580-610) and (a.a. 688-718) for human and mouse GPIba. (B) Schematic representation of the GpIba gene with CRISPR guide target site, gRNA sequence, BbvCI restriction enzyme site and Cas9 predicted cut site. Primers used to amplify the GpIba allele from genomic DNA are indicated in purple. Design of the 101 bp single stranded DNA repair template with the point mutation to introduce a codon stop eliminating the BbvCI restriction enzyme site and removing the last 24 a.a. of GPIbα is also shown. The resulting truncated a.a. sequence from GpIbaΔsig/Δsig mice is indicated in green. (C) Genomic DNA sequences from GpIba+/+ and GpIbaΔsig/Δsig mice. Successful substitution is indicated with an arrow. (D) Diagram showing the binding of the anti-GPIbα tail Ab (Biorbyt; orb 215471). (E) Platelet lysates from GpIba+/+ and GpIbaΔsig/Δsig mice were probed with the anti-GPIba tail and b-actin antibodies. Absence of band in the GPIba western-blot confirms the successful truncation of the GPIba intracellular tail in GpIbaΔsig/Δsig mice.

haematologica | 2022; 107(4)

935

A. Constantinescu-Bercu et al.

which we perfused plasma-free blood (to remove fibrinogen and outside-in activation aIIbb3) at 1,000s-1. GpIbaΔsig/Δsig platelets were recruited normally to murine VWF-coated surfaces with rolling velocities, surface coverage and platelet accumulation unaltered compared to GPIba+/+ platelets (Figure 3A to D; Online Supplementary Video S2). In order to investigate the impact of the deletion of the last 24 a.a. of GPIba on VWF signaling, we performed platelet spreading assays on murine VWF, which rely upon VWF-GPIba signaling. On VWF alone, very few GPIba+/+ or

GpIbaΔsig/Δsig platelets bound to VWF and only very few exhibited filopodia (Figure 3E and F). When these experiments were repeated in the presence of botrocetin (a snake venom that increases the affinity of VWF A1 domain for GPIba)39 a large proportion (90±2.8%) of GPIba+/+ platelets underwent shape changes and developed filopodia (Figure 3E and G; Online Supplementary Figure S2A and B), a welldescribed consequence of VWF-GPIba signaling.9,19 This process was significantly diminished in GpIbaΔsig/Δsig platelets with only 46±2.6% platelets exhibiting filopodia (Figure 3E

Figure 2. GpIbaΔsig/Δsig mice display normal bleeding loss and platelet and fibrin accumulation in the laser-induced thrombosis model. A) Platelet counts and (B) platelet size in GpIba+/+ (n=25) and GpIbaΔsig/Δsig mice (n=30) as determined by flow cytometry. (C) Surface expression of platelet receptors GPIba, GPIbb, aIIbb3 and GPVI in GpIba+/+ and GpIbaΔsig/Δsig mice (n=4 for each genotype) determined by flow cytometry and expressed as % of control. (D) Bar graph analyzing blood loss after 10 minutes following tail transection in GpIba+/+ and GpIbaΔsig/Δsig mice (n=9 for each genotype). (E-G) Mice cremaster muscle arterioles were subjected to the laserinduced thrombosis model as described in the Online Supplementary Appendix. Curves represent median integrated fluorescence intensity (IFI) from platelets (arbitrary units: AU) (E) or fibrin(ogen) (F) as a function of time after the injury (20 thrombi in 3 GpIba+/+ and 34 thrombi in 4 GpIbaΔsig/Δsig mice). (G) Representative composite fluorescence images of platelets (green) and fibrin (red) with bright field images after laser-induced injury of the endothelium of GpIba+/+ (top panels) vs. GpIbaΔsig/Δsig mice (bottom panels). Scale bar represents 10 mm. Each symbol represents one thrombus. Horizontal lines intersecting the data set represent the median. Data was analyzed using Mann Whitney test; ns: P>0.05. Also see the Online Supplementary Video S1 and the Online Supplementary Figure S1. FSC: foward scatter; Hb: hemoglobin.

936

haematologica | 2022; 107(4)

The role of GPIba in GPVI-signaling

and G; Online Supplementary Figure S2C to D).9,19 When experiments were performed in the presence of both botrocetin and GR144053, which competitively inhibits the interaction of aIIbb3 with VWF and/or fibrinogen, the number of GPIba+/+ platelets forming filopodia was not appreciably influenced (Online Supplementary Figure S2B), but the proportion of that formed >3 filopodia was significantly reduced (37±6.7% vs. 74±6.9%) (Online Supplementary Figure S2A), revealing the contribution of outside-in signaling to filopodia formation. Under these conditions, here again although GpIbaΔsig/Δsig platelets bound VWF surfaces, they had a significantly diminished ability to form filopodia (Figure 3E and H). Moreover, GR144053 had no effect upon filopodia formation in GpIbaΔsig/Δsig platelets (Online Supplementary Figure S2C), suggesting that the reduced filopodia formation in these platelets was likely

due to a defect in VWF-GPIba signaling manifest by a lack of activation of aIIbb3 in response to VWF-GPIba binding. Taken together, these results indicate that deletion of the last 24 a.a. of the intracellular tail of GPIba does not influence platelet binding to VWF, but significantly reduces VWF-GPIba downstream signaling response including aIIbb3 activation.

The intracellular tail of GPIba is important for GPVI signaling We next evaluated agonist-induced platelet activation in GpIbaΔsig/Δsig mice. In response to ADP, washed GpIbaΔsig/Δsig platelets exhibited normal aIIbb3 activation and P-selectin exposure and normal platelet aggregation (Figure 4A to D). Responses to thrombin were also normal except for a slight significant decrease in P-selectin exposure with the lowest

F E

Figure 3. GpIbaΔsig/Δsig platelets exhibit normal binding to von Willebrand factor but disrupted GPIbα-mediated signaling. (A-D) Plasma-free blood from GpIba+/+ and GpIbaΔsig/Δsig mice supplemented with anti-GPIbb-DyLight488 anitbody was perfused over murine VWF at a shear rate of 1,000 s-1. (A) Representative fluorescence images (n≥3; scale bar 10 mm) and bar graphs analyzing the integrated fluorescence intensity (IFI) (B) and the surface coverage (C) of GpIba+/+ and GpIbaΔsig/Δsig platelets captured by murine von Willebrand factor (VWF) after 3.5 minutes of flow. (D) Rolling velocities (median ± confidence interval [CI]) were calculated from (approximately 10,000) platelets rolling/adhering to murine VWF within the first 30 seconds (n≥3) (E) Representative confocal images of GpIba+/+ and GpIbaΔsig/Δsig platelets (n=3 for each genotype) spread on mVWF and stained with Phalloidin-Alexa 488, in the absence or presence of Botrocetin or Botrocetin and GR144053 (scale bar 10 mm). (F to H) Percentage of platelets from GpIba+/+ and GpIbaΔsig/Δsig mice (individual data points representing the average of 3-6 fields of view) with no filopodia, 1-3 filopodia or >3 filopodia formed on murine VWF in the absence (F; 129 GpIba+/+ platelets and 115 GpIbaΔsig/Δsig platelets analysed) or presence of Botrocetin (G; 511 GpIba+/+ platelets and 547 GpIbaΔsig/Δsig platelets analysed), or Botrocetin and GR144053 (H; 359 GpIba+/+ platelets and 480 GpIbaΔsig/Δsig platelets analysed). Data represents mean ± standard error of the mean (B, C, F to H) or median ± CI (D) and was analyzed using unpaired two-tailed Student’s t-test (B and C), unpaired Mann Whitney test (D) or using two-way ANOVA followed by Sidak’s multiple comparison test (F to H); *P<0.05, ***P<0.001, ****P<0.0001. Also see the Online Supplementary Figure S2 and the Online Supplementary Video S2.

haematologica | 2022; 107(4)

937

A. Constantinescu-Bercu et al.

Figure 4. Legend on following page.

938

haematologica | 2022; 107(4)

The role of GPIba in GPVI-signaling

Figure 4. GpIbaΔsig/Δsig platelets exhibit altered GPVI-mediated signaling. (A and B) Flow cytometric analysis of surface expression of activated aIIbb3 (A) and P-selectin (B) in GpIba+/+ and GpIbaΔsig/Δsig platelets (n=8) in response to ADP (1-20 mM), a-thrombin (20-200 mU/mL), or CRP (1-10 mg/mL). MFI: geometric mean fluorescence intensity (C) Representative aggregation traces (n=3-6) of washed platelets isolated from GpIba+/+ (blue) or GpIbaΔsig/Δsig (red) mice and stimulated with ADP (1-10 mM), a-thrombin (20-50 mU/mL) or CRP (0.5-3 mg/mL). Aggregation was monitored using a Chronolog aggregometer over 6 minutes. (D) Bar graph analysing the maximum aggregation (%) obtained in the conditions presented in (C). (E) Representative micrographs (n=3 for each genotype; 3 fields of view analyzed per condition; scale bar 10 mm) of 454 GpIba+/+ and 420 GpIbaΔsig/Δsig platelets spread on CRP and stained with Phalloidin-Alexa 488. Bar graphs quantifying the surface area (F) and percentages (G) of platelets that remained round, formed filopodia or spread on CRP. (H) Western blot analyzing tyrosine kinase phosphorylation in platelets from GpIba+/+ and GpIbaΔsig/Δsig mice, following stimulation with 3 mg/mL CRP for 0-180 seconds (s), using b-actin as a loading control (representative of n=3). (I) Western blots analyzing the levels of phosphorylated and non-phosphorylated SYK, PLCg2 and Akt in platelets from GpIba+/+ and GpIbaΔsig/Δsig mice, after 0-180 s stimulation with CRP (representative of n=3). (J-L) Bar graphs displaying the levels of phosphorylated SYK, PLCg2 and Akt in platelets from GpIba+/+ and GpIbaΔsig/Δsig mice, after 0-180 s stimulation with CRP and normalizing the intensity according to the non-phosphorylated levels of SYK, PLCg2 and Akt. For the surface area (F), the data represent the median ± confidence interval (CI) and was analyzed using the unpaired Mann Whitney test. All other data is displayed as mean ± standard error of the mean and was analyzed using two-way ANOVA followed by Sidak’s multiple comparison test. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. Also see the Online Supplementary Figures S2 and S3.

thrombin concentration (Figure 4A and B) but this did not influence thrombin-induced platelet aggregation (Figure 4C and D). How this reduced P-selectin exposure in response to low thrombin concentration is manifest remains unclear, but may reflect the findings of a previous study that suggested the importance of 14-3-3ζ binding to GPIba specifically for low-dose thrombin responses.40 Despite largely unaffected responses to ADP and thrombin, in response to collagen-related peptide (CRP), GpIbaΔsig/Δsig platelets exhibited markedly reduced aIIbb3 activation and P-selectin exposure (Figure 4A and B). Interestingly, GpIbaΔsig/Δsig platelet aggregation following CRP stimulation appeared normal (Figure 4C and D). Next, we evaluated the ability of GpIbaΔsig/Δsig platelets to spread on fibrinogen surfaces with and without prior stimulation with thrombin. Without platelet stimulation, similar to wild-type platelets, most GpIbaΔsig/Δsig platelets remained round while upon stimulation with thrombin approximately 80% platelets spread fully with no difference observed in the spread platelet area (Online Supplementary Figure S3A to E). As full spreading is highly dependent upon outside-in signaling through aIIbb3,41 this suggests that this signaling pathway is unaffected in GpIbaΔsig/Δsig platelets. We then explored the ability of platelets to spread on CRP-coated surfaces. Consistent with diminished platelet activation in response to CRP, GpIbaΔsig/Δsig platelets remained round in contrast to wild-type platelets (59±3.4% vs. 19±7%; Figure 4E and G). This effect was also quantified by a 20% reduction in bound platelet area (Figure 4F) and in the reduced incidence of filopodia formation – 16±6.3% for GpIbaΔsig/Δsig versus 52±7.1% for GPIba+/+ (Figure 4E and G). Collectively, these results reveal an appreciable defect in GPVI-mediated signaling in GpIbaΔsig/Δsig platelets. There was an overall reduction in tyrosine phosphorylation after CRP stimulation in GpIbaΔsig/Δsig platelets compared to wild-type platelets (Figure 4H). Further analysis revealed appreciably reduced Syk kinase activation in GpIbaΔsig/Δsig platelets, as measured by phosphorylation of Syk on Tyr525 and Tyr526 in response to CRP and lower phosphorylation levels of its downstream target pPLCg2 (p-Tyr 1217), although this was less marked than for those observed with pSyk (Figure 4I to K). In addition, phosphorylation levels of Akt (p-Ser 473), a known substrate of PI3K were also appreciably diminished in GpIbaΔsig/Δsig versus GpIbaΔsig/Δsig (Figure 4I and L). In order to assess whether the effect of truncation of GPIba was specific for GPVI-mediated platelet responses, or whether other tyrosine-mediated signaling pathways might also be affected, we stimulated GpIbaΔsig/Δsig and wild-type platelets with rhodocytin (C-type lectin receptor 2 [CLEC-2] agonist). Tyrosine-phosphorylation profile of GpIbaΔsig/Δsig platelets in response to rhodocytin was similar to that of GPIba+/+ platelets, with slightly haematologica | 2022; 107(4)

reduced phosphorylation of Syk (approximately 20%) (Online Supplementary Figure S4A to C). P-selectin exposure in response to rhodocytin was reduced in GpIbaΔsig/Δsig platelets while aIIbb3 activation was only diminished for the lowest concentration of the toxin without reaching statistical significance (Online Supplementary Figure S4D to E). These results suggest that the GPIba tail may also influence CLEC-2 ITAM-mediated signaling, but perhaps with reduced dependency.

The role of the GPIba intracellular tail in platelet recruitment and aggregation under flow In order to examine the consequences of the combined effects of disrupted VWF-GPIba signaling and diminished GPVI signaling in platelets in more physiological assays, we quantified platelet recruitment and aggregate formation on collagen-coated microchannels under flow. Experiments were performed at high (3,000 s-1), medium (1,000 s-1) and low (200 s-1) shear, as platelet recruitment is increasingly dependent on VWF-GPIba as shear increases while subsequent platelet aggregate formation on collagen surfaces becomes more dependent upon GPVI signaling.42-44 Perfusing whole blood at 3,000 s-1 and 1,000 s-1 over collagen, we observed a marked reduction in surface coverage of GpIbaΔsig/Δsig platelets when compared to GPIba+/+ platelets (Figure 5A and B; Figure 6A and B; Online Supplementary Videos S3 and S4). GpIbaΔsig/Δsig platelets that bound to collagen also formed smaller aggregates than GPIba+/+ platelets (Figures 5C and 6C), likely reflecting the subsequent effect of diminished collagen-GPVI signaling. Perfusing wild-type plasma-free blood (to remove soluble VWF and fibrinogen) in collagen-coated microchannels revealed a significant reduction of both platelet adhesion and thrombus growth to similar levels observed in GpIbaΔsig/Δsig samples (Figure 6A, B, D and E) showing that a small amount of VWF-independent binding to collagen occurs at 1,000 s-1. When whole blood experiments were performed in the presence of GR144053, to block aIIbb3, GPIba+/+ platelets were recruited to the collagen surface as a monolayer. However, additional platelet-platelet recruitment was abolished and therefore there was limited thrombus growth in 3D. This was measured by an increase in surface coverage with a decrease in thrombus formation (i.e., total platelet fluorescence; Figure 6A, B and D).45 Surface coverage as well as platelet accumulation of GpIbaΔsig/Δsig platelets was similar in both the absence and presence of GR144053 (Figure 6A, B and E), suggesting that lack of active aIIbb3 is part of the platelet phenotype. In order to more specifically examine the role of GPVI in this system, we performed experiments in the presence of JAQ1, an anti-murine GPVI blocking antibody. Blocking GPVI significantly reduced surface coverage and platelet accumulation in GpIbaΔsig/Δsig and GPIba+/+ 939

A. Constantinescu-Bercu et al.

platelets, revealing the important contribution of GPVI signaling at 1,000 s-1 (Figure 6F to I), in stabilizing platelet recruitment and their subsequent aggregation. At venous shear rates (200 s-1) where the dependencies on VWF and collagen are slightly different to 1,000 s-1, surface coverage of GpIbaΔsig/Δsig platelets was slightly reduced compared to GPIba+/+ platelets although it did not reach significance. However, thrombus growth was significantly diminished (Figure 7A to C; Online Supplementary Video S5). Using plasma-free blood, the surface coverage was similar for GpIbaΔsig/Δsig and GPIba+/+ platelets, mediated by direct (VWFindependent) interaction with collagen (Figure 7A and B). Similar to high-shear conditions, platelet accumulation under plasma-free conditions of GPIba+/+ platelets was significantly reduced compared to whole blood (Figure 7D) similar to those observed with GpIbaΔsig/Δsig platelets (Figure 7E). In the presence of GR144053, we saw the same increase in surface coverage of GPIba+/+ platelets with reduced localized 3D-platelet thrombi (Figure 7A and B) although the platelet accumulation was not significantly different to GPIba+/+ whole blood (Figure 7D) likely due to the increased platelet coverage. Consistent with the results obtained under high-shear conditions, the effect of increased surface coverage in the presence of GR144053 was not observed with GpIbaΔsig/Δsig platelets, nor was platelet accumulation appreciably further diminished (Figure 7A, B and E). Finally, similar to results obtained under arterial shear conditions, blocking GPVI significantly reduced surface coverage and platelet accumulation in both GpIbaΔsig/Δsig and GPIba+/+ platelets (Figure 7F to I). As removal of either VWF or blocking of GPVI had very similar effects, this suggests that VWF-GPIba and GPVI-collagen binding may act synergistically to recruit platelets at low shear.

Discussion The ability of platelet GPIba binding to VWF to transduce intraplatelet signaling is well-known, but the hemostatic role of the platelet ‘priming’ that follows has frequently been perceived as redundant due to the comparatively mild phenotypic changes in platelets that ensue when compared to other platelet agonists (e.g., thrombin, collagen). Using a novel GpIbaΔsig/Δsig mouse, we now demonstrate that the intracellular tail of GPIba is important not only for transduction of VWF-GPIba signaling, but also collagen-GPVI-mediated responses in platelets (Figure 8). The binding of GPIba to VWF, and of GPVI to collagen, are critical events for platelet plug formation.42,46,47 Previous studies reported associations between GPIba and GPVI, or its co-receptor FcRg suggesting potential interplay between these signaling pathways.25,26,28 Functional crosstalk between these signaling pathways is supported by the diminished VWF-GPIba-dependent responses in platelets deficient in GPVI and by the ability of VWF to further potentiate platelet secretion in response to CRP.13,27,28 In order to explore GPIba signaling function and its influence upon GPVI signaling, we generated GpIbaΔsig/Δsig mice by introduction of a stop codon downstream of the main filamin binding site (a.a. 668-681), but upstream of the 14-3-3 isoforms and PI3K binding regions that are important for VWF-GPIba signaling.8-12,48 (Figure 1) This resulted in uniform production of platelets that express 940

GPIba with truncated intracellular tail. This circumvented the limitations associated with studying/expressing platelet receptor complexes in heterologous cellular systems. Previously generated full knockout (GPIba-/-) and also GPIba/IL4Ra-tg mice that lack the extracellular region of GPIba do not enable analysis of VWF signaling per se, as they lack the ability to bind VWF, meaning that one cannot dissociate the effects of loss of VWF binding and/or VWF signaling upon functional effects upon the platelets.32,35 Transgenic mice (hTgY605X) that express human GPIba that lacks the terminal 6 a.a. of the intracellular tail displayed reduced megakaryocyte recovery following induced thrombocytopenia,49 but more recent in vitro studies have revealed that these mice do not lack the entire 143-3/PI3K binding region,9,10,12 suggesting that their VWF signaling function may not be fully disrupted making interpretation of the mouse phenotype difficult. GpIbaΔsig/Δsig mice had a modest reduction in platelet counts compared to GPIba+/+ littermates that is likely be attributable to the small increase in platelet size (Figure 2A and B). Interestingly, platelet size is also moderately increased in the GPIba/IL4Ra-tg mice,35 but, again, this is modest compared to the size observed in GPIba-/- or in Bernard-Soulier platelets.32,33 Although the major filamin binding site remains intact in GpIbaΔsig/Δsig mice, our findings may be consistent with CHO cell studies that suggested the presence of additional or extended filamin binding regions within the intracellular tail of GPIba.48 By themselves, the 20% reduction in platelet count and slight increase in platelet size would not impart a hemostatic defect.50 GpIbaΔsig/Δsig mice exhibited normal hemostatic responses to tail transection, and normal thrombus formation following mild laser-induced thrombosis (Figure 2D to G). We used a non-perforating endothelial cell injury that does not induce collagen exposure. Therefore, this non-ablasive model is independent of collagen-mediated signaling pathways.36,38 However, both the tail transection and laserinduced models are sensitive to VWF function.34,37 Our results reveal the normal VWF-binding function of GpIbaΔsig/Δsig platelets. Normal bleeding times were also reported in hTgY605X transgenic mice with no overt effect on platelet or coagulation functions.49 Truncation of the intracellular tail of GPIba did not alter expression of its extracellular domain (nor influence surface expression of GPIba, GPVI or aIIbb3) (Figure 2C). Consequently, GpIbaΔsig/Δsig platelet capture to mouse VWFcoated surfaces was unaffected as well their rolling velocities (Figure 3A to D). Despite normal VWF binding, deletion of the PI3K and 14-3-3 binding region in GPIba9,10,12 significantly decreased filopodia extension upon stimulation of VWF binding with botrocetin but also in the presence of an aIIbb3 antagonist that prevent outside-in signaling induced by the VWF C4 domain binding to activated aIIbb3 (Figure 3E, G to H). Normal VWF-platelet binding in GpIbaΔsig/Δsig mice is in line with previous studies showing that deletion of the 14-3-3ζ binding site in human GPIba in GPIb-IX CHO cells does not influence VWF binding, but does reduce their ability to spread.9,51 Other studies showed that a membrane-permeable inhibitor of the 14-3-3ζ-GPIba interaction (MP-aC) inhibited GPIba-dependent platelet agglutination and was protective in murine thrombosis models.11,52 However, although this peptide disrupts the interaction between 143-3ζ and GPIba, it may also influence 14-3-3ζ function haematologica | 2022; 107(4)

The role of GPIba in GPVI-signaling

independent of GPIba binding. This contention is perhaps supported by a recent study revealing that 14-3-3ζ deficient mice are protected against arterial thrombosis with normal VWF-GPIba-mediated platelet function.53 In addition to defective VWF-mediated signaling, GpIbaΔsig/Δsig platelets exhibited markedly diminished collagen-mediated signaling through GPVI evidenced by reduced surface expression of P-selectin and activation of aIIbb3, fewer filopodia upon CRP stimulation (Figure 4A, B, E to G), and severely diminished platelet aggregate formation on collagen under venous and arterial shears (Figures 5 to 7). Bernard-Soulier patient platelets have historically been reported to respond normally to collagen in aggregation assays.54 However, the thrombocytopenia and giant platelets associated with full GPIba deficiency combined with the loss of VWF-dependent platelet recruitment on collagen impair full analysis of other platelet signaling pathways under physiological flow conditions. Interestingly, although early studies on Bernard-Soulier patients reported that platelet aggregation in response to collagen was normal, their transformation into procoagulant platelets was specifically impaired in response to collagen (but not other agonists).55 More recently, a BernardSoulier patient with mutations in both GPIba and filamin A was also reported to exhibit defects in GPVI-mediated signaling responses.56 Although the authors contended that this defect might be due to the filamin A mutation, this may warrant some reappraisal in light of the data presented herein. Like Bernard-Soulier platelets, we found that GpIbaΔsig/Δsig platelets aggregated normally in response to CRP (Figure 4C to D). The signaling deficit presumably allows sufficient activation of aIIbb3 for the platelets to aggregate. This is perhaps unsurprising given that Gp6+/platelet aggregation is only affected at low collagen con-

centrations.57,58 Taken together, previous studies support the contention that Bernard-Soulier patient platelets exhibit a partial deficit in GPVI signaling that resembles the deficit in GpIbaΔsig/Δsig mouse platelets. Platelets can interact with collagen directly through GPVI and a2b1, and indirectly via GPIba binding to VWF, the latter being increasingly important as shear rates rise to first capture the platelets and enable the aforementioned direct interactions to take place.42,59 This is demonstrated in wild-type mice, similar to previous reports,43,60 by the markedly reduced binding of platelets to collagen in the absence of plasma (and therefore VWF) at medium shear rates (Figure 6A, B and D). Although we demonstrated that GpIbaΔsig/Δsig platelets bind VWF normally, we saw the largest defect in platelet coverage/accumulation when compared to wild-type mice at 3,000 s-1 (Figure 5). Based on these results, it seems likely that VWF-GPIba signaling is also important at these high shear rates, similar to the importance of GPIba binding to VWF for platelet tethering. We therefore contend that under medium/high shear conditions, VWF-GPIba platelet priming induces some rapid activation of aIIbb3, which enable the platelets to better withstand the higher shear rates, prior to their interaction/activation by collagen (Figure 8). Although most evident at the highest shear rates, GpIbaΔsig/Δsig platelets exhibited reduced accumulation at venous shear rates (Figure 7C). Given that the surface coverage on collagen was not significantly altered at 200 s-1 in GpIbaΔsig/Δsig platelets compared to wild-type platelets (Figure 7A and B), the deficit in subsequent platelet accumulation must be due to reduced reactivity of GpIbaΔsig/Δsig platelets. This is supported by the clear importance of aIIbb3 activation to this assay, demonstrated by the effects of GR144053 in preventing 3D accumulation of platelets at both 200 s-1 and 1,000 s-1 in wild-

Figure 5. GpIbaΔsig/Δsig platelets have a reduced ability to bind to collagen and form microthrombi at 3,000 s-1. Hirudin anticoagulated whole blood from GpIba+/+ and GpIbaΔsig/Δsig mice was labeled with anti-GPIba-DyLight488 antibody and perfused over fibrillar collagen type I (0.2 mg/mL) at a shear rate of 3,000 s-1 for 3 minutes. (A) Representative fluorescence images (n=6) after 3 minutes of perfusion in whole blood (WB) from GpIba+/+ and GpIbaΔsig/Δsig mice at 3,000 s-1. Platelet deposition (B) and thrombus build-up measured as integrated fluorescence intensity (IFI) (C). All data is shown as mean ± standard error of the mean and analyzed using unpaired two-tailed student’s t-test The maximal platelet IFI was used to compare the thrombus build up data. *P<0.05, ***P<0.001. Scale bar 100 mm. Also see the Online Supplementary Video S3.

haematologica | 2022; 107(4)

941

A. Constantinescu-Bercu et al.

type platelets (Figure 6A to D; Figure 7A to D). We also observed an increase in the platelet coverage in wild-type platelets in the presence of the aIIbb3 blocker. This is in line with our previous study and others showing that aIIbb3 blockade allows the formation of a platelet monolayer, but prevents thrombus growth in 3D and also lateral platelet-platelet aggregation (Figure 6B; Figure 7B).5,45,61,62 This underscores the importance in quantifying both platelet coverage and accumulation in flow assays when studying platelet signaling defects.45,61 Importantly, GR144053 did not alter these parameters when added to

GpIbaΔsig/Δsig platelets (Figure 6A, B, D and E; Figure 7A, B, D and E), demonstrating a lack of aIIbb3 activation that would be consistent with a diminished GPVI-mediated signaling response. It is important to note that this response is diminished, rather than ablated as the addition of JAQ1 led to a marked decrease in both platelet tethering and accumulation at both 1,000 s-1 and 200 s-1 shear rates (Figure 6F to I; Figure 7F to I). The question remains open as to the precise contribution of VWF-GPIba versus collagen-GPVI signaling deficits to the phenotype of GpIbaΔsig/Δsig platelets. Our data suggest that both signaling pathways likely con-

Figure 6. GpIbaΔsig/Δsig platelets have a reduced ability to bind to collagen and form microthrombi at 1,000 s-1. (A to E) Hirudin anticoagulated whole blood supplemented or not with GR144053 or plasma-free blood from GpIba+/+ and GpIbaΔsig/Δsig mice was labeled with anti-GPIba-DyLight488 antibody (Ab) and perfused over fibrillar collagen type I (0.2 mg/mL) at a shear rate of 1,000 s-1 for 3 minutes (min). (A) Representative fluorescence images (n≥3) after 3 min of perfusion in whole blood (WB), plasma-free blood (PFB) or WB + GR144053 from GpIba+/+ and GpIbaΔsig/Δsig mice. Platelet deposition (B) and thrombus build-up measured as integrated fluorescence intensity (IFI) (C to E). All data is shown as mean ± standard error of the mean and analyzed using unpaired two-tailed student’s t-test (C) or one-way ANOVA followed by Dunnett’s multiple comparison test (B, D to E). Data is compared to means from GpIba+/+ WB (B and D) or GpIbaΔsig/Δsig WB (E). The maximal platelet integrated fluorescence intensity (IFI) was used to compare the thrombus build up data. *P<0.05. Scale bar 100 mm. Also see the Online Supplementary Video S4. (F to I) Hirudin anticoagulated whole blood from GpIba+/+ and GpIbaΔsig/Δsig mice supplemented with JAQ1 or Rat-IgG control Ab (20 mg/mL) was labeled with anti-GPIbαDyLight488 Ab and perfused over fibrillar collagen type I (0.2 mg/mL) at a shear rate of 1,000 s-1 for 3 min. (F) Representative fluorescence images (n=3) after 3 min of perfusion. Platelet deposition (G) and thrombus build-up measured as IFI (H and I). All data is shown as mean ± standard error of the mean and analyzed using unpaired two-tailed student’s t-test. The maximal platelet IFI was used to compare the thrombus build up data. *P<0.05, **P<0.01. Scale bar 100 mm.

942

haematologica | 2022; 107(4)

The role of GPIba in GPVI-signaling

tribute to this, as disruption of either interaction causes a major reduction in platelet accumulation in wild-type platelets under both venous and arterial shear rates. GPVI belongs to the immunoglobulin superfamily and signals via tyrosine kinase phosphorylation pathways. In order to further investigate the defect in GPVI signaling in GpIbaΔsig/Δsig platelets, analysis of tyrosine phosphorylation downstream of GPVI revealed that SYK and PLCg2 phosphorylation was reduced in GpIbaΔsig/Δsig platelets (Figure 4H to K). Interestingly, the diminished phosphorylation was more pronounced for SYK than for PLCg2 perhaps highlighting the existence of LAT-independent mechanisms of PLCg2 phosphorylation.63 Interestingly, activation of

GpIbaΔsig/Δsig platelets via CLEC-2, another receptor that signals via an ITAM motif,64 was also affected, but perhaps to a lesser extent than those mediated by GPVI (Online Supplementary Figure S4) suggesting that the function of the GPIba intracellular tail is more important for GPVI mediated responses. Based on these findings, we hypothesize that the tail of GPIba may be important for the docking of signaling molecules such as SYK, LAT and PLCg2 that are downstream of GPVI and CLEC-2 on ITAM phosphorylated motif of the FcRg and CLEC-2 receptors and warrant further investigation. It would also be of interest to determine if the reduction in PI3K signaling in response to CRP stimulation (Figure 4I to L) is due to the lack of binding of

Figure 7. GpIbaΔsig/Δsig platelets have a reduced ability to bind to collagen and form microthrombi at 200 s-1. (A to E) Hirudin anticoagulated whole blood supplemented or not with GR144053 or plasma-free blood from GpIba+/+ and GpIbaΔsig/Δsig mice was labeled with anti-GPIba-DyLight488 Ab and perfused over fibrillar collagen type I (0.2 mg/mL) at a shear rate of 200 s-1 for 3 minutes (min). (A) Representative fluorescence images (n≥3) after 3 min of perfusion in whole blood (WB), plasma-free blood (PFB) or WB + GR144053 from GpIba+/+ and GpIbaΔsig/Δsig mice. Platelet deposition (B) and thrombus build-up measured as integrated fluorescence intensity (IFI) (C to E). All data is shown as mean ± standard error of the mean and analyzed using unpaired two-tailed student’s t-test (C) or one-way ANOVA followed by Dunnett’s multiple comparison test (B, D to E). Data is compared to means from GpIba+/+ WB (B and D) or GpIbaΔsig/Δsig WB (E). The maximal platelet integrated fluorescence intensity (IFI) was used to compare the thrombus build-up data. *P<0.05, **P<0.01. Scale bar 100 mm. Also see the Online Supplementary Video S5. (F to I) Hirudin anticoagulated whole blood from GpIba+/+ and GpIbaΔsig/Δsig mice supplemented with JAQ1 or Rat-IgG control antobodies (20 mg/mL) was labeled with antiGPIba-DyLight488 Ab and perfused over fibrillar collagen type I (0.2 mg/mL) at a shear rate of 200 s-1 for 3 min. (F) Representative fluorescence images (n=3) after 3 min of perfusion. Platelet deposition (G) and thrombus build-up measured as IFI (H and I). All data is shown as mean ± standard error of the mean and analyzed using unpaired two-tailed student’s t-test. The maximal platelet IFI was used to compare the thrombus build up data. *P<0.05. Scale bar 10 mm.

haematologica | 2022; 107(4)

943

A. Constantinescu-Bercu et al.

Figure 8. Proposed model for GPIba-GPVI cross talk. Under normal conditions, resting/circulating platelets (1) present aIIbb3 on their surface in its closed conformation. Plasma von Willebrand factor (VWF) (2) circulates in its globular conformation with its A1 domain hidden, preventing interaction with platelet GPIba. Upon vascular injury, the subendothelial extracellular matrix containing collagen becomes exposed to the blood. VWF, via its A3 domain, binds to collagen and, due to shear forces, unravels to expose its A1 domain to which platelet GPIba binds (3). Next, mechanosensitive signaling events downstream of VWF A1-GPIba that require the intracellular tail of GPIba take place leading to some activation of surface aIIbb3 (4) while the deceleration of platelets allows for the subsequent binding of platelets to collagen via several collagen receptors including GPVI (6). The intracellular tail of GPIbα is also crucial for optimal collagen/GPVI signaling that lead to platelet activation, shape change and granule release (7). Ultimately, additional circulating platelets will be recruited at the site of injury to form the hemostatic plug (8).

PI3K to the intracellular tail of GPIba or it is a consequence of diminished SYK phosphorylation.65 In summary, we generated a novel GPIba transgenic mouse in which their platelets bind VWF normally, but the subsequent VWF-GPIba signaling is disrupted. Intriguingly, these mice clearly reveal the molecular link between GPIba- and GPVI-mediated signaling in platelets and underscore the cooperative functions of these two major platelet receptors.45 Platelets in addition to their important role in thrombosis and hemostasis contribute to the host response to infection and inflammation.66-69 Our recent work suggests that VWF-GPIba-dependent platelet priming potentiates the recruitment of neutrophils, which may represent a key early event in the targeting of pathogens, but also in the development of deep vein thrombosis.5 The GpIbaΔsig/Δsig mice now provide an invaluable tool to probe the importance of the GPIba-mediated signaling in inflammatory diseases such as atherosclerosis and deep vein thrombosis, as well as in the host response to infection but also to fully decipher the molecular dependency of GPVI signaling upon GPIba. Disclosures No conflicts of interest to disclose. Contributions AC-B designed and performed experiments, analyzed data and wrote the manuscript; YAW performed experiments and

References 1. Li R. The Glycoprotein Ib-IX-V Complex. Vol. 4th. Michelson AD, Cattaneo M, Frelinger L, Newman PJ. Platelets. Elsevier/Academic Press; 2019. 2. Mazzucato M, Pradella P, Cozzi MR, De Marco L, Ruggeri ZM. Sequential cytoplasmic calcium signals in a 2-stage platelet activation process induced by the glycoprotein

944

revised the manuscript; KJW designed and performed experiments and revised the manuscript; PM and KV provided critical reagents and revised the manuscript; JTBC designed experiments, prepared the figures and wrote the manuscript; IIS-C designed and performed experiments, analyzed data, prepared the figures and wrote the manuscript. Acknowledgements The authors acknowledge the technical assistance of Alisha Miller, Elodie Ndjetehe, Ben Moyon and Zoe Webster from Central Biomedical Services and MRC transgenic group at Imperial College. We thank the LMS/NIHR Imperial Biomedical Research Center Flow Cytometry Facility for support. We would like to thank Sooriya Soman, Dr Pavarthi Sasikumar, Dr Claire Peghaire at Imperial College for technical assistance, and Nilanthi Karawitage at Imperial College Healthcare NHS trust for the use of the aggregometer. We are grateful to Professor Johannes A. Eble (University of Münster) and Dr Craig E. Hughes (University of Reading) for providing rhodocytin. Funding This work was supported by the British Heart Foundation grants FS/15/65/32036, PG/17/22/32868 and RG/18/3/33405. Data sharing statement Additional information on original data and protocols will be available upon request via email i.salles@imperial.ac.uk.

Ibalpha mechanoreceptor. Blood. 2002;100( 8):2793-2800. 3. Nesbitt WS, Kulkarni S, Giuliano S, et al. Distinct glycoprotein Ib/V/IX and integrin alpha IIbbeta 3-dependent calcium signals cooperatively regulate platelet adhesion under flow. J Biol Chem. 2002;277(4):29652972. 4. Kasirer-Friede A, Cozzi MR, Mazzucato M, et al. Signaling through GP Ib-IX-V activates alpha IIb beta 3 independently of other

receptors. Blood. 2004;103(9):3403-3411. 5. Constantinescu-Bercu A, Grassi L, Frontini M, et al. Activated alphaIIbbeta3 on platelets mediates flow-dependent NETosis via SLC44A2. Elife. 2020;9:53353. 6. Zhang W, Deng W, Zhou L, et al. Identification of a juxtamembrane mechanosensitive domain in the platelet mechanosensor glycoprotein Ib-IX complex. Blood. 2015;125(3):562-569. 7. Ju L, Chen Y, Xue L, Du X, Zhu C.

haematologica | 2022; 107(4)

The role of GPIba in GPVI-signaling

Cooperative unfolding of distinctive mechanoreceptor domains transduces force into signals. Elife. 2016;5:e15447. 8. Gu M, Xi X, Englund GD, Berndt MC, Du X. Analysis of the roles of 14-3-3 in the platelet glycoprotein Ib-IX-mediated activation of integrin alpha(IIb)beta(3) using a reconstituted mammalian cell expression model. J Cell Biol. 1999;147(5):1085-1096. 9. Mangin P, David T, Lavaud V, et al. Identification of a novel 14-3-3zeta binding site within the cytoplasmic tail of platelet glycoprotein Ibalpha. Blood. 2004;104(2):420-427. 10. Mangin PH, Receveur N, Wurtz V, et al. Identification of five novel 14-3-3 isoforms interacting with the GPIb-IX complex in platelets. J Thromb Haemost. 2009;7(9):1550-1555. 11. Dai K, Bodnar R, Berndt MC, Du X. A critical role for 14-3-3zeta protein in regulating the VWF binding function of platelet glycoprotein Ib-IX and its therapeutic implications. Blood. 2005;106(6):1975-1981. 12. Mu FT, Andrews RK, Arthur JF, et al. A functional 14-3-3zeta-independent association of PI3-kinase with glycoprotein Ib alpha, the major ligand-binding subunit of the platelet glycoprotein Ib-IX-V complex. Blood. 2008;111(9):4580-4587. 13. Wu Y, Suzuki-Inoue K, Satoh K, et al. Role of Fc receptor gamma-chain in platelet glycoprotein Ib-mediated signaling. Blood. 2001;97(12):3836-3845. 14. Li Z, Zhang G, Feil R, Han J, Du X. Sequential activation of p38 and ERK pathways by cGMP-dependent protein kinase leading to activation of the platelet integrin alphaIIb beta3. Blood. 2006;107(3):965-972. 15. Li Z, Xi X, Gu M, et al. A stimulatory role for cGMP-dependent protein kinase in platelet activation. Cell. 2003;112(1):77-86. 16. Yin H, Liu J, Li Z, et al. Src family tyrosine kinase Lyn mediates VWF/GPIb-IX-induced platelet activation via the cGMP signaling pathway. Blood. 2008;112(4):1139-1146. 17. Garcia A, Quinton TM, Dorsam RT, Kunapuli SP. Src family kinase-mediated and Erk-mediated thromboxane A2 generation are essential for VWF/GPIb-induced fibrinogen receptor activation in human platelets. Blood. 2005;106(10):3410-3414. 18. Estevez B, Stojanovic-Terpo A, Delaney MK, et al. LIM kinase-1 selectively promotes glycoprotein Ib-IX-mediated TXA2 synthesis, platelet activation, and thrombosis. Blood. 2013;121(22):4586-4594. 19. Mangin P, Yuan Y, Goncalves I, et al. Signaling role for phospholipase C gamma 2 in platelet glycoprotein Ib alpha calcium flux and cytoskeletal reorganization. Involvement of a pathway distinct from FcR gamma chain and Fc gamma RIIA. J Biol Chem. 2003;278(35):32880-32891. 20. Tsuji M, Ezumi Y, Arai M, Takayama H. A novel association of Fc receptor gammachain with glycoprotein VI and their coexpression as a collagen receptor in human platelets. J Biol Chem. 1997;272(38):2352823531. 21. Rayes J, Watson SP, Nieswandt B. Functional significance of the platelet immune receptors GPVI and CLEC-2. J Clin Invest. 2019;129(1):12-23. 22. Alshehri OM, Hughes CE, Montague S, et al. Fibrin activates GPVI in human and mouse platelets. Blood. 2015;126(13):16011608. 23. Schmaier AA, Zou Z, Kazlauskas A, et al. Molecular priming of Lyn by GPVI enables an immune receptor to adopt a hemostatic

haematologica | 2022; 107(4)

role. Proc Natl Acad Sci U S A. 2009;106(50):21167-21172. 24. Ezumi Y, Shindoh K, Tsuji M, Takayama H. Physical and functional association of the Src family kinases Fyn and Lyn with the collagen receptor glycoprotein VI-Fc receptor gamma chain complex on human platelets. J Exp Med. 1998;188(2):267-276. 25. Falati S, Edmead CE, Poole AW. Glycoprotein Ib-V-IX, a receptor for von Willebrand factor, couples physically and functionally to the Fc receptor gammachain, Fyn, and Lyn to activate human platelets. Blood. 1999;94(5):1648-1656. 26. Arthur JF, Gardiner EE, Matzaris M, et al. Glycoprotein VI is associated with GPIb-IXV on the membrane of resting and activated platelets. Thromb Haemost. 2005;93(4):716723. 27. Goto S, Tamura N, Handa S, et al. Involvement of glycoprotein VI in platelet thrombus formation on both collagen and von Willebrand factor surfaces under flow conditions. Circulation. 2002;106(2):266272. 28. Baker J, Griggs RK, Falati S, Poole AW. GPIb potentiates GPVI-induced responses in human platelets. Platelets. 2004;15(4):207214. 29. Salles-Crawley I, Monkman JH, Ahnstrom J, Lane DA, Crawley JT. Vessel wall BAMBI contributes to hemostasis and thrombus stability [Research Support, N.I.H., Extramural Research Support, Non-U.S. Gov't]. Blood. 2014;123(18):2873-2881. 30. Crawley JTB, Zalli A, Monkman JH, et al. Defective fibrin deposition and thrombus stability in Bambi(-/-) mice are mediated by elevated anticoagulant function. J Thromb Haemost. 2019;17(11):1935-1949. 31. Nakamura F, Pudas R, Heikkinen O, et al. The structure of the GPIb-filamin A complex. Blood. 2006;107(5):1925-1932. 32. Ware J, Russell S, Ruggeri ZM. Generation and rescue of a murine model of platelet dysfunction: the Bernard-Soulier syndrome. Proc Natl Acad Sci U S A. 2000;97(6):28032808. 33. Lanza F. Bernard-Soulier syndrome (hemorrhagiparous thrombocytic dystrophy). Orphanet J Rare Dis. 2006;1:46. 34. Denis C, Methia N, Frenette PS, et al. A mouse model of severe von Willebrand disease: defects in hemostasis and thrombosis [Research Support, Non-U.S. Gov't Research Support, U.S. Gov't, P.H.S.]. Proc Natl Acad Sci U S A. 1998;95(16):9524-9529. 35. Kanaji T, Russell S, Ware J. Amelioration of the macrothrombocytopenia associated with the murine Bernard-Soulier syndrome. Blood. 2002;100(6):2102-2107. 36. Stalker TJ. Mouse laser injury models: variations on a theme. Platelets. 2020;31(4):423431. 37. Dubois C, Panicot-Dubois L, Gainor JF, Furie BC, Furie B. Thrombin-initiated platelet activation in vivo is vWF independent during thrombus formation in a laser injury model. J Clin Invest. 2007;117(4):953-960. 38. Dubois C, Panicot-Dubois L, Merrill-Skoloff G, Furie B, Furie BC. Glycoprotein VI-dependent and -independent pathways of thrombus formation in vivo. Blood. 2006;107(10):3902-3906. 39. Fukuda K, Doggett T, Laurenzi IJ, Liddington RC, Diacovo TG. The snake venom protein botrocetin acts as a biological brace to promote dysfunctional platelet aggregation. Nat Struct Mol Biol. 2005;12(2):152-159. 40. Estevez B, Kim K, Delaney MK, et al.

Signaling-mediated cooperativity between glycoprotein Ib-IX and protease-activated receptors in thrombin-induced platelet activation. Blood. 2016;127(5):626-636. 41. Durrant TN, van den Bosch MT, Hers I. Integrin alphaIIbbeta3 outside-in signaling. Blood. 2017;130(14):1607-1619. 42. Ruggeri ZM. The role of von Willebrand factor in thrombus formation. Thromb Res. 2007;120 Suppl 1:S5-9. 43. Kuijpers MJ, Schulte V, Oury C, et al. Facilitating roles of murine platelet glycoprotein Ib and alphaIIbbeta3 in phosphatidylserine exposure during vWF-collagen-induced thrombus formation. J Physiol. 2004;558(Pt 2):403-415. 44. Nieswandt B, Brakebusch C, Bergmeier W, et al. Glycoprotein VI but not alpha2beta1 integrin is essential for platelet interaction with collagen. EMBO J. 2001;20(9):21202130. 45. Pugh N, Simpson AM, Smethurst PA, et al. Synergism between platelet collagen receptors defined using receptor-specific collagenmimetic peptide substrata in flowing blood. Blood. 2010;115(24):5069-5079. 46. Bergmeier W, Piffath CL, Goerge T, et al. The role of platelet adhesion receptor GPIbalpha far exceeds that of its main ligand, von Willebrand factor, in arterial thrombosis. Proc Natl Acad Sci U S A. 2006;103(45):16900-16905. 47. Nieswandt B, Pleines I, Bender M. Platelet adhesion and activation mechanisms in arterial thrombosis and ischaemic stroke. J Thromb Haemost. 2011;9 Suppl 1:92-104. 48. Feng S, Resendiz JC, Lu X, Kroll MH. Filamin A binding to the cytoplasmic tail of glycoprotein Ibalpha regulates von Willebrand factor-induced platelet activation. Blood. 2003;102(6):2122-2129. 49. Kanaji T, Russell S, Cunningham J, et al. Megakaryocyte proliferation and ploidy regulated by the cytoplasmic tail of glycoprotein Ibalpha. Blood. 2004;104(10):31613168. 50. Morowski M, Vogtle T, Kraft P, et al. Only severe thrombocytopenia results in bleeding and defective thrombus formation in mice. Blood. 2013;121(24):4938-4947. 51. David T, Strassel C, Eckly A, et al. The platelet glycoprotein GPIbbeta intracellular domain participates in von Willebrand factor induced-filopodia formation independently of the Ser 166 phosphorylation site. J Thromb Haemost. 2010;8(5):1077-1087. 52. Yin H, Stojanovic-Terpo A, Xu W, et al. Role for platelet glycoprotein Ib-IX and effects of its inhibition in endotoxemia-induced thrombosis, thrombocytopenia, and mortality. Arterioscler Thromb Vasc Biol. 2013;33(11):2529-2537. 53. Schoenwaelder SM, Darbousset R, Cranmer SL, et al. 14-3-3zeta regulates the mitochondrial respiratory reserve linked to platelet phosphatidylserine exposure and procoagulant function. Nat Commun. 2016;7:12862. 54. Andrews RK, Berndt MC. Bernard-Soulier syndrome: an update. Semin Thromb Hemost. 2013;39(6):656-662. 55. Walsh PN, Mills DC, Pareti FI, et al. Hereditary giant platelet syndrome. Absence of collagen-induced coagulant activity and deficiency of factor-XI binding to platelets. Br J Haematol. 1975;29(4):639655. 56. Li J, Dai K, Wang Z, et al. Platelet functional alterations in a Bernard-Soulier syndrome patient with filamin A mutation. J Hematol Oncol. 2015;8:79. 57. Kato K, Kanaji T, Russell S, et al. The contri-

945

A. Constantinescu-Bercu et al. bution of glycoprotein VI to stable platelet adhesion and thrombus formation illustrated by targeted gene deletion [Research Support, U.S. Gov't, P.H.S.]. Blood. 2003;102(5):1701-1707. 58. Mazharian A, Wang YJ, Mori J, et al. Mice lacking the ITIM-containing receptor G6b-B exhibit macrothrombocytopenia and aberrant platelet function. Sci Signal. 2012;5(248):ra78. 59. Nieswandt B, Watson SP. Platelet-collagen interaction: is GPVI the central receptor? Blood. 2003;102(2):449-461. 60. Kuijpers MJ, Schulte V, Bergmeier W, et al. Complementary roles of glycoprotein VI and alpha2beta1 integrin in collageninduced thrombus formation in flowing whole blood ex vivo. FASEB J. 2003;17(6):685-687.

946

61. Pugh N, Maddox BD, Bihan D, et al. Differential integrin activity mediated by platelet collagen receptor engagement under flow conditions. Thromb Haemost. 2017;117(8):1588-1600. 62. Verkleij MW, Morton LF, Knight CG, et al. Simple collagen-like peptides support platelet adhesion under static but not under flow conditions: interaction via alpha2 beta1 and von Willebrand factor with specific sequences in native collagen is a requirement to resist shear forces. Blood. 1998;91(10):3808-3816. 63. Pasquet JM, Gross B, Quek L, et al. LAT is required for tyrosine phosphorylation of phospholipase cgamma2 and platelet activation by the collagen receptor GPVI. Mol Cell Biol. 1999;19(12):8326-8334. 64. Suzuki-Inoue K, Inoue O, Ozaki Y. Novel

platelet activation receptor CLEC-2: from discovery to prospects. J Thromb Haemost. 2011;9 Suppl 1:44-55. 65. Manne BK, Badolia R, Dangelmaier C, et al. Distinct pathways regulate Syk protein activation downstream of immune tyrosine activation motif (ITAM) and hemITAM receptors in platelets. J Biol Chem. 2015;290(18):11557-11568. 66. Clark SR, Ma AC, Tavener SA, et al. Platelet TLR4 activates neutrophil extracellular traps to ensnare bacteria in septic blood. Nat Med. 2007;13(4):463-469. 67. Deppermann C, Kubes P. Platelets and infection. Semin Immunol. 2016;28(6):536-545. 68. Jenne CN, Kubes P. Platelets in inflammation and infection. Platelets. 2015;26(4):286-292. 69. Kapur R, Semple JW. Platelets as immunesensing cells. Blood Adv. 2016;1(1):10-14.

haematologica | 2022; 107(4)

ARTICLE

Platelet Biology & its Disorders

Comparative analysis of ChAdOx1 nCoV-19 and Ad26.COV2.S SARS-CoV-2 vector vaccines

Ferrata Storti Foundation

Stephan Michalik,1* Florian Siegerist,2* Raghavendra Palankar,3* Kati Franzke,4* Maximilian Schindler,2 Alexander Reder,1 Ulrike Seifert,5 Clemens Cammann,5 Jan Wesche,3 Leif Steil,1 Christian Hentschker,1 Manuela GesellSalazar,1 Emil Reisinger,6 Martin Beer,7# Nicole Endlich,2# Andreas Greinacher3# and Uwe Völker1# 1 Interfaculty Institute of Genetics and Functional Genomics, Department of Functional Ge-nomics, University Medicine Greifswald, Greifswald; 2Institute for Anatomy and Cell Biology, University Medicine Greifswald, Greifswald; 3Institute of Transfusion Medicine, University Medicine Greifswald, Greifswald; 4Institute of Infectiology, Friedrich-LoefflerInstitut, Greifswald-Insel Riems; 5Friedrich Loeffler-Institute of Medical MicrobiologyVirology, University Medicine Greifswald, Greifswald; 6Division of Tropical Medicine and Infectious Diseases, Center of Internal Medicine II, Rostock University Medical Center, Rostock, and 7Institute of Diagnostic Virology, Friedrich-Loeffler-Institut, Greifswald-Insel Riems, Germany

Haematologica 2022 Volume 107(4):947-957

*SM, FS, RP and KF contributed equally as co-first authors #

MB, NE, AG and UV contributed equally as co-senior authors

ABSTRACT

ector-based SARS-CoV-2 vaccines have been associated with vaccine-induced thrombosis with thrombocytopenia syndrome (VITT/TTS), but the causative factors are still unresolved. We comprehensively analyzed the ChAdOx1 nCoV-19 (AstraZeneca) and Ad26.COV2.S (Johnson & Johnson) vaccines. ChAdOx1 nCoV-19 contains significant amounts of host cell protein impurities, including functionally active proteasomes, and adenoviral proteins. A much smaller amount of impurities was found in Ad26.COV2.S. Platelet factor 4 formed complexes with ChAdOx1 nCoV-19 constituents, but not with purified virions from ChAdOx1 nCoV-19 or with Ad26.COV2.S. Vascular hyperpermeability was induced by ChAdOx nCoV-19 but not by Ad26.COV2.S. These differences in impurities together with EDTAinduced capillary leakage might contribute to the higher incidence rate of VITT associated with ChAdOx1 nCoV-19 compared to Ad26.COV2.S.

Introduction Vaccination is key for the control of the severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) pandemic. Adenoviral vector-, mRNA encapsulated in lipid nanoparticles-, and antigen-based vaccines are currently in use, all encoding the spike protein.1,2 Since February 2021 the rare but severe adverse reaction of vaccine-induced immune thrombotic thrombocytopenia (VITT; synonym thrombosis with thrombocytopenia syndrome [TTS]) has been observed in individuals vaccinated against SARS-CoV-2. VITT/TTS occurs 5-20 days (occasionally later) after vaccination with the ChAdOx1 nCoV-19 vaccine (produced by AstraZeneca) and the Ad26.COV2.S vector vaccine (produced by Janssen/Johnson & Johnson). The incidence rate of VITT/TTS seems to be higher for ChAdOx1 nCoV-19. The reported rate of VITT/TTS in the USA is 0.355 cases per 100,000 people vaccinated with Ad26.COV2.S,3 compared to 1 per 50,000-100,000 people vaccinated with ChAdOx1 nCoV-19 in the UK.4 In Germany both vaccines were used and within this medical system, 0.56 suspected cases per 100,000 vaccine doses for Ad26.COV2.S (3,186,297 vaccine doses administered) and 1.49 suspected cases per 100,000 vaccine doses for ChAdOx1 nCoV-19 (12,692,700 vaccine doses administered) were reported.5 VITT/TTS involves high-affinity, platelet-activating anti-platelet factor 4 (PF4) antibodies,6–8 but the mechanisms triggering these anti-PF4 antibodies are still unre-

haematologica | 2022; 107(4)

Correspondence: ANDREAS GREINACHER, andreas.greinacher@med.uni-greifswald.de UWE VÖLKER voelker@uni-greifswald.de Received: October 5, 2021. Accepted: January 5, 2022. Pre-published: January 20, 2022. https://doi.org/10.3324/haematol.2021.280154

947

S. Michalik et al.

solved. VITT/TTS shows striking similarities with another PF4-mediated adverse drug effect, heparin-induced thrombocytopenia (HIT) and autoimmune HIT. In HIT, polyanions form complexes with PF4, inducing a conformational change, which triggers anti-PF4 antibodies. This immune response is pronounced in patients with tissue trauma and inflammation. We have shown that one or more constituents of the ChAdOx1 nCoV-19 vaccine interact with PF4, forming complexes which contain PF4 and the adenovirus hexon protein. This might trigger conformational changes in the positively charged chemokine PF4 leading to the formation of a neo-antigen and then subsequent activation of B cells in a pro-inflammatory environment.9 These activated B cells then produce high avidity anti-PF4 antibodies that bind PF4 and trigger an activation cascade of platelets and granulocytes, leading to NETosis and massive thrombin production. However, it is not known which vaccine components, beside the hexon protein, interact with PF4 and which additional factors influence this interaction. Both vaccines (ChAdOx1 nCoV-19 and Ad26.COV2.S) are produced in human cell lines, T-REx-293 cells (human embryonic kidney cells, a HEK293 derivate) for ChAdOx1 nCoV-19 and PER.C6 TetR cells (human embryonic retinal cells) for Ad26.COV2.S. We and others have previously shown that the ChAdOx1 nCoV-19 vaccine contains a large number of host cell proteins (HCP).9,10 Here we report the results of a comprehensive, comparative analysis of the ChAdOx1 nCoV-19 and Ad26.COV2.S vaccines, using proteomics, transmission electron microscopy, dynamic lightscattering, single-molecule light microscopy, and an in vivo capillary leakage assay. Our data reveal substantial differences in composition and functional properties between the two vaccines, which may contribute to the different incidences of VITT/TTS.

precast NuPAGE™ 4 to 12% gel. Electrophoresis was performed at 150 V. The western blots were prepared using the Trans-Blot Turbo Transfer System from BioRad and the transfer of proteins to the polyvinylidene fluoride membrane was verified and documented using LICOR's Revert Total Protein Stain protocol (Doc # 988-19494). The specific proteins were detected with primary and secondary antibodies described in the Online Supplementary Methods.

Proteasome activity assays Chymotrypsin-like activity was assessed in vaccine or HEK293T cell lysate using 0.2 mM fluorescently tagged Suc-LLVYAMC (Bachem, Bubendorf, Switzerland) quantified with a fluorometer using a 380/460 nm filterset. To confirm proteasomal activity 100 nM bortezomib14 or 200 nM carfilzomib (Selleckchem, Houston, TX, USA)15, was added.

Dynamic light scattering and zeta potential All dynamic light scattering measurements were performed in a fixed scattering angle Zetasizer Nano-S system (Malvern Instruments Ltd., Malvern, UK). The hydrodynamic diameter (nm) was measured at 25°C, and light scattering was detected at 173°. Surface 𝜁 potential was performed in folded capillary 𝜁 cells (DTS1070, Malvern Instruments Ltd., Malvern, UK). Data were analyzed using Zetasizer software, version 7.13 (Malvern Instruments Ltd., Malvern, UK).

Immunoelectron and transmission electron microscopy Vaccines or the purified adenovirus particles were incubated with biotinylated PF4 and transferred to formvar-coated transmission electron microscopy grids. After washing, samples were labeled with an anti-adenovirus monoclonal antibody detected by a gold conjugate. The same samples were labeled with a streptavidin-gold conjugate to detect PF4-biotin. Grids were stained with 1% phosphotungstic acid and analyzed with a Tecnai-Spirit transmission electron microscope (FEI, Eindhoven, Germany).

Methods Super resolution single-molecule light microscopy Comprehensive details of the Methods are described in the Online Supplementary Material. All experiments were performed in accordance with local and national ethics standards and German animal protection legislation, overseen by the “Landesamt für Landwirtschaft, Lebensmittelsicherheit und Fischerei, Rostock” of the federal state of Mecklenburg - Western Pomerania.

Sample preparation and liquid chromatography tandem mass spectrometry. Vaccines were precipitated using salt-acetone precipitation.11 Adenovirus particles were purified using subsequent sucrosecushion and sucrose-gradient ultracentrifugation. Protein was digested with trypsin as described by Blankenburg et al.12 Liquid chromatography tandem mass spectrometry (LC-MS/MS) experiments were performed on an Orbitrap ExplorisTM 480 mass spectrometer (Thermo Scientific, Bremen, Germany) coupled to an UltimateTM 3000 RSLCnano HPLC (Dionex/ Thermo Scientific, Waltham, MA, USA). The mass spectrometry proteomics data have been deposited with the ProteomeXchange Consortium via the PRIDE18 partner repository13 with the dataset identifier PXD027344.

Diluted vaccine or purified virions were incubated with human PF4 and immobilized on cleaned coverslips. After fixation and blocking, PF4 and adenoviral hexon were visualized using secondary (PF4) and primary (hexon) immunofluorescence detected by Alexa Fluor 488 and Cy5. Coverslips were mounted in Everspark dSTORM buffer (Idylle Labs, France)16 and blinking sequences imaged on a Zeiss Elyra PS.1 super resolution system. Single-molecule localization microscopy data were processed in FIJI using NanoJ core17 and Thunderstorm18 and analyzed using custom FIJI19 scripts.

Zebrafish vascular permeability assay A novel zebrafish-based in vivo assay was developed to determine local changes of vascular permeability following intramuscular injections: Transgenic zebrafish at 5 days post-fertilization expressing a 78 kDa GFP-tagged plasma protein20,21 were injected intramuscularly with 1 nL of native vaccine, purified virions, 100 µM EDTA or 0.9% NaCl. Fluorescence intensity ratios (intravascular vs. intramuscular) were measured in the direct vicinity of the injection site at t=0 and t=10 min.

Results Sodium dodecylsulfate gel electrophoresis and western blot analysis For protein separation, one-fiftieth of one vaccine dose, as well as dilutions of HEK293 total protein lysate, were loaded onto a

948

Comparative profiling of ChAdOx1 nCoV-19 and Ad26.COV2.S vaccines Comparative profiling of ChAdOx1 nCoV-19 and haematologica | 2022; 107(4)

Comparative analysis of SARS-CoV-2 vector vaccines

Ad26.COV2.S (three different lots each) consistently revealed significant differences: (i) the total protein concentration of the ChAdOx1 nCoV-19 vaccine was approximately 3.4-times higher than that of the Ad26.COV2.S vaccine (mean: 102 ng/µL vs. 29.8 ng/µL) (Figure 1A); (ii) silvernitrate staining of vaccines separated by sodium dodecylsulphate polyacrylamide gel electrophoresis (SDS-PAGE) displayed a markedly more complex protein pattern than expected for pure virions for ChAdOx1 nCoV-19 compared to Ad26.COV2.S (Figure 1B). (iii) mass spectrometric analysis (Online Supplementary Table S1) identified a much higher proportion (44.5% to 59.2% vs. only 0.26% to 0.96%) (Figure 1D, Online Supplementary Figures S1 and S2) and = 1571±31 vs. N = 59±14; twonumber (N sided t-test P=8.709x10-6) (Online Supplementary Figure S2) of ChAdOx1 nCoV-19

Ad26.COV2.S

host cell-derived human proteins. A dilution series of a laboratory HEK293 cell line lysate, which was used instead of the cell line used for vaccine production, confirmed the quantities of host-cell proteins (54% for ChAdOx1 nCoV19 and 1.5% for Ad26.COV2.S) (Online Supplementary Figure S3). None of the top ten most abundant human proteins in ChAdOx1 nCoV-19 was detected in Ad26.COV2.S (Online Supplementary Table S2). Adenoviral proteins accounted for 40.8-55.5% (ChAdOx1 nCoV-19) and 99.04-99.74% (Ad26.COV2.S) of total ion intensity, while the SARS-CoV2 spike protein was detected only in the ChAdOx1 nCoV19 vaccine (3 different lots) (Online Supplementary Figure S1). Western blot analysis confirmed the significant abundance of 11 selected host-cell proteins in the ChAdOx1 nCoV-19 vaccine, which even exceeded the abundances detected in

Figure 1. Analysis of the protein composition of ChAdOx1 nCoV-19 and Ad26.COV2.S vaccines. (A) Determination of the protein concentration of the two vaccines (3 different lots each). Protein concentration was determined with a quantitative bicinchoninic acid (BCA) assay. Protein concentration per 500 µL (vaccination dose) and 1 µL vaccine (secondary axis) of three lots of ChAdOx1 nCoV-19 or Ad26.COV2.S vaccine, respectively, are shown. Statistical testing was performed using a twosided t-test. (B) Protein patterns of silver nitrate-stained sodium dodecylsulfate polyacrylamide gel electrophoresis of ChAdOx1 nCoV-19 or Ad26.COV2.S (3 lots each) vaccines along with a dilution series of a laboratory HEK293 cell line extract for comparison. The HEK293 cell extract was loaded onto the gel at 1.5, 1.0, 0.5, or 0.25 µg per lane, and 10 µL (1/50th of a vaccine dose) were used for each vaccine. (C) Western blot analysis of HSP90-a protein, using the same gel loading scheme as for the silver nitrate-stained gel. (D) intensity-based absolute quantification (iBAQ) protein intensities and theoretical molecular mass of identified proteins. Protein intensities of ChAdOx1 nCoV-19 or Ad26.COV2.S (exemplarily shown for lot 3) were calculated using the iBAQ algorithm (minimum of 3 unique peptides per protein) and plotted against the theoretical molecular mass. Proteins are color-coded according to their respective class. Blue dots indicate vector proteins; gray dots represent human proteins; the red dot indicates the SARS-CoV-2 spike protein. Points highlighted with a cross indicate proteins additionally analyzed by western blotting in Online Supplementary Figures S4-S7.

haematologica | 2022; 107(4)

949

S. Michalik et al.

the HEK293 cell line. None of these proteins was detected in the Ad26.COV2.S vaccine (Figure 1C, Online Supplementary Figure S4). In summary, per vaccine dose (500 µL) of ChAdOx1 nCoV-19 we detected 19.1-33.8 µg host-cell proteins and 23.3-26.3 µg chimpanzee adenovirus proteins and for the Ad26.COV2.S vaccine 0.04-0.19 µg host-cell proteins and 10.2-19.2 µg adenoviral proteins. Since the approximately 5x1010 virions per vaccine dose weigh about 12.5 µg, both vaccines contain unassembled virus proteins, mostly hexon proteins. However, the amount of approximately 10-14 µg unassembled virus proteins in ChAdOx1 nCoV-19 was again much larger compared to the 0-6.5 µg in Ad26.COV2.S.

Proteasome activity in the different vaccines Proteasome subunits were identified by mass spectrometry and verified by western blot analysis (Figure 2A, Online Supplementary Figure S5). Chymotrypsin-like activity associated with the proteasomal b-5 subunit showed lot-dependent high levels in ChAdOx1 nCoV-19, while in Ad26.COV2.S only minimal proteasome activity was found in one of three lots (Figure 2B, C). Substrate turnover

varied in the different lots according to the proteasome subunit expression level (Figure 2A, C). Inhibition of the proteasome activity by 100 nM bortezomib14 or 200 nM carfilzomib15 confirmed assay specificity (Online Supplementary Figure S6).

Platelet factor 4-vaccine cluster formation PF4 is the key protein involved in the immune response causing VITT. We assessed the interaction of PF4 with the native vaccines and the purified adenoviral particles of ChAdOx1 nCoV-19, which were obtained by sucrose cushion and gradient ultracentrifugation. The purity of isolated ChAdOx1 nCoV-19 virions was confirmed by transmission electron microscopy and silver-staining of one-dimensional SDS-PAGE (Online Supplementary Figures S7 and S8). Dynamic light scattering confirmed PF4-induced clustering of the non-purified ChAdOx1 nCoV-19 vaccine (Figure 3A, left panel). The hydrodynamic diameter increased from 88±2.4 nm up to 151±12 nm and 320±45 nm with 10 µg/mL and 50 µg/mL PF4, respectively (Figure 3A; the PF4 dose-dependent size change is shown in Online Supplementary Figure S9). This complex formation was reversible upon the addition of unfractionated heparin and

Figure 2: Analysis of proteasome proteins and activity in the vaccines. (A) Western blot analysis of proteasome 20S subunit beta 5 (the original western blot image is provided in Online Supplementary Figure S4). (B) Fifty microliters (1/10th of the vaccination dose) of different ChAdOx1 nCoV-19 (n=5) and Ad26.COV2.S (n=3) lots were analyzed for the chymotrypsin-like activity of the proteasome and compared with the activity of 0.25 µg of HEK293 cell lysate. The mean ± standard deviation of two technical replicates are shown. (C) Proteasomal activity was confirmed by inhibition with 100 nM bortezomib. The mean of two technical replicates is shown.

950

haematologica | 2022; 107(4)

Comparative analysis of SARS-CoV-2 vector vaccines

can be attributed to weakened electropositive surface potential of PF4 in the presence of highly negatively charged unfractionated heparin thus decreasing its ability to interact electrostatically with vaccine components. In contrast, the addition of PF4 (50 µg/mL) only marginally increased the particle size when incubated with purified virions from ChAdOx1 nCoV-19 (from 79.6±10.3 nm to 86.7±6.22 nm; P=0.2404) (Figure 3A, middle panel) or the Ad26.COV2.S vaccine (from 85.7±2.2 nm to 91.3±2.83 nm; P=0.0620) (see particle size-frequency distribution plots in Online Supplementary Figure S10). Complex formation of PF4 with the ChAdOx1 nCoV-19 vaccine was charge-dependent, as the negative charge of ChAdOx1 nCoV-19 (𝜁 potential -27.5±4.7) was neutralized by PF4. In comparison, both, purified ChAdOx1 nCoV-19

virions (𝜁 potential -1.7±4.7) and the untreated Ad26.COV2.S vaccine (𝜁 potential -4.5±5.7) showed only a minimal negative charge (Figure 3B). Consistent with dynamic light scattering findings and as described before,6 PF4 induced the formation of electrondense aggregates with ChAdOx1 nCoV-19 (Figure 4A), which contained unassembled hexon proteins (Online Supplementary Figure S11). In contrast, no comparable aggregates were detected after incubation of PF4 with purified virions from ChAdOx1 nCoV-19 (Figure 4B) or the Ad26.COV2.S vaccine (Figure 4C). PF4 single-molecule density analysis using single-molecule localization microscopy (Online Supplementary Figures S12 and S13) revealed that PF4 clusters formed on or in close vicinity to ChAdOx1 nCoV-19 adenoviral hexon pro-

B Figure 3. Dynamic light scattering analysis of vaccineinduced platelet factor 4 clustering. Analysis of ChAdOx1 nCoV-19 vaccine, purified ChAdOx1 nCoV-19 virions and Ad26.COV2.S vaccine by dynamic light scattering. (A) The hydrodynamic diameter (mean ± standard deviation [SD], n=9) of ChAdOx1 nCoV-19 or Ad26.COV2.S particles before and after addition of 10 µg/mL or 50 µg/mL platelet factor 4 (PF4) was determined. A dose-dependent increase in the size of ChAdOx1 nCoV-19 aggregates in the presence of PF4 was detected. This effect was markedly reduced for both purified ChAdOx1 nCoV-19 virions and Ad26.COV2.S. Addition of unfractionated heparin (UFH; 1 IU/mL) dissociated complexes between PF4 and vaccine components. (B) The ζ-potential (mean ± SD, n=9) of ChAdOx1 nCoV-19 was lower than the purified ChAdOx1 nCoV-19 virions or of Ad26.COV2.S; and largely neutralized by PF4. In the presence of UFH, a charge reversal to net negative charge was observed in both vaccines and purified virions from ChAdOx1 nCoV-19 vaccine. Statistical analysis was performed by one-way analysis of variance on ranks/KruskalWallis followed by correction for multiple comparison by two-stage Benjamini, Krieger, & Yekutieli controlling the false discovery rate procedure (n=9) and a <0.05 was considered significant.

haematologica | 2022; 107(4)

951

S. Michalik et al.

teins (PF4 single-molecule density ratio 7.09±1.38, Figure 4D, G; arrowheads in Online Supplementary Figure S12), but not on Ad26.COV2.S (1.13±0.14, P<0.0001) or purified ChAdOx1 nCoV-19 virions (2.33±0.44, P=0.0115, mean ± standard error of mean; P-values refer to comparison with ChAdOx1 nCoV-19) (Figure 4D-G). Dynamic light scattering experiments with ultracentrifugation-separated virions from the vaccines and the resulting supernatant confirmed the single-molecule density analysis of PF4 binding (Figure 5). Electron

microscopy analysis revealed absence of intact virions in the supernatant fraction of both ChAdOx1 nCoV-19 and Ad26.COV2.S, while the pellet was enriched in virions (Figure 5A, B). Intriguingly, we observed amorphous electron-dense irregularly shaped particulate material in ChAdOx1 nCoV-19 supernatant that was absent in Ad26.COV2.S. No significant PF4-dependent complex formation was detected by dynamic light scattering with the pellet fractions of ChAdOx1 nCoV-19 and Ad26.COV2.S, but the supernatant fraction of ChAdOx1

Figure 4. Ultrastructural analysis of vaccine-induced platelet factor 4 clustering. (A-C) Representative micrographs of streptavidin-gold immunoelectron microscopy (arrow); biotinylated, human platelet factor 4 (PF4) was incubated with either ChAdOx1 nCoV-19 (A), purified virions from ChAdOx1 nCoV-19 (B) or Ad26.COV2.S (C). Virions are exemplarily labeled by asterisks, and immunogoldlabeled aggregates by arrowheads; scale bars represent 200 nm. (D-F) Singlemolecule light microscopy show representative dual PF4 and hexon polypeptide reconstructions after incubation with either ChAdOx1 nCoV-19 (D), purified virions from ChAdOx1 nCoV-19 (E) or Ad26.COV2.S (F). As indicated by the green signal in close proximity to adenoviral particles in (D), PF4 was found in dense clusters after incubation with ChAdOx1 nCoV-19 whereas PF4 formed a more homogeneous layer on glass after incubation with purified ChAdOx1 nCoV-19 (E) or Ad26.COV2.S (F). For quantification, single-molecule particle analysis was performed and particle density ratios (on viral particles/on glass) analyzed (Online Supplementary Figure S10). (G) Relative PF4 particle densities, showing a statistically significant affinity of PF4 to adenoviral particles predominantly on ChAdOx1 nCoV-19 but not on purified ChAdOx1 nCoV-19 virions or Ad26.COV2.S. Statistical analysis of n=92 particles was performed with the Kruskal-Wallis test with the Dunn correction for multiple comparisons. Respective P-values are indicated in the plot, red lines and whiskers indicate mean ± standard deviation. The dashed line is at y=1 (equal affinity of PF4 for adenoviral hexon and glass). Full-field of view images and single-channel reconstructions are provided in Online Supplementary Figure S9. Scale bars represent 100 nm.

952

haematologica | 2022; 107(4)

Comparative analysis of SARS-CoV-2 vector vaccines

nCoV-19 showed clear PF4 complex formation potential (Figure 5B). However, it is important to note that under in vitro conditions, PF4 binds to both chimpanzee adenovirus Y25 (ChAdOx1) and human adenoviruses (HAdVD26 and HAdV-C5) through weak electrostatic interactions that were abrogated in the presence of fondaparinux, a heparin pentasaccharide22. Proteomic analysis showed that the host-cell protein content in the pellet fraction was reduced compared to the non-fractionated vaccine and a large proportion of the ChAdOx1 nCoV-19 host-cell proteins was located in the supernatant fraction. As expected, the viral proteins were enriched in the pellet fraction (Figure 5E, Online Supplementary Figures S14-S16).

ChAdOx1 nCoV-19 induced vascular hyperpermeability To study the effect of the two vaccines and EDTA (100 µM present in the ChAdOx1 nCoV19 vaccine) on vascular permeability, we used in vivo microscopy of transgenic zebrafish larvae expressing an eGFP-tagged plasma protein (gc-eGFP, 78kDa20). Intramuscular injections (Online Supplementary Video S1) of 1 nL of 100 µM EDTA or ChAdOx1 nCoV-19 locally increased vascular permeability, indicated by leakage of eGFP from the intravascular to the intramuscular compartment (P=0.0001), but this was not observed after injection of purified ChAdOx1 nCoV19 virions, Ad26.COV2.S or physiological saline (Figure 6).

Discussion Our comprehensive analyses revealed major differences between ChAdOx1 nCoV-19 and Ad26.COV2.S vaccines. Confirming our previous observation, a high proportion of host-cell proteins (54%) was found in the ChAdOx1 nCoV-19 vaccine, but only a very low level was found in Ad26.COV2.S (1.5%). This observation suggests very different purification approaches and purification efficiencies for the two vaccines, with a more thorough purification of adenoviruses in Ad26.COV2.S. Such differences might be caused by the use of detergent treatment of the infected production cell culture for ChAdOx1 nCoV-19, which will make subsequent purification strategies more complicated.23 The SARS-CoV-2 spike protein was detected only in the ChAdOx1 nCoV-19 vaccine. There may be several reasons for this, since different cell lines and procedures were used for production and purification of the two vaccines. Both suppliers use systems in which the TetR repressor suppresses the expression of the SARS-CoV-2 transgene during the production of the recombinant adenoviral vectors. A different degree of leakage of repression might occur in the cell lines thus allowing different residual expression of the spike protein. Another reason could be depletion of the spike protein during purification of Ad26.COV2.S. This can only be differentiated by inprocess sampling at different production steps of the vaccines. The two vaccines display differences in their ability to interact with PF4. We confirm the previously observed complex formation of PF4 with ChAdOx1 nCoV-19. Consistent with a recent study using cryo-electron microscopy of ChAdOx1 nCoV-19,22 in our study addition of PF4 to purified ChAdOx1 nCoV-19 virions or to Ad26.COV2.S also resulted in a slight increase in particle haematologica | 2022; 107(4)

size. This was reversed by addition of heparin, indicating charge-related binding of PF4 to the virions. This is again consistent with the data obtained by cryo-electron microscopy and in silico modeling of ChAdOx1 nCoV-19, which established a possible electrostatic interaction of positively charged PF4 and negatively charged adenovirus hexon polypeptide.22 However, the interaction Kd of about 300 nmol was rather weak. This is likely the reason that we could not demonstrate distinct complexes of PF4 with Ad26.COV2.S or purified ChAdOx1 nCoV-19 virion preparations. This further supports a role of impurities in ChAdOx1 nCoV-19 for the observed formation of large complexes in ChAdOx1 nCoV-19 after addition of PF4.9 We have shown that unassembled adenoviral hexon proteins are part of the PF4-complexes, but we still cannot exclude a contribution of additional constituents of the vaccine supernatant. Antibody formation against PF4 is enhanced by antigen presentation in an inflammatory environment. Recently, we have shown (in collaboration with the laboratory of Prof. Thomas Renne, Universitätsmedizin Hamburg Eppendorf, Germany) that intradermal injection of ChAdOx1 nCoV-19 leads to EDTA-induced capillary leakage in the Mile skin edema assay, thus increasing the vaccine’s intravascular distribution.9 Our zebrafish model allows intramuscular injection, which recapitulates the actual mode of vaccination. EDTA and ChAdOx1 nCoV19 but not Ad26.COV2.S rapidly induced local vascular hyperpermeability. Such an increase of local capillary leakage might facilitate direct contact of the immune system with vaccine components, as does accidental intravascular administration of the vaccine.24 Inflammation early after vaccination may also be enhanced when host-cell proteins are recognized by endogenous natural IgG.25 These natural antibodies bind proteins of degrading cells and can form immune complexes. Furthermore, proteasome activity was detectable in both vaccines, again with a remarkable difference between Ad26.COV2.S (only low proteasome activity and protein abundance in one lot) compared to substantially higher proteasome activities in almost all lots of ChAdOx1 nCoV-19. This is of particular interest since Hauler et al. showed that adenoviral capsid proteins are intracellularly processed by proteasomal degradation which is mediated by the chaperone p97/VCP.26 In our proteomic analysis, we identified VCP as one of the top five most abundant proteins of the ChAdOx1 nCoV-19 vaccine (Online Supplementary Tables S1 and S2). Proteasomal degradation of adenoviral components such as the hexon polypeptide and/or host-cell proteins might lead to a reduction in vaccine efficiency. Whether proteasomal activity may also create potentially immunogenic or immunoreactive neo-antigens remains unresolved. VITT/TTS occurs clinically after vaccination with ChAdOx1 nCoV-19 or Ad26.COV2.S and is mediated by platelet-activating anti-PF4 antibodies. Recently, we and others have shown that anti-spike protein antibodies and anti-PF4 antibodies react independently of each other, therefore making it unlikely that VITT is caused by crossreacting anti-spike protein antibodies.27,28 The current data indicate that adenovirus particles and free hexon proteins are the common features of both vaccines. We only observed formation of larger complexes of PF4 with ChAdOx1 nCoV-19. This indicates that an additional 953

S. Michalik et al. A

Figure 5. Analysis of vaccine components prepared by ultracentrifugation. (A, C) Representative transmission electron micrographs of (A) ChAdOx1 nCoV-19 and (C) Ad26.COV2.S vaccine and their respective supernatants and pellets obtained after ultracentrifugation. Asterisks and arrows indicate virions and electron dense amorphous vaccine components, respectively. Scale bar represents 200 nm. (B, D) Changes in the hydrodynamic diameter (in nm) of (B) ChAdOx1 nCoV-19 and (D) Ad26.COV2.S vaccine and their respective supernatants and pellets obtained after ultracentrifugation of vaccines before and after addition of 10 µg/mL or 50 µg/mL platelet factor 4 (PF4) assessed by dynamic light scattering. Dissociation of complexes between PF4 and the vaccine component was achieved by the addition of unfractionated heparin (UFH; 10 IU/mL) (E) Western blot analysis of vaccine and pellet or supernatant fraction is shown. Primary antibodies for viral vector (antihexon antibody) or human protein contaminants (anti-HSP90 antibody, anti-Epsilon 14-3-3 antibody, anti-PSMB5 antibody) were used. Statistical analysis was performed by one-way analysis of variance on ranks/Kruskal-Wallis followed by correction for multiple comparisons by two-stage Benjamini, Krieger, & Yekutieli controlling the false discovery rate procedure (n=9) and a<0.05 was considered significant.

954

haematologica | 2022; 107(4)

Comparative analysis of SARS-CoV-2 vector vaccines

Figure 6. Vascular hyperpermeability assay. Five days post-fertilization Tg(fabp10a:gc-eGFP) zebrafish larvae were microinjected with either physiological saline, 100 µM EDTA, ChAdOx1 nCoV-19, purified ChAdOx1 nCoV-19 or Ad26.COV2.S in four adjacent myotomes. Local fluorescence intensities of the myotomes were measured at 0 min and 10 min post-injection (p.i.) (red asterisks) and normalized to the respective intravascular fluorescence (red arrows). Injection of 100 µM EDTA, as well as ChAdOx1 nCoV-19, resulted in a significantly elevated extravascular leakage of the 78 kDa gc-eGFP compared to the saline control (A, B, C, F). However, injection of Ad26.COV2.S, purified ChAdOx1 nCoV-19 or physiological saline did not cause an increase of local vascular permeability (A, D, E, F). The scale bar in (E) represents 200 µm without zoom and 100 µm with zoom.

cofactor is needed. This cofactor is present in the supernatant of the ChAdOx1 nCoV-19 vaccine. However, this does not exclude complex formation of PF4 and Ad26.COV2.S in vivo. Several studies have shown the interaction of different adenoviruses with platelets and on the platelet surface they may also interact with PF4. In addition, hexon proteins or the virions may form complexes with PF4 when they come into contact with additional lymphatic or plasma proteins. Beyond VITT/TTS, the potential for alloantibody formation by protein contaminants in vaccines might be a matter of concern since the PregSure® vaccine, used in cattle, induced alloantibody-driven bovine neonatal pancytopenia, a vaccine-induced alloimmune disease that was observed in young calves of PregSure®-vaccinated cows and is characterized by hemorrhage, pancytopenia, and severe destruction of hematopoietic tissue. The plasma of cows that gave birth to affected calves contained alloantibodies, which were likely induced by alloantigenexpressing protein contaminants of the vaccine from the bovine kidney cell line used for vaccine production.29 Alloantibodies will only cause clinical effects in the case haematologica | 2022; 107(4)

of pregnancy or organ transplantation. Systematic screening of vaccinated individuals excluding or confirming such alloantibodies should be performed to clarify whether this theoretical concern requires further measures. Limitations of our study include its in vitro design. Furthermore, our findings do not exclude the contribution of certain cofactors (e.g., within the interstitial fluid, lymphatic system, plasma, or cell surfaces) to the induction of the anti-PF4 immune response. We also did not investigate involvement of B-cell and T-cell populations. Moreover, since both adenoviral vector-based vaccines have been discontinued in Germany, we cannot compare the immune responses among vaccinated individuals. Finally, the lack of a suitable in vivo model for induction of VITT/TTS limits the ability to reach definite conclusions regarding in vivo mechanisms. In summary, we show that process-related impurities in the form of host-cell derived proteins, active proteases and unassembled hexon proteins differ in quality and quantity between ChAdOx1 nCoV-19 and Ad26.COV2.S SARS-CoV-2 vaccines. EDTA-induced capillary leakage 955

S. Michalik et al.

and host-cell protein impurities might further facilitate induction of an anti-PF4 immune response by intravascular translocation of vaccine constituents and induction of an early inflammatory response after vaccination. These factors might explain the higher incidence rate of VITT/TTS for ChAdOx1 nCoV-19 compared to Ad26.COV2.S vaccines. However, the authors would like to point out again that only comprehensive vaccination of the human population can effectively contain the SARSCoV-2 pandemic. Disclosures AG reports personal fees and non-financial support from Aspen, Boehringer Ingelheim, Instrumentation Laboratory, and Roche; grants from Ergomed, Rovi, Sagent, Portola, Fa. Blau Farmaceutics, Prosensa/Biomarin, DRK-BSD BadenWürtemberg/Hessen, and Biokit; personal fees from Bayer Vital, Chromatec, Sanofi-Aventis, and GTH e.V; grants and personal fees from Macopharma; as well as grants and other from DRKBSD NSTOB, In addition, AG reports having a patent, application n. 2021032220550000DE, pending. Contibutions SM, FS, RP, KF, AR, US, CC, JW, LS, CH, MG-S, MB, NE, AG, and UV conceived the research. SM, FS, RP, KF, AR, US, CC, JW, LS, CH, MG-S, MB, NE, AG, and UV defined the methodology. SM, FS, RP, KF, AR, MS, US, CC, JW, LS, CH, and MG-S. conducted the analysis. SM, FS, RP, KF, AR, US, CC., JW, LS, CH, MG-S, MB, NE, AG, and U.V. interpreted the data. SM, FS, RP, KF, AR, and MS generated ways

References 1. Sadoff J, Gray G, Vandebosch A, et al. Safety and efficacy of single-dose Ad26.COV2.S vaccine against Covid-19. N Engl J Med. 2021;384(23):2187-2201. 2. Voysey M, Clemens SAC, Madhi SA, et al. Safety and efficacy of the ChAdOx1 nCoV19 vaccine (AZD1222) against SARS-CoV-2: an interim analysis of four randomised controlled trials in Brazil, South Africa, and the UK. Lancet. 2021;397(10269):99-111. 3. See I, Lale A, Marquez P, et al. Case series of thrombosis with thrombocytopenia syndrome following COVID-19 vaccination— United States, December 2020–August 2021. Available at: https://www.medrxiv.org/content/10.1101/ 2021.11.10.21266063v1. 4. Chevassut T, Hunt BJ, Pavord S. VITT, COVID-19 and the Expert Haematology Panel: the story of how the UK responded to emerging cases of vaccine-induced immune thrombocytopenia and thrombosis during the vaccination programme. Clin Med (Lond). 2021;21(6):e600-e602. 5. Paul-Ehrlich-Institut, Germany. Sicherheitsbericht: Verdachtsfälle von Nebenwirkungen und Impfkomplikationen nach Impfung zum Schutz vor COVID-19 seit Beginn der Impfkampagne am 27.12.2020 bis zum 30.09.2021 [accessed Dec 6, 2021]. Available at: https://www.pei.de/SharedDocs/Download s/DE/newsroom/dossiers/sicherheitsberichte/sicherheitsbericht-27-12-20-bis-3009-21.pdf?__blob=publicationFile&v=9. 6. Greinacher A, Thiele T, Warkentin TE,

956

to visualize the data. SM, FS, RP, KF, MB, NE, AG, and UV wrote the first draft of the manuscript. All authors contributed to the interpretation of results and manuscript editing. All authors approved the final version of the manuscript. Acknowledgments We thank Katrin Schoknecht for support in the proteomics analyses and Mandy Jörn for graphical design of electron microscopy micrographs. Funding This study was funded by Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) grants: 374031971 - A06 and A11-TRR240, 398967434 - SFB/TR261, A11 - SFB877, P6 - KFO306, B8 - SFB841, and INST 2026/13-1 FUGG, the Ministerium für Wirtschaft, Arbeit und Gesundheit MecklenburgVorpommern (project COVIDPROTECT), “Structure and Function of the Proteasome System in Platelets“ GR2232/8_1 and SE 885/2-1 (DFG), Leibniz WissenschaftsCampus – ComBioCat – W10/2018, by the Federal Ministry of Education and Research (BMBF, grant 01GM1518B, STOP- FSGS), the Südmeyer fund for kidney and vascular research (“SüdmeyerStiftung für Nieren- und Gefäßforschung”), the Dr. Gerhard Büchtemann fund, Hamburg, Germany and the PeNe_C19 study by the Ministerium für Wirtschaft, Arbeit und Gesundheit Mecklenburg-Vorpommern. Data-sharing statement The mass spectrometry proteomics data have been deposited to ProteomeXchange (dataset identifier PXD027344).

Weisser K, Kyrle PA, Eichinger S. Thrombotic thrombocytopenia after ChAdOx1 nCov-19 vaccination. N Engl J Med. 2021;384(22):2092-2101. 7. Scully M, Singh D, Lown R, et al. Pathologic antibodies to platelet factor 4 after ChAdOx1 nCoV-19 vaccination. N Engl J Med. 2021;384(23):2202-2211. 8. Huynh A, Kelton JG, Arnold DM, Daka M, Nazy I. Antibody epitopes in vaccineinduced immune thrombotic thrombocytopaenia. Nature. 2021;596(7873):565-569. 9. Greinacher A, Selleng K, Palankar R, et al. Insights in ChAdOx1 nCoV-19 vaccineinduced immune thrombotic thrombocytopenia. Blood. 2021;138(22):2256-2268. 10. Krutzke L, Roesler R, Wiese S, Kochanek S. Process-related impurities in the ChAdOx1 nCov-19 vaccine [accessed May 4, 2021]. Available at: https://www.researchsquare.com/article/rs477964/v1); DOI: 10.21203/rs.3.rs477964/v1. 11. Nickerson JL, Doucette AA. Rapid and quantitative protein precipitation for proteome analysis by mass spectrometry. J Proteome Res. 2020;19(5):2035-2042. 12. Blankenburg S, Hentschker C, Nagel A, et al. Improving proteome coverage for small sample amounts: an advanced method for proteomics approaches with low bacterial cell numbers. Proteomics. 2019;19(23):e1900192. 13. Perez-Riverol Y, Csordas A, Bai J, et al. The PRIDE database and related tools and resources in 2019: improving support for quantification data. Nucleic Acids Res. 2019;47(D1):D442-D450. 14. Berkers CR, Verdoes M, Lichtman E, et al.

Activity probe for in vivo profiling of the specificity of proteasome inhibitor bortezomib. Nat Methods. 2005;2(5):357-362. 15. Arastu-Kapur S, Anderl JL, Kraus M, et al. Nonproteasomal targets of the proteasome inhibitors bortezomib and carfilzomib: a link to clinical adverse events. Clin Cancer Res. 2011;17(9):2734-2743. 16. Provost A, Rousset C, Bourdon L, et al. Innovative particle standards and long-lived imaging for 2D and 3D dSTORM. Sci Rep. 2019;9(1):17967. 17. Laine RF, Tosheva KL, Gustafsson N, et al. NanoJ: a high-performance open-source super-resolution microscopy toolbox. J Phys D Appl Phys. 2019;52(16):163001. 18. Hoboth P, Šebesta O, Sztacho M, Castano E, Hozák P. Dual-color dSTORM imaging and ThunderSTORM image reconstruction and analysis to study the spatial organization of the nuclear phosphatidylinositol phosphates. MethodsX. 2021;8:101372. 19. Schindelin J, Arganda-Carreras I, Frise E, et al. Fiji: an open-source platform for biologicalimage analysis. Nat Methods. 2012;9(7):676682. 20. Xie J, Farage E, Sugimoto M, Anand-Apte B. A novel transgenic zebrafish model for blood-brain and blood-retinal barrier development. BMC Dev Biol. 2010;10:76. 21. Siegerist F, Zhou W, Endlich K, Endlich N. 4D in vivo imaging of glomerular barrier function in a zebrafish podocyte injury model. Acta Physiol (Oxf). 2017;220(1):167173. 22. Baker AT, Boyd RJ, Sarkar D, et al. ChAdOx1 interacts with CAR and PF4 with implications for thrombosis with thrombocytopenia syndrome. Sci Adv. 2021;7(49):eabl8213.

haematologica | 2022; 107(4)

Comparative analysis of SARS-CoV-2 vector vaccines

23. Eurpean Medicines Agency (EMA). Assessment report: COVID-19 Vaccine AstraZeneca [accessed Dec 6, 2021]. Available at: https://www.ema.europa.eu/en/documents/assessment-report/vaxzevria-previously-covid-19-vaccine-astrazeneca-eparpublic-assessment-report_en.pdf. 24. Nicolai L, Leunig A, Pekayvaz K, et al. Thrombocytopenia and splenic platelet directed immune responses after intravenous ChAdOx1 nCov-19 administration [accessed July 29, 2021]. Available at:

haematologica | 2022; 107(4)

https://www.biorxiv.org/content/10.1101/2 021.06.29.450356v1. 25. Pfueller SL, Logan D, Tran TT, Bilston RA. Naturally occurring human IgG antibodies to intracellular and cytoskeletal components of human platelets. Clin Exp Immunol. 1990;79(3):367-373. 26. Hauler F, Mallery DL, McEwan WA, Bidgood SR, James LC. AAA ATPase p97/VCP is essential for TRIM21-mediated virus neutralization. Proc Natl Acad Sci U S A. 2012;109(48):19733-19738. 27. Greinacher A, Selleng K, Mayerle J, et al.

Anti-platelet factor 4 antibodies causing VITT do not cross-react with SARS-CoV-2 spike protein. Blood. 2021;138(14):1269-1277. 28. Uzun G, Althaus K, Bakchoul T. No correlation between anti-PF4 and anti-SARS-CoV-2 antibodies after ChAdOx1 nCoV-19 vaccination. N Engl J Med. 2021;385(14):13341336. 29. Bastian M, Holsteg M, Hanke-Robinson H, Duchow K, Cussler K. Bovine neonatal pancytopenia: is this alloimmune syndrome caused by vaccine-induced alloreactive antibodies? Vaccine. 2011;29(32):5267-5275.

957

ARTICLE Ferrata Storti Foundation

Haematologica 2022 Volume 107(4):958-965

Red Cell Biology & its Disorders

Brain injury pathophysiology study by a multimodal approach in children with sickle cell anemia with no intra or extra cranial arteriopathy Valentine Brousse,1,2,3,4 Corinne Pondarre,5,6 Manoelle Kossorotoff,7 Cecile Arnaud,5 Annie Kamdem,5 Mariane de Montalembert,1,2 Benedicte Boutonnat-Faucher,1 Slimane Allali,1,2,8 Hélène Bourdeau,9 Keyne Charlot,10 Sebastien Bertil,11 Lydie da Costa,2,9,12,13 Philippe Connes,2,14,15# David Grévent16# and Suzanne Verlhac17,18# 1

AP-HP, Hôpital Necker Enfants Malades, Service de Pédiatrie Générale et Maladies infectieuses, Paris; 2LABEX GR-Ex, Paris; 3Institut National de la Transfusion Sanguine, UMR_S1134, INSERM, Paris; 4AP-HP, Hôpital Universitaire Robert Debré, ImmunoHématologie, Paris; 5Centre Intercommunal de Créteil, Service de Pédiatrie, Créteil; 6 Paris XII University, INSERM U955, Créteil; 7AP-HP, Hôpital Necker Enfants Malades, Service de Neurologie Pédiatrique, Paris; 8Laboratory of Cellular and Molecular Mechanisms of Hematological Disorders and Therapeutical Implications, Paris Descartes – Sorbonne Paris Cité University, Imagine Institute, INSERM U1163, Paris; 9 AP-HP, Hôpital Robert Debré, Service d’Hématologie Biologique, Paris; 10Unité de Physiologie des Exercices et Activités en Conditions Extrêmes, Département Environnements Opérationnels Institut de Recherche Biomédicale des Armées, Brétigny-sur-Orge; 11AP-HP, Service d’Hématologie Biologique, Hôpital Européen Georges Pompidou, Paris; 12Service d'Hématologie Biologique, Université de Paris, Paris; 13 Hematim UR 4666, Amiens; 14Laboratoire LIBM EA7424, Equipe « Biologie Vasculaire et du Globule Rouge », Université Claude Bernard Lyon 1, Villeurbanne; 15Institut Universitaire de France, Paris; 16AP-HP, Hôpital Necker Enfants Malades, Service d’Imagerie Pédiatrique, Paris; 17Centre Hospitalier Intercommunal de Créteil, Service d’Imagerie Médicale, Créteil and 18AP-HP, Hôpital Universitaire Robert Debré, Service d’Imagerie Médicale, Paris, France #

PH, DG and SV contributed equally as co-senior authors

ABSTRACT

Correspondence: VALENTINE BROUSSE valentine.brousse@gmail.com Received: December 21, 2020. Accepted: April 14, 2021. Pre-published: April 22, 2021. https://doi.org/10.3324/haematol.2020.278226

958

espite its high prevalence in children with sickle cell anemia (SCA), the pathophysiology of silent cerebral infarcts (SCI) remains elusive. The main objective of this study was to explore the respective roles of major determinants of brain perfusion in SCA children with no past or current history of intracranial or extracranial vasculopathy. We used a multimodal approach based notably on perfusion imaging arterial spin labeling (ASL) magnetic resonance imaging (MRI) and near infra-red spectroscopy (NIRS), as well as biomarkers reflecting blood rheology and endothelial activation. Out of 59 SCA patients (mean age 11.4±3.9 yrs), eight (13%) had a total of 12 SCI. Children with SCI had a distinctive profile characterized by decreased blood pressure, impaired blood rheology, increased P-selectin levels, and marked anemia. Although ASL perfusion and oximetry values did not differ between groups, comparison of biological and clinical parameters according to the level of perfusion categorized in terciles showed an independent association between high perfusion and increased sP-selectin, decreased red blood cell deformability, low hemoglobin F level, increased blood viscosity and no a-thalassemia deletion. NIRS measurements did not yield additional novel results. Altogether, these findings argue for early MRI detection of SCI in children with no identified vasculopathy and suggest a potential role for ASL as an additional screening tool. Early treatment targeting hemolysis, anemia and endothelial dysfunction should reduce the risk of this under diagnosed and serious complication.

haematologica | 2022; 107(4)

Multimodal brain study in children with SCA

Introduction Brain injury is a major complication in sickle cell anemia (SCA) patients because of its high prevalence and devastating consequences. Historically, overt stroke used to be the most distressing clinical complication in childhood but its prevalence has dropped from 10% to 1-3% in high-income countries since the implementation of a preventive screening strategy by transcranial Doppler (TCD) coupled with chronic transfusion therapy in case of abnormal results.1,2,3,4 In sharp contrast, the contribution of silent cerebral infarcts (SCI) to global neurovascular burden has emerged, following the improved ability to detect subclinical lesions by magnetic resonance imaging (MRI). SCI are found in up to 37% of SCA children before the age of 14 years, can be detected as early as 1 year old and their prevalence increases with age.3,5,6 In addition, SCI are frequently associated with neurocognitive impairment.7,8 While overt ischemic stroke mainly occurs in the arterial territory of internal carotid or middle cerebral artery (usually with an associated stenosis or occlusion), silent infarcts tend to occur in the border zone areas of the brain, mainly in the deep white matter, including in patients without large vessel arteriopathy. This observation highly suggests that SCI may be related to alterations of perfusion and oxygenation.9 However, no clear pathophysiological process in SCI genesis has been demonstrated thus far, although anemia, both chronic and acute, seems to be a significant contributor.6 Conventional MRI shows brain injury in a pattern suggestive of hemodynamic compromise but it does not directly assess cerebral hemodynamics. By contrast, perfusion analysis by arterial spin labeling (ASL) allows to better evaluate the cerebral hemodynamic risk at a regional level, in a non-invasive manner, and may serve as a screening tool to identify children at risk of SCI. Blood rheology is a key determinant of blood flow, vascular resistance and tissue perfusion.10 Increased blood viscosity in both adults and children with SCA has been associated with a risk of frequent vaso-occlusive crises. Moreover, SCA individuals with the greatest reduction in red blood cell (RBC) deformability would be at higher risk for developing leg ulcers, glomerulopathy and priapism.11,12,13 These studies also reported increased RBC aggregates strength in SCA patients with glomerulopathy or frequent priapism. Although a contribution of blood rheology has been speculated in overt stroke in SCA,14 there is no study focusing on SCI and blood rheology in this disease in the literature. Surprisingly, a study by Brown et al.15 performed in patients with paraproteinemia or leukemia observed no association between cerebral blood flow and blood viscosity. Instead, the authors reported an inverse relationship between cerebral blood flow and arterial oxygen content, showing that regulatory mechanisms can maintain cerebral oxygen transport despite increased blood viscosity. The same negative association has been reported in children with SCA.16 However, the associations between brain oxygenation, cerebral blood flow in the different brain areas and blood rheology have never been investigated, and more particularly in the context of SCI. In order to better decipher the mechanisms involved in subclinical cerebral injury in children with SCA, we sought to explore the respective roles of hemolysis, abnormal blood rheology, cerebral oxygen supply and brain perfusion characteristics in children with no past or current history of intracranial or extracranial vasculopathy. We used a multimodal approach in a multi-centric pilot study where we i) haematologica | 2022; 107(4)

compared several biomarkers (markers of endothelial injury and hemorheology), level of brain oxygenation (near infra-red spectroscopy [NIRS]) and perfusion imaging (TCD and arterial spin labeling [ASL] MRI) between children with newly discovered and without SCI and ii) tested the association between these parameters.

Methods Consecutive patients from two French referral centers for sickle cell disease (University Hospital Necker-Enfants Malades and Centre Hospitalier Intercommunal de Creteil) were screened during routine visits according to the following inclusion criteria: i) SS or S-b°thalassemia genotype; ii) no past history of abnormal or conditional transcranial and extracranial TCD; iii) at steady state (> 3 months from any vaso-occlusive event, transfusion or infection); iv) age between 5 and 17 years. The study was offered to all children meeting the inclusion criteria and regularly followed up since neonatal screening in the participating centers, between March 2015 and July 2016, with a target of 60 children (based on the expected number of SCA patients meeting the inclusion criteria in these centers and the pilot design of the study). Treatment by hydroxyurea was not an exclusion criterion, but chronic transfusion was. After written informed consent was obtained, a visit during which all investigations were performed was scheduled. The protocol was approved by the ethics committee “Comité pour la Protection des Personnes Ile de France III” (014-A01575-42) and registered in clinicaltrials gov. Identifier: NCT 02909283. Clinical parameters as well as relevant past medical events were collected and are available in the Online Supplementary Appendix. Biological parameters were measured for each patient: routine hematological and blood rheological parameters, and markers of endothelial activation (see the Online Supplementary Appendix). TCD imaging (TCDI) was performed using a LOGIQ E9 XDclear 2.0 ultrasound system (GE Healthcare, Milwaukee, WI, USA) at Necker – Enfants Malades and an Acuson S 2000 ultrasound system (Siemens Healthineers, Erlangen, Germany) at Centre Hospitalier Intercommunal de Creteil. MR imaging was performed with a GE Signa HDxt 1.5-T system (GE Medical Systems, Milwaukee, WI, USA) and a 12-channel head-neck-spine coil in non-sedated children. SCI were defined on MRI as hypersignals of cerebral parenchyma of at least 3 mm on T2 FLAIR sequence. Details on MRI sequences and image analysis including calculation of hypersignal volume are available in the Online Supplement Appendix. Cerebral blood flow was measured using a three-dimensional pseudo-continuous arterial spin labeling (pCASL) sequence (repetition time msec/echo time msec, 4,428/10.5; postlabeling delay =1,025 msec; 80 axial partitions; field of view, 240x240x4 mm; acquisition matrix, eight spiral arms in each three-dimensional partition with 512 points per arm; flip angle, 155°; acquisition time, 4 minutes 17 seconds). Transcranial NIRS was performed using a dual-channel absolute cerebral oximeter FORE-SIGHT® (Branford, CT, USA).

Statistics We first compared the different measured parameters in patients with and without SCI using an unpaired student t test. The cohort was divided into terciles of ASL values: hypo-perfusion (Low), normal perfusion (Middle) and hyper-perfusion (High) groups. the three groups were then compared using a one-way ANOVA with Tukey post hoc test. A c2 test was used for the qualitative analyses. The significance level was defined as P<0.05 (SPSS, v. 20, IBM SPSS Statistics, Chicago, IL).

959

V. Brousse et al.

Results General characteristics of the population Fifty-nine SCA patients (mean age 11.4±3.9 yrs) at steady state were enrolled, of which 34 (57.6%) were treated with hydroxyurea (HU). Patients main clinical and biological characteristics are summarized in Table 1.

Transcranial Doppler and magnetic resonance imaging analysis At the time of assessment, all TCD velocities were within normal ranges (<170 cm/s) and there were no arte-

rial abnormalities on MRA analysis. Twelve SCI (3-5 mm: n= 8; 5-15 mm: n=4, >15 mm: n=0) were observed in eight of 59 (13%) children. The median volume of SCI was 41,57 mm (range, 15,8-362 mm). SCI were located in the right anterior border zone in six patients. Five patients had a single lesion, while three patients had >1 lesion, including two with bilateral lesions.

Analysis of patients with silent cerebral infarcts Children with SCI had a lower past vaso-occlusive crisis (VOC) rate compared to children with no SCI (Table 1). Despite a comparable frequency of HU treatment in both

Table 1. General characteristics of the population. Sex (M/F) Age (yrs) Ongoing HU ttt at inclusion (N; %) Time since HU initiation (yrs) Daily dose (mg/kg/day) mean (SD) Age at beginning of HU initiation (yrs) Alpha thalassemia (N; %) G6PD deficiency (N; %) VOC rate (events/yr) ACS rate (events/yr) Transfusion history (%) Mean number of transfusions (n, SD) SpO2 (%) DBP (mmHg) SBP (mmHg) MAP (mmHg) Pulse Pressure (mmHg) Hb (g/dL) Hct (%) MCV (fl) MCHC (g/dL) Reticulocyte count (109/L) Platelets (109/L) WBC count (109/L) HbF (%) ASAT UI/L LDH UI/L Soluble E-selectin (ng/mL) Soluble P-selectin (ng/mL) CD34* cell (/mL) RBC aggregation (%) RBC disaggregation threshold (s-1) RBC deformability (a.u.) Blood viscosity (cP) Hematocrit/blood viscosity (a.u.) Right ctSO2 (%) Left ctSO2 (%)

All (n=59)

no-SCI (n = 51)

SCI (n = 8)

20/39 11.4 ± 3.9 33; 56 3.1 ± 3.1 23 ± 4.6 8.3 ± 4.3 24; 41 4; 7 0.58 ± 0.58 0.08 ± 0.10 23.7 % 8 (11) 99.0 ± 1.1 68 ± 11 110 ± 11 82 ± 9 42 ± 12 8.8 ± 1.2 25.0 ± 3.3 82 ± 12 35.0 ± 1.5 225 ± 102 343 ± 154 8.98 ± 3.61 16.0 ± 10.1 49 ± 16 484 ± 164 73.4 ± 29.6 59.3 ± 21.3 9185 ± 6738 52 ± 6 378 ± 201 0.49 ± 0.09 5.35 ± 0.93 4.72 ± 0.59 63 ± 10 63 ± 10

16/35 11.4 ± 3.7 30; 59 3.2 ± 3.2 24.1 ± 3.3 8.2 ± 3.9 23; 46 4; 8 0.65 ± 0.60 0.09 ± 0.10 23% 8 (10) 99.1 ± 1.0 69 ± 11 111 ± 11 83 ± 9 42 ± 12 8.9 ± 1.1 25.3 ± 3.2 83 ± 12 34.9 ± 1.5 224 ± 105 331 ± 141 8.78 ± 3.64 16.0 ± 10.4 48 ± 16 485 ± 173 73.2 ± 21.7 57.1 ± 20.7 9803 ± 6773 52 ± 7 359 ± 191 0.50 ± 0.09 5.40 ± 0.94 4.73 ± 0.58 63 ± 10 63 ± 10

4/4 11.6 ± 4.9 3; 38 1.6 ± 1.3 12,8 ± 0.5 9.0 ± 5.4 1; 13 0; 0 0.17 ± 0.20 0.03 ± 0.05 25% 12 (17) 98.6 ± 2.0 61 ± 7 106 ± 10 76 ± 8 45 ± 6 8.2 ± 1.1 23.0 ± 3.0 77 ± 10 35.8 ± 0.9 229 ± 88 418 ± 220 10.29 ± 3.36 16.1 ± 8.2 52 ± 13 479 ± 108 74.5 ± 39.7 73.5 ± 21.7 5788 ± 6229 51 ± 4 491 ± 236 0.44 ± 0.07 5.04 ± 0.83 4.62 ± 0.69 61 ± 9 64 ± 9

0.301 0.865 0.259 0.170 0.001 0.170 0.074 0.412 0.001 0.157 0.96 0.16 0.599 0.047 0.283 0.047 0.437 0.047 0.025 0.186 0.037 0.91 0.141 0.277 0.986 0.548 0.938 0.911 0.047 0.301 0.811 0.037 0.025 0.312 0.627 0.662 0.828

M: male; F: female; SD: standard deviation; SCI: silent cerebral infarct; HU: hydroxyurea; VOC: vaso-occlusive crisis; ACS: acute chest syndrome; DBP: diastolic blood pressure; SBP: systolic blood pressure; MAP: mean arterial pressure; MCHC: mean corpuscular hemoglobin concentration; MCV: mean corpuscular volume; Hct:hemtocrit; HbF: hemglobin F; ASAT: aspartate aminotransferase; LDH: lactate dehydrogenase; RBC: red blood cell; a.u.: artbitrary units; CP: centipoise; ctSO2: cerebral tissue hemoglobin oxygen saturation; P-values are given for the comparison between the non-SCI and SCI groups.

960

haematologica | 2022; 107(4)

Multimodal brain study in children with SCA

Table 2. Cerebral blood flow values in patients with or without silent cerebral infarct. Right hemisphere perfusion (mL/100 g/min) ACA Anterior Junctional Superficial MCA Caudate nucleus Putamen posterior Junctional Left hemisphere perfusion (mL/100 g/min) ACA Anterior Junctional Superficial MCA Caudate nucleus Putamen Posterior Junctional Posterior fossa perfusion Right Cerebellar hemisphere Left Cerebellar hemisphere

no-SCI (n = 51)

SCI (n = 8)

116.5 ± 17.8 (112.0-121.0) 122.9 ± 21.1 (117.0-129.0) 111.9 ± 17.9 (107.0-117.0) 93.3 ± 16.2 (88.8-97.8) 82.5 ± 14.2 (78.6-86.4) 120.6 ± 23.3 (114.0-127.0)

126.1 ± 17.4 (114.0-138.0) 123.8 ± 16.2 (113.0-135.0) 116.5 ± 13.8 (107.0-126.0) 98.5 ± 16.4 (87.1-110.0) 90.0 ± 13.5 (80.7-99.3) 120.5 ± 16.9 (109.0-132.0)

0.162 0.909 0.489 0.406 0.168 0.988

116.5 ± 19.4 (111.0-122.0) 121.6 ± 22.1 (116.0-128.0) 111.3 ± 18.9 (106.0-116.0) 94.3 ± 16.0 (89.9-98.7) 83.0 ± 14.4 (79.0-87.0) 119.9 ± 26.4 (113.0-127.0)

132.0 ± 23.9 (115.0-149.0) 122.8 ± 15.5 (112.0-134.0) 121.0 ± 16.1 (110.0-132.0) 102.8 ± 14.2 (93.0-113.0) 93.4 ± 11.0 (85.8-101.0) 122.5 ± 24.3 (106.0-139.0)

0.117 0.887 0.174 0.165 0.058 0.794

88.1 ± 8.2 (85.8-90.3) 89.6 ± 16.9 (85.0-94.2)

84.1 ± 8.2 (78.4-89.8) 85.4 ± 8.1 (79.8-91.0)

0.295 0.277

No significant difference in cerebral blood flow values was evidenced between the two groups. Values are given as means ± standard deviation (95% Confidence Interval): SCI: silent cerebral infarcts; MCA: middle cerebral artery; ACA: anterior cerebral artery.

Table 3. Characteristics of the population according to the level of perfusion assessed by arterial spin labeling in the right anterior cerebral artery and categorized in terciles. Age (yrs) HU (%) a thalassemia (%) G6PD deficiency (%) VOC rate (events/yr) ACS rate (events/yr) SpO2 (%) DBP (mmHg) SBP (mmHg) MAP (mmHg) Hb (g/dL) Hct (%) MCV (fL) MCHC (g/dL) Reticulocyte count (109/L) Platelets (109/L) White blood cells (109/L) HbF (%) ASAT UI/L LDH UI/L CRP mg/L E-selectin P-selectin RBC aggregation (%) RBC disaggregation threshold (s-1) RBC deformability (a.u.) Blood viscosity (cP) Hematocrit/blood viscosity (a.u.) Right ctSO2 (%) Left ctSO2 (%)

Lower CBF (N=20)

Middle CBF (N = 19)

Higher CBF (N = 20)

12.4 ± 4.2 71 46 31 0.68 ± 0.53 0.09 ± 0.11 99.3 ± 1.0 70 ± 12 114 ± 12 85 ± 10 9.2 ± 1.2 26.3 ± 3.4 82 ± 13 35.0 ± 1.3 193 ± 84 315 ± 123 7.99 ± 3.37 19.4 ± 11.3 42 ± 8 407 ± 126 5.8 ± 8.4 67.0 ± 22.3 49.3 ± 14.7 52 ± 6 390 ± 222 0.53 ± 0.06 5.75 ± 1.03 4.66 ± 0.70 64 ± 10 64 ± 13

10.5 ± 4.2 56 29 35 0.47 ± 0.43 0.06 ± 0.08 99.1 ± 1.2 67 ± 11 108 ± 12 81 ± 9 8.9 ± 1.1 25.2 ± 3.3 81 ± 11 35.2 ± 1.0 229 ± 112 314 ± 141 9.23 ± 3.89 18.4 ± 10.2 50 ± 15 515 ± 154 6.0 ± 5.0 68.5 ± 35.0 58.2 ± 21.3 53 ± 7 399 ± 225 0.51 ± 0.09 5.23 ± 0.80 4.80 ± 0.40 62 ± 9 64 ± 8

11.2 ± 3.2 43 25* 35 0.63 ± 0.73 0.09 ± 0.11 98.8 ± 1.1 66 ± 10 108 ± 10 80 ± 9 8.3 ± 1.0* 23.6 ± 2.6* 84 ± 11 34.9 ± 2.0 255 ± 105* 388 ± 159 9.91 ± 3.51 11.9 ± 6.5*# 55 ± 18* 545 ± 186 5.4 ± 9.2 85.0 ± 28.4 68.9 ± 23.1** 50 ± 5 356 ± 174 0.46 ± 0.08*# 5.18 ± 0.93* 4.67 ± 0.67 63 ± 10 64 ± 10

a.u.: arbitray unit; G6PD: glucose-6-phosphate dehydrogenase; HU: hydryurea; VOC: vaso-occlusive crisis; ACS: acute chest syndrome; DBP: diastolic blood pressure; SBP: systolic blood pressure; MAP: mean arterial pressure; Hb: hemoglobin; Hct: hematocrit; MCV: mean corpuscular volume; MCHC: mean corpuscular hemoglobin concentration; ASAT: aspartate aminotransferase; LDH: lactate dehydrogenase; RBC: red blood cell; ctSO2: cerebral tissue hemoglobin oxygen saturation. Significantly different from the lower cerebral blood flow (CBF) group: *P<0.05; **P<0.01. Significantly different from the Middle CBF group: #P<0.05.

haematologica | 2022; 107(4)

961

V. Brousse et al.

groups, significant differences in biological profiles were evidenced: children with SCI had increased plasma P-selectin level and RBC disaggregation threshold, and lower hemoglobin (Hb) level and RBC deformability. Of note, daily dose of HU was significantly lower in patients with SCI compared to those with no SCI at the time of data collection but the sample size was too small (n=3 of 8) to further interpret this finding. In addition, diastolic blood pressure (DBP) and mean arterial pressure (MAP) were lower in patients with SCI. Oximetry data (right and left cerebral oxygenation levels) showed no significant difference between patients with

and without SCI. Of note, there was no association between cerebral blood flow (CBF) values and the presence of SCI (Table 2).

Arterial spin labeling perfusion imaging In the whole group of patients, CBF values showed no major asymmetry and correlated across all territories (r ranging from 0.46 to 0.91; P ranging from <0.01 to <0.001). Interestingly, comparison of patient subgroups according to CBF terciles in the arterial territories showed marked differences (see Table 3 for the right anterior cerebral artery territory, as an illustration). The group with

Figure 1. Regions of interest on the arterial spin labeling sequence. Left column: T1 weighted imaging. Right column: arterial spin labeling (ASL) imaging. First row: axial section tangential to the upper wall of the lateral ventricles, second row: axial section crossing the anterior and posterior white commissures (AC-PC), third row: axial section at the level of the internal auditory canals. All axial images are parallel to CA-CP. ACA: anterior cerebral artery area; AW: anterior watershed area; MCA: middle cerebral artery area; PW: posterior watershed area; C: caudate nuclei; LN: lenticulostriate nuclei; PF: posterior fossa.

962

haematologica | 2022; 107(4)

Multimodal brain study in children with SCA

higher CBF had increased markers of hemolytic anemia (lower Hb, increased aspartate aminotransferase (ASAT), lactate dehydrogenase [LDH] and reticulocyte count), lower HbF level, increased marker of endothelial activation (sP-selectin value), lower RBC deformability and blood viscosity compared to the group with lower CBF. In addition, the frequency of patients with co inherited αthalassemia was lower. An ordinal multivariate analyses was performed to test the independent associations between the perfusion level (tercile groups) and the main parameters influencing blood flow. Blood viscosity was retained for the model because it is a key determinant of blood flow.10 Since blood viscosity is highly dependent on hematocrit/Hb and colinearity was indeed very strong between blood viscosity and hematocrit and Hb (variance inflation factor [VIF] = 36.6 and 37.4, respectively), hematocrit and Hb were not included in the multivariate model. Likewise, ASAT and reticulocyte count were not considered for the model as, like LDH, they reflect hemolysis. In the multivariate model, which was highly s i g n i f i c a n t (P<0.001), all the parameters except LDH (P=0.058), were independently associated with the level of perfusion (sP-selectin: P<0.01; RBC deformability: P<0.01; HbF level: P<0.05; blood viscosity: P<0.05; a-thalassemia: P<0.05).

Oximetry measurements In the whole population, cerebral oximetry results showed a significant positive correlation with Hb, SpO2, RBC deformability, blood viscosity and HbF and a negative correlation with markers of hemolysis (total bilirubin and LDH) as well as the RBC disaggregation threshold (Table 4). Of note, blood pressure, age, reticulocyte count, level of selectins and TCD velocities did not correlate with oximetry results. A multivariate analysis was performed for the right and left cerebral tissue hemoglobin oxygen saturation including SpO2, Hb, RBC deformability, HbF, LDH and the RBC disaggregation threshold. Hb was preferred to blood viscosity as tissue oxygenation is thought to be highly dependent on the number of RBC and Hb concentration. None of the parameters was significant (P=0.38 and P=0.35 for the right and left ctSO2, respectively), which may be explained by the fact that all are interrelated to affect brain oxygenation.

Effect of hydroxyurea treatment HU treatment did not significantly impact the results of neurovascular explorations in this cohort, except for sEand sP-selectin levels that were significantly lower in children receiving treatment. Treated children received a daily dose of HU that was not a maximum tolerated dose but was a fairly high dose (23+/-4.6 mg/kg/day) and both groups had indeed comparable levels of Hb or HbF. Although compliance was not specifically addressed, the decreased level of selectins in treated children argues for an effect of HU on alleviating endothelial injury, in addition to its known effect on hemolytic anemia. The crosssectional design of the study, however, does not allow to actually compare groups at baseline, i.e., before treatment initiation and precludes further interpretation. In line, the specific effect of HU in children with SCI (or the lack of, given the significantly low dose) is limited by the small sample size (n=3).

haematologica | 2022; 107(4)

Table 4. Correlations between cerebral oxygenation (ctSO2) and biological parameters

Left ctSO2 (%) Right ctSO2 (%) Age (yrs) SpO2 (%) DBP (mmHg) SBP (mmHg) MAP (mmHg) Hb (g/dL) Hct (%) MCV (fL) MCHC (g/dL) Reticulocyte count (109/L) Platelets (109/L) White blood cells (109/L) HbF (%) ASAT UI/L LDH UI/L CRP mg/L E-selectin P-selectin RBC aggregation (%) RBC disaggregation threshold (s-1) RBC deformability (a.u.) Blood viscosity (cP) Hematocrit/blood viscosity (a.u.)

r = -0.12 r = 0.35* r = -0.08 r = -0.09 r = -0.09 r = 0.48*** r = 0.53*** r = 0.07 r = -0.26 r = -0.06 r = -0.17 r = -0.26 r = 0.44** r = -0.37* r = -0.51*** r = -0.07 r = -0.16 r = 0.12 r = 0.03 r = -0.30* r = 0.54*** r = 0.30* R = 0.21

r = -0.22 r = 0.43** r = 0.14 r = -0.08 r = 0.08 r = 0.53*** r = 0.56*** r = 0.17 r = -0.30* r = -0.12 r = -0.14 r = -0.24 r = 0.45** r = -0.24 r = -0.43** r = -0.24 r = -0.20 r = 0.07 r = -0.10 r = -0.27* r = 0.51*** r = 0.30* R = 0.17

DBP: diastolic blood pressure; SBP: systolic blood pressure; Hb: hemoglobin; Hct: hemtocrit; MCV: mean corpuscular volume; MCHC: mean corpuscular hemoglobin concentration; ASAT: aspartate aminotransferase; LDH: lactate dehydrogenase; CRP: C-reactive protein; a.u.: arbitrary units; RBC: red blood cell; ctSO2: cerebral tissue hemoglobin oxygen saturation Significant correlations: *P< 0.05; **P< 0.01; ***P< 0.001.

Discussion In this study of 59 young children highly selected for no present or past history of cerebral and extra cranial vasculopathy, a prevalence of eight of 59 (13%) children with SCI was found, illustrating the residual burden of neurovascular injury in patients having received recommended follow-up. SCI predominated in the fronto-parietal white matter at the junction between the anterior and the middle cerebral artery territories, as previously described.17 Expectedly, on T2 FLAIR sequence, no SCI >15 mm were found which would have presumably translated in clinically detectable symptoms. Likewise, volumes of SCI were within previously described ranges.18 Consistent with the inclusion criteria of no past conditional or abnormal TCD or transient vasculopathy, there was no arterial abnormality on MRA analysis. SCI and more specifically their size, volume and localization may impact cognitive function. In this study cognitive testing was not performed and subsequently cognitive consequences related to SCI cannot be inferred. Nevertheless, the frequency of SCI in these highly selected children plead for early identification of silent lesions in order to further explore cognitive functions for early implementation of supportive learning skills. In line with recent recommendations, MRI screening, is therefore highly recommended in all children with SCA, including in those with no identified vasculopathy.19

963

V. Brousse et al.

Given the absence of large vessel vasculopathy in these children, SCI are presumably unrelated to ischemia occurring downstream of a large vessel stenosis. SCI may nevertheless share common risk factors with large vessel vasculopathy and result from impaired perfusion. Indeed, we found distinctive clinical and biological features in children with SCI. In line with previous reports, a markedly increased hemolytic and anemic profile favoring SCI was evidenced, consistent with a lower pain rate.20,21 HU treatment did not significantly impact these data, but interpretation is very limited given the small sample size of treated children with SCI (n=3), and the low daily dose at the time of data collection. Interestingly, despite significantly greater anemia in children with SCI, blood viscosity did not significantly differ across groups because RBC deformability was reduced in the former group, which exerts opposite effects on blood viscosity.22 Intravascular hemolysis is a major determinant of vascular and endothelial dysfunction.23 Consistent with this, an increased level of sP-selectin, a marker of endothelial activation, was evidenced and was associated with SCI. sP-selectin could potentially serve as a biomarker of cerebral injury in children.6 In contrast with previous studies,24 a lower diastolic and mean arterial pressure was found in children with SCI. Altogether, this data strengthens the hemodynamic pathophysiology of SCI, whereby a further drop in pressure and/or blood flow in children with severe anemia results in ischemia in the border zone region, characterized by terminal arterial supply. Furthermore, hemorheological exploration of these patients suggests that increased RBC aggregate strength and decreased red cell deformability may also influence the risk of SCI. Deformable RBC mostly flow in single file in capillaries and RBC aggregates need to be fully dispersed before entering into the microcirculation.22,10 Consequently, decreased RBC deformability and increased RBC aggregates strength may both impair blood flow at the entry of small capillaries and affect tissue perfusion. Cerebral blood flow begins at a low level in the perinatal period, increases to a peak value at 3-8 years of age and then gradually decreases to adult levels with a negative correlation with age.25 Several reports have documented elevated CBF in the grey matter of children with SCA, compared to normal controls,26-28 a finding attributed to the compensatory increase in blood velocity and flow secondary to baseline anemia. This study allowed further insight by showing that, regardless of concomitant HU treatment, the level of perfusion was independently associated with blood viscosity (which is negatively correlated to anemia), HbF level, endothelial activation, a-thalassemia status, and marginally associated with the level of hemolysis. ASL perfusion may therefore serve as an additional screening tool that integrates all these parameters to identify children at risk of subclinical injury, beyond the known risk factor, and regardless of cerebral vasculopathy. Thresholds for risk assessment will need to be determined by further prospective studies. Oximetry analysis yielded expected and coherent results. NIRS is a non-invasive measurement of cerebral oxygenation, that varies with SpO2 and Hb. In line with previous reports, we show that cerebral oxygenation is low in SCA patients.29,30,31 Our results also confirmed an association between RBC rheological parameters and cerebral oxygenation, suggesting that patients with the most deformable RBC and less robust RBC aggregates 964

would have increased cerebral perfusion and oxygenation. However, because NIRS is unable to measure cerebral oxygenation at a depth of interest and is limited to the anterior territories, its additional predictive value remains to be demonstrated, particularly in patients at risk of white matter SCI. This study has a number of limitations. Its cross-sectional design does not allow associations to be interpreted as causalities and its sample size was small, particularly regarding the number of patients with SCI treated by HU. It is possible that the lack of association between elevated CBF values and the presence of SCI was due to the low power of the study, for instance. Regarding imaging techniques, ASL MRI has also inherent limitations. In particular the transit time is reduced in SCA patients relative to non-anemic children. Thus, the labeling efficiency and postlabeling delay may not be adequate in every patient and may modify the perfusion signal measured by ASL.28 Another parameter, the fixed value of the T1-longitudinal relaxation time of blood used for CBF quantification may not be adequate in every individual patient, given its dependence on hematocrit and blood composition. Another limitation is that CBF measurements were made in grey matter, while silent infarcts are located preferentially in white matter. At a magnetic field strength of 1.5 Tesla, the signal-to-noise ratio of the resulting perfusion map is too low in the cerebral white matter to allow accurate measurements. Despite these limitations, however, our ASL results adequately reflected the level of perfusion in different territories of the brain. In fact, our results were consistent with previous reports regarding the influence of both anemia and age and further allowed novel coherent results regarding the influence of endothelial activation and blood rheology for instance, and more generally the possible pathophysiology of SCI genesis. Altogether the findings of this study suggest that SCI may, like overt stroke, preferably occur in otherwise paucisymptomatic children with marked hemolytic anemia and endothelial dysfunction. Cerebral blood flow measurements by ASL MRI may help assess the quality of perfusion at a microvascular level in children with no vasculopathy but nevertheless at risk of subclinical injury. An early disease-modifying treatment like HU, which improves all afore-mentioned factors associated with SCI may therefore decrease SCI risk as well, in addition to its known beneficial effect on TCD velocities and stroke risk.27,32,33 Decreasing hemolytic anemia, improving RBC rheological characteristics and limiting endothelial injury will help avoid cerebral injury, particularly in case of further aggression such as arterial stenosis, acute hypoxia, drop in Hb or increased metabolic demand. Large prospective trials evaluating the protective effect of HU on SCI are ongoing and will hopefully confirm such beneficial effect. It is possible that new therapeutic approaches such as P-selectin blockade by monoclonal or pan antibodies, may have further protective effects in this particular complication. Disclosures VB, CP and MdM report honoraria and expert/consultancy testimony for Addmedica. Contributions VB, MK, DG and SV designed the study; VB, CP, CA, AK, MdM, BBF and SA enrolled patients; HB and LdC were responsible for biological data, except for E- and P-selectins that were measured by SB; SV and DG were responsible for imaging haematologica | 2022; 107(4)

Multimodal brain study in children with SCA

data; KC and PC were responsible for oximetry and hemorheological data interpretation; VB, PC, CP, MK, DG, and SV analyzed the data; PC was responsible for statistical analysis; VB and PC wrote the manuscript. All authors reviewed, edited and approved the manuscript. Acknowledgments We wish to thank Wiam BHIA, Nicholas Renaud, Elisabeth

References 1. Ohene-Frempong K, Weiner SJ, Sleeper LA, et al. Cerebrovascular accidents in sickle cell disease: rates and risk factors. Blood. 1998;91(1):288-294. 2. Brousse V, Arnaud C, Lesprit E, et al. Evaluation of outcomes and quality of care in children with sickle cell disease diagnosed by newborn screening: a real-world nation-wide study in France. J Clin Med. 2019;8(10):1594. 3. Bernaudin F, Verlhac S, Arnaud C, et al. Impact of early transcranial Doppler screening and intensive therapy on cerebral vasculopathy outcome in a newborn sickle cell anemia cohort. Blood. 2011;117(4): 1130-1140. 4. Adams RJ, McKie VC, Hsu L, et al. Prevention of a first stroke by transfusions in children with sickle cell anemia and abnormal results on transcranial Doppler ultrasonography. N Engl J Med. 1998; 339(1):5-11. 5. McCarville MB, Goodin GS, Fortner G, et al. Evaluation of a comprehensive transcranial doppler screening program for children with sickle cell anemia. Pediatr Blood Cancer. 2008;50(4):818-821. 6. DeBaun MR, Kirkham FJ. Central nervous system complications and management in sickle cell disease. Blood. 2016;127(7):829838. 7. Bernaudin F, Verlhac S, Freard F, et al. Multicenter prospective study of children with sickle cell disease: radiographic and psychometric correlation. J Child Neurol. 2000;15(5):333-343. 8. DeBaun MR, Schatz J, Siegel MJ, et al. Cognitive screening examinations for silent cerebral infarcts in sickle cell disease. Neurology. 1998;50(6):1678-1682. 9. Fields ME, Guilliams KP, Ragan DK, et al. Regional oxygen extraction predicts border zone vulnerability to stroke in sickle cell disease. Neurology. 2018;90(13):e1134e1142. 10. Nader E, Skinner S, Romana M, et al. Blood rheology: key parameters, impact on blood flow, role in sickle cell disease and effects of exercise. Front Physiol. 2019;10:1329. 11. Lamarre Y, Romana M, Lemonne N, et al. Alpha thalassemia protects sickle cell anemia patients from macro-albuminuria through its effects on red blood cell rheological properties. Clin Hemorheol Microcirc. 2014;57(1):63-72.

haematologica | 2022; 107(4)

Hullier-Amar and Isabelle Buffet from the Fondation Institut Imagine. Funding This research was funded by LABEX-GR-Ex, grant number GR-EX14/ParisDiderot/ and an additional grant was obtained from AddMeddica.

12. Connes P, Lamarre Y, Hardy-Dessources M-D, et al. Decreased hematocrit-to-viscosity ratio and increased lactate dehydrogenase level in patients with sickle cell anemia and recurrent leg ulcers. PLoS One. 2013;8(11):e79680. 13. Cita K-C, Brureau L, Lemonne N, et al. Men with sickle cell anemia and priapism exhibit increased hemolytic rate, decreased red blood cell deformability and increased red blood cell aggregate strength. PLoS One. 2016;11(5):e0154866. 14. Connes P, Verlhac S, Bernaudin F. Advances in understanding the pathogenesis of cerebrovascular vasculopathy in sickle cell anaemia. Br J Haematol. 2013;161(4):484498. 15. Brown MM, Marshall J. Regulation of cerebral blood flow in response to changes in blood viscosity. Lancet. 1985;1(8429):604609. 16. Bush AM, Borzage MT, Choi S, et al. Determinants of resting cerebral blood flow in sickle cell disease. Am J Hematol. 2016;91(9):912-917. 17. Ford AL, Ragan DK, Fellah S, et al. Silent infarcts in sickle cell disease occur in the border zone region and are associated with low cerebral blood flow. Blood. 2018; 132(16):1714-1723. 18. van der Land V, Hijmans CT, de Ruiter M, et al. Volume of white matter hyperintensities is an independent predictor of intelligence quotient and processing speed in children with sickle cell disease. Br J Haematol. 2015;168(4):553-556. 19. DeBaun MR, Jordan LC, King AA, et al. American Society of Hematology 2020 guidelines for sickle cell disease: prevention, diagnosis, and treatment of cerebrovascular disease in children and adults. Blood Adv. 2020;4(8):1554-1588. 20. Bernaudin F, Verlhac S, Arnaud C, et al. Chronic and acute anemia and extracranial internal carotid stenosis are risk factors for silent cerebral infarcts in sickle cell anemia. Blood. 2015;125(10):1653-1661. 21. Kinney TR, Sleeper LA, Wang WC, et al. Silent cerebral infarcts in sickle cell anemia: a risk factor analysis. The Cooperative Study of Sickle Cell Disease. Pediatrics. 1999;103(3):640-645. 22. Baskurt OK, Meiselman HJ. Blood rheology and hemodynamics. Semin Thromb Hemost. 2003;29(5):435-450. 23. Sundd P, Gladwin MT, Novelli EM. Pathophysiology of sickle cell disease.

Annu Rev Pathol. 2019;14:263-292. 24. DeBaun MR, Sarnaik SA, Rodeghier MJ, et al. Associated risk factors for silent cerebral infarcts in sickle cell anemia: low baseline hemoglobin, sex, and relative high systolic blood pressure. Blood. 2012;119(16):36843690. 25. Petcharunpaisan S, Ramalho J, Castillo M. Arterial spin labeling in neuroimaging. World J Radiol. 2010;2(10):384-398. 26. Oguz KK, Golay X, Pizzini FB, et al. Sickle cell disease: continuous arterial spin-labeling perfusion MR imaging in children. Radiology. 2003;227(2):567-574. 27. Helton KJ, Paydar A, Glass J, et al. Arterial spin-labeled perfusion combined with segmentation techniques to evaluate cerebral blood flow in white and gray matter of children with sickle cell anemia. Pediatr Blood Cancer. 2009;52(1):85-91. 28. Gevers S, Nederveen AJ, Fijnvandraat K, et al. Arterial spin labeling measurement of cerebral perfusion in children with sickle cell disease. J Magn Reson Imaging. 2012; 35(4):779-787. 29. Waltz X, Pichon A, Mougenel D, et al. Hemorheological alterations, decreased cerebral microvascular oxygenation and cerebral vasomotion compensation in sickle cell patients. Am J Hematol. 2012; 87(12):1070-1073. 30. Nahavandi M, Nichols JP, Hassan M, Gandjbakhche A, Kato GJ. Near-infrared spectra absorbance of blood from sickle cell patients and normal individuals. Hematology. 2009;14(1):46-48. 31. Charlot K, Antoine-Jonville S, Moeckesch B, et al. Cerebral and muscle microvascular oxygenation in children with sickle cell disease: Influence of hematology, hemorheology and vasomotion. Blood Cells Mol Dis. 2017;65:23-28. 32. Ware RE, Davis BR, Schultz WH, et al. Hydroxycarbamide versus chronic transfusion for maintenance of transcranial doppler flow velocities in children with sickle cell anaemia-TCD with transfusions changing to hydroxyurea (TWiTCH): a multicentre, open-label, phase 3, non-inferiority trial. Lancet. 2016;387(10019):661670. 33. Bernaudin F, Verlhac S, Peffault de Latour R, et al. Association of matched sibling donor hematopoietic stem cell transplantation with transcranial Doppler velocities in children with sickle cell anemia. JAMA. 2019; 321(3):266-276.

965

LETTERS TO THE EDITOR Loss of 5-hydroxymethylcytosine expression is near-universal in B-cell lymphomas with variable mutations in epigenetic regulators Epigenetic alterations are increasingly recognized in human malignancies, with loss of 5-hydroxymethylcytosine (5hmC) expression by immunohistochemistry as a potential correlational proxy for malignant epigenetic change. In this study, nearly 100 cases of a broad range of B-cell lymphomas (BCL) studied via 5hmC immunohistochemistry showed pervasive loss of 5hmC expression. Whole exome sequencing (WES) performed on a subset of diffuse large B-cell lymphoma (DLBCL) showed a significant fraction of cases (58.8%) with missense mutations in such genes CREBBP, IDH1, IDH2, TET1, TET2, and WT1. Epigenetic pathways related to gene expression, including DNA methylation, are often dysregulated in human cancers. 5hmC is increasingly gaining recognition as an important epigenetic mediator, indicative of active demethylation and corresponding gene activation.1,2 While patterns of gene expression resulting from epigenetic modifications are often variable, 5hmC loss (as a surrogate for TET loss-of-function and 5mC persistence) is associated with some human cancers, including melanoma, mesothelioma, acute myeloid leukemias, and myelodysplastic syndrome.3 TET2 mutations in lymphoid cells exert pleiotropic effects depending on the cellular compartment. Conditional knockout of TET2 in immature B cells leads to development of lymphoblastic leukemias,4 whereas

966

TET2-deficiency in hematopoietic stem cells predisposes to myeloid leukemias.5 Mutations in epigenetic modifying genes, such as DNMT3A, TET2, or IDH2, are associated with lack of 5hmC expression and recurrent in lymphomas of follicular helper T-cell origin.6 Mutations in epigenetic modifying genes (viz., EZH2) are also common in BCL, including follicular lymphoma (FL) and germinal center-derived DLBCL, supporting the rationale for 5hmC evaluation in BCL. An earlier study of TET1deficient mice associated loss of 5hmC immunostaining with development of BCL, suggesting a relationship between TET1 expression, 5hmC content, and lymphomagenesis.7 With this background, we evaluated a spectrum of BCL for 5hmC expression and the association of expression with mutations in epigenetic pathway-related genes. After review of histology, 92 cases of BCL (5 whole section and 87 on tissue microarray [TMA]) were stained for 5hmC (polyclonal 1:1,000, active motif, see Figure 1). Five normal lymph nodes were also examined in tandem. A subset of the whole section cases were subject to double stain performed for CD3 (red)/5hmC (brown) to assess 5hmC in the T-cell compartment. All staining was performed on the Leica Bond IIITM autostainer. Staining was scored dichotomously as absent or present. WES 17 DLBCL cases from whole sections was also performed. Tumor DNA from 17 large-cell lymphoma samples within the above cohort was isolated using the AllPrep DNA/RNA FFPE kit (Qiagen) and matched germline DNA was obtained using peripheral blood with the DNeasy Blood/Tissue kit (Qiagen). One case showing slightly increased large cells, morphologically closer

Figure 1. 5-hydroxymethylcytosine expression by lymphoma subtypes. (A) Lymphomas were organized into three groups: high-grade B-cell lymphoma (HGBCL) /diffuse large B-cell lymphoma (DLDCL) low-grade (including mantle cell lymphoma [MCL]), and Hodgkin lymphoma [HL]; with constituent subtypes specified, and average percent 5-hydroxymethylcytosine (5hmC) loss per group tabulated by total numbers of cases. Overall, 94% of large B-cell lymphomas, 94% low-grade B-cell lymphomas (including MCL) and 88.5% of HL showed loss of 5hmC expression. One very weakly stained MCL case was considered within the negative group. The HGBCL cases included 1 double-hit lymphoma and one HGBCL-not otherwise specified (NOS) as per World Health Organization 2017 schema. Graphs were generated within Plotly and Visual Studio code.14 (B) Three sample cases of lymphomas demonstrating loss of expression in the neoplastic cells (black arrowhead) depicting Hodgkin/ReedSternberg cell, diffuse DLBCL as well as neoplastic follicles of follicular lymphoma (FL). Background endothelial cells and reactive lymphoid cells demonstrate strong retained expression of 5hmC (blue arrows). NLPHL: nodular lymphocyte predominant Hodgkin lymphoma; cHL: classical Hodgkin lymphoma; B-NHL: B-cell non-Hodgkin lymphoma; LPL: lymphoplasmacytic lymphoma; MZL: marginal zone lymphoma; SLL: small lymphocytic lymphoma; THRBCL: T-cell/histiocyte-rich B-cell lymphoma; PMBL: primary mediastinal (thymic) large B-cell lymphoma; RT: Richter transformation.

haematologica | 2022; 107(4)

Letters to the Editor

to nodular lymphocyte predominant Hodgkin lymphoma (NLPHL), was excluded from WES analysis to maintain homogeneity. The Agilent SureSelect XT Human All Exon v6 Kit captured whole-exome and untranslated regions (UTR), with reads generated using Illumina HiSeq2500 and NovaSeq6000 at Theragen Bio Co., Ltd, due to performance of sequencing over two separate experiments. See the Online Supplementary Table S1 for additional methods on read alignment. Somatic mutations were then called using Mutect2 through GATK4 either paired germline DNA or best practices provided panel of normals. Identified mutations were then post-processed by filtering according to best practices and annotated using the vcf2maf tool,8 which integrates the annotation tool Variant Effect Predictor (VEP) from Ensembl and a format conversion step. The final annotated mutations from WES were then analyzed in R (version 4.0.3) using tidyverse (version 1.3.0) best practices and the package maftools (version 2.4.12).9 See the Online Supplementary Table S1 and Online Supplementary Figure S1 for results pertaining to cell of origin status and predicted significance of observed mutations and their locations. The study was approved by the UOC Institutional Review Board (IRB13-1297). Normal lymph nodes retained 5hmC in the mantle, marginal, and paracortical areas with isolated loss only in the germinal center B (GCB) cells with scattered follicular dendritic cells showing retained 5hmC. Notably, a

significant majority of intrafollicular (likely CD4 follicular helper T cells) and extrafollicular T cells showed loss of 5hmC on the CD3/5hmC double stain (Figure 2A). Among all lymphoma cases (n=92), the majority of high-grade, low-grade B-cell and Hodgkin lymphomas (HL) showed loss of 5hmC in neoplastic cells (94%, 94%, and 88.5% of cases, respectively, Figure 1B, Figure 2B and D). 5hmC loss occurred in over 90% of lymphoma cells in any given case. Partial/variable loss only occurred in four classical HL (cHL) (2 were considered 5hmC retained, while 2 with extensive loss in over 90% of cells were considered 5hmC lost), demonstrating occasional weak staining in a subset of Hodgkin cells. When the DLBC-not otherwise specified (NOS) with available cell of origin (COO) data were stratified by COO, there were 11 GCB cell, 11 non-GCB cell, and one undetermined COO, with all cases showing 5hmC loss. Cases with weak staining and partial loss was not observed, except in one mantle cell lymphoma (MCL). Rare cases of high-grade BCL (HGBCL) with retained 5hmC expression consisted of one primary mediastinal BCL and one DLBCL-Richter transformation (RT) from chronic lymphocytic leukemia and small lymphocytic lymphoma (CLL/SLL). The only two low-grade BCL with retained 5hmC expression were both CLL/SLL. HL with retained 5hmC expression were split between two cHL (9% of cHL) and one NLPHL (25% of NLPHL). Only two DLBCL-RT from CLL/SLL were included within the DLBCL-NOS set. One case showed retained

Figure 2. T-cell panel: CD3 (red)/5-hydroxymethylcytosine (brown) double stain on reactive node and B-cell lymphoma. (A) Low power magnification of germinal center-mantle interface. Noted normal mantle cells positive for 5-hydroxymethylcytosine (5hmC) while perifollicular T cells are in red. High power magnification of the paracortical areas (panel on right top) show several T cells with only CD3 (red) while 5hMC (brown) is negative in these cells. Germinal center areas at high power with scattered T cells (likely CD4+ follicular helper T cells) with loss of 5hmC. Normal follicular dendritic cells (large doublets) express strong 5hmC while germinal center B cells are also negative. (B, C and D) Correspond to 1 case each of classical Hodgkin Lymphoma (cHL), diffuse large B-cell lymphoma (DLBCL), and T-cell/histiocyte-rich B-cell lymphoma (THRBCL) with neoplastic cells negative for 5hMC (black arrowheads) with a significant component of microenvironment T cells also negative for nuclear 5hmC.

haematologica | 2022; 107(4)

967

Letters to the Editor

5hmC expression and the other showed concordant 5hmC loss in both the high-grade DLBCL and residual low-grade CLL/SLL components. In most lymphomas, reactive background small lymphoid cells with strong retained 5hmC expression served as internal controls. However, in some Hodgkin and DLBCL cases (including transformed FL), a significant fraction of non-neoplastic background milieu also demonstrated loss of 5hmC expression (Figures 2B and C; Online Supplementary Figure S1). WES analysis interrogating for mutations in epigenetic regulators (IDH1, IDH2, TET1, TET2, KMT2D, EZH2 and CREBBP) showed missense mutations in ten of 17 DLBCL tested (58.8% of cases) with CREBBP seen most frequently (Figure 3). Details on the nature of these mutations and effect on protein are detailed in the Online Supplementary Figure S2. This study demonstrated near-universal loss of 5hmC expression by immunohistochemistry across a wide range of HGBCL and LGBCL (>90%) and HL (88%). The observations support results from prior studies in BCL by Matsuda et al. and Siref et al., and extends the observation to additional BCL including MCL (typical and blastoid variants) as well as DLBCL-NOS.10,11 The Matsuda study evaluated four subtypes of BCL (follicular lymphoma [FL], chronic lymphocytic leukemia [CLL], MCL, and Burkitt lymphoma [BL]) and noted uniform loss of 5hmC expression in FL and BL.10 Lack of staining in FL is congruent with our findings. We included two cases of HGBCL (1 double hit, another without),

and both showed uniform loss of 5hmC expression. Their study also found that all CLL and most MCL cases retained 5hmC expression, with only two of 11 MCL demonstrating loss. In contrast, our study noted loss of expression in the majority of CLL cases (6/8) and all four MCL cases (Figure 1A). Their study included one DLBCL-RT, but staining pattern details in the transformed component were not reported. We expected an increased likelihood of 5hmC loss in the transformed component, but noted an inverse pattern with retained 5hmC in one DLBCL-RT and loss in one CLL component, suggesting that loss is not correlated with progression in BCL. Our data in HL cases is aligned with the observations of Siref et al. which reported near-universal loss of expression in cHL.11 Additionally, we assessed 17 DLBCL cases via WES for missense mutations explanative of 5hmC expression loss in the aforementioned epigenome-related genes (DNMT3A, TET1, TET2, IDH1, IDH2 and WT1). While about half of these cases demonstrated mutations in one or more of these genes, most without alterations still showed 5hmC loss. This aligns with observations by Lemonnier et al. in T-cell lymphoma where a significant proportion of cases with 5hmC loss did not show mutations in TET2 or DNMT3A.6 From a mechanistic perspective, BCL carry mutations in these epigenetic regulator genes less frequently. Rather, a subset of FL and DLBCL (enriched for germinal center COO) harbor mutations in the epigenetic modulator EZH2 that promotes increased suppressive trimethylation via H3K27me, affecting 5mC

Figure 3. Whole exome sequencing data looking at epigenetic regulators (IDH1, IDH2, TET1, TET2) mutations in 17 diffuse large B-cell lymphomas. (A) Distribution of cases stratified by mutations and cell of origin (COO) status showing that most cases with mutations were enriched in the non-germinal center COO. For more information, see the Online Supplementary Figure S2.

968

haematologica | 2022; 107(4)

Letters to the Editor

hydroxylation.12,13 The minimal number of mutated cases impedes speculation since we focused on just DLBCL to ensure homogeneity. However, non-GCB predominance in mutation-positive DLBCL in our study suggests that 5hmC loss is likely independent of EZH2 mutation status, although the exact mechanism remains poorly understood. From a translational perspective, it was recently demonstrated that 5hmC in circulating cellfree DNA assessed by chemical labeling-based sequencing technology correlated with prognosis in newly-diagnosed DLBCL and hence examining TET1, TET2 in conjunction with 5hmC and 5mC may have prognostic utility in the setting of BCL.2 In summary, we corroborate previously published data and extend current insights by demonstrating loss of 5hmC expression in most BCL. This loss may be diagnostically useful in establishing a malignant B-cell phenotype in limited samples without flow cytometry/molecular data. However, the loss of 5hmC in reactive background T cells (in normal and malignant nodes) indicates that 5hmC loss in T cells is not a surrogate of aberrant/neoplastic phenotype. Kevin S. Tanager,1* Jovian Yu,2* Brian C-H Chiu,3 Timothy C. Carll,1 Alexandra H. Tatarian,4 Peter Riedell,2 Sonali Smith,2 Justin Kline2# and Girish Venkataraman1# 1 The University of Chicago Medicine, Department of Pathology, Chicago, IL; 2The University of Chicago Medicine, Department of Hematology/Oncology, Chicago, IL; 3The University of Chicago Medicine, Department of Public Health Sciences, Chicago, IL and 4 SUNY Upstate College of Medicine, Syracuse, NY, USA *KST and JY contributed equally as co-first authors # JK and GV contributed equally as co-senior authors Correspondence: GIRISH VENKATARAMAN - girish.venkataraman@uchospitals.edu doi:10.3324/haematol.2021.279648 Received: July 19, 2021. Accepted: November 30, 2021. Pre-published: December 16, 2021. Disclosures: no conflicts of interest to disclose Contributions:: KST and JY conducted the research and wrote manuscript; BC, PR and SS designed the study and wrote parts of the manuscript; TCC and AHT wrote the manuscript and conducted a smaller portion of the study; GV and JK designed the study,

haematologica | 2022; 107(4)

conducted the research and edited significant portions of the manuscript. Acknowledgements: we thank Veronique Saada and Cyril Quivoron, Department of Translational Hematology and Pathology, Gustave Roussy Cancer Campus, Villejuif, France for help with proofing the manuscriptand Jessica Robertson of the lymphoma program for help with logistics of the Hoogland Lymphoma Biobank tissue array data. Funding: the study was funded in part by the Hoogland Lymphoma Biobank at the University of Chicago Medicine.

References 1. Branco MR, Ficz G, Reik W. Uncovering the role of 5-hydroxymethylcytosine in the epigenome. Nat Rev Genet. 2011;13(1):7-13. 2. Chiu BC, Zhang Z, You Q, et al. Prognostic implications of 5-hydroxymethylcytosines from circulating cell-free DNA in diffuse large Bcell lymphoma. Blood Adv. 2019;3(19):2790-2799. 3. Chapel DB, Husain AN, Krausz T. Immunohistochemical evaluation of nuclear 5-hydroxymethylcytosine (5-hmC) accurately distinguishes malignant pleural mesothelioma from benign mesothelial proliferations. Mod Pathol. 2019;32(3):376-386. 4. Mouly E, Ghamlouch H, Della-Valle V, et al. B-cell tumor development in Tet2-deficient mice. Blood Adv. 2018;2(6):703-714. 5. Moran-Crusio K, Reavie L, Shih A, et al. Tet2 loss leads to increased hematopoietic stem cell self-renewal and myeloid transformation. Cancer Cell. 2011;20(1):11-24. 6. Lemonnier F, Poullot E, Dupuy A, et al. Loss of 5-hydroxymethylcytosine is a frequent event in peripheral T-cell lymphomas. Haematologica. 2018;103(3):e115-e118. 7. Cimmino L, Dawlaty MM, Ndiaye-Lobry D, et al. TET1 is a tumor suppressor of hematopoietic malignancy. Nat Immunol. 2015;16(6):653-662. 8. Van der Auwera GA & O'Connor BD. Genomics in the Cloud: Using Docker, GATK, and WDL in Terra. Sebastopool (CA): O'Reilly Media; 2020. 9. Mayakonda A, Lin DC, Assenov Y, Plass C, Koeffler HP. Maftools: efficient and comprehensive analysis of somatic variants in cancer. Genome Res. 2018;28(11):1747-1756. 10. Matsuda I, Imai Y, Hirota S. Distinct global DNA methylation status in B-cell lymphomas: immunohistochemical study of 5-methylcytosine and 5-hydroxymethylcytosine. J Clin Exp Hematop. 2014;54(1):67-73. 11. Siref A, McCormack C, Huang Q, Lim W, Alkan S. Diminished expression of 5hmc in Reed-Sternberg cells in classical Hodgkin lymphoma is a common epigenetic marker. Leuk Res. 2020;96:106408. 12. Dobashi A. Molecular pathogenesis of diffuse large B-cell lymphoma. J Clin Exp Hematop. 2016;56(2):71-78. 13. Schmitz R, Wright GW, Huang DW, et al. Genetics and pathogenesis of diffuse large B-cell lymphoma. N Engl J Med. 2018;378(15):13961407. 14. Sievert C. Interactive Web-Based Data Visualization with R, Plotly, and Shiny. Chapman and Hall/CRC, 2020.

969

Letters to the Editor

Immunophenotypic changes in leukemic blasts in children with relapsed/refractory B-cell precursor acute lymphoblastic leukemia after treatment with CD19-directed chimeric antigen receptor (CAR)expressing T cells Implementation of CD19 targeting with the bispecific engager blinatumomab and T cells harboring chimeric antigen receptor (CAR-T) has resulted in significant improvement in therapy results in children with relapsed/refractory (R/R) B-cell precursor acute lymphoblastic leukemia (BCP-ALL).1,2 However, some patients do not respond to therapy or experience relapse.1,3 Minimal residual disease (MRD) persistence or reappearance during CD19-directed therapy is a key sign predicting failure of pan-B-cell antigen targeting. Thus, accurate detection of residual tumor cells has emerged as a key tool in evaluating the efficacy of such immunotherapy. Loss of targeted antigen, which is the main escape mechanism for BCP-ALL under selective pressures of CD19-directed treatment, is a significant obstacle for MRD monitoring by multicolor flow cytometry (MFC). As MFC-MRD assessment is ordinarily based on CD19positive compartment analysis,4 possible CD19 negativity of tumor cells increases the significance of other immunophenotypic markers for MFC data analysis. Nevertheless, antigens applicable for immunophenotypic aberrations evaluation are also known to undergo expression modulation during therapy.5 Recently, we demonstrated that each antigen useful for MFC-MRD detection can be either increased or decreased in children treated with blinatumomab.6 As longer CAR-T cell therapy persistence leads to prolonged pressure on leukemic cells compared to blinatumomab,3 this phenomenon could probably also affect the immunophenotypic stability of the leukemia, making MFC-MRD assessment in such patients very challenging. The current report briefly summarizes our data on MFC-MRD and relapse detection in patients treated with CAR-T cell therapy with a main focus on changes in the expression of markers that are relevant for MFC-MRD investigation. We carried out a retrospective review of 39 pediatric patients with R/R ВCP-ALL who received CAR-T cell therapy between February 2018 and September 2020 and in whom tumor blasts were detectable in the bone marrow (BM) at least once after CAR-T cell infusion. Among them, there were four patients resistant to CART cell therapy, 16 who experienced relapse (>5% blast cells by MFC) and 19 children with blasts in the BM on MFC-MRD level. The characteristics of the patients, including cytogenetic data, are presented in the Online Supplementary Table S1. The study was approved by the Ethics Committee of the Dmitry Rogachev National Medical Research Center of Pediatric Hematology, Oncology and Immunology, and informed consent for the collection and investigation of samples and for participation in current study was obtained from patients’ parents or legal guardians. Subjects reported here were either included in a prospective trial of CD19 CAR-T cell therapy with the 41BB costimulatory domain (clinicaltrials gov. Identifier: NCT03467256) or treated on a compassionate use basis. Morphological examination and MFC-MRD detection in BM aspirates were performed before CAR-T cell infusion, on days 14 and 28, and 2, 3, 4, 6, 8, 10, 12, and 24 months after the beginning of CAR-T cell therapy. The immunophenotypic stability of leukemic cells was eval970

uated only after one CAR-T cell therapy course. All patients underwent routine diagnostic immunophenotyping and MRD detection by 8-10-color flow cytometry according to the standard protocols of the Moscow-Berlin group.7,8 During the study period, MFC was performed on FACS Canto II, FACS Celesta (both from Becton Dickinson, BD, US), CytoFlex and Navios (both from Beckman Coulter, BC, US) flow cytometers. EuroFlow guidelines for machine performance monitoring were used.9 Cytometer Setup and Tracking beads (BD), Flow-Check Pro Fluorospheres (BC) and CytoFLEX Daily QC Fluorospheres (BC) were used for daily cytometer optimization. Normal lymphocytes were used as the control for positivity/negativity definition. The tumor cell immunophenotype was analyzed with a focus on markers applicable for MFC-MRD investigation. The Online Supplementary Table S2 provides a list of monoclonal antibodies used for MFC-MRD monitoring. CD22 and CD24 were additionally studied as suggested by S. Cherian et al.10 MFC-MRD results were controlled with IG/TR gene rearrangements and fusion genes monitoring by molecular techniques with high concordance obtained (85% of qualitatively concordant MRD results for CD19-negative relapses and 79% for those who retained CD19 positivity). Expression of surface antigens was deemed positive if the antigen was expressed on more than 20% of tumor cells.7 An increase/decrease in the expression of each single antigen was defined as a change in the percentage of positive cells of more than 25%. Proportions of patients with stable and changed expression of each single antigen between CD19-negative and CD19-positive entities were compared using Fisher’s exact test. The CD19 expression status on leukemic cells after CAR-T cell therapy is depicted in Figure 1. The blasts of all patients except one were totally CD19-positive prior to CAR-T administration. In the remaining patient, only 30% of BM cells expressed the target antigen, while in cerebrospinal fluid (CSF), all leukemic BCP were CD19positive. In five patients, the leukemic cells at relapse were CD19-positive, and in 11 relapsed patients, CD19 negativity was found. Two of 11 CD19-negative relapses displayed a small CD19-positive subpopulation (5% and 15% respectively). Patients who were resistant to CART cell therapy had preserved CD19 expression. The blasts of children who had leukemic cells in the BM on MFC-MRD level only (n=19) were either CD19-positive (n=10) or CD19 negative (n=9). CD19-negative leukemia appeared a little less frequent in children who have already received CD19 targeting (8 of 16, 50%) comparing to those who were not pretreated with blinatumomab (12 of 19, 63%, P=0.506). The immunophenotypic changes of leukemic cells in the 16 relapsed patients are presented in Figure 2A. For all antigens applicable for MFC-MRD assessment except CD58, expression changes were demonstrated, either increased or decreased expression (Figure 3). On the other hand, the proportion of patients with noted changes was relatively small: CD34 was found to be the most unstable antigen (changes in 4 of 16 patients, 25%). We did not find differences in the frequencies of changes in marker expression among CD19-negative and CD19-positive relapses (Figure 2A). The expression of CD22 and CD24, which are suggested to be candidates for CD19 substitution,10 displayed remarkable stability, although CD24 was tested only in part of the study group. We analyzed 19 patients who did not relapse but had leukemic cells at the MRD level in the BM by MFC at least once during the follow-up period. In haematologica | 2022; 107(4)

Letters to the Editor

Figure 1. CD19 expression in the studied patients (n=39). The study group included resistant patients (n=4), patients with relapse (5 CD19-positive and 11 CD19-negative) and patients with blasts detected by multicolor flow cytometry (MFC) at the minimal residual disease (MRD) level in the bone marrow at least once (n=19).

addition to explicable downmodulation of CD19 in nearly half of the patients (9 of 19), as in the relapsed patients, CD34 displayed the highest instability (expression changes in 10 of 19 patients), while other antigens again were rather stable, with no changes noted in the expression of CD38, CD58 and CD22 (Figure 2B). The leukemic cells in four resistant patients had a rather stable immunophenotypic profile with very rare cases of antigen expression changes (Figure 2C). As CD19 is the best known target for immunotherapy in BCP-ALL,1,2 CD19-directed CAR-T cells are widely used as the salvage approach for R/R patients.3 Thus, lack or preservation of CD19 expression is crucial for further treatment strategy choice. D. Libert et al. have shown that CD19 negativity occurs in nearly 65% of relapses after CAR-T cell application11 with a massive disproportion in frequency between 4-1BB- and CD28containing platforms (85% vs. 22% of relapsed patients, respectively).11 In addition, as discussed previously,6 loss of CD19 could break the well-established conventional algorithm of MFC-MRD gating, as cytometric residual leukemia detection is based on B-cell compartment investigation.4 In our study, 4-1BB CAR-T cells were used, and the frequency of CD19-negative relapses was 68.8%. Taking these patients together with the resistant patients (all remained CD19-positive) and the patients with only MFC-detected MRD tumor cells, the total incidence of CD19 negativity was 51.3% in all children in whom tumor cells were detectable in the BM during follow-up after the first CAR-T cell therapy course. In all three patients in whom CD19-positive relapse was preceded by MFC-MRD reappearance, the residual leukemia had preserved expression of CD19 (Online haematologica | 2022; 107(4)

Supplementary Figure S1). Among the 11 CD19-negative relapses, six were preceded by CD19-negative MFCMRD, and one was preceded by a progressive decrease in CD19 expression on MFC-MRD, while in others, relapse developed directly after MFC-MRD negativity (Online Supplementary Figure S1). As CD19 downregulation was noted in more than half of the patients, for antigens useful for primary B-lineage gating (CD10, CD22, CD24)10,12 not only the stability but also homogeneous positivity on tumor cells is vital. In the current study, neither of these antigens displayed significant changes, although their expression prior to CAR-T cell infusion was not always satisfactory. Indeed, in six of 33 and eight of 25 patients, less than 90% of the leukemic population before the CAR-T cell course was CD22-positive (median 64%, range, 4-73%) and CD24positive (median 0%, range, 0-79%), respectively. As expected,13 in four patients with low CD22 and in five patients with low CD24, rearrangements involving the KMT2A gene (KMT2A-r) were found. CD10 negativity/low expression (median 9%, range, 0-52%) was also not restricted to KMT2A-r cases: it was found in 13 patients (7 with KMT2A-r) and remained stable in nine children. Moreover, in two of three cases of CD10 upmodulation, this marker expression after CAR-T cell therapy was still not total. Previously, we showed the results of a similar study that included patients with R/R BCP-ALL treated with blinatumomab only.6 Compared to that after blinatumomab treatment, the proportion of cells with CD19 loss after CAR-T cell treatment was significantly higher (51.3% vs. 27.1%), although no lineage switches were noted in the current work (comparing 3 of 30 relapses 971

Letters to the Editor

Figure 2. Frequency of changes in the immunophenotype of leukemic blasts. (A) Relapsed patients. (B) Patients who had detectable blasts in the bone marrow only at the minimal residual disease (MRD) level. (C) Resistant patients.

972

haematologica | 2022; 107(4)

Letters to the Editor

after blinatumomab). Generally, the antigen profile of leukemic cells after CAR-T treatment displayed relatively higher stability, although CD58 was the most stable and CD34 was the most unstable marker irrespective of the type of immunotherapy given. Moreover, no differences in the immunophenotype of CD19-negative and CD19-positive relapses and MFC-MRD persistence were noted after CAR-T cell therapy, while after BiTE treatment, CD45 and CD38 changed more frequently in the case of CD19 loss.6 With its peculiar immunophenotype and tendency towards lineage switch, KMT2A-r BCP-ALL cases require special solutions for MFC-MRD monitoring after CD19 targeting. In our cohort, there were seven patients with various types of KMT2A-r. Before CAR-T cell therapy, they were either CD10-negative (n=6) or borderline CD10-positive (n=1, CD10 decreased at MFC-MRD reappearance). Among six children with available CD22 and CD24 expression, only two were totally CD22-positive and one was CD24-positive. Moreover, in three patients of this subgroup, ALL recurrence was CD19negative. Fortunately, no lineage switches were regis-

tered in this study (probably due to relatively few patients investigated), although they were rather frequent in case of blinatumomab application.6 Described immunophenotypic features make this genetic subgroup a very hard target for MFC-MRD technique. Being notably rare in the general BCP-ALL population, KMT2A-r are present in a significant part of R/R patients14,15 (17.9% in the current study). Considering this, addition of at least one early B-lineage antigen (iCD79a preferably) and careful analysis of the whole immunophenotype with searching both for initial antigenic patterns and for cells different from normal counterparts (with respect to specific regeneration patterns)16 could lead to reliable results, highly comparable with molecular techniques data (86% with IG/TR rearrangements monitoring by next-generation sequencing and 98% with polymerase chain reaction-based fusion genes transcript detection). Moreover, additional MRD confirmation tools, such as fluorescence in situ hybridisation or chimerism studies17 from sorted suspicious cells, can help in MFC-MRD monitoring of such a complicated BCP-ALL subgroup.

Figure 3. Representative examples of immunophenotypic plasticity in patients after T cells harboring chimeric antigen receptor. (A) Case of CD19-negative relapse. (B) Case of reappearance of CD19-negative minimal residual disease (MRD). (C) Case of phenotypically stable resistance to the immunotherapy. CAR-T: CD19-directed chimeric antigen receptor (CAR)expressing T cells.

haematologica | 2022; 107(4)

973

Letters to the Editor

Despite the more stable antigen profile of leukemic cells after CAR-T cell therapy than after blinatumomab treatment, more frequent loss of CD19, which is the cornerstone of B-lineage gating during conventional MFCMRD detection, necessitates usage of additional B-lineage antigens such as CD22, CD24, CD10 and iCD79a for primary B-cell compartment gating. Therefore, the final search for MFC-MRD becomes more sophisticated6,18 than conventional CD19-based methods, especially taking into account specific BM regeneration patterns that are more visible after CD19 targeting.16 Thus, considering the frequent loss of CD19 as well as the modulation of the expression of all other antigens relevant to MFC-MRD monitoring, a large panel of antibodies including additional B-lineage markers (CD22, CD24, iCD79a, etc.) combined with modified gating strategy and consideration of specific background variations, should be applied to increase the effectiveness of MFC-MRD detection in BCP-ALL patients after CD19 targeting by CAR-T cell therapy. Ekaterina Mikhailova, Olga Illarionova, Larisa Shelikhova, Elena Zerkalenkova, Olga Molostova, Yulia Olshanskaya, Galina Novichkova, Alexey Maschan, Michael Maschan and Alexander Popov Dmitry Rogachev National Medical Research Center of Pediatric Hematology, Oncology and Immunology, Moscow, Russian Federation Correspondence: ALEXANDER M POPOV - uralcytometry@gmail.com doi:10.3324/haematol.2021.279677 Received: July 21, 2021. Accepted: December 2, 2021 Pre-published: December 16, 2021. Disclosures: no conflicts of interest to disclose. Contributions: EM and AP designed the study, acquired and analyzed cytometric data and wrote the paper; OI acquired and analyzed cytometric data; LSh collected clinical data and wrote the paper; EZ acquired and analyzed cytogenetic and molecular genetic data, and wrote the paper; OM collected clinical data; YuO acquired and analyzed cytogenetic and molecular genetic data; GN and MM designed the study and acted as the general supervisor; AM designed the study, acted as the general supervisor and wrote the paper. Funding: the KMT2A rearrangement assessment study was supported by RFBR grant number 7-29-06052 and Presidential grant number MK-1645.2020.7 and 075-15-2020-338. Flow cytometric minimal residual disease evaluation was supported by RFBR grant number 18-29-09132.

References 1. Topp MS, Gokbuget N, Zugmaier G, et al. Phase II trial of the antiCD19 bispecific T cell-engager blinatumomab shows hematologic and molecular remissions in patients with relapsed or refractory Bprecursor acute lymphoblastic leukemia. J Clin Oncol. 2014;32(36):4134-4140. 2. Maude SL, Frey N, Shaw PA, et al. Chimeric antigen receptor T cells

974

for sustained remissions in leukemia. N Engl J Med. 2014;371(16):1507-1517. 3. Maude SL, Laetsch TW, Buechner J, et al. Tisagenlecleucel in children and young adults with B-cell lymphoblastic leukemia. N Engl J Med. 2018;378(5):439-448. 4. Dworzak MN, Gaipa G, Ratei R, et al. Standardization of flow cytometric minimal residual disease evaluation in acute lymphoblastic leukemia: multicentric assessment is feasible. Cytometry B Clin Cytom. 2008;74(6):331-340. 5. Dworzak MN, Gaipa G, Schumich A, et al. Modulation of antigen expression in B-cell precursor acute lymphoblastic leukemia during induction therapy is partly transient: evidence for a drug-induced regulatory phenomenon. Results of the AIEOP-BFM-ALL-FLOWMRD-Study Group. Cytometry B Clin Cytom. 2010;78(3):147-153. 6. Mikhailova E, Gluhanyuk E, Illarionova O, et al. Immunophenotypic changes of leukemic blasts in children with relapsed/refractory B-cell precursor acute lymphoblastic leukemia, who have been treated with Blinatumomab. Haematologica. 2020;106(7):2009-2012. 7. Novikova I, Verzhbitskaya T, Movchan L, et al. Russian-Belarusian multicenter group standard guidelines for childhood acute lymphoblastic leukemia flow cytometric diagnostics. Oncohematology. 2018;13(1):73-82. 8. Popov A, Belevtsev M, Boyakova E, et al. Standardization of flow cytometric minimal residual disease monitoring in children with Bcell precursor acute lymphoblastic leukemia. Russia–Belarus multicenter group experience. Oncohematology. 2016;11(4):64-73. 9. Kalina T, Flores-Montero J, Lecrevisse Q, et al. Quality assessment program for EuroFlow protocols: summary results of four-year (2010-2013) quality assurance rounds. Cytometry A. 2015;87(2): 145-156. 10. Cherian S, Miller V, McCullouch V, et al. A novel flow cytometric assay for detection of residual disease in patients with B-lymphoblastic leukemia/lymphoma post anti-CD19 therapy. Cytometry B Clin Cytom. 2018;94(1):112-120. 11. Libert D, Yuan CM, Masih KE, et al. Serial evaluation of CD19 surface expression in pediatric B-cell malignancies following CD19-targeted therapy. Leukemia. 2020;34(11):3064-3069. 12. Mejstrikova E, Hrusak O, Borowitz MJ, et al. CD19-negative relapse of pediatric B-cell precursor acute lymphoblastic leukemia following blinatumomab treatment. Blood Cancer J. 2017;7(12):659. 13. De Zen L, Bicciato S, te Kronnie G, Basso G. Computational analysis of flow-cytometry antigen expression profiles in childhood acute lymphoblastic leukemia: an MLL/AF4 identification. Leukemia. 2003;17(8):1557-1565. 14. Queudeville M, Schlegel P, Heinz AT, et al. Blinatumomab in pediatric patients with relapsed/refractory B-cell precursor acute lymphoblastic leukemia. Eur J Haematol. 2021;106(4):473-483. 15. Winters AC, Bernt KM. MLL-rearranged leukemias-an update on science and clinical approaches. Front Pediatr. 2017;5:4. 16. Mikhailova E, Semchenkova A, Illarionova O, et al. Relative expansion of CD19-negative very-early normal B-cell precursors in children with acute lymphoblastic leukaemia after CD19 targeting by blinatumomab and CAR-T cell therapy: implications for flow cytometric detection of minimal residual disease. Br J Haematol. 2021;193(3):602-612. 17. Semchenkova A, Brilliantova V, Shelikhova L, et al. Chimerism evaluation in measurable residual disease-suspected cells isolated by flow cell sorting as a reliable tool for measurable residual disease verification in acute leukemia patients after allogeneic hematopoietic stem cell transplantation. Cytometry B Clin Cytom. 2020;100(5):568-573. 18. Cherian S, Stetler-Stevenson M. Flow cytometric monitoring for residual disease in B lymphoblastic leukemia post T cell engaging targeted therapies. Curr Protoc Cytom. 2018;86(1):e44.

haematologica | 2022; 107(4)

Letters to the Editor

Early testicular maturation is sensitive to depletion of spermatogonial pool in sickle cell disease Cryopreservation of testicular tissues before hematopoietic stem cell transplant (HSCT) is offered by specialized centers worldwide to boys with sickle cell disease (SCD). In order to elucidate whether hydroxyurea (HU) therapy or disease-related factors affect the spermatogonial quantity of SCD patients, we collected clinical data of 29 boys (age range, 2.8-15.1 years) and calculated a Z-score. Our results show that most spermatogonial numbers (n=17) were below the reference values of healthy boys. There was a correlation between the number of spermatogonia and the age at HU initiation (P=0.029, r=0.476). This study suggests that besides factors intrinsic to SCD, an HU exposure in early life may lead to further depletion of spermatogonia reducing the potential for successful fertility preservation. SCD is the most common corpuscular anemia, with approximately 300,000 affected babies born anually.1 SCD manifests in hemolytic anemia and episodes of microvascular vaso-occlusion leading to end-organ ischemia-reperfusion injury and organ damage. Treatment to alleviate complications involves HU.14 In high-resource countries, allogeneic HSCT is proposed as a cure. However, there are significant risks, including acute toxicity and late effects, e.g., infertility caused by previous gonadotoxic conditioning.3 As fertility preservation, a growing number of centers worldwide cryopreserve either sperm or testicular tissues of prepubertal boys prior to HSCT.4,5 Cryopreservation of testicular tissue containing undifferentiated germ cells, termed spermatogonia, remains an experimental approach as protocols to differentiate spermatogonia into

sperm are under development.6 In fertility preservation programs, patients with SCD constitute up to 35% of all patients with non-malignant diseases.4 Based on limited case series, even before HU therapy, most young men with SCD have semen quality below the World Health Organization standard minimum criteria and spermatogonial numbers in prepubertal boys have been shown to be reduced.5,7-8 A recent study with a small cohort size concluded that spermatogonial numbers in HU exposed patients (n=17) were not significantly different from those without HU exposure (n=13).9 The current study was designed to evaluate if dose, exposure time, age at HU initiation, iron load or disease severity affects spermatogonial quantity in SCD patients. Insights enable clinicians to prospectively counsel patients and parents about the risks for reduced spermatogonial numbers prior to HSCT. Testicular tissues from 29 boys with SCD (age range, 2.8-15.1 years) participating in the fertility preservation programs Androprotect (German network, n=19) and Nordfertil (Nordic network, n=10) between 2013-2020 were included. Also, testicular tissue from two adult SCD patients (age range, 21.5-48.5 years, University Hospital Münster) were available. All age-appropriate patients and the guardians gave written informed consent for the use of testicular tissues for research. All procedures were in accordance with the Declaration of Helsinki and Ethical approval was obtained from the responsible Ethics Boards. Hormone levels, testicular volume, and markers of chronic hemolysis were determined, information detailing transfusions, HU therapy and SCD complications was collected, and severity scores were calculated (Table 1; Online Supplementary Table S1).10,11 Patients underwent open testicular biopsy during

Table 1. Clinical characteristics and Spearman correlations for mean spermatogonial numbers per round tubular cross section and fertility index Z-scores of the 24 patients with testicular samples containing a minimum of 25 evaluated tubules.

Clinical characteristics Age Hematological parameters HbS level, % of total Hb HbF level, % of total Hb MCV, fL Platelets, G/L Neutrophils, G/L Biomarkers of chronic hemolysis Hb, mg/L LDH, U/L Biomarker of iron status Ferritin, mg/L HU therapy† Dose, mg/kg Exposure time, d Age at HU start, y Dose multiplied by exposure time, mg x d Complications Pain crisis per year, n Number of ATS, n Score (Sebastiani et al.)10 Score C (Van den Tweel et al.)11

S/T Z-score P r

Median (range) 7.1 (2.8-15.1)

0.015*

16 13 15 22 20

54.2 (9.4-85.3) 12.0 (3.0-34.0) 86.0 (76.0-123.0) 294.0 (104.0-1,103.0) 4.2 (0.5-7.9)

0.778. 0.553 0.512 0.497 0.925

23 20

100.0 (74.0-134.0) 158.0 (86.2-262.3)

0.492*

FI Z-score P r

0.073

0.373

- 0.077 0.181 0.184 0.153 0.023

0.571 0.734 0.838 0.990 0.853

- 0.153 0.105 - 0.058 0.003 0.044

0.667 0.405

- 0.095 0.197

0.645 0.046*

- 0.101 0.451*

305.0 (51.0-3,244.0)

0.304

0.249

0.209

0.302

24 24 24

22.5 (0.0-45.0) 750.5 (0.0-2,839.0) 5.0 (1.0-9.9)

0.766 0.856. 0.029*

- 0.064 - 0.039 0.476*

0.537 0.565 0.024*

- 0.132 - 0.124 0.490*

17,434.0 (0.0-1,01636.0)

0.984

0.004

0.720

- 0.077

19 24 24 24

0.4 (0.0-4.0) 0.0 (0.0-3.0) 0.4 (0.3-0.6) 50.0 (5.0-80.0)

0.486 0.703 0.439 0.168

0.170 - 0.082 - 0.166 0.291

0.158 0.802 0.748 0.149

0.337 0.054 - 0.069 0.304

ATS indicates acute thorax syndrome; FI: feritlity index; S/T: spermatogonia per round tubular cross-section; Hb: hemoglobin; HbF: fetal hemoglobin; HbS: sickle hemoglobin; HU: hydroxyurea; y: year; x d: per day; LDH: lactate dehydrogenase; MCV: mean corpuscular volume; P: P-value and r, Spearman correlation coefficient. *P<0.05. † Washout period of 2 weeks - 4 months was used prior to biopsy for three HU-exposed patients. All 3 patients received transfusions.

haematologica | 2022; 107(4)

975

Letters to the Editor

which less than 20% of the testicular volume of one testis was sampled. Two-thirds of the tissues were transported to the research center, partly overnight, and were cryopreserved. The remaining third was fixed in formalin in the Nordfertil program. Tissues from the Androprotect program and adult patients were fixed in Bouin's solution. After embedding in paraffin, all tissues were sectioned (3-5 µm), and two independent sections (distance >15 µm) were immunostained with MAGEA4, following published protocols.12,13 At Karolinska Institutet, a fluorescence microscope (Eclipse E800, Nikon; Japan) was employed for analysis. In Münster, images were captured using the PreciPoint M8 microscope/scanner and subsequently analysed using the ViewPoint light software (1.0.0.9628, PreciPoint, Freising, Germany). Spermatogonia were identified based on their morphology (size, shape), location and MAGEA4 expression (Online Supplementary Figure S1). Using a blinded approach, all round tubular cross-sections within the tissue sections were quantified (mean: 94, range, 0-344) and classified as tubules with spermatogonia and tubules only

containing somatic Sertoli cells. Mean spermatogonial numbers per round tubular cross section (S/T) were assessed to obtain comparable spermatogonial numbers across samples (Online Supplementary Table S1). Moreover, the fertility index (FI) as the percentage of tubular cross-sections containing spermatogonia was determined. In order to control physiological variation in spermatogonial numbers during development, Z-scores were calculated for S/T and FI using reference means.14 For statistical analysis, only samples with >25 round tubules were considered, resulting in the inclusion of n=24 prepubertal/pubertal patient samples (Online Supplementary Table S1). Spearman correlation coefficient (r) and ROC analysis were performed to determine the relationships between spermatogonial quantity, age and treatment characteristics using IBM SPSS Statistics V26.0 software (IBM Corporation, Armonk, NY; US) and GraphPad Prism Version 8.4.3(471) (GraphPad Software, San Diego, CA, US). Out of the 29 prepubertal/pubertal samples, the major-

Figure 1. Number of spermatogonia per round tubular cross-section and fertility index in patients with sickle cell disease. (A) Representative images showing MAGEA4-positive spermatogonia (immunohistochemical staining) in patients with sickle cell disease (SCD) and controls (n=3, each) of different ages. Scale bars: 20 μm (4-10 years), 50 μm (42-48 years), arrows indicate spermatogonia and arrowheads Sertoli cells. Scale bars were added to the images using Adobe Photoshop CS2 (Adobe Systems, California, US). Objective: Olympus PlanC N 60x/0.80 (PreciPoint, Freising, Germany). (B) Spermatogonia per round tubular cross-section (S/T) and (C) S/T Z-score as well as the fertilty index (FI) (D) and FI Z-scores (E) by age in 29 patients with SCD. In (B) and (D) data are plotted on lines corresponding Z-scores for the mean reference values (Adapted from Funke et al.).14 A significant correlation exists between younger age and lower S/T Zscores (P=0.015, r=0.492). SD: standard deviation.

976

haematologica | 2022; 107(4)

Letters to the Editor

ity (n=17) scored below previously published reference values of S/T (Z-score <-3), but only four were devoid of spermatogonia (Figure 1A to C). Both adult patients showed spermatogonial numbers below the reference values (Online Supplementary Table S1). Our observations confirm previous reports of severely decreased spermatogonial numbers.5,7,9 Additionally, most patients (n=21/29) had FI Z-scores below the reference values (Figure 1D and E), illustrating that spermatogonia were only focally present. In line with Gille et al.,9 no difference was observed between the numbers of spermatogonia in HU-exposed (n=21) and non-exposed groups (n=3) although the small sample size results in limited informative value (P=0.9999). Three patients with a wash out period of 2 weeks to 4 months without HU prior to biopsy were included in the HU-exposed group. There was no correlation between spermatogonial numbers and HU dose or exposure time (Table 1; Figure 2A and B). Importantly, younger age correlated significantly with lower S/T (P=0.015, r=0.492) Z-score (Table 1; Figure 1C). Since 2014, an international consensus has recommended HU therapy for all infants with SCD aged 9 months or older to reduce complications.2 Our observations reflect this recommendation; younger patients had an earlier HU initiation (P<0.001, r=0.713), leading to HU exposure at an earlier time of testicular development compared to patients diagnosed before 2014. The young age at the HU initiation further correlated with lower S/T (P=0.029, r=0.476) and FI Z-scores (P=0.024, r=0.490) (Figure 2C and D). By using ROC analysis, we were able to unveil that an age of 2.4 years at HU initiation showed a good diagnostic value (area under the curve [AUC]:

0.84, 95% confidence interval [CI]: 0.67–1.00) with 75% sensitivity and 82% specificity to identify testicular samples containing very low spermatogonial numbers (S/T Z-score <-7). Therefore, testicular maturation during the first 3 years of life may be especially sensitive to depletion of the spermatogonial pool in SCD patients. Significant correlation however does not prove that early HU initiation determines the low S/T Z-score at very young age. Older age correlated significantly with higher S/T Zscores (P=0.015, r=0.492) and a cut-off age of 8.8 years identified with 100% sensitivity and 80% specificity (AUC 0.84, 95% CI: 0.64–1.00) S/T Z-scores within the normal range (>-3). Our data confirm that the pubertal increase in spermatogonial numbers does occur in SCD patients in line with a previous report.8 Also in cases of early HU exposure we cannot exclude that spermatogonia maintain their capacity of expansion at puberty. Besides the potentially harmful effects of early HU exposure, other factors intrinsic to SCD need to be considered. There is wide variability in the phenotypic severity of SCD. This variation can be explained partly by differences in the total hemoglobin concentration, the mean corpuscular hemoglobin concentration, iron load, coinheritance of a-thalassemia and fetal hemoglobin persistence.15 In line with this, low hemoglobin values observed in the present study correlated with a high frequency of pain crises (P=0.026, r=-0.524). However, we could not show any correlation between spermatogonial numbers and the severity of SCD indicated by the total number of pain crises per year, the number of acute thorax syndromes or SCD severity scores (Table 1; Figure 2E and F). This absence of correlation may be due to the limited

Figure 2. The impact of hydroxyurea dose, age at hydroxyurea therapy initiation and the number of pain crises as disease severity marker on the numbers of spermatogonia per round tubular cross-section and fertility index in patients with sickle cell disease. Graphical display of the numbers of spermatogonia per round tubular cross section (S/T) and fertility index (FI) Z-score by hydroxyurea (HU) dose (A and B), age at HU therapy initiation (C and D) and the numbers of pain crises (E and F). In linear regression analysis, the age at HU therapy initiation reached statistical significance. P indicates P-value and r Spearman correlation coefficient.

haematologica | 2022; 107(4)

977

Letters to the Editor

number of patients enrolled. Altogether seven patients were under regular transfusion regimes and six patients received pre-operative transfusions prior to testicular biopsy (Table 1). With limited informative value, the serum ferritin showed no correlation with spermatogonial numbers suggesting that iron load plays no significant role in the testicular toxicity of this patient population. In summary, most boys with SCD in this study had spermatogonia within their testicular tissue although at reduced numbers. Importantly, an association between reduced spermatogonial numbers and the young age was found while boys older than 8.8 years showed spermatogonial numbers within the normal range. Reduced spermatogonial numbers were further shown to correlate with the age at HU initiation. An age of 2.4 years or less at initiation associated with severely decreased spermatogonial numbers suggesting that spermatogonia and/or testicular somatic cells may be especially sensitive to disturbances such as HU exposure or disease complications. The potential risk of further compromising the already limited fertility by early HU initiation, must be weighed against the overwhelming evidence that HU significantly decreases morbidity and increases the lifespan among SCD patients.2 For all boys that are mature enough, it is most important that they are offered sperm cryopreservation. Moreover, testicular function should be studied if HU therapy must be discontinued to increase our knowledge of possible testicular recovery. For prepubertal boys, it is essential that they are given the opportunity to cryopreserve immature testicular tissue prior to HSCT after appropriate counseling regarding potentially decreased spermatogonial numbers. Klara M. Benninghoven-Frey,1* Nina Neuhaus,1* Atte K. Lahtinen,2,3 Claudia Krallmann,1 Joana M.D. Portela,4 Andrea Jarisch,5 Verena Nordhoff,1 Armin Soave,6 Hajar A.M. Ba Omar,7 Mikael Sundin,8,9 Cecilia Langenskiöld,10 Sabine Kliesch,1 Jan-Bernd Stukenborg7 and Kirsi Jahnukainen7,11 1 Center of Reproductive Medicine and Andrology, University of Münster and University Clinic Münster, Münster, Germany; 2Applied Tumor Genomics Research Program, Faculty of Medicine, University of Helsinki, Helsinki, Finland; 3Department of Medical and Clinical Genetics / Medicum, Faculty of Medicine, University of Helsinki, Helsinki, Finland; 4Center for Reproductive Medicine, Research Institute Reproduction and Development, Amsterdam Universitair Medische Centra, University of Amsterdam, Amsterdam, the Netherlands; 5Division of Stem Cell Transplantation and Immunology, Department of Children and Adolescent Medicine, University Hospital Frankfurt, Johann Wolfgang Goethe University, Frankfurt am Main, Germany; 6Department of Urology, University Medical Center Hamburg-Eppendorf, Hamburg, Germany; 7NORDFERTIL Research Lab Stockholm, Department of Women’s and Children s Health, Karolinska Institutet, and Karolinska University Hospital, Stockholm, Sweden; 8Division of Pediatrics, Department of Clinical Science, Intervention and Technology, Karolinska Institutet, Stockholm, Sweden; 9Section of Hematology, Immunology and HCT, Astrid Lindgren Children’s Hospital, Karolinska University Hospital, Stockholm, Sweden; 10Department of Pediatric Oncology, The Queen Silvia Children’s Hospital, Gothenburg, Sweden and 11New Children’s Hospital, Pediatric Research Center, University of Helsinki and Helsinki University Hospital, Helsinki, Finland *KMBF and NN contributed equally as co-first authors Correspondence: KIRSI JAHNUKAINEN- Kirsi.Jahnukainen@ki.se doi:10.3324/haematol.2021.279253 978

Received: July 5, 2021. Accepted: December 7, 2021. Pre-published: December 9, 2021. Disclosures: no conflicts of interest to disclose. Contributions: KMBF collected and analyzed data, prepared figures, drafted the manuscript; NN and JBS prepared samples, set up the experimental design, collected, analyzed and interpreted data, and drafted the manuscript; AJ set up the experimental design and collected data; AKL, CL, HAMBO, JMDP, MS and VN collected data; CK and AS collected data and were responsible for clinical patient care, SK supervised clinical patient care, clinical and laboratory testing and documentation in the University Hospital of Mun̈ ster; KJ set up the experimental design, collected and interpreted data, and drafted the manuscript. All authors had significant intellectual contribution in reviewing the manuscript and approved the final article. Acknowledgments: the authors would like to thank Dr. E. Oliver (Karolinska Institutet, and Karolinska University Hospital, Stockholm, Sweden) for language editing. Moreover, the authors would like to thank all clinical units involved in the study, especially Prof. Dr. U. Kontny, Dr. L. Lassay, Prof. Dr. N. Wagner (RWTH Aachen, University Hospital, Aachen, Germany), Dr. B. Bernbeck, Prof. Dr. D. T. Schneider (Klinikum Dortmund, Dortmund, Germany), Prof. Dr. A. Borkhardt, Prof. Dr. R. Meisel, Dr. F. Schuster (Department of Pediatric Oncology, Hematology and Clinical Immunology, HeinrichHeine-University, Duesseldorf, Germany), Dr. M. Höfs, Prof. Dr. D. Reinhardt, Dr. S. Schönberger, B. Schwert (Essen University Hospital, Essen, Germany), Prof. Dr. P. Bader, Dr. A. Barnbrock (University Hospital Frankfurt, Frankfurt, Germany), Dr. M. Bleeke (Center for Obstetrics and Pediatrics, Sektion für Pädiatrische Stammzelltransplantation und Immunologie University Medical Center Hamburg-Eppendorf, Hamburg, Germany), Prof. Dr. M. Fisch (Department of Urology, University Medical Center HamburgEppendorf, Hamburg, Germany), Dr. M. Bleeke and Prof. Dr. I. Müller (Center for Obstetrics and Pediatrics, Sektion für Pädiatrische Stammzelltransplantation und Immunologie, University Medical Center Hamburg-Eppendorf, Hamburg, Germany), Prof. Dr. S. Rutkowski (Center for Obstetrics and Pediatrics, University Medical Center Hamburg-Eppendorf, Hamburg, Germany), PD Dr. A. Claviez (University Medical Center Schleswig-Holstein, Kiel, Germany), and PD Dr. T. Nüßlein (Gemeinschaftsklinikum Mittelrhein, Koblenz, Germany). This project was supported by J. Salzig and J. Dabel (Center of Reproductive Medicine and Andrology, University of Münster and University Clinic Münster, Münster, Germany), who provided excellent technical support. Funding:this study was supported by grants from The Swedish Childhood Cancer Foundation (PR2019-0123; TJ2020-0023; PR2015-0073, TJ2015-0046) (JBS, KJ), the Jane and Dan Olssons Foundation (2016-33) (JBS), the Finnish Cancer Society (KJ), the Finnish Foundation for Pediatric Research (KJ), the Magnus Bervalls Foundation (JBS), the Swedish Research Council (2018-03094; 201601296) (JBS, KJ), the Birgitta and Carl-Axel Rydbeck’s Research Grant for Pediatric Research (2020-00348; 2020-00335) (JBS, KJ), and the Väre Foundation for Pediatric Cancer Research (AKL, KJ). This work was also supported by the German Research Foundation Grand CRU326 (NN). Moreover, we thank the Faculty of Medicine of the University of Münster for the financial support of KMBF.

References 1. Piel FB, Simon IH, Gupta S, Weatherall DJ, Williams TN. Global burden of sickle cell anaemia in children under five, 2010–2050: modelling based on demographics, excess mortality, and interventions. PLoS Med. 2013;10(7):e1001484. 2. Yawn BP, Buchanan GR, Afenyi-Annan AN, et al. Management of sickle cell disease: summary of the 2014 evidence-based report by expert panel members. JAMA. 2014;312(10):1033-1048." 3. Walters MC, Hardy K, Edwards S, et al. Pulmonary, gonadal and central nervous system status after bone marrow transplantation for sickle cell disease. Biol Blood Marrow Transplant.

haematologica | 2022; 107(4)

Letters to the Editor

2010;16(2):263-272. 4. Goossens E, Jahnukainen K, Mitchell RT, et al. Fertility preservation in boys: recent developments and new insights †. Hum Reprod Open. 2020;2020(3):hoaa016. 5. Valli-Pulaski H, Peters KA, Gassei K, et al. Testicular tissue cryopreservation: 8 years of experience from a coordinated network of academic centers. Hum Reprod. 2019;34(6):966-977. 6. Neuhaus N, Schlatt S. Stem cell-based options to preserve male fertility. Science. 2019;363(6433):1283-1284. 7. Stukenborg JB, Alves-Lopes JP, Kurek M, et al. Spermatogonial quantity in human prepubertal testicular tissue collected for fertility preservation prior to potentially sterilizing therapy. Hum Reprod. 2018;33(9):1677-1683. 8. DeBaun MR. Hydroxyurea therapy contributes to infertility in adult men with sickle cell disease: a review. Expert Rev Hematol. 2014;7(6):767-773. 9. Gille AS, Pondarré C, Dalle JH, et al. Hydroxyurea does not affect the spermatogonial pool in prepubertal patients with sickle cell disease. Blood. 2021;137(6):856-859.

haematologica | 2022; 107(4)

10. Sebastiani P, Nolan VG, Baldwin CT, et al. A network model to predict the risk of death in sickle cell disease. Blood. 2007;110(7):27272735. 11. van den Tweel XW, van der Lee JH, Heijboer H, Peters M, Fijnvandraat K. Development and validation of a pediatric severity index for sickle cell patients. Am J Hematol. 2010;85(10):746-751. 12. Alves-Lopes JP, Soder O, Stukenborg JB. Testicular organoid generation by a novel in vitro three-layer gradient system. Biomaterials. 2017;130:76-89. 13. Albert S, Wistuba J, Eildermann K, et al. Comparative marker analysis after isolation and culture of testicular cells from the immature marmoset. Cells Tissues Organs. 2012;196(6):543-554. 14. Funke M, Yang Y, Lahtinen A, et al. Z-scores for comparative analyses of spermatogonial numbers throughout human development. Fertil Steril. 2021;116(3):713-720. 15. Quinn CT. Minireview: clinical severity in sickle cell disease: the challenges of definition and prognostication. Exp Biol Med (Maywood). 2016;241(7):679-688.

979

Letters to the Editor

Targeting B-cell maturation antigen increases sensitivity of multiple myeloma cells to MCL-1 inhibition Interfering with the mechanisms by which malignant plasma cells develop drug resistance is critical for preventing relapse in multiple myeloma (MM). Downregulation of target antigen is one of the escape mechanisms limiting the success of promising therapeutic approaches such as B-cell maturation antigen (BCMA)-targeted immunotherapy.1,2 Despite high response rates and depth of responses after treatment with BCMA-targeting agents, patients eventually relapse.1–3 In most cases, the residual MM cells persisting after BCMA-based immunotherapy express reduced BCMA levels,1,2 suggesting that eliminating this BCMAlow MM cell pool could significantly delay or prevent relapse. An optimal therapeutic approach may therefore involve sequential treatment strategies targeting the critical anti-apoptotic pathways in refractory MM cells selected right after BCMA-based immunotherapy. However, ex vivo identification of potentially relevant sequential treatment strategies involving culture of primary MM cells for several days, has been limited due to the low viability of malignant plasma cells outside the bone marrow niche. Here, we describe a reproducible culture system for primary MM cells based on a synthetic hydrogel where viability of primary myeloma cells is preserved even in the absence of additional stromal subsets. Drug screening of MM samples in this three-dimensional (3D) platform revealed a dynamic interplay between BCMA and MCL-1, showing that BCMAlow MM cells are highly sensitive to MCL-1 inhibitors (MCL1i), and that pretreatment with BCMA-blocking antibodies significantly increases MCL-1i efficacy in MM cells. The BCMA (TNFRSF17) surface receptor is selectively present on plasma cells, and its expression is significantly higher in high- versus low-risk MM and in relapsed/refractory versus newly diagnosed patients.4 This progressive BCMA increase is due to loss of cells expressing a relatively lower level of this receptor,4 suggesting that malignant cells with highest BCMA expression may have a selective advantage over the course of the disease. Multiple BCMA-targeting approaches are being tested in the clinic for the treatment of MM, including antibody-drug conjugates, bispecific BCMAxCD3 antibodies, and anti-BCMA chimeric antigen receptor (CAR) T cells. 1–3 Signaling through the APRIL-BCMA axis promotes MM cell proliferation and survival, and induces the expression of the anti-apoptotic proteins BCL-2 and MCL-1.5 MCL-1 is a key pro-survival factor for healthy and malignant plasma cells, and elevated MCL-1 expression in MM is associated with chemoresistance and shorter event-free survival.6 Several compounds targeting MCL-1 are currently tested in clinical trials, including the potent MCL-1 inhibitor (MCL1i) S63845.7 MM cell sensitivity to MCL-1i treatment is highly variable between patients,8 as is observed with BCMA-targeting immunotherapy,2,9 stressing the need for primary MM cell-based ex vivo studies on the factors conditioning the response to these treatment modalities. Our culture approach for patient-derived MM cells is based on Puramatrix (PMX) hydrogel, which has previously been validated for MM cell co-culture with mesenchymal stromal cells (MSC).10 A major advantage of PMX over other gels such as Matrigel or fibrin scaffolds is its chemically-defined composition (R-A-D-A repeats), which eliminates batch-to-batch variability issues. MM cells (patient information is listed in the Online 980

Supplementary Table S1) were cultured in PMX supplemented with the prosurvival cytokines interleukin 6 (IL6) and APRIL, which are found at high concentrations in the bone marrow and plasma of MM patients.5 The combination of both cytokines preserved MM cell viability significantly better than in unstimulated controls (Online Supplementary Figure S1A). Primary MM cell survival in PMX was significantly higher than in two-dimensional (2D) cultures, as determined by cell viability percentages and by absolute cell number (Figure 1A to C). Compound diffusion in PMX was evaluated by exposing PMX-cultured MM cells to molecules added to the supernatant. In antibody-dependent cellular cytotoxicity (ADCC) and complement-dependent cytotoxicity (CDC) assays with the anti-CD38 antibody Daratumumab, surface-bound anti-CD38 was detected in all cells (Online Supplementary Figure S1B), and MM cell line sensitivity to CD38 targeting in PMX was comparable to that in 2D culture (Online Supplementary Figure S1C), showing that both small and large molecules diffuse efficiently in this hydrogel. Importantly, MM cells retained expression of plasma cell-associated markers when cultured in PMX. BCMA expression significantly increased over time, while CD38 levels remained mostly stable (Figure 1D and data not shown). Due to limited plasma cell viability in 2D culture, previous reports testing ex vivo chemosensitivity of primary MM cells are based on their co-culture with either MSC or MS-5 cells. 8,10,11 In order to relate to these previous studies, we compared MM drug responses in PMX supplemented with either cytokines or MSC. Specific apoptosis induced by MCL-1i in PMX + IL6/APRIL was similar to that measured in PMX + MSC co-cultures (Figure 1E). Furthermore, it has been reported that primary MM cells with a t(11;14) translocation are highly sensitive to BCL-2 inhibition,12 and we confirmed this observation in PMX-based drug screenings (Online Supplementary Figure S1D). Over culture in the presence of IL-6 and APRIL, sensitivity of MM cells to MCL-1i decreased significantly (Figure 1F), evocative of drug sensitivity loss in the physiological niche. Both IL-6 and APRIL have been linked to acquisition of drug resistance in MM, while previous reports rely mostly on the study of cell lines.5,13,14 Next, we assessed BCMA levels on MM cells that persisted after each treatment. Remarkably, BCMA expression on malignant cells that survived MCL-1 inhibition was significantly higher than on cells left untreated or exposed to other agents (Figure 2A). Distribution of BCMA expression in viable MM cells indicated that MCL-1 inhibition preferentially spares BCMAhi cells (Figure 2B). Taken together, these data reveal a connection between BCMA expression and dependence on MCL-1, and indicates that MCL-1i treatment eliminates BCMAlow MM cells. This observation is especially relevant considering that residual MM cells remaining after BCMA CAR-T cell therapy show significantly reduced BCMA levels, suggesting immune selection for BCMAdim/negative clonal variants.1,2 Our data indicates that combining MCL-1 inhibitors with BCMA-targeting immunotherapies may increase their efficacy by depleting residual BCMAlow cells. We next addressed the relation between BCMA blockade and MCL-1i-induced apoptosis by treating primary MM samples with a BCMA-targeting antibody (antiBCMA). MM cell sensitivity to single-agent anti-BCMA was heterogeneous and did not correlate with BCMA expression in MM cells before treatment (Online Supplementary Figure S2A and B). Importantly, there was a clear inverse correlation between MCL-1i and antihaematologica | 2022; 107(4)

Letters to the Editor

BCMA sensitivity in MM samples: MCL-1i-resistant MM cells were highly sensitive to anti-BCMA, and vice versa (Figure 3A). We did not observe changes in MCL-1, BCL-2 or BCL-XL protein levels after anti-BCMA treat-

ment, neither by fluorescence-activated cell sorting (FACS) nor by western blot analysis (Figure 3B; Online Supplementary Figure S2C and D). Considering that MCL1i spares MM cells with highest BCMA expression,

Figure 1. Primary multiple myeloma cell viability, phenotype, and drug sensitivity testing after culture in Puramatrix hydrogel. (A) Representative flow cytometry plots of multiple myeloma (MM) cells after culture. MM bone marrow mononuclear cells were seeded in Puramatrix (PMX) in the presence of IL-6 and APRIL (100 ng/mL each), for 7 days. All experiments involving primary MM cell culture were performed in this setting, unless otherwise stated. MM cells were gated as CD38+ CD138+ (left), and viable cells were identified as DiOC6+ TOPRO3- cells (right). (B) Frequency and (C) absolute cell number (expressed as foldincrease relative to two-dimensional [2D] controls) of MM cells after 7 days in culture in 2D or PMX (n=18). Bars indicate mean + standard error of the mean. (D) Mean fluorescence intensity (MFI) of Bcell maturation antigen (BCMA) in MM cells was measured by flow cytometry after 2 or 7 days in PMX culture (n=12). (E) Simple linear regression analysis comparing specific apoptosis (%) induced by the MCL-1 inhibitor (MCL-1i) S63845, in n=12 MM samples cultured in PMX supplemented with either IL-6 + APRIL or mesenchymal stromal cells (MSC, 80,000/well 10). MCL-1i (1,000 nM) was added on day 1, and MM viability was measured by flow cytometry 24 hours later. Specific apoptosis (%) was calculated by applying the following formula: [(%viable Nil - %viable treated)/ %viable Nil] x 100. (F) Specific apoptosis induced by MCL-1i in PMX-cultured MM cells (n=12). MCL-1i (100 nM) was added on day 1 or 6, and viability was measured 24 hours later (day 2 or 7, respectively). Each dot represents an individual sample. Statistical differences between 2 groups were analyzed using paired t-tests. *P<0.05; **P<0.01.

haematologica | 2022; 107(4)

Figure 2. BCMAlow multiple myeloma cells are sensitive to MCL-1 inhibition. (A) Mean fluorescence intensity (MFI) of B-cell maturation antigen (BCMA) (expressed as x-fold of MFI in untreated controls) as measured by flow cytometry in alive multiple myeloma (MM) cells after treatment with either 4 nM Bortezomib (Bor), 1,000 nM dexamethasone (Dex), 100 nM BCL-2 inhibitor (BCL-2i; ABT-199), or 100 nM MCL-1i (S63845) for 24 hours (n=4-9). Each dot represents an individual sample. Statistical differences between groups were analyzed using a one-way ANOVA with Bonferronis’ multiple comparison test. **P<0.01. (B) Representative histograms showing BCMA expression in MM cells from 4 different primary samples (MM1-MM4) cultured in Puramatrix (PMX). Histograms show BCMA in alive MM cells untreated (black) or after a 24-hour treatment with 100 nM MCL-1i (red).

981

Letters to the Editor

which may largely rely on BCMA signaling for survival, we next evaluated if co-treatment with MCL-1i and antiBCMA has synergistic effects on MM cell apoptosis. In five of seven primary MM samples tested, MCL-1i + anti-BCMA combination had a more than additive effect on MM cell killing (Figure 3C and D), but we did not observe consistent synergy when combining these drugs with primary MM samples. Next, we addressed whether increased MCL-1i efficacy in a context of BCMA blockade could be further enhanced by following a sequential treatment strategy. To this end, primary MM cells were

cultured in the presence of anti-BCMA for 6 days before a 24-h culture with MCL-1i. Strikingly, apoptosis induced by MCL-1 inhibition was significantly higher after a 6-day pretreatment with anti-BCMA, as compared to treatment with an isotype control or a shorter (24 hours) BCMA blockade period (Figure 3E and F). Increased MM cell susceptibility towards sequential BCMA and MCL-1 targeting was also observed following co-culture with MSC in PMX (Online Supplementary Figure S2E to H). Pretreatment with either Daratumumab (anti-CD38) or tocilizumab (anti-IL6R) did not sensitize

Figure 3. Pretreatment with anti-BCMA maturation antigen increases MCL-1 inhibitor efficacy in multiple myeloma cells. (A) Simple linear regression analysis comparing specific apoptosis after 24-hour treatment with either 100 nM MCL-1i or 5 ug/mL anti-B-cell maturation antigen (anti-BCMA) (a-BCMA; Vicky-1) (n=12). (B) Flow cytometry analysis of MCL-1 (left) and BCL-2 (right) expression in primary multiple myeloma (MM) cells cultured for 2 or 7 days with IL-6 + APRIL (100 ng/mL each) in the presence of 5 ug/mL isotype control antibody (black) or anti-BCMA (grey). Mean + standard error of the mean (n=3). (C) Specific apoptosis induced by 100 nM MCL-1i (black), 5 ug/mL anti-BCMA (grey), or their combination (blue) in 7 primary MM samples cultured in Puramatrix (PMX). Drugs were added on day 6, and MM cell viability was measured 24 hours later. (D) Plots comparing expected (EXP) to observed (OBS) specific apoptosis induced by combining anti-BCMA 5 ug/mL and MCL-1i (100 nM) for 24h (n=7). Hypothetical expected (EXP) specific apoptosis assumes an additive effect of the two combined drugs, and was calculated by using the formula: [(apoptosis drug A + apoptosis drug B) – (apoptosis drug A x apoptosis drug B)/100]. In 5 of 7 samples, this combination showed a more than additive pro-apoptotic effect. (A and D) Each dot represents an individual sample. (E) Representative flow cytometry plots showing the proportion of alive MM cells (CD38+ TOPRO3-) after the indicated treatments. MM samples were cultured for 6 days in the presence of either 5 ug/mL isotype control antibody (black), anti-BCMA (red), or no antibody (blue). After this time, cells were treated with 100 nM MCL-1i (black, red) or co-treated with MCL-1i + anti-BCMA (blue). MM viability was analyzed 24 hours later (i.e., on day 7) by fluorescence-activated cell sorting. (F) Cumulative plots showing apoptosis induced by MCL-1i in the conditions specified in (E). Mean + standard error of the mean (n=5). Statistical differences between 2 groups (D) were analyzed using paired t-tests. Statistical differences between 3 or more groups (F) were analyzed using a one-way ANOVA with Bonferronis’ multiple comparison test. *P<0.05; ** P<0.01.

982

haematologica | 2022; 107(4)

Letters to the Editor

MM cells to MCL-1 inhibition, suggesting that increased sensitivity to MCL-1i is specifically related to blockade of the APRIL-BCMA axis and not a general consequence of ADCC (Online Supplementary Figure S2I and J). Limiting the potential toxicities associated with MCL1i therapy is important: MCL-1 is expressed on different healthy tissues, including cardiomyocytes, hematopoietic stem cells, oocytes, and lymphocytes.15 It is, therefore relevant to find strategies to enhance the efficacy of MCL-1 inhibitors specifically in MM cells, which may lower the required doses for a persistent therapeutic effect. Our results suggest that blocking the APRILBCMA axis may render MM cells more dependent on pro-survival IL-6 signaling,5 forcing a scenario where cells are more sensitive to MCL-1 inhibition. Interestingly, patients with 1q21 amplification are highly sensitive to MCL-1 targeting, likely due to higher relative MCL1 expression resulting from amplification of 1q21.8 The IL-6 receptor (IL-6R) locus is also located in this chromosomal region, suggesting that 1q21+ MM cells may largely rely on the IL-6 – MCL-1 axis for survival. Taken together, our data reveals the potential of MCL-1 inhibition as a strategy to eliminate BCMAlow residual cells following BCMA-directed immunotherapy, and shows that anti-BCMA (pre-)treatment significantly enhances MCL-1i-induced apoptosis in MM cells. Marta Cuenca,1 Niels van Nieuwenhuijzen,1,2 Laura M. Moesbergen,1 Andries Bloem,3 Monique C. Minnema2 and Victor Peperzak1 1 Center for Translational Immunology, University Medical Center Utrecht, Utrecht University; 2Department of Hematology, University Medical Center Utrecht, Utrecht University and 3Central Diagnostic Laboratory, University Medical Center Utrecht, Utrecht, the Netherlands Correspondence: VICTOR PEPERZAK-: v.peperzak@umcutrecht.nl doi:10.3324/haematol.2021.279517 Received: June 26, 2021. Accepted: December 7, 2021. Pre-published: December 9, 2021. Disclosures: VP received royalty payments related to venetoclax. MCM received research funding from Celgene and honoraria from Celgene, Alnylam, BMS, Janssen-Cilag and Gilead. The remaining authors declare no competing interests. Contributions: MC and VP designed the research; MC, NvN and LMM performed the experiments and analyzed the results; AB provided resources; MC, NvN, AB, MCM and VP contributed to interpretation and discussion; MC and VP wrote and revised the manuscript; VP supervised the study. All authors critically reviewed and approved the final manuscript. Acknowledgements: the authors would like to thank the support facilities of the University Medical Center Utrecht (UMCU) and the

haematologica | 2022; 107(4)

Dutch Parelsnoer Institute for providing MM bone marrow samples. We are grateful to D. van den Blink, C. Steenhuis, N.J.G. WissingBlokland, and M.J.M. Dijkstra-Boerkamp from the Central Diagnostic Laboratory (CDL) of the UMCU.We thank S.J. Vastert for providing Tocilizumab. We thank L. Abbink for her help with ADCC and CDC assays. We thank Servier for providing the MCL-1-specific inhibitor S63845. We thank R. Raijmakers, M. Jak, T. Kimman and all VP laboratory members for helpful discussions. Funding: this investigation was supported by a Bas Mulder Award to VP from the Dutch Cancer Foundation (KWF)/Alped’HuZes foundation (award no. UU 2015-7663) and a project grant to VP from the Dutch Cancer Foundation (KWF)/Alped’HuZes foundation (grant no. 11108). MC was supported in part by a postdoctoral grant from the Ramón Areces Foundation.

References 1. Brudno JN, Maric I, Hartman SD, et al. T cells genetically modified to express an anti–B-Cell maturation antigen chimeric antigen receptor cause remissions of poor-prognosis relapsed multiple myeloma. J Clin Oncol. 2018;36(22):2267-2280. 2. Cohen AD, Garfall AL, Stadtmauer EA, et al. B cell maturation antigen–specific CAR T cells are clinically active in multiple myeloma. J Clin Invest. 2019;129(6):2210-2221. 3. Raje N, Berdeja J, Lin Y, et al. Anti-BCMA CAR T-cell therapy bb2121 in relapsed or refractory multiple myeloma. N Engl J Med. 2019;380(18):1726-1737. 4. Seckinger A, Delgado JA, Moser S, et al. Target expression, generation, preclinical activity, and pharmacokinetics of the BCMA-T cell bispecific antibody EM801 for multiple myeloma treatment. Cancer Cell. 2017;31(3):396-410. 5. Moreaux J, Legouffe E, Jourdan E, et al. BAFF and APRIL protect myeloma cells from apoptosis induced by interleukin 6 deprivation and dexamethasone. Blood. 2004;103(8):3148-3157. 6. Wuillème-Toumi S, Robillard N, Gomez P, et al. Mcl-1 is overexpressed in multiple myeloma and associated with relapse and shorter survival. Leukemia. 2005;19(7):1248-1252. 7. Kotschy A, Szlavik Z, Murray J, et al. The MCL1 inhibitor S63845 is tolerable and effective in diverse cancer models. Nature. 2016;538(7626):477-482. 8. Slomp A, Moesbergen LM, Gong JN, et al. Multiple myeloma with 1q21 amplification is highly sensitive to MCL-1 targeting. Blood Adv. 2019;3(24):4202-4214. 9. Lonial S, Lee HC, Badros A, et al. Belantamab mafodotin for relapsed or refractory multiple myeloma (DREAMM-2): a two-arm, randomised, open-label, phase 2 study. Lancet Oncol. 2020;21(2): 207-221. 10. Jakubikova J, Cholujova D, Hideshima T, et al. A novel 3D mesenchymal stem cell model of the multiple myeloma bone marrow niche: biologic and clinical applications. Oncotarget. 2016;7(47): 77326-77341. 11. Braham MVJ, Minnema MC, Aarts T, et al. Cellular immunotherapy on primary multiple myeloma expanded in a 3D bone marrow niche model. Oncoimmunology. 2018;7(6):e1434465. 12. Kumar S, Kaufman JL, Gasparetto C, et al. Efficacy of venetoclax as targeted therapy for relapsed/refractory t(11;14) multiple myeloma. Blood. 2017;130(22):2401-2409. 13. Tai YT, Acharya C, An G, et al. APRIL and BCMA promote human multiple myeloma growth and immunosuppression in the bone marrow microenvironment. Blood. 2016;127(25):3225-3236. 14. Ogiya D, Liu J, Ohguchi H, et al. The JAK-STAT pathway regulates CD38 on myeloma cells in the bone marrow microenvironment: therapeutic implications. Blood. 2020;136(20):2334-2345. 15. Lin VS, Xu ZF, Huang DCS, Thijssen R. Bh3 mimetics for the treatment of b-cell malignancies—insights and lessons from the clinic. Cancers. 2020;12(11):3353.

983

Letters to the Editor

Phase Ib dose-escalation study of the selective, noncovalent, reversible Bruton’s tyrosine kinase inhibitor vecabrutinib in B-cell malignancies Several Bruton’s tyrosine kinase inhibitors (BTKi) have been approved for the treatment of B-cell malignancies and are particularly active in chronic lymphocytic leukemia (CLL), where they have transformed the treatment paradigm. However, the activity of currently available BTKi (ibrutinib, acalabrutinib, and zanubrutinib) requires covalent bond formation with cysteine 481 (C481) of BTK; hence resistance to covalent BTKi may be mediated through mutations which remove C481. In order to overcome this resistance, and potentially prevent the proliferation of C481 mutant cells, non-covalent BTKi that do not require interaction with the C481 residue were developed.1 This phase Ib/II trial investigated the safety and clinical activity of vecabrutinib, a reversible, non-covalent inhibitor of BTK that inhibits both wild-type and C481-mutated BTK,2 in patients with advanced, BTKiresistant B-cell malignancies. The results of the completed phase Ib dose-escalation portion of the study demonstrated that vecabrutinib was well-tolerated up to 410 mg twice daily (BID), the highest dose studied. Evidence of clinical benefit was observed, with a best response of partial response (PR) in one CLL patient and stable disease (SD) in 13 patients. Pharmacokinetics (PK) were approximately dose-proportional and sustained reductions in serum cytokine concentrations were observed at higher dose levels, suggesting BTK inhibition. However, the association between vecabrutinib dose, pharmacodynamics (PD), and clinical activity was inconsistent and the activity observed was considered insufficient for phase II expansion in patients with BTKi-resistant CLL. Patient and disease characteristics are summarized in the Online Supplementary Table S1. Thirty-nine patients with histologically confirmed, relapsed/refractory CLL or other B-cell malignancies and ≥2 prior lines of standard systemic therapy, including progression during BTKi therapy, were enrolled and treated across seven dose-levels. The majority (77%) of patients had CLL. The enrolled population was high-risk, with a median of four prior therapies (range, 29), 17p deletion or TP53 mutation in 74% of all patients, and BTK C481 mutations in 55% of CLL patients. Vecabrutinib capsules were administered orally BID at a starting dose of 20.5 mg/dose (41 mg total daily dose). Vecabrutinib was well-tolerated up to 410 mg BID (820 mg total daily dose), the highest dose studied. One patient, treated at the 41 mg BID dose level, experienced a doselimiting toxicity (DLT), consisting of failure to receive >80% of planned vecabrutinib doses due to adverse events (AE) of grade 3 alanine aminotransferase (ALT) and grade 2 aspartate aminotransferase (AST) elevations. Upon expansion of this cohort, no additional DLT were observed and dose escalation continued. The maximum-tolerated dose of vecabrutinib was not reached. The most common treatment-emergent AE were anemia (31%) and nausea, fatigue, headache and dyspnea (21% each). The most common AE considered treatmentrelated by the investigator were anemia and fatigue (10% each). Grade ≥3 AE were mainly hematologic, including anemia (23%), neutropenia (13%) and thrombocytopenia (10%) (Online Supplementary Table S2). Grade ≥3 AE considered treatment-related by the investigator consisted of leukocytosis in two patients (5.1%) and anemia, neutropenia, and increased AST in one patient (2.6%) each. No obvious pattern of dose-dependent toxicity was observed, with no grade ≥3 AE observed at the two highest dose lev984

els. One or more serious AE (SAE) were reported in seven patients and consisted of cellulitis (in 2 patients) and lymphocytosis, intestinal perforation, myelitis, sepsis, hematuria, pleural effusion, and deep vein thrombosis (in 1 patient each). No SAE were considered related to study treatment per the investigator. Two patient deaths were associated with AE: perforated bowel in one patient treated at the 41 mg BID dose level (who had mantle cell lymphoma [MCL] with bowel involvement), and sepsis in one patient treated at the 164 mg BID dose level. Neither event was considered related to study treatment. There were no cardiac events or clinically significant electrocardiogram findings. Vecabrutinib showed modest evidence of clinical benefit, with one PR observed in a patient with CLL (treated at the 246 mg BID dose level) and SD in 13 patients (31%; 11 CLL, 1 MCL, 1 marginal zone lymphoma). A waterfall plot of percent change in tumor burden from baseline in all patients is displayed in Figure 1. Among 14 patients with a best response of PR or SD, six had BTK C481 mutations, eight had 17p deletion and/or TP53 mutations, and the median number of prior therapies was three (range, 2-9). Patients with PR or SD remained on study treatment for a median of 28.5 weeks (range, 6.1-56+ weeks). Eight of these patients received vecabrutinib for ≥6 months, including five who discontinued treatment due to termination of the study and not for progressive disease, suggesting some durable clinical benefit. Pharmacokinetic data were available for 38 patients. Concentration-time profiles and linear regression analysis indicated that both exposure and median steady-state Ctrough concentrations generally increased in an approximately dose-proportional manner. Exposure to vecabrutinib was maintained across the ~12 hour dosing interval, supporting BID dosing, with cycle 1 day 8 trough values at dose levels ≥164 mg BID expected to provide >90% inhibition of BTK signaling based an earlier single-dose phase I study.3 Pharmacodynamic activity of vecabrutinib was assessed in CLL patients who completed cycle 1 (n=25) via measurement of serum cytokine levels. CCL3, CCL4, and TNFa have previously been shown to be inhibited by other BTKi.4–6 For all three cytokines, a sustained reduction in serum concentration was evident in most patients after one cycle of vecabrutinib treatment at higher dose levels (246, 328 and 410 mg), suggesting inhibition of BTK activity. Mean reductions at these dose levels ranged from 3462% for CCL3, 33-59% for CCL4, and 24-57% for TNFa (Figure 2). In regression analyses, there was a trend for greater reduction in serum cytokine levels with increasing Cmax and AUClast as well as vecabrutinib dose, suggestive of an exposure-response effect (Online Supplementary Figure S1). Chemokine reduction was associated with clinical benefit, with decreased serum cytokine levels demonstrated in all but one patient with a clinical response of PR or SD. However, the extent of inhibition was generally less than that observed in BTKi-naïve patients treated with ibrutinib (which produced median decreases ≥80% for CCL3, CCL4 and TNFa),6 consistent with the limited clinical activity observed for vecabrutinib. The question arises as to why BTK inhibition by vecabrutinib did not translate to a clinical response despite strong preclinical evidence and promising early (phase Ia) clinical PK/PD data in healthy subjects, particularly in CLL patients in whom substantial clinical activity has been observed with other non-covalent BTKi. In vitro cellular assays were performed to evaluate the half maximal inhibitory concentration (IC50) values and BTK residence time (i.e., the time a BTKi remains bound to BTK) for haematologica | 2022; 107(4)

Letters to the Editor

Figure 1. Percent change in tumor burden from baseline in patients treated with vecabrutinib. Percent change in tumor burden (sum of the product of the diameters [SPD]) from baseline at time of best response assessment is shown by patient for all patients who underwent post-treatment imaging-based disease assessment. Disease type, dose (in mg twice daily [BID]), response assessment per the Investigator, baseline molecular characteristics (Bruton’s tyrosine kinase [BTK] C481X mutation status, presence of PLCg2 mutation, complex karyotype, TP53 mutation, or 17p deletion), and number of prior regimens received are indicated for each patient below the graph. U: indicates unknown

vecabrutinib compared with other BTKi to look for potential correlation with outcome (Table 1). IC50 values for vecabrutinib (18.4 nM) and ARQ 531 (32.9 nM) were similar, however, ARQ 531 demonstrated greater clinical efficacy; conversely, fenebrutinib demonstrated greater in vitro potency (7.04 nM) but showed limited clinical activity in patients previously treated with ibrutinib.7–10 The residence time observed for vecabrutinib (15 minutes [min]) was much shorter than that observed for ARQ 531 (128 min) and fenebrutinib (557 min) and was also shorter relative to reported values for pirtobrutinib (LOXO-305; 314 min). Other possible explanations for the limited clinical activity observed are that vecabrutinib is highly protein bound (98.7%), which may have affected the availability of the free drug, or that vecabrutinib may not have been consistently distributed from blood to disease sites; either of these possibilities may have resulted in levels insufficient to provide adequate BTK inhibition. Furthermore, PK properties differ among non-covalent BTKi: the effective agents, pirtobrutinib and ARQ 531, have longer half-lives (approximately 20 hours and 55 hours, respectively) than the agents with limited clinical activity, vecabrutinib, fenehaematologica | 2022; 107(4)

brutinib and dasatinib (reported half-lives ranging between 4 and 14 hours).3,7,9–12 Although no single property aligned consistently with the observed clinical activity, these attributes provide possible explanations regarding the limited clinical activity observed with vecabrutinib compared to other reversible BTK inhibitors. Overall, vecabrutinib was well-tolerated and demonstrated some evidence of clinical benefit. However, despite dose-proportional PK, the association between vecabrutinib dose, PD, and clinical response was inconsistent. Increasing the dose from 246 to 410 mg BID did not uniformly correlate with increased PD activity though there was an overall trend towards improved inhibition with dose. Assessment of clinical activity by dose may have been confounded by the impact of baseline patient characteristics: clinical benefit (i.e., PR or SD lasting >6 months) was most commonly observed in patients who were less heavily pretreated and had better prognostic factors as identified by Ahn et al.,13 such as lower baseline lactate dehydrogenase levels and wild-type TP53, regardless of dose. These results suggest that the potency of singleagent vecabrutinib was not sufficient to control disease in 985

Letters to the Editor

Figure 2. Dose-pharmacodynamics relationship for CCL3, CCL4, and TNFa. Serum cytokine levels at baseline (cycle [C]1, day [D]1) and after C1 of treatment (predose on C2D1 [day 29]) were measured by enzyme-linked immunosorbant assay in 25 chronic lymphocytic leukemia (CLL) patients who completed C1. Dot plots show mean change (with standard deviation) from baseline in CCL3 (A), CCL4 (B), and TNFa (C) serum levels. BID: twice daily.

Table 1. In vitro assessment of Bruton’s tyrosine kinase (BTK) residence time and half maximal inhibitory concentration values for BTK engagement for vecabrutinib and other BTK inhibitors.

Vecabrutinib BTK WT IC50 for BTK engagement, nM BTK residence time, minutes BTK C481S IC50 for BTK engagement, nM BTK residence time, minutes

Ibrutinib

ARQ 531

Fenebrutinib

Dasatinib

Pirtobrutinib

18.4 15

1.65 >1,000

32.9 128

7.04 557

34.8 61

3.7a 314a

34.6 16

229 31

102 228

13.1 900

78.8 120

8.5a 231a

BTK: Bruton’s tyrosine kinase; BTKi: BTK inhibitor; IC50: half-maximal inhibitory concentration. aIC50 values and residence time values (calculated as the reciprocal of kd [1/kd]) for pirtobrutinib (LOXO-305) are as reported by Gomez et al.15; WT: wild-type.

refractory patients; however vecabrutinib in combination with other agents, including BCL2 inhibitors,14 may result in improved efficacy. Based on the dose-escalation results, the activity observed in BTKi-resistant patients at the dose levels studied was considered insufficient for phase II expansion of this patient cohort. Future directions for vecabrutinib may include indications such as chronic graftversus-host disease or in combination with chimeric antigen receptor T-cell therapies, where dual inhibition of BTK and IL-2 Inducible T-cell Kinase (ITK) may contribute to clinical outcomes. John N. Allan,1 Javier Pinilla-Ibarz,2 Douglas E. Gladstone,3 Krish Patel,4 Jeff P. Sharman,5 William G. Wierda,6 Michael Y. Choi,7 Susan M. O’Brien,8 Mazyar 986

Shadman,9 Matthew S. Davids,10 John M. Pagel,4 Habte A. Yimer,11 Renee Ward,12 Gary Acton,12 Pietro Taverna,12 Daniel L. Combs,13 Judith A. Fox,12 Richard R. Furman1 and Jennifer R. Brown10 1 Weill Cornell Medicine, Department of Medicine, New York, NY; 2Moffitt Cancer Center, Tampa, FL; 3Johns Hopkins Sidney Kimmel Comprehensive Cancer Center, Baltimore, MD; 4Swedish Cancer Institute, Seattle, WA; 5Willamette Valley Cancer Institute/US Oncology, Eugene, OR; 6MD Anderson Cancer Center, Houston, TX; 7Moores Cancer Center, University of California San Diego, La Jolla, CA; 8Chao Family Comprehensive Cancer Center, University of California Irvine, Orange, CA; 9Fred Hutchinson Cancer Research Center, Seattle, WA; 10CLL Center, Medical Oncology, Dana-Farber Cancer Institute, Boston, MA; 11 Texas Oncology/US Oncology Research, Tyler, TX; 12Sunesis haematologica | 2022; 107(4)

Letters to the Editor

Pharmaceuticals, South San Francisco, CA and 13Combs Consulting Service, Mountain View, CA, USA Correspondence: JENNIFER R. BROWN - jennifer_brown@dfci.harvard.edu doi:10.3324/haematol.2021.280061 Received: September 24, 2021. Accepted: December 7, 2021. Pre-published: December 23, 2021. Disclosures: JNA served as a consultant to AbbVie, Acerta, Ascentage Pharma, AstraZeneca, Bayer, BeiGene, Epizyme, Genentech, Janssen, Pharmacyclics, Sunesis Pharmaceuticals, TG Therapeutics, and Verastem Oncology; received honoraria from AbbVie, BeiGene, Janssen, Pharmacyclics; and received research funding from AstraZeneca, Celgene, Genentech, and Janssen. JP-I served as a consultant to AbbVie, Bristol-Meyers Squibb, Janssen, Novartis, Takeda, Teva, and TG Therapeutics; and served on the speakers bureau for AbbVie, Bayer, Janssen, Sanofi, and Takeda. DEG reports no conflict of interest. KP served as a consultant to AstraZeneca, Celgene, Genentech, Pharmacyclics/Janssen and Sunesis; received research funding from AstraZeneca; and was a member of the Speakers Bureau for AstraZeneca, Celgene, Genentech, and Pharmacyclics/Janssen. JPS served as a consultant to and received honoraria and research funding from AbbVie, Acerta, AstraZeneca, Genentech, Janssen, Pharmacyclics, and TG Therapeutics. WGW received research funding from AbbVie, Acerta, Cyclcel, Genentech, Gilead Sciences, GSK/Novartis’ Janssen, Juno Therapeutics, KITE pharma, Loxo Oncology, Miragen, Oncternal Therapeutics, Pharmacyclics, Sunesis, and Xencor. MYC served as a consultant to AbbVie, Genentech, Gilead, Pharmacyclics, and Rigel; received research funding from AbbVie, Oncternal Therapeutics, Pharmacyclics, and Rigel; and was a member of the Speakers Bureau for AbbVie, Genentech, Gilead, and Pharmacyclics. SMO served as a consultant to AbbVie, Alexion, Amgen, Aptose Biosciences, Astellas, Celgene, Eisai, Gilead, GlaxoSmithKline, Janssen, Pfizer, Pharmacyclics, Sunesis, TG Therapeutics, Vaniam Group, and Verastem Oncology; received honoraria from AbbVie, Janssen, and Pfizer; and received research funding from Acerta, Gilead, KITE pharma, Pfizer, Pharmacyclics, Regeneron, Sunesis, and TG Therapeutics. MS served as a consultant to AbbVie, ADC Therapeutics, AstraZeneca, Atara Biotherapeutics, Genentech, Gilead, Pharmacyclics, Sound Biologics, and Verastem; and received research funding from AbbVie, Acerta Pharma, BeiGene, Celgene, Genentech, Gilead, Mustang Bio, Pharmacyclics, Sunesis, TG Therapeutics, and Verastem. MSD served as a consultant to AbbVie, Adaptive Biotechnologies, Ascentage Pharma, AstraZeneca, BeiGene, BMS, Celgene, Eli Lilly, Genentech, Janssen, MEI Pharma, Merck, Novartis, Takeda, and Verastem; received research funding from Ascentage Pharma, AstraZeneca, BMS, Genentech, MEI pharma, Novartis, Pharmacyclics, Surface Oncology, TG Therapeutics, and Verastem; and received honoraria from Research to Practice. JMP served as a consultant to AstraZeneca, Gilead Sciences, and Pharmacyclics. HAY was a member of the Speakers Bureau for Amgen, AstraZeneca, BeiGene, Janssen, Karyopharm, Pharmacylics, and Sanofi; and holds publicly-traded stock in Karyopharm. RW served as a consultant to Sunesis Pharmaceuticals. GA served as a consultant to Sunesis Pharmaceuticals. PT was employed by Sunesis Pharmaceuticals. JAF was employed by Sunesis Pharmaceuticals. DLC served as a consultant to Sunesis Pharmaceuticals. RRF served as a consultant to AbbVie, Acerta Pharma, AstraZeneca, BeiGene, Genentech, Incyte, Janssen, Loxo Oncology, Morphosys, Oncotracker, Pharmacyclics, Sanofi, Sunesis, TG Therapeutics, and Verastem. JRB served as a consultant to AbbVie, Acerta Pharma, AstraZeneca, BeiGene, Catapult Therapeutics, Dynamo, Genentech, Gilead, Juno/Celgene, Kite, Loxo, Novartis, Octapharma, Pfizer, Pharmacyclics, Sunesis, TG Therapeutics, and Verastem; received honoraria from Janssen and Teva, received research funding from Gilead, Kite, Loxo, Sun Pharmaceuticals, and Verastem; and was a haematologica | 2022; 107(4)

member of the data safetymonitoring board for Ivectys and Morphosys. Contributions: JNA, JP-I, DEG, KP, JPS, WGW, MYC, SMO, MS, MSD, JMP, HAY, RW, GA, PT, JAF, RRF, and JRB contributed to study concept and study supervision; PT and JAF were responsible for study administration; JNA, JP-I, DEG, KP, JPS, WGW, MYC, SMO, MS, MSD, JMP, HAY, RRF, and JRB provided patients and conducted the investigation; DC did a formal analysis of the data; DC, PT, and JAF prepared the data presentation; JA, GA, PT, JAF, and JRB wrote the original draft of the manuscript. All authors reviewed and edited the manuscript. Acknowledgments: medical writing support was provided by Janis Leonoudakis, PhD, and was funded by Sunesis Pharmaceuticals, Inc., San Francisco, CA. Funding: financial support for this study was provided by Sunesis Pharmaceuticals, Inc., South San Francisco, CA. Trial registration: clinicaltrials.gov Identifier: NCT03037645; EU clinical trials register identifier: EudraCT # 2018-000108-41

References 1. Gu D, Tang H, Wu J, Li J, Miao Y. Targeting Bruton tyrosine kinase using non-covalent inhibitors in B cell malignancies. J Hematol Oncol. 2021;14(1):40. 2. Allan JA, Patel K, Mato AR, et al. Preliminary results of a phase 1b/2 dose-escalation and cohort-expansion study of the noncovalent, reversible Bruton’s tyrosine kinase inhibitor (BTKi), vecabrutinib, in B-cell malignancies. HemaSphere. 2019;3(suppl 1):520. 3. Neuman L, Ward R, Arnold D, et al. First-in-human phase 1a study of the safety, pharmacokinetics, and pharmacodynamics of the noncovalent bruton tyrosine kinase (BTK) inhibitor SNS-062 in healthy subjects. Blood. 2016;128(Suppl):S2032. 4. Ponader S, Chen S-S, Buggy JJ, et al. The Bruton tyrosine kinase inhibitor PCI-32765 thwarts chronic lymphocytic leukemia cell survival and tissue homing in vitro and in vivo. Blood. 2012;119(5):1182-1189. 5. Byrd JC, Harrington B, O’Brien S, et al. Acalabrutinib (ACP-196) in relapsed chronic lymphocytic leukemia. N Engl J Med. 2016;374(4):323-332. 6. Niemann CU, Herman SEM, Maric I, et al. Disruption of in vivo chronic lymphocytic leukemia tumor-microenvironment interactions by ibrutinib - findings from an investigator-initiated phase II study. Clin Cancer Res. 2016;22(7):1572-1582. 7. Byrd JC, Smith S, Wagner-Johnston N, et al. First-in-human phase 1 study of the BTK inhibitor GDC-0853 in relapsed or refractory Bcell NHL and CLL. Oncotarget. 2018;9(16):13023-13035. 8. Crawford JJ, Johnson AR, Misner DL, et al. Discovery of GDC-0853: a potent, selective, and noncovalent Bruton’s tyrosine kinase inhibitor in early clinical development. J Med Chem. 2018;61(6):2227-2245. 9. Herman AE, Chinn LW, Kotwal SG, et al. Safety, pharmacokinetics, and pharmacodynamics in healthy volunteers treated with GDC0853, a selective reversible Bruton’s tyrosine kinase inhibitor. Clin Pharmacol Ther. 2018;103(6):1020-1028. 10. Woyach J, Stephens DM, Flinn IW, et al. Final results of phase 1, dose escalation study evaluating ARQ 531 in patients with relapsed or refractory B-cell lymphoid malignancies. Blood. 2019;134(Suppl 1):S4298. 11. Christopher LJ, Cui D, Wu C, et al. Metabolism and disposition of dasatinib after oral administration to humans. Drug Metab Dispos. 2008;36(7):1357-1364. 12. Mato AR, Shah NN, Jurczak W, et al. Pirtobrutinib in relapsed or refractory B-cell malignancies (BRUIN): a phase 1/2 study. Lancet. 2021;397(10277):892-901. 13. Ahn IE, Tian X, Ipe D, et al. Prediction of outcome in patients with chronic lymphocytic leukemia treated with ibrutinib: development and validation of a four-factor prognostic model. J Clin Oncol. 2021;39(6):576-585. 14. Jebaraj B, Müller A, Dheenadayalan R, et al. Evaluation of vecabrutinib as a model for non-covalent BTK/ITK inhibition for treatment of chronic lymphocytic leukemia. Blood. 2022;139(6):859-875. 15. Gomez EB, Isabel L, Rosendahal MS, Rothenberg SM, Andrews SW, Brandhuber BJ. Loxo-305, a highly selective and non-covalent next generation BTK inhibitor, inhibits diverse BTK C481 substitution mutations. Blood. 2019;135(Suppl 1):S4644.

987

Letters to the Editor

In vitro and in vivo effects of short-term cold storage of platelets in PAS-C Cold (4°C)-stored platelets (CSP) were the standard of care in the 1960s and 1970s but fell out of favor when their short in vivo survival was discovered. Since then, room temperature-stored platelets (RSP) have been the standard of care. Septic transfusion reactions from bacterially contaminated RSP remain the most common transfusiontransmitted infection. In addition, accumulating data questioning the efficacy and safety of RSP, together with a short shelf life, highlight an unmet medical need for an alternative product. Currently, CSP are being re-evaluated for bleeding patients, for whom immediate function matters more than long circulation time. An added benefit of CSP is that bacterial growth is markedly reduced at 4°C. However, how to store CSP best is poorly understood. Some groups reported increased wastage due to aggregates in CSP stored in plasma. Platelet additive solutions (PAS) were developed to reduce transfusion reactions and limit metabolic damage, but one group showed that PAS prevented aggregates in CSP.1 Other groups reported decreas-

ing platelet counts despite the use of PAS during cold storage, suggesting persistent microaggregates.2-4 While numerous in vitro studies on CSP in PAS exist,4-8 only one study looked at the effect of PAS on in vivo kinetics of transfused CSP, but lacked fresh comparators.9 In the current study, we investigated CSP in PAS (PASCSP) and compared them to CSP in plasma (P-CSP) and room temperature-stored platelets in plasma (P-RSP). All platelets were stored for 5 days. We collected a standard single apheresis platelet unit from six healthy subjects stored in either 100% plasma at 22°C or 4°C, or 65% PASC (Intersol) and 35% plasma at 4°C. In this study, we included historical controls for P-CSP and P-RSP,10,11 but all units were collected by apheresis and stored in the same fashion as described below. Concerns regarding risks for healthy human volunteers and their safety, in addition to the high costs of in vivo radiolabeling studies, made repeating control groups that have already been studied and published both redundant and ethically burdensome for this small study. We included PAS-C because it is currently licensed in the USA and we previously obtained in vivo data that favored PAS-C over PAS-F (Isoplate) for CSP.9 All three groups

988

Figure 1. In vitro platelet characteristics. Platelets stored at room temperature in plasma (P-RSP) (Plasma, 22°C, solid black bars), platelets stored cold in plasma (P-CSP) (Plasma, 4°C, striped bars), and platelets stored cold in platelet additive solution (PAS-CSP) (PAS-C, 4°C, solid gray bars) were tested fresh and on day 5 after storage. (A) Platelet count measured by an ABX Hemanalyzer, P-RSP (n=18), P-CSP (n=5), PAS-CSP (n=6). (B) Glucose measured by a blood gas analyzer, P-RSP (n=6), P-CSP (n=5), PAS-CSP (n=6). (C) Lactate measured by a blood gas analyzer, P-RSP (n=6), P-CSP (n=5), PAS-CSP (n=6). Data are shown as percentage of fresh and as mean + standard error of mean. *P<0.05, **P<0.01, ***P<0.001. ns = not significant

Figure 2. In vivo platelet characteristics. Healthy human subjects received autologous radiolabeled fresh platelets or platelets stored for 5 days in plasma at room temperature (P-RSP) (Plasma, 22°C, solid black bars), in plasma at 4°C (P-CSP) (Plasma, 4°C, striped bars), or in plasma additive solution at 4°C (PAS-CSP) (PAS-C, 4°C, solid gray bars). (A) Recovery of transfused platelets after 2 hours, P-RSP (n=21), P-CSP (n=5), PAS-CSP (n=5). (B) Survival of transfused platelets, P-RSP (n=21), P-CSP (n=5), PAS-CSP (n=5). Data shown as percentage of fresh, mean + standard error of mean. ***P<0.001, ns = not significant

haematologica | 2022; 107(4)

Letters to the Editor

(PAS-CSP, P-CSP, and P-RSP) comprised different cohorts (no matching between groups). The annexin V and Pselectin data were obtained from a separate group of four volunteers whose platelets were collected by apheresis and stored in mini-storage bags to clarify a role of these platelet activation parameters. All apheresis units were collected and bags stored at room temperature were agitated as per standard blood banking protocols. Cold-stored units were stored without agitation. We radiolabeled platelets as previously described with minor modifications.12 The Western Institutional Review Board approved the research, and all human participants gave written informed consent. The study was conducted in accordance with the Declaration of Helsinki and registered with ClinicalTrials.gov identifier NCT02754414. We assessed statistical significance by one way analysis of variance (ANOVA) with the Tukey correction for multiple comparisons. To minimize biological variability, and following recommendations by Murphy et al., we present the normalized stored data (percentage of fresh values). The absolute data are shown in Online Supplementary Figures S1-S3. As previously described, CSP counts were significantly lower than P-RSP counts. PAS-CSP counts were significantly higher than those of P-CSP, corroborating findings from others and our group (Figure 1A).1,11 Consistent with high metabolic activity at room temperature, glucose levels at day 5 were lowest in P-RSP. Post-storage levels of glucose in PAS-CSP were significantly lower than in P-CSP (Figure 1B). While P-RSP showed the highest lactate levels, cold storage reduced lactate production, with a trend for lower levels in PAS-CSP than in P-CSP (Figure 1C).

These findings hint at persistent, detectable metabolic activity up to 5 days, even though the metabolism is markedly slowed at 4°C. As expected, platelet in vivo markers decreased significantly, but the recovery of PASCSP and P-CSP did not differ significantly (Figure 2A). We observed a trend for longer survival in PAS-CSP than in PCSP (Figure 2B). To obtain more insights into the biological health of stored platelets, we studied mitochondrial membrane potential as an early marker of apoptosis. We observed significantly better-preserved membrane potential in P-CSP than in P-RSP. PAS-CSP and P-CSP values did not differ significantly as percentage of fresh values, but the absolute data showed significantly better preservation in PAS-CSP (Figure 3A, and Online Supplementary Figure S2D). All cells responded appropriately to the uncoupler CCCP (2-[2-(3-chlorophenyl)hydrazinylyidene]propanedinitrile) (Online Supplementary Figure S2C, D). There was a trend for more caspase activation in P-CSP, but overall, we did not see significant differences in this marker of late apoptosis (Figure 3B). Adding ABT 737 induced caspase activation before and after storage, indicating that platelets had the capacity to undergo apoptosis (Online Supplementary Figure S2E, F). We did not find significant differences in procoagulant activity, but there was a trend to higher levels at room temperature (Figure 3C). Similarly, we observed higher P-selectin levels at room temperature, but although this was significant when compared to PASCSP, it was not when compared to P-CSP (Figure 3D). Integrin activation was greatest in P-CSP, significantly more than in P-RSP. The PAS-CSP integrin response was lower than the P-CSP one, but the difference was not statistically significant (Figure 3E-G). The largest difference

Figure 3. In vitro apoptosis and activation parameters. Apoptosis and parameters of platelet activation were measured in fresh platelets and platelets stored for 5 days in plasma at room temperature (P-RSP) (Plasma, 22 °C, solid black bars), in plasma at 4°C (P-CSP) (Plasma, 4°C, striped bars), or in plasma additive solution at 4°C (PAS-CSP) (PAS-C, 4 °C, solid gray bars). (A) Platelet mitochondrial membrane potential measured by JC-1 dye red (FL2) to green (FL-1) ratio, PRSP (n=7), P-CSP (n=5), PAS-CSP (n=5). (B) Caspase 3,7 activation measured by flow cytometry, P-RSP (n=5), P-CSP (n=5), PAS-CSP (n=5). (C) Procoagulant activity measured by annexin V binding by flow cytometry. P-RSP (n=4), P-CSP (n=4), PAS-CSP (n=4). (D) a-granule degranulation was measured by CD62P binding by flow cytometry. P-RSP (n=4), P-CSP (n=4), PAS-CSP (n=4). Platelet aIIbb3 integrin activation was measured by PAC-1 antibody binding by flow cytometry at baseline (BL) and after stimulation (E) with collagen (C), P-RSP (n=5), P-CSP (n=7), PAS-CSP (n=5). (F) with arachidonic acid (AA), P-RSP (n=5), P-CSP (n=5), PAS-CSP (n=5). or (G) with ADP (A) P-RSP (n=5), P-CSP (n=5), PAS-CSP (n=5). (A-C) Data are shown as percentage of values for fresh platelets, mean + standard error of mean. (E-G) Absolute data, mean + standard error of mean. *P<0.05, **P<0.01, ***P≤0.001.

haematologica | 2022; 107(4)

989

Letters to the Editor

between P-CSP and PAS-CSP was after stimulation with arachidonic acid (Figure 3F). After only 5 days, we found that PAS prevented a coldinduced decrease in platelet count, likely by preventing microaggregates.1,9 One report suggests that aggregate formation is prevented by continuous agitation during cold storage.8 We observed a platelet count decrease independently of agitation in preliminary studies (Online Supplementary Figure S3C). In a previous study including over 20 units, we saw one large proteinaceous aggregate in one single unit, while a study with frequent manipulation and rewarming led to much more frequent detection of aggregates (Online Supplementary Figure S3D). Other investigators attempted to store cold platelets with repeated rewarming episodes (temperature cycling).13 To prevent aggregates, rewarming had to be accompanied by agitation, an approach we have thus far not incorporated in our studies. Metabolically active platelets under normal storage conditions convert most of the supernatant glucose into lactate in P-RSP. Replacing plasma with PAS-C removes sugars but adds acetate, which modifies glucose utilization and can suppress lactate generation in RSP. Accordingly, the level of lactate in PAS-CSP stored for 5 days was lower than that in P-CSP or P-RSP, indicating that replacing plasma with PAS had a beneficial effect.10,11 We did not observe activation differences between P-CSP and PAS-CSP after stimulation with various agonists similar to what has been described before for PAS/plasma CSP in aggregometry experiments1 and 100% plasma CSP.11 However, whether in vitro responses of CSP to agonists predict in vivo hemostasis is not well understood. Nevertheless, the fact that CSP in PAS and plasma have similar responses suggests that both media provide the right environment for platelet function testing in vitro. In our study, the mitochondrial membrane potential was best preserved in P-CSP and best predicted integrin activation, highlighting the need for an intact energy supply for platelet activation.10,11 It is likely that our 5-day storage time was not enough to induce later stages of apoptosis: a recent study of CSP in PAS-F did not find significant differences up to day 15, compared to baseline, in some markers of apoptosis.6 Our study provides a limited analysis of the complex apoptotic process and further studies are needed to explore this in further detail. Small sample sizes and donor-to-donor variability may also have affected the outcomes of this study. An additional limitation is the lack of an earlier testing time point before the 5-day maximum. PAS may have additional benefits by reducing allergic and febrile adverse reactions, although this has not been systematically studied with CSP. During storage, the levels of inflammatory mediators are lower in CSP than in RSP. Whether PAS further reduces inflammatory levels in CSP remains to be investigated, but data from RSP support this idea. Most European countries utilize pathogen reduction technology (PRT). Two previous studies looked into the effects of combining PAS/PRT with 4°C storage.14,15 The authors found mostly comparable results with the notable exception of reduced clot retraction and reduced GPIba-levels with PRT. There may also be a benefit in promoting coagulation with PRT. 14,15 In summary, we found that PAS had positive effects on in vitro parameters while not negatively affecting in vivo kinetics. Our data along with the results from other groups suggest that CSP in PAS are a safe and efficacious product and could have a practice-changing impact on the blood banking industry in the coming years. S. Lawrence Bailey,1 Lydia Y. Fang,1 Lynda Fitzpatrick,1 Daire Byrne,1 Esther Pellham1 and Moritz Stolla1,2 990

1 Bloodworks Northwest Research Institute, Seattle, WA and University of Washington Medical Center, Department of Medicine, Division of Hematology, Seattle, WA, USA Correspondence: MORITZ STOLLA - mstolla@bloodworksnw.org doi:10.3324/haematol.2021.279865 Received: August 25, 2021. Accepted: December 16, 2021. Pre-published: December 23, 2021. Disclosures: MS received research funding from Cerus Corp. and Terumo BCT Contributions: SLB analyzed the data, established assays, and performed experiments, EP, DB and LYF performed experiments. LF performed apheresis collection procedures. MS designed the study, analyzed the data and wrote the manuscript. Acknowledgments: the authors would like to thank Dr. Sherrill Slichter and the members of the cold-stored platelet interest group organized by the Department of Defense for helpful discussions. We thank Renetta Stevens and Tena Petersen for administrative support. Funding: this project received funding support from the Department of Defense, award n.. W81XWH-12-1-0441

References 1. Getz TM, Montgomery RK, Bynum JA, Aden JK, Pidcoke HF, Cap AP. Storage of platelets at 4 degrees C in platelet additive solutions prevents aggregate formation and preserves platelet functional responses. Transfusion. 2016;56(6):1320-1328. 2. Johnson L, Tan S, Wood B, Davis A, Marks DC. Refrigeration and cryopreservation of platelets differentially affect platelet metabolism and function: a comparison with conventional platelet storage conditions. Transfusion. 2016;56(7):1807-1818. 3. Johnson L, Vekariya S, Wood B, Tan S, Roan C, Marks DC. Refrigeration of apheresis platelets in platelet additive solution (PASE) supports in vitro platelet quality to maximize the shelf-life. Transfusion. 2021;61(Suppl 1):S58-S67. 4. Reddoch-Cardenas KM, Montgomery RK, Lafleur CB, Peltier GC, Bynum JA, Cap AP. Cold storage of platelets in platelet additive solution: an in vitro comparison of two Food and Drug Administration-approved collection and storage systems. Transfusion. 2018;58(7):1682-1688. 5. Marini I, Aurich K, Jouni R, et al. Cold storage of platelets in additive solution: the impact of residual plasma in apheresis platelet concentrates. Haematologica. 2019;104(1):207-214. 6. Reddoch-Cardenas KM, Peltier GC, Chance TC, et al. Cold storage of platelets in platelet additive solution maintains mitochondrial integrity by limiting initiation of apoptosis-mediated pathways. Transfusion. 2021;61(1):178-190. 7. Braathen H, Sivertsen J, Lunde THF, et al. In vitro quality and platelet function of cold and delayed cold storage of apheresis platelet concentrates in platelet additive solution for 21 days. Transfusion. 2019;59(8):2652-2661. 8. Hegde S, Wellendorf AM, Zheng Y, Cancelas JA. Antioxidant prevents clearance of hemostatically competent platelets after longterm cold storage. Transfusion. 2021;61(2):557-567. 9. Stolla M, Fitzpatrick L, Gettinger I, et al. In vivo viability of extended 4 degrees C-stored autologous apheresis platelets. Transfusion. 2018;58(10):2407-2413. 10. Zimring JC, Slichter S, Odem-Davis K, et al. Metabolites in stored platelets associated with platelet recoveries and survivals. Transfusion. 2016;56(8):1974-1983. 11. Stolla M, Bailey SL, Fang L, et al. Effects of storage time prolongation on in vivo and in vitro characteristics of 4 degrees C-stored platelets. Transfusion. 2020;60(3):613-621. 12. The Biomedical Excellence for Safer Transfusion (BEST) Collaborative. Platelet radiolabeling procedure. Transfusion. 2006;46(Suppl 3):59S-66S. 13. Vostal JG, Gelderman MP, Skripchenko A, et al. Temperature cycling during platelet cold storage improves in vivo recovery and survival in healthy volunteers. Transfusion. 2018;58(1):25-33. 14. Agey A, Reddoch-Cardenas K, McIntosh C, et al. Effects of Intercept pathogen reduction treatment on extended cold storage of apheresis platelets. Transfusion. 2020;61(1):167-177. 15. Six KR, Devloo R, Compernolle V, Feys HB. Impact of cold storage on platelets treated with Intercept pathogen inactivation. Transfusion. 2019;59(8):2662-2671.

haematologica | 2022; 107(4)

Letters to the Editor

Conventional interferon-a 2b versus hydroxyurea for newly-diagnosed patients with polycythemia vera in a real world setting: a retrospective study based on 286 patients from a single center Polycythemia vera (PV) is a myeloproliferative neoplasm (MPN) characterized by clonal proliferation of multipotent bone marrow progenitors.1 A clinical trial investigating the efficacy of pegylated interferon (IFN) for PV and essential thrombocytosis is ongoing in China. However, as only conventional IFN and hydroxyurea (HU) are covered by Chinese basic medical insurance, these cytoreductive agents are recommended as first-line treatment by the consensus of Chinese experts for the diagnosis and treatment of PV, including low-risk patients.2 As the difference in efficacy between conventional IFN and HU for newly-diagnosed PV is undefined, we retrospectively analyzed data of 286 newly-diagnosed PV patients who were treated at the Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Sciences between June 1, 2007 and February 28, 2020. All patients received conventional IFN-a 2b or HU for at least 6 months. Patients were excluded if they changed groups. The flowchart for patient selection is shown in the Online Supplementary Figure S1A. Cases were diagnosed in accordance with the 2016 World Health Organization diagnostic definitions.3 Conventional IFN-a 2b was recommended first for young (age <60 years old) patients and older patients without contraindications. HU was usually recommended for other patients. In total, 82 and 204 patients received single-agent conventional IFN-a 2b (IFN cohort) and single-agent HU (HU cohort), respectively. Generally, the initial dose of conventional IFN-a 2b was 3×106 IU three times per week; the initial dose of HU was 20 mg/kg/day. Treatment schedules were adjusted by monitoring peripheral blood counts with the target of hematocrit (HCT) <45%. Quantitative measurements of the JAK2 V617F variant allele frequency (VAF) were performed by real-time poly-

merase chain reaction (PCR) as previously described.4 Hematologic and molecular responses were evaluated in accordance with the revised response criteria of the European LeukemiaNet (ELN) and International Working Group-Myeloproliferative Neoplasms Research and Treatment (IWG-MRT).5 Complete hematologic remission (CHR) was defined as HCT <45% without phlebotomy, white blood cell (WBC) <10×109/L, and platelets ≤400×109/L. Complete molecular response (CMR) was defined as indetectable JAK2 V617F mutation. Partial molecular response (PMR) applied only to patients with a JAK2 V617F VAF ≥20% before treatment and was defined as a ≥50% decrease in allele burden after treatment.5 The clinical and laboratory features of subjects in the IFN and HU cohorts are displayed in Table 1. The median treatment duration in the IFN and HU cohorts were 51 months (interquartile range [IQR], 24–83 months) and 53 months (IQR, 31–84 months), respectively. The duration of exposure to IFN or HU for each patient is shown in the Online Supplementary Figure S1B and C. Compared with the HU cohort, a higher proportion of patients in the IFN cohort achieved CHR (65% vs. 43%; P=0.001), control of HCT (72% vs. 43%; P=0.06), control of platelets (88% vs. 77%; P=0.04), and control of WBC (89% vs. 72%; P=0.002; Figure 1A) during follow-up. A higher proportion of low-risk subjects who received IFN achieved CHR compared with those who received HU (64% vs. 32%; P=0.001; Figure 2A). Consistently, high-risk subjects who received IFN also had a higher CHR rate than those who received single-agent HU (68% vs. 47%; P=0.06; Figure 2C). Interestingly, the median increase in mean corpuscular volume (MCV) from baseline, when patients achieved the best control of HCT, was 21.2 fL (IQR, 9.1–31.9 fL) in the HU cohort, which was much higher than in the IFN cohort (3.9 fL; IQR, −2.8–9.5; P<0.001). Because HCT equals the erythrocyte count multiplied by MCV, some patients in the HU cohort might have been phlebotomized with normal RBC. The median duration from starting treatment to achiev-

Table 1. Clinical features of the interferon and hydroxyurea cohorts at baseline.

Variables Age, years Sex, female Palpable splenomegaly Disease duration, month; median (range) Baseline hemoglobin, g/L Baseline RBC, ×1012/L Baseline hematocrit, % Baseline WBC, ×109/L Baseline platelet, ×109/L Baseline MCV, fL JAK2 V617F mutation Baseline JAK2 V617F VAF, %, (n=209)* Abnormal cytogenetics, % (n/N) Thrombosis pretreatment (n=406) Thrombosis risk stratification (n=406) Low risk High risk Follow-up from start of treatment, months

IFN (N = 82)

HU (N = 204)

P-value

51 (44-57) 50 (61%) 20 (26%) 0 (0-2) 189 (177-209) 7.0 (6.3-7.6) 58 (54-63) 12.6 (9.4-15.1) 464 (339-623) 84.0 (80.7-89.6) 77 (94%) 56 (35-73） 4% (2/49) 23 (29%)

61 (52-67) 104 (51%) 52 (27%) 0 (0-2) 197 (187-210) 7.2 (6.5-7.8) 61 (57-65) 13.1 (9.8-18.1) 424 (324-572) 85.4 (79.8-89.9) 191 (94%) 59 (33-73) 4% (4/114) 69 (35%)

52 (65%) 28 (35%) 52 (35-91)

59 (30%) 141 (70%) 55 (33-84)

<0.001 0.13 0.83 0.31 0.01 0.02 0.003 0.15 0.40 0.99 0.93 0.62 1.00 0.36 <0.001 0.82

Data are presented as median (interquartile range [IQR]) or n (%), unless otherwise indicated. IFN: interferon; HU: hydroxyurea; RBC: red blood cell; WBC: white blood cell; MCV: median corpuscular volume; VAF: variant allele frequency; JAK2 V617F VAF in JAK2 V617F-mutated patients.

haematologica | 2022; 107(4)

991

Letters to the Editor

Figure 1. Comparison of hematologic and molecular responses between the interferon and hydroxyurea cohorts. (A) Hematologic responses, (B) complete hematologic remission (CHR) rates over time, (C) molecular responses, and (D) dynamics of JAK2 V617F variant allele frequencies (VAF) over time are compared between the interferon (IFN) and hydroxyuera (HU) cohorts. In (D) the horizontal lines indicate median values; bars represent minimum and maximum values; boxes represent values included between the 25% and 75% percentiles. (E) JAK2v V617F VAF waterfall plot in the IFN (n=22) and HU (n=31) cohorts; the y-axis indicates the absolute change of the JAK2 V617F VAF from baseline to the best molecular response; each bar represents a patient; dotted lines represent median changes of the JAK2 V617F VAF in each group. IFN: conventional (non-pegylated) interferon; HCT: hematocrit; PLT: platelet; WBC: white blood cell; PMR: partial molecular response; Pts: patients. *P<0.05; **P<0.01; ***P<0.001.

992

haematologica | 2022; 107(4)

Letters to the Editor

ing CHR was 11 months (IQR, 7–23 months) in the IFN cohort, which was shorter than in the HU cohort (19 months; IQR, 10–47 months; P=0.001). Among the patients who achieved CHR, two (4%) and seven (7%) were lost during follow-up in the IFN and HU cohorts, respectively. Compared with the HU cohort, CHR rates in the IFN cohort were higher throughout the treatment duration and became significantly better after 3 years of continuous treatment (88% vs. 44%; P<0.001; Figure 1B), which was consistent with the results of the PROUD-PV and CONTINUATION-PV studies, which used pegylated IFN.6 In addition to hematologic responses, the IFN cohort also showed better molecular responses than the HU cohort. In total, 31 and 40 patients had data regarding molecular responses in the IFN and HU cohorts, respectively. The median JAK2 V617F VAF at baseline were not significantly different between the IFN and HU cohorts (68% [IQR, 51–78%] vs. 62% [IQR, 40–70%]; P=0.21). Only one patient in the IFN cohort achieved CMR. The percentage of patients who obtained a JAK2 V617F VAF

<10% was higher in the IFN (65%) than in the HU (33%) cohort (P=0.007; Figure 1C). Among patients with baseline JAK2 VAFs ≥20%, 95% (19/20) achieved PMR in the IFN and 59% (17/29) in the HU cohorts (P=0.007; Figure 1C). The median change in JAK2 V617F VAF from baseline to the best molecular response in the IFN and HU cohorts was −58% (IQR, −69% to −34%) and −30% (IQR, −51% to –0.4%) (P=0.001; Figure 1E). Finally, the JAK2 V617F VAF in the IFN cohort was significantly lower than in the HU cohort after 3 years of continuous treatment (Figure 1D). Because the IFN cohort was younger than the HU cohort, we compared treatment responses between patients in the IFN and HU cohorts matched for age and sex. The baseline peripheral blood counts, JAK2 V617F allele burdens, follow-ups, and thrombosis risk stratifications were balanced between the two matched cohorts (Online Supplementary Figure S2A). The CHR rate (66% [44/67] vs. 34% [23/67]; P=0.001; Online Supplementary Figure S2B), control of HCT rate (73% [49/67] vs. 54% [36/67]; P=0.03; Online Supplementary Figure S2B), and

Figure 2. Comparison of hematologic and molecular responses between the interferon and hydroxyurea cohorts stratified by thrombosis risk. Hematologic (A) and molecular (B) responses of low-risk patients. Hematologic (C) and molecular (D) responses of high-risk patients. IFN: interferon; HU: hydroxyurea; RBC: red blood cell; WBC: white blood cell; HCT: hematocrit; PLT: platelet; VAF: variant allele frequency; IQR: interquartile range; *JAK2 V617F VAF in JAK2 V617F-mutated patients; CHR: complete hematologic remission; PMR: partial molecular response. *P<0.05; **P<0.01; ***P<0.001.

haematologica | 2022; 107(4)

993

Letters to the Editor

PMR rate (95% [18/19] vs. 62% [8/13]; P=0.029; Online Supplementary Figure S2C) were significantly higher in the IFN cohort than in the HU cohort when matched for age and sex. When patients were stratified by thrombosisrisk, the CHR rate in the IFN cohort was higher than in the HU cohort for age- and sex-matched low-risk (63% [24/38] vs. 26% [9/35]; P=0.002) and high-risk (70% [19/27] vs. 45% [14/31]; P=0.067) patients. In total, 14 of 82 subjects (17%) discontinued IFN treatment for the following reasons: normalized peripheral blood counts (n=8, 57%), adverse effects (n=2, 14%), disease progression (n=1, 7%), and unknown reasons (n=3, 21%). Fever was the most common adverse effect of IFN, which was reported in 23% (14/62) of patients, followed by bone pain in 11% (7/62) of patients. Post-treatment thrombotic events occurred in two (2%) and six (3%) patients in the IFN and HU cohorts. The thrombosis rates were 0.5% (95% confidence interval [CI]: 0–1.1) patients per year for the IFN cohort and 0.7% (95% CI: 0.2–1.2) for the HU cohort. These rates were much lower than those published in a previous study (2.62%; 95% CI: 2.34–2.94 patients per year).7 In our study, the thrombosis rate in the low-risk cohort (95% CI: 0.6% [0.2–0.9]) was lower than that of low-risk PV patients treated by phlebotomy, as reported by Barbui et al. (95% CI: 2.0% [1.5-2.5]).8 The lower incidence of thrombosis in this study compared with previous studies might be related to racial differences in thromboembolism between Asian and Western populations. This is related to differences in genetic polymorphisms and environmental factors, such as obesity and healthcare facilities.9,10 For instance, a study reported that Japanese patients with paroxysmal nocturnal hemoglobinuria (PNH) had a significantly lower incidence of thrombosis than American PNH patients.11 Moreover, the ECLAP study and a matched study of 951 patients with PV reported a benefit-risk profile of HU therapy over phlebotomy with respect to the lower rate of arterial thrombosis.12,13 Our findings suggested that early intervention with cytoreductive treatments for low-risk subjects rather than phlebotomy might also correlate with lower thrombosis rates. Thrombosis-free survival rates were not significantly different between the IFN and HU cohorts (P=0.81), similar results were found when adjusted by age (P=0.73; Online Supplementary Figure S3A). There was no significant difference in overall survival (P=0.99; Online Supplementary Figure S3B) or myelofibrosis-free survival (P=0.98; Online Supplementary Figure S3C) between the IFN and HU cohorts when adjusted by age. A previous retrospective study of PV patients reported that IFN reduced the risk of mortality and transformation into myelofibrosis compared with HU or phlebotomy.14 The different conclusions that we report might be due to the relatively short follow-up in our study. Finally, there was no significant difference in thrombosis-free survival (P=0.40), overall survival (P=0.55), or myelofibrosis-free survival (P=0.26) between patients who achieved PMR or not. A recent meta-analysis reported that CHR rates, thrombotic complications, and treatment discontinuations owing to adverse events were not significantly different between pegylated and conventional IFN.15 In our study, conventional IFN-a 2b was a good choice for PV, showing better efficacy than HU and acceptable tolerance. In conclusion, this study found that the hematologic and molecular responses of newly-diagnosed PV to conventional IFN-a 2b were better than to HU. There are limitations to this study, such as it being a retrospective study from a single center with a short follow-up, a mixture of 994

low- and high-risk patients, and only a few patients who were tested for molecular responses, which are all sources of potential bias. Dan Liu,1,2* Zefeng Xu,1,2* Peihong Zhang,1,3 Tiejun Qin,1,2 Xiujuan Sun,1 Shiqiang Qu,1,2 Lijuan Pan,1,2 Jiao Ma,1,3 Wenyu Cai,1,3 Jinqin Liu,1 Huijun Wang,1,3 Qi Sun,1,3 Zhongxun Shi,1,2 Huijun Huang,1,2 Gang Huang,4 Robert Peter Gale,5 Bing Li,1,2 Raajit K. Rampal 6 and Zhijian Xiao1,2,3 1 State Key Laboratory of Experimental Hematology, National Clinical Research Center for Blood Diseases, Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Tianjin, China; 2MDS and MPN Center, Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Tianjin, China; 3Hematologic Pathology Center, Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Tianjin, China; 4Divisions of Experimental Hematology and Cancer Biology, Cincinnati Children's Hospital Medical Center, Cincinnati, OH, USA; 5Hematology Section, Division of Experimental Medicine, Department of Medicine, Imperial College London, London, UK and 6 Leukemia Service, Department of Medicine, Memorial Sloan Kettering Cancer Center, New York, NY, USA. *DL and ZX contributed equally as co-first authors. Correspondence: BING LI - libing@ihcams.ac.cn RAAJIT K RAMPAL - rampalr@mskcc.org ZHIJIAN XIAO - zjxiao@ihcams.ac.cn doi:10.3324/haematol.2021.280080 Received: September 24, 2021. Accepted: December 24, 2021. Prepublished: January 6, 2022. Disclosures: RPG is a consultant to BeiGene Ltd., Fusion Pharma LLC, LaJollaNanoMedical Inc., Mingsight Pharmaceuticals Inc., and CStone Pharmaceuticals; is an advisor to Antegene Biotech LLC; medical director of FFF Enterprises Inc.; is a partner of AZAC Inc.; is a member of the Board of Directors of the Russian Foundation for Cancer Research Support; and member of the Scientific Advisory Board of StemRad Ltd. RKR has received grants and personal fees from Constellation, Incyte, Celgene/BMS, and Stemline, and personal fees from Promedior, CTI, Blueprint, Jazz Pharmaceuticals, Galecto, Pharmaessentia, AbbVie, and Novartis. All other authors declare no conflicts of interest. Contributions: ZJX designed the study; DL and ZFX collected and interpreted the data and performed statistical analysis; PHZ and QS contributed to analysis of bone marrow histology; TJQ, SQQ, LJP, WYC, JQL, HJW, XJS, MJ, and QYG contributed to recruiting patients and collecting data; DL wrote the manuscript with contributions from ZJX, ZFX, BL, GH, RPG, RKR, ZXS, and HJH. All authors reviewed the manuscript during its preparation and approved the final version of the manuscript. Funding: this study was supported in part by National Natural Science Funds (No. 81530008, 81870104, and 82070134), Tianjin Natural Science Funds (18JCZDJC34900, 16JCQNJC11400, and 19JCQNJC09400), and CAMS Initiative Fund for Medical Sciences (No. 2016-I2M-1-001, 2020-I2M-C&T-A-020, and 2020-I2M-C&T-B-090).

References 1. Grinfeld J, Nangalia J, Baxter EJ, et al. Classification and personalized prognosis in myeloproliferative neoplasms. N Engl J Med. 2018;379(15):1416-1430. 2. Leukemia and Lymphoma Group, Chinese Society of Hematology,

haematologica | 2022; 107(4)

Letters to the Editor

Chinese Medical Association. [Chinese expert consensus on the diagnosis and treatment of polycythemia vera (2016)]. Zhonghua Xue Ye Xue Za Zhi. 2016;4(4):265-268. 3. 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. 4. Tashi T, Swierczek S, Kim SJ, et al. Pegylated interferon Alfa-2a and hydroxyurea in polycythemia vera and essential thrombocythemia: differential cellular and molecular responses. Leukemia. 2018;32(8):1830-1833. 5. Barosi G, Mesa R, Finazzi G, et al. Revised response criteria for polycythemia vera and essential thrombocythemia: an ELN and IWG-MRT consensus project. Blood. 2013;121(23):4778-4781. 6. Gisslinger H, Klade C, Georgiev P, et al. Ropeginterferon alfa-2b versus standard therapy for polycythaemia vera (PROUD-PV and CONTINUATION-PV): a randomised, non-inferiority, phase 3 trial and its extension study. Lancet Haematol. 2020;7(3):e196-e208. 7. Tefferi A, Rumi E, Finazzi G, et al. Survival and prognosis among 1545 patients with contemporary polycythemia vera: an international study. Leukemia. 2013;27(9):1874-1881. 8. Barbui T, Vannucchi AM, Carobbio A, et al. The effect of arterial hypertension on thrombosis in low-risk polycythemia vera. Am J

haematologica | 2022; 107(4)

Hematol. 2017;92(1):E5-E6. 9. Hamasaki N, Kuma H, Tsuda H. Activated protein C anticoagulant system dysfunction and thrombophilia in Asia. Ann Lab Med. 2013;33(1):8-13. 10. Zakai NA, McClure LA. Racial differences in venous thromboembolism. J Thromb Haemost. 2011;9(10):1877-1882. 11. Nishimura JI, Kanakura Y, Ware RE, et al. Clinical course and flow cytometric analysis of paroxysmal nocturnal hemoglobinuria in the United States and Japan. Medicine (Baltimore). 2004;83(3):193-207. 12. Barbui T, Vannucchi AM, Finazzi G, et al. A reappraisal of the benefit-risk profile of hydroxyurea in polycythemia vera: a propensitymatched study. Am J Hematol. 2017;92(11):1131-1136. 13. Barbui T, De Stefano V, Ghirardi A, et al. Different effect of hydroxyurea and phlebotomy on prevention of arterial and venous thrombosis in polycythemia vera. Blood Cancer J. 2018;8(12):124. 14. Abu-Zeinah G, Krichevsky S, Cruz T, et al. Interferon-alpha for treating polycythemia vera yields improved myelofibrosis-free and overall survival. Leukemia. 2021;35(9):2592-2601. 15. Bewersdorf JP, Giri S, Wang R, et al. Interferon alpha therapy in essential thrombocythemia and polycythemia vera - a systematic review and meta-analysis. Leukemia. 2020;35(6):1643-1660.

995

Letters to the Editor

Daratumumab with or without chemotherapy in relapsed and refractory acute lymphoblastic leukemia. A retrospective observational Campus ALL study The anti-CD38 antibody daratumumab, currently approved for the treatment of patients with multiple myeloma, is also being explored for patients with acute lymphoblastic leukemia (ALL), whose blasts commonly express high levels of CD38.1 Patients with relapsed or refractory (R/R) disease, as well as those with positive measurable residual disease (MRD) have consistently shown unfavorable outcomes and, especially for T-lineage ALL, therapeutic options beyond first-line treatment remain limited. Preclinical studies have demonstrated that daratumumab has significant activity in human xenograft models of ALL, both as a single agent and in combination with chemotherapy.2-4 However, the clinical experience is so far very limited. A few case reports have suggested that daratumumab has anti-leukemic activity in R/R and MRD-positive ALL cases,5-10 but the small numbers of patients and the positive-outcome bias, due to the propensity to publish mainly positive results, prevent any robust conclusions on the clinical impact of this drug in ALL. We hereby report a retrospective, observational study performed in patients with R/R or MRD-positive ALL who received daratumumab in Italy in order to provide further information on the safety and efficacy of this drug in a real-life context. In this study we included adult and pediatric patients with R/R or MRD-positive T- or B-lineage ALL or lymphoblastic lymphoma, who received at least one dose of daratumumab between December 2018 and December 2020 at 17 Italian hematology centers. Data were retrospectively collected in an anonymous database. This study was performed in the context of the Campus ALL national framework and in agreement with the Declaration of Helsinki. Patients received daratumumab either off-label or in the context of a compassionate use program, kindly supported by Janssen-Cilag Spa. Daratumumab was administered at the approved schedule for multiple myeloma (i.e., 16 mg/kg weekly for 8 doses, then every 2 weeks for 8 doses, then monthly until disease progression), either alone or in combination with chemotherapy. The co-primary endpoints were overall response rate and overall survival of patients after daratumumab. Additional endpoints were safety and bridge to allogeneic hematopoietic cell transplant (HCT) after daratumumab. Complete response was defined as a bone marrow blast count <5% without evidence of extramedullary manifestations, partial response was defined as a bone marrow blast count ≥5% and <25% with a reduction of leukemic involvement of at least 50%. For lymphoblastic lymphoma, the Lugano criteria were applied. MRD was monitored either by flow cytometry and/or by real-time quantitative polymerase chain reaction in centralized laboratories. The overall response rate was defined as the proportion of patients who obtained a partial response, a complete response or, only for patients who were MRD positive, MRD negativity. Any systemic anti-neoplastic treatment started at diagnosis or with R/R disease counted as a line of therapy, except for allogeneic HCT. Survival was estimated using the Kaplan-Meier method and overall survival was calculated from the date of the first daratumumab infusion to the date of death or the last follow-up. The association of baseline variables with response was explored using the Fisher exact, c2 or 996

Mann-Whitney test, as appropriate. The cut-off date for this analysis was March 31, 2021 and data were analyzed with STATA 12.1 software (Stata Corporation, College Station, TX, USA). We included 20 patients (85% males) in the study. Thirteen had T-ALL, four had B-ALL, one had mixed phenotype acute leukemia and two had lymphoblastic lymphoma (B-lineage in 1, T-linage in the other); 11 patients were treated front-line according to the GIMEMA LAL1913 intensive pediatric-like protocol.11 The patients’ characteristics are summarized in Table 1. Daratumumab was administered at a median time of 13 months after diagnosis and patients had received a median of three prior lines of therapy. The median age of the patients at the start of daratumumab treatment was 35 years (range, 8-73) and three patients were below the age of 18. Nine patients had already undergone an allogeneic HCT, with a median time from transplantation to daratumumab treatment of 6 months (range, 2-20). At the start of daratumumab treatment, 80% of patients had a bone marrow relapse, either isolated or with concomiTable 1. Patients’ characteristics.

Variable

N. or median

Male sex 17 Type of ALL 18 T 13 ETP 4 B 4 MPAL 1 Type of LBL 2 T 1 B 1 Firstline treatment LAL-1913 11 Hyper-CVAD 3 AIEOP-BFM 4 Others (EWALL, BFM) 2 Allo-HCT before daratumumab 9 Lines of treatment before daratumumab 3 Age at daratumumab start, years 35 Below 18 years 3 Disease status at daratumumab start Isolated BM relapse 8 Extramedullary and BM 8 Extramedullary and MRD positivity 2 Extramedullary only 1 CR, MRD-positive 1 Disease characteristics at daratumumab start White blood cells, x109/L 3.34 Platelets, x109/L 34 Peripheral blood blasts, % 14 Bone marrow blasts, % 45 ECOG at daratumumab start 2 Concomitant chemotherapy 9 Time from diagnosis to daratumumab, months 13

% or range 85 80 65 20 20 5 10 5 5 55 15 20 10 45 1-4 8 - 73 15

40 40 10 5 5

0.1 - 39 1 - 233 0 - 98 0 - 100 0-4 45 7 - 28

ALL: acute lymphoblastic leukemia; ETP: early-T-precursor; MPAL: mixed phenotype acute leukemia; LBL: lymphoblastic lymphoma; allo-HCT: allogeneic hematopoietic cell transplant; BM: bone marrow; MRD: measurable residual disease; CR: complete response; ECOG: performance status according to the Eastern Cooperative Oncology Group scale.

haematologica | 2022; 107(4)

Letters to the Editor

tant extramedullary disease. Extramedullary sites involved were the lymph nodes in four patients, the central nervous system in three, the mediastinum in two, the breast in one and the gut in one. The median performance status according to the Eastern Cooperative Oncology Group (ECOG) was 2. Daratumumab was administered alone in 11 cases (in 1 case after a short dexamethasone pre-phase), while nine patients received concomitant chemotherapy (Online Supplementary Table S1). The overall response rate was 20%, with two patients achieving a MRD-negative complete response, one a complete response with persistent MRD and one a partial response (Table 2). Patients responded after two to six infusions of daratumumab and the median time to response was 4 weeks. Three of the four responses were observed in patients with T-ALL, who were treated with daratumumab as a single agent. Two patients (both with T-ALL) were alive at the last follow-up, one patient died after relapse and one died of treatment-related complications after allogeneic HCT. The characteristics of the responding patients are summarized in Online Supplementary Table S2. Four patients (2 responders, 2 refractory) proceeded to allogeneic HCT after daratumumab. Next, we explored the potential factors associated with response. Patients with a bone marrow hematologic relapse (P=0.013), lower platelet count (P=0.019) and higher circulating blast percentage (P=0.034) were less likely to respond, while those with a better ECOG performance status (P=0.019) and who had received fewer prior lines of therapy (P=0.022) responded better (Table 3). Consistently, bone marrow MRD positivity, with or without extramedullary involvement, was associated with a better overall response rate, without however the difference reaching statistical significance (P=0.088). Finally, we evaluated the potential association of CD38 expression on lymphoblasts with response. Among the 18 evaluable cases, CD38 positivity and mean fluorescence intensity did not differ significantly between responders and non-responders (median 96.5% vs. 95.6%, P=0.9 and 16,800 vs. 12,800; P=0.51, respectively) At the last follow-up, all but one patient had stopped treatment and two patients remained alive and in complete remission. The median overall survival of the whole cohort was 4 weeks, with a 3-month overall survival rate of 25% (Online Supplementary Figure S1). No unexpected toxicities were observed and there was only one grade 2 infusion reaction. Table 2. Outcome after daratumumab treatment.

Outcome

N. or median

% or range

4 2 1 1 16 2 12 2 4 2 2 19

20 10 5 5 80 10 60 10 20 10 2-120 95

Response to daratumumab Responders CR, MRD-negative CR, MRD-positive PR Non-responders Stable disease Progressive disease Not evaluable Allo-HCT post-daratumumab in CR/PR Treatment duration, weeks Discontinued treatment

CR: complete remission; MRD: measurable residual disease; PR: partial remission; Allo-HCT: allogeneic hematopoietic cell transplant.

haematologica | 2022; 107(4)

To our knowledge, this is the largest series of ALL patients treated with daratumumab reported so far and further underlines the potential activity of this drug in R/R and MRD-positive cases. While the advent of monoclonal antibodies and chimeric antigen receptor T-cell therapy is progressively changing the therapeutic scenario in B-lineage ALL, the approved treatment options for R/R and MRD-positive TALL remain unsatisfactory, as highlighted by the large prevalence of T-lineage diseases in our cohort (14/20 total patients). Nelarabine has been confirmed to be active in this setting,12 but responses are short-lived and half of the patients are resistant to the drug. More recently, the AKR1C activated prodrug OBI342413 and the BCL2 inhibitor venetoclax have been tested in R/R T-ALL, and the combination of venetoclax and navitoclax with chemotherapy appears particularly promising.14 However, data on these new agents are still immature and new therapeutic approaches are urgently needed. Following preclinical data and a few positive case reports, daratumumab has started to be used in patietns with advanced ALL without other therapeutic options, but data from unselected cohorts are lacking. In our series, in which we included all patients who received at

Table 3. Predictors of response to daratumumab.

Variable*

Number (%) or median (range) Responders Non-responders

Sex Male 4 (23.5) Female 0 (0) Age, years 34 (25 - 45) T-lineage 3 (21.4) B-lineage 1 (20) Lymphoblastic lymphoma 1 (50) Extramedullary disease 2 (18.2) BM MRD° 2 (66.7) BM relapse 1 (6.2) Previous allo-HCT 2 (22.2) Previous lines of treatment 1 2 (100) 2 1 (25) 3 1 (9.1) 4 0 (0) White blood cells, x109/L 3.36 (3 - 4.3) Hemoglobin, g/dL 10 (10 - 11) Platelets, x109/L 151 (70 - 233) PB blasts, % 0 (0 - 0) BM blasts, % 2 (0 - 78) ECOG score 0 2 (100) 1 2 (40) 2 0 3 0 4 0 Concomitant chemotherapy 1 (11.1)

13 (76.5) 3 (100) 35.5 (8 - 73) 11 (78.6) 4 (80) 1 (50) 9 (81.8) 1 (33.7) 15 (93.8) 7 (77.8) 0 (0) 3 (75) 10 (90.9) 3 (100) 4.66 (0.1 - 39.4) 9.5 (8 - 13) 27 (1 - 199) 24 (0 - 98) 50 (1 - 100) 0 (0) 3 (60) 4 (100) 7 (100) 1 (100) 8 (88.9)

P= 1

0.92 1 1 0.37 1 0.088 0.013 1 0.022

0.91 0.25 0.019 0.034 0.099 0.019

0.37

*Disease status and patients’ characteristics evaluated at the time of starting daratumumab therapy. °Includes the patient in complete remission with isolated measurable residual disease positivity and those with extramedullary relapse and measurable residual disease positivity in the bone marrow. Bold values denote statistical significance at the P<0.05 level. BM: bone marrow; MRD: measurable residual disease; Allo-HCT: allogeneic hematopoietic cell transplant; ECOG: performance status according to the Eastern Cooperative Oncology Group scale.

997

Letters to the Editor

least one dose of the drug, we observed a relatively low overall response rate of 20% and limited survival. However, most patients were heavily pre-treated, with a poor ECOG performance status and a high disease burden. Indeed, several of these patients would be excluded from any clinical trial. Responses were obtained rapidly, after two to six infusions of the drug, either alone or in combination with chemotherapy, and interestingly also in cases with extramedullary disease. Although limited by the small numbers, we could analyze potential predictive factors of response to daratumumab. We observed that patients with a high ALL burden (i.e., those with a bone marrow hematologic relapse and circulating blasts) were unlikely to benefit from the treatment, while daratumumab proved to be effective in patients with a good performance status and less advanced disease. These findings are in agreement with the current literature, with positive case reports mostly describing patients treated for MRD positivity or with low disease burden and in good clinical conditions. Indeed, a better selection of patients, as well as earlier use of the compound (e.g., in MRD-positive cases) appear crucial to obtain meaningful results. We also evaluated CD38 expression on lymphoblasts before the start of daratumumab therapy and found no significant association with response. Sample investigation was not centralized and so this exploratory analysis was limited by the heterogeneity of antibodies and analytical techniques employed by different flow cytometry laboratories. We observed a patient who responded to daratumumab despite high disease burden, but in this case the antibody was used in combination with chemotherapy. Recently, a case report outlined the feasibility of this approach,15 which is currently being tested in a clinical trial evaluating daratumumab in combination with chemotherapy in younger ALL patents (NCT03384654). This strategy might be the best option in the presence of a full-blown relapse, but the best chemotherapy regimen to combine with daratumumab remains to be defined. Combinations with innovative drugs with promising activity in ALL, such as venetoclax or bortezomib,16 could also be tested following the experience in multiple myeloma,17 as these patients are often chemorefractory. We could confirm the safety of daratumumab in the setting of ALL, with a lower than expected rate of infusion reactions compared to those occurring in multiple myeloma and no unexpected toxicities. Although limited by its retrospective nature and the heterogeneity of the patients, our study provided data that could help to design new clinical trials aimed at testing daratumumab in ALL and to selecting patients who may benefit more from its use. In conclusion, our data confirm the potential activity and safety of daratumumab in R/R and MRD-positive ALL, and suggest that this compound should be possibly used earlier, rather than after several lines of salvage treatment. Further studies are needed to clarify whether daratumumab could be another game-changer in this disease. Marco Cerrano,1 Massimiliano Bonifacio,2,3 Matteo Olivi,1,4 Antonio Curti,5 Michele Malagola,6 Michelina Dargenio,7 Anna Maria Scattolin,8 Cristina Papayannidis,5 Fabio Forghieri,9 Carmela Gurrieri,10 Ilaria Tanasi,2,3 Patrizia Zappasodi,11 Roberta La Starza,12 Nicola Stefano Fracchiolla,13 Patrizia Chiusolo,14,15 Luisa Giaccone,4,16 Maria Ilaria Del Principe,17 Fabio Giglio,18 Marzia Defina,19 Claudio Favre,20 Carmelo Rizzari,21 Barbara Castella,22 Giovanni Pizzolo,2,3 Felicetto Ferrara,23 Sabina Chiaretti24 and Robin Foà24 998

1 Divisione di Ematologia, Dipartimento di Oncologia, AOU Città della Salute e della Scienza, Torino; 2UOC di Ematologia, Azienda Ospedaliera Universitaria Integrata di Verona, Verona; 3Dipartimento di Medicina, Sezione di Ematologia, Università di Verona, Verona; 4 Dipartimento di Biotecnologie Molecolari e Scienze per la Salute, Università di Torino, Torino; 5IRCCS Azienda OspedalieroUniversitaria di Bologna, Istituto di Ematologia “Seràgnoli”, Bologna; 6 Università di Brescia, Unità Operativa di Ematologia e Centro Trapianti, ASST Spedali Civili di Brescia, Brescia; 7Unità Operativa di Ematologia e Trapianto, Ospedale Vito Fazzi, Lecce; 8Unità Operativa di Ematologia e Trapianto, Ospedale dell'Angelo, Mestre; 9 Unità Operativa Complessa di Ematologia, Azienda OspedalieroUniversitaria di Modena, Modena; 10UO Ematologia ed Immunologia Clinica, Azienda Ospedale-Università Padova, Padova; 11Unità Operativa di Ematologia, Fondazione IRCCS Policlinico San Matteo, Pavia; 12Unità Operativa di Ematologia e Trapianto, AOU-Ospedale S. Maria della Misericordia, Perugia; 13UOC Ematologia Fondazione IRCCS Ca' Granda - Ospedale Maggiore Policlinico, Milano; 14 Dipartimento di Diagnostica per Immagini, Radioterapia Oncologica ed Ematologia, Fondazione Policlinico Universitario A. Gemelli IRCCS, Roma; 15Sezione di Ematologia, Dipartimento di Scienze Radiologiche ed Ematologiche, Università Cattolica del Sacro Cuore, Roma; 16Unità di Trapianto Allogenico di Cellule Staminali, Dipartimento di Oncologia, A.O.U. Città della Salute e della Scienza, Torino; 17 Ematologia, Dipartimento di Biomedicina e Prevenzione, Università di Roma Tor Vergata, Roma; 18Unità Operativa di Ematologia e Trapianto di Midollo Osseo, IRCCS Ospedale San Raffaele, Milano; 19 Unità Operativa Complessa di Ematologia, AOUS, Università di Siena, Siena; 20Unità Operativa di Oncologia ed Ematologia Pediatrica, Ospedale Pediatrico Meyer, Firenze; 21Unità Operativa di Ematologia Pediatrica, Università di Milano-Bicocca, Fondazione MBBM, Monza; 22Laboratorio Analisi, AO S. Croce e Carle, Cuneo; 23 UOC di Ematologia, AORN Cardarelli, Napoli and 24Ematologia, Dipartimento di Medicina Traslazionale e di Precisione, Università Sapienza, Roma, Italy Correspondence: MARCO CERRANO - cerranomarco@gmail.com ROBIN FOÀ - rfoa@bce.uniroma1.it doi:10.3324/haematol.2021.279851 Received: September 6, 2021. Accepted: Junuary 4, 2022. Pre-published: January 13, 2022. Disclosures: no conflicts of interest to disclose Contributions: MC, MB, SC and RF designed the study. MC assembled and analyzed the data. MO collected data and contributed to their analysis. BC analyzed flow cytometry data. MC, MB, SC and RF drafted the manuscript. All authors contributed to data collection, revised the manuscript, and accepted its final version. Acknowledgments: we thank Janssen-Cilag Spa Italy for providing daratumumab for compassionate use. Funding: the work was partly supported by the Associazione Italiana per la Ricerca sul Cancro (AIRC), Metastases Special Program, N° 21198, Milan, Italy (to RF).

References 1. Leong S, Inglott S, Papaleonidopoulou F, et al. CD1a is rarely expressed in pediatric or adult relapsed/refractory T-ALL: implications for immunotherapy. Blood Adv. 2020;4(19):4665-4668. 2. Bride KL, Vincent TL, Im S-Y, et al. Preclinical efficacy of daratumumab in T-cell acute lymphoblastic leukemia. Blood. 2018;131(9):995999. 3. Vogiatzi F, Winterberg D, Lenk L, et al. Daratumumab eradicates minimal residual disease in a preclinical model of pediatric T-cell acute lymphoblastic leukemia. Blood. 2019;134(8):713-716. 4. Naik J, Themeli M, de Jong-Korlaar R, et al. CD38 as a therapeutic target for adult acute myeloid leukemia and T-cell acute lymphoblastic leukemia. Haematologica. 2019;104(3):e100-e103.

haematologica | 2022; 107(4)

Letters to the Editor

5. Ofran Y, Ringelstein-Harlev S, Slouzkey I, et al. Daratumumab for eradication of minimal residual disease in high-risk advanced relapse of T-cell/CD19/CD22-negative acute lymphoblastic leukemia. Leukemia. 2020;34(1):293-295. 6. Mirgh S, Ahmed R, Agrawal N, et al. Will daratumumab be the next game changer in early thymic precursor-acute lymphoblastic leukaemia? Br J Haematol. 2019;187(2):e33-e35. 7. Ganzel C, Kharit M, Duksin C, Rowe JM. Daratumumab for relapsed/refractory Philadelphia-positive acute lymphoblastic leukemia. Haematologica. 2018;103(10):e489-e490. 8. Zhang Y, Xue S, Liu F, Wang J. Daratumumab for quick and sustained remission in post-transplant relapsed/refractory acute lymphoblastic leukemia. Leuk Res. 2020;91:106332. 9. Cerrano M, Castella B, Lia G, et al. Immunomodulatory and clinical effects of daratumumab in T-cell acute lymphoblastic leukaemia. Br J Haematol. 2020;191(1):e28-e32. 10. Ruhayel SD, Valvi S. Daratumumab in T-cell acute lymphoblastic leukaemia: a case report and review of the literature. Pediatr Blood Cancer. 2021;68(5):e28829. 11. Chiaretti S, Messina M, Della Starza I, et al. Philadelphia-like acute lymphoblastic leukemia is associated with minimal residual disease persistence and poor outcome. First report of the minimal residual disease-oriented GIMEMA LAL1913. Haematologica. 2021;106(6):1559-1568. 12. Candoni A, Lazzarotto D, Ferrara F, et al. Nelarabine as salvage ther-

haematologica | 2022; 107(4)

apy and bridge to allogeneic stem cell transplant in 118 adult patients with relapsed/refractory T-cell acute lymphoblastic leukemia/lymphoma. A CAMPUS ALL study. Am J Hematol. 2020;95(12):14661472. 13. Evans K, Duan J, Pritchard T, et al. OBI-3424, a novel AKR1C3-activated prodrug, exhibits potent efficacy against preclinical models of T-ALL. Clin Cancer Res. 2019;25(14):4493-4503. 14. Pullarkat VA, Lacayo NJ, Jabbour E, et al. Venetoclax and navitoclax in combination with chemotherapy in patients with relapsed or refractory acute lymphoblastic leukemia and lymphoblastic lymphoma. Cancer Discov. 2021;11(6):1440-1453. 15. Fulcher J, Berardi P, Christou G, Villeneuve PJA, Bredeson C, Sabloff M. Nelarabine-containing regimen followed by daratumumab as an effective salvage therapy and bridge to allogeneic hematopoietic stem cell transplantation for primary refractory early T-cell precursor lymphoblastic leukemia. Leuk Lymphoma, 2021;62(9):2295-2297. 16. Kaspers GJL, Niewerth D, Wilhelm BAJ, et al. An effective modestly intensive re-induction regimen with bortezomib in relapsed or refractory paediatric acute lymphoblastic leukaemia. Br J Haematol, 2018;181(4):523-527. 17. Bahlis NJ, Baz R, Harrison SJ, et al. Phase I study of venetoclax plus daratumumab and dexamethasone, with or without bortezomib, in patients with relapsed or refractory multiple myeloma with and without t(11;14). J Clin Oncol. 2021;39(32):3602-3612.

999

Letters to the Editor

T-cell immune responses following vaccination with mRNA BNT162b2 against SARS-CoV-2 in patients with chronic lymphocytic leukemia: results from a prospective open-label clinical trial The Coronavirus disease 2019 (COVID-19) pandemic has seriously affected patients with chronic lymphocytic leukemia (CLL).1 Fatalities exceeding 30% has been reported among hospitalized patients in international surveys,2,3 and in consecutively identified CLL patients.4 Additionally, many patients with CLL do not achieve seroconversion upon mRNA vaccination.5,6 We recently confirmed those observations in the course of a prospective clinical trial involving the BNT1622b2 mRNA vaccine.7 The study included five equally sized cohorts of patients with different types of immunocompromised disorders (total n=449) and 90 healthy controls. Sixtythree percent of patients within the specific CLL cohort (n=90) seroconverted after two vaccine doses. The lowest seroconversion rate was found in patients on ibrutinib followed by those who had stopped Bruton’s tyrosine kinase inhibitor (BTKi)-therapy.7 Even though T-cell responses may occur in most patients with CLL following SARS-CoV-2 infection,4 there is still limited data on T-cell immunity following vaccination of many immunocompromised patient groups including CLL. In the first report on hematological malignancies, nine of 18 patients developed SARS-CoV-2-specific T-cell reactivity.8 Similar results were seen in a large patient group applying a whole blood interferon-g (IFN-g) release assay.9 An early study10 of T-cell responses identified IFN-g as a key cytokine produced by spike-specific CD4+ and CD8+ T cell in BNT162b1 mRNA-vaccinated individuals. We report here on T-cell immunity in patients with CLL from our above mentioned prospective clinical trial7 using IFN-g ELISpot, a validated quantitative assay to measure T-cell responses against SARS-CoV-2-specific peptides.4,11 Inclusion criteria and monitoring of our vaccine clinical trial has been described earlier.7 Patients with a previous history or signs of COVID-19, or who tested positive for SARS-CoV-2 or had spike protein-specific antibodies at baseline were excluded. No patient with CLL developed break-through infection during the study or early followup. Fifty-two predefined patients from the CLL cohort

(n=90) were subjected to repeated analysis of T-cell immunity against SARS-CoV-2 specific peptides. Baseline characteristics of the patients and controls are shown in Table 1. Patient groups included i) previously untreated (n=14), ii) individuals with ongoing ibrutinib therapy (n=14), iii) individuals who had stopped ibrutinib ≥2 months ago due to remission as part of another study addressing intermittent ibrutinib treatment (n=10), iv) or individuals who had received CD20 monoclonal antibody (mAb)-containing chemoimmunotherapy 6-30 months ago (of which 11 received their last dose >12 months ago) (n=14). Cellular and humoral immunity was measured at day (d) 0 (baseline), d10 (10 days after first dose), d21 and d35 (21 days after first dose, day of second dose). Seroconversion and antibody titres were analyzed as reported7 with d35 data available in 48 of the 52 predefined patients who were analyzed for T-cell immunity. Additionally, 41 of 90 healthy controls in the trial7 were preselected for T- cell analysis (Table 1). The ELISpot assay was applied as previously described,11 using plates and reagents from a human IFNg ELISpot kit (3420-2APT-2, Mabtech). Briefly, 2.5x105 peripheral blood mononuclear cells (PBMC)/well were seeded and supplemented with 0.15 µg/mL of co-stimulatory anti-CD28/CD49d (347690, BD Biosciences). Cells were stimulated for 20 hours (h) with a peptide pool covering the SARS-CoV-2 spike glycoprotein (0.5 µg/mL, LB01792; peptides&elephants) and equivalent dimethyl sulfoxide (DMSO) in unstimulated wells. Spot-forming units (SFU) were counted using the IRIS (Mabtech) automated reader system. Data points are presented as background corrected SFU/106 cells, calculated by subtracting the mean value of the corresponding duplicate unstimulated wells from the mean value of duplicate spike stimulated wells. Negative values after background correction are set to 1. The threshold for positive response corresponds to the average SFU/106 cells of unstimulated wells + 2 standard deviations (30 SFU/106 cells). Data were excluded when unstimulated wells had >100 SFU/106 cells. All analyses were prespecified as per protocol. Data is summarized using descriptive statistics such as counts, percentages, medians and range. Categorical variables are presented as cross-tabulations and distributional differences were tested using the Chi-squared test.

Table 1. Baseline characteristics and immunological results 2 weeks after the second dose (d35) of the mRNA BNT162b2 vaccine in relation to chronic lymphocytic leukemia patient subgroup and healthy controls.

Age, years median (range) Male sex n (%) Seroconverted d35 n (%) Ab titres d35, U/mL* median (range) ELISpot d35, SFU/106 PBMC median (range) Positive T-cell response n (%)

Entire cohort N=52

70 (23–86) 36 (69%) 29/48 (60%) 24.6a (0.4–3,320)a 10 (1–1,096) 15/52 (29%)

Indolent untreated N=14

71 (49–82) 6 (43%) 11/13 (85%) 81.6 (0.4–3,320) 6 (1–526) 4/14 (29%)

Previous CD20 mAb N=14

71 (56–84) 13 (93%) 11/12 (92%) 42.4 (0.4–1,343) 10 (1–490) 2/14 (14%)

Previous ibrutinib N=10

71 (54–86) 6 (60%) 5/9 (56%) 35.5 (0.4–559) 53 (2–1,096) 7/10 (70%)

Ongoing ibrutinib N=14

70 (23–84) 11 (79%) 2/14 (14%) 0.4 (0.4–170) 5 (2–70) 2/14 (14%)

Healthy controls N=41

52 (25–79) 19 (46%) 41/41 (100%) 2696 (766–14,269) 48 (1-1,526) 24/41 (59%)

*Ab baseline = 0.4 U/mL; seropositive >0.8 U/mL. an = 48 (with d35 data). Ab: antibodies; ELISpot: enzyme-linked immunospot; SFU: spot forming units; PBMC: peripheral blood mononuclear cells; d35: day 35.

1000

haematologica | 2022; 107(4)

Letters to the Editor

Significance between time points with missing values was assessed using Kruskal-Wallis test with Dunn’s posttest. Correlation analysis was done using non-parametric Spearman rank correlation. P-values <0.05 were considered significant. Graphs and associated statistical tests were performed in Prism v.9 (GraphPad Software Inc.). Seroconversion rates were in line with our full clinical trial report7 and are summarized in Table 1 and Figure 1A. Seroconversion occurred in 29 of 48 patients (60%) (4/52 patients had missing serology data at d35) compared to 41 of 41 (100%) of controls. The time kinetics of seroconversion are shown in Figure 1A. Only 4% and 18% of patients had seroconverted at d10 and d21, respectively, followed by 60% at d35. Patients’ and controls’ responses showed similar kinetics. There was no

significant increase in SARS-CoV-2-specific immunoglobulin G (IgG) after vaccination on d10 in either group. At d21 and at d35, both groups had a significant response compared to baseline (CLL: d0 vs. d21, P<0.05, d0 vs. d35, P<0.0001; controls: d0 vs. d21, P<0.0001; d0 vs. d35, P<0.0001). Subgroup analysis revealed that seroconversion occurred in 11 of 13 (85%) of previously untreated patients, 11 of 12 (92%) of previously CD20 mAb-treated patients, five of nine (56%) of previously ibrutinib-treated and two of 14 (14%) of patients with ongoing ibrutinib therapy. The difference between patients on or off ibrutinib was significant (P=0.036). The median antibody titer at d35 in each subgroup as above was 81.6 U/mL (range, 0.4-3,320), 42.4 U/mL (range, 0.4-1,343), 35.5

Figure 1. Humoral and cellular immune response in chronic lymphocytic leukemia patients. Longitudinal assessment (day 0, 10, 21, 35 post-vaccination) of SARS-CoV-2-specific immunglobulin G (IgG) (A) and IFN-g T cells (B) after spike glycoprotein stimulation, with summarized number and frequency of patients tested positive. Subgroup analysis of chronic lymphocytic leukemia (CLL) patients at day 35 (C) and correlation day 35 (D). The dashed line indicates positive threshold for SARS-CoV-2-specific IgG and IFN-g spot forming units (SFU)/106 cells, 0.8 U/mL and 30 SFU/106 cells respectively. The dotted line represents the lower limit of detection of both assays. Each dot represents one patient. Rs=Spearman r value. Kruskal-Wallis test with Dunn’s correction for multiple comparisons. *P<0.05, **P< 0.01, ***P<0.001, ****P<0.0001.

haematologica | 2022; 107(4)

1001

Letters to the Editor

U/mL (range, 0.4-559) and 0.4 U/mL (0.4-170) respectively. This compared to a median antibody titer of 2,696 U/ml (range, 766-14,269) in controls (Table 1). Longitudinal assessment of T-cell immunity against SARS-CoV-2 spike peptides (ELISpot) is shown in Figure 1B and summarized in Table 1. At d35, 15 of 52 patients (29%) had a specific T-cell response (P<0.05 vs. baseline, Figure 1B) compared to 24 of 41 (59%) in controls (P<0.01). Pre-existing spike-cross-reactive T cells12 was observed at baseline in five of 50 patients; four patients showed no vaccine response and one patient showed a marginal increase in T-cell response (mean spot count from 44 to 70 SFU/106 PBMC at d35). A positive T-cell response was observed in nine of 50 tested patients (18%) at d10 and in six of 51 (12%) at d21. CLL subgroup results are shown in Figure 1C and Table 1. IFN-g positivity was observed in seven of ten patients at d35 who were off ibrutinib, whereas only two of 14 patients on ibrutinib developed T-cell immunity (P<0.01). The corresponding numbers were four of 14 among previously untreated patients and two of 14 if previously treated with CD20 mAb (P<0.05 and P<0.01, respectively, vs. patients off ibrutinib). Finally, we analyzed correlation between seroconversion and T-cell response (Figure 1D). A weak but significant correlation was observed (r=0.2861, P=0.049). Fifteen patients (29%) were double-negative i.e., neither mounted a T-cell response nor seroconverted, whereas nine (18%) came out positive in both assays. Twenty patients (39%) were positive in serology only and only three patients (2 in the off ibrutinib group) had an IFN-g response in the absence of seroconversion. Most doublenegative patients (11/15) were found among patients on ibrutinib. Double-positive patients were most frequent in those off ibrutinib (4/9). Of the 20 seroconverted patients with no T-cell response, one patient was found in the ongoing ibrutinib and one in the previously ibrutinib treated group, while eight were previously untreated and ten were previously treated with CD20 mAb. Following natural COVID-19 infection, durable immunity including both antibodies and T cells seem to occur in most healthy individuals13 as well as in patients with CLL.4 Most healthy individuals mount T-cell responses following mRNA vaccination.14 This was reported also in patients with solid tumors.15 Lower numbers were recently reported in patients with hematological malignancies.8,9 The present study shows that, compared to healthy controls, half as many of patients with CLL developed IFN-g T-cell response (28% vs. 59%) after two doses of mRNA vaccine. A limitation of the present T-cell assay11 is capturing of only IFN-g positive cells e.g., missing-out on other cytokine-secreting antigen-specific T cells. Despite this, we were able to capture both temporal and group dynamic changes of the SARS-CoV-2 spikespecific T-cell response, and were able to make comparison of patients with CLL with healthy controls. CLL subgroup results were driven by patients who were off ibrutinib (7/10 responded) whereas other CLL sub-groups had few T-cell responders. However, the data must be viewed with caution due to the open-label trial design and the small numbers within each subset. Thus, our subgroup analysis should be confirmed in extended studies. Even though our healthy controls were younger (median age 52 years) age did not impact on their T-cell response (data not shown). Double-negativity was found in most patients on ibrutinib who remain of major concern, suggesting that temporary cessation of BTKi may be explored onwards in future studies. Of note, Ehmsen et al.9 found T-cell responses in 26% of seronegative hema1002

tology patients whereas we found it only in three of 52 patients with CLL (6%). A third dose is currently explored in CLL and its effect on T-cell immunity, even though its additional effect on T cells was limited in solid tumors.15 CLL remain as a group of special concern in the ongoing pandemic. Lisa Blixt,1,2* David Wullimann,3* Soo Aleman,4,5 Jeanette Lundin,1,2 Puran Chen,3 Yu Gao,3 Angelica Cuapio,3 Mira Akber,3 Joshua Lange,3 Olga Rivera-Ballesteros,3 Marcus Buggert,3 Hans-Gustaf Ljunggren,3 Lotta Hansson1,2 and Anders Österborg1,2 for the COVAXID clinical study group*** 1 Department of Hematology, Karolinska University Hospital Solna; 2 Deptartment of Oncology-Pathology, Karolinska Institutet; 3Center for Infectious Medicine, Department of Medicine Huddinge, Karolinska Institutet; 4Department of Infectious Diseases, Karolinska University Hospital and 5Department of Medicine Huddinge, Infectious Diseases, Karolinska Institutet, Stockholm, Sweden *LB and DW contributed equally as co-first authors Correspondence: LOTTA HANSSON- lotta.hansson@regionstockholm.se doi:10.3324/haematol.2021.280300 Received: November 9, 2021. Accepted: January 5, 2022. Prepublished: January 20, 2022. Disclosures: MB is a consultant for Oxford Immunotech. All other authors have no conflicts of interest to disclose.. Contributions: LH, AÖ, MB and HGL contributed to conceptualization, funding acquisition and discussion of data; DW, PC, YG, AC, JL, ORB and MA contributed to sample processing throughout the COVAXID clinical trial; DW performed experiments and analyzed data; LB, LH and AÖ recruited CLL study participants, conducted investigation through recruitment of the study participants and conducted management of participants during the trial and analyzed data; SA was the PI of the COVAXID clinical trial, contributed to conceptualization, funding acquisition and discussion of data; JL recruited CLL study participants including those off BTK inhibitor-treatment and discussed data; LB, DW, AÖ, LH, MB and HGL wrote the original draft of the manuscript. All authors reviewed and edited revisions of the manuscript and had final responsibility for the decision to submit for publication. Acknowledgements: we thank all patients who donated blood for this study and Leila Relander and Sonja Sönnert-Husa for technical assistance. Funding: this study was supported by grants from the SciLifeLab National COVID-19 Research Program, financed by the Knut and Alice Wallenberg Foundation, the Swedish Research Council, Region Stockholm, the Swedish Blood Cancer Foundation and Karolinska Institutet. Clinical trial information: The COVAXID clinical trial (EudraCT no.2021-000175-37) was approved by the Swedish Medical Product Agency (ID 5.1-2021-5881) and the Swedish Ethical Review Authority (ID 2021-00451) ***The COVAXID Clinical Study Group: Peter Bergman, MD, Dept. of Infectious Diseases, Karolinska University Hospital and Dept. of Laboratory Medicine, Clinical Microbiology, Karolinska Institutet, Stockholm; Ola Blennow, MD, Dept. of Infectious Diseases, Dept. of Transplantation, Karolinska University Hospital and Dept. of Clinical Science, Intervention and Technology, Karolinska Institutet, Stockholm; Lotta Hansson, MD, Dept. of Hematology, Karolinska University Hospital Solna and Dept. of Oncology-Pathology, Karolinska Institutet, Stockholm; Stephan Mielke, MD, Dept of Laboratory Medicine, Biomolecular and Cellular Medicine, Karolinska Institutet and Dept. of Cellular Therapy and Allogeneic Stem Cell Transplantation (CAST), Karolinska University Hospital Huddinge, Stockholm; Piotr Nowak, MD, Dept. of Infectious Diseases, haematologica | 2022; 107(4)

Letters to the Editor

Karolinska University Hospital and Dept. of Medicine Huddinge, Infectious Diseases, Karolinska Institutet, Stockholm, and Laboratory for Molecular Infection Medicine Sweden MIMS, Umeå University, Umeå; Puran Chen, MD, Dept. of Medicine Huddinge, Center for Infectious Medicine, Karolinska Institutet, Stockholm; Gunnar Söderdahl, MD, Dept. of Transplantation, Karolinska University Hospital and Dept. of Clinical Science, Intervention and Technology, Karolinska Institutet, Stockholm; Gordana Bogdanovic, MD, Dept. of Clinical Microbiology, Karolinska University Hospital, Stockholm; Anders Österborg, MD, Dept. of Hematology, Karolinska University Hospital Solna and Dept. of Oncology-Pathology, Karolinska Institutet, Stockholm; C. I. Edvard Smith, MD, Dept. of Laboratory Medicine, Karolinska Institutet, Stockholm; Marcus Buggert, PhD, Dept. of Medicine Huddinge, Center for Infectious Medicine, Karolinska Institutet, Stockholm; Hans-Gustaf Ljunggren, MD, Dept. of Medicine Huddinge, Center for Infectious Medicine, Karolinska Institutet, Stockholm; Per Ljungman, MD, Dept. of Medicine Huddinge, Hematology, Karolinska Institutet and Dept. of Cellular Therapy and Allogeneic Stem Cell Transplantation (CAST), Karolinska University Hospital Huddinge, Stockholm; Soo Aleman, MD, Dept. of Infectious Diseases, Karolinska University Hospital and Dept. of Medicine Huddinge, Infectious Diseases, Karolinska Institutet, Stockholm, Sweden

References 1. Langerbeins P, Eichhorst B. Immune dysfunction in patients with chronic lymphocytic leukemia and challenges during COVID-19 pandemic. Acta Haematol. 2021;144(5):508-518. 2. Mato AR, Roeker LE, Lamanna N, et al. Outcomes of COVID-19 in patients with CLL: a multicenter international experience. Blood. 2020;136(10):1134-1143. 3. Scarfo L, Chatzikonstantinou T, Rigolin GM, et al. COVID-19 severity and mortality in patients with chronic lymphocytic leukemia: a joint study by ERIC, the European Research Initiative on CLL, and CLL Campus. Leukemia. 2020;34(9):2354-2363. 4. Blixt L, Bogdanovic G, Buggert M, et al. Covid-19 in patients with

haematologica | 2022; 107(4)

chronic lymphocytic leukemia: clinical outcome and B- and T-cell immunity during 13 months in consecutive patients. Leukemia. 2022;36(2):476-481. 5. Herishanu Y, Avivi I, Aharon A, et al. Efficacy of the BNT162b2 mRNA COVID-19 vaccine in patients with chronic lymphocytic leukemia. Blood. 2021;137(23):3165-3173. 6. Roeker LE, Knorr DA, Thompson MC, et al. COVID-19 vaccine efficacy in patients with chronic lymphocytic leukemia. Leukemia. 2021;35(9):2703-2705. 7. Bergman P, Blennow O, Hansson L, et al. Safety and efficacy of mRNA BNT162b2 vaccine against SARS-CoV-2 in five groups of immunocompromised patients and healthy controls in a prospective open-label clinical trial. EBioMedicine. 2021;9;74:103705. 8. Monin L, Laing AG, Munoz-Ruiz M, et al. Safety and immunogenicity of one versus two doses of the COVID-19 vaccine BNT162b2 for patients with cancer: interim analysis of a prospective observational study. Lancet Oncol. 2021;22(6):765-778. 9. Ehmsen S, Asmussen A, Jeppesen SS, et al. Antibody and T cell immune responses following mRNA COVID-19 vaccination in patients with cancer. Cancer Cell. 2021;39(8):1034-1036. 10. Sahin U, Muik A, Derhovanessian E, et al. COVID-19 vaccine BNT162b1 elicits human antibody and TH1 T cell responses. Nature. 2020;586(7830):594-599. 11. Sekine T, Perez-Potti A, Rivera-Ballesteros O, et al. Robust T cell immunity in convalescent individuals with asymptomatic or mild COVID-19. Cell. 2020;183(1):158-168. 12. Loyal L, Braun J, Henze L, et al. Cross-reactive CD4(+) T cells enhance SARS-CoV-2 immune responses upon infection and vaccination. Science. 2021;374(6564):eabh1823. 13. Cohen KW, Linderman SL, Moodie Z, et al. Longitudinal analysis shows durable and broad immune memory after SARS-CoV-2 infection with persisting antibody responses and memory B and T cells. Cell Rep Med. 2021;2(7):100354. 14. Painter MM, Mathew D, Goel RR, et al. Rapid induction of antigenspecific CD4(+) T cells is associated with coordinated humoral and cellular immunity to SARS-CoV-2 mRNA vaccination. Immunity. 2021;54(9):2133-2142. 15. Shroff RT, Chalasani P, Wei R, et al. Immune responses to two and three doses of the BNT162b2 mRNA vaccine in adults with solid tumors. Nat Med. 2021;27(11):2002-2011.

1003

Letters to the Editor

Acute lymphoblastic leukemia cells are able to infiltrate the brain subventricular zone stem cell niche and impair neurogenesis In pediatric acute lymphoblastic leukemia (ALL), the most common hematologic malignancy in childhood, central nervous system (CNS) relapse is a major clinical problem, accounting for about one-third of the relapses.1 Since the early autopsy studies, CNS leukemia has been described primarily as a leptomeningeal disease which can be accompanied by infiltration of different brain parenchyma areas by ALL cells.2,3 The CNS is therefore considered to act as a sanctuary for ALL, but the specific neural microenvironments in which leukemic cells can stay for prolonged periods of time as extramedullary minimal residual disease and be responsible for CNS relapses are still poorly defined. Recently, we and others have reported that the choroid plexus stroma and the leptomeningeal stromal cell network are two of those neural microenvironments able to lodge ALL cells and promote their survival and acquisition of quiescence and chemoresistance.4,5 Neurogenic niches, such as the subventricular zone (SVZ) and the subgranular zone, are areas of the brain in which neurogenesis takes place throughout life. Concretely, the SVZ is located along the walls of the lateral brain ventricles and represents the largest neurogenic niche in the postnatal and adult mammalian brain.

Neural stem cells residing in the SVZ divide slowly and give rise to rapidly proliferating cells, called transit amplifying progenitors, which then differentiate to neuroblasts. These immature neuronal progenitors further migrate along a pathway, called the rostral migratory stream, towards the olfactory bulb where they differentiate into olfactory interneurons and integrate in the existing neuronal circuitry involved in odor discrimination.6 Nevertheless, the neurogenic niches, apart from their ability to support and maintain neural stem cells, can also serve as refuge for neoplastic cells, and the migratory mechanisms of neural stem cells can be utilized by tumor cells;7 therefore, in this study we investigated whether the SVZ can also harbor ALL cells. A xenograft model of non-obese diabetic/severe combined immunodeficiency/IL-2Rgnull (NSG) mice injected with the Nalm-6 human pre-B ALL cell line was used to analyze the presence of leukemia cells in the brain SVZ. When symptoms of CNS involvement (such as hind limb paralysis) appeared, animals were deeply anesthetized and transcardially perfused immediately before the SVZ were carefully dissected and dissociated. Flow cytometry analysis of cells recovered from dissociated SVZ revealed that all mice with leukemic infiltration in the CNS showed CD19+ human blasts in this brain location (Figure 1A), and these ALL cells could represent up to 30% of total leukemic cells invading the nervous parenchyma (Figure 1B). Similar results were obtained

Figure 1. Acute lymphoblastic leukemia cells infiltrate the subventricular zone neurogenic niche. (A) Percentage of CD19+ leukemia cells detected in subventricular zones (SVZ) from mice xenografted with pre-B acute lymphoblastic leukemia (ALL) (n=8). When disease symptoms were evident, mice were deeply anesthetized and transcardially perfused with cold 0.1 M phosphate-buffered saline (pH 7.4) to clear circulating leukemia cells prior to euthanasia. Brains were then removed, washed several times with cold Dulbecco phosphate-buffered saline and SVZ were carefully microdissected, dissociated and processed for flow cytometry. Representative dot plots show, after gating out the non-neurogenic cell populations, the expression of human CD19 (hCD19) versus murine CD24 (mCD24), a neuroblast cell marker, in SVZ from leukemic and healthy mice. (B) Percentages (mean ± standard deviation [SD]) represent the leukemic cells present in the SVZ or the brain parenchyma out of total CD19+ cells infiltrated into the brain. (C) Proportion of CD19+ leukemia cells detected by flow cytometry in SVZ from mice xenografted with ALL cell lines REH and RS4;11 (n=4-8). (D) Percentages of CD19+ leukemia cells expressing Ki-67 in the SVZ and meninges from xenografted mice (n=3). Representative histograms show the expression of Ki-67 antigen in CD19+ leukemic cells found in the SVZ and the meninges (*P≤0.05; t-test). (E) Mice were injected with Nalm-6 cells and after successful engraftment and randomization, the leukemic mice were intraperitoneally treated with methotrexate (5 mg/kg) or saline twice a week for 4 weeks (n=6). Percentages of CD19+ cells present in the SVZ and the rest of the brain, including the meninges, were determined by flow cytometry (**P≤0.01; t-test).

1004

haematologica | 2022; 107(4)

Letters to the Editor

with mice xenografted with other ALL cell lines, REH and RS4;11 (Figure 1C). However, no correlation was seen between the degree of leukemic infiltration in brain parenchyma and the proportion of ALL cells present in the SVZ. Immunofluorescence studies in brain cryosections from xenografted mice showed that leukemic cells could be seen, apart from in the SVZ niche, also along the rostral migratory stream (Online Supplementary Figure S1). These data indicate that the SVZ can provide a favorable microenvironment in which ALL cells can survive and be maintained over time. Supporting this, the study of the leukemia proliferation rate using Ki-67 staining showed that ALL cells found in the SVZ niche exhibit a much lower proliferative activity than those leukemic cells isolated from the meninges (Figure 1D). Furthermore, leukemic cells infiltrating the SVZ niche were shown to have higher chemoresistance after methotrexate treatment of xenografted mice (Figure 1E). The above results showed that leukemic invasion of the SVZ neurogenic niche is a common event in the xenograft model of ALL, so we analyzed the effects of this infiltration on the differentiation of neural stem cells. The proportion of the different SVZ populations was determined by flow cytometry using a combination of multiple specific cell markers, as we previously described.8 Non-neurogenic cells were first discarded from the study, and the remaining neurogenic lineage cell pool was subdivided according to the expression of the glial marker GLAST, the neuroblast marker CD24, the tetraspanin CD9 and the proliferation-associated recep-

tor EGFR. The population of neural stem cells was defined as GLAST+CD24-/lowCD9high and further classified by EGFR expression and GLAST intensity into quiescent, primed quiescent and activated neural stem cells. Transit amplifying progenitors were defined as GLAST-CD24/low EGFR+ cells, and the GLAST-CD24high population included EGFR+ proliferating (NB1 or early) and EGFR-/low migrating (NB2 or late) neuroblasts. As can be seen in Figure 2A, the percentage of total neural stem cells was notably increased in xenografted mice, with quiescent neural stem cells being the main subset responsible for this rise. Concomitantly, the proportion of transit amplifying progenitors and late neuro-blasts was reduced in these animals. These results suggest that SVZ neurogenesis is impaired in leukemia-bearing mice at the expense of an increase in quiescence, and the effect appears to be a direct consequence of the leukemic cell infiltration in the SVZ since the most affected animals were those showing the highest numbers of CD19+ cells in the neurogenic niche. Figure 2B shows that the percentages of CD19+ leukemia cells correlated directly with the accumulation of quiescent neural stem cells, and inversely with the proportions of late neuroblasts (Figure 2C). In line with these data, and since the generation of new olfactory bulb neurons from the SVZ is required for novel odor discrimination,9 xenografted mice displayed altered olfactory discrimination capacities (Figure 2D). To analyze the effects of leukemia on neural precursors directly, we first generated SVZ neurospheres, floating cellular aggregates clonally derived from neurosphere-ini-

Figure 2. Subventricular zone cell populations are affected by the presence of acute lymphoblastic leukemia cells. (A) Bars represent the percentages (mean ± standard deviation [SD]) of total, quiescent (q), primed quiescent (p) and activated (a) neural stem cells (NSC), as well as transit amplifying progenitors (TAP) and proliferating (NB1) and migrating (NB2) neuroblasts present in the subventricular zones (SVZ) from leukemic (gray) and healthy (black) mice (n=8). All these neurogenic cell populations were defined according to the expression of the glial marker GLAST, the neuroblast marker CD24, the tetraspanin CD9 and the proliferation-associated receptor EGFR (*P ≤ 0.05; t-test). (B,C) The percentages of CD19+ cells found in the SVZ are represented as a function of the corresponding (B) increases in the proportion of quiescent NSC and (C) decreases in the proportion of migrating EGFR- neuroblasts. P values of the Pearson correlation are provided. (D) Olfactory habituation-dishabituation tests of healthy (black circles) and leukemic (gray triangles) mice were performed at week 3, before the typical disease symptoms (including rough hair, lethargy, hunched-back posture, loss of motor functions and hind limb paralysis) were observed. Exploration time (in seconds) of successive cotton swabs soaked in octanal (O), heptanal (H), or anisole (A) is shown. After exposure to octanal-soaked swabs, healthy mice reacted to heptanal- and anisole-soaked swabs; however leukemic mice displayed lower olfactory exploration and no reaction to the new odor stimuli. Asterisks represent statistically significant differences (*P≤0.05; t-test). ALL: acute lymphoblastic leukemia.

haematologica | 2022; 107(4)

1005

Letters to the Editor

tiating neural stem cells which constitute an ideal system to evaluate modifications in proliferation and self-renewal. Single cells dissociated from neurospheres were either cultured with medium conditioned by leukemic cells or co-cultured with ALL cells using transwell inserts. In both cases, although no change in the number of new neurospheres was found after 10 days (Figure 3A), a significant reduction in neurosphere sizes could be clearly observed (Figure 3B), suggesting that leukemia-derived factors limit growth but not survival of neurosphere cells. In agreement, the inhibition of the expansion potential of neurospheres in the presence of leukemia cells was detected throughout the culture period using MTS proliferation assays (Figure 3C). To analyze whether ALLmediated effects included effects on self-renewal, cells obtained from neurospheres that had been grown in the presence of soluble factors secreted by leukemic blasts

1006

were re-plated in fresh growth medium without leukemia-derived factors and neurosphere formation was evaluated. In these cultures, the numbers of secondary neurospheres were not altered but significant changes in sphere diameters could be newly detected, indicating that leukemia cells can reduce activation without altering self-renewal (Figure 3D, E). ALL cells have been reported to be able to induce a proinflammatory microenvironment in different locations.4,10 The expression of pro-inflammatory factors was therefore analyzed in the SVZ of healthy and leukemic mice. As shown in Figure 3F, the levels of IL-1b, IL-6 and TNFa cytokines as well as CCL2 and CXCL10 chemokines were notably upregulated in the leukemia-invaded SVZ niches. All these inflammatory mediators have been described as negative regulators of neurogenesis.11 However, we have recently reported that TNF-a, which

Figure 3. Acute lymphoblastic leukemia cells impair neurogenesis. (A) Number of subventricular zone (SVZ) primary neurospheres grown in neural stem cell (NSC) complete medium previously conditioned by Nalm-6 leukemic cells for 48 h (CM; grey bars) or unconditioned growth medium (CTL; black bars). Data represent mean ± standard deviation (SD) from three independent experiments. (B) Percentage of small (<60 mm diameter), medium (60-180 mm diameter) and large (>180 mm diameter) primary neurospheres grown in conditioned or unconditioned NSC complete medium. Data represent mean ± SD from three independent experiments (*P≤0.05, **P≤0.01; t-test). (C) Proliferation of neurospheres was measured by MTS assays after 1, 3 and 6 days of culture in unconditioned (black circles) or leukemia-conditioned medium (gray squares). Data represent mean ± SD from three independent experiments (***P≤0.001; Mann–Whitney test). (D) Percentages of small, medium and large secondary neurospheres generated after dissociation of primary neurospheres, grown in conditioned and unconditioned medium, and replated in fresh NSC complete medium. Data represent mean ± SD from three independent experiments (*P≤0.05; t-test). (E) Micrographs show floating neurospheres grown in unconditioned and leukemia-conditioned medium. Scale bar: 100 mm. (F) The expression of the indicated cytokines and chemokines was analyzed by reverse transcriptase quantitative polymerase chain reaction in SVZ cells from healthy (black bars) and leukemic mice (gray bars). Fold induction relative to cells from healthy SVZ is shown, and the mean ± SD of four independent experiments is presented (*P≤0.05; Mann–Whitney test). ALL: acute lymphoblastic leukemia.

haematologica | 2022; 107(4)

Letters to the Editor

underwent one of the highest increases in expression, reduces neuroblast generation because it induces a transient activation of neural stem cells followed by their entry into quiescence.8 Transgenic mice overexpressing IL-6 in astrocytes exhibit reduced cycling of neural stem cells in the subgranular zone niche, suggesting that it acts as a negative regulator of proliferation,12 and IL-1b is reportedly secreted by choroid plexus cells to the cerebrospinal fluid and induces the upregulation of VCAM-1 levels in SVZ neural stem cells, reducing their proliferation and preventing lineage progression.13 Importantly, the levels of IL-1b, IL-6, TNF-a, CCL2 and CXCL10 have also been described to be increased in blood and cerebrospinal fluid of ALL patients, promoting the survival and quiescence of leukemic cells.14,15 Taken together, the results of the present study show that infiltration of the SVZ may be a common event in childhood ALL with CNS involvement, suggesting that SVZ is a sanctuary in which ALL cells could lodge, survive for prolonged periods of time and be responsible for future CNS relapses. Our results also show that leukemic infiltration of the SVZ neurogenic niche impairs neurogenesis, which likely leads to deleterious effects on brain functions. It is important to note that in human infants and young children not all neuroblasts born in the SVZ migrate to the olfactory bulb, but many of them migrate into the ventromedial prefrontal cortex as well as multiple regions of the frontal cortex, such as the cingulate gyrus.16,17 The late incorporation of inhibitory interneurons into those regions of the developing human brain has been proposed to constitute a mechanism of delayed postnatal plasticity and, therefore, injuries affecting neuronal recruitment during this period could contribute to neurocognitive deficits and sensorimotor disturbances,16,17 such as those reported in ALL patients at diagnosis, before treatment initiation.18 Lidia M. Fernández-Sevilla,1,2 Germán Belenguer,3 Beatriz Martí-Prado,3 Paula Ortiz-Sánchez,1 Manuel Ramírez,4 Alberto Varas,1,2 Isabel Fariñas3 and Ángeles Vicente1,2 1 Department of Cell Biology, Faculty of Medicine, Complutense University, Madrid; 2Instituto de Investigación Sanitaria del Hospital Clínico San Carlos (IdISSC), Madrid; 3CIBERNED, Departamento de Biología Celular, Biología Funcional y Antropología Física, Instituto de Biotecnología y Biomedicina, Universitat de València, Valencia and 4 Department of Pediatric Hematology and Oncology, Advanced Therapies Unit, Niño Jesús University Children's Hospital, Madrid, Spain Correspondence: ÁNGELES VICENTE - avicente@ucm.es ISABEL FARIÑAS - isabel.farinas@uv.es doi:10.3324/haematol.2021.279383 Received: June 8, 2021. Accepted: January 7, 2022. Pre-published: January 20, 2022. Disclosures: MR has received a research grant from Orgenesis Inc. Contributions: LMFS performed experiments, acquired the data, analyzed results, interpreted data and wrote the manuscript; GB, BMP and POS performed experiments, acquired the data and analyzed results; MR and IF co-designed the study and interpreted data; AVa and AVi conceived and designed the research, interpreted data and wrote the manuscript.

haematologica | 2022; 107(4)

Funding: this work was supported by grants RTI2018-093899-BI00 and SAF2017-86690-R (Spanish Ministry of Economy and Competitiveness), RD16/0011/0002 and CB06/05/0086 (Institute of Health Carlos III, Spain), B2017/BMD-3692 AvanCell-CM (Community of Madrid), and Beca I-UnoEntreCienMil (Uno Entre Cien Mil Foundation). LMFS was supported by a pre doctoral fellowship (CT45/15 CT46/15) from the Complutense University of Madrid. POS is supported by a pre-doctoral fellowship (CT63/19CT64/19) from the Complutense University.

References 1. Frishman-Levy L, Izraeli S. Advances in understanding the pathogenesis of CNS acute lymphoblastic leukaemia and potential for therapy. Br J Haematol. 2017;176(2):157-167. 2. Thomas LB. Pathology of leukemia in the brain and meninges: postmortem studies of patients with acute leukemia and of mice given inoculations of L1210 leukemia. Cancer Res. 1965;25(9):1555-1571. 3. Kinjyo I, Bragin D, Grattan R, Winter SS, Wilson BS. Leukemiaderived exosomes and cytokines pave the way for entry into the brain. J Leukoc Biol. 2019;105(4):741-753. 4. Fernandez-Sevilla LM, Valencia J, Flores-Villalobos MA, et al. The choroid plexus stroma constitutes a sanctuary for paediatric B-cell precursor acute lymphoblastic leukaemia in the central nervous system. J Pathol. 2020;252(2):189-200. 5. Jonart LM, Ebadi M, Basile P, Johnson K, Makori J, Gordon PM. Disrupting the leukemia niche in the central nervous system attenuates leukemia chemoresistance. Haematologica. 2020;105(8):21302140. 6. Fontán-Lozano A, Morcuende S, Davis-López de Carrizosa MA, et al. To become or not to become tumorigenic: subventricular zone versus hippocampal neural stem cells. Front Oncol. 2020;10:602217. 7. Sinnaeve J, Mobley BC, Ihrie RA. Space invaders: brain tumor exploitation of the stem cell niche. Am J Pathol. 2018;188(1):29-38. 8. Belenguer G, Duart-Abadia P, Jordan-Pla A, et al. Adult neural stem cells are alerted by systemic inflammation through TNF-alpha receptor signaling. Cell Stem Cell. 2021;28(2):285-299. 9. Alonso SB, Reinert JK, Marichal N, et al. An increase in neural stem cells and olfactory bulb adult neurogenesis improves discrimination of highly similar odorants. EMBO J. 2019;38(6):e98791. 10. Vilchis-Ordonez A, Contreras-Quiroz A, Vadillo E, et al. Bone marrow cells in acute lymphoblastic leukemia create a proinflammatory microenvironment influencing normal hematopoietic differentiation fates. Biomed Res Int. 2015;2015:386165. 11. Voloboueva LA, Giffard RG. Inflammation, mitochondria, and the inhibition of adult neurogenesis. J Neurosci Res. 2011;89(12):19891996. 12. Brett FM, Mizisin AP, Powell HC, Campbell IL. Evolution of neuropathologic abnormalities associated with blood-brain barrier breakdown in transgenic mice expressing interleukin-6 in astrocytes. J Neuropathol Exp Neurol. 1995;54(6):766-775. 13. Kokovay E, Wang Y, Kusek G, et al. VCAM1 is essential to maintain the structure of the SVZ niche and acts as an environmental sensor to regulate SVZ lineage progression. Cell Stem Cell. 2012;11(2):220230. 14. Gomez AM, Martinez C, Gonzalez M, et al. Chemokines and relapses in childhood acute lymphoblastic leukemia: a role in migration and in resistance to antileukemic drugs. Blood Cells Mol Dis. 2015;55(3):220-227. 15. Perez-Figueroa E, Sanchez-Cuaxospa M, Martinez-Soto KA, et al. Strong inflammatory response and Th1-polarization profile in children with acute lymphoblastic leukemia without apparent infection. Oncol Rep. 2016;35(5):2699-2706. 16. Sanai N, Nguyen T, Ihrie RA, et al. Corridors of migrating neurons in the human brain and their decline during infancy. Nature. 2011;478(7369):382-386. 17. Paredes MF, James D, Gil-Perotin S, et al. Extensive migration of young neurons into the infant human frontal lobe. Science. 2016;354(6308):aaf7073. 18. Reinders-Messelink H, Schoemaker M, Snijders T, et al. Motor performance of children during treatment for acute lymphoblastic leukemia. Med Pediatr Oncol. 1999;33(6):545-550.

1007

CASE REPORTS Immune-mediated thrombotic thrombocytopenic purpura following administration of Pfizer-BioNTech COVID-19 vaccine Coronavirus disease 2019 (COVID-19) is an infectious disease caused by a recently discovered coronavirus called severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). The disease affected well over a hundred million people all over the world and mortality rates were greater in the elderly population. The quick development of safe and effective vaccines represents an important step in the management of the ongoing pandemic and, so far, four vaccines have been approved by the European Medicines Agency, including mRNA (Pfizer-BioNTech and Moderna) and adenovirus-based vaccines (AstraZeneca and Johnson & Johnson). Immune-mediated thrombotic thrombocytopenic purpura (TTP) is a rare disease (annual incidence between 1.5 and 6.0 cases per million)1 characterized by microangiopathic hemolytic anemia, thrombocytopenia and ischemic end-organ injury due to microvascular plateletrich thrombi. The formation of microvascular thrombi is caused by a deficiency of ADAMTS13, a von Willebrand factor-cleaving protease, due to the presence of antiADAMTS13 autoantibodies.2 Cases of immune-mediated TTP following the administration of vaccines have been previously described and recently reviewed (Table 1).3 However, only one case of newly diagnosed immunemediated TTP following COVID-19 vaccination has been reported in the literature: it occurred in a 62-year-old female patient 37 days after receiving the adenovirusbased AstraZeneca vaccine.4 Additinoally, a case of immune-mediated TTP relapse was recently described in a 48-year-old female patient after the second dose of the Pfizer-BioNTech vaccine.5 Here we report two cases of immune-mediated TTP following the first dose of the Pfizer-BioNTech COVID-19 vaccine. Informed consent was obtained from the patients regarding the report of their clinical scenarios. Case 1. In April 2021, an 83-year-old female patient was admitted to the emergency room with severe anemia and macrohematuria, in the absence of fever, neurological signs and renal impairment. On clinical examination, diffuse petechiae and venipuncture hematomas were observed. The patient suffered from undifferentiated connective tissue disease treated with low-dose steroids and steroid-induced diabetes mellitus. The woman had been administered the first dose of Pfizer-BioNTech COVID-19 vaccine 14 days prior to the admission. One week before the admission, the patient was treated briefly at another center because of fatigue and the appearance of petechiae. A complete blood count revealed grade 3 anemia (hemoglobin 6.1 g/dL) and thrombocytopenia (platelet count 46x109/L), requiring

transfusion support. However, the patient refused hospital admission and was discharged. Seven days later, on admission to our center, the complete blood count again showed severe anemia (hemoglobin 5.6 g/dL) and thrombocytopenia (platelet count 23x109/L) with a normal white blood cell count. Markers of hemolysis were present, including increased reticulocytes, increased lactate dehydrogenase (1905 U/L, normal values [n.v:] 0-248), increased unconjugated bilirubin (5.5 mg/dL, n.v: 0.301.20) and reduced haptoglobin (<7 mg/dL) (Table 2). Coagulation and renal function tests were within normal limits. Both direct and indirect antiglobulin tests were negative. Examination of a peripheral blood smear revealed an increased number of schistocytes (10% per field). The PLASMIC score (6 points) classified the patient as being at high risk of severe ADAMTS13 deficiency.6 Tumor markers and infectious screening for hepatitis B virus, hepatitis C virus, human immunodeficiency virus, cytomegalovirus and Epstein-Barr virus resulted negative. Autoimmunity screening revealed the presence of anti-nuclear antibodies (titer 1:640, n.v. <1:80). A rapid ADAMTS13 test was performed demonstrating markedly reduced activity (below 10%) with a high titer of antiADAMTS13 antibodies according to enzyme-linked immunosorbent assay (ELISA) in serum (40 U/mL, n.v. 12-15), thus confirming the diagnosis of immune-mediated TTP. The patient was promptly started on intravenous methylprednisolone 1 mg/kg and daily sessions of plasma exchange in combination with the humanized antivon Willebrand factor nanobody, caplacizumab. Caplacizumab was administered intravenously at the dose of 10 mg before the first plasma-exchange, followed by 10 mg subcutaneous injections after each plasmaexchange. The patient was transfused with several units of concentrated red blood cells. After an initial clinical benefit with resolution of hematuria and early signs of hematologic recovery with a platelet count of 30x109/L, the patient died after only 2 days of treatment, probably due to a sudden cardiovascular event. Case 2. In June 2021, a 30-year-old woman, a b-thalassemia carrier, was admitted to the emergency room because of the appearance of diffuse petechiae, intense headache and fatigue. The patient had received her first dose of the Pfizer-BioNTech COVID-19 vaccine 18 days before the admission. On admission, a complete blood count revealed the presence of anemia (hemoglobin 8.9 g/dL) and thrombocytopenia (platelet count 11x109/L), with a normal white blood cell count (9.2x109/L). Total body computed tomography was negative. A peripheral blood smear showed the presence of schistocytes (5-10% per field). Investigations for hemolysis were positive, while both direct and indirect Coombs tests were negative. Coagulation, hepatic and renal function tests were all within normal limits (Table 3). Tumor markers,

Table 1. Cases of immune-mediated thrombotic thrombocytopenic purpura following the administration of vaccines.

Vaccine type Rabies Pneumococcal Influenza Influenza H1N1 Influenza

n 1 1 1 1 1 1

Age (years) 28 68 Unknown 54 56 23

Sex

Time of development

Male Female Unknown Male Male Female

14 day 15th day Unknown 5th day 13th day 14th day

Literature Kadikoylu et al., 2014 Kojima et al., 2014 Ramakrishnan et al., 1998 Dias and Gopal, 2009 Hermann et al., 2010 Brown et al., 1973

Source: Yavaşoğlu I.Vaccination and thrombotic thrombocytopenic purpura. Turkish J Hematol. 2020;37(3):218-219.

1008

haematologica | 2022; 107(4)

Case Reports

Table 2. Case 1: clinical and laboratory characteristics at diagnosis.

Table 3. Case 2: clinical and laboratory characteristics at diagnosis.

Characteristic

Age 83 years Sex Female Comorbidities UCTD, DM Medications Low-dose steroids, insulin Time from vaccination to admission, days 14 Hemoglobin, g/dL 5.6 (n.v. 12-16) Platelets, x109/L 23 (n.v. 130-400) 9 WBC, x10 /L 9.260 (n.v. 4.8-10.8) INR 1.14 (n.v. 0.8-1.2) aPTT, sec 24 (n.v. 24-36) Fibrinogen, mg/dL 141 (n.v. 170-400) Reticulocytes, % 28 (n.v. 0-25) LDH, U/L 1905 (n.v. 0-248) Haptoglobin, mg/dL <7 (n.v. 36-145) Unconjugated bilirubin, mg/dL 5.5 (n.v. 0.30-1.20) Creatinine, mg/dL 0.88 (n.v. 0.51-0.95)

Age Sex Comorbidities Medications Time from vaccination to admission, days Hemoglobin, g/dL Platelets, x109/L WBC, x109/L INR aPTT, sec Fibrinogen, mg/dL Reticulocytes, % LDH, U/L Haptoglobin, mg/dL Unconjugated bilirubin, mg/dL Creatinine, mg/dL

UCTD: undifferentiated connective tissue disease; DM: steroid-induced diabetes mellitus; n.v.: normal values; WBC: white blood cell count; INR: International Normalized Ratio; aPTT: activated partial thromboplastin time; LDH: lactate dehydrogenase.

n.v.: normal values; WBC: white blood cell count; INR: International Normalized Ratio; aPTT: activated partial thromboplastin time; LDH: lactate dehydrogenase.

autoimmune and infectious screening resulted negative. The PLASMIC score (6 points) classified the patient as high risk A rapid ADAMTS13 test revealed reduced activity (below 10%), while a high titer of anti-ADAMTS 13 antibodies was confirmed by ELISA (77.6 U/mL; n.v. 12-15). The woman was treated promptly with daily sessions of plasma exchange in combination with caplacizumab and intravenous methylprednisolone 1 mg/kg, and responded well to treatment. Her platelet count normalized on day 5 (platelet count, 158x109/L) with a hemoglobin value of 8.1 g/dL. Daily plasma exchange was continued for 8 consecutive days and she was discharged on day 8. On day 14 and on day 30 ADAMTS13 activity was 0 U/mL (n.v. 0.4-1.3), while antiADAMTS13 antibody titer progressively reduced, being 47 U/mL and 30 U/mL on day 14 and on day 30, respectively. The patient continued therapy with caplacizumab for 30 days after stopping daily plasma-exchange treatment. To the best of our knowledge, this is the first report of newly diagnosed immune-mediated TTP following the first dose of COVID-19 Pfizer-BioNTech vaccine. Immune-mediated TTP was not mentioned among the adverse events in the pivotal study leading to the approval of this vaccine.7 Given the immune origin of the TTP and the short latency period between the COVID19 vaccine and disease onset, a hypothesized temporal association is plausible. Although many cases of immune-mediated TTP following vaccinations have been reported previously, including those following influenza, pneumococcal, rabies and a recently published COVID19 adenovirus vector-based vaccines,3-8,10 the underlying mechanism is still unknown. In our first case, the patient also suffered from undifferentiated connective tissue disorder, an autoimmune disease that might have been a predisposing factor to post-vaccination TTP. In the literature, there is evidence of vaccine-induced autoimmunity, adjuvant-induced autoimmunity and antibody crossreaction in both experimental models as well as human patients.8 Furthermore, other cases of immune-mediated

disease onset or its flare were described after COVID-19 vaccination,9-11 including cases of post-vaccination immune thrombocytopenic purpura.12,13 A lot of attention has recently been given to the thrombotic risk of COVID-19 vaccination. In particular, a new syndrome called vaccine-induced immune thrombotic thrombocytopenia (VITT) following administration of the adenovirus-based vaccine AstraZeneca has been described. This syndrome is characterized by thrombosis at unusual sites, thrombocytopenia and the presence of high levels of antibodies to platelet factor 4 (PF4) in the absence of heparin treatment.14 In our cases, another disease characterized by an increased thrombotic risk developed following administration of an mRNA COVID-19 vaccine. Furthermore, the clinical cases we have described confirm that even a single administration of vaccine can induce the development of autoimmune manifestations especially in predisposed subjects. The consequences of developing antibodies against ADAMTS13 can be very serious and even fatal. It is, therefore, always necessary to take a thorough history before the administration of COVID-19 vaccines, and careful clinical surveillance in the post-vaccine period must be taken into consideration in patients with autoimmune diseases or a clinical or family history leading to the suspicion of an autoimmune tendency.

haematologica | 2022; 107(4)

30 years Female None None 18 8.9 (n.v. 12-16) 11 (n.v. 130-400) 9.2 (n.v. 4.8-10.8) 1.03 (n.v. 0.8-1.2) 26 (n.v. 24-36) 156 (n.v. 170-400) 29 (n.v. 0-25) 900 (n.v. 0-248) <7 (n.v. 36-145) 2.5 (n.v. 0.30-1.20) 0.90 (n.v. 0.51-0.95)

Gaetano Giuffrida,1* Annalisa Condorelli,1,2* Mary Ann Di Giorgio,1,2 Uros Markovic,1-3 Roberta Sciortino,1,2 Daniela Nicolosi1 and Francesco Di Raimondo1 *These authors contributed equally to the work 1 Division of Hematology, AOU "Policlinico G. Rodolico-San Marco", Via Santa Sofia 78, 95124 Catania, Italy; 2Postgraduate School of Hematology, University of Catania, Italy and 3Unità Operativa di Oncoematologia e BMT Unit, Istituto Oncologico del Mediterraneo, Viagrande, Italy Correspondence: ANNALISA CONDORELLI condorelli.1312@gmail.com 1009

Case Reports

doi:10.3324/haematol.2021.279535 Received: June 30, 2021. Accepted: July 21, 2021. Pre-published: August 12, 2021. Disclosures: no conflicts of interest to disclose. Ethical statement: informed consent was obtained from the patients regarding the report of their clinical scenario. Contributions: AC, MDG, UM, RS wrote the original draft; DN performed diagnostic tests; GG and FDR revised the paper critically and approved the final version for submission.

References 1. Miesbach W, Menne J, Bommer M, et al. Incidence of acquired thrombotic thrombocytopenic purpura in Germany: a hospital level study. Orphanet J Rare Dis. 2019;14(1):260. 2. Sukumar S, Lämmle B, Cataland SR. Thrombotic thrombocytopenic purpura: pathophysiology, diagnosis, and management. J Clin Med. 2021;10(3):536. 3. Yavaşoğlu I. Vaccination and thrombotic thrombocytopenic purpura. Turkish J Hematol. 2020;37(3):218-219. 4. Yocum A, Simon EL. Thrombotic thrombocytopenic purpura after Ad26.COV2-S vaccination. Am J Emerg Med. 2021:441.e3441.e4. 5. Sissa C, Al-Khaffaf A, Frattini F, et al. Relapse of thrombotic throm-

1010

bocytopenic purpura after COVID-19 vaccine. Transfus Apher Sci. 2021:103145. 6. Oliveira DS, Lima TG, Benevides FLN, et al. Plasmic score applicability for the diagnosis of thrombotic microangiopathy associated with ADAMTS13-acquired deficiency in a developing country. Hematol Transfus Cell Ther. 2019;41(2):119-124. 7. Polack FP, Thomas SJ, Kitchin N, et al. Safety and efficacy of the BNT162b2 mRNA Covid-19 vaccine. N Engl J Med. 2020; 383(27):2603-2615 8. Guimarães LE, Baker B, Perricone C, Shoenfeld Y. Vaccines, adjuvants and autoimmunity. Pharmacol Res. 2015;100:190-209. 9. Watad A, De Marco G, Mahajna H, et al. Immune-mediated disease flares or new-onset disease in 27 subjects following mRNA/DNA SARS-CoV-2 vaccination. Vaccines. 2021;9(5):435. 10. Ïremli BG, Şendur SN, Ünlütürk U. Three cases of subacute thyroiditis following SARS-CoV-2 vaccine: post-vaccination ASIA syndrome. J Clin Endocrinol Metab. 2021;106(9):2600-2605. 11. Condorelli A, Markovic U, Sciortino R, Di Giorgio MA, Nicolosi D, Giuffrida G. Immune thrombocytopenic purpura cases following COVID-19 vaccination. Mediterr J Hematol Infect Dis. 2021; 13(1):e2021047. 12. Helms JM, Ansteatt KT, Roberts JC, et al. Severe, refractory immune thrombocytopenia occurring after SARS-CoV-2 vaccine. J Blood Med. 2021;12:221-224. 13. Candelli M, Rossi E, Valletta F, De Stefano V, Francheschi F. Immune thrombocytopenic purpura after SARS-CoV-2 vaccine. Br J Haematol. 2021;194(3):547-549. 14. Cines DB, Bussel JB. SARS-CoV-2 vaccine-induced immune thrombotic thrombocytopenia. N Engl J Med. 2021;384(23):2254-2256.

haematologica | 2022; 107(4)

Case Reports

VEXAS syndrome in a female patient with constitutional 45,X (Turner syndrome) VEXAS (vacuoles, E1 enzyme, X-linked, autoinflammatory, somatic) syndrome is a newly-defined autoinflammatory disorder that arises from somatic mutations affecting UBA1, a major E1 enzyme that initiates ubiquitinylation and is important for the maturation of autophagic vacuoles.1,2 Specifically, these mutations occur in hematopoietic stem cells and in the erythroid and granulocyte precursors in the bone marrow.1 This results in an autoinflammatory syndrome characterized by recurrent fevers, cytopenias, chondritis, vasculitis, pulmonary inflammation, and neutrophilic dermatoses. This inflammatory syndrome is typically treatment-refractory, with patients being persistently steroid-dependent, and is often fatal: 40% of patients were deceased at the time of inclusion in the original study.1 There is also a strong association with hematologic malignancy. In the original series, 24% of patients developed a low-grade myelodysplastic syndrome (MDS) while 20% had multiple myeloma or a monoclonal gammopathy.1 Patients who progress to MDS tend to have more prominent thrombocytopenia and neutropenia, and one of the most commonly co-occurring somatic mutations appears to be in DNMT3A.3-5 Interestingly, the bone marrow morphology demonstrates a characteristic cytoplasmic vacuolation restricted to the erythroid and granulocyte precursors.6 The UBA1 gene lies on the X-chromosome, making VEXAS an X-linked syndrome. Consistent with this, the majority of reported cases have occurred in biological males. The presence of a second UBA1 allele in females is thought to mitigate the impact of the presence of the mutant allele should it arise.1 Part of the reason the second allele is protective against VEXAS in women is related to the fact that UBA1 does not undergo X-inactivation; in contrast, other disease-causing mutations in genes on the X-chromosome that do undergo inactivation (e.g., PIGA mutations in Paroxysmal Nocturnal Haemoglobinuria) manifest disease at similar rates in males and females.7,8 However, VEXAS has been described in a small number of women (N=4).9-11 These women were observed to have acquired monosomy X due to age-related mosaicism in the X chromosome.12 In contrast to acquired mosaicism of the X chromosome, Turner syndrome refers to a constitutional loss of one X chromosome (i.e., 45,X karyotype). This occurs in between 1 in 2,000 to 1 in 2,500 live female births and

results in short stature, skeletal abnormalities, primary ovarian failure, and several other end-organ complications.13 As with acquired X chromosome mosaicism, constitutional loss of the X chromosome may also predispose to female VEXAS. In this case report, we present the first case of female VEXAS syndrome diagnosed in a patient with constitutional 45,X (Turner syndrome). A 67 year-old female was referred with a diagnosis of myelodysplastic syndrome with multilineage dysplasia (MDS-MLD), originally low-risk by IPSS (score 0) and very low-risk by IPSS-R (score 1), which was diagnosed 36 months prior. Peripheral blood counts at relevant timepoints are demonstrated in Table 1. Her bone marrow cytogenetics demonstrated 45,X (20/20) and a targeted capture panel demonstrated variants of uncertain significance in DNMT3A (VAF 28.0%) and SMC3 (48.0%). The patient’s medical history was notable for relapsing polychondritis, diagnosed five years previous, and constitutional 45,X (Turner syndrome) that was diagnosed in her teenage years after a failure of pubertal development. Her relapsing polychondritis had presented with a migratory inflammatory arthritis and auricular inflammation. This was treated initially with prednisone; she had trialed multiple steroid-sparing agents (methotrexate, azathioprine, cyclosporine) but was unable to wean from steroids. She had remained dependent on a low dose of prednisone (10mg) from the time of her original diagnosis; her disease had remained stable with a combination of steroids and etanercept. In addition, she had recently been found to have a monoclonal gammopathy of unclear significance (MGUS), with a very faint (unquantifiable) IgG lambda monoclonal band on immunofixation. Her low-risk MDS was observed until she developed transfusion dependence 24-months after her initial diagnosis. This was managed with erythropoietin injections, which reduced her transfusion requirements for eight months’ time. After failing erythropoietin, a repeat bone marrow biopsy was performed (32 months); this marrow was hypercellular (95%) and demonstrated more prominent trilineage dysplasia than her previous diagnostic marrow. Interestingly, her bone marrow morphology demonstrated multiple vacuolated erythroid and granulocytic precursor cells (Figure 1A). The number of blast cells was unchanged (2% of total nucleated cells), hence her diagnosis was rendered as persistence of MDS-MLD. Her cytogenetics were unchanged (45,X [20/20]) (Figure 1B) and her targeted capture panel re-demonstrated variants

Table 1. Peripheral blood counts at relevant timepoints throughout the disease course of a female VEXAS patient with Myelodysplastic syndrome (MDS)

Parameter Timepoint (months) Hemoglobin (g/L) Mean cell volume (fL) Platelets (x109/L) White blood cell (x109/L) Neutrophils (x109/L) Lymphocytes (x109/L) Monocytes (x109/L) Eosinophils (x109/L) Basophils (x109/L) BM blasts Karyotype

MDS diagnosis 0 107 109.1 137 6.60 5.60 0.80 0.10 0.10 0.00 1% 45,X

MDS progression 32 101 101 35 5.30 3.98 1.06 0.05 0.11 0.00 2% 45,X

Transplant referral 36 92 99 40 6.90 5.10 0.90 0.30 0.00 0.10 NA NA

On treatment (HMA) 43 188 100 102 2.60 1.50 1.00 0.10 0.10 0.00 1% 45,X

Abbreviations: VEXAS: vacuoles, E1 enzyme, X-linked, autoinflammatory, somatic; HMA: hypomethylating agent; BM: bone marrow.

haematologica | 2022; 107(4)

1011

Case Reports

Figure 1. Representative diagnostic features of a female patient with constitutional 45,X and VEXAS syndrome (A) A diagnostic bone marrow aspirate was collected and stained with Giemsa-Wright. There is trilineage dysplasia, consistent with the known diagnosis of myelodysplastic syndrome with multilineage dysplasia. There is notable vacuolation of the erythroid and granulocytic precursors, indicated by the black arrows. (B) Karyotyping with G-banding demonstrated X monosomy, consistent with the patient’s reported history of constitutional 45,X (Turner syndrome). No clonal evolution was identified. A single X-chromosome is indicated by the black arrow. (C) Sanger sequencing for the UBA1 locus performed on peripheral blood demonstrated the presence of a somatic UBA1 p.Met41Thr mutation (c.122T>C) with approximately equal allele frequency as the reference allele, confirming the diagnosis of VEXAS syndrome; the Sanger chromatogram is demonstrated.

of uncertain significance in DNMT3A (VAF 30.0%) and SMC3 (VAF 43.8%). With the deterioration in her peripheral counts, her IPSS-R score was now intermediate-risk (score 3.5) and her IPSS score intermediate-1 (score 0.5). Given the recently described association between lowrisk MDS/MGUS, bone marrow vacuolation, and a steroid-dependent autoinflammatory syndrome resembling relapsing polychondritis (VEXAS syndrome) caused by somatic mutations in the X-linked E1 enzyme UBA1, further diagnostic testing was pursued by referral of this patient to the National Institutes of Health (NIH). Sanger sequencing was performed at the NIH; genomic DNA was prepared from peripheral blood using the Maxwell 16 Blood DNA purification kit (Promega). Coding exons of UBA1 were sequenced using the BigDye Terminator v1.1 Cycle Sequencing Kit (Applied Biosystems), and sequencing data analyzed using Sequencher (Gene Codes) and 4Peaks (Mekentosj). Sanger sequencing results were positive for the presence of a somatic UBA1 p.Met41Thr mutation (c.122T>C) with approximately equal allele frequency as the reference allele, (Figure 1C) confirming the diagnosis of VEXAS given her X monosomy. With her transfusion dependence, the increase in IPSS-R score to intermediate, and the persistent steroid requirement due to VEXAS, it was recommended that the patient be initiated on hypomethylating agent therapy with 5-azacitidine. She tolerated 5-azacitidine well and achieved transfusion independence with an improvement in both hemoglobin and platelets (43 months). Human leukocyte antigen (HLA) typing was initiated for the patient and her siblings and an unrelated donor search activated. Given her relatively young age, a sibling-donor allogeneic hematopoietic stem cell transplant is planned as a definitive therapy for both her MDS and VEXAS syndrome. Although VEXAS syndrome primarily affects males as an X-linked disease, it is important for the clinician to realize that there are several factors that can lead VEXAS (and other X-linked diseases) to manifest in females. Mosaicism of the X chromosome is an age-related phenomenon that primarily affects the inactivated X chromosome; it has been shown to occur in 0.11% of 50 yearold females, with this increasing to 0.45% of 75 yearolds.12 This has been the most commonly reported reason 1012

underlying a diagnosis of VEXAS in a female. In contrast, the underlying reason our patient was predisposed to developing VEXAS was the presence of a constitutional 45,X karyotype from birth, which was diagnosed in her teenage years after a failure of pubertal development. There are other mechanisms that can lead to the functional loss of one X chromosome, such as uniparental disomy and skewed X-inactivation; in the future we may see these reported as an underlying risk factor in female VEXAS patients. While clinicians are increasingly aware of VEXAS syndrome as a diagnostic entity, it is still largely characterized as an X-linked disorder affecting only males. However, as our case demonstrates, it is important to be aware that VEXAS can affect females and further investigations should be pursued in the appropriate context for female patients. In our case, the diagnosis of Turner syndrome was established long before the development of her VEXAS syndrome. The underlying risk factor for female VEXAS may not always be obvious, however, as in patients with acquired X chromosome mosaicism. It is important to be alert for the presence of disorders that result in the inactivation of the X-chromosome when assessing female patients with a possible X-linked disorder, such as VEXAS syndrome. Ryan J. Stubbins,1,2 Eric McGinnis,3 Bhupinder Johal,4 Luke YC Chen,2 Lorena Wilson,5 Daniela Ospina Cardona5 and Thomas J. Nevill1,2 1 Leukemia/BMT Program of BC, BC Cancer, Vancouver, BC, Canada; 2Division of Hematology, Department of Medicine, University of British Columbia, Vancouver, BC, Canada; 3Department of Pathology and Laboratory Medicine, University of British Columbia, Vancouver, BC, Canada; 4Department of Pathology, Kelowna General Hospital, Kelowna, BC, Canada and 5National Human Genome Research Institute, National Institutes of Health, Bethesda, MD, USA Correspondence: RYAN J. STUBBINS- ryan.stubbins1@bccancer.bc.ca doi:10.3324/haematol.2021.280238 Received: October 25, 2021. Accepted: December 3, 2021. haematologica | 2022; 107(4)

Case Reports

Pre-published: December 16, 2021. Disclosures: no conflicts of interest to disclose. Contributions: RJS and TJN conceived the study; RJS, EM, BJ, LW and DOC collected data and images; RJS wrote the manuscript. All authors read, critically assessed, and approved the final manuscript. Acknowledgments: we thank Dr. David Beck and the National Institutes of Health for their collaboration and assistance with this manuscript. Funding: RJS is supported by grants from the Leukaemia Lymphoma Society of Canada (20200LFC-439884), Canadian Institutes of Health Research, and the University of British Columbia Clinician Investigator Program. Data-sharing statement: data is available, upon request, from the corresponding author. Trial registration and ethics: informed consent was obtained from the patient for this case report.

References 1. Beck DB, Ferrada MA, Sikora KA, et al. Somatic mutations in UBA1 and severe adult-onset autoinflammatory disease. N Engl J Med. 2020;383(27):2628-2638. 2. Lenk SE, Dunn WA, Trausch JS, Ciechanover A, Schwartz AL. Ubiquitin-activating enzyme, E1, is associated with maturation of autophagic vacuoles. J Cell Biol. 1992;118(2):301-308. 3. Obiorah IE, Patel BA, Groarke EM, et al. Benign and malignant hematologic manifestations in patients with VEXAS syndrome due to somatic mutations in UBA1. Blood Adv. 2021;5(16):3203-3215.

haematologica | 2022; 107(4)

4. Georgin-Lavialle S, Terrier B, Guedon AF, et al. Further characterization of clinical and laboratory features occurring in VEXAS syndrome in a large-scale analysis of multicenter case-series of 116 French patients. Br J Dermatol., October 10, 2021, https://doi.org/10.1111/bjd.20805, [Epub ahead of print]. 5. van der Made CI, Potjewijd J, Hoogstins A, et al. Adult-onset autoinflammation caused by somatic mutations in UBA1: A Dutch case series of patients with VEXAS. J Allergy Clin Immunol. 2022;149(1):432-439.e4. 6. Dehghan N, Marcon KM, Sedlic T, Beck DB, Dutz JP, Chen LYC. Vacuoles, E1 enzyme, X-linked, autoinflammatory, somatic (VEXAS) syndrome: fevers, myalgia, arthralgia, auricular chondritis, and erythema nodosum. Lancet. 2021;398(10300):621. 7. Carrel L, Clemson CM, Dunn JM, et al. X inactivation analysis and DNA methylation studies of the ubiquitin activating enzyme E1 and PCTAIRE-1 genes in human and mouse. Hum Mol Genet. 1996;5(3):391-401. 8. Luzzatto L, Risitano AM, Notaro R. Mutant UBA1 and Severe AdultOnset Autoinflammatory Disease. N Engl J Med. 2021;384(22):2164. 9. Arlet JB, Terrier B, Kosmider O. Mutant UBA1 and severe adult-onset autoinflammatory disease. N Engl J Med. 2021;384(22):2163. 10. Tsuchida N, Kunishita Y, Uchiyama Y, et al. Pathogenic UBA1 variants associated with VEXAS syndrome in Japanese patients with relapsing polychondritis. Ann Rheum Dis., March 31, 2021, https://doi.org/10.1136/annrheumdis-2021-220089 [Epub ahead of print]. 11. Barba T, Jamilloux Y, Durel CA, et al. VEXAS syndrome in a woman. Rheumatology (Oxford). 2021;60(11):e402-e403. 12. Machiela MJ, Zhou W, Karlins E, et al. Female chromosome X mosaicism is age-related and preferentially affects the inactivated X chromosome. Nat Commun. 2016;7:11843. 13. Gravholt CH, Viuff MH, Brun S, Stochholm K, Andersen NH. Turner syndrome: mechanisms and management. Nat Rev Endocrinol. 2019;15(10):601-614.

1013

Case Reports

A case series of primary cutaneous B-cell lymphomas with atypical presentations: diagnostic and therapeutic challenges Primary cutaneous B-cell lymphomas (PCBCL) are defined as B-cell lymphomas of the skin without nodal, bone marrow, or visceral involvement at the time of diagnosis.1 They represent approximately 25% of primary cutaneous lymphomas.1,2 The histopathological diagnosis of PCBCL can be challenging in certain instances in which overlapping features are present. Nevertheless, identification of the correct subtype of PCBCL is imperative for determining the prognosis and avoiding inappropriate aggressive treatments which could lead to unnecessary morbidity.3 Here we present a series of three cases that highlight the distinguishing features between two subtypes of PCBCL. The first case was a 57-year-old African American man with multiple pruritic nodules on his abdomen that appeared and rapidly progressed in size 6 months prior to presentation (Figure 1A). Review of systems was negative for pain, weight loss, night sweats or fever. Blood flow cytometry analysis was normal, and positron emission tomography (PET) scan was negative for metabolically active lymph nodes or systemic disease. Skin biopsy demonstrated sheets of large atypical lymphoid aggregates with centroblast morphology extending into the entire thickness of the dermis with an accompanying significant reactive small-sized lymphocytic infiltrate (Figure 1B). The large cells expressed CD20 and BCL6 but were negative for CD10, BCL2, and MUM-1 on immunostaining (Figure 1D-H). Ki67 stain showed a proliferation rate of 30-40%. B-cell receptor clonality assay (ClonoSeq) identified two dominant immunoglobulin heavy chain sequences. An initial diagnosis of diffuse large B-cell lymphoma (DLBCL) was considered, but a complete histopathological review with clinical correlation led to a final diagnosis of primary cutaneous follicle center lymphoma (PCFCL) with diffuse pattern. Three large lesions were excised and local radiation therapy provided total regression of the rest of the abdominal tumors without

recurrence to date (2 years). The second case was an 83-year-old Caucasian male presenting with an erythematous, asymptomatic growth on the forehead that had increased in size to 6 x 7 cm over 6 months (Figure 2A). The patient had no systemic symptoms of fever, night sweats, weight loss, or lymphadenopathy. Punch biopsy of the lesion demonstrated a dense lymphocytic neoplasm composed of centroblasts and immunoblasts positive for CD10, CD20, and BCL6 but negative for MUM1. Small-sized reactive lymphocytes were BCL2-positive (Figure 2B, D-H). Immunoglobulin heavy chain gene-rearrangement studies revealed a monoclonal population. Blood flow cytometry analysis and PET scan were negative for blood or systemic involvement. This case was also initially diagnosed as DLBCL, but a secondary histopathological review combined with clinical correlation led to a diagnosis of PCFCL with diffuse pattern. Radiation therapy provided complete regression of the tumor without recurrence to date (1.5 years). The third case was a 72-year-old man with a history of chronic kidney disease and heart failure who presented with a tender pink tumor on his scalp that abruptly appeared as a small papule but rapidly grew in size to 4 x 4 cm over 1 month (Figure 3A). His clinical history was negative for fever, lymphadenopathy, fatigue, night sweats, and systemic symptoms. Punch biopsy revealed a sheet-like diffuse dense infiltrate composed of cells with immunoblastic morphology and high mitotic activity (Figure 3B). The atypical lymphocytes stained positive for CD10, CD20, BCL2, BCL6, and MUM1 (Figure 3D-H) with a more than 80% proliferative population based on Ki-67 positivity. Fluorescent in-situ hybridization (FISH) was positive for a BCL6 gene rearrangement (18% of nuclei) and negative for rearrangement of MYC, CCND1, and BCL2. A dominant immunoglobulin heavy chain sequence present in 99% of all nucleated cells was identified by immunosequencing (ClonoSeq). PET scan revealed a hypermetabolic scalp lesion with no evidence of lymph node or systemic involvement, and blood flow cytometry analysis was normal. The patient was diagnosed with primary cutaneous diffuse large B-cell lym-

Figure 1. Primary cutaneous follicle center lymphoma identified by various features. (A) Three firm, ill-defined, pink tumors, 2 to 3 cm in size, with surrounding erythematous plaques and significant induration on the mid abdomen. (B, C) Dense lymphocytic infiltrate of small to medium sized lymphocytes with condensed nuclei (hematoxylin & eosin: [B] 40x, [C] 100x) that are (D) CD10-negative, (E) CD20-positive, (F) BCL2-negative, (G) BCL6-positive, and (H) MUM1-negative.

1014

haematologica | 2022; 107(4)

Case Reports

Figure 2. A different presentation of primary cutaneous follicle center lymphoma. (A) A single firm erythematous tumor on the right of the forehead. (B, C) Dense infiltrate of medium-sized centroblasts and immunoblasts in the dermis (hematoxylin & eosin: [B] 40x, [C] 100x) that are (D) CD10-positive, (E) CD20positive, (F) BCL2-negative on large cells and BCL2-positive on reactive cells, (G) BCL6-positive, and (H) MUM1-negative.

phoma, leg type (PCDLBCL, LT). Combination chemotherapy with rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) was not initiated due to the patient’s poor ejection fraction (48%). Instead, given his comorbidities and life expectancy, he received radiation therapy with complete resolution of the lesion, confirmed with PET scan. Due to his aggressive diagnosis, the patient was closely monitored by oncology without recurrence for 2 years until he died from a cardiac arrest due to his comorbidities. The three cases presented here highlight the challenge of distinguishing between PCFCL and PCDLBCL-LT, two of the three main subtypes of PCBCL. The first two cases of PCFCL were originally diagnosed as DLBCL without specification. The correct diagnosis of diffuse PCFCL was made after a secondary histological consultation along with clinical correlation.4,5 The third case illustrates the fact that a PCDLBCL, LT can present on the scalp. For the first two cases, the absence of a follicular pattern and diffuse sheets of atypical large cells gave an initial impression of DLBCL while the lack of expression of MUM1 and the presence of a reactive infiltrate clearly pointed to the diagnosis of PCFCL.5,6 Although all cases showed a diffuse infiltrate on histology, their histomorphology, immunophenotype, pattern of molecular aberration on FISH analysis, and clinical presentation distinguished the correct diagnosis. This underscores the importance of recognizing PCFCL with diffuse pattern to avoid overcalling DLBCL and the resulting unnecessary aggressive treatment. Histomorphology should be investigated in detail to arrive at the correct diagnosis among cases of PCBCL.3 Although diffuse sheets of cells were seen in all cases, a close evaluation of cellular morphology clearly distinguishes DLBCL from other subtypes. Large cells with multiple mitotic figures and nuclear atypia in PCDLBCL, LT contrast directly with the smaller cells and condensed nuclei seen in PCFCL cases (Figures 1C, 2C, and 3C).6 Immunohistochemistry is an essential tool in diagnosing subtypes of PCBCL. MUM-1 positivity precludes PCFCL and must be investigated before making a diagnohaematologica | 2022; 107(4)

sis of any subtype of PCBCL. BCL-2 is not expressed by malignant cells in PCFCL, but it may be present in reactive T cells.3 In the second case of the series presented here, BCL-2 was originally called positive but upon further evaluation it was clear that only reactive cells expressed BCL-2. FISH studies can be utilized in cases of PCDLBCL, LT, and we found a positive BCL6 gene rearrangement in our case. Overall, these cases demonstrate the architectural, histomorphological, and immunohistochemical features that can distinguish PCFCL from PCDLBCL, LT and highlight the diagnostic challenges that arise as a result of overlapping characteristics. The clinical impact of this overlap is most acutely felt in PCDLBCL, LT, because of its more aggressive course and the fact that radiation therapy alone is generally considered inadequate.7,8 While there is currently no evidence-based standard of care, most cases of PCDLBCL, LT are treated as systemic DLBCL, with R-CHOP chemoimmunotherapy, often with central nervous system prophylaxis, due to the high risk of central nervous system dissemination.9 The addition of radiation therapy to chemoimmunotherapy was found to be important in a recent case series.10 Despite historical data showing that the outcomes of patients with PCDLBCL,LT have improved since the introduction of modern chemoimmunotherapy, outcomes remain relatively poor.11 In addition, many patients are unfit for chemotherapy, due to age or comorbidities. In the cohort reported by Grange et al. in 2014 about 50% of the patients were older than 80 years. At the moment, frontline radiation therapy, especially for localized, unifocal disease, is an acceptable option for elderly and frail patients, and some patients, including the third case presented here, have durable responses and long-progression free survival.12,13 The clinical and histological findings of B-cell lymphomas can vary widely. The clinical picture in the first case reminds readers that it is possible to have multiple lesions in PCFCL, including multifocal lesions, although it is often thought to present as a solitary lesion. The second case emphasizes that histological and immunohisto1015

Case Reports

Figure 3. Primary cutaneous diffuse large B-cell lymphoma, leg type is defined by features seen in this figure. (A) A well-defined tumor on the left parietal scalp with (B, C) diffuse proliferation of large polygonal lymphocytes with high mitotic activity, large nuclei, and little cytoplasm (hematoxylin & eosin: [B] 40x, [C] 100x) that are (D) CD10-positive, (E) CD20-positive, (F) BCL2-positive, (G) BCL6-positive and (H) MUM1-positive.

chemical results must be assessed together, without relying on one over the other. Case three cautions physicians that PCDLBCL, LT can occur elsewhere on the body while having overlapping histological features with PCFCL. Due to the complexity of cutaneous lymphomas, it is imperative for physicians to work together in a multidisciplinary team with dermatology, oncology, dermatopathology and radiation oncology in order to provide the best care for these patients. Emily Correia,1 Jisun Cha,1 Shalini Krishnasamy,1,2 Megan O’Donnell,1 Wenyin Shi,3 Pierluigi Porcu2 and Neda Nikbakht1 1 Department of Dermatology and Cutaneous Biology, Thomas Jefferson University, Philadelphia, PA; 2Division of Hematologic Malignancies and HSCT, Department of Medical Oncology, Sidney Kimmel Cancer Center, Thomas Jefferson University, Philadelphia, PA and 3Department of Radiation Oncology, Thomas Jefferson University, Philadelphia PA, USA. Correspondence: NEDA NIKBAKHT - neda.nikbakht@jefferson.edu PIERLUIGI PORCU - pierluigi.porcu@jefferson.edu doi:10.3324/haematol.2021.279992 Received: September 13, 2021. Accepted: December 2, 2021 Pre-published: December 16, 2021. Disclosures: no conflicts of interest to disclose Contributions: EC performed the literature review, wrote the manuscript, and collected illustrations; JC performed histology, helped to write the manuscript, and led discussions of the manuscript; SK and MOD wrote parts of the manuscript; WS wrote part of the manuscript and edited it; PP edited the manuscript; and NN wrote part of the manuscript, edited it, helped with the histology, and led discussions of the manuscript. Funding: NN is supported by a Skin Cancer Foundation Todd Nagel Memorial Research Grant.

1016

References 1. Suárez AL, Pulitzer M, Horwitz S, Moskowitz A, Querfeld C, Myskowski PL. Primary cutaneous B-cell lymphomas: part I. Clinical features, diagnosis, and classification. J Am Acad Dermatol. 2013;69(3):329. 2. Mehta-Shah N, Horwitz SM, Ansell S, et al. NCCN guidelines insights: primary cutaneous lymphomas, version 2.2020. J Natl Compr Canc Netw. 2020;18(5):522-536. 3. Willemze R, Cerroni L, Kempf W, et al. The 2018 update of the WHO-EORTC classification for primary cutaneous lymphomas. Blood. 2019;133(16):1703-1714. 4. Malachowski SJ, Sun J, Chen PL, Seminario-Vidal L. Diagnosis and management of cutaneous B-cell lymphomas. Dermatol Clin. 2019;37(4):443-454. 5. Kodama K, Massone C, Chott A, Metze D, Kerl H, Cerroni L. Primary cutaneous large B-cell lymphomas: clinicopathologic features, classification, and prognostic factors in a large series of patients. Blood. 2005;106(7):2491-2497. 6. Felcht M, Klemke CD, Nicolay JP, et al. Primary cutaneous diffuse large B-cell lymphoma, NOS and leg type: Clinical, morphologic and prognostic differences. J Dtsch Dermatol Ges. 2019;17(3):275-285. 7. Senff NJ, Hoefnagel JJ, Neelis KJ, et al. Results of radiotherapy in 153 primary cutaneous B-cell lymphomas classified according to the WHO-EORTC classification. Arch Dermatol. 2007;143(12):15201526. 8. Haverkos B, Tyler K, Gru AA, et al. Primary cutaneous B-cell lymphoma: management and patterns of recurrence at the Multimodality Cutaneous Lymphoma Clinic of The Ohio State University. Oncologist. 2015;20(10):1161-1166. 9. Sundriyal D, Arya L, Srivastava R, Walia M, Sehrawat A. Leptomeningeal relapse in primary ccutaneous DLBCL: implications for a prophylactic CNS therapy. Cancer Rep (Hoboken). 2021;4(1):e1295. 10. Kraft RM, Ansell SM, Villasboas JC, et al. Outcomes in primary cutaneous diffuse large B-cell lymphoma, leg type. J Clin Oncol. 2021;39(15_suppl):e19547. 11. Grange F, Joly P, Barbe C, et al. Improvement of survival in patients with primary cutaneous diffuse large B-cell lymphoma, leg type, in France. JAMA Dermatol. 2014;150(5):535-541. 12. Graham PM, Richardson AS, Schapiro BL, Saunders MD, Stewart DM. Spontaneous regression of primary cutaneous diffuse large Bcell lymphoma, leg type with significant T-cell immune response. JAAD Case Rep. 2018;4(4):305-309. 13. Rodriguez-Pinilla SM, Santonja C, Stewart P, et al. Indolent clinical behaviour of primary cutaneous diffuse large B-cell lymphoma, leg type, with double MYC and BCL6 gene rearrangement. Br J Haematol. 2020;191(3):e83-e86.

haematologica | 2022; 107(4)

haematologica — Vol. 107 n. 4 — April 2022 —781-1016