Haematologica, Volume 107, Issue 8 by Haematologica

Table of Contents Volume 107, Issue 8: August 2022 About the Cover 1736

Images from the Haematologica Atlas of Hematologic Cytology: congenital dyserythropoietic anemia type II Rosangela Invernizzi https://doi.org/10.3324/haematol.2022.281481

Landmark Papers in Hematology 1737

The polymerase chain reaction, so simple, so clever: the discovery that made minimal residual disease come true Hélène Cavé https://doi.org/10.3324/haematol.2022.281413

1739

Expanding approaches to detect clonal hematopoiesis M.A. Wasay Khan and Alexander G. Bick https://doi.org/10.3324/haematol.2021.279818

1741

Kinetics matter: prognostic implications of early bone marrow assessment in acute myeloid leukemia Florian Perner https://doi.org/10.3324/haematol.2021.280133

1743

Decoding m6A, one reader at a time Hanzhi Luo and Michael G. Kharas https://doi.org/10.3324/haematol.2021.280166

1746

Acute Lymphoblastic Leukemia Rational drug combinations with CDK4/6 inhibitors in acute lymphoblastic leukemia Karen L. Bride et al. https://doi.org/10.3324 haematol.2021.279410

1758

Acute Myeloid Leukemia The retinoic acid receptor co-factor NRIP1 is uniquely upregulated and represents a therapeutic target in acute myeloid leukemia with chromosome 3q rearrangements Sarah Grasedieck et al. https://doi.org/10.3324/haematol.2020.276048

1773

Acute Myeloid Leukemia Acute myeloid leukemia: negative prognostic impact of early blast persistence can be in part overcome by a later remission prior to post-induction therapy Jana Ihlow et al. https://doi.org/10.3324/haematol.2021.279134

Editorials

Articles

Haematologica | 107 - August 2022

III

1786

Acute Myeloid Leukemia A novel CD34-specific T-cell engager efficiently depletes acute myeloid leukemia and leukemic stem cells in vitro and in vivo Lucas C. M. Arruda et al. https://doi.org/10.3324/haematol.2021.279486

1796

Chronic Lymphocytic Leukemia B-cell antigen receptor expression and phosphatidylinositol 3-kinase signaling regulate genesis and maintenance of mouse chronic lymphocytic leukemia Vera Kristin Schmid et al. https://doi.org/10.3324/haematol.2021.279924

1815

Hematopoiesis Utility of plasma cell-free DNA for de novo detection and quantification of clonal hematopoiesis Fernanda Gutierrez-Rodrigues et al. https://doi.org/10.3324/haematol.2021.279230

1827

Hemostasis Syntaxin 5 determines Weibel-Palade body size and von Willebrand factor secretion by controlling Golgi architecture Marije Kat et al. https://doi.org/10.3324/haematol.2021.280121

1840

Immunology Antibody response after vaccination against SARS-CoV-2 in adults with hematological malignancies: a systematic review and meta-analysis Nico Gagelmann et al. https://doi.org/10.3324/haematol.2021.280163

1850

Non-Hodgkin Lymphoma Mutational landscape of high-grade B-cell lymphoma with MYC-, BCL2 and/or BCL6 rearrangements characterized by whole-exome sequencing Axel Künstner et al. https://doi.org/10.3324/haematol.2021.279631

1864

Non-Hodgkin Lymphoma Immune pathway upregulation and lower genomic instability distinguish EBV-positive nodal T/NK-cell lymphoma from ENKTL and PTCL-NOS Cho Mar Myint Wai et al. https://doi.org/10.3324 haematol.2021.280003

1880

Non-Hodgkin Lymphoma Programmed cell death ligand 1 expression in aggressive pediatric non-Hodgkin lymphomas: frequency, genetic mechanisms, and clinical significance Kevin E. Fisher et al. https://doi.org/10.3324/haematol.2021.280342

1891

Plasma Cell Disorders Comprehensive genomic analysis of refractory multiple myeloma reveals a complex mutational landscape associated with drug resistance and novel therapeutic vulnerabilities Nicola Giesen et al. https://doi.org/10.3324/haematol.2021.279360

1902

Platelet Biology & its Disorders Sorting nexin 24 is required for α-granule biogenesis and cargo delivery in megakaryocytes Joanne Lacey et al. https://doi.org/10.3324/haematol.2021.279636

1914

Red Cell Biology & its Disorders Effects of corticosteroids in patients with sickle cell disease and acute complications: a systematic review and meta-analysis Julien Lopinto et al. https://doi.org/10.3324/haematol.2021.280105

Haematologica | 107 - August 2022

Letters to the Editor 1924

m6A reader Ythdf3 protects hematopoietic stem cell integrity under stress by promoting the translation of Foxm1 and Asxl1 transcripts Qinglin Dang et al. https://doi.org/10.3324/haematol.2021.279300

1928

Cell-free DNA sequencing as a potential screening tool for phase I targeted treatment in refractory/relapse diffuse large B-cell lymphoma Cyril Quivoron et al. https://doi.org/10.3324/haematol.2021.280464

1933

Full versus prophylactic-intermediate doses of anticoagulants in COVID-19: a meta-analysis Lorenzo Loffredo et al. https://doi.org/10.3324/haematol.2022.280652

1940

The use of hydroxyurea pretreatment in chronic myeloid leukemia in the current tyrosine kinase inhibitor era Camille C.B. Kockerols et al. https://doi.org/10.3324/haematol.2021.280501

1944

Preclinical evaluation of the preservation of red blood cell concentrates by hypoxic storage technology for transfusion in sickle cell disease Laura Bencheikh et al. https://doi.org/10.3324/haematol.2021.279721

1950

Hemostatic and protein C pathway dysfunction in the pathogenesis of experimental cerebral malaria Niamh O’Regan et al. https://doi.org/10.3324/haematol.2021.280450

1955

COVID-19 infection in acute lymphoblastic leukemia over 15 months of the pandemic. A Campus ALL report Sabina Chiaretti et al. https://doi.org/10.3324/haematol.2021.280289

1960

A novel BCMA CAR-T-cell therapy with optimized human scFv for treatment of relapsed/refractory multiple myeloma: results from phase I clinical trials Min Yang et al. https://doi.org/10.3324/haematol.2022.280629

1966

Low-dose tyrosine kinase inhibitors in patients with chronic myeloid leukemia: a retrospective study in China Yilin Chen et al. https://doi.org/10.3324/haematol.2022.280637

1971

Copy number alterations define outcome in Philadelphia chromosome-positive acute lymphoblastic leukemia Helena Hohtari et al. https://doi.org/10.3324/haematol.2021.280578

1977

PD-1/PD-L1 expression is frequent and correlated with lymphocyte density in Erdheim-Chester disease Fréderic Charlotte et al. https://doi.org/10.3324/haematol.2021.280312

1981

Blood cell and marrow changes in patients with Kikuchi disease Shan-Chi Yu et al. https://doi.org/10.3324/haematol.2022.280746

Haematologica | 107 - August 2022

Comment 1986

Alloimmunization against Fy3 is a serious threat in the era of cell therapy Baptiste Lemaire and Sophie Waldvogel Abramowski https://doi.org/10.3324/haematol.2022.280632

Response to comment 1988

Intricacies of GATA-ca, continued Christine Lomas-Francis et al. https://doi.org/10.3324/haematol.2022.280876

1989

Occurrence of a paroxysmal nocturnal hemoglobinuria clone in an essential thrombocythemia: a link between PIGV and MPL Alexej Knaus et al. https://doi.org/10.3324/haematol.2021.279804

1994

Functional testing of relapsed chronic lymphocytic leukemia guides precision medicine and maps response and resistance mechanisms. An index case Sigrid S. Skånland et al. https://doi.org/10.3324//haematol.2021.280393

1999

Distinct genetic alterations in Burkitt-like lymphoma with 11q aberration and Burkitt lymphoma: a novel case report of composite lymphoma Meejeong Kim et al. https://doi.org/10.3324//haematol.2021.280543

2004

Treatment of leptomeningeal disease in blastic plasmacytoid dendritic cell neoplasm following tagraxofusp-erzs induction Deepak B. Vangala et al. https://doi.org/10.3324//haematol.2022.280843

Case Reports

Haematologica | 107 - August 2022

ABOUT THE COVER

R. Invernizzi

Images from the Haematologica Atlas of Hematologic Cytology: congenital dyserythropoietic anemia type II Rosangela Invernizzi University of Pavia, Pavia, Italy E-mail: rosangela.invernizzi@unipv.it https://doi.org/10.3324/haematol.2022.281481

Congenital dyserythropoietic anemias belong to a group of rare/very rare inherited conditions characterized by a maturation arrest during erythropoiesis with reduced reticulocyte production, which contrasts with the erythroid hyperplasia in the bone marrow. All these disorders are characterized by distinct morphological abnormalities of erythroblasts and result in congenital anemia of variable degree. Congenital dyserythropoietic anemia type II (CDA-II) is the most common of these inherited disorders. In CDA-II patients, anemia is generally normocytic and peripheral blood smears show moderate to marked anisopoikilocytosis including elliptocytes, spherocytes, dacryocytes, schistocytes, basophilic stippling and occasional nucleated red cells. Bone marrow smears are characterized by marked erythroid hyperplasia. The unique feature of CDA-II is the presence of a large number of bi- or multi-nucleated late erythroid precursors. The percentage of erythroblasts with more than one nucleus ranges from 10 to 30%. The figure shows that in this patient most binucleate cells are orthochromatic erythroblasts, but there are also a few more immature erythroblasts with two nuclei. Note also the lobulated nucleus of the erythroblast in the center. These abnormalities clearly indicate that incomplete cytokinesis is one of the features of erythroid cells in this condition. CDA-II is an autosomal recessive disease caused by biallelic mutations in the SEC238 gene. SEC23B is an essential component of coat protein complex II (COPII)-coated vesicles that transport secretory proteins from the endoplasmic reticulum to the Golgi complex in erythroid cells. Abnormalities in SEC23B activity may affect both erythroblast maturation/division in the bone marrow and red cell survival in the circulation.1 Disclosures No conflicts of interest to disclose.

References 1. Bianchi P, Invernizzi R, Holman C, et al. Congenital dyserythropoietic anemias. Haematologica. 2020;105(Suppl 1):177-183. Haematologica | 107 August 2022

1736

LANDMARK PAPER IN HEMATOLOGY

H. Cavé

The polymerase chain reaction, so simple, so clever: the discovery that made minimal residual disease come true Hélène Cavé Department of Genetics, University Hospital of Robert Debré and Université Paris-Cité, Paris, France E-mail: helene.cave@aphp.fr https://doi.org/10.3324/haematol.2022.281413 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

TITLE

Specific enzymatic amplification of DNA in vitro: the polymerase chain reaction.

AUTHORS Mullis K, Faloona F, Scharf S, Saiki R, Horn G, Erlich H. JOURNAL Cold Spring Harbor Symposia on Quantitative Biology 1986;51(Pt 1):263-273. PMID: 3472723. Today, every child with acute lymphoblastic leukemia (ALL) can benefit from minimal residual disease (MRD) monitoring, which enables therapy to be tailored according to the patient’s individual response to treatment. Before MRD monitoring become possible, hematologists already suspected that residual cells persisted in chil-

dren in cytological remission from their leukemia. But how can you detect what you cannot see? The answer came with the polymerase chain reaction (PCR).1 The first goal of Kary Mullis, a self-described ‘generalist with a chemical prejudice’ was to develop a prenatal diagnostic test for the b-globin mutation in sickle cell di-

Continued on following page. Haematologica | 107 August 2022

1737

LANDMARK PAPER IN HEMATOLOGY

H. Cavé

Figure 1. Polymerase chain reaction analysis for minimal residual disease assessment. Left: principle of minimal residual disease (MRD) detection based on clono-specific immunoglobulin (IG) and T-cell receptor (TCR) gene rearrangements. Characterization of IG/TCR rearrangements present in acute lymphoblastic leukemia (ALL) blasts at diagnosis enables a clono-specific probe or primer to be designed. At remission, this probe or primer can be used to detect patient’s residual blasts using either allele-specific polymerase chain reaction (PCR) or hybridization of PCR products with the allele-specific probe. Right: the Perkin-Elmer Cetus DNA Thermal Cycler, first commercial system automating the thermal cycling required in the PCR, introduced in 1987, and the principle of PCR as first reported by K Mullis.3

sease. Southern blot was cumbersome, with limited sensitivity. Kary Mullis began to think about how to make DNA-based diagnosis more practical. There were two possibilities: amplify the signal by improving the probing system or amplify the target. The latter option turned out to be the right one. This was the birth of the concept of PCR, which consists of the exponential amplification of a specific target DNA delimited by the position of two primers on a template DNA molecule.1 By ‘extracting’ the fragment of interest from the source DNA and amplifying it, PCR provides unlimited amounts of precise genetic material for diagnosis. Thanks to its principle of exponential amplification, PCR not only enabled genetic screening for sickle cell disease, but also provided a tool for detecting nucleic targets with an unprecedented level of sensitivity. The idea soon arose that it could make the detection of residual disease possible. But what PCR target should be used? What PCR target is specific to leukemia cells? Fusion genes had still only been identified in a minority of cases of ALL. Why not use T-cell receptor and immunoglobulin rearrangements? Such rearrangements, which are inherent to lymphocyte physiology, are found in virtually all ALL thus making it possible to track down the leukemic clones (Figure 1).

A sensitive technique, a specific marker, the first assay for monitoring MRD in leukemia was born.2 This enabled the detection of one leukemia cell among hundreds of thousands of normal cells. It would soon be shown that residual disease is still present in patients in cytological remission and is one of the most powerful prognostic factors in ALL. Is the invention of PCR a revolution? Probably not. Although considered one of the most important inventions of the late 20th century, it did not lead to a new paradigm. Actually, PCR emerged from a surprisingly simple conceptual idea. Let’s bet with Kary Mullis that many molecular biologists have since asked themselves with a tinge of regret, ‘Why didn't I think of that?’3 PCR is a fine illustration of a breakthrough made possible by a lot of ingenuity and multidisciplinarity, not forgetting... a moonlit walk in the mountains of California.3 Beyond MRD monitoring, PCR has certainly revolutionized diagnostic practice in hematology in many ways. It is for this reason, and because we must always remember the fruitful importance of multidisciplinarity, that the discovery of PCR deserves to be given the status of "landmark in hematology". Disclosures No conflicts of interest to disclose.

References 1. Mullis K, Faloona F, Scharf S, Saiki R, Horn G, Erlich H. Specific enzymatic amplification of DNA in vitro: the polymerase chain reaction. Cold Spring Harb Symp Quant Biol. 1986;51(Pt 1):263-273. 2. d'Auriol L, Macintyre E, Galibert F, Sigaux F. In vitro amplification

of T cell gamma gene rearrangements: a new tool for the assessment of minimal residual disease in acute lymphoblastic leukemias. Leukemia. 1989;3(2):155-158. 3. Mullis KB. The unusual origin of the polymerase chain reaction. Sci Am. 1990;262(4):56-61, 64-65.

Haematologica | 107 August 2022

1738

EDITORIAL

M.A.W. Khan and A.G. Bick

Expanding approaches to detect clonal hematopoiesis M.A. Wasay Khan and Alexander G. Bick Division of Genetic Medicine, Department of Medicine, Vanderbilt University School of Medicine and Vanderbilt University Medical Center, Nashville, TN, USA E-mail: alexander.bick@vumc.org https://doi.org/10.3324/haematol.2021.279818 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

Clonal hematopoiesis (CH) occurs when a single hematopoietic stem cell acquires a mutation that gives it a competitive advantage over other stem cells. CH is often driven by somatic mutations in genes that are recurrently mutated in myeloid malignancies such as DNMT3A, ASXL1, or TET2. CH is present at an appreciable fraction of the blood in more than 10% of people older than 70 years and is a precursor to hematologic cancers. The rate of development of neoplasia in patients with CH is estimated to be approximately 0.5% to 1% per year.1 The contributing factors to clonal progression other than the selective pressure conferred by the driver mutation2 are incompletely understood but proposed mechanisms include inflammation,3 short telomeres leading to chromosomal instability,4 an aging bone marrow microenvironment that favors expansion of clonal stem cells5 and germline genetic predisposition.6 Unexpectedly, CH was also found to increase the risk of diverse forms of cardiovascular disease.7 Accumulating evidence supports a mechanism of accelerated atherogenesis because of vascular inflammation driven by clonally derived monocytes and macrophages.

Figure 1. Clonal hematopoiesis is a risk factor for multiple hematologic malignancies and other diseases of aging. Serial plasma samples are often available from large epidemiological cohorts; however, blood is frequently collected at only a single timepoint. Gutierrez-Rodriquez et al. demonstrated that it is possible to detect clonal hematopoiesis from plasma cell-free DNA, but the clonal fraction from cell-free DNA is not always consistent with measurements of clonal hematopoiesis from blood-derived DNA.

Many ongoing studies seek to profile the diverse disease consequences of CH and how CH evolution is associated with both malignant and non-malignant disease consequences. Existing collections of samples frequently comprise serial plasma samples to measure changes in biomarkers over time. However, blood for genotyping is usually only collected at a single time point (Figure 1). Prior studies have observed that plasma cell-free DNA (cfDNA) may contain CH mutations. For example, CH is a frequently observed “contaminant” in the context of liquid biopsies when using peripheral blood to detect solid tumor mutations.8 In this issue of Haematologica, Gutierrez-Rodriquez et al.9 highlight both the potential and shortcomings of utilizing cfDNA to identify individuals with CH. In their study, cfDNA samples from patients with aplastic anemia, myelodysplastic syndrome, and healthy individuals from the Baltimore Longitudinal Study of Aging were screened for somatic variants in CH driver genes using a chemiluminescent immunoassay commercial targeted sequencing panel. Results were compared to those of DNA derived from paired samples of blood cells from the same individuals. The authors found excellent concordance between cfDNA and paired blood samples among individuals with a hematologic malignancy, particularly with large CH clones. However, they observed poor concordance between cfDNA and paired blood samples among the general healthy aging cohort. It was found that 85% of healthy subjects, 36% of patients with aplastic anemia and 74% of those with myelodysplastic syndrome had somatic cfDNA variants, most frequently in DNMT3A, TET2, ASXL1 and SF3B1. Importantly, there was relatively low agreement between clonal fraction of blood as measured by cfDNA and matched blood cells. The cfDNA CH assay used in this study can be applied to detect CH from historical collections of serial plasma samples. In doing so, this study paves the way for future expanded studies of clonal evolution. However, this study leaves several questions unresolved. In particular, the methodological details of cfDNA isolation and targeted sequencing assays vary considerably from laboratory to laboratory, so it is unclear whether the observations are generally true of other CLIA cfDNA assays. Second, it is unclear whether the clonal fraction estimates from blood and cfDNA are a more reliable estimate of the underlying hematopoietic stem cell clonality that exists in the bone mar-

Haematologica | 107 August 2022

1739

EDITORIAL

M.A.W. Khan and A.G. Bick

row or whether these two estimates have different prognostic relevance. Taken together, the results from this study show both the promise and the limitations of using cfDNA for studying CH in research and clinical settings. From a research perspective, the data presented here suggest that it is possible to detect CH in cfDNA, but rigorous assay validation is required. From a clinical perspective, a physician making a diagnosis of CH based on a cfDNA test would be wise to

perform confirmatory testing in peripheral blood to quantify the clonal fraction accurately, given the discordance between the clonal fraction in cfDNA and blood. Disclosures AGB is a consultant to Foresite Labs. Contributions MAWK and AGB co-wrote the manuscript.

References 1. Hammond D, Loghavi S. Clonal haematopoiesis of emerging significance. Pathology. 2021;53(3):300-311. 2. Watson CJ, Papula AL, Poon GY, et al. The evolutionary dynamics and fitness landscape of clonal hematopoiesis. Science. 2020;367(6485):1449-1454. 3. Marnell CS, Bick A, Natarajan P. Clonal hematopoiesis of indeterminate potential (CHIP): linking somatic mutations, hematopoiesis, chronic inflammation and cardiovascular disease. J Mol Cell Cardiol. 2021;161:98-105. 4. Nakao T, Bick AG, Taub MA, et al. Bidirectional Mendelian randomization supports bidirectional causality between telomere length and clonal hematopoiesis of intermediate potential. medRxiv. 2021. https://www.medrxiv.org/content/10.1101/2021.02.26.21252199 [preprint, not peer-reviewed].

5. Nachun D, Lu AT, Bick AG, et al. Clonal hematopoiesis associated with epigenetic aging and clinical outcomes. Aging Cell. 2021;20(6):e13366. 6. Silver AJ, Bick AG, Savona MR. Germline risk of clonal haematopoiesis. Nat Rev Genet. 2021;22(9):603-617. 7. Bhattacharya R, Bick AG. Clonal hematopoiesis of indeterminate potential: an expanding genetic cause of cardiovascular disease. Curr Atheroscler Rep. 2021;23(11):66. 8. Razavi P, Li BT, Brown DN, et al. High-intensity sequencing reveals the sources of plasma circulating cell-free DNA variants. Nat Med. 2019;25(12):1928-1937. 9. Gutierrez-Rodrigues F, Beerman I, Groarke EM, et al. Utility of plasma cell-free DNA for de novo detection and quantification of clonal hematopoiesis. Haematologica. 2022;107(8):1815-1826.

Haematologica | 107 August 2022

1740

EDITORIAL

F. Perner

Kinetics matter: prognostic implications of early bone marrow assessment in acute myeloid leukemia Florian Perner1,2 1 Department of Pediatric Oncology, Dana Farber Cancer Institute, Harvard Medical School, Boston, MA, USA and 2Innere Medizin C, Universitätsmedizin Greifswald, Greifswald, Germany E-mail: Florian_Perner@DFCI.harvard.edu https://doi.org/10.3324/haematol.2021.280133 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

Despite recent advances in molecular stratification and targeted therapies, including the availability of BCL2-, FLT3- and IDH1/2-inhibitors, curative treatment of patients with acute myeloid leukemia (AML) remains dependent on intensive treatment regimens based on high-dose chemotherapy and consolidation with allogenic hematopoietic stem cell transplantation (HSCT). In order to ensure optimal outcomes for patients it is essential to balance the intensity of the therapeutic interventions and properly target the malignant cells but limit toxicity if possible. A well-defined risk stratification system by the European LeukemiaNet (ELN) classifies patients into

“favorable”, “intermediate” or “adverse” risk groups based on predictive genetic determinants and provides riskadjusted therapeutic recommendations.1 This classification system has proven to be effective in clinical routine and is internationally accepted. Nevertheless, the currently established therapeutic decision trees are fairly static, incorporating mainly genetic factors and the evaluation of response at the end of each treatment cycle. Several studies have proposed, that the kinetics of remission induction may be an independent prognostic factor to guide the decision of whether treatment should be intensified to improve long-term outcomes.2-4 Due to

Figure 1. Prognostic relevance of early bone marrow assessment in acute myeloid leukemia. Schematic depicting the key findings of Ihlow et al. regarding the prognostic relevance of early bone marrow assessment within the population of patients with acute myeloid leukemia, who reach complete remission after intensive induction chemotherapy. A complete clearance of blasts between day 14 and 21 is a favorable prognostic factor, while early resistant disease is an adverse factor even when a complete remission is reached before consolidation. Importantly, the negative prognostic impact of a partial response at early bone marrow assessment may be overcome if patients are consolidated with allogenic stem-cell transplantation but not conventional chemotherapy. allo-HSCT: allogeneic hematopoietic stem cell transplantation. Haematologica | 107 August 2022

1741

EDITORIAL

F. Perner

the relatively small numbers of patients examined in most of these studies and/or the lack of long-term follow-up data no clear consensus has yet been established in the field on whether the kinetics of blast clearance should be taken into consideration for therapeutic decision-making. In this issue of Haematologica, Ihlow et al. report on their findings from a retrospective study in 1,008 patients undergoing intensive treatment for AML.5 The authors aimed to clarify the question whether early bone marrow assessment (day 14-21) during the first cycle of induction chemotherapy has a predictive value for long-term clinical outcomes, specifically within the majority of patients who reach a complete remission before consolidation (Figure 1). Due to high patient numbers and long-term follow-up, Ihlow and co-authors were able to perform sophisticated analyses of various subgroups of patients and add significant granularity to the concept of early response assessment in AML. Not surprisingly, early blast clearance was a favorable prognostic factor both in patients who underwent consolidation by chemotherapy and those who underwent allogenic HSCT. Similarly expected, patients who failed to clear leukemia cells during induction therapy had inferior outcomes compared to patients who reached complete remission across all subgroups. Interestingly, Ihlow et al. discovered relevant prognostic ramifications within the subgroup of patients who showed blast persistence at the point of early bone marrow assessment (day 14-21) but eventually reached complete remission before consolidation. The authors were able to demonstrate, that patients with “early resistant” disease had significantly worse outcomes after both allogeneic HSCT and chemotherapy consolidation, even when reaching complete remission after the last cycle of induction therapy. Of note, patients with blast persistence reaching the status of a partial remission during early bone marrow assessment

followed by a complete remission after induction did not have a significant survival disadvantage over patients with early blast clearance when consolidated with allogeneic HSCT. In contrast, an early partial remission followed by complete remission after induction was associated with a significantly decreased progressionfree survival when patients were consolidated with chemotherapy (Figure 1). Implications of the work presented by Ihlow et al. will be particularly relevant as instrument for improved decision making in the subset of patients diagnosed with ELN “intermediate”-risk AML, for whom the recommendations for allogeneic HSCT are not strictly defined. According to the authors, the application of a decision-making concept incorporating the findings from their current study would have changed therapeutic decisions in about a third of patients with “intermediate”-risk AML. Furthermore, one could speculate that even patients with “favorable”-risk AML may benefit from allogeneic HSCT in case of early blast persistence. Going forward, the utility of these exciting concepts must be confirmed in prospective clinical trials to allow a precise assessment of which groups of patients will benefit from allogeneic HSCT based on early blast persistence. The fact that allogeneic HSCT but not consolidation with conventional chemotherapy was able to overcome the negative prognostic impact of an early partial remission implies that early response assessment primarily reflects cell-intrinsic differences in molecular networks affecting sensitivity to cytotoxic agents. It is tempting to speculate which genetic and/or epigenetic factors may be drivers of this increased resistance to chemotherapy and how far molecular profiling may be useful in the future to further stratify these patients and objectify therapeutic decisions based on genetic markers. Disclosures No conflicts of interest to disclose.

References 1. Dohner H, Estey E, Grimwade D, et al. Diagnosis and management of AML in adults: 2017 ELN recommendations from an international expert panel. Blood. 2017;129(4):424-447. 2. Kern W, Haferlach T, Schoch C, et al. Early blast clearance by remission induction therapy is a major independent prognostic factor for both achievement of complete remission and longterm outcome in acute myeloid leukemia: data from the German AML Cooperative Group (AMLCG) 1992 Trial. Blood. 2003;101(1):64-70. 3. Hussein K, Jahagirdar B, Gupta P, Burns L, Larsen K, Weisdorf D.

Day 14 bone marrow biopsy in predicting complete remission and survival in acute myeloid leukemia. Am J Hematol. 2008;83(6):446-450. 4. Estey EH, Shen Y, Thall PF. Effect of time to complete remission on subsequent survival and disease-free survival time in AML, RAEB-t, and RAEB. Blood. 2000;95(1):72-77. 5. Ihlow J, Gross S, Busack L, et al. Acute myeloid leukemia: negative prognostic impact of early blast persistence can be in part overcome by a later remission prior to post-induction therapy. Haematologica. 2022;107(8):1773-1785.

Haematologica | 107 August 2022

1742

EDITORIAL

H. Luo and M.G. Kharas

Decoding m6A, one reader at a time Hanzhi Luo and Michael G. Kharas Molecular Pharmacology Program, Center for Cell Engineering, Center for Stem Cell Biology, Center for Experimental Therapeutics, Center for Hematologic Malignancies, Memorial Sloan Kettering Cancer Center, New York, NY, USA. E-mail: kharasm@mskcc.org https://doi.org/10.3324/haematol.2021.280166 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

In this issue of Haematologica, Dang et al. demonstrated that the N6-methyladenosine (m6A) RNA methylation reader protein YTHDF3 protects hematopoietic stem cell (HSC) integrity under stress conditions.1 The mRNA methylation pathway has emerged as one of the most important co-transcriptional regulatory processes that controls both normal and malignant hematopoiesis. A full accounting of the different m6A mediators and their distinct roles remains to be established. The study by Dang et al. identifies a new reader that maintains HSC self-renewal under stress. The m6A mRNA modification is the most abundant co-transcriptional mark identified in eukaryotes.2 A set of m6A regulators, including “writers”, “erasers” and “readers”, have been identified to dynamically modify target mRNA and regulate their fate. The m6A writer complex is composed of the core methyltransferase METTL3 which forms a heterodimer with METTL14 and adaptor proteins, including WTAP, RBM15 and ZC3H13. The erasers FTO and ALKBH5 are m6A demethylases that remove the m6A mark, while the readers bind to m6A-marked transcripts. The readers include forms of YTHDF (YTHDF1,2,3), YTHDC (YTHDC1,2), and IGF2BP (IGF2BP1,2,3).2 These m6A modulators have been implicated in mRNA stability, splicing, nuclear export, and translational efficiency. Although RNA methylation has been implicated in a variety of tissues and diseases, there has been significant attention to both normal and malignant hematopoiesis. It was found that m6A writers METTL3 and METTL14 control myeloid differentiation and are essential for development of acute myeloid leukemia (AML).2-4 Suggesting that the balance of m6A methylation is important, the erasers FTO and ALKBH5 are both upregulated in several subtypes of AML and maintain leukemia stem cell activity.2,5,6 Additionally, the reader YTHDF2 contributes to leukemia stem cell self-renewal and maintenance while the nuclear reader YTHDC1 forms condensates with m6A marked transcripts and promotes AML cell survival.7,8 Therefore, targeting the m6A RNA methylation program has been proposed as a new therapeutic strategy against AML. Most excitingly, small molecules targeting m6A regulators, including METTL3 and FTO, have now demonstrated therapeutic efficacy in AML models.9,10 While m6A regulators seem to converge on an oncogenic role in AML, their roles in normal hematopoiesis are more complex. Loss of the m6A writer METTL3 results in a symmetric commitment defect in HSC and a failure to differentiate, while deletion of the eraser ALKBH5 has minimal

effects on normal hematopoiesis.5,6,11 Suppression of the reader YTHDF2 promotes HSC self-renewal and results in HSC expansion at a young age but is detrimental to HSC function at an old age or under hematopoietic stress.12 Inhibiting another reader, YTHDC1, results in defects in murine HSC self-renewal, but has minimal impact on CD34+ human hematopoietic stem and progenitor cell (HSPC) survival.8,13 Together, these studies imply that the m6A program in normal blood cells and AML cells is cell-context dependent and requires further delineation. In this issue of Haematologica, Dang et al. describe their investigation of role of the m6A reader YTHDF3 in normal hematopoiesis.1 YTHDF3 is the third member of the YTHDF family, sharing 65% protein sequence identity with both YTHDF1 and YTHDF2. These YTHDF paralogs share 90% homology in the YTH domain, which recognizes and binds to m6A sites. Using Ythdf3 knockout mice, Dang et al. first showed that YTHDF3 is dispensable for steady-state hematopoiesis. Germline Ythdf3 knockout (KO) mice displayed comparable hematopoietic lineage compositions in the peripheral blood, bone marrow and spleen. Loss of YTHDF3 resulted in equivalent HSPC populations, and normal cell cycle and death percentages as compared to those in littermate controls. To determine the role of YTHDF3 under stress conditions, the authors performed bone marrow transplantation assays. At transplantation, Ythdf3 KO bone marrow cells engraft comparably to control cells. However, a significant reduction in engraftment in the HSPC compartment was observed. These data prompted the authors to further examine the impact of hematopoietic stress on Ythdf3 KO bone marrow cells. Under a 5-fluorouracil challenge, which ablates mature myeloid cells and pushes dormant HSC into the cycle, mice that received Ythdf3 KO bone marrow were more susceptible to death compared to the recipients of control bone marrow. Additionally, Ythdf3 KO bone marrow cells exhibited a two-fold reduction in repopulating capacity in secondary transplantation experiments. This defect was also observed in HSC and without a lineage output bias in the Ythdf3 KO cells. Together, these data suggest that YTHDF3 is required for HSC self-renewal and maintenance during stress hematopoiesis. To address the molecular mechanism underlying the role of YTHDF3 in HSPC, the authors analyzed the publicly available YTHDF3 cross-linking immunoprecipitation (CLIP)-sequencing data in Hela cells. Given the established roles of Foxm1

Haematologica | 107 August 2022

1743

EDITORIAL

H. Luo and M.G. Kharas

Figure 1. YTHDF3 is required in hematopoietic stem cell maintenance under stress conditions. KO: knockout; HSC: hematopoietic stem cells, HSPC: hematopoietic stem and progenitor stem cells., 5FU: 5-fluorouracil.

and Asxl1 in HSC maintenance the authors focused on these transcripts. Both Foxm1 and Asxl1 transcripts were m6A-modified and were direct binding targets of YTHDF3 in HPC-7 cells. Using either siRNA-mediated knockdown in HPC-7 cells or Ythdf3 KO primary c-Kit+ bone marrow progenitor cells, the authors found that YTHDF3 loss did not affect the mRNA levels but significantly reduced the protein abundance of FOXM1 and ASXL1. The authors then examined whether YTHDF3 regulates Foxm1 and Asxl1 translation and provided two pieces of evidence that support this hypothesis. First, they showed that the binding of the translation initiation factor EIF3A to the Foxm1 and Asxl1 transcripts was reduced upon YTHDF3 knockdown. Second, the half-life of FOXM1 and ASXL1 proteins remained unchanged upon YTHDF3 knockdown, suggesting the defect is during protein synthesis. Interestingly, this effect was not associated with YTHDF2 since YTHDF2 depletion had no effect on either RNA levels or protein abundance of these targets. These results suggest that YTHDF3, but not YTHDF2, promotes the translation of Foxm1 and Asxl1 in hematopoietic cells. Both specific and shared functions for each YTHDF paralog have been proposed. Some studies suggested that YTHDF1 and YTHDF3 promote translation by recruiting translation initiation factors; while YTHDF2 enhances mRNA degradation, by recruiting the CCR4–NOT deadenylase complex.2 Recently, a new model has been proposed, in which the YTHDF paralogs act redundantly on the same set of mRNA targets to regulate RNA stability.14 This model is supported by YTHDF single knockouts and triple knockout studies in

human Hela and mouse embryonic stem cells in which the YTHDF paralogs are all similarly expressed. However, whether the YTHDF paralogs also act redundantly in hematopoiesis remains to be tested. Thus, the paper by Dang et al. provides new insights into the role of another YTHDF paralog YTHDF3 in hematopoiesis. In contrast to YTHDF2, which can expand HSC numbers, YTHDF3 is only required during stress hematopoiesis. Moreover, the authors found that YTHDF3, but not YTHDF2, translationally regulates Foxm1 and Asxl1 transcripts. These results suggest that YTHDF3 and YTHDF2 may act on different sets of mRNA targets in HPC-7 cells and indicate that the YTHDF paralogs may act non-redundantly in hematopoiesis. Future studies are needed to globally compare the direct binding targets of the YTHDF paralogs in HSPC and their redundant and non-redundant functions. Overall, the study by Dang et al. demonstrates a role for the m6A reader YTHDF3 in HSC self-renewal and maintenance under stress conditions (Figure 1). It also implies cell-context dependency for individual YTHDF paralogs. Further studies on additional m6A regulators and their downstream targets will uncover the complex m6A network in normal and malignant hematopoiesis. Disclosures No conflicts of interest to disclose. Contributions HL and MK both contributed to this editorial.

Haematologica | 107 August 2022

1744

EDITORIAL

H. Luo and M.G. Kharas

References 1. Dang Q, Wu Q, Sheng Y, et al. m6A reader Ythdf3 protects hematopoietic stem cell integrity under stress by promoting the translation of Foxm1 and Asxl1 transcripts. Haematologica. 2022;107(8)1922-1927. 2. Vu LP, Cheng Y, Kharas MG. The biology of m(6)A RNA methylation in normal and malignant hematopoiesis. Cancer Discov. 2019;9(1):25-33. 3. Barbieri I, Tzelepis K, Pandolfini L, et al. Promoter-bound METTL3 maintains myeloid leukaemia by m(6)A-dependent translation control. Nature. 2017;552(7683):126-131. 4. Vu LP, Pickering BF, Cheng Y, et al. The N(6)-methyladenosine (m(6)A)-forming enzyme METTL3 controls myeloid differentiation of normal hematopoietic and leukemia cells. Nat Med. 2017;23(11):1369-1376. 5. Wang J, Li Y, Wang P, et al. Leukemogenic chromatin alterations promote AML leukemia stem cells via a KDM4CALKBH5-AXL signaling axis. Cell Stem Cell. 2020;27(1):81-97.e8. 6. Shen C, Sheng Y, Zhu AC, et al. RNA demethylase ALKBH5 selectively promotes tumorigenesis and cancer stem cell self-renewal in acute myeloid leukemia. Cell Stem Cell. 2020;27(1):64-80.e9. 7. Paris J, Morgan M, Campos J, et al. Targeting the RNA m(6)A reader YTHDF2 selectively compromises cancer stem cells in

acute myeloid leukemia. Cell Stem Cell. 2019;25(1):137-148.e6. 8. Cheng Y, Xie W, Pickering BF, et al. N(6)-methyladenosine on mRNA facilitates a phase-separated nuclear body that suppresses myeloid leukemic differentiation. Cancer Cell. 2021;39(7):958-972.e8. 9. Yankova E, Blackaby W, Albertella M, et al. Small-molecule inhibition of METTL3 as a strategy against myeloid leukaemia. Nature. 2021;593(7860):597-601. 10. Su R, Dong L, Li Y, et al. Targeting FTO suppresses cancer stem cell maintenance and immune evasion. Cancer Cell. 2020;38(1):79-96.e11. 11. Cheng Y, Luo H, Izzo F, et al. m(6)A RNA methylation maintains hematopoietic stem cell identity and symmetric commitment. Cell Rep. 2019;28(7):1703-1716.e6. 12. Mapperley C, van de Lagemaat LN, Lawson H, et al. The mRNA m6A reader YTHDF2 suppresses proinflammatory pathways and sustains hematopoietic stem cell function. J Exp Med. 2021;218(3):e20200829. 13. Sheng Y, Wei J, Yu F, et al. A critical role of nuclear m6A reader YTHDC1 in leukemogenesis by regulating MCM complexmediated DNA replication. Blood. 2021;138(26):2838-2852. 14. Zaccara S, Jaffrey SR. A unified model for the function of YTHDF proteins in regulating m(6)A-modified mRNA. Cell, 2020;181(7):1582-1595.e18.

Haematologica | 107 August 2022

1745

ARTICLE - Acute Lymphoblastic Leukemia

Rational drug combinations with CDK4/6 inhibitors in acute lymphoblastic leukemia Karen L. Bride,1* Hai Hu,2* Anastasia Tikhonova,3* Tori J. Fuller,4 Tiffaney L. Vincent,4 Rawan Shraim,4 Marilyn M. Li,4 William L. Carroll,2 Elizabeth A. Raetz,2 Iannis Aifantis2# and David T. Teachey4# Department of Pediatrics, Division of Hematology/Oncology and Cellular Therapy, Cohen Children's Medical Center, New Hyde Park, New York, NY, USA; 2Perlmutter Cancer Center and Department of Pediatrics, NYU Langone Health, New York, NY, USA; 3Princess Margaret Cancer Center, University of Toronto, Toronto, Ontario, Canada and 4Department of Pediatrics, Division of Oncology, Children's Hospital of Philadelphia and University of Pennsylvania Perelman School of Medicine, Philadelphia, PA, USA 1

*KLB and AT contributed equally as co-first authors.

Correspondence: David T. Teachey teacheyd@chop.edu Received: June 15, 2021. Accepted: December 16, 2021. Prepublished: December 23, 2021. https://doi.org/10.3324 haematol.2021.279410 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

IA and DTT contributed equally as co-senior authors.

Abstract Despite improvements in outcomes for children with B- and T-cell acute lymphoblastic leukemia (B-ALL and T-ALL), patients with resistant or relapsed disease fare poorly. Previous studies have demonstrated the essential role of cyclin D3 in T-ALL disease initiation and progression and that targeting of the CDK4/6-cyclin D complex can suppress T-ALL proliferation, leading to efficient cell death in animal models. Studies in leukemia and other malignancies, suggest that schedule is important when combining CDK4/6 inhibitors (CDKi) with cytotoxic agents. Based on these observations, we broadened evaluation of two CDKi, palbociclib (PD-0332991, Pfizer) and ribociclib (LEE011, Novartis) in B- and T-ALL as single agent and in combination with conventional cytotoxic chemotherapy, using different schedules in preclinical models. As monotherapy, CDKi caused cell cycle arrest with a significant decrease in S phase entry and were active in vivo across a broad number of patient-derived xenograft samples. Prolonged monotherapy induces resistance, for which we identified a potential novel mechanism using transcriptome profiling. Importantly, simultaneous but not sequential treatment of CDKi with conventional chemotherapy (dexamethasone, L-asparaginase and vincristine) led to improved efficacy compared to monotherapy in vivo. We provide novel evidence that combining CDKi and conventional chemotherapy can be safe and effective. These results led to the rational design of a clinical trial.

Introduction Acute lymphoblastic leukemia (ALL), the most common pediatric malignancy, is a biologically heterogeneous disease with multiple different subtypes characterized by genetic alterations that result in deregulation of hematopoietic transcription factors, epigenetic modifiers, cytokine receptors and tyrosine kinases.1 While the overall prognosis for children and young adults with ALL has improved, outcomes for patients with some ALL subtypes remain poor despite intensification of cytotoxic chemotherapy. Deregulated cell cycle progression is a hallmark feature of cancer cells. Cyclin-dependent kinases, CDK4 and CDK6 (CDK4/6) compose the core machinery governing the progression through the early G1 phase of the cell cycle. In complex with cyclin D, CDK4/6 promote cell cycle progression through at least two functions: hyper-phosphorylation of the retinoblastoma protein (RB) thereby uncoupling the CDK4/6-cyclin D complex from the E2F transcription fac-

tors; and interaction with cell cycle inhibitors, p21cip1 and p27kip, which promotes activation of the CDK2/cyclin E complex.2 The catalytic activity of CDK4 or CDK6 regulates a critical checkpoint for the G1-S transition and the commitment to cell division. As a consequence, gain of function of the G1-S checkpoint’s cyclin-CDK complexes is a major driver in a large number of human cancers. CDK4/6 inhibition has been explored as a means to bypass the deregulated cell cycle progression common in cancer pathogenesis. While CDK4/6 is rarely mutated in human cancers, aberrations that lead to the loss of proliferative controls include overexpression of CDK4 itself, amplification of the D-type cyclins, downregulation of p16INK4A, mutations in CDK4 that prevent p16INK4A binding to the enzyme, and deletion or mutation of RB1 itself.3-6 The frequency of these alterations implies that abrogation of the G1 checkpoint or progression through the CDK4/cyclin D pathway provides a distinct advantage to cancer cell proliferation and survival. Deletions of CDKN2A/B and RB1

Haematologica | 107 August 2022

1746

ARTICLE - Chemotherapy and CDKi in ALL

K.L. Bride et al.

are common in ALL, with CDKN2A/B deletions occurring in 70% of T-ALL and 36% of B-ALL. Moreover, RB1 suppressor gene deletions occur in 12% in T-ALL and 8% in B-ALL.7-11 Deletions involving CDKN2A/B are also common in BCR-ABL1+ ALL, and genetic alterations involving the RB1 gene are hallmarks of low hypodiploid ALL.12,13 Based on these observations, a number of CDK4/6 selective inhibitors have been developed, including ribociclib (LEE011) and palbociclib (PD0332991). Both are more selective for the ATP-binding pocket of the kinase domain compared to previous non-selective cell cycle inhibitors, with significantly less off-target effects and fewer doselimiting toxicities.14,15 Palbociclib has progressed furthest in the clinic, having received Food and Drug Administration approval for estrogen receptor-positive (ER-positive), HER2-negative post-menopausal breast cancer patients in combination with endocrine therapy.16 Previous studies suggest induction of early G1 arrest by CDK inhibitors (CDKi) requires retinoblastoma (Rb), and this arrest is reversible in vitro and in vivo.4,17 Therefore, amplification of CDK6 and cyclin E1 are able to bypass dependency on cyclin D1/CDK4 signaling, while RB1 loss provides intrinsic resistance to CDKi in preclinical studies. Overall, the selectivity and reversibility of cell-cycle inhibition by CDKi and their clinical efficacy, provide a safe and effective approach to target specific phases of the cell cycle in cancer. Since the majority of conventional chemotherapeutics target actively cycling cells, some studies suggest combinations of CDKi with cytotoxic chemotherapy could be antagonistic.14,18-20 Thus, it is crucial to determine efficacy of CDKi, and understand the optimal schedule in order to combine them safely and effectively with chemotherapy. We investigated the activity of CDKi in ALL preclinical models. We found that B-ALL and T-ALL blasts were sensitive to these inhibitors via concentration-dependent cell cycle arrest in vitro. We then tested various dosing schedules of CDKi with conventional cytotoxic chemotherapy agents typically used in upfront therapy for B-ALL and TALL, including glucocorticoids, vincristine and L-asparaginase in vivo using patient-derived xenograft (PDX) models. Importantly, we provide divergent evidence from prior in vitro studies evaluating timing of CDKi and conventional chemotherapy and find that CDKi enhance combination therapy in ALL when administered concurrently rather than in sequence. These data prompted the phase I clinical trial (clinicaltrials gov. Identifier: NCT03792256) evaluating palbociclib in combination with chemotherapy for children with relapsed ALL or lymphoblastic lymphoma.

von Mikroorganismen und Zellkulturen (DSMZ, Braunschweig, Germany) and maintained in RPMI 1640 supplemented with 10-20% heat-inactivated fetal bovine serum (Gibco, Thermo-Scientific). All cell lines were frozen and low passage cells were used for all experiments. Western blotting Cells were lysed and 10-15 mg were prepared,22 resolved by electrophoresis on NuPAGE 4-12% SDS-PAGE gradient gels (Invitrogen) and transferred to polyvinylidene fluoride membranes (Invitrogen). Immunoblotting was performed with primary antibodies: RB, phosphorylated RB (pRB)S807/801; pRBS807/810; cyclin D3, p21 (Cell Signaling Technologies). Bound antibody was detected using Western Chemiluminescence Reagent (PerkinElmer). Compounds Ribociclib (LEE011) was generously provided by Novartis Inc. (Basel, Switzerland). Palbociclib was provided by Pfizer Inc. (New York, NY). Clinical grade chemotherapy agents were purchased from commercial vendors: vincristine, Teva (Petah Tikva, Israel); L-asparaginase, Jazz pharmaceuticals (Dublin, Ireland); dexamethasone, Fresenius Kabi (Bad Homburg vor der Hohe, Germany); and everolimus, Novartis. Vincristine, L-asparaginase and dexamethasone dosing per previous studies.21,22 Everolimus was dosed at 2.5 mg/kg and 5 mg/kg by oral gavage based on prior reports in preclinical models of ALL.23 Flow cytometry Cells were treated with vehicle versus palbociclib or ribociclib for 48-72 hours, labeled with BrdU for 1 hour and processed for flow cytometry, on an Accuri CSampler flow cytometer (BD Accuri Cytometers), and analysis was performed with FlowJo software (Tree Star, Ashland, OR). Engraftment and serial disease evaluation were determined by flow cytometric analysis of blood using antibodies against human CD19 and CD45 for B-ALL and CD45 for T-ALL.24 Xenograft therapeutic trials Patient samples

Banked diagnostic specimens (lymphoblasts from peripheral blood or bone marrow) were obtained from patients enrolled on the Children’s Oncology Group (COG) study P9906 or from the Stem Cell and Leukemia Core at the Children’s Hospital of Philadelphia or Columbia Presbyterian Hospital. All samples were obtained after informed consent was given in accordance with the Declaration of Helsinki.24,25 Patient-derived xenografts

Methods Reagents and cell lines Human cell lines were obtained from Deutsche Sammlung

PDX models using non-obese diabetic/severe combined immunodeficiency (NOD/SCID/Il2rgtm1wjl/SzJ) (NSG) mice were developed as previously described.24-26 All experiments were conducted in accordance to protocols ap-

Haematologica | 107 August 2022

1747

ARTICLE - Chemotherapy and CDKi in ALL

K.L. Bride et al.

Figure 1. CDK inhibitors induce G1 arrest in B-cell acute lymphoblastic leukemia cell lines in vitro. (A) We performed western blot analysis of a panel of B-cell acute lymphoblastic leukemia (B-ALL) cell lines (labeled in blue) and primary patient samples (labeled in green), which demonstrate high protein expression of CDK4, CDK6 and CCND3. Inhibitor p21 is also highly expressed. A T-ALL cell line, Jurkat (labeled in black), was used as a positive control, given prior published data, however, it is known to be negative for p21.40 (B) Samples were treated for a minimum of 48 hours with either palbociclib (B) or ribociclib (C) prior to harvesting for flow cytometry. Percentage of cells arrested in G1 with varying doses of CDK inhibitors (CDKi) in 3 B-ALL cell lines: REH, MHH CALL4, and KOPN8. (D) Corresponding dose-dependent inhibition by CDKi reflects loss of phosphorylated retinoblastoma (p-RB), with no effect on total retinoblastoma (Rb). Phosphorylated S6, a downstream signaling marker was differentially affected with CDK inhibition in REH and KOPN8, but not MHH CALL4. Haematologica | 107 August 2022

1748

ARTICLE - Chemotherapy and CDKi in ALL

K.L. Bride et al.

proved by the Institutional Animal Care and Use Committee and Institutional Review Board of the Children’s Hospital of Philadelphia (CHOP) and New York University School of Medicine (NYU). Samples were derived from patients with de novo ALL. RNA sequencing and data analysis RNA sequencing (RNA-seq) of native Jurkat and palbociclib-resistant Jurkat cells were performed at the NYU School of Medicine Genome Technology Core. STAR 2.4.2a was applied to align the RNA-seq samples to the reference mouse genome (mm9) and count the number of reads that map to each gene in the ensemble GRCm38.80 gene model. R (v.3.5.1) (http://www.R-project.org/) and DESeq2 package (v1.10.0) were used to perform differential gene expression analysis among different sample groups.

and KOPN8) (Figure 1B and C; Online Supplementary Figure S1). Jurkat was used as a positive control. In spite of robust phosphorylated RB expression, the sensitivity to CDK4/6 inhibition was variable in cell lines. In order to confirm that the growth inhibition observed in sensitive cell lines reflected targeted impairment of CDK4/6 signaling, we analyzed the levels of phosphorylated RB following treatment with either palbociclib or ribociclib (Figure 1D). Correspondingly, the amount of phosphorylated RB decreased with increasing concentrations of palbociclib or ribociclib. While both CDKi induced cell cycle arrest via inhibition of phosporylated RB, palbociclib is more potent at lower concentrations compared to ribociclib in vitro. However, based on expected comparable PK/PD in humans, Table 1. Panel of B-cell and T-cell acute lymphoblastic leukemia cell lines and patient samples included in this study.

Statistical analysis Statistical analyses were performed using GraphPad Prism version 7. Results were analyzed for statistical significance using repeated measures 2-way analysis of variance (ANOVA) of peripheral blast counts (P-values represent the interaction between time and treatment). Spleen blast counts were analyzed by two-sided t-test.

Xeno ID (USI)

Results CDK4/6 signaling is hyperactive in B-cell acute lymphoblastic leukemia In order to confirm that genomic aberrations translate to aberrant CDK4/6 signaling within the cyclin D/CDK4/CDK6/RB pathway, we examined the activation status of RB in a comprehensive panel of well-characterized human B-ALL cell lines and patient samples. As shown in Figure 1A, robust phosphorylation of RB at serines 780 and 807/810 – residues directly targeted by CDK4 and CDK6 – was observed in all tested cell lines, and protein-level expression of CDK4, CDK6 and cyclin D3 (CCND3) occurred in the majority of cell lines and patient samples. Jurkat, a T-ALL cell line was used as a positive control given previously published data demonstrating TALL sensitivity to CDKi.20 B-cell acute lymphoblastic leukemia is sensitive to CDK4/6 inhibition in vitro Given the observation that CDK4/6 signaling is highly active in B-ALL, thus maintaining hyperphosphorylated RB and supporting cell cycle progression through the G1-S checkpoint, we evaluated the effect of selective CDK4/6 inhibition on B-ALL cell lines in vitro (Table 1). Palbociclib (Figure 1B) and ribociclib (Figure 1C) induce significant growth arrest in G1 across a broad range of concentrations (0.5 to 4.0 mM) in a panel of cell lines (REH, MHH CALL4,

Biology

697

Cell line

SUP-B15

Cell line

REH

Cell line

MHH-CALL4

Cell line

KOPN8

Cell line

E2A-PBX t(1;19) Ph+ t(9;22) ETV6-RUNX1 t(12;21) Hypodiploid MLL-MLLT1 translocation

CDKN2A/B deletion No (32) Yes (33) No (32) No (34) Yes (35)

JH652 (PALLSD)

Ph-like: Jak mut Primary B-ALL (R683G), high CRLF2, IKZF1

Yes (36)

JH561 (PAKRSL)

Ph-like: Jak mut Primary B-ALL (R683G), high CRLF2, IKZF1

Yes (36)

POG200521761 NH011 (PAKVKK)

Primary B-ALL Primary B-ALL

NL482A (PAKYEP)

Primary B-ALL

240

Primary B-ALL

359

Primary B-ALL

2365

Primary B-ALL

ALL16 (T-ALL1)

Primary T-ALL

Cul76

Primary T-ALL

ETP8 (T-ALL2) Primary T-ALL

MLL Ph-like: NUP214-ABL IKZF1, BCR/JAK2; wild-type JAK with low CRLF2 del20 Adult BCR-ABL (t9;22) Relapsed BCRABL (t9;22) Non-ETP T-ALL Non-ETP T-ALL (STIL-TAL+) ETP T-ALL

No* Yes (37)

Yes (37) Yes* Yes* Yes* Yes (36) Yes (38) No (39)

A variety of acute lymphoblastic leukemia (ALL) cell lines and patient samples were evaluated in these studies. *Indicates samples that underwent somatic genetic profiling as previously described.31 See the Online Supplementary Appendix for details. ETP: early T-cell precursor; Xeno: xenograft.

Haematologica | 107 August 2022

1749

ARTICLE - Chemotherapy and CDKi in ALL

K.L. Bride et al.

these agents are considered equivalent in efficacy at their respective appropriate doses.27 Further, these two drugs are similar in their selectivity of CDK4/6 inhibition. CDK4/6 inhibition causes growth inhibition in vivo We assayed for single agent in vivo efficacy using PDX models of T-ALL and B-ALL. In the two T-ALL samples, CDKi treatment markedly decreased the number of human CD45+ (hCD45+) cells (Figure 2C) with a correlated prolonged overall survival in comparison to vehicle (Figure 2E). Similar to in vitro studies, this treatment regimen appropriately induced profound cell cycle arrest of hCD45+ leukemic cells in vivo (Figure 2D). Of note, the difference in palbociclib dosing was related to increased diarrhea in the B-ALL PDX models that was not seen in the T-ALL PDX. In B-ALL, leukemia progression was significantly inhibited in some of the samples but not all (Figure 3). In six of the seven unique PDX samples tested, tumor growth was significantly impacted in the spleen and/or blood. The seven PDX represented a heterogeneous group of high-risk B-

ALL patient samples (Table 1). All of the tested samples had high levels of Rb therefore did not vary in the activation status of cyclin D, nor had any obvious biomarker to predict efficacy. In both subtypes (B- and T-ALL xenografts), the dosing strategy was well tolerated with no evidence of weight loss of other signs of toxicity. Cyclin D3 overexpression confers palbociclib resistance In order to explore potential de novo mechanisms of resistance to palbociclib treatment, we treated a human TALL cell line with low dose (0.1 mM) palbociclib in vitro, but gradually escalated to a final dose of 2.8 mM, a process that lasted 52 weeks, generating palbociclib resistant T-ALL cells (Figure 4A). Strikingly, resistant cells have higher rates of cycling cells (Figure 4B). This finding suggests that the resistant cells are adapting to treatment by identifying alternative ways to boost cell cycle progression. In agreement with this notion, we were able to demonstrate that resistant cells do not decrease levels of phosphorylated RB in response to palbociclib treatment (Figure 4C). In fact,

Figure 2. CDK inhibitor suppress human primary T-cell acute lymphoblastic leukemia xenograft proliferation in vivo. (A) Schematic representation of CDK inhibitor (CDKi) treatment course with single agent palbociclib at 75 mg/kg gavage vs. vehicle. Dosing determined based on prior studies.28 (B) A representation of flow cytometric analysis of T-cell acute lymphoblastic leukemia xenograft populations (hCD45+) in peripheral blood following 3 weeks of CDKi treatment. (C) Flow cytometry analysis of T-ALL xenograft populations (hCD45+) in peripheral blood over the course of treatment. (D) DAPI staining of 2 different human xenografts (ALL16 and ETP8) following 48-hour treatment with CDKi in vivo. (E) Kaplan-Meier survival graph of the 2 T-ALL xenografts treated with CDKi. Haematologica | 107 August 2022

1750

ARTICLE - Chemotherapy and CDKi in ALL

K.L. Bride et al.

Figure 3. CDK inhibitors suppress human primary B-cell acute lymphoblastic leukemia xenograft proliferation in vivo. (A) Schematic representation of CDK inhibitor (CDKi) administration with single agent palbociclib at 35 mg/kg or ribociclib at 150 mg/kg, gavage vs. vehicle. Ribociclib dosing determined based on prior studies, and palbociclib dosing based on internal communication with Pfizer.28,41,42 (B) Mice (5 mice per treatment group) were randomized to palbociclib (red hashed bars), or to ribociclib (solid red bars) vs. vehicle (blue bars), for 5 consecutive days per week until sacrifice (21-28 days). Peripheral blood blast cell (PBC) counts are shown in the left panel. Splenic blast burden is shown in the right panel, with corresponding P-values. Two patient-derived xenograft (PDX) models (2,365 and 240), demonstrated significant inhibition of disease progression in blood and spleen, two PDX (359, JH561), demonstrated significant inhibition of disease progression in the peripheral blood but not in the spleen; and two PDX, JH562 and POG/MLL demonstrated significant disease inhibition in the spleen but not in the blood; 1 PDX model (NL482A) had no response. hB-ALL: human B-cell acute lymphoblastic leukemia.

we detected significantly elevated basal levels of cyclin D3 protein (CCND3), which suggests a potential mechanism of resistance. In order to further probe for such resistance mechanisms we performed RNA-seq of sensitive and resistant Jurkat cells at steady-state and after palbociclib treatment (Figure 4D). We found sensitive Jurkat cells at baseline (steady-state) had lower levels of cell cycle related genes as compared with resistant cells. Moreover, we found a significant downregulation of cell cycle related genes including cyclins A2, B1, B2 and E2; CDK1, 2 and 4 and E2F transcription factors (E2F1, 2, 7), in sensitive cells, similar to previous work in other malignancies.28 In resistant cells, expression of cell cycle related genes was only modestly decreased when treated with palbociclib. These experiments suggest that palbociclib-resistant cells have adapted to treatment by boosting cell cycle entry and progression through the upregulation of a number of essential regulators, including CCND3, CDK and members of the E2F family.

Combination therapy of palbociclib was enhanced with concurrent administration of standard chemotherapy compared to sequential treatment Given the suggestion that timing may be a critical component of efficacy with combination therapy, we tested pretreatment of palbociclib versus concurrent administration on disease burden in two high risk B-ALL PDX (JH561 and NH011) and one T-ALL xenograft (Cul76). These samples were chosen given their known risk for having inferior prognosis and modest response to conventional cytotoxic chemotherapy. The treatment schema for combinations in B-ALL is shown in Figure 5. In order to closely mimic agents typically used in conventional therapy for B-ALL we tested dexamethasone, vincristine and L-asparaginase. Single agent vincristine (1 mg/kg) or dexamethasone (7.5 mg/kg) were too effective on their own to delineate a differential impact of combination therapy, when they were given at a dose that mirrors the PK/PD of the dose used in children with ALL.

Haematologica | 107 August 2022

1751

ARTICLE - Chemotherapy and CDKi in ALL

K.L. Bride et al.

Figure 4. Cell cycle regulators are upregulated in CDK inhibitor-resistant T-cell acute lymphoblastic leukemia. (A) Schematic representation of the method of generating de novo CDK inhibitor (CDKi)‐resistant cells using palbociclib dose escalation. (B) Cell cycle analysis of 1 mM palbociclib (palbo) treated parental and palbociclib-resistant cells. (C) Western blotting of pRB and Cyclin D3 (CCND3) in parental and palbociclib-resistant cells. (D) Differential expression of cell cycle genes in treatment of parental and CDKi-resistant cells.

Therefore, we performed a titration of single drug in vivo, determining the concentration where single agent decreased disease burden by about 50% (half maximal inhibitory concentration [IC50]) of the maximum dose in the peripheral blood. Average IC50 of three independent determinations were performed (vincrisitine at 0.5, 0.25, 0.1 and 0.05 mg/kg; dexamethasone at 4, 2 and 1 mg/kg; data not shown). IC50 doses for each drug include vincristine at 0.1 mg/kg intraperitoneally (IP) weekly, dexamethasone at 2 mg/kg IP daily and L-asparaginase at 5,000 IU/kg IP weekly. As shown in Figure 6 (and Online Supplementary Figure S2), simultaneous administration of combination therapy (cytotoxic agent + palbociclib) was superior to monotherapy with any of the individual cytotoxics (dexamethasone, vincristine, or L-asparaginase) or palbociclib alone in both B-ALL PDX models. In contrast, sequential treatment with palbociclib or vehicle followed by vincristine or L-asparaginase or vehicle did not exhibit an augmented disease response in combination. However, palbociclib followed by dexametha-

sone did show superiority of the combination over either single agent. In order to verify the benefit of the simultaneous administration of palbociclib plus cytotoxics in T-ALL, we set up four cohorts of human T-ALL xenografted animals that received combination treatment for 4 weeks (Figure 7A). We found that combination therapy of palbociclib concurrently with vincristine led to a significant difference in disease burden (Figure 7B) and survival (Figure 7C) compared to monotherapy. Combination CDK4/6 and mTOR inhibitors is not effective in vivo against B-cell acute lymphoblastic leukemia The PI3K/AKT/mTOR pathway is a key upstream regulator of cyclin D-CDK4/6 activity. Thus, agents that target PI3K/Akt/mTOR signaling including mTOR inhibitors have been studied in a number of malignancies.19 CDK4/6 inhibition has been demonstrated to overcome resistance

Haematologica | 107 August 2022

1752

ARTICLE - Chemotherapy and CDKi in ALL

K.L. Bride et al.

to PI3K inhibition and endocrine therapy, providing a strong rationale for the combination of an mTOR inhibitor with CDKi. Recent in vitro data suggest the combination of everolimus and ribociclib may be more effective than either agent alone in ALL cell lines.20 We tested simultaneous combination therapy with palbociclib or ribociclib with everolimus in four different B-ALL PDX models (JH561, 240, POG/MLL, and NL482B) (Online Supplementary Figure S3). Inhibition of the mTOR pathway did not improve the response to CDK4/6 inhibition in any of the models.

Discussion We provide evidence of single agent efficacy of palbociclib and ribociclib, two selective CDK4/6 inhibitors, in a heterogeneous range of B- and T-ALL samples, in vitro and in vivo. We show that RB phosphorylation via CDK4/6 signaling is prevalent in B-ALL cell lines and primary patient samples, reflected by high expression of CDK4, CDK6 and CCND1, but there are likely other mechanisms of CDK4/6 hyperactivation. We demonstrate growth inhibition reflects G1 arrest and depletion of phosphorylated RB. Similar to other studies, RB expression appeared to be essential for sensitivity. However, each sample had varying levels of sensitivity suggesting there may be additional relevant biomarkers yet to be determined, to predict sensitivity to CDKi in B-ALL. Genomic alterations in the CDK pathway are common in ALL, with deletions in CDKN2A/B occurring with an incidence of 30-50% in childhood and adult ALL.29 The prognostic value has been widely investigated but the results remain unclear. While there is some suggestion that CDKN2A/B deletions have an adverse impact on overall survival (OS), these studies have been unable to determine if CDK alterations are independently prognostic of minimal residual disease (MRD) or other sentinel genetic lesions, such as BCR-ABL, ETV6-RUNX1, in multi-variant analyses. In an attempt to associate CDKN2A/B deletion status to CDKi sensitivity (Table 1; Figure 3B), we did not identify a correlation, however were likely underpowered to make any definitive conclusions. We did identify a novel mechanism of acquired resistance to palbociclib treatment, related to upregulation of CCND3, CDK and members of the E2F family. Future studies will need to investigate the mechanisms of intrinsic resistance to CDKi. We believe large scale genomic testing embedded in future clinical trials will be necessary to confirm and/or identify additional relevant biomarkers that confer sensitivity to these agents, and to determine if CDK pathway lesions are independently prognostic when considering other clinical and biologic variables. Not surprisingly, a number of clinical studies across ma-

Figure 5. Treatment schema of pretreatment versus concurrent regimens in vivo. Treatment schema depicting the timing of each drug (vincristine, dexamethasone or L-asparaginase and/or CDK inhibitor [CDKi]) vs. vehicle in the pretreatment and concurrent schema. Veh: vehicle; chemo: chemotherapy; hB-ALL: human B-cell acute lymphoblastic leukemia; palbo: palbociclib.

lignancies suggest single agent CDKi fail to provide durable responses, indicating combination therapies are needed. Efficacy was more pronounced in T-ALL as compared to B-ALL in our preclinical models, suggesting CDKi may have single agent activity in T-ALL, but should only be tested in combination with other drugs in B-ALL in clinical trials. In fact, we noted higher baseline expression of CCND3 in B-ALL, which may partly explain the increased sensitivity of T-ALL blasts to monotherapy with CDKi as compared with B-ALL. Pikman et al. provided recent evidence in T-ALL of synergistic combinations between ribociclib and corticosteroids and everolimus, while demonstrating ribociclib antagonizes a number of cytotoxic chemotherapy agents, including mercaptopurine and asparaginase in in vitro studies of T-ALL cell lines.20,30 Importantly, our data highlight the need to investigate drug schedules using in vivo models, as in vitro models cannot replicate the complexities of drug metabolism, drug-drug interactions, bioavailability, and the microenvironment. In contrast to other groups using in vitro models, we

Haematologica | 107 August 2022

1753

ARTICLE - Chemotherapy and CDKi in ALL

K.L. Bride et al. Figure 6. Simultaneous and not sequential treatment of palbociclib with cytotoxic chemotherapy inhibits disease significantly burden in two Philadelphia chromosome-like B-cell acute lymphoblastic leukemia samples in vivo. Pretreatment with palbociclib (palbo) vs. vehicle (veh) for 7 days by oral gavage, followed by treatment by three common chemotherapy agents: vincristine (VCR, green), L-asparaginase (Asp, blue) or dexamethasone (Dex, red). Concurrent treatment of palbo vs. veh with chemotherapy agents: VCR, Asp or Dex for the duration of the trial of 21 days. Peripheral blood blast count (PBC) over time in two B-cell acute lymphoblastic leukemia (B-ALL) patient-derived xenograft (PDX) (JH561 and NH011). Graphed are means and standard error of PBCx106 per mL on the vertical axis and days of treatment on the X axis. Samples following the pretreatment schema are depicted in the left column, while concurrent treatment is shown in the right column. P-values were calculated using ANOVA to detect the difference between single agent either chemotherapy or CDK inhibitors compared to the combination of both drugs.

Haematologica | 107 August 2022

1754

ARTICLE - Chemotherapy and CDKi in ALL

K.L. Bride et al.

C Figure 7. Simultaneous and not sequential treatment of palbociclib with cytotoxic chemotherapy significantly inhibits disease burden in one T-cell acute lymphoblastic leukemia sample in vivo. (A) Schematic representation of CDK inhibitors (CDKi) and vincristine (VCR) combination regimen in a T-cell acute lymphoblastic leukemia (T-ALL) xenograft (Cul76). Peripheral blood is taken at specific time points as shown. (B) Flow cytometry analysis of peripheral blood analysis of human primary T-ALL patient-derived xenograft (PDX), Cul76, following palbociclib (palbo) treatment. (C) Kaplan-Meier survival graph of T-ALL PDX treated with the combination therapy regimen over a period of 42 days.

found simultaneous administration of CDKi with three different cytotoxic agents was superior to single agent alone, yet a similar benefit was not seen when CDKi were administered prior to cytotoxic chemotherapeutics. Importantly in order to determine if there was a benefit to combination therapy, the doses of some of the cytotoxic agents had to be reduced since they were too effective as single agents. This reflects one constraint of murine models as myelosuppression cannot be abrogated with supportive care such as transfusion, limiting dose intensity. Future studies should also include combinations of CDKi with other cytotoxic agents, including DNA damaging agents such as anthracyclines. We also evaluated the combination of everolimus with CDKi as an alternative approach to treatment for patients with B-ALL. This is under clinical investigation in a phase I clinical trial at the Dana Farber Cancer Institute for relapsed ALL (clinicaltrials gov. Identifier: NCT03740334), based on data provided by Pikman et al.20 Importantly, their initial preclinical study evaluated the combination of these drugs in T-ALL. Our data would suggest this combination is unlikely to be of benefit in B-ALL despite preclinical efficacy of single agent CDKi or mTORi in B-ALL preclinical studies. We found CDKi influenced phosphorylated S6 expression, as seen in Figure 1D, which may explain the lack of benefit from combination therapy. In summary, this work provides a number of novel observations that may impact clinical care and trial design. We provide novel evidence of combination therapy with CDKi to enhance cytotoxic chemotherapy when used simultaneously but not in sequence. We demonstrate the com-

bination of CDKi and mTORi does not merit exploration in B-ALL, despite promising preclinical data in T-ALL by other groups. We identified a novel mechanism of resistance to CDKi. Finally, we found CDKi were effective in a biologically heterogeneous cohort of B-ALL preclinical models. The results of this work led to a recent early phase clinical trial for children and young adults with relapsed/refractory ALL (clinicaltrials gov. Identifier: NCT03792256) and demonstrates the importance of performing rigorous preclinical studies to inform trials that include multi-agent therapy. Disclosures DTT serves on advisory boards for Sobi, BEAM, and Janssen. DTT receives laboratory research funding from BEAM and NeoImmune Tech. All other authors have no conflicts of interest to disclose. Contributions KLB, HH, TF, TLV, AT, IA and DTT designed the study, analyzed data and wrote the manuscript; KLB, HH, TF, TLV, RS, MML and AT contributed to data acquisition and manuscript editing; WLC and EAR contributed to study design and manuscript editing. Funding This work was supported by an LLS SCOR grant (to WLC, EAR and DTT), and grants: R01CA193776, R01CA264837, R03CA256550, X01HD100702-01 and UG1CA233249, Children’s Oncology Group, Alex’s Lemonade Stand Foundation for Childhood Cancer, CHOP Frontiers Program Immune Dysregulation Team, and Cookies for Kids Cancer (to DTT).

Haematologica | 107 August 2022

1755

ARTICLE - Chemotherapy and CDKi in ALL

K.L. Bride et al.

References 1. Iacobucci I, Mullighan CG. Genetic basis of acute lymphoblastic leukemia. J Clin Oncol. 2017;35(9):975-983. 2. Sherr CJ, Roberts JM. Living with or without cyclins and cyclindependent kinases. Genes Dev. 2004;18(22):2699-2711. 3. Sellers WR, Kaelin WG, Jr. Role of the retinoblastoma protein in the pathogenesis of human cancer. J Clin Oncol. 1997;15(11):3301-3312. 4. Fry DW, Harvey PJ, Keller PR, et al. Specific inhibition of cyclindependent kinase 4/6 by PD 0332991 and associated antitumor activity in human tumor xenografts. Mol Cancer Ther. 2004;3(11):1427-1438. 5. Bartkova J, Lukas J, Bartek J. Aberrations of the G1- and G1/Sregulating genes in human cancer. Prog Cell Cycle Res. 1997;3:211-220. 6. Hall M, Peters G. Genetic alterations of cyclins, cyclindependent kinases, and Cdk inhibitors in human cancer. Adv Cancer Res. 1996;68:67-108. 7. Girardi T, Vicente C, Cools J, De Keersmaecker K. The genetics and molecular biology of T-ALL. Blood. 2017;129(9):1113-1123. 8. Carrasco Salas P, Fernandez L, Vela M, et al. The role of CDKN2A/B deletions in pediatric acute lymphoblastic leukemia. Pediatr Hematol Oncol. 2016;33(7-8):415-422. 9. Strefford JC, Worley H, Barber K, et al. Genome complexity in acute lymphoblastic leukemia is revealed by array-based comparative genomic hybridization. Oncogene. 2007;26(29):4306-4318. 10. Mullighan CG, Goorha S, Radtke I, et al. Genome-wide analysis of genetic alterations in acute lymphoblastic leukaemia. Nature. 2007;446(7137):758-764. 11. Kuiper RP, Schoenmakers EF, van Reijmersdal SV, et al. Highresolution genomic profiling of childhood ALL reveals novel recurrent genetic lesions affecting pathways involved in lymphocyte differentiation and cell cycle progression. Leukemia. 2007;21(6):1258-1266. 12. Bernt KM, Hunger SP. Current concepts in pediatric Philadelphia chromosome-positive acute lymphoblastic leukemia. Front Oncol. 2014;4:54. 13. Schwab CJ, Chilton L, Morrison H, et al. Genes commonly deleted in childhood B-cell precursor acute lymphoblastic leukemia: association with cytogenetics and clinical features. Haematologica. 2013;98(7):1081-1088. 14. Jin D, Tran N, Thomas N, Tran DD. Combining CDK4/6 inhibitors ribociclib and palbociclib with cytotoxic agents does not enhance cytotoxicity. PLoS One. 2019;14(10):e0223555. 15. Chen P, Lee NV, Hu W, et al. Spectrum and degree of CDK drug interactions predicts clinical performance. Mol Cancer Ther. 2016;15(10):2273-2281. 16. O'Leary B, Finn RS, Turner NC. Treating cancer with selective CDK4/6 inhibitors. Nat Rev Clin Oncol. 2016;13(7):417-430. 17. Huang X, Di Liberto M, Jayabalan D, et al. Prolonged early G(1) arrest by selective CDK4/CDK6 inhibition sensitizes myeloma cells to cytotoxic killing through cell cycle-coupled loss of IRF4. Blood. 2012;120(5):1095-1106. 18. Bortolozzi R, Mattiuzzo E, Trentin L, et al. Ribociclib, a Cdk4/Cdk6 kinase inhibitor, enhances glucocorticoid sensitivity in B-acute lymphoblastic leukemia (B-All). Biochem Pharmacol. 2018;153:230-241. 19. Franco J, Witkiewicz AK, Knudsen ES. CDK4/6 inhibitors have potent activity in combination with pathway selective therapeutic agents in models of pancreatic cancer. Oncotarget.

2014;5(15):6512-6525. 20. Pikman Y, Alexe G, Roti G, et al. Synergistic drug combinations with a CDK4/6 Inhibitor in T-cell acute lymphoblastic leukemia. Clin Cancer Res. 2017;23(4):1012-1024. 21. Liem NL, Papa RA, Milross CG, et al. Characterization of childhood acute lymphoblastic leukemia xenograft models for the preclinical evaluation of new therapies. Blood. 2004;103(10):3905-3914. 22. Guo QL, Wu MS, Chen Z. Comparison of antitumor effect of recombinant L-asparaginase with wild type one in vitro and in vivo. Acta Pharmacol Sin. 2002;23(10):946-951. 23. Dubois M, Le Joncour V, Tonon MC, et al. Evaluation of the impact of the cancer therapy everolimus on the central nervous system in mice. PLoS One. 2014;9(12):e113533. 24. Maude SL, Tasian SK, Vincent T, et al. Targeting JAK1/2 and mTOR in murine xenograft models of Ph-like acute lymphoblastic leukemia. Blood. 2012;120(17):3510-3518. 25. Teachey DT, Obzut DA, Cooperman J, et al. The mTOR inhibitor CCI-779 induces apoptosis and inhibits growth in preclinical models of primary adult human ALL. Blood. 2006;107(3):1149-1155. 26. Suryani S, Carol H, Chonghaile TN, et al. Cell and molecular determinants of in vivo efficacy of the BH3 mimetic ABT-263 against pediatric acute lymphoblastic leukemia xenografts. Clin Cancer Res. 2014;20(17):4520-4531. 27. Yang C, Boyson CA, Di Liberto M, et al. CDK4/6 inhibitor PD 0332991 sensitizes acute myeloid leukemia to cytarabinemediated cytotoxicity. Cancer Res. 2015;75(9):1838-1845. 28. Sawai CM, Freund J, Oh P, et al. Therapeutic targeting of the cyclin D3:CDK4/6 complex in T cell leukemia. Cancer Cell. 2012;22(4):452-465. 29. Zhang W, Kuang P, Liu T. Prognostic significance of CDKN2A/B deletions in acute lymphoblastic leukaemia: a meta-analysis. Ann Med. 2019;51(1):28-40. 30. Carroll WL, Aifantis I, Raetz E. Beating the clock in T-cell acute lymphoblastic leukemia. Clin Cancer Res. 2017;23(4):873-875. 31. Surrey LF, MacFarland SP, Chang F, et al. Clinical utility of custom-designed NGS panel testing in pediatric tumors. Genome Med. 2019;11(1):32. 32. Takahashi K, Inukai T, Imamura T, et al. Anti-leukemic activity of bortezomib and carfilzomib on B-cell precursor ALL cell lines. PLoS One. 2017;12(12):e0188680. 33. Mariani SA, Minieri V, De Dominici M, et al. CDKN2A-independent role of BMI1 in promoting growth and survival of Ph+ acute lymphoblastic leukemia. Leukemia. 2016;30(8):1682-1690. 34. Ghandi M, Huang FW, Jane-Valbuena J, et al. Next-generation characterization of the Cancer Cell Line Encyclopedia. Nature. 2019;569(7757):503-508. 35. Quentmeier H, Pommerenke C, Dirks WG, et al. The LL-100 panel: 100 cell lines for blood cancer studies. Sci Rep. 2019;9(1):8218. 36. Rokita JL, Rathi KS, Cardenas MF, et al. Genomic profiling of childhood tumor patient-derived xenograft models to enable rational clinical trial design. Cell Rep. 2019;29(6):1675-1689. 37. Roberts KG, Morin RD, Zhang J, et al. Genetic alterations activating kinase and cytokine receptor signaling in high-risk acute lymphoblastic leukemia. Cancer Cell. 2012;22(2):153-166. 38. Furness CL, Mansur MB, Weston VJ, et al. The subclonal complexity of STIL-TAL1+ T-cell acute lymphoblastic leukaemia. Leukemia. 2018;32(9):1984-1993. 39. Zhang J, Ding L, Holmfeldt L, et al. The genetic basis of early

Haematologica | 107 August 2022

1756

ARTICLE - Chemotherapy and CDKi in ALL

K.L. Bride et al.

T-cell precursor acute lymphoblastic leukaemia. Nature. 2012;481(7380):157-163. 40. Shenker BJ, Walker LM, Zekavat A, et al. The cell-cycle regulatory protein p21(CIP1/WAF1) is required for cytolethal distending toxin (Cdt)-induced apoptosis. Pathogens. 2020;9(1):38.

41. Rader J, Russell MR, Hart LS, et al. Dual CDK4/CDK6 inhibition induces cell-cycle arrest and senescence in neuroblastoma. Clin Cancer Res. 2013;19(22):6173-6182. 42. Teh JL, Aplin AE. Arrested developments: CDK4/6 inhibitor resistance and alterations in the tumor immune microenvironment. Clin Cancer Res. 2019;25(3):921-927.

Haematologica | 107 August 2022

1757

ARTICLE - Acute Myeloid Leukemia

The retinoic acid receptor co-factor NRIP1 is uniquely upregulated and represents a therapeutic target in acute myeloid leukemia with chromosome 3q rearrangements Sarah Grasedieck,1 Ariene Cabantog,2 Liam MacPhee,2 Junbum Im,2 Christoph Ruess,3 Burcu Demir,3 Nadine Sperb,3 Frank G. Rücker,3 Konstanze Döhner,3 Tobias Herold,4 Jonathan R. Pollack,5 Lars Bullinger,6 Arefeh Rouhi2# and Florian Kuchenbauer2# University of British Columbia, Department of Microbiology & Immunology, Vancouver, British Columbia, Canada; 2Terry Fox Laboratory, BC Cancer Agency, Vancouver, British Columbia, Canada; 3Ulm University Hospital, Department of Internal Medicine III, Ulm, Germany; 4Department of Medicine III, University Hospital, LMU Munich, Munich, Germany; 5 Department of Pathology, Stanford University School of Medicine, Stanford, CA, USA and 6 Department of Hematology, Oncology and Tumor Immunology, Charité University Medicine, Berlin, Germany. 1

Correspondence: Florian Kuchenbauer fkuchenbauer@bccrc.ca Received: November 16, 2020. Accepted: November 25, 2021. Prepublished: December 2, 2021. https://doi.org/10.3324/haematol.2020.276048 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

AR and FK contributed equally as co-senior authors.

Abstract Aberrant expression of Ecotropic Viral Integration Site 1 (EVI1) is a hallmark of acute myeloid leukemia (AML) with inv(3) or t(3;3), which is a disease subtype with especially poor outcome. In studying transcriptomes from AML patients with chromosome 3q rearrangements, we identified a significant upregulation of the Nuclear Receptor Interacting Protein 1 (NRIP1) as well as its adjacent non-coding RNA LOC101927745. Utilizing transcriptomic and epigenomic data from over 900 primary samples from patients as well as genetic and transcriptional engineering approaches, we have identified several mechanisms that can lead to upregulation of NRIP1 in AML. We hypothesize that the LOC101927745 transcription start site harbors a context-dependent enhancer that is bound by EVI1, causing upregulation of NRIP1 in AML with chromosome 3 abnormalities. Furthermore, we showed that NRIP1 knockdown negatively affects the proliferation and survival of 3qrearranged AML cells and increases their sensitivity to all-trans retinoic acid, suggesting that NRIP1 is relevant for the pathogenesis of inv(3)/t(3;3) AML and could serve as a novel therapeutic target in myeloid malignancies with 3q abnormalities.

Introduction Ectopic activation of the Ecotropic Viral Integration Site 1 (EVI1) gene is associated with a dismal outcome in patients with acute myeloid leukemia (AML) or myelodysplastic syndrome (MDS), who have an average survival of only 10 months after diagnosis.1 EVI1 is located in the MDS1 And EVI1 Complex Locus (MECOM) on chromosomal band 3q26.2, encodes a DNA-binding protein with two zinc finger domains, and is expressed in hematopoietic stem and progenitor cells. In an inducible mouse model, Evi1 overexpression led to the suppression of erythropoiesis and lymphopoiesis, driving a pre-leukemic expansion of myeloid cells which ultimately led to leukemic transformation,2 suggesting that activation of EVI1 drives myeloid leukemias. Moreover, ectopic activation of the EVI1 gene through vector integration was associated with the development of AML in a gene therapy trial.3 Although upregulated EVI1 is the defining molecular char-

acteristic of AML with inv(3) or t(3;3), high EVI1 expression has been reported in approximately 11% of all adult AML cases, in which it was suggested to be an independent adverse prognostic factor.1,4-7 Currently, there are no targeted therapies or additional prognostic indicators available for myeloid malignancies with abnormal 3q or high EVI1 expression and the mechanisms that cause or contribute to EVI1 upregulation remain largely unclear. In 2012, Haferlach et al. reported seven AML cases with translocation t(3;21)(q26;q11). This led to the formation of an EVI1 fusion protein with the Nuclear Receptor Interacting Protein 1 (NRIP1), which the authors found to be associated with an especially poor prognosis.8 Additionally, a more recent report described a poor prognosis, therapy-induced childhood AML with a cryptic t(3;21)(q26;q11), leading to NRIP1-EVI1 fusion, which displayed high EVI1 expression.9 While studying transcriptomes from AML and MDS patients with or without 3q rearrangements, we discovered that both NRIP1 as well

Haematologica | 107 August 2022

1758

ARTICLE -NRIP1 in EVI1-expressing AML

S. Grasedieck et al.

as its neighboring non-coding gene, LOC101927745, were upregulated in MECOM-rearranged AML without t(3;21) fusions, such as inv(3) and t(3;3). NRIP1 and its neighboring gene LOC101927745, share a topologically associating domain (TAD). TAD are genomic regions that are delimited and insulated from external regulatory influences by CCCTC-binding factor (CTCF)- and cohesinbound sites.10 Genes and regulatory elements within a TAD were shown to physically interact with each other more frequently than with sequences located outside their TAD,10 prompting us to investigate whether a potential co-regulatory relationship exists between NRIP1 and LOC101927745 which could lead to upregulation of NRIP1 in MECOM-rearranged AML. Although NRIP1 has been implicated in differentiation processes via modulation of retinoic acid (RA) receptors11,12,13 and regulation of energy metabolism,14 its role in myeloid malignancies is largely unknown. Hypothesizing that NRIP1 could function as a yet undescribed proto-oncogene in AML cells, we utilized public transcriptomic and epigenomic data collected from more than 900 AML and MDS patients and conducted genetic and transcriptional engineering approaches to: (i) identify potential mechanisms that can lead to upregulation of NRIP1 and (ii) investigate how the perturbation of NRIP1 expression would affect inv(3)/t(3;3) AML cells.

Methods Patients’ data and experimental datasets Most datasets that were analyzed for this study are available either through the National Cancer Institute’s Genomic Data Commons portal (TCGA-LAML,15 Beat AML,16 https://portal.gdc.cancer.gov/), the International Human Epigenome Consortium data portal (BLUEPRINT primary AML samples,17 sample identities: ERS699839, ERS699842, ERS699843, ERS753996; https://epigenomesportal.ca/ihec/) or the National Center for Biotechnology Information’s Gene Expression Omnibus database (https://www.ncbi.nlm.nih.gov/gds) under the following identifiers: GSE114922: healthy and MDS CD34+ cells, RNA-seq (n=130);18 GSE106291: AMLCG-2008 & AMLG-1999 AML samples, RNAseq (n=250);19 GSE104099: cytogenetically normal AML, AMLSG_07-04, RNA-seq (n=46);20 GSE35159: expression profiling in 12 human myeloid cell lines, microarray;21 GSE123255: murine leukemic stem cell enriched cells (LSCe) +/- ATRA;22 GSE31477: ENCODE TF and co-factors in various cell lines, ChIP-seq;23 GSE32465: ENCODE TF and co-factors in various cell lines, ChIP-seq;24

GSE36030: murine B10 cell line, Rad21 ChIP-seq (ENCODE mouse project); GSE136488: murine E14 cell line, Ctcf ChIP-seq (ENCODE mouse project); GSE55407: THP-1 AML cell line, CTCF and Rad21 ChIP-seq;25 GSE87286: SKH-1 AML cell line, ChIP-seq and RNA-seq;26 GSE72816: Gm12878 cell line ChIA-PET cluster data;27 GSE63525: Gm12878 and K562 Hi-C chromatin contact data;28 PRJNA385337: THP-1 Hi-C chromatin contact data;29 GSE52457: H1-derived hMSC Hi-C chromatin contact data;30 GSE84662: keratinocytes Hi-C chromatin contact data.31 Survival analyses Cox proportional hazard and Kaplan-Meier analyses for association with overall or event-free survival were calculated using the R survival32 and survminer33 packages. Patients were dichotomized into groups expressing the gene of interest at a high or low level based on maximally selected rank statistics.34 Patients’ samples and cell lines RNA sequencing was performed on primary samples of viably frozen bone marrow from patients (n=65) with a complex karyotype (CK-AML) (Online Supplementary Methods). Forty of the 65 (62%) patients were treated on consecutive multicenter treatment trials of the AML Study Group (AMLSG), applying age-adjusted intensive chemotherapy: AMLHD98A (n=4; NCT00146120) and AMLSG07-04 (n=22; NCT00151242) for younger patients (16 to 60 years); AMLSG06-04 (n=14; NCT00151255) for elderly patients (>60 years). All studies were approved by local ethics committees, and all patients gave informed consent to treatment, cryopreservation of samples, and molecular analyses according to the Declaration of Helsinki. All the cell lines used were obtained from the German Collection of Microorganisms and Cell Cultures (DSMZ), except for the Cas9-expressing OCI-AML5 cells which were a gift from Dr. Jan Krönke (Ulm University Hospital, Germany). All cells were maintained in adherence to the culturing conditions recommended by the DSMZ. All-trans retinoic acid treatment and analysis All-trans retinoic acid (ATRA; Sigma-Aldrich, Germany) was prepared in dimethylsulfoxide (DMSO) at 100 mM and further diluted in phosphate-buffered saline. Cell lines that were transfected with NRIP1-targeting GapmeR or control or with NRIP1-targeting shRNA or control after selection (see Online Supplementary Methods) were seeded at 0.25x106 cells per well in 2 mL culture medium at a final concentration of 0.5 mM ATRA or a concentration-matched DMSO control. Cells were analyzed after 24 and 72 h for cell proliferation by counting trypan-negative cells, for gene expression of NRIP1, MECOM, and LOC101927745, for NRIP1 protein expression (see Online Supplementary Methods), and for apoptosis after

Haematologica | 107 August 2022

1759

ARTICLE -NRIP1 in EVI1-expressing AML

S. Grasedieck et al.

staining with an annexin-V-APC antibody (Biolegend, USA) abnormalities of inv(3)/t(3;3) (Online Supplementary Figure and SYTOX-Blue dead stain (Invitrogen, USA) using flow cyto- S2). The association of high NRIP1 or LOC101927745 expression metry. with poorer outcome was further corroborated in the AMLCG-2008 cohort19 as well as in RNA-sequencing data from CK-AML patients37 (Figure 2B, C, Online Supplementary Methods). Of note, LOC101927745 and NRIP1 transcript levels Results exceeded the prognostic stratification capacity of EVI1 levels The NRIP1-EVI1 fusion positions the EVI1 open reading in all analyzed datasets both when applying a dichotomizaframe under control of the NRIP1 gene regulatory elements tion approach based on maximally selected rank statistics34 To understand which functional domains of the NRIP1 and (Figure 1A-C, Online Supplementary Figure S2) or the median EVI1 proteins were lost and retained in a fusion event as re- (data not shown). ported by Haferlach et al.8 and D'Angiò et al.,9 we studied the In line with other studies reporting EVI1 expression in the abexact t(3;21)(q26;q11) breakpoints and found that the NRIP1- sence of detectable EVI1 fusion events or inv(3), only 11/126 EVI1 fusion does not generate a novel chimeric protein, but (8.7%) and 5/28 (17.8%) of EVI1-expressing patients carried instead removes most of the upstream regulatory elements molecularly detectable 3q abnormalities in the analyzed Beat of EVI1 and places the complete EVI1 coding sequence (exons AML and CK-AML cohorts, respectively (2x dupl(3q26), 3 to 16) under the control of the three putative NRIP1 pro- t(3;21)(q26;q22), t(3;3)(q26;q21), and t(3;6)(q26;p22)), in which moters and additional upstream regulatory elements (Figure expression was defined as EVI1 >5 transcripts per kilobase 1A). Similar to what was observed in inv(3)(q21q26) or million. Although expression of LOC101927745 and NRIP1 was t(3;3)(q21;q26) AML cases, in which EVI1 was reported to ap- particularly enriched in inv(3) AML (i.e., EVI1high), and both tranpropriate multiple enhancer sites of the GATA2 gene,35,36 we scripts showed a high degree of correlation in EVI1high healthy found that a t(3;21)(q26;q11) event effectively places the com- and MDS CD34+ blast cells (R=0.78 and 0.89) (Figure 2D), we plete EVI1 open reading frame under the control of a different observed that in AML this correlation was perturbed (R=0.31 transcriptional network. and 0.64) (Figure 2E). Compared to healthy and MDS samples, in which all NRIP1-expressing cells also expressed Expression of NRIP1 and LOC101927745 is associated with LOC101927745, NRIP1 and LOC101927745 RNA transcripts only poor survival in acute myeloid leukemia showed correlation in 46% of AML cases. However, AML paAnalysis of published RNA-sequencing datasets, comprising tients who expressed LOC101927745 always expressed NRIP1 sorted healthy human donor bone marrow cell populations (Figure 2E). Considering this highly specific pattern of ex(n=56),18 blasts from MDS patients (n=74),18 and primary AML pression, we next investigated a potential co-regulatory resamples (n=950, including TCGA L-AML,15 AMLCG-2008/1999,19 lationship between LOC101927745 and NRIP1 in AML. an in-house generated cytogenetically normal AML20 and a CK-AML cohort37 as well as data from the Beat AML trial,16 The LOC101927745 transcription start site contains an showed that transcript levels of NRIP1, located on chromo- NRIP1 gene regulatory element some 21q, as well as its neighboring gene, LOC101927745, NRIP1 and its neighboring gene LOC101927745 share a TAD were highly expressed in CD34+ hematopoietic stem and pro- (Online Supplementary Figure S3A, B), suggesting a co-regugenitor cells from healthy donors and MDS patients (Online latory relationship between these genes. In support of this Supplementary Figure S1A) and were significantly upregulated hypothesis, we found that the exonic regions of in inv(3)/t(3;3) AML patients compared to all other cytogen- LOC101927745, its genomic location and its orientation etic AML subgroups (Padj.=0.01 and 0.003, respectively) (Figure relative to the murine Nrip1 homolog as well as their 1B). We were able to confirm this exceptionally high ex- Ctcf:Rad21 TAD boundaries are conserved between human pression of LOC101927745 and NRIP1 in EVI1high- compared to and mouse genomes (Online Supplementary Figure S3B-D). EVI1low-expressing AML cell lines (total of n=4 vs. n=17) (Online Of note, the region overlapping the putative transcription Supplementary Figure 1B-D). start site (TSS) of LOC101927745 is annotated as GH21J015439 Based on the report by Haferlach et al., which associated the in most recent genome browser versions, due to its classifipresence of an NRIP1/EVI1 fusion gene with an especially poor cation as a candidate enhancer of NRIP1 by the GeneHancer prognosis,8 we next examined whether NRIP1 and project38 (Online Supplementary Figure S4A, based on eQTL LOC101927745 transcript levels were linked to outcome in and Hi-C data). A defining feature of enhancers is that they AML patients. Indeed, survival analyses using the Beat AML establish measurable physical contact with the promoters RNA-sequencing dataset16 showed that high NRIP1 as well as of their regulated genes. As there are currently no chromatin high LOC101927745 expression were significantly associated conformation capture data available from cells with chrowith poorer overall survival (Figure 2A), independently of the mosome 3 abnormalities, we instead compared Hi-C data inclusion of inv(3)/t(3;3), complex karyotype and also of from NRIP1low leukemia cell lines (THP-1, K562 and Gm12878) del(5q)/del(7q) cases, which represent the most common co- to data generated from human skin samples, which show Haematologica | 107 August 2022

1760

ARTICLE -NRIP1 in EVI1-expressing AML

S. Grasedieck et al.

strong NRIP1 as well as EVI1 expression when compared to mature blood cells (Figure 3A). When we subtracted normalized chromatin contact counts obtained in NRIP1low AML cell lines from counts recorded in NRIP1high/EVI1high human skin cells, we found that contact was 70- to 330-fold increased specifically between the LOC101927745 TSS (GH21J015439) and the NRIP1 genomic locus in NRIP1high-expressing cells

(Figure 3B, blue areas depict regions where contact intensities in NRIP1/EVI1high tissues exceed contact in NRIP1low AML cells). Collectively, these observations suggest that LOC101927745 likely harbors an NRIP1-controlling regulatory element. To further explore the transcription factor (TF) binding and chromatin state at the LOC101927745/NRIP1 locus in AML cells

Figure 1. LOC101927745 and NRIP1 are upregulated in EVI1high acute myeloid leukemia. (A) Schematic representation of the EVI1 and NRIP1 gene loci including the breakpoints and resulting fusion product in t(3;21)(q26;q11) cases based on the report by Haferlach et al.8 (B) Transcript levels of LOC101927745, NRIP1, and EVI1 (MECOM) in a subset of acute myeloid leukemia patients with recurrent cytogenetic and molecular abnormalities from the Beat AML Master Trial.16 Data were TMM-normalized and are presented as counts per million. P-values were calculated comparing individual groups to all cases, using the Welch two sample t-test with 95% confidence and adjusted for multiple hypothesis testing. The group t(3;other) comprises a t(2;3)(p13;q25~26) and a t(3;11)(p21.3;p11.2) sample. CDS: coding sequence; UTR: untranslated region; CPM: counts per million; CN: cytogenetically normal; MLLr: MLL gene locus rearrangement; TMM: trimmed mean of M values. Haematologica | 107 August 2022

1761

ARTICLE -NRIP1 in EVI1-expressing AML

S. Grasedieck et al.

Continued on the following page. Haematologica | 107 August 2022

1762

ARTICLE -NRIP1 in EVI1-expressing AML

S. Grasedieck et al.

Figure 2. LOC101927745 and NRIP1 transcript levels are negatively associated with survival in patients with acute myeloid leukemia. (A-C) Graphical representation of Kaplan-Meier estimates based on EVI1 (MECOM), NRIP1 and LOC101927745 expression calculated in: (A) the adult acute myeloid leukemia (AML) subset of the Beat AML Master Trial cohort;16 (B) the AMLCG-2008 cohort;19 and (C) a subset of patients from the complex karyotype (CK)-AML cohort37 for whom survival information was available. P-values were calculated using the log-rank test. Patients were dichotomized into groups with high or low expression of the gene of interest using maximally selected rank statistics.34 (D) Pairwise correlations of TMM-normalized RNA-sequencing data for LOC101927745 and NRIP1 from healthy CD34+ hematopoietic stem and progenitor cells (GSE114922), primary blast cells from patients with myelodysplastic syndrome (GSE114922) and complex karyotype-AML patients’ AML cells37 with the respective Pearson R, where R=1 describes a perfect correlation. (E) Pairwise correlation of TMM-normalized RNA-sequencing data for LOC101927745 and NRIP1 in the AMLCG-2008 cohort with the respective Pearson R, where R=1 describes a perfect correlation. Data were voom-transformed (variance modeling at the observational level, see Online Supplementary Methods). OS: overall survival; CK: complex karyotype; MDS: myelodysplastic syndrome; TMM: trimmed mean of M values.

with intact chromosome 3 and with low NRIP1 expression, we analyzed public chromatin immunoprecipitation (ChIP)sequencing data assessing TF occupancy and ChIP-sequencing data from AML patients’ blasts and cell lines (data from BLUEPRINT17 and ENCODE23,24). NRIP1low AML cell lines and primary cells displayed a universal lack of TF binding and were devoid of activating histone modifications in the LOC101927745/NRIP1 TAD (data not shown). Instead, these cells displayed an accumulation of repressive histone marks such as H3K27me3 (associated with promotor repression) and H3K9me3 (associated with permanent heterochromatin formation) (Online Supplementary Figure S4B). In summary, these findings support our hypothesis that the LOC101927745 genomic site is of relevance for the transcriptional regulation of NRIP1. NRIP1 transcription is independently regulated by retinoic acid signaling and the GH21J015439 enhancer in acute myeloid leukemia cells The promoter sequences of NRIP1 were reported to be rich in RA receptor binding sites.39 Another hint towards the relevance of RA signaling for NRIP1 transcription is that t(15;17) AML patients’ samples and the t(15;17) RA receptor dysfunctional NB-4 cell line do not transcribe NRIP1 at all (Figure 1B, Online Supplementary Figure S1A). Therefore, we aimed to assess whether an external stimulation of RA receptor signaling would induce NRIP1 expression in an NRIP1low-expressing AML model. Concurrently, we also aimed to determine whether the deletion of the putative regulatory element GH21J015439, which is embedded within the LOC101927745 TSS, would have an impact on any hypothetical RA-mediated effects on NRIP1 transcription. We, therefore, deleted a 470 bp genomic region spanning GH21J015439, LOC101927745 exon 1, and part of exon 2 (Online Supplementary Figure S4A, C) using a CRISPR/Cas9guided approach in the human OCI-AML5 AML cell line (LOCKO), which inherently displays low NRIP1 and lack of LOC101927745 expression (Online Supplementary Figure S1A). Genomic deletion led to a modest but significant reduction of NRIP1 mRNA levels in LOC-KO compared to wild-type OCIAML5 control cells (LOC-WT) (P=0.05) (Figure 4A), confirming that the deleted site likely harbors an NRIP1-enhancing regulatory element. Treatment of both LOC-KO and LOC-WT cell lines with the RA receptor agonist ATRA led to a pronounced

upregulation of NRIP1 expression in both cell lines (Figure 4A). As ATRA is known to induce differentiation in AML cells, including OCI-AML5 (Online Supplementary Figure S5), we next tested whether expression of NRIP1 was truly the result of RA receptor signaling or merely a side effect of differentiation. Thus, we independently treated both cell lines with the protein kinase C-activator 12-O-tetra-decanoylphorbol-13-acetate (TPA), which induces differentiation of OCI-AML5 cells in an RA-independent manner.40 TPA treatment did not affect NRIP1 transcription, suggesting that NRIP1 was indeed induced by RA receptor signaling and is differentiation independent. Of note, genomic deletion of GH21J015439 also led to a significant reduction of NRIP1 mRNA levels in LOC-KO compared to LOC-WT cells treated with TPA (P=0.03) (Figure 4A) again highlighting the positive regulatory effect of this element on NRIP1 transcription. In addition, LOC-KO cells displayed increased expression of the mature myeloid surface marker CD11c and a reduced proliferation rate compared to LOC-WT cells (Online Supplementary Figure S5B, C). We were able to confirm the pronounced upregulation of NRIP1 in response to ATRA treatment in an RNA-sequencing dataset generated by Nguyen et al. in MLL-AF9-expressing murine leukemic stem cell-enriched fractions that were treated with 1 mM ATRA or control for 24 h.22 In these cells, ATRA treatment resulted in an 8.8-fold increase of Nrip1 expression (P=0.006) (Figure 4B) and, interestingly, ATRA-induced expression of Nrip1 was antagonized by simultaneous shRNA-mediated Evi1-knock-down (7.6-fold increase relative to vehicle-treated sh-control), which reduced Nrip1 levels even more strongly in vehicle controls alone (6.6-fold decrease, P=0.07) (Figure 4B). This observation hints at a relevance of NRIP1 expression for EVI1high AML and suggests that NRIP1 could be under the direct transcriptional control of the EVI1 TF complex. NRIP1 expression is regulated by the oncogenic EVI1 transcription factor network To explore whether EVI1 contributes directly to the transcriptional control of NRIP1, we first ensured that there are EVI1 binding sites present within the NRIP1/LOC101927745 TAD and identified 16 sites with a significant z-score that are conserved between human, mouse and rat (Online Supplementary Figure S6). No EVI1 binding motifs were present within

Haematologica | 107 August 2022

1763

ARTICLE -NRIP1 in EVI1-expressing AML

S. Grasedieck et al.

Figure 3. Contact between the LOC101927745 transcription start site and the NRIP1 genomic site is specifically enriched in NRIP1high/EVI1high tissues. (A) RNA expression of NRIP1, LOC101927745 and EVI1/MECOM in the indicated tissues. Expression data were obtained from the GTEx Portal, dbGaP accession number phs000424.vN.pN in March 2021. (B) Heatmaps generated by subtracting sequencing-based Hi-C chromatin contact quantification data which were obtained in NRIP1high/EVI1high skin tissues: human H1 ESC-derived fibroblast-like mesenchymal stromal cells (H1-MSC, left) and human epidermal keratinocytes (right), from Hi-C quantification data obtained in NRIP1low/EVI1negative leukemia cell lines (THP-1, K562, and Gm12878). Blue areas on heatmaps represent sites of high contact in NRIP1high/EVI1high tissues whereas red areas have greater contact intensities in leukemia cells. The black triangle indicates the region connecting the LOC101927745 transcription start site and the NRIP1 gene body. All presented Hi-C data were visualized using the Yue Lab 3D Genome Browser, available at: http://3dgenome.org. Data from H1-MSC were generated by Dixon et al.,30 from keratinocytes by Rubin et al.,31 from THP-1 by Phanstiel et al.,29 and from Gm12878 by Rao et al.28 TPM: transcripts per kilobase million. Haematologica | 107 August 2022

1764

ARTICLE -NRIP1 in EVI1-expressing AML

S. Grasedieck et al.

the NRIP1 promoter but a total of six conserved EVI1 binding sites were present within the LOC101927745 gene, including the putative NRIP1 enhancer GH21J015439, which overlaps the LOC101927745 TSS. To confirm that any of these motifs are actually bound by EVI1 in AML cells, we analyzed ChIP-sequencing data produced by Loke et al. from EVI1-knockdown

vs. control SKH-1 AML cells, which harbor a t(3;21)(q26;q22) translocation. The exact SKH-1 breakpoint in chromosome 21q is located more than 19 Mb upstream of the NRIP1 genomic locus which is therefore retained in this cell line.41 Translocation t(3;21) causes expression of a fused RUNX1EVI1 TF, which was found to form an abnormal transcription-

Figure 4. NRIP1 transcription is induced by retinoic acid signaling and reduced upon knockdown of EVI1. (A) Quantitative realtime polymerase chain reaction of NRIP1 in LOC101927745 transcription start site-knockout (LOC-KO) and control OCI-AML5 cells treated with either 0.1 mM ATRA or 1 nM TPA for 72 h in three independent experiments. Total RNA was extracted before, after 24 h and after 72 h of treatment, DNase-digested and reverse transcribed. 2ddCt was calculated relative to SDHC. P-values were calculated using a paired two-tailed Student t-test. (B) Normalized RNA-sequencing data for Nrip1 and Mecom, generated by Nguyen et al.22 in MLL-AF9-expressing murine leukemic stem enriched cell fractions treated with 1 mM ATRA or dimethylsulfoxide control for 24 h. (C) Normalized RNA-sequencing data quantifying NRIP1 and EVI1 transcripts, both generated from t(3;21)(q26;q22) SKH1 AML cells treated with either EVI1-targeting- or control siRNA. Presented data combined all replicates analyzed 48 h (n=2) and 96 h (n=2) after siRNA treatment to increase statistical power. Data were generated by Loke et al.26 (D) Chromatin immunoprecipitation-sequencing footprinting data for RUNX1, EVI1 and GATA-2, which form an abnormal transcription-activating complex in t(3;21) AML. Data was generated by Loke et al.26 as described in C. P-values were calculated using the Welch two sample t-test. LOC-KO: LOC101927745 transcription start site-knockout; ATRA: all-trans retinoic acid; (LOCKO); TPA: 12-O-tetra-decanoylphorbol-13-acetate; FC: fold change; KD: knockdown; DMSO: dimethylsulfoxide; RPKM: reads per kilobase million; FPKM: fragments per kilobase per million. Haematologica | 107 August 2022

1765

ARTICLE -NRIP1 in EVI1-expressing AML

S. Grasedieck et al.

activating complex together with GATA-2 and ETS factors.26 When analyzing EVI1, RUNX1, and GATA-2 ChIP-sequencing data produced in SKH-1 cells, we found that they displayed strong binding at two EVI1 motifs within the LOC101927745 gene, with the strongest binding of all three TF at the GH21J015439 putative NRIP1 regulatory site as well as high NRIP1 transcript levels in the corresponding RNA-sequencing data (Figure 4C, D). Upon shRNA-mediated EVI1:RUNX1 knockdown, both TF binding and NRIP1 expression were significantly reduced in SKH-1 cells compared to controls (Figure 4C, D), supporting our hypothesis that the NRIP1 gene is under the direct control of the abnormal EVI1 TF complex that drives this AML phenotype. Assuming that LOC101927745 upregulation is indeed controlled by an oncogenic EVI1 TF complex, LOC101927745 and EVI1 transcript levels would be expected to correlate in patients’ samples. Indeed, similar to our observations regarding a correlation between LOC101927745 and NRIP1 RNA, transcript levels of LOC101927745 and EVI1 showed a high degree of correlation in normal and MDS hematopoietic stem and progenitor cells (R=0.87) (Figure 5A). According to AML data, 2.9% to 10.5% of all patients expressed both transcripts (LOC+/EVI1+), 66.9% to 78% expressed neither (LOC-/EVI1-), 3.4% to 8.7% expressed only EVI1 in the absence of LOC101927745 (LOC-/EVI1+) and 13.9% to 15.7% of patients only expressed LOC101927745 (LOC+/EVI1-), suggesting that LOC101927745 and therefore NRIP1 transcription is not exclusively regulated through EVI1 in AML patients (Figure 5A). When comparing survival among these four groups, patients who expressed LOC101927745 RNA had significantly worse outcome or response to treatment, independently of EVI1 expression (P=0.04 and 0.0018 in the AMLCG-2008 and Beat AML cohort, respectively) (Figure 5B). Of note, expression of the LOC101927745 transcript was able to further sub-stratify adverse-risk EVI1high AML patients, highlighting cases with especially poor outcome. NRIP1 knockdown affects proliferation, viability, and response to all-trans retinoic acid in chromosome 3 rearranged acute myeloid leukemia cells As NRIP1 is strongly upregulated in AML cases with chromosome 3q rearrangements, we further assessed the dependence of EVI1-expressing AML blasts on NRIP1 expression. Therefore, we performed both transient antisense- and stably integrated shRNA-mediated knockdown of NRIP1 in EVI1high and EVI1negative AML cell lines with NRIP1 expression. Stable NRIP1 knockdown significantly affected growth and viability in t(3;3) UCSD-AML1 and HNT-34 (Figure 6A, B) but to a lesser extent or not at all in chromosome 3 intact OCIAML3, Kasumi-1, and K562 cells (Online Supplementary Figure S7A, B). Knockdown of NRIP1 was confirmed at RNA and protein levels in all cell lines (Figure 6C, D, Online Supplementary Figures S7C, S8 and S9). NRIP1 knockdown rendered t(3;3) cells significantly more sensitive to ATRA treatment, as

exemplified by decreased proliferation (HNT-34 88% and UCSD-AML1 39% reduction) (Figure 6A) and higher levels of apoptosis after 72 h of treatment compared to controls (HNT-34 46% and UCSD-AML1 37% increase in apoptotic cells) (Figure 6B). In line with our earlier findings in OCI-AML5 cells, ATRA treatment induced expression of NRIP1 and, of note, also resulted in increased EVI1 transcription in t(3;3) cells as well as in the EVI1-negative cell line OCI-AML3 (Online Supplementary Figures S10 and S11). Furthermore, NRIP1knockdown resulted in significantly elevated transcription of LOC101927745 RNA exclusively in t(3;3) cells (Figure 6C, Online Supplementary Figure S10). In all EVI1-expressing cell lines, including chromosome 3 normal K562 cells, NRIP1-knockdown resulted in reduced EVI1 transcription (Figure 6C, Online Supplementary Figure S7). Proliferation and apoptosis data recorded in t(3;3) UCSD-AML1 and HNT-34 cells as well as in chromosome 3 normal OCI-AML3 and NRIP1negative NB-4 cells after transient transfection with NRIP1-targeting GapmeRs, confirmed our observation that cell lines with chromosome 3q rearrangements were more vulnerable to NRIP1-knockdown, especially in combination with ATRA treatment (Online Supplementary Figure S11A-C). In contrast, GapmeR-mediated knockdown of LOC101927745 RNA in UCSD-AML1 did not affect RIP1 transcription or proliferation and proliferation and cell viability, supporting a regulatory model in which the LOC101927745 genomic site - i.e., its function as an enhancer - but not its RNA transcript, is relevant for the control of NRIP1 expression (Online Supplementary Figure S11D).

Discussion In a large collection of transcriptomic and epigenomic datasets, we found that expression of LOC101927745 and NRIP1 on chromosome 21 is markedly upregulated in AML patients with chromosome 3q abnormalities. Although the majority of AML patients do not express LOC101927745 at comparably high levels, overall between 18.6% and 24.4% of AML cases had detectable LOC101927745 transcription in the Beat AML and AMLCG-2008 cohorts, and this was associated with especially poor outcome in AML and was independent of EVI1 transcription status. In all the datasets we studied, we found that whenever LOC101927745 is transcribed, so too is NRIP1 and that their expression correlates significantly. Based on our findings, we propose the following functional interaction model between NRIP1 and LOC101927745 in AML: blast cells that do not express either LOC101927745 or NRIP1 display a repressive heterochromatin state in the LOC101927745/NRIP1 TAD (Figure 7A). In EVI1high MDS and AML cells, NRIP1 transcription is regulated predominantly through usage of the GH21J015439 enhancer site, which is embedded within the LOC101927745 TSS. Binding of the enhancer is at least in part mediated by the

Haematologica | 107 August 2022

1766

ARTICLE -NRIP1 in EVI1-expressing AML

S. Grasedieck et al.

Figure 5. Transcription of LOC101927745 is associated with negative outcomes independently of EVI1 status in patients with acute myeloid leukemia. (A) Left: pairwise correlations of TMM-normalized RNA-sequencing data for EVI1 (MECOM) and LOC101927745 in primary blast cells from patients with myelodysplastic syndrome (GSE114922) with the respective Pearson R, where R=1 describes a perfect correlation. Right: Kaplan-Meier plots based on LOC101927745 and EVI1 (MECOM) expression calculated in the subset of patients with available survival information. (B) Left: pairwise correlations of TMM-normalized RNA-sequencing data for EVI1 (MECOM) and LOC101927745 from the AMLCG-2008 cohort19 with respective Pearson R. Data were voom-transformed (variance modeling at the observational level, see Online Supplementary Methods). Right: Kaplan-Meier plot based on LOC101927745 and EVI1 (MECOM) transcript levels stratified as indicated by the color coding in the subset of patients with available survival information of the same dataset. (C) Left: pairwise correlations of TMM-normalized RNA-sequencing data for EVI1 (MECOM) and LOC101927745 from the Beat AML trial16 with the respective Pearson R. Axes are log10 transformed. Right: Kaplan-Meier plot based on LOC101927745 and EVI1 (MECOM) expression stratified as indicated by the colour coding in the subset of patients with available survival information of the same dataset. MDS: myelodysplastic syndrome; TMM: trimmed mean of M values; OS: overall survival. Haematologica | 107 August 2022

1767

ARTICLE -NRIP1 in EVI1-expressing AML

S. Grasedieck et al.

Continued on the following page. Haematologica | 107 August 2022

1768

ARTICLE -NRIP1 in EVI1-expressing AML

S. Grasedieck et al.

Figure 6. Knockdown of NRIP1 induces apoptosis and increases sensitivity to all-trans retinoic acid in t(3;3) acute myeloid leukemia cells. (A, B) t(3;3) HNT-34 and UCSD-AML1 cells were lentivirally transduced with constructs encoding either NRIP1targeting or two different control shRNA in three independent experiments. Puromycin-selected cells were seeded at 0.25x106 cells/mL and treated with with either 0.5 mM all-trans retinoic acid or dimethylsulfoxide control and a total of three replicates per condition were harvested and analyzed after 24 h and 72 h. (A) Cells were resuspended, stained with trypan blue and counted at the indicated time points. The figure shows the total number of trypan-negative (live) cells per well over time. (B) Cells were stained with annexin-V-APC and Sytox viability dye at the indicated time points and 10,000 cells per sample were recorded via flow cytometry. The figure shows the percentage of annexin/Sytox double negative (i.e., non-apoptotic) cells relative to day 0 of the experiment. (C) RNA was extracted from three individual treatment samples per condition, DNase-digested and reverse-transcribed to cDNA. In the same sample, RNA levels of NRIP1, EVI1 (MECOM), and the housekeeping gene ABL1 were quantified using commercially available TaqMan assays in technical triplicates. Transcript levels of LOC101927745 (exon 1) were quantified using SYBR Green chemistry and compared to the predetermined levels of the housekeeping gene SDHC. Fold changes were calculated using the 2ddCt method with sh-contr. samples set to 1. (D) Total protein was extracted, BSA-quantified and western blots were performed with NRIP1- and b-actin targeting antibodies from three individually generated cell lines per condition. The signal intensity of NRIP1 relative to b-actin bands was quantified using Fiji/Image J. All P-values shown were calculated using paired, two-tailed Student ttests for unequal variance. ATRA: all-trans retinoic acid; DMSO: dimethylsulfoxide.

Figure 7. Schematic model of NRIP1 transcriptional regulation in acute myeloid leukemia. (A) Cells that neither express LOC101927745 nor NRIP1 display tightly coiled heterochromatin that is marked by repressive histone modifications and is inaccessible to transcription factors (TF). (B) In EVI1-expressing acute myeloid leukemia (AML) cells, the oncogenic EVI1 TF complex binds the GH21J015439 enhancer located in the LOC101927745 transcription start site, which causes the enhancer complex to loop onto the NRIP1 promoter sites. This process delivers the RNA transcription machinery to the site of contact, resulting in simultaneous transcription of the LOC101927745 and NRIP1 genes. (C) In NRIP1-expressing cells that do not express the EVI1 TF, the NRIP1 promoters are bound and transcription is activated directly, without the involvement of GH21J015439, by other TF (as was shown for retinoic acid- and estrogen receptors). Top: while the NRIP1 gene locus is accessible, the LOC101927745 genomic site is in a repressed, inaccessible state. Bottom: an alternative mode of NRIP1 transcriptional activation in EVI1negative/NRIP1low-high cells involves binding of TF other than EVI1 to the GH21J015439 enhancer promoting NRIP1 transcription. (D) CRISPR-Cas9-mediated excision of GH21J015439 (LOC-KO) in EVI1negative/NRIP1low expressing AML cells only modestly affected NRIP1 transcription and did not prevent promotermediated activation of the NRIP1 gene through retinoic acid receptor (RAR):retinoid X receptor (RXR) heterodimers, which were induced via external stimulation with all-trans retinoic acid. AML: acute myeloid leukemia; RA: retinoic acid; ER: estrogen receptor; ATRA: all-trans retinoic acid; TF: transcription factor; RAR: retinoic acid receptor; RXR: retinoid X receptor. Haematologica | 107 August 2022

1769

ARTICLE -NRIP1 in EVI1-expressing AML

S. Grasedieck et al.

EVI1 oncogenic TF complex, which does not bind the NRIP1 promoter directly but instead forms a looped chromatin contact as is common for enhancer:promoter interactions42 (Figure 7B). This process recruits the transcriptional machinery to the site of contact which results in simultaneous transcription of both the LOC101927745 and NRIP1 genes. Hence, the expression of the LOC101927745 transcript, regardless of its functionality, is indicative of the accessibility and activation of the NRIP1 enhancer GH21J015439. LOC101927745 and EVI1 transcript levels correlate significantly in EVI1high healthy and MDS CD34+ cells. Thus, NRIP1 transcriptional control through usage of the GH21J015439 enhancer might reflect how NRIP1 expression is physiologically regulated in immature hematopoietic cells and in other EVI1-expressing tissues such as fibroblasts. In more mature cells and in about 50% of NRIP1-expressing AML patients, NRIP1 transcription is activated via additional mechanisms that are not mediated through GH21J015439 and therefore do not cause LOC101927745 transcription (Figure 7C). Comparing survival of AML patients stratified according to their expression patterns of NRIP1, EVI1 and LOC101927745, we found that transcription of LOC101927745 was the most reliable prognostic factor among all groupings and that it was highly associated with a dismal overall outcome in the presence or absence of EVI1 transcription. Due to the high degree of functional redundancy in the regulatory landscape of physiologically relevant genes, deletion of a single enhancer site is not necessarily expected to have an impact on gene expression as this greatly depends on the presence of transcriptional regulators such as TF and histone modifiers. Nevertheless, we observed a modest but significant reduction of NRIP1 transcript levels after excising GH21J015439 in an NRIP1low-expressing AML cell line, which enhanced its response to differentiation-inducing drugs. However, removal of this enhancer site did not hinder the transcriptional upregulation of NRIP1 through ATRA-stimulated RA signaling (Figure 7D), which we have identified to serve as an independent mechanism that induces NRIP1 expression in human AML cells. Utilizing two different knockdown strategies in combination with ATRA-mediated induction of NRIP1 transcription, we observed that a forced downregulation of NRIP1 is harmful to t(3;3) EVI1high AML cells. Furthermore, NRIP1-knockdown rendered chromosome 3 rearranged cell lines vulnerable to ATRA treatment, resulting in decreased growth and induction of apoptosis. After 72 h of ATRA treatment, we detected an upregulation of EVI1 in t(3;3) and, surprisingly, in the chromosome 3 normal cell line OCI-AML3. Increased EVI1 expression also resulted in an upregulation of LOC101927745 RNA in t(3;3) cells which, according to our proposed model of NRIP1 regulation in EVI1high AML, likely reflected increased enhancer usage in an attempt to upregulate NRIP1. In line with the findings of a study by Nguyen et al.,22 reporting that ATRA enhances the oncogenic effects of EVI1, these observations

highlight the relevance of NRIP1 for EVI1-expressing AML cells and suggest that NRIP1 might contribute to resistance of EVI1high AML to RA agonists such as ATRA. In our study of the exact t(3;21)(q26;q11) breakpoints, as reported by Haferlach et al.8 and D'Angiò et al.,9 we discovered that instead of forming a novel fusion protein, this translocation results in a repositioning of the complete EVI1 open reading frame under the control of the NRIP1 gene promoters and upstream regulatory elements containing multiple RA responsive elements. In chromosome 21 intact cells, these elements mediate a strong induction of NRIP1 transcription upon stimulation with ATRA. Treating t(3;21)(q26;q11) cases with ATRA, as suggested in a recent clinical study,43 would therefore coordinate a similarly pronounced upregulation of the EVI1 oncogene in these special cases, presumably with devastating consequences. Another finding of our analyses of NRIP1 regulation is that the absence of RA receptor signaling - as in t(15;17) AML - abrogates NRIP1 expression. Transferring this knowledge to the aberrant NRIP1-abstracted upstream control of EVI1 that is unique to t(3;21)(q26;q11) AML would therefore theoretically open a therapeutic window for RA receptor antagonists, which might help to reduce or even abrogate expression of the EVI1 oncogene in these cases. Our data and those from Nguyen et al. as well as a recent clinical study do not convincingly show a benefit for adding ATRA to the treatment of EVI1-expressing AML patients.43 As knockdown of NRIP1 negatively affected the proliferation and survival of EVI1-expressing AML cells, our findings warrant further investigation of NRIP1 as a therapeutic target in myeloid diseases with EVI1 activation. Disclosures No conflicts of interest to disclose. Contributions SG performed experiments, conducted data analyses and wrote the manuscript. AC and LM established and performed the knockdown experiments and NRIP1 RNA and protein quantification. JI performed flow cytometry and western blots, and quantified relative band intensities. CR, BD, and NS helped to perform the CRISPR experiments and quantitative reverse transcriptase polymerase chain reaction measurements. FR, KD and LB provided clinical samples, clinical data, and input to the manuscript. JRP contributed the funding and access to facilities for RNA sequencing to generate the CK-AML dataset and revised the manuscript. TH gave advice on the study design and data analyses and revised the manuscript. AR and FK supervised the study and revised the manuscript. Acknowledgments The authors would like to thank Dr. Dirk Heckl for providing

Haematologica | 107 August 2022

1770

ARTICLE -NRIP1 in EVI1-expressing AML

S. Grasedieck et al.

the pL40C-CRISPR.EFS.PAC and pL-CRISPR.EFS.tRFP vectors was supported by the DFG (SFB 1074, project A5), the BC and Dr. Jan Krönke for providing the Cas9-expressing OCI- Cancer Foundation and the Leukemia & Lymphoma Society of AML-5 cell line. Canada. Funding SG is supported by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation – project 446251518), the Michael Smith Foundation for Health Research (MSFHR) and the Lotte & John Hecht Memorial Foundation (project RT2020-0578). AR was supported by the DFG (SFB 1074, project A5, and the gender equality program SFB 1074, project Z2). FK

Data-sharing statement All datasets that were analyzed for the current study are available in the NCBI’s GEO repository under the indicated identifiers, which are listed in the Methods section and cited throughout the main text and figure legends. RNA-sequencing data from the CK-AML cohort will be made available upon request. Please contact the corresponding author.

References 1. Gröschel S, Lugthart S, Schlenk RF, et al. High EVI1 expression predicts outcome in younger adult patients with acute myeloid leukemia and is associated with distinct cytogenetic abnormalities. J Clin Oncol. 2010;28(12):2101-2107. 2. Ayoub E, Wilson MP, McGrath KE, et al. EVI1 overexpression reprograms hematopoiesis via upregulation of Spi1 transcription. Nat Commun. 2018;9(1):4239. 3. Stein, S, Ott M, Schultze-Strasser S, et al. Genomic instability and myelodysplasia with monosomy 7 consequent to EVI1 activation after gene therapy for chronic granulomatous disease. Nat Med. 2010;16(2):198-204. 4. van Doorn SBW, Erpelinck CAJ, van Putten WLJ, et al. High EVI1 expression predicts poor survival in acute myeloid leukemia: a study of 319 de novo AML patients. Blood. 2003;101(3):837-845. 5. Lugthart S, van Drunen E, van Norden Y, et al, High EVI1 levels predict adverse outcome in acute myeloid leukemia: prevalence of EVI1 overexpression and chromosome 3q26 abnormalities underestimated. Blood. 2008;111(8):4329-4337. 6. Rockova V, Abbas S, Wouters BJ, et al. Risk stratification of intermediate-risk acute myeloid leukemia: integrative analysis of a multitude of gene mutation and gene expression markers. Blood. 2011;118(4):1069-1076. 7. Haas K, Kundi M, Sperr WR, et al. Expression and prognostic significance of different mRNA 5′-end variants of the oncogene EVI1 in 266 patients with de novo AML: EVI1 and MDS1/EVI1 overexpression both predict short remission duration. Genes Chromosomes Cancer. 2008;47(4):288-298. 8. Haferlach C, Bacher U, Grossmann, et al. Three novel cytogenetically cryptic EVI1 rearrangements associated with increased EVI1 expression and poor prognosis identified in 27 acute myeloid leukemia cases. Genes Chromosomes Cancer. 2012;51(12):1079-1085. 9. D'Angiò M, Fazio G, Grioni A, et al. High EVI1 expression due to NRIP1/EVI1 fusion in therapy-related acute myeloid leukemia: description of the first pediatric case. Hemasphere. 2020;17;4(5):e471. 10. Pombo A, Dillon N. Three-dimensional genome architecture: players and mechanisms. Nat Rev Mol Cell Biol. 2015;16(4):245-257. 11. L'Horset F, Dauvois S, Heery DM, et al. RIP-140 interacts with multiple nuclear receptors by means of two distinct sites. Mol Cell Biol. 1996;16(11):6029-6036. 12. Vivante A, Mann N, Yonath H, et al. A dominant mutation in nuclear receptor interacting protein 1 causes urinary tract malformations via dysregulation of retinoic acid signaling. J Am Soc Nephrol. 2017;28(8):2364-2376. 13. Cabezas-Wallscheid N, Buettner F, Sommerkamp P et al.

Vitamin A-retinoic acid signaling regulates hematopoietic stem cell dormancy. Cell. 2017;169(5):807-823. 14. Augereau P, Badia E, Carascossa S, et al. The nuclear receptor transcriptional coregulator RIP140. Nucl Recept Signal. 2006;4:e024. 15. The Cancer Genome Atlas Research Network. Genomic and epigenomic landscapes of adult de novo acute myeloid leukemia. N Engl J Med. 2013;368(22):2059-2074. 16. Tyner JW, Tognon CE, Bottomly D, et al. Functional genomic landscape of acute myeloid leukaemia. Nature. 2018;562(7728):526-531. 17. Martens JHA, Stunnenberg HG. BLUEPRINT: mapping human blood cell epigenomes. Haematologica. 2013;98(10):1487-1489. 18. Pellagatti A, Armstrong RN, Steeples V, et al. Impact of spliceosome mutations on RNA splicing in myelodysplasia: dysregulated genes/pathways and clinical associations. Blood. 2018;132(12):1225-1240. 19. Herold T, Jurinovic V, Batcha AMN, et al. A 29-gene and cytogenetic score for the prediction of resistance to induction treatment in acute myeloid leukemia. Haematologica. 2018;103(3):456-465. 20. Hirsch S, Blätte TJ, Grasedieck S, et al. Circular RNAs of the nucleophosmin (NPM1) gene in acute myeloid leukemia. Haematologica. 2017;102(12):2039-2047. 21. Saito Y, Nakahata S, Yamakawa, et al. CD52 as a molecular target for immunotherapy to treat acute myeloid leukemia with high EVI1 expression. Leukemia. 2011;25(6):921-931. 22. Nguyen CH, Bauer K, Hackl H, et al. All-trans retinoic acid enhances, and a pan-RAR antagonist counteracts, the stem cell promoting activity of EVI1 in acute myeloid leukemia. Cell Death Dis. 2019;10(12):944. 23. ENCODE Project Consortium. An integrated encyclopedia of DNA elements in the human genome. Nature. 2012;489(7414):57-74. 24. Gertz J, Savic D, Varley KE, et al. Distinct properties of celltype-specific and shared transcription factor binding sites. Mol Cell. 2013;52(1):25-36. 25. Rousseau M, Ferraiuolo MA, Crutchley JL, et al. Classifying leukemia types with chromatin conformation data. Genome Biol. 2014;15(4):R60. 26. Loke J, Assi SA, Imperato MR, et al. RUNX1-ETO and RUNX1-EVI1 differentially reprogram the chromatin landscape in t(8;21) and t(3;21) AML. Cell Rep. 2017;19(8):1654-1668. 27. Tang Z, Luo OJ, Li X, et al. CTCF-mediated human 3D genome architecture reveals chromatin topology for transcription. Cell. 2015;163(7):1611-1627. 28. Rao S, Huntley MH, Durand NC, et al. A 3D map of the human

Haematologica | 107 August 2022

1771

ARTICLE -NRIP1 in EVI1-expressing AML

S. Grasedieck et al.

genome at kilobase resolution reveals principles of chromatin looping. Cell. 2014;159(7):1665-1680. 29. Phanstiel DH, Van Bortle K, Spacek D, et al. Static and dynamic DNA loops form AP-1-bound activation hubs during macrophage development. Mol Cell. 2017;67(6):1037-1048. 30. Dixon JR, Jung I, Selvaraj S, et al. Chromatin architecture reorganization during stem cell differentiation. Nature. 2015;518(7539):331-336. 31. Rubin A, Barajas B, Furlan-Magaril M, et al. Lineage-specific dynamic and pre-established enhancer-promoter contacts cooperate in terminal differentiation. Nat Genet. 2017;49(10):1522-1528. 32. Kassambara A, Kosinski M. (2018). survminer: drawing survival curves using 'ggplot2'. R package version 0.4.2. https://CRAN.Rproject.org/package=survminer. 33. Therneau T (2015). A package for survival analysis in S. version 2.38,. https://CRAN.R-project.org/package=survival. 34. Hothorn T, Lausen B. On the exact distribution of maximally selected rank statistics. Comput Stat Data Anal. 2002;43(2):121-137. 35. Gröschel S, Sanders MA, Hoogenboezem R, et al. A single oncogenic enhancer rearrangement causes concomitant EVI1 and GATA2 deregulation in leukemia. Cell. 2014;157(2):369-381. 36. Yamazaki H, Suzuki M, Otsuki A, et al. A remote GATA2 hematopoietic enhancer drives leukemogenesis in

inv(3)(q21;q26) by activating EVI1 expression. Cancer Cell. 2014;25(4):415-427. 37. Rücker FG, Gong X, Dolnik A, et al. Identification of novel gene fusions in acute myeloid leukemia with complex karyotype by transcriptome analysis using RNA sequencing. Haematologica. 2017;102(s2):39-40. 38. Fishilevich S, Nudel R, Rappaport N, et al. GeneHancer: genomewide integration of enhancers and target genes in GeneCards. Database (Oxford). 2017;2017:bax028. 39. Kerley JS, Olsen SL, Freemantle SJ, et al. Transcriptional activation of the nuclear receptor corepressor RIP140 by retinoic acid: a potential negative-feedback regulatory mechanism. Biochem Biophys Res Commun. 2001;285(4):969-975. 40. Tohda S, Kurokawa H, Nara N. Relation of protein kinase A and protein kinase C to signaling pathways of hematopoietic factors in leukemia cell lines. Int J Oncol. 1996;8(3):521-524. 41. Huret JL. t(3;21)(q26;q22). Atlas Cytogenet Oncol Haematol. 2014;18(1):53-56. Online version: http://AtlasGeneticsOncology.org/Anomalies/t0321ID1009.html. 42. Schoenfelder S, Fraser P. Long-range enhancer–promoter contacts in gene expression control. Nat Rev Genet. 2019;20(8):437-455 43. Paubelle E, Plesa A, Hayette S, et al. Efficacy of all-transretinoic acid in high-risk acute myeloid leukemia with overexpression of EVI1. Oncol Ther. 2019;7(2):121-130.

Haematologica | 107 August 2022

1772

J. Ihlow et al.

ARTICLE - Acute Myeloid Leukemia

Acute myeloid leukemia: negative prognostic impact of early blast persistence can be in part overcome by a later remission prior to post-induction therapy Jana Ihlow,1,2 Sophia Gross,1 Leonie Busack,1 Anne Flörcken,1,3 Julia Jesse,1 Michaela Schwarz,1 Nina Rosa Neuendorff,1 Ann-Christin von Brünneck,2 Ioannis Anagnostopoulos,2 Seval Türkmen,3,4 Igor Wolfgang Blau,1,3 Thomas Burmeister,1,3 David Horst,2 Lars Bullinger1,3 and Jörg Westermann1,3 Department of Hematology, Oncology and Tumor Immunology, Campus Virchow Clinic, Charité–Universitätsmedizin Berlin, corporate member of Freie Universität Berlin and Humboldt-Universität zu Berlin; 2Institute of Pathology, Charité–Universitätsmedizin Berlin, corporate member of Freie Universität Berlin and Humboldt-Universität zu Berlin; 3Labor Berlin Charité Vivantes GmbH and 4Department of Medical Genetics and Human Genetics, Charité–Universitätsmedizin Berlin, corporate member of Freie Universität Berlin, Humboldt-Universität zu Berlin, Berlin, Germany 1

Correspondence: Jörg Westermann joerg.westermann@charite.de Received: May 2, 2021. Accepted: October 7, 2021. Prepublished: November 11, 2021 https://doi.org/10.3324/haematol.2021.279134 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

Abstract In acute myeloid leukemia, there is an ongoing debate on the prognostic value of the early bone marrow assessment in patients receiving intensive therapy. In this retrospective study, we analyzed the prognostic impact of the early response in 1,008 patients with newly diagnosed acute myeloid leukemia, who were treated at our institution with intensive chemotherapy followed by consolidation chemotherapy and/or allogeneic hematopoietic stem cell transplantation (HSCT). We found that early blast persistence has an independent negative prognostic impact on overall survival, eventfree survival and relapse-free survival. This negative prognostic impact may only be overcome in patients showing at least a partial remission at the early bone marrow assessment and who subsequently achieve blast clearance by additional induction chemotherapy prior to consolidation therapy with allogeneic HSCT. In accordance, we propose that the time slope of remission is an additional leukemia-related dynamic parameter that reflects chemosensitivity and thus may inform post-induction therapy decision-making. In addition to patient-related factors, European LeukemiaNet risk group, measurable residual disease monitoring and donor availability, this may particularly apply to European LeukemiaNet intermediate-risk patients, for whom a decision between consolidation chemotherapy and allogeneic HSCT remains challenging in many cases.

Introduction In acute myeloid leukemia (AML), an early bone marrow (BM) assessment is widely performed during the induction therapy to guide further decisions regarding therapy.1-4 However, a clear consensus concerning its prognostic impact on long-term survival and the optimal time point to perform an early BM assessment does not exist so far.1,3,5-11 While several studies negate the prognostic value of an early BM assessment,8,9,11 many others discuss early blast clearance as a favorable prognostic parameter with regard to both remission rates and longterm survival.1,12-14 Conversely, early blast persistence has been linked to an unfavorable outcome in patients treated intensively for AML.10,12,13,15 However, it is still unclear whether a potential negative prognostic impact

of early blast persistence can be overcome during subsequent therapy of AML. The present study was conducted in a large cohort of intensively treated AML patients (n=1,008) with the aims of (i) analyzing the prognostic impact of early blast clearance on overall survival (OS), event-free survival (EFS) and relapse-free survival (RFS), and (ii) of evaluating the longterm prognosis in patients with early blast persistence.

Methods Clinical characteristics, treatment and endpoints We have treated 1,340 patients aged ≥18 years with newly diagnosed AML at our clinic within the past two decades (January 1st, 2000 - December 31st, 2018). After application

Haematologica | 107 August 2022

1773

ARTICLE - Early blast persistence in AML

J. Ihlow et al.

of the exclusion criteria (Figure 1), 1,008 patients were eligible for this retrospective and non-interventional study, which is in line with local ethical guidelines and the Declaration of Helsinki and was approved by the local ethics committee (EA1/038/21). Standard first-line induction therapy consisted of cytarabine-/daunorubicin-based chemotherapy according to the “7+3” regimen. Some patients received comparable induction therapy with Idarubicin, cytarabine and etoposide (ICE), thioguanine, cytarabine and daunorubicin (TAD9) or high-dose cytarabine and mitoxantrone (HAM) within particular clinical trial protocols. Targeted therapies such as midostaurin or gemtuzumab-ozogamicin were applied in some patients in addition to “7+3”-based regimens within clinical trials. Consolidation chemotherapy was performed with intermediate- to high-dose cytarabine-based therapy with or without mitoxantrone or TAD9 in particular clinical trials (more details concerning chemotherapy are provided in the Online Supplementary Methods). Allogeneic hematopoietic stem cell transplantation (HSCT) following either myeloablative (MAC) or reduced-intensity conditioning (RIC) regimens was used as consolidation therapy in first remission or in relapsed/refractory patients (further details are given in the Online Supplementary Methods). BM assessment was performed at baseline, on day 14-21 of the first induction cycle, and prior to post-induction

therapy. BM assessment was performed by both morphology (cytology and/or histopathology) and multiparametric flow cytometry (see Online Supplementary Methods). The 2010 European LeukemiaNet (ELN) classification was applied for the assessment of the remission status (see Online Supplementary Methods).16 For this analysis, combined remission was defined as a combination of complete remission plus complete remission with incomplete hematologic recovery plus morphological leukemia-free state (MLFS). Early partial remission (PR) was defined by a decrease of bone marrow blasts by at least 50% to a blast percentage in the range of 5%-25%. Cytogenetic and molecular risk was defined using the ELN risk stratification of 2010 (due to a lack of some molecular data that are mandatory for the 2017 ELN risk classification).16 The patients' general condition was measured by the Eastern Cooperative Oncology Group (ECOG) performance score.17 Comorbidity was assessed using the Charlson Comorbidity Index.18 OS, EFS, RFS, risk of relapse and non-relapse mortality were defined as clinical endpoints by applying the Cheson criteria and the response criteria of the European Society for Blood and Bone Marrow Transplantation.19,20 Statistical analysis Data were curated and retrospectively analyzed using SPSS 23.0 software (IBM® , 2015, Armonk, NY, USA).

Figure 1. Study design and clinical endpoints. AML: acute myeloid leukemia; APL; acute promyelocytic leukemia; OS: overall survival; EFS: event-free survival; RFS: relapse-free survival; CS-HR: cause-specific hazard ratio: RR: risk of relapse; NRM: nonrelapse mortality. Haematologica | 107 August 2022

1774

ARTICLE - Early blast persistence in AML

J. Ihlow et al.

Baseline characteristics were analyzed using the Kruskal-Wallis-H test and the c2 test followed by post-hoc testing and Bonferroni adjustment. The median followup was estimated by the reverse Kaplan-Meier method. Survival was analyzed by the Kaplan-Meier method. The log-rank test was used to detect survival differences between groups. Subsequently, univariate and multivariate Cox regression models, which included factors with a significance level of P≤0.1, were applied. In order to define a hazard ratio (HR) and a cause-specific hazard ratio (CS-HR), the variables were transformed into categorical dichotomous data. To estimate the relapse

risk and non-relapse mortality in patients with blast clearance, a multivariate cause-specific Cox proportional hazards model that included confounding factors with a significant impact on relapse and survival was used based on an etiological approach. Within this model, death and relapse were defined as competing events and hence treated as censored observations.21 Post-hoc survival analysis was conducted using the Benjamini-Hochberg procedure. A P<0.05 was considered statistically significant. For graphical presentation, Graph Pad Prism 8 (GraphPad Software.Inc) was applied.

Table 1. Baseline characteristics and therapeutic approach in 1,008 acute myeloid leukemia patients with regard to remission status at interim bone marrow assessment. Early blast clearance

Early PR

Early resistant AML

572 (57)

196 (19)

240 (24)

ECOG, median (IQR)

1 (0-1)

0 (0-1)

0.144

CCI, median (IQR)

0 (0-1)

0.237

56 (46 - 64)

54 (43 - 61)

57 (45 - 65)

0.060

Characteristics N (% entire cohort)

Age (years), median (IQR) Subtype of AML

P-value

0.019

de novo AML, N (% remission subgroup)

387 (68)

136 (69)

137 (57)

sAML, N (% of remission subgroup)

115 (20)

42 (22)

74 (31)

tAML, N (% of remission subgroup)

67 (11)

16(8)

26 (11)

3 (1)

2 (1)

3 (1)

unknown, N (% of remission subgroup) ELN 2010 risk group

< 0.001

favorable, N (% of remission subgroup)

97 (17)

19 (10)

8 (3)

intermediate I/II, N (% of remission subgroup)

280 (49)

96 (49)

123 (51)

adverse, N (% of remission subgroup)

121 (21)

67 (34)

90 (38)

unknown, N (% of remission subgroup)

74 (13)

14 (7)

19 (8)

225 (39)

86 (44)

90 (38)

0.423*

double induction prior to allo-HSCT, N (% allo-HSCT 1st CR/CRi/MLFS within remission subgroup)

120 (53)

79 (92)

85 (94)

0.002*

consolidation chemotherapy prior to allo-HSCT N (% allo-HSCT 1st CR/CRi/MLFS within remission subgroup)

158 (70)

53 (62)

49 (54)

0.230*

Allo-HSCT in 2nd remission or as salvage therapy, N (% of remission subgroup)

156 (27)

63 (32)

67 (28)

0.317*

Consolidation chemotherapy without allo-HSCT, N (% of remission subgroup)

191 (33)

47 (24)

83 (35)

0.070*

double induction prior to scheduled consolidation chemotherapy, N (% of non-allo-HSCT within remission subgroup)

87 (46)

38 (81)

59 (71)

0.021*

CR/CRi/MLFS after double induction, N (% of all nonallo-HSCT patients with double induction)

69 (79)

29 (76)

29 (49)

0.045*

Allo-HSCT in 1st remission, N (% of remission subgroup)

*Significance level adjusted with the Bonferroni correction P=0.008. PR: partial remission; AML: acute myeloid leukemia; n: number of patients; IQR: interquartile range; ECOG: Eastern Cooperative Oncology Group performance status; CCI: Charlson Comorbidity Index; sAML: secondary acute myeloid leukemia; tAML: therapy-related acute myeloid leukemia; ELN: European LeukemiaNet; allo-HSCT: allogeneic hematopoietic stem cell transplantation; CR: complete remission; CRi: complete remission with incomplete hematologic recovery; MLFS: morphological leukemia-free state. Haematologica | 107 August 2022

1775

ARTICLE - Early blast persistence in AML

J. Ihlow et al.

Results Clinical characteristics A total of 1,008 patients, who had undergone intensive therapy, were eligible for this analysis. The median followup was 63.1 months (95% confidence interval [95% CI]: 55.3-71.0 months). Fifty-seven percent of the entire cohort showed early blast clearance (n=572), whereas 43% had blast persistence (n=436). Within the latter group, 45% (196/436) had an early PR and 55% (240/436) showed early resistant disease without any response. The distribution of baseline characteristics within the entire cohort and the three “remission groups” (early blast clearance, early PR and early resistant AML) are shown in Table 1. As expected, ELN risk stratification (P<0.001) and subtype of AML differed significantly between the three groups (P=0.019). However, there were no further significant differences with regard to baseline characteristics. The further treatment of patients with early blast persistence beyond induction 1 is outlined in Table 1 and Figure 2. Early blast clearance and early blast persistence are prognostic in the entire cohort The entire cohort had a 5-year OS of 35% with a median OS of 28.5 months (95% CI: 24.4 - 32.6 months). The evaluation of early BM results revealed a significant decrease in OS in patients with early blast persistence as compared to those with blast clearance (P<0.001) (Table

2). The 5-year OS of patients with early blast clearance was 41% as compared to 30% for those with early blast persistence (P<0.001). This observation maintained its significance within a multivariate model (HR=1.4, P<0.001) (Table 2) that included all factors with a significant impact on survival within the univariate analysis (ECOG status >1: P=0.004; Charlson Comorbidity Index ≥2: P=0.003; ELN risk group intermediate/adverse: P<0.001 (including FLT3ITD mutational status); age ≥60 years: P<0.001; and subtype of AML: P<0.001). In the entire cohort, 5-year EFS and RFS were 24% and 25%. The median EFS and RFS were 13.9 months (95% CI: 12.4-15.5 months) and 13.9 months (95% CI: 11.9-15.9 months), respectively. The negative prognostic impact of early blast persistence also translated into an effect on EFS. Early blast clearance was associated with a 5-year EFS of 26% as compared to 18% in patients with early blast persistence (P=0.001) (Table 2). This significant difference was also maintained within the multivariate analysis (HR=1.3, P=0.001). Comparable results were observed for RFS with a hazard ratio of 1.2 in the multivariate analysis in the presence of the other biologically relevant risk factors that are mentioned above (P=0.031) (Table 2). The negative prognostic impact of early blast persistence can be overcome if a response is achieved prior to post-induction therapy In the entire cohort, the combined remission rate was 81%

Figure 2. Further treatment in 436 acute myeloid leukemia patients with early blast persistence. For further details regarding chemotherapy, see the Online Supplementary Methods. AML: acute myeloid leukemia; PR: partial remission; n: number of patients; “7+3”: cytarabine-/daunorubicin-based chemotherapy according to the “7+3” regimen; HAM: high-dose cytarabine and mitoxantrone; IdaFLAG: idarubicin, fludarabine, cytarabine; granulocyte colony-stimulating factor; MitoFLAG: mitoxantrone, fludarabine, cytarabine; granulocyte colony-stimulating factor; allo-HSCT:, allogeneic hematopoietic stem cell transplantation. Haematologica | 107 August 2022

1776

ARTICLE - Early blast persistence in AML

J. Ihlow et al.

(813/1,008) prior to scheduled post-induction therapy. The combined remission rate was 89% (508/572) in patients with early blast clearance and 70% (305/436) in patients with early blast persistence after additional therapy (P<0.001). The negative prognostic impact of early blast persistence was maintained in patients who achieved blast clearance during further induction therapy. The 5year OS and RFS were 43% and 26% in patients with an early blast clearance as compared to 31% and 23% in patients with early blast persistence who achieved a remission prior to post-induction therapy (P=0.016 and P=0.013) (Table 2). The negative prognostic impact of early blast persistence was also maintained in the multivariate model that included relevant risk factors for OS (HR=1.3, P=0.024) and RFS (HR=1.4, P=0.002) (Table 2). Furthermore, patients with early blast persistence showed an increased risk of relapse in the cause-specific hazard model which included the same covariates (CS-HR=1.3,

P=0.039) (Table 2). Moreover, in patients with early blast persistence, there was a strong trend towards a higher risk of non-relapse mortality, even in the presence of other risk factors (CS-HR=1.4, P=0.069) (Table 2). Interestingly, in the group with early blast persistence, survival was very heterogeneous depending on whether the patients had at least an early PR or showed early resistant disease (Figure 3). The survival of patients with early PR and subsequent combined remission prior to consolidation therapy was comparable to the survival of patients with early blast clearance (5-year OS: 45% vs. 44%, P=0.618), whereas early resistant AML maintained its negative prognostic impact throughout the analysis (5year OS: 28%, P<0.001). Comparable results were observed for RFS (Figure 3). Similarly, early resistant AML (but not early PR) remained prognostically unfavorable in the multivariate analysis for both OS (HR=1.5, 95% CI: 1.22.0; P=0.001) and RFS (HR=1.4, 95% CI: 1.1 - 1.7; P=0.012).

Table 2. Impact of early blast persistence on survival in the entire cohort and in patients with combined remission prior to consolidation therapy. Early blast clearance

Early blast persistence (PR & resistant AML)

572 (57)

436 (43)

OS (months), median (95% CI)

35.8 (29.4-42.3)

18.0 (14.2-21.8)

< 0.001

RFS (months), median (95% CI)

15.7 (13.2-18.2)

11.5 (9.1-13.9)

0.100

EFS (months), median (95% CI)

16.2 (13.8-18.6)

11.3 (10.0-12.6)

0.001

Survival Entire cohort N (% entire cohort)

P-value

MV-HR* for OS with early blast persistence

1.42 (1.18-1.71)

< 0.001

MV HR* for RFS with early blast persistence

1.22 (1.02-1.50)

0.031

MV-HR* for EFS with early blast persistence

1.34 (1.13-1.59)

0.001

Combined remission prior to post-induction therapy N (% of group with post-induction 508 (62) combined remission) OS (months), median (95% CI) 42.2 (31.8-52.6)

29.3 (20.7-37.8)

0.016

RFS (months), median (95% CI)

13.0 (10.5-15.6)

0.013

305 (38)

18.8 (15.8-21.9-37.8)

MV-HR* for OS with early blast persistence MV-HR* for RFS with early blast persistence MV CS-HR* for NRM with early blast persistence CS-HR* for relapse with early blast persistence

1.29 (1.03-1.60)

0.024

1.35 (1.11-1.65)

0.002

1.42 (0.97-2.07)

0.069

1.30 (1.01-1.62)

0.039

*The multivariate analysis included the following dichotomized parameters: Eastern Cooperative Oncology Group score ≤1 vs. >1, Charlson Comorbidity Index <2 vs. ≥2, European LeukemiaNet risk group favorable vs. intermediate/adverse, age <60 years vs. ≥60 years, subtype of acute myeloid leukemia de novo vs. secondary/therapy-related. CR: complete remission; CRi: complete remission with incomplete hematologic recovery; MLFS: morphological leukemia-free state; n: number of patients; PR: partial remission; AML: acute myeloid leukemia; 95% CI: 95% confidence interval; OS: overall survival; EFS: event-free survival; RFS: relapse-free survival; MV: multivariate; HR: hazard ratio; NRM: non-relapse mortality; CS-HR: cause-specific hazard ratio.

Haematologica | 107 August 2022

1777

Haematologica | 107 August 2022

1778

Figure 3. Entire cohort. Survival outcomes after achievement of combined remission prior to post-induction therapy (allogeneic transplantation or consolidation chemotherapy) with regard to early remission status at the interim bone marrow assessment. (A) Overall survival (OS) and relapse-free survival (RFS) with early blast clearance versus early partial remission. (B) OS and RFS with early blast clearance versus early resistant acute myeloid leukemia. (C) OS and RFS with early partial remission versus early resistant acute myeloid leukemia. AML: acute myeloid leukemia; PR: partial remission; OS: overall survival; RFS: relapse-free survival; n: number of patients; 95% CI: 95% confidence interval.

ARTICLE - Early blast persistence in AML J. Ihlow et al.

ARTICLE - Early blast persistence in AML

J. Ihlow et al.

Prognostic impact of early blast clearance in patients who underwent allogeneic hematopoietic stem cell transplantation as post-remission therapy Of all patients, 68% (687/1,008) underwent allogeneic HSCT. Among these transplanted patients, 5-year OS and RFS were 44% and 28%, and median OS and RFS were 41.8 months (95% CI: 33.1-50.3 months) and 16.5 months (95% CI: 14.1-19.0 months), respectively. Of these patients, 58% (401/687) underwent allogeneic HSCT as consolidation therapy in first remission, whereas the remaining 42% (286/687) received their transplant beyond first remission (Table 1). Patients who underwent allogeneic HSCT in first remission (n=401) had a 5-year OS of 51% and a 5-year RFS of 47% with a median OS of 62.1 months (95% CI: 33.1-91.0 months) and a median RFS of 38.6 months (95% CI: 17.359.9 months). Regarding OS and RFS, there was no significant difference between patients with early blast clearance and early PR in this particular subset of patients (Figure 4A). In contrast, patients with early resistant AML showed both inferior RFS and OS, even after having achieved combined remission prior to allogeneic HSCT (Figure 4B). Precisely, 5-year OS was 57% with early blast clearance, and 46% with early PR (P=0.267) as compared to 37% with early resistant AML (Figure 4B, C, P=0.002). The 5-year RFS was 51% with early blast clearance, 42% with early PR (P=0.333) and 32% with early resistant AML (P=0.001). Considering the multivariate analysis, early resistant AML was an unfavorable prognostic factor in patients who underwent allogeneic HSCT as consolidation therapy (OS: P=0.011, HR=1.6, 95% CI: 1.1-2.4; RFS: P=0.006, HR=1.7, 95% CI: 1.2-2.4). In patients who had been transplanted in first remission, we further analyzed the influence of both the type of transplant conditioning (MAC vs. RIC) and the type of donor (matched sibling donor [MSD] vs. matched unrelated donor [MUD]) on the clinical outcome. In fact, 136/401 patients received MAC, whereas 261/401 patients were treated with RIC (information on the conditioning regimen was not available in 4 cases). Comparing MAC and RIC, we found a significant difference in median OS (61.2 months vs. 46.8 months, P=0.012) and RFS (58.4 months vs. 33.5 months, P=0.013). However, this finding was mainly attributable to major differences in median age (39 years vs. 58 years, P<0.001) and Charlson Comorbidity Index (upper quartile 0 points vs. 1 point, P<0.001) between both groups. In contrast, ELN subgroups (P=0.182) and median ECOG status (P=0.866) did not differ significantly between patients given MAC or RIC. More importantly, there was no significant difference in early remission status between patients given MAC or RIC since early blast clearance and blast persistence were similarly distributed between these subgroups (early blast clearance 51% with MAC vs. 58% with RIC, early blast persist-

ence 49% with MAC vs. 42 % with RIC, P=0.153) and, vice versa, MAC and RIC application were equally distributed within the remission subgroups. Regarding the type of donor (MSD vs. MUD), there was a trend towards better OS and RFS in patients with MSD (OS 59.4 months vs. 47.0 months, P=0.058; RFS: 51.3 months vs. 34.3 months, P=0.091) However, this seemed to be caused again by differences in median age (50 vs. 53 years, interquartile range 37-58 vs. 43-62 years, P=0.003) or HLA-mismatch (full match vs. mismatch, P<0.001). However, early blast clearance and blast persistence were equally distributed within the subgroups with MSD and MUD (early blast clearance: 56% with MSD and MUD, early blast persistence: 44% with MSD and MUD, P=0.968) and vice versa. In conclusion, early resistant AML remained an independent unfavorable prognostic factor in the multivariate analysis of patients with allogeneic HSCT as consolidation therapy (OS: P=0.011, HR=1.6, 95% CI: 1.1-2.4; RFS: P=0.006, HR=1.7, 95% CI: 1.2-2.4). Prognostic impact of early blast persistence in patients who received chemotherapy as post-remission therapy In patients who did not undergo allogeneic HSCT (321/1,008), 5-year OS and RFS were 20% and 17% with a median OS of 12.0 months (95% CI: 9.9-14.1 months) and a median RFS of 8.4 months (95% CI: 6.8-9.9 months). Fifty-one percent (165/321) of the non-transplanted patients had achieved blast clearance prior to consolidation chemotherapy. Of these latter patients, 24% (39/165) had been treated with one cycle of induction chemotherapy and 76% (126/165) had received two cycles of induction therapy. Within the latter subgroup of non-transplanted patients who had received two cycles of induction therapy, 5-year OS and RFS were 32% and 23% with a median OS and RFS of 17.8 months (95% CI: 9.5-26.1 months) and 9.7 months (95% CI: 6.9-12.5 months), respectively. In these patients, early blast clearance was comparable to early PR with regard to OS (5-year OS 40% vs. 32%, P=0.401) (Figure 5A). In contrast, RFS with early PR was significantly worse than with early blast clearance in the univariate analysis (5-year RFS 32% vs. 15%, P=0.037) (Figure 5A), and there was a clear trend towards inferior survival in the multivariate model (HR=1.6, 95% CI: 1.0-2.7, P=0.058), suggesting an adverse prognostic impact of early PR on RFS which was most likely compensated by subsequent salvage therapy with regard to OS.

Discussion Whether an early remission during AML induction therapy is of any prognostic value has remained a matter of debate over the past decade. Even in the era of minimal/ measurable residual disease (MRD)-guided therapeutic decision-

Haematologica | 107 August 2022

1779

Haematologica | 107 August 2022

1780

Figure 4. Allogeneic hematopoietic stem cell transplant cohort in first remission. Survival outcomes after achievement of a combined remission prior to allogeneic hematopoietic stem cell transplantation with regard to early remission status at the interim bone marrow assessment. (A) Overall survival (OS) and relapse-free survival (RFS) with early blast clearance versus early partial remission (B) OS and RFS with early blast clearance versus early resistant acute myeloid leukemia. (C) OS and RFS with early partial remission versus early resistant acute myeloid leukemia. AML: acute myeloid leukemia; PR: partial remission; RD: refractory disease; OS: overall survival; RFS: relapse-free survival; n: number of patients; 95% CI: 95% confidence interval; allo-HSCT: allogeneic hematopoietic stem cell transplantation.

ARTICLE - Early blast persistence in AML J. Ihlow et al.

Haematologica | 107 August 2022

1781

Figure 5. Non-transplanted cohort. Surival after achievement of combined remission after double induction and prior to scheduled consolidation chemotherapy with regard to early remission status at the interim bone marrow assessment. (A) Overall survival (OS) and relapse-free survival (RFS) with early blast clearance versus early partial remission. (B) OS and RFS with early blast clearance versus early resistant acute myeloid leukemia. (C) OS and RFS with early partial remission versus early resistant acute myeloid leukemia. AML: acute myeloid leukemia; PR: partial remission; OS: overall survival; RFS: relapse-free survival; n: number of patients; 95% CI: 95% confidence interval; allo-HSCT: allogeneic hematopoietic stem cell transplantation.

ARTICLE - Early blast persistence in AML J. Ihlow et al.

ARTICLE - Early blast persistence in AML

J. Ihlow et al.

Figure 6. Potential treatment algorithm in European LeukemiaNet intermediate-risk acute myeloid leukemia after implementation of the early response as an additional prognostic parameter (in a notional scenario). Clinical responses and therapy decisions in 499 patients with intermediate-risk acute myeloid leukemia (according to the European LeukemiaNet classification) with and without implementation of the early bone marrow assessment into further therapeutic decision making. In our cohort, 153/343 patients with at least an early partial remission underwent allogeneic hematopoietic stem cell transplantation (HSCT) and 132/343 were treated with consolidation chemotherapy. In these groups, the implementation of the early response as a prognostic parameter (in addition to minimal residual disease, which was not available in our cohort) would have possibly led to consolidation chemotherapy instead of allogeneic transplantation in 153/343 patients, if they had achieved minimal residual disease negativity. Seventy-seven of 420 patients had early resistant disease. Within this subgroup, 45/77 underwent allogeneic HSCT and 14/77 consolidation chemotherapy. The implementation of the early response as a prognostic parameter (in addition to minimal residual disease) would have possibly led to allogeneic HSCT as consolidation therapy in an additional 14/77 patients. Thus, in our cohort, the implementation of early response would have possibly changed treatment decision in 33% of ELN intermediate-risk patients and in 17% of all patients. ELN: European LeukemiaNet; n: number of patients; PR: partial remission; allo-HSCT: allogeneic hematopoietic stem cell transplantation.

making, this controversy has not been resolved, since early blast clearance can indicate a therapy response at a very early time point when MRD assessment is not yet part of the routine management. Furthermore, there are also patients in whom an adequate MRD marker cannot be established. In these cases, early BM assessment may inform therapeutic decision-making, particularly for those in whom the choice of consolidation therapy (i.e., conventional chemotherapy vs. allogeneic HSCT) is challenging. Over the past decade, there has been an extensive discussion not only on the general value, but also on the most appropriate time point, of the early BM assessment. Recommendations vary from omitting early BM assessment completely (due to a lack of prognostic information) to its implementation during induction therapy between day 6 and day 21.3,5-9 At our institution, the early BM assessment was generally performed between day 14 and 21, as previous studies had shown that there is no substantial difference between BM evaluation on day 14 and 21 and, thus, results obtained within this interval were merged.3,7 Certainly, there is some heterogeneity with re-

gard to induction and consolidation therapy within our cohort of AML patients. However, the different treatment protocols were prospectively compared within the German Intergroup trial and no relevant outcome differences were detected22 and thus they should be comparable with regard to long-term survival. Our large cohort of 1,008 intensively treated patients with newly diagnosed AML does now confirm a negative prognostic impact of early blast persistence on both OS and RFS. While in our cohort survival was slightly above the upper range of international studies,23,24 which might be explained by the large number of patients who underwent allogeneic HSCT, the survival in the transplanted cohort was in the range of other studies in AML.25-28 Thus, the favorable impact of early blast clearance observed in our cohort is in line with previously published data,1,12 and it seems conceivable that this effect is due to chemosensitivity of AML cells in vivo. Vice versa, the negative impact of early persistent AML most likely reflects resistance to conventional chemotherapy.13,15 This assumption is emphasized by the comparison of cause-

Haematologica | 107 August 2022

1782

ARTICLE - Early blast persistence in AML

J. Ihlow et al.

specific hazards for relapse and non-relapse mortality in our cohort. When comparing patients with early blast persistence to patients with early blast clearance, we observed a significant increase in the risk of relapse, but not in the risk of non-relapse mortality. This finding strongly suggests that the unfavorable prognostic impact of early blast persistence is mainly driven by disease relapse and less by toxicity caused by additional therapy. Interestingly, the achievement of a combined remission prior to post-induction therapy outperforms the negative impact of early blast persistence on OS, if the early BM assessment shows at least a PR. In contrast, the negative prognostic impact of early resistant AML cannot be completely overcome, even if a combined remission is obtained prior to consolidation therapy. Notably, in patients with a later combined remission during induction, the poor prognostic impact of early blast persistence can be in part compensated by consolidation with allogeneic HSCT. This is possibly due to an additional immunological graft-versus-leukemia-effect that may compensate for a lower extent of chemosensitivity in these cases.29-32 This hypothesis is underlined by the observation that the adverse prognostic impact of early blast persistence cannot be overcome in AML patients who do not proceed to allogeneic HSCT in first remission after induction therapy. Notably, early blast persistence only translates into inferior RFS in this subgroup, whereas OS is not significantly different, most likely because of subsequent salvage therapy. There is general consensus on the key prognostic value of MRD both under intensive chemotherapy and in the setting of allogeneic HSCT.33-37 In addition, our study suggests that it is not only the static remission status at a particular time point after therapy, but also the time slope of remission which is of prognostic value. This seems reasonable since a rapid initial response (i.e., with early blast clearance on day 14-21) is a surrogate for chemo-responsiveness, whereas the need for a second induction cycle due to early blast persistence indicates at least some degree of chemoresistance. In this regard, it is intriguing that an early PR has a negative impact on RFS in non-transplant patients, whereas this is not observed in patients undergoing allogeneic HSCT. Therefore, early blast clearance from the BM can be considered a dynamic and easily attainable parameter indicative of therapeutic response in addition to the ELN risk stratification and MRD monitoring. This might primarily apply to patients within the intermediate ELN risk group (Figure 6) due to its biological heterogeneity and the challenging choice of the most appropriate consolidation treatment for individual patients in this subgroup. Potential risks and benefits of allogeneic HSCT need to be considered, such as the risk of relapse versus non-relpase mortality or treatment-related mortality.27,35,38-50 Interestingly, Venditti et al. recently showed in

patients with intermediate-risk AML, according to the National Comprehensive Cancer Network (NCCN) classification, that consolidation with allogeneic HSCT in MRD-positive patients is practically equivalent to consolidation with autologous HSCT in patients who are MRDnegative. Using this strategy, the investigators found that disease-free survival in the group of intermediate-risk MRD-negative AML patients was comparable to that of NCCN favorable-risk AML patients.50 The important role of MRD stratification in intermediate-risk AML patients was also observed in a survey of the NCRI-AML17 trial.48 Thus, early treatment response (besides general health condition, MRD levels and donor availability) might serve as an additional parameter for risk stratification in intermediate-risk AML. In our cohort, such an early response assessment would have possibly influenced the post-induction therapy decision in approximately 33% of all ELN intermediate-risk patients (Figure 6). Therefore, the adverse impact of early blast persistence might be overcome in the subgroup with at least an early PR by additional induction therapy resulting in subsequent remission prior to post-induction therapy that should comprise allogeneic HSCT consolidation. Unfortunately, MRD assessment by molecular methods and/or highly sensitive multicolor flow cytometry, particularly within the ELN intermediate-risk group, could not be included in our analysis, since these data were not available for all patients throughout the study period. In the future, additional evidence provided by large datasets from molecularly characterized AML cohorts consolidated with either intensive chemotherapy or allogeneic HSCT might pave the way for such a strategy which would then also need further prospective evaluation. Disclosures No conflicts of interest to disclose. Contributions JI and JW designed the study. AF, MS, NRN, AB, IA, ST, IWB, TB, DH, LAB and JW performed the clinical and diagnostic workup required for this study. JI, SG, LEB and JW collected, analyzed, and interpreted the data. JI and JW wrote the manuscript draft. All authors critically revised the manuscript and approved the final version. Acknowledgments We thank Margrit Stodder, Alma Herneth, Kacper Adamiak and Annabel Sick for their administrative support. Furthermore, we owe thanks to Sven Bischoff for his excellent support with the statistical analysis. Data-sharing statement Data for this study are not publicly available due to ethical restrictions.

Haematologica | 107 August 2022

1783

ARTICLE - Early blast persistence in AML

J. Ihlow et al.

References 1. Campuzano-Zuluaga G, Deutsch Y, Salzberg M, et al. Routine interim disease assessment in patients undergoing induction chemotherapy for acute myeloid leukemia: can we do better? Am J Hematol. 2016;91(3):277-282. 2. 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. 3. Yanada M, Borthakur G, Ravandi F, Bueso-Ramos C, Kantarjian H, Estey E. Kinetics of bone marrow blasts during induction and achievement of complete remission in acute myeloid leukemia. Haematologica. 2008;93(8):1263-1265. 4. National Comprehensive Cancer Network Clinical Practice Guidelines in Oncology (NCCN Guidelines): Acute Myeloid Leukemia. Version 1 (2017), available at: https://www.nccn.org/professionals/physician_gls/pdf/aml.pdf., date accessed: 03.09.2021 5. Percival M-E, Lai C, Estey E, Hourigan CS. Bone marrow evaluation for diagnosis and monitoring of acute myeloid leukemia. Blood Rev. 2017;31(4):185-192. 6. Griffin PT, Komrokji RS, Sweet K, et al. Bone marrow cellularity at day 14 is the most important predictive factor for response in patients with AML who require double-induction chemotherapy: analysis from a large, single institution experience. Am J Hematol. 2017;92(3):232-237. 7. Pullarkat V, Aldoss I. Prognostic and therapeutic implications of early treatment response assessment in acute myeloid leukemia. Crit Rev Oncol Hematol. 2015;95(1):38-45. 8. Ofran Y, Leiba R, Ganzel C, et al. Prospective comparison of early bone marrow evaluation on day 5 versus day 14 of the “3 + 7” induction regimen for acute myeloid leukemia. Am J Hematol. 2015;90(12):1159-1164. 9. Alsaleh K, Aleem A, Almomen A, Anjum F, Alotaibi GS. Impact of day 14 bone marrow biopsy on re-induction decisions and prediction of a complete response in acute myeloid leukemia cases. Asian Pac J Cancer Prev. 2018;19(2):421-425. 10. Rowe JM, Kim HT, Cassileth PA, et al. Adult patients with acute myeloid leukemia who achieve complete remission after 1 or 2 cycles of induction have a similar prognosis: a report on 1980 patients registered to 6 studies conducted by the Eastern Cooperative Oncology Group. Cancer. 2010;116(21):5012-5021. 11. Terry CM, Shallis RM, Estey E, Lim SH. Day 14 bone marrow examination in the management of acute myeloid leukemia. Am J Hematol. 2017;92(10):1079-1084. 12. Kern W, Haferlach T, Schoch C, et al. Early blast clearance by remission induction therapy is a major independent prognostic factor for both achievement of complete remission and longterm outcome in acute myeloid leukemia: data from the German AML Cooperative Group (AMLCG) 1992 Trial. Blood. 2003;101(1):64-70. 13. Hussein K, Jahagirdar B, Gupta P, Burns L, Larsen K, Weisdorf D. Day 14 bone marrow biopsy in predicting complete remission and survival in acute myeloid leukemia. Am J Hematol. 2008;83(6):446-450. 14. Estey E Shen Y, Thall PF. Effect of time to complete remission on subsequent survival and disease-free survival time in AML, RAEB-t, and RAEB. Blood. 2000;95(1):72-77. 15. Bertoli S, Bories P, Béné MC, et al. Prognostic impact of day 15 blast clearance in risk-adapted remission induction chemotherapy for younger patients with acute myeloid leukemia: long-term results of the multicenter prospective LAM-2001 trial by the GOELAMS study group. Haematologica.

2014;99(1):46-53. 16. Döhner H, Estey EH, Amadori S, et al. Diagnosis and management of acute myeloid leukemia in adults: recommendations from an international expert panel, on behalf of the European LeukemiaNet. Blood. 2010;115(3):453-474. 17. Oken MM, Creech RH, Tormey DC, et al. Toxicity and response criteria of the Eastern Cooperative Oncology Group. Am J Clin Oncol. 1982;5(6):649-656. 18. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chron Dis. 1987;40(5):373-383. 19. 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. 20. Iacobelli S, on behalf of the ESC. Suggestions on the use of statistical methodologies in studies of the European Group for Blood and Marrow Transplant. Bone Marrow Transplant. 2013;48(1):S1-S37. 21. Andersen PK, Geskus RB, de Witte T, Putter H. Competing risks in epidemiology: possibilities and pitfalls. Intern J Epidemiol. 2012;41(3):861-870. 22. Büchner T, Schlenk RF, Schaich M, et al. Acute myeloid leukemia (AML): different treatment strategies versus a common standard arm--combined prospective analysis by the German AML Intergroup. J Clin Oncol. 2012;30(29):3604-3610. 23. Howlader N NA, Krapcho M, Miller D, et al. SEER Cancer Statistics Review, 1975-2013, National Cancer Institute. Bethesda, MD, available from: http://seer.cancer.gov/csr/1975_2013/, based on November 2015 SEER data submission, posted to the SEER web site, 04. 2016. date accessed:03.09.2021 24. Song X, Peng Y, Wang X, et al. Incidence, survival, and risk factors for adults with acute myeloid leukemia not otherwise specified and acute myeloid leukemia with recurrent genetic abnormalities: analysis of the Surveillance, Epidemiology, and End Results (SEER) database, 2001-2013. Acta Haematol. 2018;139(2):115-127. 25. Shimoni A, Labopin M, Savani B, et al. Long-term survival and late events after allogeneic stem cell transplantation from HLAmatched siblings for acute myeloid leukemia with myeloablative compared to reduced-intensity conditioning: a report on behalf of the acute leukemia working party of European group for blood and marrow transplantation. J Hematol Oncol. 2016;9(1):118. 26. Bertoli S, Tavitian S, Huynh A, et al. Improved outcome for AML patients over the years 2000-2014. Blood Cancer J. 2017;7(12):635. 27. Koreth J, Schlenk R, Kopecky KJ, et al. Allogeneic stem cell transplantation for acute myeloid leukemia in first complete remission: systematic review and meta-analysis of prospective clinical trials. JAMA. 2009;301(22):2349-2361. 28. Schlenk RF, Döhner K, Mack S, et al. Prospective evaluation of allogeneic hematopoietic stem-cell transplantation from matched related and matched unrelated donors in younger adults with high-risk acute myeloid leukemia: German-Austrian trial AMLHD98A. J Clin Oncol. 2010;28(30):4642-4648. 29. Martino R, Caballero MaD, Pérez Simón JA, et al. Evidence for a graft-versus-leukemia effect after allogeneic peripheral blood stem cell transplantation with reduced-intensity conditioning in

Haematologica | 107 August 2022

1784

ARTICLE - Early blast persistence in AML

J. Ihlow et al.

acute myelogenous leukemia and myelodysplastic syndromes. Blood. 2002;100(6):2243-2245. 30. Sweeney C, Vyas P. The graft-versus-leukemia effect in AML. Front Oncol. 2019;9:1217. 31. Weisdorf D, Zhang M-J, Arora M, Horowitz MM, Rizzo JD, Eapen M. Graft-versus-host disease induced graft-versus-leukemia effect: greater impact on relapse and disease-free survival after reduced intensity conditioning. Biol Blood Marrow Transplant. 2012;18(11):1727-1733. 32. Gale RP, Horowitz HM. Graft-versus-leukemia in bone marrow transplantation. The Advisory Committee of the International Bone Marrow Transplant Registry. Bone Marrow Transplant. 1990;6(Suppl 1):94-97. 33. Schuurhuis GJ, Heuser M, Freeman S, et al. Minimal/measurable residual disease in AML: a consensus document from the European LeukemiaNet MRD Working Party. Blood. 2018;131(12):1275-1291. 34. Jongen-Lavrencic M, Grob T, Hanekamp D, et al. Molecular minimal residual disease in acute myeloid leukemia. N Engl J Med. 2018;378(13):1189-1199. 35. Balsat M, Renneville A, Thomas X, et al. Postinduction minimal residual disease predicts outcome and benefit from allogeneic stem cell transplantation in acute myeloid leukemia with NPM1 mutation: a study by the Acute Leukemia French Association Group. J Clin Oncol. 2016;35(2):185-193. 36. Goldberg AD, Famulare C, Devlin SM, et al. Molecular predictors and current management of minimal residual disease (MRD) following induction chemotherapy for acute myeloid leukemia (AML). Blood. 2018;132(Suppl 1):292-292. 37. Kayser S, Benner A, Thiede C, et al. Pretransplant NPM1 MRD levels predict outcome after allogeneic hematopoietic stem cell transplantation in patients with acute myeloid leukemia. Blood Cancer J. 2016;6(7):e449. 38. Li D, Wang L, Zhu H, Dou L, et al. Efficacy of allogeneic hematopoietic stem cell transplantation in intermediate-risk acute myeloid leukemia adult patients in first complete remission: a meta-analysis of prospective studies. PLoS One. 2015;10(7):e0132620. 39. Suciu S, Mandelli F, de Witte T, et al. Allogeneic compared with autologous stem cell transplantation in the treatment of patients younger than 46 years with acute myeloid leukemia (AML) in first complete remission (CR1): an intention-to-treat analysis of the EORTC/GIMEMAAML-10 trial. Blood. 2003;102(4):1232-1240.

40. Slovak ML, Kopecky K, Cassileth PA, et al. Karyotypic analysis predicts outcome of preremission and postremission therapy in adult acute myeloid leukemia: a Southwest Oncology Group/Eastern Cooperative Oncology Group study. Blood. 2000;96(13):4075-4083. 41. Cornelissen JJ, van Putten WLJ, Verdonck LF, et al. Results of a HOVON/SAKK donor versus no-donor analysis of myeloablative HLA-identical sibling stem cell transplantation in first remission acute myeloid leukemia in young and middle-aged adults: benefits for whom? Blood. 2007;109(9):3658-3666. 42. Yanada M, Matsuo K, Emi N, Naoe T. Efficacy of allogeneic hematopoietic stem cell transplantation depends on cytogenetic risk for acute myeloid leukemia in first disease remission. Cancer. 2005;103(8):1652-1658. 43. Brunet S, Esteve J, Berlanga J, et al. Treatment of primary acute myeloid leukemia: results of a prospective multicenter trial including high-dose cytarabine or stem cell transplantation as post-remission strategy. Haematologica. 2004;89(8):940-949. 44. Tsimberidou A-M, Stavroyianni N, Viniou N, et al. Comparison of allogeneic stem cell transplantation, high-dose cytarabine, and autologous peripheral stem cell transplantation as postremission treatment in patients with de novo acute myelogenous leukemia. Cancer. 2003;97(7):1721-1731. 45. Pfirrmann M, Ehninger G, Thiede C, et al. Prediction of postremission survival in acute myeloid leukaemia: a post-hoc analysis of the AML96 trial. Lancet Oncol. 2012;13(2):207-214. 46. Stelljes M, Krug U, Beelen DW, et al. Allogeneic transplantation versus chemotherapy as postremission therapy for acute myeloid leukemia: a prospective matched pairs analysis. J Clin Oncol. 2014;32(4):288-296. 47. Burnett AK, Wheatley K, Goldstone AH, et al. The value of allogeneic bone marrow transplant in patients with acute myeloid leukaemia at differing risk of relapse: results of the UK MRC AML 10 trial. Br J Haematol. 2002;118(2):385-400. 48. Freeman SD, Hills RK, Virgo P, et al. Measurable residual disease at induction redefines partial response in acute myeloid leukemia and stratifies outcomes in patients at standard risk without NPM1 mutations. J Clin Oncol. 2018;36(15):1486-1497. 49. Ivey A, Hills RK, Simpson MA, et al. Assessment of minimal residual disease in standard-risk AML. N Engl J Med. 2016;374(5):422-433. 50. Venditti A, Piciocchi A, Candoni A, et al. GIMEMA AML1310 trial of risk-adapted, MRD-directed therapy for young adults with newly diagnosed acute myeloid leukemia. Blood. 2019;134(12):935-945.

Haematologica | 107 August 2022

1785

ARTICLE - Acute Myeloid Leukemia

A novel CD34-specific T-cell engager efficiently depletes acute myeloid leukemia and leukemic stem cells in vitro and in vivo Lucas C. M. Arruda,1 Arwen Stikvoort,1 Melanie Lambert,2 Liqing Jin,3 Laura Sanchez Rivera,4 Renato M. P. Alves,4 Tales Rocha de Moura,5 Carsten Mim,5 Sören Lehmann,2 Rebecca Axelsson-Robertson,1,6 John E. Dick,3,7 Jonas Mattsson,3,8 Björn Önfelt,4,9 Mattias Carlsten2,10 and Michael Uhlin1,4,6 Department of Clinical Science, Intervention and Technology, Karolinska Institute, Stockholm, Sweden; 2Center for Hematology and Regenerative Medicine, Department of Medicine, Huddinge, Karolinska Institutet, Stockholm, Sweden; 3Princess Margaret Cancer Centre, University Health Network, Toronto, Ontario, Canada; 4Department of Applied Physics, Science for Life Laboratory, KTH Royal Institute of Technology, Stockholm, Sweden; 5Department for Biomedical Engineering and Health Systems, KTH Royal Institute of Technology, Stockholm, Sweden; 6Department of Immunology and Transfusion Medicine, Karolinska University Hospital, Stockholm, Sweden; 7Department of Molecular Genetics, University of Toronto, Ontario, Canada; 8Gloria and Seymour Epstein Chair in Cell Therapy and Transplantation, Princess Margaret Cancer Centre, University Health Network, Toronto, Ontario, Canada; 9Department of Microbiology, Tumor and Cell Biology, Science for Life Laboratory, Karolinska Institutet, Stockholm, Sweden and 10Center for Cell Therapy and Allogeneic Stem Cell Transplantation, Karolinska University Hospital, Stockholm, Sweden. 1

Correspondence: Lucas C. M. Arruda lucas.arruda@ki.se Received: June 23, 2021. Accepted: February 1, 2022. Prepublished: February 10, 2022. https://doi.org/10.3324/haematol.2021.279486 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

Abstract Less than a third of patients with acute myeloid leukemia (AML) are cured by chemotherapy and/or hematopoietic stem cell transplantation, highlighting the need to develop more efficient drugs. The low efficacy of standard treatments is associated with inadequate depletion of CD34+ blasts and leukemic stem cells, the latter a drug-resistant subpopulation of leukemia cells characterized by the CD34+CD38- phenotype. To target these drug-resistant primitive leukemic cells better, we have designed a CD34/CD3 bi-specific T-cell engager (BTE) and characterized its anti-leukemia potential in vitro, ex vivo and in vivo. Our results show that this CD34-specific BTE induces CD34-dependent T-cell activation and subsequent leukemia cell killing in a dose-dependent manner, further corroborated by enhanced T-cell-mediated killing at the singlecell level. Additionally, the BTE triggered efficient T-cell-mediated depletion of CD34+ hematopoietic stem cells from peripheral blood stem cell grafts and CD34+ blasts from AML patients. Using a humanized AML xenograft model, we confirmed that the CD34-specific BTE had in vivo efficacy by depleting CD34+ blasts and leukemic stem cells without side effects. Taken together, these data demonstrate that the CD34-specific BTE has robust antitumor effects, supporting development of a novel treatment modality with the aim of improving outcomes of patients with AML and myelodysplastic syndromes.

Introduction Acute myeloid leukemia (AML) and high-risk myelodysplastic syndromes (MDS) are poor prognosis hematologic malignancies characterized by abnormal hematopoiesis and dysfunctions of the hematopoietic stem cell (HSC) system.1,2 Chemotherapy remains the standard of care but is associated with side effects and often high rates of relapse. At present, less than a third of patients diagnosed with AML are cured.3 Resistance to standard treatment has recently been attributed to inadequate depletion of leukemic stem cells (LSC), a self-renewing population of

leukemic cell progenitors characterized, typically but not exclusively, by the CD34+CD38- phenotype.4-6 Relapsefated LSC have been identified in combined stem cell/clonal evolution studies of paired diagnosis-relapse samples7 and the survival of phenotypic LSC after chemotherapy is associated with minimal residual disease and subsequent relapse.8 LSC-specific gene expression signatures9 and high LSC burden are associated with high risk of relapse and death.10-13 Hence, LSC represent optimal targets for new therapies aimed at improving the outcome of patients with AML/MDS.14 Bi-specific T-cell engagers (BTE) are promising immuno-

Haematologica | 107 August 2022

1786

ARTICLE - CD34-specific BTE for AML/MDS treatment

L.C.M. Arruda et al.

therapeutic agents intended for cancer treatment. BTE are small molecules constructed of two single chain variable fragments (scFv) connected in tandem by a flexible linker that acts by retargeting T cells against tumor cells.15 One scFv binds to CD3, while the second scFv binds to a tumor-associated antigen.16 This structure and specificity allow a BTE construct to physically link a T cell to a tumor cell, stimulating effector cell activation and ultimately leading to cytokine production and tumor killing.17 The arrangement of antigen-binding sites and the flexible linker forces target and effector cells into very close proximity, leading to the formation of a cytolytic immune synapse that can counteract the multiple mechanisms of immune evasion presented by cancer cells, such as MHC downregulation.15,16 Several studies have shown that LSC are resistant to conventional chemotherapy and radiation-based therapies and to NK-/T-cell-mediated cytotoxicity18-20 due to drug extrusion20 and low immunogenicity, resulting from low expression of major immune response molecules.18 A LSC-targeting BTE could efficiently bypass this and mediate immune activation towards LSC, thereby being a more specific and less toxic treatment than those currently available.15 The CD19/CD3 BTE blinatumomab has been used to treat B-cell acute lymphoblastic leukemia with great success,21–23 representing a great advance on the use of BTE to treat blood cancers. Leukemic cells express several antigens that are under investigation for targeted immunotherapies, such as CD33. Gemtuzumab (an anti-CD33 monoclonal antibody) was recently approved by the Food and Drug Administration for the treatment of relapsed/refractory CD33+ AML,24 being a very significant step toward defining new immunotherapeutic treatment regimens in AML. Several BTE have been investigated to treat AML and MDS, targeting CD33,25 CD123 26 and WT1,27 but no CD34-specific BTE has been developed. CD34 is a molecule expressed almost exclusively by normal stem and progenitor cells but is also found in AML/MDS blasts, and to a lesser degree on renal vessel walls. CD34 is highly expressed by AML blasts and is associated with several genetic aberrancies characteristic of the development of AML.28 CD34 expression has been linked to increased resistance to apoptosis and to multidrug resistance,29 protecting CD34+ blasts from the immune system and chemotherapeutic drugs. Because of such characteristics, CD34 expression is associated with low survival and high relapse rates.11-13,28 Indeed, the presence of CD34+ leukemic cells was associated with an approximately 4-fold lower event-free survival than in patients with absence of CD34+ leukemic cells,28 indicating that this molecule is a promising target to improve AML treatment. Here, we demonstrate the preclinical efficacy of a CD34specific BTE to target CD34+ AML cells in vitro and in vivo. We observed that the BTE promoted specific killing of

AML cell lines and AML blasts by T-cell-dependent mechanisms. The antibody showed no toxicity in AML-bearing immune-deficient mice infused with human T cells, making this bi-specific antibody a very attractive clinical candidate with potential to improve the outcome of patients with AML and MDS.

Methods T-cell isolation and in vitro killing assay T-cell cytotoxicity was assessed by fluorescence activated cell sorting (FACS) using purified T cells from peripheral blood mononuclear cells as effector cells and cancer cell lines (KG1a, Kasumi 1, NALM-6), primary AML blasts or hCMEC/D3 as target cells. Cells were co-cultured at an effector-to-target ratio (E.T) of 3:1 for 48 h (cancer cell lines and hCMEC/D3) and 72 h (primary AML samples), and serial dilutions of αCD34, CD34/CD3 and RSV/CD3 BTE. Target cell killing was assessed by gating on the CellTrace+ fraction and evaluating the annexin V and 7-aminoactinomysin D staining.30,31 For more information see the Online Supplementary Appendix. Proliferation assay Purified T cells were stained with 2 mM of CellTrace violet (ThermoFisher) according to the manufacturer’s instructions and incubated with target cells (KG1a, Kasumi 1, NALM-6) and serial dilutions of αCD34, CD34/CD3 and RSV/CD3 BTE for 5 days at 37°C in complete RPMI medium. The fraction of CellTracelow proliferating T cells was assessed by FACS. Single-cell killing assay Time-lapse live single-cell imaging was performed as previously described.32,33 A detailed description is provided in the Online Supplementary Appendix. Human primary acute myeloid leukemia samples The blood samples from AML patients were collected at Princess Margaret Cancer Center (Toronto, Canada) and CD34+ AML blasts were used as target cells in the killing assays. Purified T cells from donor lymphocyte infusions were used as effector cells. Bone marrow samples from three AML patients were collected at Karolinska University Hospital (Huddinge, Sweden). Cells were co-cultured with T cells (n= 5) isolated from heathy donor buffy coats in an E:T of 3:1 for 72 h in the presence of serial dilutions of CD34/CD3 BTE. All patients provided informed consent in accordance with the Declaration of Helsinki and with approval of the ethics committees in the respective centers (ethical approval 2010-1496-31-3). After incubation, AML blast cell depletion and T-cell activation were assessed by FACS. See the Online Supplementary Appendix for details.

Haematologica | 107 August 2022

1787

ARTICLE - CD34-specific BTE for AML/MDS treatment

L.C.M. Arruda et al.

Human hematopoietic stem cell depletion assay Unmanipulated peripheral blood stem cell samples were collected during preparation of grafts for hematopoietic stem cell transplantation (HSCT) and processed as previously described.34 HSC were purified by positive selection of CD34+ cells using the CD34 MicroBead Kit UltraPure (Miltenyi Biotec). T cells were purified from the same graft using negative selection (Pan T-Cell Isolation Kit; Miltenyi Biotec). Immediately after purification, T cells and HSC were seeded at a 3:1 ratio and cultured for 48 h with serial dilutions of αCD34, CD34/CD3 and RSV/CD3 BTE. T-cell activation was assessed by CD25 and CD69 expression and depletion of CD34+ HSC by gating the CD45dimCD34+ population. See the Online Supplementary Appendix for details. In vivo xenograft model The antitumor efficacy and safety of a CD34-specific BTE were evaluated in 6- to 8-week-old NOD.CgPrkdc scid il2rg tm1Wjl Tg(CMV-IL3,CSF2,KITLG)1Eav/MloySzJ (NSG-SGM3) mice (Taconic). All the experiments were performed according to the Swedish Animal Welfare Ordinance and approved by the local animal ethics committee (ethical approval ID1533). See the Online Supplementary Appendix for details. Statistical analysis In vitro data were analyzed using the Wilcoxon matched pairs signed rank-test. In vivo data from mice were compared using an unpaired Student t-test. Correlations were assessed using a non-parametric Spearman rank correlation coefficient. GraphPad Prism 9.0 was used, with significance set at P≤0.05.

Results Design and expression of CD34/CD3 bi-specific T-cell engagers and controls CD34/CD3 and RSV/CD3 BTE were assembled from two scFv domains by recombinant DNA technology (Online Supplementary Figure S1A, B). The scFv domain binding to CD34 is positioned N-terminally, and the scFv binding to CD3ɛ C-terminally, followed by a hexa-histidine sequence. FACS analysis demonstrated that the CD34/CD3 BTE is capable of binding to the CD34+ AML cell lines KG1a and Kasumi 1 as well as to purified human T cells (Online Supplementary Figure S1C, D). We found no cross-reactivity against mouse CD34 (Online Supplementary Figure S1E). The RSV/CD3 control BTE have exactly the same CD3ɛ scFv binding domain as that of the CD34/CD3 BTE in addition to the RSV-specific scFv domain, allowing us to assess whether target specificity is required for the T-cell activation. The

αCD34 unit, which binds to the target protein but does not trigger CD3 activation, was also developed as a control. The CD34-specific bi-specific T-cell engager redirects T-cell effector function towards acute myeloid leukemia cell lines We next addressed whether the CD34/CD3 BTE was able to redirect primary human T cells to deplete CD34+ human AML cell lines. Initially, we assessed the T-cell activation and killing dynamics by co-culturing effector T cells and target AML/acute lymphoblastic leukemia cell lines at an E:T ratio of 3:1 for a time range between 24 h and 120 h (Online Supplementary Figure S2). We observed that 48 h was the optimal timing to assess Tcell effector function and the earliest timepoint at which maximum killing is observed, being adopted thereafter. By co-culturing T cells and target cell lines for 48 h in the presence of increasing concentrations of BTE or controls, we observed that CD34-BTE efficiently triggered T-cell-mediated depletion of the CD34hi KG1a and CD34low Kasumi 1 cell lines, while both RSV/CD3 and αCD34 controls killed none of the target cell lines (Figure 1A, B). The CD34/CD3 BTE did not trigger any unspecific killing of CD34- cells as shown using the CD34- human cell line NALM-6 (Figure 1C), indicating that CD34/CD3 BTE killing is specific to CD34+ cells. The CD34 bi-specific T-cell engager mediates cellcell interactions and target cell killing In order to understand better how the CD34 BTE affects T-cell dynamics we used a single-cell microchip-based method for live cell imaging (Online Supplementary Figure S3).32,33 Effector T cells and target KG1a cells were stained with distinct dyes and seeded into a siliconglass microchip containing thousands of wells, in which interactions between effector and target cells can be followed by live imaging (Online Supplementary Figure S3). T-cell interaction dynamics and target cell lysis were assessed without or in the presence of CD34/CD3 or RSV/CD3. The presence of the CD34/CD3 BTE led to significant increases in cytotoxicity (Figure 2A) and fraction of lytic contacts (Figure 2B) as compared to RSV/CD3 or untreated conditions. The CD34/CD3 BTE also resulted in longer lytic and non-lytic contacts between T cells and target cells as compared to the other conditions (Figure 2C). The BTE also resulted in fewer contacts being formed (Figure 2D), which is a sign of the BTE mediating more stable contacts, and a shorter time until the T cell makes the first contact with target cells (Figure 2E). Taken together, our data show that the BTE led to formation of stable contacts between T cells and target cells often resulting in target cell lysis.

Haematologica | 107 August 2022

1788

ARTICLE - CD34-specific BTE for AML/MDS treatment

L.C.M. Arruda et al.

Figure 1. The CD34-targeting bi-specific T-cell engager mediates CD34+ target-cell killing. (A-C) Purified T cells were co-cultured with KG1a (A), Kasumi1 (B) and NALM6 (C) at an effector-to-target ratio of 3:1 in the presence of anti-CD34 antibody (black), RSV/CD3 (blue) or CD34/CD3 (red) bi-specific T-cell engagers. Dose response killing was assessed 48 h later by fluorescence activated cell sorting (n=3, mean ± standard error of mean).

Figure 2. CD34-specific bi-specific T-cell engager mediates stronger cell-cell contact and T-cell killing. Primary T cells were purified and seeded with KG1a target cells onto a silicon-glass microchip containing CD34-specific (bi-specific T-cell engager, BTE) or RSV-specific antibodies (RSV) or left untreated (none). Images were acquired every 5 minutes over 21 hours. (A) Percentage of cytotoxic T cells (n=6; mean ± standard error of mean [SEM]). (B) Percentage of all contacts that led to target cell death (n=6; mean ± SEM). (C) Duration of the contacts for the different treatments (none, RSV, BTE). For the BTE condition contacts have been divided into non-lytic (no target cell death) and lytic (target cell death). Individual dots of the same color represent single contact events from one donor, and the bigger circles of the same color represent the median time for that donor (n=6; mean ± SEM), differences between conditions were evaluated by a paired t-test using the median values. (D) Number of contacts per T cell (each circle depicts the median number of events per donor, n=6; mean ± SEM). (E) Time from the beginning of the experiment until the T cell makes the first contact with the target cell (individual dots of the same color represent the first contact in a specific well from one donor, and the bigger circles of the same color represent the median time for that donor (n=6; mean ± SEM), differences between conditions were evaluated by a paired t-test using the median values). Each color represents a different donor (n=6), and color coding is maintained in all figures. Statistical significance: *P< 0.05. **P< 0.01. ns: non-significant (paired t-test). Haematologica | 107 August 2022

1789

ARTICLE - CD34-specific BTE for AML/MDS treatment

L.C.M. Arruda et al.

Figure 3. Bi-specific T-cell engager-induced interaction and killing leads to T-cell activation and proliferation. Purified T cells were co-cultured with KG1a, Kasumi1 and NALM6 in the presence of anti-CD34 antibody, RSV/CD3 or CD34/CD3 bi-specific Tcell engager (BTE) at the concentration of 100 ng/mL. (A) CD3+CD4+(CD8+)CD25+CD69+ activated T cells were assessed after 48 h of co-culture (mean ± standard error of mean [SEM], n=3). (B) Representative histogram overlays of proliferation-induced CellTrace dilution on CD3+ T cells after 5 days of co-culture. (C) Proliferation quantification gated as CellTracelow T cells (mean ± SEM, n=3). *P<0.05, **P<0.01, ***P<0.001 (Wilcoxon test).

Bi-specific T-cell engager-mediated T-cell activation and proliferation are CD34-dependent Next, we examined T-cell activation and proliferation. We observed that both CD4+ and CD8+ T cells presented high levels of CD25/CD69 expression when the CD34+ cell lines KG1a and Kasumi-1 were co-cultured with T cells in the presence of CD34/CD3 BTE (Figure 3A). No unspecific activation was found when the CD34- NALM6 cell line was used as target cells. In the same way, T cells only presented significant expansion when cultured with CD34-expressing target cells with CD34/CD3 BTE (Figure 3B, C), indicating that the T-cell activation is CD34-dependent. The CD34-bi-specific T-cell engager depletes hematopoietic stem cells Since CD34 is constitutively expressed by HSC, the CD34specific BTE may deplete not only CD34+AML blasts but also healthy HSC. To test this, effector T cells and HSC were purified from the same peripheral blood stem cell grafts and co-cultured in the presence of either the αCD34 antibody, RSV/CD3 BTE or the CD34/CD3 BTE. After co-culture, a significant depletion of CD34+ HSC was observed for the CD34/CD3 BTE, while no significant killing was observed in the presence of the αCD34 antibody or the RSV/CD3 BTE (Figure 4A, B). This was accompanied by a dose-dependent increase of activated CD4+CD25+CD69+ and CD8+CD25+CD69+ T cells in the CD34-BTE-treated conditions, while no activation was observed in the controls (Figure 4C, D). These data indicate that the CD3/CD34 BTE induces T-cell-mediated HSC killing.

The CD34-bi-specific T-cell engager depletes human CD34+ blast cells from acute myeloid leukemia patients ex vivo CD34+ blast count in the peripheral blood and bone marrow is used for diagnosis and prognosis of AML.1 In order to test the ability of the CD34-specific BTE to deplete CD34+ blasts, we co-cultured effector T cells isolated from donor lymphocyte infusions with target CD34+ AML blasts from primary patients’ samples. We initially compared an incubation time of 48 h versus 72 h in the presence of several concentrations of CD34 BTE and corresponding controls, with the later timepoint presenting clearer results regarding T-cell activation and blast killing, being adopted thereafter (Online Supplementary Figure S4). We observed no CD34+ blast depletion in either the αCD34- or RSV/CD3-treated group (Figure 5A), despite the increase of activated T cells in the RSV/CD3 control group (Figure 5B). The CD34/CD3 BTE triggered both significant T-cell activation and killing of CD34+ blasts (Figure 5A, B), resulting in significant blast reduction (Figure 5C). No correlation was observed between blast reduction and CD34 expression (Figure 5D). Dose-dependent T-cell activation and blast depletion were observed when T cells from five different donors were cultured with bone marrow-derived CD34+ blasts from AML patients (Online Supplementary Figure S5), together suggesting that the BTE can trigger efficient T-cell-mediated depletion of primary CD34+ AML blasts ex vivo. Next, we used the cell line hCMEC/D3 as the target to test whether the developed CD34 BTE can promote the killing

Haematologica | 107 August 2022

1790

ARTICLE - CD34-specific BTE for AML/MDS treatment

L.C.M. Arruda et al.

Figure 4. The CD34-specific bi-specific T-cell engager efficiently depletes hematopoietic stem cells. T cells and CD34+ hematopoietic stem cells (HSC) were isolated from the same peripheral blood stem cell grafts and co-cultured at an effector:target ratio of 3:1 in the presence of anti-CD34 antibody, RSV/CD3, CD34/CD3 from 102 to 104 ng/mL or left untreated. (A) Representative plot of CD45dimCD34+ HSC quantification after 48 h of co-culture. (B) Quantification of CD45dimCD34+ HSC depletion (mean ± standard error of mean [SEM], n=6). (C) CD3+CD4+CD25+CD69+ and (D) CD3+CD8+CD25+CD69+ activated T cells separated by fluorescence activated cell sorting (mean ± SEM, n=6). *P<0.05, **P<0.01, ***P<0.001 (Wilcoxon test).

of human CD34+ endothelial cells. We found that, even at high concentrations (103 ng/mL), the CD34 BTE was not able to deplete the endothelial cells (Online Supplementary Figure S6), indicating that the BTE does not crossreact with CD34 protein expressed by endothelial cells. In vivo efficacy of the bi-specific T-cell engager To address the potential of the anti-CD34 BTE in vivo, we next established the hCD34+ KG1a cell line in NSG-SGM3 mice via intravenous injection and then randomized them into three different groups. Two groups of mice received two consecutive cycles of one intraperitoneal injection of freshly isolated human T cells followed by daily intravenous injections of either BTE or phosphate-buffered saline (PBS) (Figure 6A). One group only received daily intravenous injections of PBS, at the time point corresponding to that of BTE, but no T cells. The mice were

euthanized at the predetermined time on day 21, at which point the AML burden was measured, and the T cells quantified. No side effects of the treatment, including after BTE administration, were observed. At day 21 there were statistically significant reductions of leukemia burden in both bone marrow (Figure 6B) and spleen (Figure 6C) in mice receiving T cells and BTE compared to those that received T cells only and PBS control. Of note, the leukemia clearance was near complete in the bone marrow of mice receiving the combination of T cells and BTE (Figure 6B). As we had observed tumor reduction in both these groups compared to the tumor burden in the group given PBS only, we hypothesized that this was primarily attributed to the presence of T cells. To address this, we next analyzed the T-cell compartment in the harvested organs by flow cytometry. The data revealed persistence of T cells in mice treated with the CD34/CD3 BTE,

Haematologica | 107 August 2022

1791

ARTICLE - CD34-specific BTE for AML/MDS treatment

L.C.M. Arruda et al.

Figure 5. Bi-specific T-cell engager treatment leads to depletion of patients’ CD34+ acute myeloid leukemia blasts. (A) Purified T cells from donor lymphocyte infusions were co-cultured with CD34+ blasts isolated from patients with acute myeloid leukemia in the presence of anti-CD34 antibody, RSV/CD3 or CD34/CD3 bi-specific T-cell engagers (BTE) at a concentration of 2.5 mg/mL for 72 h (n=14). (B) Quantification of CD3+CD25+CD69+ activated T cells (n=14). (C) Percentage CD34+ blast reduction as compared to untreated (No Ab) control. (D) Correlation between blast reduction and CD34 expression (Spearman rank correlation coefficient). Dots represent individual patients and the connecting lines the paired samples. *P<0.05, ***P<0.001 (Wilcoxon test).

while the PBS-treated mice had a very low T-cell frequency (<1% in all mice) but, in contrast, high counts of CD34+ blasts (Figure 6C, D). Collectively, these data show that hCD34+ AML-engrafted mice treated with the combination of primary human T cells and the CD3/CD34 BTE have significantly reduced tumor burden associated with T-cell persistence, without any side effects in the mice.

Discussion Leukemia-specific BTE have been developed intensively in the last years with prominent in vitro and in vivo effects by targeting CD33,25 CD12326 and WT127 in AML and CD19 in B-cell acute lymphoblastic leukemia.21 We have developed a CD34-specific BTE aiming to deplete both CD34-expressing HSC and leukemic blasts, including LSC. Here we have shown that the CD34/CD3 BTE is able to promote Tcell activation and killing of CD34-expressing target cells with high efficacy in vitro and in vivo, supporting the translation of this drug into clinical trials.

In patients with high-risk MDS/AML, HSCT is the only curative treatment. Disease relapse, infections and graftversus-host disease are still the major causes of treatment failure, with few improvements in the last years.35 This indicates that adjuvant therapies are necessary to make HSCT safer and new approaches are required to treat relapsed patients after HSCT. CD34+ leukemic cell frequency has been associated with high relapse rates,10-13 implying that the use of leukemia-depleting drugs in HSCT protocols could be beneficial to reduce the tumor burden. In this scenario, treatment with CD34-targeting BTE prior to HSCT would trigger the patient’s T cells to deplete CD34+ leukemic blasts, LSC and HSC. As a consequence, this adjuvant treatment would decrease the use of cytotoxic and cytostatic conditioning drugs before HSCT, reducing life-threatening complications such as infections. Recently, an anti-CD117 monoclonal antibody has been used with non-myeloablative conditioning in patients with minimal residual disease-positive MDS/AML undergoing allogeneic HSCT.36 The treatment has been shown to be safe, well-tolerated and efficient in clearing MDS/AML minimal residual disease, promoting complete donor

Haematologica | 107 August 2022

1792

ARTICLE - CD34-specific BTE for AML/MDS treatment

L.C.M. Arruda et al.

Figure 6. In vivo efficacy of the bi-specific T-cell engager against acute myeloid leukemia in a xenograft mouse model. (A) Schematic overview of the protocol, in which NSG-SGM3 mice were inoculated with the hCD34+ acute myeloid leukemia (AML) cell line KG1a intravenously before receiving two cycles of freshly isolated human T cells intraperitoneally (days 3 and 10) and either 1 mg/kg of bi-specific T-cell engager (BTE) or phosphate-buffered saline (PBS) daily intravenously (day 3 to 21). AML burden and T-cell persistence were assessed by organ harvest and flow cytometry at the predetermined termination day 21. Quantification of hCD34+ leukemia cells in (B) bone marrow and (C) spleen. (D) Quantification on hCD3+ T cells in spleen. Five mice per group were analyzed. **P<0.01, ***P<0.001, ****P<0.0001 (unpaired t-test).

chimerism.36 Less than 25% of patients who could benefit from HSCT actually undergo transplantation because of the high toxicity associated with conditioning regimens.37 There are indications that adjuvant antibody-based therapies, such as the here presented CD34 BTE, can reduce conditioning toxicities and expand the use of HSCT to older and more fragile patients. CD34 is expressed in other tissues besides HSC and blasts, and a potential side effect of targeting CD34 is the recognition of CD34+ endothelial cells.38 Considering that BTE have a relatively short half-life (≈2h), possible side effects could be circumvented by halting drug administration. Moreover, we show here that the BTE developed did not present significant activity against a human CD34+ endothelial cell line even at high concentrations, indicating that our CD34-specific BTE is specific for CD34 expressed by cells from the hematologic lineage rather than from the endothelial lineage. More studies, including patientderived xenograft models, humanized mice models (hCD34+) and non-human primates are necessary to confirm these observations. This lack of reactivity could be explained by the presence of several CD34 epitopes that are differentially expressed in specific tissues as results of heavy glycosylation,39,40 and therefore can be targeted by specific antibodies.41 Additionally, further engineering of the CD34-specific scFv might improve BTE affinity to CD34 expressed by leukemic blasts and HSC rather than endothelial CD34. The CD34-specific BTE developed could also be used in the treatment of a wide range of non-malignant diseases as an HSC-specific depleting agent in a non-myeloablative conditioning treatment regimen, as shown by the use of anti-CD117 antibodies.42,43 This would be of particular interest in cases of inborn mutations that can be cured by gene therapy or HSCT in which reduced intensity

chemotherapy is desired. In this case, the CD34-specific BTE could be used as adjuvant therapy prior to HSC, and combined with anti-T-cell antibodies, it may be possible to significantly reduce chemotherapy intensity during conditioning therapy. This would result in significantly reduced short- and long-term side effects of chemotherapy following HSCT. In this regard, there are recent reports that the use of anti-CD117 as non-toxic single-agent transplant conditioning may supplant conventional conditioning for newly diagnosed infants with severe combined immunodeficiency, thereby avoiding the toxicities of chemotherapy.44,45 Currently, HSCT is still restricted to patients with otherwise incurable malignant diseases and it is not indicated for most of the patients who could benefit from it because of its high risks and toxicity.37 This is largely due to the use of non-specific cytotoxic chemotherapy and the irradiation conditioning necessary to enable engraftment of donor HSC. Regular conditioning regimens lead to longterm immune deficiency and multi-organ damage, which is associated with a high risk of infections.46 This is particularly severe in children, in whom conditioning regimens are associated with infertility, hormonal dysfunction and growth problems.47 The developed CD34specific BTE could help to eliminate such severe conditioning regimens and dramatically improve HSCT and expand its use. Disclosure JED receives royalties from Trillium Therapeutics Inc and reports receiving a commercial research grant from Celgene. The other authors declare that they have no competing financial interests.

Haematologica | 107 August 2022

1793

ARTICLE - CD34-specific BTE for AML/MDS treatment

L.C.M. Arruda et al.

Contributions LCMA, AS and RAR performed in vitro experiments with cell lines and human hematopoietic stem cells. AS and LJ conducted the in vitro studies with primary acute myeloid leukemia samples. ML performed the in vivo experiments. LSR and RA performed the single-cell killing experiments. TRM and CM purified the antibodies. SL, JED, JM, BO, MC, and MU provided financial support. LCMA and MU designed the antibodies and wrote the manuscript. All authors critically reviewed the manuscript and approved the final ver-

sion. AS, ML and LJ contributed equally to this work. Funding This work was supported by research funding from Stiftelsen Felix Mindus Bidrag till Leukemiforskningen, David och Astrid Hageléns Stiftelse, Wallenberg, Cancerfonden, and Barncancerfonden. Vetenskapsrådet (2021-01755) Data-sharing statement Data are available upon request to the author.

References 1. Short NJ, Rytting ME, Cortes JE. Acute myeloid leukaemia. Lancet. 2018;392(10147):593-606. 2. Doll DC, List AF. Myelodysplastic syndromes. West J Med. 1989;151(2):161-167. 3. Cancer Stat Facts: leukemia — acute myeloid leukemia (AML). https://seer.cancer.gov/statfacts/html/amyl.html. 4. Vetrie D, Helgason GV, Copland M. The leukaemia stem cell: similarities, differences and clinical prospects in CML and AML. Nat Rev Cancer. 2020;20(3):158-173. 5. Lapidot T, Sirard C, Vormoor J, et al. A cell initiating human acute myeloid leukaemia after transplantation into SCID mice. Nature. 1994;367(6464):645-648. 6. Bonnet D, Dick JE. Human acute myeloid leukemia is organized as a hierarchy that originates from a primitive hematopoietic cell. Nat Med. 1997;3(7):730-737. 7. Shlush LI, Mitchell A, Heisler L, et al. Tracing the origins of relapse in acute myeloid leukaemia to stem cells. Nature. 2017;547(7661):104-108. 8. Ho TC, LaMere M, Stevens BM, et al. Evolution of acute myelogenous leukemia stem cell properties after treatment and progression. Blood. 2016;128(13):1671-1678. 9. Ng SWK, Mitchell A, Kennedy JA, et al. A 17-gene stemness score for rapid determination of risk in acute leukaemia. Nature. 2016;540(7633):433-437. 10. Van Rhenen A, Feller N, Kelder A, et al. High stem cell frequency in acute myeloid leukemia at diagnosis predicts high minimal residual disease and poor survival. Clin Cancer Res. 2005;11(18):6520-6527. 11. Jentzsch M, Geus U, Grimm J, et al. Pretreatment CD34+/CD38– cell burden as prognostic factor in myelodysplastic syndrome patients receiving allogeneic stem cell transplantation. Biol Blood Marrow Transplant. 2019;25(8):1560-1566. 12. Jentzsch M, Bill M, Nicolet D, et al. Prognostic impact of the CD34+/CD38− cell burden in patients with acute myeloid leukemia receiving allogeneic stem cell transplantation. Am J Hematol. 2017;92(4):388-396. 13. Zeijlemaker W, Grob T, Meijer R, et al. CD34+CD38− leukemic stem cell frequency to predict outcome in acute myeloid leukemia. Leukemia. 2019;33(5):1102-1112. 14. Saygin C, Matei D, Majeti R, Reizes O, Lathia JD. Targeting cancer stemness in the clinic: rrom hype to hope. Cell Stem Cell. 2019;24(1):25-40. 15. Velasquez MP, Bonifant CL, Gottschalk S. Redirecting T cells to hematological malignancies with bi-specific antibodies. Blood. 2018;131(1):30-38. 16. Labrijn AF, Janmaat ML, Reichert JM, Parren PWHI. Bispecific

antibodies: a mechanistic review of the pipeline. Nat Rev Drug Discov. 2019;18(8):585-608. 17. Einsele H, Borghaei H, Orlowski RZ, et al. The BTE (bi-specific T-cell engager) platform: development and future potential of a targeted immuno-oncology therapy across tumor types. Cancer. 2020;126(14):3192-3201. 18. Costello RT, Mallet F, Gaugler B, et al. Human acute myeloid leukemia CD34+/CD38- progenitor cells have decreased sensitivity to chemotherapy and Fas-induced apoptosis, reduced immunogenicity, and impaired dendritic cell transformation capacities. Cancer Res. 2000;60(16):4403-4411. 19. She M, Niu X, Chen X, et al. Resistance of leukemic stem-like cells in AML cell line KG1a to natural killer cell-mediated cytotoxicity. Cancer Lett. 2012;318(2):173-179. 20. Raaijmakers MHGP, De Grouw EPLM, Heuver LHH, et al. Breast cancer resistance protein in drug resistance of primitive CD34+38- cells in acute myeloid leukemia. Clin Cancer Res. 2005;11(6):2436-2444. 21. Gökbuget N, Dombret H, Bonifacio M, et al. Blinatumomab for minimal residual disease in adults with B-cell precursor acute lymphoblastic leukemia. Blood. 2018;131(14):1522-1531. 22. Bargou R, Leo E, Zugmaier G, et al. Tumor regression in cancer patients by very low doses of a T cell-engaging antibody. Science. 2008;321(5891):974-977. 23. Löffler A, Kufer P, Lutterbüse R, et al. A recombinant bi-specific single-chain antibody, CD19 x CD3, induces rapid and high lymphoma-directed cytotoxicity by unstimulated T lymphocytes. Blood. 2000;95(6):2098-2103. 24. Norsworthy KJ, Ko C-W, Lee JE, et al. FDA approval summary: mylotarg for treatment of patients with relapsed or refractory CD33-positive acute myeloid leukemia. Oncologist. 2018;23(9):1103-1108. 25. Aigner M, Feulner J, Schaffer S, et al. T lymphocytes can be effectively recruited for ex vivo and in vivo lysis of AML blasts by a novel CD33/CD3-bi-specific BTE antibody construct. Leukemia. 2013;27(5):1107-1115. 26. Hutmacher C, Volta L, Rinaldi F, et al. Development of a novel fully-human anti-CD123 antibody to target acute myeloid leukemia. Leuk Res. 2019;84:106178. 27. Dao T, Pankov D, Scott A, et al. Therapeutic bi-specific T-cell engager antibody targeting the intracellular oncoprotein WT1. Nat Biotechnol. 2015;33(10):1079-1086. 28. Zeijlemaker W, Kelder A, Wouters R, et al. Absence of leukaemic CD34+ cells in acute myeloid leukaemia is of high prognostic value: a longstanding controversy deciphered. Br J Haematol. 2015;171(2):227-238. 29. Suárez L, Vidriales MB, García-Laraña J, et al. CD34+ cells from

Haematologica | 107 August 2022

1794

ARTICLE - CD34-specific BTE for AML/MDS treatment

L.C.M. Arruda et al.

acute myeloid leukemia, myelodysplastic syndromes, and normal bone marrow display different apoptosis and drug resistance - associated phenotypes. Clin Cancer Res. 2004;10(22):7599-7606. 30. Robinson HR, Qi J, Cook EM, et al. A CD19/CD3 bi-specific antibody for effective immunotherapy of chronic lymphocytic leukemia in the ibrutinib era. Blood. 2018;132(5):521-532. 31. Hipp S, Tai Y-T, Blanset D, et al. A novel BCMA/CD3 bi-specific Tcell engager for the treatment of multiple myeloma induces selective lysis in vitro and in vivo. Leukemia. 2017;31(8):1743-1751. 32. Sarhan D, Brandt L, Felices M, et al. 161533 TriKE stimulates NKcell function to overcome myeloid-derived suppressor cells in MDS. Blood Adv. 2018;2(12):1459-1469. 33. Guldevall K, Brandt L, Forslund E, et al. Microchip screening platform for single cell assessment of NK cell cytotoxicity. Front Immunol. 2016;7(119):1-7. 34. Arruda LCM, Gaballa A, Uhlin M. Graft γδ TCR sequencing identifies public clonotypes associated with hematopoietic stem cell transplantation efficacy in acute myeloid leukemia patients and unravels cytomegalovirus impact on repertoire distribution. J Immunol. 2019;202(6):1859-1870. 35. Horowitz M, Schreiber H, Elder A, et al. Epidemiology and biology of relapse after stem cell transplantation. Bone Marrow Transplant. 2018;53(11):1379-1389. 36. Muffly LS, Chin M, Kwon H-S, et al. Early results of phase 1 study of JSP191, an anti-CD117 monoclonal antibody, with nonmyeloablative conditioning in older adults with MRD-positive MDS/AML undergoing allogeneic hematopoietic cell transplantation. J Clin Oncol. 2021;39(15_suppl):7035. 37. Yao S, Hahn T, Zhang Y, et al. Unrelated donor allogeneic hematopoietic cell transplantation is underused as a curative therapy in eligible patients from the United States. Biol Blood Marrow Transplant. 2013;19(10):1459-1464. 38. Uhlén M, Fagerberg L, Hallström BM, et al. Proteomics. tissuebased map of the human proteome. Science.

2015;347(6220):1260419. 39. Croockewit AJ, Raymakers RA, Preijers FW, Vierwinden G, De Witte TJ. The role of the different CD34 epitopes in detection and positive selection of CD34 + bone marrow and peripheral blood stem cells. Scand J Immunol. 1998;47(1):82-90. 40. Steen R, TjØnnfjord GE, Gaudernack G, Brinch L, Egeland T. Differences in the distribution of CD34 epitopes on normal haemopoietic progenitor cells and leukaemic blast cells. Br J Haematol. 1996;94(4):597-605. 41. Sidney LE, Branch MJ, Dunphy SE, Dua HS, Hopkinson A. Concise review: evidence for CD34 as a common marker for diverse progenitors. Stem Cells. 2014;32(6):1380-1389. 42. Chhabra A, Ring AM, Weiskopf K, et al. Hematopoietic stem cell transplantation in immunocompetent hosts without radiation or chemotherapy. Sci Transl Med. 2016;8(351):351ra105. 43. Czechowicz A, Palchaudhuri R, Scheck A, et al. Selective hematopoietic stem cell ablation using CD117-antibody-drugconjugates enables safe and effective transplantation with immunity preservation. Nat Commun. 2019;10(1):617. 44. Agarwal R, Weinberg KI, Kwon H-S, et al. First report of nongenotoxic conditioning with JSP191 (anti-CD117) and hematopoietic stem cell transplantation in a newly diagnosed patient with severe combined immune deficiency. Blood. 2020;136(Suppl 1):10. 45. Agarwal R, Dvorak CC, Kwon H-S, et al. Non-genotoxic antiCD117 antibody conditioning results in successful hematopoietic stem cell engraftment in patients with severe combined immunodeficiency. Blood. 2019;134(Suppl_1):800. 46. Bearman SI, Appelbaum FR, Buckner CD, et al. Regimen-related toxicity in patients undergoing bone marrow transplantation. J Clin Oncol. 1988;6(10):1562-1568. 47. Matthes-Martin S, Lamche M, Ladenstein R, et al. Organ toxicity and quality of life after allogeneic bone marrow transplantation in pediatric patients: a single centre retrospective analysis. Bone Marrow Transplant. 1999;23(10):1049-1053.

Haematologica | 107 August 2022

1795

ARTICLE - Chronic Lymphocytic Leukemia

B-cell antigen receptor expression and phosphatidylinositol 3-kinase signaling regulate genesis and maintenance of mouse chronic lymphocytic leukemia Vera Kristin Schmid,1 Ahmad Khadour,1 Nabil Ahmed,2 Carolin Brandl,3 Lars Nitschke,3 Klaus Rajewsky,4 Hassan Jumaa1 and Elias Hobeika1 Institute of Immunology, University Medical Center Ulm, Ulm; Uniklinik RWTH Aachen, Aachen; 3Division of Genetics, Department of Biology, Friedrich Alexander University Erlangen-Nürnberg, Erlangen and 4Max Delbrück Center for Molecular Medicine, Berlin, Germany 1

Correspondence: Elias Hobeika elias.hobeika@uni-ulm.de Received: September 1, 2021. Accepted: January 4, 2022. Prepublished: January 13, 2022. https://doi.org/10.3324/haematol.2021.279924 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

Abstract Chronic lymphocytic leukemia (CLL) is a frequent lymphoproliferative disorder of B cells. Although inhibitors targeting signal proteins involved in B-cell antigen receptor (BCR) signaling constitute an important part of the current therapeutic protocols for CLL patients, the exact role of BCR signaling, as compared to genetic aberration, in the development and progression of CLL is controversial. In order to investigate whether BCR expression per se is pivotal for the development and maintenance of CLL B cells, we used the TCL1 mouse model. By ablating the BCR in CLL cells from TCL1 transgenic mice, we show that CLL cells cannot survive without BCR signaling and are lost within 8 weeks in diseased mice. Furthermore, we tested whether mutations augmenting B-cell signaling influence the course of CLL development and its severity. The phosphatidylinositol-3-kinase (PI3K) signaling pathway is an integral part of the BCR signaling machinery and its activity is indispensable for B-cell survival. It is negatively regulated by the lipid phosphatase PTEN, whose loss mimics PI3K pathway activation. Herein, we show that PTEN has a key regulatory function in the development of CLL, as deletion of the Pten gene resulted in greatly accelerated onset of the disease. By contrast, deletion of the gene TP53, which encodes the tumor suppressor p53 and is highly mutated in CLL, did not accelerate disease development, confirming that development of CLL was specifically triggered by augmented PI3K activity through loss of PTEN and suggesting that CLL driver consequences most likely affect BCR signaling. Moreover, we could show that in human CLL patient samples, 64% and 81% of CLL patients with a mutated and unmutated IgH VH, respectively, show downregulated PTEN protein expression in CLL B cells if compared to healthy donor B cells. Importantly, we found that B cells derived from CLL patients had higher expression levels of the miRNA-21 and miRNA-29, which suppresses PTEN translation, compared to healthy donors. The high levels of miRNA-29 might be induced by increased PAX5 expression of the B-CLL cells. We hypothesize that downregulation of PTEN by increased expression levels of miR-21, PAX5 and miR-29 could be a novel mechanism of CLL tumorigenesis that is not established yet. Together, our study demonstrates the pivotal role for BCR signaling in CLL development and deepens our understanding of the molecular mechanisms underlying the genesis of CLL and for the development of new treatment strategies.

Introduction Chronic lymphocytic leukemia (CLL) is the most frequent type of leukemia in Western countries.1 Like most neoplastic B-cell malignancies, CLL cells maintain their Bcell antigen receptor (BCR) expression.2 This selective

pressure to maintain a functional BCR is linked to the fact that malignant B cells profit from the proliferation and survival signals triggered by the BCR.3 Several lines of evidence support a key role of BCR signaling in the pathogenesis of CLL. Thus, CLL with hyper-mutated immunoglobulin heavy chain variable region (IgH VH) genes

Haematologica | 107 August 2022

1796

ARTICLE - BCR and PI3K in CLL cells

V.K. Schmid et al.

(mutated [M]-CLL) show a more favorable prognosis than those with unmutated IgH VH genes (unmutated [U]-CLL).4 Indeed, patients with fewer than 2% mutations in the IgH VH genes present a more polyreactive BCR and have a more aggressive disease with shorter survival.5 Furthermore, the IgH VH repertoire is highly restricted leading to different groups of CLL patients with stereotypic BCR.6 In contrast to diffuse large B-cell lymphoma (DLBCL), the BCR-activating signaling is not due to mutations in BCR signaling components7 but to autonomously activated BCR signaling initiated by the ability of CLL BCR to bind each other.8 More recently the crystal structures of a set of CLL BCR identified the regions involved in this binding. Small molecule inhibitors against BTK, like ibrutinib and acalabrutinib show anti-tumor activity in clinical studies of relapsed/refractory CLL.9 In line with this, the inhibition of BTK kinase activity through targeted genetic inactivation and inhibition of BTK by ibrutinib in Eµ-TCL1 mouse models significantly delays the outbreak of CLL, demonstrating that BCR signaling is critical for CLL development and expansion.10 The remarkable clinical effectiveness of BCR signaling inhibitors underscores the importance of BCR signaling and of BCR-associated kinases in the proliferation and homing of CLL cells, making this class of agents the treatment of choice for CLL patients.11 The prolonged survival of CLL cells is in part associated with defective apoptosis triggered by the phosphatidylinositol 3-kinase (PI3K)/ protein kinase B (PKB/AKT) and NF-κB pathways, which, among other pathways, are downstream of the BCR.12 PI3K exerts its effects by generating phosphatidylinositol-3,4,5-trisphosphate (PI(3,4,5)P3) via phosphorylating the 3-position of the inositol ring of phosphatidylinositol-4,5-bisphosphate (PI(4,5)P2).13 On the contrary, the phosphatase and tensin homolog deleted on chromosome 10 (PTEN) can dephosphorylate PI(3,4,5)P3 at the 3-position of the inositol ring converting PI(3,4,5)P3 back to PI(4,5)P2.13,14 Loss of PTEN results in accumulation of PI(3,4,5)P3 mimicking the effect of PI3K activation and triggering the activation of its downstream effector AKT.15 TCL1 is an oncoprotein contributing to occurrence of Tcell prolymphocytic leukemia, as a result of chromosomal translocations and inversions at 14q31.2.16 Although such a defect is not found in CLL, TCL1 is expressed in more than 90% human CLL patients.17 In order to facilitate the development of novel therapeutics for B-cell malignancies, an in vivo model, which recapitulates the human disease, is required. Several mouse models have provided important insights into CLL pathogenesis.18 These particularly include the widely studied Eµ-TCL1 model.19 Aging transgenic mice that overexpress TCL1 under the control of the µ immunoglobulin (Igµ) gene enhancer, develop a CD5-positive B-cell lymphoproliferative disorder mimicking human CLL and im-

plicating TCL1 in the pathogenesis of CLL.20 Given the importance of intrinsic BCR signaling in survival and progression of CLL, the establishment of a mouse model that can provide a genetic answer to the importance of the BCR in CLL would be a benefit for the understanding of its pathogenesis. In this study, we investigated the role of the BCR in the development of a CLL-like B-cell tumor disease in the mouse. The transformed B cells are here referred to as CLL cells. We also explored the impact of the PI3K signaling on the progression of the disease and found that Ptendeletion accelerated the onset of leukemogenesis. Moreover, we revealed that PTEN is downregulated in human CLL patients, which might be caused by increased expression levels of the microRNA family miR-29 and the protein PAX5. Thus, by generating inducible mouse models allowing the inactivation of BCR components we established a tool for the investigation of CLL signal transduction and treatment modalities. Our present data identify the BCR as a uniquely important regulator of CLL viability, confirm that an increased PI3K-signaling pathway supports CLL development and maintenance and describe PTEN as a potential target for therapeutic intervention.

Methods Mouse models In order to delete the mb-1 gene, encoding the Igα protein in B cells, the previously described mouse strain IgαTMF (Igαfl/fl),21 was crossed with the mb1-CreERT2 strain (Cd79atm3(cre/ERT2)Reth), which expresses a tamoxifen-inducible form of the Cre recombinase under the control of the mb-1 promoter region.22 Mb1-CreERT2; Igαfl/fl mice were crossed to the Eµ-TCL1 mouse strain20 to generate the compound mouse mb1CreERT2;Igαfl/fl; Eµ-TCL1. The mb1-CreERT2;Eµ-TCL1 served as a control. The Ptenfl/fl mouse strain23 was crossed with the mb1CreERT2;Eµ-TCL1. Mice heterozygous for Pten (Ptenfl/+) were generated by crossing Ptenfl/fl mice24 to the mb1-Cre mouse (Cd79atm1(cre)Reth/Ehobj) to generate the mb1Cre;Ptenfl/+ mouse strain. Treatment of the mice with antiIL7R antibody (Ab) and tamoxifen as well as transfer and propagation of CLL cells in Rag2−/−;γc−/− mice is described in the Online Supplementary Appendix. All mice were bred on Black6 background. All animal studies were carried out in accordance with the German Animal Welfare Act, having been reviewed by the regional council and approved under the license (#1288 L1-L18). Culture of splenic cells The splenic cells of Tam-treated mb1-CreERT2;Ptenfl/fl;EµTCL1 mice were cultured in complete medium (ISCOVE’s,

Haematologica | 107 August 2022

1797

ARTICLE - BCR and PI3K in CLL cells

V.K. Schmid et al.

10% fetal calf serum, 50 mg/mL gentamycin, 50 mM 2mercapto-ethanol) without any additional growth factors (like BAFF) and incubated at 37°C in the presence of 7.5% CO2. Only in the case of ex vivo inactivation of PTEN expression in purified primary splenic B2 cells from mb1CreERT2;Ptenfl/fl;Eµ-TCL1 mice and mb1-CreERT2;Eµ-TCL1 controls, the cells were cultured in the presence of 50 ng/mL recombinant human BAFF and were treated with 1 µM Tam (4-OHT). Flow cytometry For flow cytometric analysis, cell suspensions were pretreated with α-CD16/CD32 Fc Block (2,4G2; BD Biosciences). Dead cells were excluded by staining with Fixable Viability Dye eFluor 450 (eBioscience). Intracellular (IC) flow cytometry staining was performed using the ADG Fix&Perm Kit (Dianova). The detailed IC and extracellular staining procedure including the respective antibodies is provided in the Online Supplementary Appendix. Cells were acquired at a FACS Canto II flow cytometer (BD Biosciences). Analysis was performed using the FlowJo software (Tree Star). Ca2+ influx measurement 0.5–1×106 cells preloaded with the calcium-sensitive fluorescent dye Indo-1 (Invitrogen) were analyzed by flow cytometry (LSR Fortessa, BD Biosciences) upon application of 10 µg/mL anti-IgM F(ab')2 fragments (Jackson ImmunoResearch). Chronic lymphocytic leukemia patient and healthy donor sample analysis CLL samples were obtained from the Department of Internal Medicine III, University Hospital Ulm. Peripheral blood mononuclear cells from healthy donors (HD) were obtained from the Institute for Clinical Transfusion Medicine and Immunogenetics at Ulm University Medical Center. All samples were obtained with informed consent and used in full compliance with institutional regulations (no. 456/19). Primary CLL patient and HD lymphocytes were isolated from peripheral blood using Ficoll-Paque PLUS (GE Healthcare, 17-440-03). The human samples were MACSed for CD19-positive B cells before microRNA (miRNA) and total RNA isolation for real-time quantitative polymerase chain reaction (qRT-PCR) analysis. More details are provided in the Online Supplementary Appendix. Statistical analysis Unpaired two-tailed Student’s t-tests (with n between 3 and 5 mice per group) were carried out using Prism 9 software (GraphPad Software Inc) to determine the statistical differences between groups.

Results Efficient inducible deletion of the Igα-encoding mb-1 gene in B cells of mb1-CreERT2; Igαbfl/fl mice In order to investigate the efficiency of mb1-CreERT2 in the deletion of Igα in mature B cells we crossed mb1-CreERT2 with Igαfl/fl mice, in which the exon 3 and 4 of the mb-1 gene are flanked by loxP sites (floxed). All splenic B cells isolated from the tamoxifen (Tam)-treated mb1CreERT2;Igαfl/fl mice lacked both IgM and IgD expression while splenic B cells from the mb1-CreERT2 mice treated in the same manner still expressed the BCR (Figure 1A and B). Furthermore, 8 weeks after the start of the Tam-treatment and additional treatment with anti-IL-7R Ab, which block the influx of newly generated B cells from the bone marrow (BM), absolute splenic B-cell numbers from mb1CreERT2;Igαfl/fl mice were significantly reduced (up to 40×) when compared to those from the spleens of the mb1CreERT2 mice (Figure 1C). Additionally, peripheral blood (PBL), lymph nodes (LN) and peritoneal cavity (PC) of these mice contained only a few surviving B cells (Figure 1D). In line with other studies these results demonstrate that mature B cells absolutely require the expression of Igα and consequently of the BCR for their survival in the periphery. Inactivation of Igα in mouse chronic lymphocytic leukemia reverts the disease phenotype Although inhibitors against prominent signaling molecules downstream of the BCR, like ibrutinib and idelalisib, successfully eradicate CLL cells in patients, pointing to the quintessential role of the BCR in CLL cell survival, there is no genetic evidence for the involvement of the BCR in the maintenance of CLL cells. Therefore, we generated a mouse model, which enables us to address this question. In order to investigate the role of the BCR in mouse CLL we crossed the mb1-CreERT2; Igαfl/fl mice with the Eµ-TCL1 mouse strain, which is a well-accepted model for mouse CLL, generating mb1-CreERT2;Igαfl/fl;Eµ-TCL1 mice. Overexpression of TCL1 in B cells drives the development of CLL cells over time. The disease is first detected in 6month-old mice and its incidence increases with age, reaching its maximum at 12 months. The main features of mouse CLL are the expression of CD5 (Figure 2A; right panel) and the deregulation of B220 and IgD expression. Before and after the beginning of Tam treatment, we assessed the development and survival of CLL cells in the blood of the transgenic mice by flow cytometry using the mouse CLL key markers CD19+ CD93- B220low CD5+ IgM+ IgD−. Fully diseased mb1-CreERT2;Igαfl/fl;Eµ-TCL1 and control mb1-CreERT2; Eµ-TCL1 (12-month-old) carrying mainly CLL cells (Figure 2A) were sacrificed 8 weeks after the beginning of the Tam treatment and without additional treatment with anti-IL-7R Ab. The frequencies and abso-

Haematologica | 107 August 2022

1798

ARTICLE - BCR and PI3K in CLL cells

V.K. Schmid et al.

Figure 1. Efficient deletion of the Igα gene leads to the loss of mature B cells. (A) Flow cytometric analysis of B cells from the spleens of mb1-CreERT2;Igαfl/fl (left) and mb1-CreERT2 control mice (right) treated with tamoxifen (Tam) as described in the Materials and Methods section. Shown are dot plots of the anti-IgM vs. anti-IgD staining after gating on CD19+CD93− mature B cells, 2 weeks post Tam treatment. The gated regions in the dot plots correspond to mature follicular (MF) (IgMlow IgDhigh) and marginal zone (MZ) (IgMhighIgDlow) B cells. The numbers in the dot plots indicate the mean relative frequency of cells in the gate. (B) Quantification of the relative cell count of the Igα-deficient splenic B cells 2 weeks after the beginning of the Tam treatment: bars (left) represent % of cells obtained from mb1-CreERT2;Igαfl/fl mice and bars (right) from mb1-CreERT2 mice. Graphs are presented as mean ± standard error of the mean (SEM). Four asterisks (****) indicate P<0.0001, P-values were obtained using two-tailed Student's t-test. Cell numbers of 5 mice per group are shown. (C) Statistical analysis of absolute cell numbers of mature splenic B cells 8 weeks after Tam and anti-IL-7R treatment: filled bars represent cells obtained from mb1-CreERT2;Igαfl/fl mice and open bars from mb1-CreERT2 mice. Two asterisks (**) indicate P<0.01, P-values were obtained using two-tailed Student's t-test. Cell numbers of 5 mice per group are shown. (D) Flow cytometric analysis of B cells peripheral blood (PBL) (left), lymph nodes (LN) (middle) and peritoneal cavity PC (right) of mb1-CreERT2;Igαfl/fl mice treated with Tam and anti-IL-7R as described in the Materials and Methods section. Shown are dot plots of the anti-CD19 vs. CD93 staining. The numbers in the dot plots indicate the mean relative frequency of cells in the gate. Data shown are representative of 5 mice.

Haematologica | 107 August 2022

1799

ARTICLE - BCR and PI3K in CLL cells

V.K. Schmid et al.

lute B-CLL cell numbers in the spleen differed dramatically and significantly between mb1-CreERT2;Igαfl/fl;Eµ-TCL1 and mb1-CreERT2;Eµ-TCL1 control mice (Figure 2A and C). The 100-fold decrease of cellularity after deletion of Igα resulted in a significant reduction of the spleen size and weight in mb1-CreERT2;Igαfl/fl;Eµ-TCL1 mice compared to the Igα-sufficient controls (Figure 2B). In the PBL and the BM (Online Supplementary Figure S1A to C), the frequencies and absolute cell numbers of mouse CLL cells differed significantly between mb1-CreERT2;Igαfl/fl;Eµ-TCL1 and mb1-CreERT2;Eµ-TCL1 control mice. In order to further investigate the impact of Igα gene deletion on the overall survival of the diseased mice, immunodeficient Rag2−/−;γc−/− mice were intraperitoneally (i.p.) transplanted with splenic cells (1×107) from either mb1CreERT2;Igαfl/fl;Eµ-TCL1 or mb1-CreERT2;Eµ-TCL1 mice, which were previously sequentially transferred in Rag2−/−;γc−/−. The mice were treated with Tam three-times every third day, without additional treatment of anti-IL7R Ab, and were analyzed at day 15 after the start of the treatment. This model has the advantage of producing a CLL-like phenotype including peripheral blood leukemia and splenomegaly in a short period of time, compared to the long period in the original TCL1 model. While Rag2−/−;γc−/− mice transplanted with mb1-CreERT2;Eµ-TCL1 CLL cells died already between day 12 and 18, the mice transplanted with the mb1-CreERT2;Igαfl/fl; Eµ-TCL1 CLL cells survived up to day 30 (Figure 2D). Notably, as shown in the previous experiments, BCR-deficient CLL cells of the mb1-CreERT2;Igαfl/fl;Eµ-TCL1 genotype were significantly reduced in the spleens of the Tam-treated Rag2−/−;γc−/− mice (Figure 2E and F) as well as in the BM, PC (Figure 2F) and PBL (data not shown), while the CLL cells of the mb1-CreERT2;Eµ-TCL1 genotype showed massive accumulation in the spleens and also in the BM and PC (Figure 2F) of the recipient Rag2−/−;γc−/− mice leading to their early death. Additionally, the spleens from Rag2−/−;γc−/− mice transplanted with mb1-CreERT2;Igαfl/fl;EµTCL1-derived CLL cells were smaller compared to the control (Online Supplementary Figure S1D and E). Collectively, these findings show that ablation of the BCR in CLL cells is associated with reduced tumor size and increased overall survival demonstrating a clear dependence of CLL cells on BCR. PI3K activity is reduced in Igα-deficient chronic lymphocytic leukemia cells PI3K signaling is activated downstream of the BCR and is quintessential for the survival of healthy B cells.25 In order to investigate the effects of BCR deficiency on BCR signaling and the PI3K pathway in CLL cells, we assessed the phosphorylation of the PI3K target AKT and the phosphorylation of other signaling factors downstream of the BCR including LYN, SYK and BTK by intracellular flow cyto-

metry in CLL cells 10 days after induced Igα deletion. The time point of 10 days was selected, because we already observed a complete loss of Igα and BCR expression at this time but enough cells for flow cytometric analysis were still alive. Ten days after induced Igα deletion, the BCR-deficient CLL cells showed decreased AKT phosphorylation at S473 and T308 compared to the BCR-sufficient CLL control cells as well as significantly reduced LYN phosphorylation (Y396/Y507), SYK phosphorylation (Y525/526) and BTK phosphorylation (Y223) (Figure 3A and B). In addition, we analyzed whether the CLL cells could be stimulated by IgM F(ab')2 treatment after Igα deletion. Using a calcium influx assay, we observed that, in contrast to mb1-CreERT2; Eµ-TCL1 CLL control cells, CLL cells from mb1-CreERT2;Igαfl/fl;Eµ-TCL1 mice did not release Ca2+ upon BCR stimulation with IgM F(ab')2 fragments 10 days after Tam treatment (Figure 3E). Moreover, 5 minutes after stimulation with anti-IgM F(ab')2 fragments, the Igα-deficient CLL cells show less increase of SYK-phosphorylation at Y525 and Y526 compared to mb1-CreERT2;Eµ-TCL1 CLL control cells (Online Supplementary Figure S2G). The decrease in AKT phosphorylation after Igα deletion points to a reduced PI3K activity in the absence of the BCR. Moreover, the decreased phosphorylation of the BCR-proximal kinases LYN, SYK and BTK indicates that BCR signaling is downregulated in CLL B cells in consequence of Igα deletion. Constitutively active PI3K signaling may lead to increased BCL-2 expression.26 Therefore, we investigated the expression of BCL-2 in BCR-deficient and BCR-expressing B cells. BCRdeficient CLL cells from mb1-CreERT2;Igαfl/fl;Eµ-TCL1 had significantly decreased BCL-2 expression 10 days after Tam treatment as compared to control cells (Figure 3C and D). In line with this, we observed a slight downregulation of the anti-apoptotic protein myeloid cell leukemia sequence 1 (MCL-1) in CLL cells 10 days after induced Igα deletion (Online Supplementary Figure S2A and F; left). Considering that BCL2 is also regulated by NFκB, we tested whether ablation of the BCR in CLL cells from mb1-CreERT2;Igαfl/fl;Eµ-TCL1 mice also resulted in reduced NFκB activity by analyzing IKKα/b phosphorylation at S176/180 and phosphorylation of NFκB p65 at S536 10 days after induced Igα deletion. Interestingly, our analysis revealed that IKKα/b phosphorylation at S176/180, phosphorylation of NFκB p65 at S536 and thus NFκB activity is slightly downregulated in CLL cells with induced Igα deletion compared to Eµ-TCL1 CLL control cells (Online Supplementary Figure S2B, C and F). Together, these data suggest that BCR-mediated activation of PI3K signaling is essential for the survival of CLL B cells and that reduced NFκB activity and BCL-2 upregulation may be an important part of this regulation. PTEN-loss augments PI3K activity and results in early onset of chronic lymphocytic leukemia Next, we explored the role of the PI3K signaling pathway

Haematologica | 107 August 2022

1800

ARTICLE - BCR and PI3K in CLL cells

V.K. Schmid et al.

Continued on following page. Haematologica | 107 August 2022

1801

ARTICLE - BCR and PI3K in CLL cells

V.K. Schmid et al.

Figure 2. The B-cell antigen receptor (BCR) is indispensable for the survival of mouse chronic lymphocytic leukemia cells and inactivation of the BCR prolongs the survival of adoptive transfer recipient mice. (A) Flow cytometric analysis of B cells from the spleens of mb1-CreERT2;Eµ-TCL1 control mice and mb1-CreERT2;Igαfl/fl;Eµ-TCL1 mice before and 8 weeks after tamoxifen (Tam) treatment. Shown are dot plots of the anti-B220 vs. anti-CD5 staining after gating on CD19+ CD93- B cells (gating shown). The gated regions in the anti-B220 vs. anti-CD5 dot plots correspond to chronic lymphocytic leukemia (CLL) cells (CD19+ CD93B220low CD5+ IgM+ IgD−) and mature healthy B cells (CD19+ CD93- B220+ CD5− IgM+ IgD+). The numbers in the dot plots indicate the mean relative frequency of cells in the gate. (B) Spleen (SP) pictures (left) and quantification of the SP weight (right) obtained from mb1-CreERT2;Igαfl/fl;Eµ-TCL1 mice and mb1-CreERT2;Eµ-TCL1 control mice 8 weeks after the beginning of the Tam treatment (left). The graphs (right) are presented as mean ± standard error of the mean (SEM). Four asterisks (****) indicate P<0.0001, Pvalues were obtained using two-tailed Student's t-test. Cell numbers of 5 mice per group are shown (right). (C) Statistical analysis of absolute cell numbers of CLL cells 8 weeks after Tam treatment: filled circles indicating cells obtained from mb1CreERT2;Igαfl/fl;Eµ-TCL1 mice (left) and from mb1-CreERT2;Eµ-TCL1 mice (right). Four asterisks (****) indicate P<0.0001, P-values were obtained using two-tailed Student's t-test. Cell numbers of 10 mice per group are shown with each circle representing an individual animal. (D) Kaplan-Meier survival curve showing the survival of recipient Rag2−/−;γc−/− mice transferred with 107 splenic CLL-like B cells derived from the mb1-CreERT2;Eµ-TCL1 control or the mb1-CreERT2;Igαfl/fl mice and treated with Tam at day 3 post transfer. Every point represents an individual mouse with n=10. The P-value (P<0.0001) was determined by Mantel-Cox logrank test. (E) Flow cytometric analysis of B cells from SP of Rag2−/−;γc−/− mice derived from the mb1-CreERT2;Igαfl/fl;Eµ-TCL1 (left) and the mb1-CreERT2;Eµ-TCL1 (right) mice +Tam. Shown are dot plots of the anti-CD19 vs. anti-CD5 staining at day 10 post transfer. The gated regions in the dot plots correspond to the individual B-cell populations. The numbers in the dot plots indicate the mean relative frequency of cells in the gate. (F) Statistical analysis of absolute cell numbers from SP (left), bone marrow (BM, middle) and peritoneal cavity (PC, right) of Rag2−/−;γc−/− mice transplanted with mb1-CreERT2;Igαfl/fl;Eµ-TCL1 (left bar) or mb1CreERT2;Eµ-TCL1 (right bar) mouse CLL B cells. Graphs are presented as mean ± SEM. Four asterisks (****) indicate P<0.0001, Pvalues were obtained using two-tailed Student's t-test. Cell numbers of 5 mice per group are shown.

in the onset and maintenance of mouse CLL by inactivating PTEN, the negative regulator of the PI3K signaling pathway. To this end, we generated a mouse model with a B cell-specific and Tam-induced deletion of the Pten gene. In order to investigate the role of PTEN in the onset of mouse-CLL, we crossed the mb1-CreERT2;Eµ-TCL1 mice to the Ptenfl/fl mouse strain to generate mb1CreERT2;Ptenfl/fl;Eµ-TCL1 mice. These mice were treated with Tam at a young age (6 weeks) before the detection of any CLL B cells in the peripheral blood or the outbreak of mouse CLL that are usually evident at 3 to 6 months in mb1-CreERT2;Ptenfl/fl;Eµ-TCL1 and mb1-CreERT2;Eµ-TCL1 mice, respectively. We assessed the development of CLL cells using the mouse CLL key markers CD19+ B220low CD5+ IgM+ IgD−. mb1-CreERT2;Eµ-TCL1 mice served as controls. Eight and 16 weeks after the beginning of the Tam treatment the mice were analyzed by flow cytometry for the absolute cell number of CD19+ B220low CD5+ IgM+ IgD− B cells. At both time points the number of CLL cells was increased (8 weeks: >2×106, Figure 4A; 16 weeks: >2×107, Figure 4B) in the mb1-CreERT2; Ptenfl/fl;Eµ-TCL1 mice as compared to those from the spleens of the mb1CreERT2;Eµ-TCL1 control mice. These data show that after loss of PTEN, the resulting constitutive activity of the PI3K pathway leads to accelerated development of CLL in young mice as shown by the accumulation of CLL B cells with time. Recent studies reported that either germline deletion27 or insertion of a single mutation28 in the Tp53 gene (encoding the tumor suppressor p53), a gene, which is often mutated in CLL,29 on Eµ-TCL1 background also accelerated the development of the CLL disease. We deleted the Tp53 gene in mb1-CreERT2;Eµ-TCL1 mice and found out that 24 weeks post treatment 94 % of B cells in the spleens of the mb1-CreERT2;Ptenfl/fl;Eµ-TCL1 mice became

CLL cells, whereas the mb1-CreERT2;Tp53fl/fl;Eµ-TCL1 mice showed 2% of CLL cells (Figure 4C). By assessing the expression of p53 by flow cytometry we confirmed that Tp53 was indeed deleted in mb1-CreERT2;Tp53fl/fl;Eµ-TCL1 mice (Figure 4D). In addition, we compared the development of CLL in mice lacking Tp53 expression (mb1-Cre;Tp53fl/fl;EµTCL1) to mice with a constitutive heterozygous loss of Pten (mb1-Cre;Ptenfl/+;Eµ-TCL1). At 32 weeks, the mb1Cre;Ptenfl/+;Eµ-TCL1 mice showed 94% of CLL cells whereas mb1-Cre;Tp53fl/fl;Eµ-TCL1 mice developed only 27% of CLL B cells in the blood (Figure 4E). Moreover, a combined deletion of Tp53 and Pten in mb1Cre;Ptenfl/+;Tp53fl/fl;Eµ-TCL1 mice did not accelerate the outbreak of CLL when compared to mb1-Cre;Ptenfl/+;EµTCL1 mice (Online Supplementary Figure S3A). This confirms that the accelerated development of the disease is specifically driven by the partial loss of PTEN and the subsequent activation of the PI3K pathway. As further signs of elevated PI3K signaling, we found an increased phosphorylation of S473 of AKT and Y223 of BTK in CD19+ B220low CD5+ IgM+ IgD− CLL cells from the spleens of mb1-CreERT2;Ptenfl/fl;Eµ-TCL1 mice compared to CLL cells from the mb1-CreERT2;Eµ-TCL1 control mice (Figure 4F; Online Supplementary Figure S3 D and E). In order to investigate, whether PTEN deficiency resulted in constitutive activation of the PI3K/AKT pathway, we stimulated splenic CLL cells from mb1-CreERT2;Ptenfl/fl;Eµ-TCL1 mice and mb1-CreERT2;Eµ-TCL1 control mice with anti-IgM F(ab')2 fragments and monitored the AKT phosphorylation at S473 and T308 as well as BTK phosphorylation at Y223 after 5 minutes of stimulation. In both cases, we could not observe any significant increase in phosphorylation after BCR stimulation indicating that both, wild-type TCL1 leukemia cells and PTEN-deficient leukemia cells exhibit

Haematologica | 107 August 2022

1802

Haematologica | 107 August 2022

1803

Figure 3. AKT phosphorylation and BCL-2 expression are reduced in Igα-deficient chronic lymphocytic leukemia cells. (A) Flow cytometric analysis of AKT phosphorylation (S473) and (T308), LYN phosphorylation (Y396) and (Y507), SYK phosphorylation (Y525/526), BTK phosphorylation (Y223) on splenic chronic lymphocytic leukemia (CLL) cells from tamoxifen (Tam)-treated mice (day 10): mb1-CreERT2;Eµ-TCL1 (red line) and mb1-CreERT2;Igαfl/fl;Eµ-TCL1 (blue line). (B) Quantification of the intracellular mean fluorescence intensity (MFI) in splenic mouse CLL B cells from mb1-CreERT2;Igαfl/fl;Eµ- TCL1 (blue filled bars) and mb1-CreERT2;Eµ-TCL1 (red filled bars) mice. MFI of AKT phosphorylation (S473) and (T308), LYN phosphorylation (Y396) and (Y507), SYK phosphorylation (Y525/526) and BTK phosphorylation (Y223) are shown. Graphs are presented as mean ± standard error of the mean (SEM). P-values were obtained using the two-tailed Student's t-test (**P<0.01; ***P<0.001; ****P<0.0001). Results from 5 mice per group are shown. (C) Flow cytometric analysis of BCL-2 expression on splenic CLL cells from Tam-treated mice (day 10): mb1CreERT2;Eµ-TCL1 (red line) and mb1-CreERT2;Igαfl/fl;Eµ-TCL1 (blue line). (D) Quantification of the intracellular BCL-2 MFI in splenic mouse CLL B cells from mb1CreERT2;Igαfl/fl;Eµ-TCL1 (blue filled bars) and mb1-CreERT2;Eµ-TCL1 (red filled bars) mice. Graphs are presented as mean ± SEM. Four asterisks (****) indicate P<0.0001, P-values were obtained using two-tailed Student's t-test. Results from five mice per group are shown. (E) Flow cytometric analysis of the intracellular Ca2+ influx mature splenic B cells from the Tam-treated mb1-CreERT2;Igαfl/fl;Eµ-TCL1 (blue line)) and mb1-CreERT2;Eµ-TCL1 (red line) after treatment with 10 µg/mL anti-IgM F(ab')2 fragments. Data shown are representative of 3 independent experiments.

ARTICLE - BCR and PI3K in CLL cells V.K. Schmid et al.

ARTICLE - BCR and PI3K in CLL cells

V.K. Schmid et al.

a constitutive active PI3K/AKT signaling, that cannot be significantly increased by BCR stimulation (Online Supplementary Figure S3B and C). Additionally, BCL-2 expression was higher in PTEN-deficient CD19+ B220low CD5+ IgM+ IgD− cells relative to the control (Figure 4F, right; Online Supplementary Figure S3F). Statistical analysis confirmed that the increased AKT- and BTK-phosphorylation as well as the increased BCL-2 expression in the PTEN-deficient CLL B cells was significant compared to the PTEN-sufficient control (Figure 4F). Furthermore, we affirmed by flow cytometry that the Pten gene was efficiently deleted in splenic CLL cells from mb1-CreERT2;Ptenfl/fl;Eµ-TCL1 mice and mb1-CreERT2;EµTCL1-derived cells served as a control (Online Supplementary Figure S3G). In conclusion, in the absence of PTEN, the CLL cells exhibit significant increase in AKT and BTK phosphorylation likely leading to higher activity of these molecules. These findings provide evidence for enhanced PI3K signaling in these PTEN-deficient cells. Furthermore, in this model, conditional inactivation of PTEN resulted in the accelerated onset of CLL. Deletion of Pten leads to autonomous survival of chronic lymphocytic leukemia cells In order to test the tumorigenic potential of the primary

splenic CD19+ B220low CD5+ IgM+ IgD− PTEN-deficient CLL cells from Tam-treated mb1-CreERT2;Ptenfl/fl;Eµ-TCL1 mice, we transferred them into Rag2−/−;γc−/− mice. Two weeks post engraftment, a population of CD19+ B220low CD5+ IgM+ IgD− cells was detected in the spleen (Figure 5A). Surprisingly, the primary splenic PTEN-deficient CLL cells from Tam-treated mb1-CreERT2;Ptenfl/fl;Eµ-TCL1 mice grew autonomously in culture and preserved the features of mouse CLL cells retaining their CD19+ B220low CD5+ IgM+ IgD− phenotype. The cells were cultured in a medium supplemented only with fetal calf serum and in the absence of any additional growth factors or antigen (Figure 5B). The PTEN deficiency of these cell lines was confirmed by genotyping and flow cytometry (Figure 5C; Online Supplementary Figure S4A, right). We generated five different CLL-like cell lines, which were heterogenous, but grew similarly in culture over prolonged period of time without growth factor supplements. After growing in culture for 4 months, functional analysis revealed that some of these cell lines mobilized intracellular Ca2+ release upon stimulation with a polyclonal antiIgM F(ab')2 antibody (Figure 5D). CD19+ B220low CD5+ IgM+ IgD− CLL cells phenotypically resemble B-1 B cells. Since BCR of some B1 B cells are specific to phosphatidyl choline (Ptc), we assessed by flow cytometry whether the

Continued on following page. Haematologica | 107 August 2022

1804

ARTICLE - BCR and PI3K in CLL cells

V.K. Schmid et al.

Figure 4. Deletion of Pten but not Tp53 leads to accelerated mouse chronic lymphocytic leukemia.(A and B) Statistical analysis of absolute cell numbers of chronic lymphocytic leukemia (CLL) cells (CD19+ B220low CD5+) (A) 8 weeks and (B) 16 weeks after tamoxifen (Tam) treatment of 6-week-old mice: filled circles indicating cells obtained from mb1-CreERT2;Ptenfl/fl;Eµ-TCL1 mice (left) and from mb1-CreERT2;Eµ-TCL1 mice (right). Four asterisks (****) indicate P<0.0001, P-values were obtained using twotailed Student's t-test. Cell numbers from 5 mice per group are shown with each circle representing an individual animal. (C) Flow cytometric analysis of B cells from the spleens of mb1-CreERT2;Ptenfl/fl;Eµ-TCL1 (left ), and mb1-CreERT2; Tp53fl/fl;Eµ-TCL1 (right) 24 weeks after Tam treatment of 6-week-old mice. Shown are dot plots of the anti-B220 vs. anti-CD5 staining after gating on CD19+ B cells. The gated regions in the dot plots correspond to CLL cells (CD19+ B220low CD5+) and healthy B cells (CD19+ B220+ CD5−). The numbers in the dot plots indicate the mean relative frequency of cells in the gate. (D) Flow cytometric analysis of p53-expression in splenic CLL cells from Tam-treated mb1-CreERT2;Tp53+/fl;Eµ-TCL1 mice (red line) and splenic CLL cells from mb1-CreERT2;Eµ-TCL1 mice (blue line). (E) Flow cytometric analysis of B cells isolated from the blood of mb1-Cre;Ptenfl/+;Eµ-TCL1 and mb1-Cre;Tp53fl/fl;Eµ-TCL1 mice at the age of 10 weeks and 32 weeks. The graph shows dot-plots of the anti-B220 vs. antiCD5 staining after gating on viable CD19+ B lymphocytes. The numbers in the dot plots indicate the mean relative frequency of cells in the gate. (F) Quantification of p-AKT (left) and p-BTK (middle), and BCL-2 (right) mean fluorescence intensity (MFI) in splenic mouse CLL B cells from mb1-CreERT2;Ptenfl/fl;Eµ-TCL1 (blue filled bars) and mb1-CreERT2;Eµ-TCL1 (red filled bars) mice. Mean ± standard deviation are shown. Four asterisks (****) indicate P<0.0001, P-values were obtained using two-tailed Student's t-test. Results from 5 mice per group are shown.

CLL-like CreERT2;Ptenfl/fl;Eµ-TCL1 cells were capable of binding Ptc liposomes. We found that some PTEN-deficient CLL-like cell lines could not bind Ptc with their BCR (Online Supplementary Figure S4A, blue line) as compared to PTEN-sufficient CLL-like CreERT2;Eµ-TCL1 control cells (Online Supplementary Figure S4A, red line). In addition, we concluded that PTEN-deficient CLL-like cell lines were

polyclonal because we detected multiple Vb-Jb recombination events within one population (Online Supplementary Figure S4B). Taken together, constitutive activation of PI3K signaling through loss of PTEN accelerates CLL development and allows efficient engraftment and maintenance of CLL-like B cell in vivo and in vitro. Because BCL-2 expression was increased in splenic PTEN-

Haematologica | 107 August 2022

1805

Haematologica | 107 August 2022

1806

B F

Figure 5. Deletion of Pten leads to autonomous survival of chronic lymphocytic leukemia B cells ex vivo. (A) Flow cytometric analysis of B cells from the spleens of Rag2−/−;γc−/− mice transplanted with splenic B cells from mb1-CreERT2;Ptenfl/fl;Eµ-TCL1 derived from the experiment depicted in (Figure 4C, left). Dot plot of the antiCD19 vs. anti-CD5 (left) and anti-IgD vs. anti-IgM (right) staining is shown. (B) Flow cytometric analysis of B cells from the spleens of mb1-CreERT2;Ptenfl/fl;Eµ-TCL1 cultured ex vivo. Dot plots of the anti-B220 vs. anti-CD5 (left) and anti-IgM vs. anti-IgD (right) staining after gating on CD19+ B cells. The gated regions in the dot plots correspond to chronic lymphocytic leukemia (CLL) cells (CD19+ B220low CD5+) and healthy mature B cells (CD19+ B220+ CD5−). The numbers in the dot plots indicate the mean relative frequency of cells in the gate. (C) Analysis of the Pten locus recombination in genomic DNA (gDNA) from the mb1-CreERT2;Ptend/d;Eµ-TCL1 cells in comparison to mb1-CreERT2;Ptenfl/fl;Eµ-TCL1, mb1-CreERT2;Ptenfl/+;Eµ-TCL1, and mb1-CreERT2;Eµ-TCL1. Left panel: floxed (fl) and wt (+) alleles. Right panel: the recombined deleted (d) allele. (D) Flow cytometric analysis of the intracellular Ca2+ influx in cultured splenic cells from the tamoxifen (Tam)-treated mb1CreERT2;Ptenfl/fl;Eµ-TCL1. Cells were treated with 10 µg/mL anti-IgM F(ab')2 fragments (red line) or left untreated (blue line). Data shown are representative of 3 independent experiments. (E) Survival plot of mb1-CreERT2;Ptend/d;Eµ-TCL1 (left) and mb1-CreERT2;Eµ-TCL1 (right) chronic lymphocytic leukemia (CLL)-like culture cells after treating the cells for 4 days with the vehicle control (dimethyl sulfoxide [DMSO]; black) or the inhibitors LY294002 (10 µM; white), zerumbone (10 µM, green), ibrutinib (1µM; red), R406 (5 µM; pink) and PP2 (10 µM; blue). Graphs are presented as mean ± standard error of the mean; dots indicate the percentage of survived cells normalized to vehicle control (100%). Three technical replicates per group are shown. (F) Flow cytometric analysis of proliferation dye intensity of labeled mb1CreERT2;Ptend/d;Eµ-TCL1 and mb1-CreERT2;Eµ-TCL1 culture cells at day 0 (black), day 2 (dark red), day 3 (red), day 4 (orange) after treatment with the inhibitors LY294002 (10 µM), zerumbone (10 µM), ibrutinib (1 µM), R406 (5 µM) and PP2 (10 µM).

ARTICLE - BCR and PI3K in CLL cells V.K. Schmid et al.

ARTICLE - BCR and PI3K in CLL cells

V.K. Schmid et al.

deficient CLL cells from mb1-CreERT2;Ptenfl/fl;Eµ-TCL1 mice, we treated the cells cultured for 4 months with the BCL-2 inhibitor ABT-199 (also known as venetoclax). The relative numbers of the PTEN-deficient as well as PTENsufficient cells were 5-fold decreased at a concentration of 10 µM compared to the vehicle (dimethyl sulfoxide [DMSO]) control (Online Supplementary Figure S4C). Moreover, in order to test whether the PTEN-deficient CLL-like cells in culture were still dependent on BCR signaling for survival and proliferation, we treated the cells with inhibitors blocking BCR downstream signaling and assessed the survival and proliferation of the cells in culture for 4 days. We used LY294002 (10 µM) as a selective PI3K inhibitor, zerumbone (10 µM) to block NFκB activity, ibrutinib (1 µM) to inhibit BTK activity, R406 (5 µM) as a SYK inhibitor and PP2 (10 µM) to block Src-family tyrosine kinases. All of the inhibitors significantly reduced the survival and proliferation of both mb1-CreERT2;Ptend/d;Eµ-TCL1 and mb1-CreERT2;Eµ-TCL1 culture cells when compared to the cells treated with vehicle control (DMSO) (Figure 5E and F). This indicates that the PTEN-deficient cells that are autonomously growing in culture are still dependent on BCR signaling and the activity of NFκB. Interestingly, we observed higher survival and proliferation of Pten-deleted CLL-like cells treated with inhibitors blocking BCR signaling (LY294002, ibrutinib, R406, PP2) when compared to the treated mb1-CreERT2;Eµ-TCL1 leukemic cells (Online Supplementary Figure S4D and E). Vice versa, mb1-CreERT2;EµTCL1 cells treated with the NFκB inhibitor zerumbone showed better survival and proliferation when compared to the PTEN-deficient mb1-CreERT2;Ptend/d;Eµ-TCL1 culture cells (Online Supplementary Figure S4D and E). In order to test whether the CD19+ B220low CD5+ IgM+ IgD− CLL cells could be generated if Pten was deleted ex vivo on Eµ-TCL1 background, we treated purified primary splenic B2 cells from mb1-CreERT2;Ptenfl/fl;Eµ-TCL1 mice and mb1-CreERT2;Eµ-TCL1 controls with Tam (4-OHT) in culture in the presence of the B-cell activating factor (BAFF). The ex vivo-generated PTEN-deficient CLL cells were more abundant compared to the PTEN-sufficient control (26% to 8%) (Online Supplementary Figure S5A and B). However, these cells could not survive for an extended time period in culture. Heterozygous loss of Pten accelerates chronic lymphocytic leukemia development In order to strengthen the evidence that PTEN-deficiency and subsequent enhanced PI3K-activity, monitored by AKT-phosphorylation, lead to accelerated onset of mouse CLL, we intercrossed the mb1-Cre;Eµ-TCL1 with the Ptenfl/+ strains to generate the mb1-Cre; Ptenfl/+;Eµ-TCL1 which allows for a constitutive heterozygous deletion of the Pten gene in B cells. Only one “floxed” Pten allele was introduced because constitutive Pten deletion on both al-

leles leads to a block of B-cell development at the proB-cell stage due to the inability to express a µHC.30 In these mice, we detected an early development of CLL cells at 8 weeks of age (Figure 6A). This was supported by the quantification of five independent experiments with five individual mice each showing that CLL cells accumulated in the spleens of young (8 weeks old) mb1Cre;Ptenfl/+;Eµ-TCL1 mice (Figure 6B, blue bar) but not in control mb1-Cre;Eµ-TCL1 mice (Figure 6B, red bar). The CD19+ B220low CD5+ CLL cells expressed only IgM and no IgD (Figure 6C, left) as compared to the CD19+ B220+ CD5− cell population (IgM+ IgD+) from the same mice (Figure 6C, right). Consistent with the heterozygous loss of Pten, the splenic CD19+ B220low CD5+ IgM+ IgD− CLL cells from mb1Cre;Ptenfl/+;Eµ-TCL1 displayed significantly more AKT phosphorylation than splenic B cells from the mb1Cre;Eµ-TCL1 mice (Figure 6D and E, left). Together, these results indicate that the constitutive loss of one Pten allele in combination with overexpression of TCL1 significantly accelerates the onset of CLL. Heterozygous Pten deletion does not lead to Richter’s transformation It was recently shown by Kohlhaas et al. that constitutive activation of AKT in Eµ-TCL1 mice results in Richter’s transformation (RT), an aggressive lymphoma which occurs upon progression from CLL.31 So, we investigated if mb1-Cre;Ptenfl/+;Eµ-TCL1 mice with a heterozygous Pten deletion develop RT and histologically analyzed spleens from diseased mice for features of RT. RT cells can be distinguished from CLL cells by morphological abnormalities, large lymphoid cells and increased proliferation. Histone H3 phosphorylation on S10 is specific to mitosis and phosphorylated histone H3 (PHH3) proliferation markers are increasingly being used to evaluate proliferation in various tumors.32 Therefore, we analyzed the number of PHH3-positive cells in spleens of diseased mb1Cre;Ptenfl/+;Eµ-TCL1 and mb1-Cre;Eµ-TCL1 control mice by fluorescence microscopy. However, no significant difference in PHH3 positive cells could be observed (Figure 6F; Online Supplementary Figure S5C) as well as no difference in size of the CLL cells (Online Supplementary Figure S5E). Hematoxylin and eosin (H&E) staining on paraffin embedded splenic sections of diseased mb1-Cre;Ptenfl/+;EµTCL1 and mb1-Cre;Eµ-TCL1 mice also revealed no significant changes after Pten deletion in splenic CLL cells (Online Supplementary Figure S5D). As a second marker for proliferation, we measured Ki-67 levels in splenic B cells of mb1-Cre;Ptenfl/+;Eµ-TCL1 and mb1-Cre;Eµ-TCL1 control mice by flow cytometric analysis. Again, no significant difference in Ki-67 expression levels could be observed (Figure 6E, right; Online Supplementary Figure S5F). So, we assume that loss of PTEN expression does not promote CLL transformation towards RT.

Haematologica | 107 August 2022

1807

ARTICLE - BCR and PI3K in CLL cells

V.K. Schmid et al.

Continued on following page. Haematologica | 107 August 2022

1808

ARTICLE - BCR and PI3K in CLL cells

V.K. Schmid et al.

Figure 6. A single deleted Pten allele is enough to accelerate the onset of chronic lymphocytic leukemia. (A) Flow cytometric analysis of B cells from the spleens of mb1-Cre;Ptenfl/+;Eµ-TCL1 mice. Dot plot of the anti-B220 vs. anti-CD5 staining after gating on CD19+ B cells. The gated regions in the dot plots correspond to chronic lymphocytic leukemia (CLL) cells (CD19+ B220low CD5+) and healthy mature B cells (CD19+ B220+ CD5−). The numbers in the dot plots indicate the mean relative frequency of cells in the gate. (B) Quantification of the relative number of (CD19+ B220low CD5+) splenic mouse B cells from mb1-Cre;Ptenfl/+;Eµ-TCL1 (blue filled bars) and mb1-Cre;Eµ-TCL1 (red filled bars) mice. Graphs are presented as mean ± standard error of the mean (SEM). Four asterisks (****) indicate P<0.0001, P-values were obtained using two-tailed Student's t-test. Results from 5 mice per group are shown. (C) Dot plot of the anti-IgM vs. anti-IgD staining after gating on CD19+ B220low CD5+ (left) and CD19+ B220+ CD5− (right). The gated regions in the dot-plots correspond to CLL cells (IgMhighIgD−) and healthy mature B cells (IgM+IgDhigh). The numbers in the dot-plots indicate the mean relative frequency of cells in the quadrants. (D) Flow cytometric analysis showing histograms overlays of AKT-phosphorylation (left) and size (right) of splenic B cells from mb1-Cre;Eµ-TCL1 (red line) and mb1-Cre;Ptenfl/+;EµTCL1 mice (blue line). (E) Quantification of AKT phosphorylation (left) and Ki-67 (right) mean fluorescence intensity (MFI) in splenic mouse CLL B cells from mb1-Cre;Ptenfl/+;Eµ-TCL1 (blue filled bars) and mb1-Cre;Eµ-TCL1 (red filled bars) mice. Graphs are presented as mean ± SEM. Four asterisks (****) indicate P<0.0001, P-values were obtained using two-tailed Student's t-test (****P<0.0001; ns=not significant). Results from 5 mice per group are shown. (F) Quantification of PHH3 positive cells of the splenic sections from diseased mb1-Cre;Ptenfl/+;Eµ-TCL1 and mb1-CreERT2; Eµ-TCL1 mice. Graphs are presented as mean ± SEM; bars show the mean value of 6 analyzed pictures per spleen. P-values were obtained using two-tailed Student's t-test (ns=not significant). Results from 4 mice per group are shown.

PTEN is downregulated in human chronic lymphocytic leukemia cells In order to test whether PTEN plays a role in the survival of human CLL, we analyzed its protein expression in B cells from human CLL blood samples. The patient-derived CLL samples were classified according to the presence (mutated, M-CLL) or absence (unmutated, U-CLL) of the mutation in the immunoglobulin heavy chain variable region gene (IGHV). An unmutated IGHV gene is a molecular marker in human CLL associated with poorer prognosis and shorter survival of the patients. As a control we used blood samples from HD (age 60+ years). Overall, we analyzed 14 M-CLL and 21 U-CLL samples as well as 16 HD samples for their PTEN protein expression in B cells using flow cytometry. Statistical analysis revealed a significantly decreased PTEN protein expression in CLL cells when compared to B cells from HD B1 and B2 (Figure 7B and E). In nine of 14 M-CLL (64.3 %) and 17 of 21 U-CLL samples (81 %) there was lower PTEN expression in CLL cells when compared to the mean of the HD B1 control cells (Figure 7B, E and F). The remaining CLL samples showed either similar or slightly higher PTEN protein levels relative to the mean of the HD samples (Figure 7F). As a control, the same samples showed no difference in size or when stained with the secondary Ab alone (Figure 7C). In order to assess whether the downregulation of PTEN protein expression in the 64.3 % M-CLL and 81 % U-CLL samples was controlled in a post-transcriptional or post-translational manner, we measured the PTEN mRNA transcript levels by RT-qPCR. The results revealed that the PTEN mRNA levels were significantly increased in 71.5 % of the M-CLL and in 73.3 % of the U-CLL samples when compared to the HD controls (Figure 7E and F). Therefore, we assumed that the PTEN protein levels in approximately two thirds of the analyzed CLL patients are likely downregulated by either translational repression via microRNA (miRNA) or in a post-translational manner. Indeed, the

repression of PTEN transcripts by miRNA has been already reported in many diseases.33,34 Downregulation of PTEN expression by miRNA-21, miRNA-29 and PAX5 Among others, the miRNA miR-21 is known to target the tumor suppressor PTEN as the knockdown of miR-21 in a DLBCL cell line resulted in increased PTEN protein expression but did not affect the level of PTEN mRNA.35 In order to address the miR-21 expression in human CLL, we performed miRNA isolation both from HD, M-CLL and UCLL patients. The analysis by RT-qPCR revealed a significant increase in the amount of miR-21 in M-CLL and U-CLL B cells compared to HD control B cells (Figure 8A). The miR-29 family of miRNA, consisting of three members miR-29a, miR-29b and miR-29c, is highly expressed in cells of the adaptive immune response and has also been shown to regulate PTEN expression leading to an increased PI3K activity.36 The ablation of miR-29 specifically in B lymphocytes results in an increase in PTEN expression and a decrease of the PI3K activity in mature B cells. We analyzed the expression levels of miR-29a, miR29b and miR-29c in the HD, M-CLL and U-CLL patients’ samples and could observe a significantly increased expression level of all three miR-29 family members in B cells from both M-CLL and U-CLL patients if compared to the HD controls (Figure 8B). Consequently, the expression of miR-21 and the miR-29 family was shown to be upregulated in human CLL cells, possibly accounting for the reduced PTEN protein levels in these cells. A recent study has reported that the expression of miR29 is controlled by PAX5.37 Hence, we analyzed the protein expression of PAX5 in B cells from M-CLL and U-CLL patients as well as HD controls by flow cytometry. Indeed, we could observe significantly higher PAX5 protein expression levels in B cells derived from CLL patients compared to HD (Figure 8C and D, left). As a control, the same samples did not show a difference in size or in the isotype

Haematologica | 107 August 2022

1809

ARTICLE - BCR and PI3K in CLL cells

V.K. Schmid et al.

Figure 7. PTEN expression is downregulated in human chronic lymphocytic leukemia B cells. (A and B) Flow cytometric analysis of PTEN expression (B) in human blood samples from healthy donors (HD) B1 cells (black), B2 cells (green) and mutated chronic lymphocytic leukemia (M-CLL, blue) and unmutated CLL (U-CLL, red) patients. The histograms indicate the fluorescence intensity of PTEN protein expression of representative samples after gating on CD19+ and CD5+ cells from the U-/M-CLL patients and HD B1 cells (A). For the HD B2 cells, we gated on CD19+ and CD5- cells (A, left). (C) Representative isotype control (C, left) and size control (C, right) of the flow cytometric analysis of PTEN expression in human blood samples from HD B1 cells (black), B2 cells (green) and M-CLL (blue) and U-CLL (red) patients. (D) Quantification of human blood samples from HD (age 60+) and CLL patients with and without a mutation in the IGHV gene (U-CLL and M-CLL). The plot indicates the mean fluorescence intensity (MFI) of PTEN protein expression after gating on CD19+ and CD5+ cells for the M-/U-CLL patients and the HD B1 cells. For the HD B2 cells we gated on CD19+ and CD5- cells. The bars include median and standard deviation. P-values were obtained using two-tailed Student's t-test (*P<0.05; **P<0.01; ***P<0.001). Every dot or triangle represents an individual. (E) Real-time quantitative polymerase chain reaction (qRT-PCR) analysis of human blood samples from HD and U-/M-CLL. The plot indicates the levels of PTEN mRNA expression relative to the expression of the housekeeping gene GAPDH, including median and standard deviation. P-values were obtained using two-tailed Student's t-test (**P<0.01; ****P<0.0001). Every dot or triangle represents an individual. (F) Relative PTEN mRNA and protein expression in B cells from U-/M-CLL patients compared to the averaged expression of HD control. Red: samples show high PTEN mRNA/protein expression relative to HD, black: PTEN expression is similar relative to HD, grey: CLL patients have lower PTEN/mRNA levels relative to HD.

control staining (Figure 8D, right). Together, these data strengthen our hypothesis, that PTEN expression in CLL cells are post-transcriptionally repressed by high levels of the miRNA miR-21 and miR-29, the second upregulated by increased levels of PAX5 itself (Figure 8E).

Discussion In this study we present a novel mouse model, which allows the Tamoxifen-inducible inactivation of Igα (subsequently preventing BCR assembly and expression on the cell surface) in a mouse CLL model.

Our findings demonstrate that loss of mature B cells after the ablation of the BCR in combination with the anti-IL7R treatment is an intrinsic feature of B cells and not due to their reduced production. Consistent with previous findings,25,38-40 we confirm here that mature peripheral B cells rely on their BCR for survival, as almost the complete B cell pool is lost within 2 months after Igα inactivation. We took advantage of the mb1-CreERT2 mouse strain’s efficient recombination of the Igα locus and applied it to the Eµ-TCL1 CLL mouse model, in which we could successfully ablate Igα. One major finding of this study is that in the Eµ-TCL1 mouse model, the maintenance of CLL cells requires BCR

Haematologica | 107 August 2022

1810

ARTICLE - BCR and PI3K in CLL cells

V.K. Schmid et al.

expression because in the absence of Igα 100-fold fewer CLL cells are maintained compared to the Igα-sufficient control. Based on our results, we suggest that in the absence of the BCR PI3K-signaling is reduced, as demonstrated by a decrease in AKT phosphorylation and lower levels of BCL-2 and MCL-1. In addition, the loss of BCR expression in CLL cells results in slightly downmodulated NFκB activity what might additionally contribute to the decreased BCL-2 expression levels. BCL-2 is important for the development and survival of mature naive B cells,41

and its overexpression partially rescues BCR-deficient B cells.42 We further show that ablation of the BCR reduces tumor size and prolongs the overall survival of mice with fully developed CLL. Srinivasan et al. have shown that BCR-dependent signaling via the PI3K provides the crucial “tonic signal”, which is indispensable for the maintenance of resting mature B cells.25 Although other studies have attributed the microenvironment with a role in CLL development and progression,43 our results demonstrate that, the BCR alone

Figure 8. miR-21, miR-29 and PAX5 expression is increased in human chronic lymphocytic leukemia B cells. (A) Expression of miR-21 in chronic lymphocytic leukemia (CLL) patients (grouped: mutated and unmutated IgH VH) and in the healthy donor (HD) control group (age: 60+ years) assessed by real-time quantitative polymerase chain reaction (qRT-PCR). The plot indicates the expression level of the microRNA (miRNA) miR-21 relative to the expression of the small-nucleolar RNA RNU44 used as endogenous control. Mean ± standard error of the mean (SEM); numbers in the graph indicate the mean of each group. P-values were obtained using Mann-Whitney U-test (**P<0.01; ***P< 0.001). Every dot or triangle represents an individual. (B) Expression of the miR-29 family in CLL patients (grouped: mutated and unmutated IgH VH) and in the HD control group (age: 60+ years) assessed by qRT-PCR. The plot indicates the expression levels of the miR-29 family members miR-29a, miR-29b, miR-29c relative to the expression of the small-nucleolar RNA RNU44 used as endogenous control. Mean ± SEM; numbers in the graph indicate the mean of each group. P-values were obtained using Mann-Whitney U-test (*P<0.05; **P<0.01; ***P<0.001; ****P<0.0001; ns=not significant). Every dot or triangle represents an individual. (C) Flow cytometric analysis of human blood samples gained from HD (age 60+ years ) and CLL patients (grouped: mutated and unmutated IgH VH). The plot indicates the mean fluorescence intensity (MFI) of PAX5 protein expression after gating on CD19+ and CD5+ cells for the mutated/unmutated CLL patients and the HD B1 cells. For the HD B2 cells we gated on CD19+ and CD5- cells. Median ± SEM. P-values were obtained using two-tailed Student's t-test (*P<0.05; **P< 0.01; ***P<0.001). Every dot or triangle represents an individual. (D) Flow cytometry analysis of PAX5 expression (left) in human blood samples of HD B1 cells (black), B2 cells (green) and CLL patients with mutated (blue) and unmutated (red) IgH VH. The histograms indicate the fluorescence intensity of PAX5 protein expression of representative samples after gating on CD19+ and CD5+ cells for the CLL mutated/unmutated patients and the HD B1 cells. For the HD B2 cells we gated on CD19+ and CD5- cells. Representative isotype control (middle) and size control (right) of the flow cytometric analysis are also shown. (E) Schematic representation of our working hypothesis. Upregulated PAX5 protein expression leads to higher expression of miR-29 family miRNA, which in turn inhibit the transcription of PTEN decreasing PTEN protein expression. This in turn increases the PI3K activity, which might lead to the development of CLL. Haematologica | 107 August 2022

1811

ARTICLE - BCR and PI3K in CLL cells

V.K. Schmid et al.

dictates the fate of the CLL cells in the Eµ-TCL1 mouse model by activating the PI3K pathway, the key component of chronic active BCR signaling in mouse CLL.44 This points out the importance of the BCR as a scaffold and platform for signaling emanating from different stimuli. Therefore, our results establish that BCR expression per se is required and indispensable to keep mouse CLL cells alive. To our knowledge, this study provides the first direct genetic evidence that the maintenance of mouse CLL cells depends on the BCR. The second major finding of this study is that conditional deletion of the Pten gene and subsequent constitutively active PI3K signaling lead to an accelerated onset of CLL development in mice. CLL in Eµ-TCL1 transgenic mice develops after long latencies.20 This indicates that high expression of TCL1 is insufficient to drive transformation and that other genetic or epigenetic changes are presumably required. PTEN deficiency alone does not drive tumorigenesis in mature B cells;26 however, based on our data, inactivation of PTEN with simultaneous overexpression of TCL1 accelerates the onset of CLL pathogenesis in young mice. We show that in PTEN-deficient TCL1-transgenic CLL cells AKT phosphorylation is increased compared to B cells only overexpressing TCL1. As PTEN-deficient splenic CLL cells are not susceptible to BCR stimulation with anti-IgM F(ab’)2 fragments, we believe that PTEN deficiency results in constitutive activation of the PI3K/AKT pathway. However, although Kohlhaas et al. recently reported that constitutive activation of AKT in a EµTCL1 mouse model results in RT31 we were unable to show this phenomenon in Ptenfl/+;mb1-Cre;Eµ-TCL1 mice with heterozygous Pten deletion. These results might be explained by differential progression and development of CLL in the respective mice. While we observe a significantly increased number of CLL cells in Ptenfl/+;mb1Cre;Eµ-TCL1 mice compared to mb1-Cre;Eµ-TCL1 control mice already at an age of 8 weeks (Figure 6A and B), this difference could not be detected in Eµ-TCL1 mice with constitutive AKT activation before the mice reached an age of 7 months. Spontaneous apoptosis of CLL cells in vitro has hampered the in-depth investigation of the mechanisms behind CLL maintenance. Cells from spleens of aged mb1CreERT2;Ptenfl/fl;Eµ-TCL1 mice with high tumor load proliferated in culture without addition of growth factors. Flow cytometric analysis showed that the cells maintained the phenotype of the primary leukemia even after prolonged in vitro culture. To the best of our knowledge, Eµ-TCL1 leukemia-derived cell lines have not been described to date. Recently, a similar phenomenon was observed by Chakraborty et al. wherein murine Eµ-TCL1 leukemia cells exhibiting biallelic inactivation of TP53, CDKN2A and CDKN2B were also found to proliferate spontaneously in vitro.39 Notably, cell lines from mb1-CreERT2;Ptenfl/fl;Eµ-

TCL1 were as susceptible to venetoclax as control cells from the mb1-CreERT2;Eµ-TCL1 mice, showing that the loss of PTEN does not confer resistance to apoptosis. Moreover, the PTEN-deficient CLL-like culture cells still relied on BCR signals for proliferation as BCR signaling inhibitors caused decreased survival and proliferation of the cells in vitro. Interestingly, significantly increased cell death and decreased proliferation of PTEN-deficient mb1CreERT2;Ptend/d;Eµ-TCL1 culture cells in comparison to mb1-CreERT2;Eµ-TCL1 leukemic cells could be observed due to NFκB inhibitor treatment, indicating that the NFκB pathway plays an important role in these PTEN-deficient CLL-like culture cells. These PTEN-deficient CLL-like cell lines may be well suited for high-throughput screening of novel compounds for CLL treatment. Furthermore, the PTEN-deficient cells can be transplanted into immunodeficient mice and may be used in further in vivo studies. The rapid development of CLL in these mice may help to dissect signaling mechanisms of CLL cells within a reasonable time frame in contrast to the slow disease development in Eµ-TCL1 mice. Therefore, this specifically provides a tool to develop novel treatment options for drug-resistant CLL. It is remarkable that although splenic CLL cells from mb1CreERT2;Ptenfl/fl;Eµ-TCL1 mice developed in culture after ex vivo deletion of Pten, they were not immortalized like the cells, in which Pten had been deleted in vivo. This may suggest that additional factors or additive mutations are required to promote transformation in vivo. PTEN is tightly regulated by various non-genomic mechanisms including epigenetic silencing, post-transcriptional regulation by non-coding RNA, and post-translational modification.45 Due to the high PTEN mRNA transcript levels in the analyzed human CLL samples which stand in contrast to the decreased PTEN protein expression, we assume that the PTEN downregulation in two thirds of overall 35 analyzed CLL samples might be regulated in a post-transcriptional manner mediated by non-coding RNA. miRNA comprise a large family of small non-coding RNA that emerged as post-transcriptional regulators of gene expression.46 The microRNA miR-21, miR-155, miR-1792 or miR-19 and miR-29, for instance, are post-transcriptional regulators of PTEN expression, which directly target PTEN and contribute to its reduced expression in CLL.47 Moreover, several studies demonstrated that microRNA expression profiles can be used to distinguish normal B cells from malignant CLL cells and that miRNA signatures are associated with prognosis and progression of CLL.48 Among other miRNA that have also been shown to regulate PTEN expression the miR-29 family is one of the critical miRNA that play a role in cancer pathogenesis.49 It was revealed that Eµ-miR-29 transgenic mice overexpressing miR-29 in B cells exhibit an expanded CD5+ B-cell population with 20% of the mice developed leukemia indicating

Haematologica | 107 August 2022

1812

ARTICLE - BCR and PI3K in CLL cells

V.K. Schmid et al.

a role of miR-29 in the pathogenesis of B-CLL.50 Our analysis of miR-29 expression in human CLL cells revealed that all three miR-29 family miRNA were significantly overexpressed in both M-CLL and U-CLL patient samples. Moreover, we could show that PAX5 expression is upregulated in human CLL patient samples, which might induce the upregulation of the miR-29 family miRNA. This stands in line with the recently published findings of Calderón et al. who identified PAX5 as an enhancer of PI3K signaling that downregulates PTEN expression in mature B cells, likely by controlling the abundance of PTEN-targeting miRNA.37 In summary, our data illustrate that the loss of PTEN in murine CLL cells results in the accelerated development of the disease. This observation underscores the significance of PI3K signaling in the pathogenesis of CLL. In addition, the significant downregulation of PTEN expression in B cells of around two third of the analyzed CLL patients suggests an important role of PTEN in the development and maintenance of human CLL. It is conceivable that the decreased PTEN expression is induced by the increased levels of miR-21, miR-29 family and PAX5 expression in the analyzed human CLL cells as PAX5 is believed to restrain PTEN expression in B cells by controlling the expression of PTENtargeting miRNA.37 Conclusively, the work described in this paper strengthens the important role of PTEN as a tumor suppressor in CLL and raises a number of interesting questions that may help to design the potential use of PTENtargeting miRNA inhibitor strategies for CLL. Furthermore, we show that BCR loss fully abrogates the survival of CLL cells in mice. We therefore conclude that

PTEN expression sets a threshold for malignant transformation in the presence of BCR. Targeting the BCR itself, however, may be a major future achievement in combatting CLL, since antibodies against Igβ have been shown to be potent in depleting autoimmune and malignant B cells in mouse models and in preclinical studies. Disclosure No conflicts of interest to disclose. Contributions VS, AK, NA, and EH performed experiments and analyzed data. LN provided the IgαTMF mice; KR provided ES cells targeted with IgαTMF and discussed the study; EH and HJ designed the study and proposed the experiments; EH supervised the work and wrote the manuscript with VS. All the authors read the manuscript and discussed the results. Acknowledgments We thank Duygu Yağdıran, Karoline Lodd, Katharina Goehring, Lisa Gögler, Selina Fahrenholz, Stefanie Brey, and Andrea Schneider for their technical support and Ella Levit-Zerdoun for proof-reading the manuscript. Funding This work was supported by the DFG through SFB1074 (Experimental Models and Clinical Translation in Leukemia) projects A9, A10 and through TRR130 (B cells: Immunity and Autoimmunity) projects P01, P02, P04, P08 and C03, and through EXC294, and ERC advanced grants 694992 to HJ.

References 1. Li Y, Wang Y, Wang Z, Yi D, Ma S. Racial differences in three major NHL subtypes: descriptive epidemiology. Cancer Epidemiol. 2015;39(1):8-13. 2. Stevenson FK, Krysov S, Davies AJ, Steele AJ, Packham G. Bcell receptor signaling in chronic lymphocytic leukemia. Blood. 2011;118(16):4313-4320. 3. Seda V, Mraz M. B-cell receptor signalling and its crosstalk with other pathways in normal and malignant cells. Eur J Haematol. 2015;94(3):193-205. 4. Damle RN, Wasil T, Fais F, et al. Ig V gene mutation status and CD38 expression as novel prognostic indicators in chronic lymphocytic leukemia. Blood. 1999;94(6):1840-1847. 5. 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. 6. Stamatopoulos K, Agathangelidis A, Rosenquist R, Ghia P. Antigen receptor stereotypy in chronic lymphocytic leukemia. Leukemia. 2017;31(2):282-291. 7. Philippen A, Diener S, Zenz T, Dohner H, Stilgenbauer S, Mertens D. SYK carries no activating point mutations in patients with chronic lymphocytic leukaemia (CLL). Br J Haematol. 2010;150(5):633-636.

8. Duhren-von Minden M, Ubelhart R, Schneider D, et al. Chronic lymphocytic leukaemia is driven by antigen-independent cellautonomous signalling. Nature. 2012;489(7415):309-312. 9. Patel V, Balakrishnan K, Bibikova E, et al. Comparison of acalabrutinib, a selective Bruton tyrosine kinase inhibitor, with ibrutinib in chronic lymphocytic leukemia cells. Clin Cancer Res. 2017;23(14):3734-3743. 10. Woyach JA, Bojnik E, Ruppert AS, et al. Bruton's tyrosine kinase (BTK) function is important to the development and expansion of chronic lymphocytic leukemia (CLL). Blood. 2014;123(8):1207-1213. 11. Burger JA, Chiorazzi N. B cell receptor signaling in chronic lymphocytic leukemia. Trends Immunol. 2013;34(12):592-601. 12. Ten Hacken E, Burger JA. Molecular pathways: targeting the microenvironment in chronic lymphocytic leukemia - focus on the B-cell receptor. Clin Cancer Res. 2014;20(3):548-556. 13. Werner M, Hobeika E, Jumaa H. Role of PI3K in the generation and survival of B cells. Immunol Rev. 2010;237(1):55-71. 14. Okkenhaug K, Vanhaesebroeck B. PI3K in lymphocyte development, differentiation and activation. Nat Rev Immunol. 2003;3(4):317-330. 15. Downes CP, Ross S, Maccario H, Perera N, Davidson L, Leslie NR. Stimulation of PI 3-kinase signaling via inhibition of the tumor

Haematologica | 107 August 2022

1813

ARTICLE - BCR and PI3K in CLL cells

V.K. Schmid et al.

suppressor phosphatase, PTEN. Adv Enzyme Regul. 2007;47:184-194. 16. Thick J, Metcalfe JA, Mak YF, et al. Expression of either the TCL1 oncogene, or transcripts from its homologue MTCP1/c6.1B, in leukaemic and non-leukaemic T cells from ataxia telangiectasia patients. Oncogene. 1996;12(2):379-386. 17. Herling M, Patel KA, Khalili J, et al. TCL1 shows a regulated expression pattern in chronic lymphocytic leukemia that correlates with molecular subtypes and proliferative state. Leukemia. 2006;20(2):280-285. 18. Chen SS, Chiorazzi N. Murine genetically engineered and human xenograft models of chronic lymphocytic leukemia. Semin Hematol. 2014;51(3):188-205. 19. Bresin A, D'Abundo L, Narducci MG, et al. TCL1 transgenic mouse model as a tool for the study of therapeutic targets and microenvironment in human B-cell chronic lymphocytic leukemia. Cell Death Dis. 2016;7(1):e2071. 20. Bichi R, Shinton SA, Martin ES, et al. Human chronic lymphocytic leukemia modeled in mouse by targeted TCL1 expression. Proc Natl Acad Sci U S A. 2002;99(10):6955-6960. 21. Kraus M, Alimzhanov MB, Rajewsky N, Rajewsky K. Survival of resting mature B lymphocytes depends on BCR signaling via the Igalpha/beta heterodimer. Cell. 2004;117(6):787-800. 22. Hobeika E, Levit-Zerdoun E, Anastasopoulou V, et al. CD19 and BAFF-R can signal to promote B-cell survival in the absence of Syk. EMBO J. 2015;34(7):925-939. 23. Suzuki A, Yamaguchi MT, Ohteki T, et al. T cell-specific loss of Pten leads to defects in central and peripheral tolerance. Immunity. 2001;14(5):523-534. 24. Hobeika E, Thiemann S, Storch B, et al. Testing gene function early in the B cell lineage in mb1-cre mice. Proc Natl Acad Sci U S A. 2006;103(37):13789-13794. 25. Srinivasan L, Sasaki Y, Calado DP, et al. PI3 kinase signals BCRdependent mature B cell survival. Cell. 2009;139(3):573-586. 26. Setz CS, Hug E, Khadour A, et al. PI3K-mediated Blimp-1 activation controls B Cell selection and homeostasis. Cell Rep. 2018;24(2):391-405. 27. Liu J, Chen G, Feng L, et al. Loss of p53 and altered miR15a/16-1short right arrowMCL-1 pathway in CLL: insights from TCL1-Tg:p53(-/-) mouse model and primary human leukemia cells. Leukemia. 2014;28(1):118-128. 28. Lee HJ, Gallardo M, Ma H, et al. p53-independent ibrutinib responses in an Emu-TCL1 mouse model demonstrates efficacy in high-risk CLL. Blood Cancer J. 2016;6(6):e434. 29. Landau DA, Tausch E, Taylor-Weiner AN, et al. Mutations driving CLL and their evolution in progression and relapse. Nature. 2015;526(7574):525-530. 30. Alkhatib A, Werner M, Hug E, et al. FoxO1 induces Ikaros splicing to promote immunoglobulin gene recombination. J Exp Med. 2012;209(2):395-406. 31. Kohlhaas V, Blakemore SJ, Al-Maari M, et al. Active Akt signaling triggers CLL toward Richter transformation via overactivation of Notch1. Blood. 2021;137(5):646-660. 32. Kim JY, Jeong HS, Chung T, et al. The value of phosphohistone H3 as a proliferation marker for evaluating invasive breast cancers: a comparative study with Ki67. Oncotarget. 2017;8(39):65064-65076.

33. Sandhu SK, Fassan M, Volinia S, et al. B-cell malignancies in microRNA Emu-miR-17~92 transgenic mice. Proc Natl Acad Sci U S A. 2013;110(45):18208-18213. 34. Battistella M, Romero M, Castro-Vega LJ, et al. The high expression of the microRNA 17-92 cluster and its paralogs, and the downregulation of the target gene PTEN, is associated with primary cutaneous B-cell lymphoma progression. J Invest Dermatol. 2015;135(6):1659-1667. 35. Bai H, Wei J, Deng C, Yang X, Wang C, Xu R. MicroRNA-21 regulates the sensitivity of diffuse large B-cell lymphoma cells to the CHOP chemotherapy regimen. Int J Hematol. 2013;97(2):223-231. 36. Hines MJ, Coffre M, Mudianto T, et al. miR-29 sustains B cell survival and controls terminal differentiation via regulation of PI3K signaling. Cell Rep. 2020;33(9):108436. 37. Calderon L, Schindler K, Malin SG, et al. Pax5 regulates B cell immunity by promoting PI3K signaling via PTEN downregulation. Sci Immunol. 2021;6(61):eabg5003. 38. Levit-Zerdoun E, Becker M, Pohlmeyer R, et al. Survival of Igalpha-deficient mature B cells requires BAFF-R function. J Immunol. 2016;196(5):2348-2360. 39. Chakraborty S, Martines C, Porro F, et al. B-cell receptor signaling and genetic lesions in TP53 and CDKN2A/CDKN2B cooperate in Richter transformation. Blood. 2021;138(12):1053-1066. 40. Iacovelli S, Hug E, Bennardo S, et al. Two types of BCR interactions are positively selected during leukemia development in the Eµ-TCL1 transgenic mouse model of CLL. Blood. 2015;125(10):1578-1588. 41. Nakayama K, Nakayama K, Negishi I, et al. Disappearance of the lymphoid system in Bcl-2 homozygous mutant chimeric mice. Science. 1993;261(5128):1584-1588. 42. Lam KP, Kuhn R, Rajewsky K. In vivo ablation of surface immunoglobulin on mature B cells by inducible gene targeting results in rapid cell death. Cell. 1997;90(6):1073-1083. 43. Ten Hacken E, Burger JA. Microenvironment interactions and Bcell receptor signaling in chronic lymphocytic leukemia: implications for disease pathogenesis and treatment. Biochim Biophys Acta. 2016;1863(3):401-413. 44. Brown JR. The PI3K pathway: clinical inhibition in chronic lymphocytic leukemia. Semin Oncol. 2016;43(2):260-264. 45. Song MS, Salmena L, Pandolfi PP. The functions and regulation of the PTEN tumour suppressor. Nat Rev Mol Cell Biol. 2012;13(5):283-296. 46. Bartel DP. MicroRNAs: target recognition and regulatory functions. Cell. 2009;136(2):215-233. 47. Musilova K, Mraz M. MicroRNAs in B-cell lymphomas: how a complex biology gets more complex. Leukemia. 2015;29(5):1004-1017. 48. Calin GA, Liu CG, Sevignani C, et al. MicroRNA profiling reveals distinct signatures in B cell chronic lymphocytic leukemias. Proc Natl Acad Sci U S A. 2004;101(32):11755-11760. 49. Pekarsky Y, Croce CM. Is miR-29 an oncogene or tumor suppressor in CLL? Oncotarget. 2010;1(3):224-227. 50. Santanam U, Zanesi N, Efanov A, et al. Chronic lymphocytic leukemia modeled in mouse by targeted miR-29 expression. Proc Natl Acad Sci U S A. 2010;107(27):12210-12215.

Haematologica | 107 August 2022

1814

ARTICLE - Hematopoiesis

Utility of plasma cell-free DNA for de novo detection and quantification of clonal hematopoiesis Fernanda Gutierrez-Rodrigues,1 Isabel Beerman,2 Emma M. Groarke,1 Bhavisha A. Patel,1 Nina Spitofsky,1 Laura W. Dillon,1 Diego Quinones Raffo,1 Christopher S. Hourigan,1 Sachiko Kajigaya,1 Luigi Ferrucci2 and Neal S. Young 1

Correspondence: Fernanda Gutierrez Rodrigues fernanda.rodrigues@nih.gov

Hematology Branch, National Heart, Lung, and Blood Institute (NHLBI), NIH, Bethesda and Translational Gerontology Branch, National Institute on Aging, NIH, BRC, Baltimore, MD, USA.

Received: May 18, 2021. Accepted: August 23, 2021. Prepublished: September 30, 2021. https://doi.org/10.3324/haematol.2021.279230

Abstract Although cell-free DNA (cfDNA) tests have emerged as a potential non-invasive alternative to bone marrow biopsies for monitoring clonal hematopoiesis in hematologic diseases, whether commercial cfDNA assays can be implemented for the detection and quantification of de novo clonal hematopoiesis in place of blood cells is uncertain. In this study, peripheral plasma cfDNA samples available from patients with aplastic anemia (n=25) or myelodysplastic syndromes (n=27) and a healthy cohort (n=107) were screened for somatic variants in genes related to hematologic malignancies using a Clinical Laboratory Improvement Amendments-certified panel. Results were further compared to DNA sequencing of matched blood cells. In reported results, 85% of healthy subjects, 36% of patients with aplastic anemia and 74% of patients with myelodysplastic syndromes were found to have somatic cfDNA variants, most frequently in DNMT3A, TET2, ASXL1 and SF3B1. However, concordance between cfDNA and blood cell findings was poor for the detection of clonal hematopoiesis when the allele frequency of the variants was <10%, which was mostly observed in the healthy and aplastic anemia cohorts but not in patients with myelodysplastic syndromes. After filtering data for potential artifacts due to low variant allele frequency and sequencing depth, the frequency of clonal hematopoiesis in cfDNA from healthy individuals and patients with aplastic anemia decreased to 52% and 20%, respectively. cfDNA and matched blood cells were not interchangeable for tracking changes in allele burdens as their agreement by Bland-Altman analysis was poor. A commercial cfDNA assay had good performance for de novo detection of clonal hematopoiesis in myelodysplastic syndromes, but showed no advantage over blood cells in diseases with low allele burdens or in healthy individuals.

Introduction Plasma circulating cell-free DNA (cfDNA) has been used as a source of tumor-derived DNA in peripheral blood (PB) for liquid biopsies.1,2 Commercial liquid biopsies based on massively parallel sequencing have been utilized in practice as they are minimally invasive, and tumor-derived cfDNA shows a good correlation with tissue genotypes.3-5 However, more than 50% of somatic variants identified in circulating DNA derive from blood cells and are consistent with clonal hematopoiesis (CH),4,6 a phenomenon in which hematopoietic blood cells acquire variants in genes recurrently mutated in hematologic malignancies.7-10 CH occurs with normal aging, and healthy individuals have CH of indeterminate potential (CHIP), defined as a clonal population at variant allele frequency (VAF) ≥2% in blood, and mostly associated with a small subset of genes: DNMT3A, TET2, ASXL1, JAK2 and TP53.11 The prevalence of CHIP is agedependent, with the condition being rare until middle age but ubiquitous in individuals older than 60 years when error-

correcting sequencing approaches are used for screening.7,12 CH has been used as a predictor for development of blood cancers and associated with an increased risk of cardiovascular disease.7,10,13-15 The particular genes mutated, the number of variants and clone sizes influence the likelihood of healthy individuals with CHIP progressing to develop a hematologic malignancy.10,14-16 In hematologic diseases, CH has been used as a predictor of clinical outcomes and for risk stratification of patients.17-19 In aplastic anemia (AA), pathogenic variants in PIGA and BCOR correlate with a better response to immunosuppressive therapy whereas patients with variants in RUNX1, TP53, DNMT3A and ASXL1 have worse outcomes; typically the median VAF of these variants is 9%.19 Genomic data are also incorporated into prognostic models for myelodysplastic syndromes (MDS); variants are commonly found at higher VAF (>30%) and in genes such as ASXL1, DNMT3A, NRAS, RUNX1, U2AF1 and TP53, which are associated with poor prognosis.17-20 In this context, the use of plasma cfDNA sequencing assays for the detection and quantification of CH has emerged as a

Haematologica | 107 August 2022

1815

ARTICLE - Utility of cfDNA for clonal hematopoiesis detection potential non-invasive alternative to bone marrow (BM) biopsies in MDS and AA. In addition, plasma is underused in many laboratories and may be a source of DNA for many retrospective CH studies in asymptomatic individuals when blood samples are not available. A few studies have shown that both PB and BM cells can be used for detection and quantification of CH,10 however, it is not yet known whether cfDNA is equally accurate. Previous experiences with liquid biopsies have demonstrated that highly sensitive assays are required to reliably detect small clonal populations, and poor congruence between Clinical Laboratory Improvement Amendments (CLIA)-certified commercial cfDNA assays has often been reported.21-23 These findings imply that detection of de novo CH in cfDNA may be troublesome, as CHIP variants can often be seen at low frequency and no matched samples or previous genotyping data would be available for validating cfDNA results. To investigate the cfDNA performance of a commercial CLIAapproved targeted sequencing panel of genes related to CH and hematologic malignancy in detecting and quantifying de novo CH variants, we screened healthy individuals from the Baltimore Longitudinal Study of Aging (BLSA) and patients with AA and MDS. Reported cfDNA results were validated against DNA-sequencing from matched blood cells.

Methods Cohorts and samples EDTA-PB plasma for cfDNA assay was available for 107 healthy individuals, including 96 participants of BLSA at the National Institute of Aging (NIA, NCT00233272) and 11 healthy volunteers at the National Heart, Lung, and Blood Institute (NHLBI). EDTA-PB samples were also collected at diagnosis from patients with AA (n=25) or MDS (n=27) from the Hematology Branch marrow failure clinic at the NHLBI (NCT00932156 and NCT01623167). Matched PB or BM cells collected at the same time as peripheral plasma samples were used for comparative analyses. In the healthy cohort, only 25 individuals had matched samples from PB mononuclear cells (PBMC, n=15) or granulocytes (n=10). In the AA and MDS cohorts, paired BM cells from diagnosis and serial time-points were used as matched samples (AA, n=53 and MDS, n=55) (Online Supplementary Figure S1). This study was approved by the Institutional Review Boards of the NHLBI and NIA. All samples were collected according to the Declaration of Helsinki and written consent was obtained from all participants or their legal guardians.

F. Gutierrez-Rodrigues et al.

a commercial CLIA panel that includes genes commonly associated with CHIP (CHIP genes) and cancer and hematologic neoplasia (non-CHIP genes) (Online Supplementary Table S1).16,24 The somatic origin of variants at VAF >35% reported by the commercial assay was confirmed by targeted DNA sequencing of sorted CD3+ cells or PBMC/granulocytes; variants present in these fractions at similar VAF to cfDNA were considered germline (Online Supplementary Tables S2-S4). Of note, the cfDNA GTC-Hematology profile assay was CLIA-validated for detecting abnormalities in MDS and AA but not for non-myeloid neoplasms; therefore, samples from healthy individuals were screened for research purposes only. Comparative analysis of cell-free DNA and matched peripheral blood or bone marrow samples Sensitivity, specificity and positive predictive value of the CLIA-certified cfDNA test were assessed in comparison to matched PB or BM samples. Likely germline variants were also included in this comparative analysis. Concordance among matched samples was evaluated at a variant level with each variant classified as true positive, false positive, false negative and true negative (Figure 2A). The performance of the cfDNA assay in quantifying both somatic and likely germline variants was assessed by linear regression and the Bland-Altman analysis, a more accurate method to evaluate the agreement between paired values.25 Statistical analysis was performed using GraphPad Prism v5 (GraphPad Software Inc, CA, USA). Stringent criteria for filtering likely false positive and false negative variants Ultra-deep error-correcting sequencing was performed in our institution as previously described to validate the CLIA results of selected samples from healthy individuals for whom we had available materials.26 Results obtained from genetic reports were refined by stringent criteria to filter variants that were below the VAF and read depth thresholds of confirmed negative variants by the ultra-deep error-correcting sequencing. Detailed experimental procedures and analytical methods are described in the Online Supplementary Data.

Results

Clonal landscapes of cell-free DNA variants in healthy individuals, and patients with aplastic anemia or myelodysplastic syndromes Commercial massively parallel sequencing of targeted Among 107 healthy subjects (median age 72 years; range, genes 24-96) screened for somatic variants in cfDNA, 91 (85%) cfDNA and PB/BM samples were screened for somatic vari- had at least one variant. The majority (75/107 subjects, ants in hematologic neoplasm-related genes using the GTC- 71%) had one (n=27) or two or more (n=48) variants in Hematology profile (Genomic Testing Cooperative, CA, USA), CHIP genes, most frequently in DNMT3A, TET2, TP53 and Haematologica | 107 August 2022

1816

ARTICLE - Utility of cfDNA for clonal hematopoiesis detection ASXL1 (Figure 1A, Online Supplementary Figure S2). However, 50/107 subjects (47%) also had somatic variants in non-CHIP genes, and 16 of them (16/50) had variants in non-CHIP genes only (Online Supplementary Table S2). The frequency of somatic cfDNA variants in CHIP genes, DTA genes (DNMT3A, TET2 and ASXL1) and non-CHIP genes increased with age and CH was observed in more than 60% of individuals older than 40 years (Figure 1B). Among the 25 patients with AA (median age 51 years; range, 13-82), nine (36%) were found to have somatic cfDNA variants. Most of them had one (n=1) or two (n=5) cfDNA variants in CHIP genes (6/25, 24%), most commonly in BCOR and SF3B1 (Figure 1C). Among 27 MDS patients (median age 63 years; range, 35-85), 20 (74%) had somatic cfDNA variants, including 16/27 (60%) with variants in CHIP genes. Patients had one (n=10) or two or more (n=6) cfDNA variants, TET2, SF3B1 and ASXL1 being most frequent (Figure 1D). Also, 5/32 AA and 9/27 MDS patients were found to have cfDNA variants in non-CHIP genes (Figure 1C, D). Four MDS and three AA patients had cfDNA variants exclusively in non-CHIP genes (Online Supplementary Tables S3 and S4). Concordance between cell-free DNA and blood cells was limited by the allele frequency of the variants In paired samples from healthy individuals (n=25), most cfDNA variants identified were not found in matched cells (i.e, they were false-positives), and true positives were likely germline variants (median VAF 47% in all cfDNA, PBMC and granulocytes) (Figure 2B). Sensitivity was defined as a proportion of variants that was correctly detected in cfDNA, and specificity was defined as the probability that a given patient was correctly reported with no variants. The positive predictive value was defined as the probability of a variant detected in cfDNA being also detected in blood cells. In this cohort, sensitivity, specificity and positive predictive value of the cfDNA assay were low (32%, 26% and 15%, respectively), and most discordant variants had median VAF <10% (Table 1, Figure 2B). Here, the median VAF of variants identified in the healthy cohort was 2.5%±10% (95% confidence interval [CI]: 2-4] and both the sensitivity and positive predictive value of the assay for detecting variants at VAF <10% were poor (Figure 2E, Table 1). In contrast to healthy subjects, most cfDNA variants found in AA and MDS patients were true positives with high concordance between paired samples. Sensitivity, specificity and the positive predictive value of cfDNA in comparison to BM cells were greater than 70% and 85% in AA and MDS patients, respectively (Table 1). Most of the discordant pairs from AA- and MDS-matched samples were at lower VAF (Figure 2B). In all cohorts, the median VAF of discordant pairs was significantly lower than the VAF of true positives (7.6% [95% CI: 4-10] vs. 33% [95% CI:

F. Gutierrez-Rodrigues et al.

28-37]; t-test, P<0.001), and most discordant variants were at VAF <10% (upper CI of VAF from discordant pairs). Indeed, discordance observed in paired samples correlated with a proportion of discordant variants at VAF <10% in each cohort, regardless of disease types (Table 1). Sensitivity, specificity and positive predictive value were very high in MDS, in which most variants were seen at higher VAF (median VAF=32%, 95% CI: 27-42) (Figure 2E). In contrast, the lowest concordance was seen in healthy subjects, in whom most discordant variants were seen at VAF <10% (Figure 2B). Analytical factors associated with cell-free DNA screening were related to assay discordance We next investigated whether discordance also correlated with sequencing depth, an analytical factor that can lead to high rates of sequencing artifacts. In all cohorts, read depth of deduplicated reads in true positives was significantly higher (482X, 95% CI: 368X–578X) (Figure 2D) than in false positives and false negatives (110X, 95% CI: 89X– 143X). To confirm these findings, we validated variants found in PB cells from two healthy subjects using an ultra-deep error-correcting sequencing panel with 13 genes: plasma samples were not available for this inhouse validation. Four out of six variants initially identified in these PB cells at a median VAF of 2.5% (range, 2-3.6%) and a median read depth of 246X (range, 104X-347X) were all confirmed negative (Figure 2D). In this validation, 95% of confirmed negative variants were at VAF <3.5% and were sequenced at coverages below <350X, which may represent a limit of detection of this commercial assay (Figure 2D). Of note, as stated by the vendor, the assay has a typical sensitivity of 5% for overall variant detection and 3% for detecting specific somatic variants in “hotspots.” However, somatic variants with an allele frequency as low as 0.9% were reported by the laboratory, and included in our concordance analysis. We found no specific genes or types of variants associated with assay discordance. When combining all three cohorts, false positive and false negative cfDNA variants were found mostly in ASXL1, TET2, SF3B1 and DNMT3A, and included missense, stop-gain and frameshift variants. The remaining discordant variants were seen in a broad spectrum of genes (Online Supplementary Figure S3). Stringent criteria to filter out discordant variants improved cell-free DNA concordance with blood cells To improve concordance between paired samples, we further refined our analysis by filtering out variants that were below the VAF and read depth thresholds of confirmed negative variants by error-correcting sequencing (likely artifacts from the cfDNA CLIA assay). We first discarded variants at VAF <1% and read depth <700X, and variants at VAF <3.5% and read depth <400X (more strin-

Haematologica | 107 August 2022

1817

ARTICLE - Utility of cfDNA for clonal hematopoiesis detection

F. Gutierrez-Rodrigues et al.

Continued on following page. Haematologica | 107 August 2022

1818

ARTICLE - Utility of cfDNA for clonal hematopoiesis detection

F. Gutierrez-Rodrigues et al.

Figure 1. Clonal landscapes of somatic cell-free DNA variants detected in healthy individuals and patients with aplastic anemia or myelodysplastic syndromes. (A) The most frequently mutated genes in 107 healthy subjects obtained from genetic reports. Cell-free (cfDNA) variants were mostly found in DNMT3A, TET2, TP53 and ASXL1 (full lists of genes are shown in Online Supplementary Figure S2). However, during analysis, the original data from genetic reports were filtered out to remove variants that were likely false positive and false negative, due to low variant allele frequency (VAF) and low sequencing coverage. Variants originally reported by the Clinical Laboratory Improvement Amendments (CLIA) assay were filtered out from analysis if present at VAF <10% and had been sequenced at a read depth <100X, or were present at VAF <3.5% and had been sequenced at a read depth <400X. After data filtering, cfDNA variants were mostly found in DNMT3A, TET2, ASXL1 (DTA genes) and KIT. (B) Frequency of cfDNA variants by age ranges. Frequency of variants in genes commonly associated with clonal hematopoiesis of indeterminate potential (CHIP), including DTA genes, and non-CHIP genes. Approximately 60% of individuals aged 40 years or older were found to have clonal hematopoiesis in cfDNA. (C) The most frequently mutated genes in 25 patients with aplastic anemia at enrollment in the NCT00932156 trial or at diagnosis. (D) The most frequently mutated genes in 27 patients with myelodysplastic syndromes at enrollment in the NCT01623167 trial. (E) Frequency of cfDNA variants by age ranges after original data were filtered to remove variants that were likely false positive and false negative, due to low VAF and low sequencing coverage. Frequency of variants in both CHIP genes, including DTA genes alone but also in non-CHIP genes, increased with aging. Approximately 30% of individuals aged 40 years or older were found to have clonal hematopoiesis in cfDNA. (F) The most frequently mutated genes in 25 patients with aplastic anemia after data filtering. BM: bone marrow; PB: peripheral blood; AA: aplastic anemia; MDS: myelodysplastic syndromes. Table 1. Concordance between cell-free DNA and bone marrow or peripheral blood cells for detection of clonal hematopoiesis.

TP N

FN N

FP N

TN N

VAF>10%

VAF<10%

VAF>10%

VAF<10%

VAF>10%

VAF<10%

VAF>10%

VAF<10%

VAF>10%

VAF<10%

VAF>10%

VAF<10%

Sensitivity Specificity % %

PPV %

Discordant variants %

89.1

42.2

46.9

100

83.7

60.5

23.3

100

14.4

12.4

100

2.1

Data originally reported by the CLIA- certified assay Healthy individuals

Aplastic anemia

Myelodysplastic syndromes

100

Concordance after data filtering* Healthy individuals

Aplastic anemia

Myelodysplastic syndromes

*Variants originally reported by the CLIA assay were filtered out from analysis if they were present at VAF <10% and had been sequenced at read depth <100X, or present at VAF <3.5% and had been sequenced at read depth <400X. No variants were found in four healthy individuals, and 33 AA and 11 MDS patients’ paired samples (labeled as true negatives). TP: true positive; FN: false negative; FP: false positive; TN: true negatives; PPV: positive predictive value; CLIA: Clinical Laboratory Improvement Amendments; VAF: variant allele frequency.

gent than the minimum of 350X observed as the upper 95% CI for confirmed negative variants). cfDNA variants at a VAF <10% were only included in the analysis if the read depth was >100X (Figure 2D). Of note, the read depth of a variant at the lowest VAF (0.96%) found in both cfDNA and blood cells was 750X. Overall concordance between cfDNA and matched samples was improved for both healthy individuals and AA patients, but there were no significant changes in MDS paired samples (Table 1).

In the healthy cohort, once filtering criteria were applied, the frequency of subjects with somatic cfDNA variants was lower than that in the original analysis (56/107; 52% vs. 92/107; 86%); the frequency of CHIP variants in individuals >40 years old also decreased (30% vs. 60%) (Figure 1B, E). In addition, the frequency of cfDNA variants in non-CHIP genes was no longer related to increased age, and the median allele frequency of the variants increased substantially to 12% (95% CI: 10-15) (Figures 1E and 2E, F).

Haematologica | 107 August 2022

1819

ARTICLE - Utility of cfDNA for clonal hematopoiesis detection

F. Gutierrez-Rodrigues et al.

Figure 2. Concordance analysis of variants detected in cell-free DNA and in matched peripheral blood or bone marrow cells. (A) Schematic criteria used to classify cell-free DNA (cfDNA) variants as (i) true positive (TP), if detected in both cfDNA and a matched sample, or a matched sample at a different time-point; (ii) false positive (FP), if detected only in cfDNA; (iii) false negative (FN), if absent in cfDNA but present in the matched-sample; and (iv) true negative (TN), if no variants were found in either cfDNA or the paired samples. (B) Analysis with original data from genetic reports. The figure shows the variant allele frequency (VAF) of TP, FN and FP variants detected in paired samples from healthy individuals (n=25), patients with aplastic anemia (AA, n=53) and patients with myelodysplastic syndromes (MDS, n=55). VAF of FN variants corresponded to VAF found in peripheral blood (PB) or bone marrow (BM) cells. (C) VAF of TP, FP and FN detected in paired samples that remained in the analysis after original Clinical Laboratory Improvement Amendments (CLIA) data were filtered out to remove variants that were likely FP and FN, due to low VAF and low sequencing coverage. (D) VAF and coverage (read depth) of discordant variants (FP and FN) in comparison to TP variants. Single variants found in both cfDNA and matched blood cells are represented by dark blue circles (TP). Single discordant variants in cfDNA and matched blood cells are indicated with light blue circles (FP and FN). Variants identified in PB cells from healthy individuals and confirmed negative by ultra-deep error-correcting sequencing are represented by red circles. VAF of 10% and minimal read depth 100X were thresholds that delimited the 95% confidence interval (95% CI) observed for discordant variants. VAF below 3.5% and read depth below 400X were the minimum thresholds of the assay sensitivity based on the VAF and read depth of confirmed negative variants (limits of detection of the assay). VAF and read depth of FN variants represent VAF and coverage of variants detected in PB or BM cells. (E) Median VAF with 95% CI of cfDNA variants reported in the CLIA cfDNA assay. Median VAF of cfDNA in healthy controls, and individuals with AA or MDS were 2.5% (2%-4%), 2.6% (2%-4%) and 32% (27%-42%), respectively. (F) Median VAF with 95% CI of cfDNA variants that remained in the analysis after original CLIA data were filtered out to remove variants that were likely FP and FN, due to low VAF and low sequencing coverage. Median VAF of cfDNA in healthy controls, and individuals with AA or MDS were 12% (10%-15%), 6.7% (2%-45%) and 37% (31%-42%), respectively. Haematologica | 107 August 2022

1820

ARTICLE - Utility of cfDNA for clonal hematopoiesis detection

F. Gutierrez-Rodrigues et al.

Figure 3. Performance of cell-free DNA for quantification of variants in comparison to matched peripheral blood or bone marrow cells. Linear regression was used to assess a correlation between quantification of clones in paired samples. Agreement analysis by the Bland-Altman method was performed to assess how accurate and precise the quantification of variants in cellfree DNA (cfDNA) were in comparison to matched blood cells. This method calculates the bias, the mean difference between the variant allele frequency (VAF) of matched samples and its standard deviation (SD) to more accurately evaluate agreement between paired values. Bias and SD were later used to define limits of agreement (LoA; mean±2 SD) which represent the 95% confidence interval of mean differences. Both analyses were performed with paired variants from healthy subjects (A), patients with aplastic anemia (B) and patients with myelodysplastic syndromes (C) who remained in the analysis after data filtering. For Bland-Altman analysis, differences of VAF of individual paired variants detected in cfDNA and peripheral blood or bone marrow cells are plotted against their average VAF (right panels). Mean differences (bias) of VAF and LoA (mean±2 SD) are shown in the graphic. The closer the mean was to zero and the narrower the LoA was, the better agreement between VAF of variants detected in paired samples. PB: peripheral blood; grans: granulocytes; PBMC: peripheral blood mononuclear cells; BM: bone marrow; AA: aplastic anemia; MDS: myelodysplastic syndromes. Haematologica | 107 August 2022

1821

ARTICLE - Utility of cfDNA for clonal hematopoiesis detection The frequency of CH in AA also decreased after data filtering (9/25; 36% vs. 5/25; 25%). Although the median allele frequency of the variants and the most commonly mutated genes changed slightly in AA after the filtering process (6.7%, 95% CI: 2%-45%) (Figures 1F and 2E, F), no changes were seen in MDS as most variants had high VAF and sequencing coverage (Figure 2B, C). In MDS, a variant previously found in matched samples and classified as a true positive was reclassified as a false positive after data filtering; in the BM, this variant was of low quality and systematically filtered with stringent criteria (Figure 2C).

F. Gutierrez-Rodrigues et al.

15%, the former had a diagnosis of prostate cancer with no evidence of hematologic disease at the time his sample was collected, and this variant may have been present in the bloodstream due to his solid organ malignancy. Similarly, HC-64 had a high VAF TP53 variant, for which a somatic status was not confirmed. This individual had no history of hematologic disease or a solid malignancy but had cytopenias of unknown significance for 9 years before his last follow-up. The remaining cfDNA variants were at very small allele fractions (median VAF 1.2% with the subjects’ median age of 65 years) and none of the individuals was known to have myeloid malignancy at Quantification of clonal hematopoiesis in cell-free DNA last follow-up, suggesting that the clinical utility of the was not comparable to that in blood cells cfDNA assays for CHIP surveillance in healthy individuals We further investigated whether the VAF of filtered cfDNA was limited. variants accurately replicated the VAF in blood cells. Linear regression showed a high correlation between paired samples in all cohorts (R2 >0.7) (Figure 3). However, agreeDiscussion ment between the quantification of CH in cfDNA and matched blood samples by the Bland-Altman method The questionable analytical validity of cfDNA assays has was poor for all the cohorts (Figure 3A-C). An average dif- raised concerns regarding their clinical utility. Two major ference (bias) and standard deviation (SD) of a clone size issues facing routine implementation of cfDNA tests are (VAF%) found in paired samples was 9.3%±9.5% in healthy the high frequency of discordant variants in matched individuals, 2.0%±5.7% in AA and -1.9%±6.5% in MDS. samples and the lack of characterization of the cfDNA geThese large variations in the bias and SD translated to a nomic landscapes in both healthy individuals and in spewide range of limits of agreement (mean ± 2 SD); the lar- cific disease states.2 As most studies in this field have gest variations were seen in healthy individuals and MDS focused on precision oncology and detection of tumorsamples as cfDNA variants over- and under-estimated derived DNA, the clinical utility of cfDNA assays for CH clone sizes by up to 28% and 21%, respectively (Figure 3A, detection and surveillance remains poorly characterized. C). Also, in all cohorts, the bias of some paired values was Using a single CLIA-certified assay, we showed that cfDNA not within the limits of agreement, suggesting that vari- assay performance in detecting CH was limited by anaations larger than the confidence intervals may often be lytical factors, which may preclude its use for CH detecobserved in cfDNA, regardless of disease type. tion and surveillance; and the reliability of a single commercial cfDNA assay was limited when variants were Healthy individuals with cell-free DNA variants at VAF <10%. in prognostically adverse genes Assay sensitivity among massive parallel sequencing platWe next assessed whether the presence of variants in forms is a key limitation for accurate detection of variants genes highly related to an increased risk of developing with low allele frequency in cfDNA.4,6 Invariably, all studies myeloid malignancies (TP53, RUNX1, IDH1, IDH2 and U2AF1) showing high cfDNA accuracy for tumor genotyping overwas associated with clinical outcomes in healthy individ- come potential sequencing errors by utilizing deep erroruals followed in the BLSA study. Nineteen individuals correcting sequencing techniques with original depths of were initially found to have variants in prognostically ad- 6200X-60000X to achieve error-correcting sequencing verse genes, but most of these variants had no clinical deduplication ratios of 360X-2400X and optimized protosignificance as they were later filtered out due to their cols to mitigate sequencing errors.4-6,27 In practice, there low VAF and poor sequencing coverage (especially TP53 is still no evidence of clinical utility of cfDNA assays for and DNMT3A variants) (Figure 4A). After data filtering, only detection and monitoring of variants at very low allele two of ten individuals found with variants in prognos- burdens.2 Discordance among CLIA assays has also been tically adverse genes developed hematologic diseases reported for liquid biopsies when patients’ paired (Figure 4B). HC-26 and HC-64 were found with pathogenic samples were assessed in parallel in different laboravariants in NRAS, U2AF1 and TP53 at VAF >30% which were tories.21-23 Similar to our results, discordance was mainly associated with longstanding cytopenias. HC-26 died 2 observed in variants at average VAF of 1% (and up to 3%) years following the data collection, reportedly from acute whereas higher concordance is observed at VAF >10%.21-23 leukemia. Although HC-04 and HC-104 were also found to In contrast to the situation in malignant diseases, unbihave pathogenic variants in U2AF1 and IDH1 at VAF around ased CH detection and quantification in asymptomatic inHaematologica | 107 August 2022

1822

ARTICLE - Utility of cfDNA for clonal hematopoiesis detection dividuals or in a disease with a low allele burden may be more challenging because cfDNA levels are typically lower in healthy individuals and are proportional to an allele burden and cell turnover.28,29 Additionally, a cfDNA yield in AA is low because of the lack of hematopoietic cells. As seen in our MDS cohort, sensitivity may not be critical when cfDNA is used for molecular profiling of malignant

F. Gutierrez-Rodrigues et al.

hematologic diseases with high allele burdens. In contrast, for asymptomatic individuals with CHIP and for AA patients, a minimal threshold of 10% for detecting de novo variants may not be useful, and reliable CH detection and quantification may be better achieved by screening of blood cells. The frequency of CH in cfDNA from healthy individuals

Figure 4. Healthy individuals with cell-free DNA variants in prognostically adverse genes. (A) The number of cell-free DNA (cfDNA) variants in prognostically adverse genes that were reported in the original genetic reports (black columns) and filtered by stringent criteria (red columns). (B) Retrospective clinical data from individuals enrolled in the Baltimore Longitudinal Study of Aging who were found to have variants in clonal hematopoiesis of indeterminate potential (CHIP) genes associated with an increased risk of myeloid malignancies. VAF: variant allele frequency. Haematologica | 107 August 2022

1823

ARTICLE - Utility of cfDNA for clonal hematopoiesis detection was higher than that in blood cells reported in previous studies (60% of individuals >40 years old vs. 10% of individuals >60 years old).9,12 However, this discrepancy may be explained by analytical factors related to the cfDNA screening, such as the use of a high sensitivity DNA sequencing approach and the increased number of false positive variants found at very low VAF in cfDNA. After we filtered the data to remove likely discordant variants, the frequency of CH in cfDNA from healthy individuals >40 years old was 30%, which was consistent with the data from error-correcting sequencing studies.14,16,24 Interestingly, after data filtering healthy individuals were predominantly found with single CHIP variants, as opposed to the initial data showing they were more frequently observed with two or more variants. Since we classified cfDNA variants as false positives solely based on their presence in matched blood cells and the cfDNA pool including circulating DNA released from other tissues, it is possible that variants labeled as false positive may be truly present but not be of a hematopoietic origin. Thus, an accurate and precise assay as well as characterization of the clonal cfDNA landscape in healthy people is critical for interpretation of results. In individuals from BLSA, who had long-term follow-up and from whom a comprehensive clinical history was available, cfDNA variants likely to be related to a malignant disease were not found in the most commonly mutated CHIP genes. True positive variants at VAF <10% were mainly observed in AA (Figure 2B, C), but increased concordance in this specific setting may be explained by the fact that their cfDNA and BM samples were the only ones analyzed and reported in parallel by the laboratory. The knowledge of parallel BM results or previous data may have influenced cfDNA results as opposed to a true de novo detection (independently sequenced, analyzed and reported). Accurate CH quantification in cfDNA is also critical for CHIP surveillance and therapeutic decisions in both AA and MDS. Although cfDNA often correlates with blood cells for quantification of clones,4,6,30 agreement by BlandAltman analysis showed poor agreement in quantification of mutant allele burdens between cfDNA and blood cells, suggesting that these DNA sources were not interchangeable for monitoring clonal dynamics. These findings could mislead correct interpretation of patients’ clonal dynamics as the clone size is an important genomic marker of progression to myeloid malignancies. We could not assess association of the cfDNA profile with clinical outcomes and a cardiovascular risk, due to small sample sizes of the cohorts. Lack of paired samples from healthy individuals and other clinical samples for an in-house deep error-correcting sequencing validation was the major limitation of this study. Here, the hematopoietic origins of cfDNA variants found in BLSA individuals were only validated in PBMC. Al-

F. Gutierrez-Rodrigues et al.

though previous studies have shown that either PB and BM cells can be used for the detection and quantification of CH,10,31,32 whether PBMC and granulocytes mirror BM as matched samples is still uncertain. We did not investigate whether biological differences accounted for cfDNA discordance with blood cells, nor did we assess the contribution of copy number variations for CH as cfDNA was not suitable for its detection. Of note, the commercial cfDNA assay used in this study was only validated for screening of myeloid malignancies and AA, not of healthy individuals. Nevertheless, lack of concordance was also seen in AA and MDS, suggesting that biological factors were not the main explanation of the assay discordance. To our knowledge, this is the first study that assessed the performance of a commercial cfDNA assay for the detection and quantification of CH in AA and MDS patients in parallel with healthy individuals. Although cfDNA tests have emerged as a non-invasive alterative to BM biopsies, our data suggest that their performance was not superior to the use of BM cells for CH profiling of diseases marked by variants present at low allele burdens. However, our data were derived from the use of a single commercial cfDNA assay, and its performance may not be comparable to that of other commercial assays. Nevertheless, the clinical use of cfDNA for detection and surveillance of CH in healthy individuals showed no advantage over the use of PB cells (additionally, both of these methods require the same PB sampling). The use of a cfDNA assay for de novo CH detection may be limited to diseases with a high allele burden (such as MDS) or retrospective studies for which plasma is the only source of DNA. Still, clinical reports must be interpreted with caution, especially when reporting de novo cfDNA variants at low levels; up to 50% of variants from the cfDNA reports of low VAF were likely artifacts. Ultra-sensitive assays with robust sequencing coverage and error-correction methodology may be required to overcome assay discordance; however, the commercial implementation of such assays may still be prohibited by costs. Although the primary focus of this study was not the screening of circulating tumor DNA, insights into methodology illustrated in our work could be applied to solid tumors when the circulating tumor DNA burden in plasma is low and detection of very small clones is required. Disclosures No conflicts of interest to disclose. Contributions FGR, IB, LF, SK and NSY made substantial contributions to the conception or design of the work; FGR, IB, DQR, EMG, BAP and NS contributed to the acquisition, analysis, or interpretation of data; LWD and CSH contributed to the vali-

Haematologica | 107 August 2022

1824

ARTICLE - Utility of cfDNA for clonal hematopoiesis detection dation of commercial results; FGR, IB, EMG, BAP and NSY drafted the work or revised it critically for important intellectual content; FGR, IB, LF and NSY contributed to the final approval of the version to be published.

F. Gutierrez-Rodrigues et al.

of the NHLBI/NIH and by the Intramural Research Program of the National Institute on Aging, NIH Baltimore. Research funding was provided by Novartis by way of a Cooperative Research and Development Agreement (CRADA).

Acknowledgments Data-sharing The authors would like to thank the DNA Sequencing and The human sequence data generated in this study are not Genomics Core Facility at NHLBI, and Toshiko Tanaka and publicly available due to patient privacy requirements but Linda Zukley at the NIA for their assistance. are available from the corresponding author upon reasonable request. Other data generated in this study are availFunding able within the article and its Online Supplementary Data This work was funded by the Intramural Research Program Files.

References 1. Wan JCM, Massie C, Garcia-Corbacho J, et al. Liquid biopsies come of age: towards implementation of circulating tumour DNA. Nat Rev Cancer. 2017;17(4):223-238. 2. Merker JD, Oxnard GR, Compton C, et al. Circulating tumor DNA analysis in patients with cancer: American Society of Clinical Oncology and College of American Pathologists joint review. J Clin Oncol. 2018;36(16):1631-1641. 3. Condoluci A, Rossi D. The future of cell-free DNA testing to guide therapeutic decisions in B-cell lymphomas. Curr Opin Hematol. 2019;26(4):281-287. 4. Razavi P, Li BT, Brown DN, et al. High-intensity sequencing reveals the sources of plasma circulating cell-free DNA variants. Nat Med. 2019;25(12):1928-1937. 5. Li BT, Janku F, Jung B, et al. Ultra-deep next-generation sequencing of plasma cell-free DNA in patients with advanced lung cancers: results from the Actionable Genome Consortium. Ann Oncol. 2019;30(4):597-603. 6. Liu J, Chen X, Wang J, et al. Biological background of the genomic variations of cf-DNA in healthy individuals. Ann Oncol. 2019;30(3):464-470. 7. 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. 8. Young AL, Challen GA, Birmann BM, Druley TE. Clonal haematopoiesis harbouring AML-associated mutations is ubiquitous in healthy adults. Nat Commun. 2016;7:12484. 9. Jaiswal S, Libby P. Clonal haematopoiesis: connecting ageing and inflammation in cardiovascular disease. Nat Rev Cardiol. 2020;17(12):828. 10. Miller PG, Steensma DP. Implications of clonal hematopoiesis for precision oncology. JCO Precis Oncol. 2020;6(4):639-646. 11. Steensma DP. The clinical challenge of idiopathic cytopenias of undetermined significance (ICUS) and clonal cytopenias of undetermined significance (CCUS). Curr Hematol Malig Rep. 2019;14(6):536-542. 12. Genovese G, Kähler AK, Handsaker RE, et al. Clonal hematopoiesis and blood-cancer risk inferred from blood DNA sequence. N Engl J Med. 2014;371(26):2477-2487. 13. Jaiswal S, Natarajan P, Silver AJ, et al. Clonal hematopoiesis and risk of atherosclerotic cardiovascular disease. N Engl J Med. 2017;377(2):111-121. 14. Abelson S, Collord G, Ng SWK, et al. Prediction of acute myeloid leukaemia risk in healthy individuals. Nature. 2018;559(7714):400-404. 15. Desai P, Mencia-Trinchant N, Savenkov O, et al. Somatic

mutations precede acute myeloid leukemia years before diagnosis. Nat Med. 2018;24(7):1015-1023. 16. Jaiswal S, Ebert BL. Clonal hematopoiesis in human aging and disease. Science. 2019;366(6465):eaan4673. 17. Bejar R. Clinical and genetic predictors of prognosis in myelodysplastic syndromes. Haematologica. 2014;99(6):956-964. 18. Bejar R, Stevenson KE, Caughey B, et al. Somatic mutations predict poor outcome in patients with myelodysplastic syndrome after hematopoietic stem-cell transplantation. J Clin Oncol. 2014;32(25):2691-2698. 19. Yoshizato T, Dumitriu B, Hosokawa K, et al. Somatic mutations and clonal hematopoiesis in aplastic anemia. N Engl J Med. 2015;373(1):35-47. 20. Steensma DP. How I use molecular genetic tests to evaluate patients who have or may have myelodysplastic syndromes. Blood. 2018;132(16):1657-1663. 21. Stetson D, Ahmed A, Xu X, et al. Orthogonal comparison of four plasma NGS tests with tumor suggests technical factors are a major source of assay discordance. JCO Precis Oncol. 2019;3(3):1-9. 22. Torga G, Pienta KJ. Patient-paired sample congruence between 2 commercial liquid biopsy tests. JAMA Oncol. 2018;4(6):868-870. 23. Kuderer NM, Burton KA, Blau S, et al. Comparison of 2 commercially available next-generation sequencing platforms in oncology. JAMA Oncol. 2017;3(7):996-998. 24. Jaiswal S, Libby P. Clonal haematopoiesis: connecting ageing and inflammation in cardiovascular disease. Nat Rev Cardiol. 2020;17(3):137-144. 25. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986;1(8476):307-310. 26. Hourigan CS, Dillon LW, Gui G, et al. Impact of conditioning intensity of allogeneic transplantation for acute myeloid leukemia with genomic evidence of residual disease. J Clin Oncol. 2020;38(12):1273-1283. 27. Rossi D, Diop F, Spaccarotella E, et al. Diffuse large B-cell lymphoma genotyping on the liquid biopsy. Blood. 2017;129(14):1947-1957. 28. Alborelli I, Generali D, Jermann P, et al. Cell-free DNA analysis in healthy individuals by next-generation sequencing: a proof of concept and technical validation study. Cell Death Dis. 2019;10(7):534. 29. Schwarzenbach H, Hoon DS, Pantel K. Cell-free nucleic acids as

Haematologica | 107 August 2022

1825

ARTICLE - Utility of cfDNA for clonal hematopoiesis detection biomarkers in cancer patients. Nat Rev Cancer. 2011;11(6):426-437. 30. Albitar A, Townsley D, Ma W, et al. Prevalence of somatic mutations in patients with aplastic anemia using peripheral blood cfDNA as compared with BM. Leukemia. 2018;32(1):227-229. 31. Guermouche H, Ravalet N, Gallay N, et al. High prevalence of

F. Gutierrez-Rodrigues et al.

clonal hematopoiesis in the blood and bone marrow of healthy volunteers. Blood Adv. 2020;4(15):3550-3557. 32. Mohamedali AM, Gäken J, Ahmed M, et al. High concordance of genomic and cytogenetic aberrations between peripheral blood and bone marrow in myelodysplastic syndrome (MDS). Leukemia. 2015;29(9):1928-1938.

Haematologica | 107 August 2022

1826

ARTICLE - Hemostasis

Syntaxin 5 determines Weibel-Palade body size and von Willebrand factor secretion by controlling Golgi architecture Marije Kat,1* Ellie Karampini,1,2* Arie J. Hoogendijk,1 Petra E. Bürgisser,3 Aat A. Mulder,4 Floris P.J. van Alphen,1 Jenny Olins,1 Dirk Geerts,5 Maartje van den Biggelaar,1 Coert Margadant,6 Jan Voorberg1,7 and Ruben Bierings3 Molecular Hematology, Sanquin Research and Landsteiner Laboratory, Amsterdam University Medical Center, University of Amsterdam, Amsterdam, the Netherlands; 2Vascular Biology, Royal College of Surgeons, Dublin, Ireland; 3Hematology, Erasmus University Medical Center, Rotterdam, the Netherlands; 4Molecular Cell Biology, Leiden University Medical Center, Leiden, the Netherlands; 5Medical Biology, Amsterdam University Medical Center, location AMC, University of Amsterdam, Amsterdam, the Netherlands; 6Angiogenesis Laboratory, Cancer Center Amsterdam, Amsterdam University Medical Center, location VUmc, Amsterdam, the Netherlands and 7Experimental Vascular Medicine, Amsterdam University Medical Center, University of Amsterdam, Amsterdam, the Netherlands. 1

Correspondence: Ruben Bierings r.bierings@erasmusmc.nl Received: September 30, 2021. Accepted: January 19, 2022. Prepublished: January 27, 2022. https://doi.org/10.3324/haematol.2021.280121 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

*MK and EK contributed equally as co-first authors.

Abstract von Willebrand factor (VWF) is a multimeric hemostatic protein primarily synthesized in endothelial cells. VWF is stored in endothelial storage organelles, the Weibel-Palade bodies (WPB), whose biogenesis strongly depends on VWF anterograde trafficking and Golgi architecture. Elongated WPB morphology is correlated to longer VWF strings with better adhesive properties. We previously identified the SNARE SEC22B, which is involved in anterograde endoplasmic reticulum-to-Golgi transport, as a novel regulator of WPB elongation. To elucidate novel determinants of WPB morphology we explored endothelial SEC22B interaction partners in a mass spectrometry-based approach, identifying the Golgi SNARE Syntaxin 5 (STX5). We established STX5 knockdown in endothelial cells using shRNA-dependent silencing and analyzed WPB and Golgi morphology, using confocal and electron microscopy. STX5-depleted endothelial cells exhibited extensive Golgi fragmentation and decreased WPB length, which was associated with reduced intracellular VWF levels, and impaired stimulated VWF secretion. However, the secretion-incompetent organelles in shSTX5 cells maintained WPB markers such as Angiopoietin 2, P-selectin, Rab27A, and CD63. In brief, we identified SNARE protein STX5 as a novel regulator of WPB biogenesis.

Introduction von Willebrand factor (VWF) is a hemostatic glycoprotein that is primarily synthesized in endothelial cells (EC) and acts as a factor VIII chaperone as well as an adhesive grid for thrombus formation.1 Decreased VWF plasma levels or mutations in the VWF gene can cause von Willebrand disease (VWD), the most common bleeding disorder.2 VWF undergoes a multistep maturation process that involves dimerization in the endoplasmic reticulum (ER), followed by multimerization and proteolytic processing in the Golgi.1 Smaller VWF multimers are continuously secreted (primarily) at the basolateral surface via the constitutive pathway, while larger VWF multimers are condensed into

storage organelles emerging from the trans-Golgi network (TGN): the Weibel-Palade bodies (WPB).3 WPB biogenesis is tightly linked to VWF synthesis, which is highlighted by the absence of WPB in VWF knockout EC, and their de novo formation in non-EC by ectopic VWF expression.4–6 Besides VWF, WPB contain inflammatory and angiogenic proteins, and recruit essential transport and exocytotic machinery.7,8 WPB exocytosis occurs via basal (continuous) or regulated (stimulus-induced) secretion pathways, both predominantly targeting the apical surface facing the blood vessel lumen.3 WPB play an important role during primary hemostasis as their release ensures the immediate delivery of VWF (and other molecules) in the vessel lumen in response to injury, whereupon VWF

Haematologica | 107 August 2022

1827

ARTICLE - Syntaxin 5 controls Weibel-Palade body size

M. Kat et al.

tubules unfurl into long VWF strings on the apical surface, which subsequently become decorated by platelets.8 WPB have a distinct, elongated morphology: the cigarshaped structure is composed of densely packed helical tubules of VWF multimers running along the length of the organelle enwrapped by a tightly fitted endomembrane.9,10 Both the length and adhesive properties of VWF strings correlated with WPB length; shorter WPB generate shorter VWF strings, with lower adhesive capacity for platelets and plasma VWF.11,12 However, what drives their distinct morphology is still largely unknown. The range in WPB size was defined by the VWF quanta model, which describes how during biogenesis VWF nanoclusters of a discrete length (i.e., quanta) are co-packaged in variable numbers at the Golgi, ultimately determining the length of the WPB.13 Although WPB length is known to be determined by Golgi ribbon architecture as well as by levels of VWF synthesis,13,14 only recently has the control of VWF progression through the early secretory pathway been appreciated as a determinant of WPB length.15–17 In anterograde transport vesicles bud off at ER exit sites, containing specific cargo en route to the Golgi. During this process soluble N-ethylmaleimide-sensitive fusion protein attachment protein receptors (SNARE) are incorporated in the vesicle membrane (i.e., v-SNARE), which can form complexes with Golgi-associated target membrane SNARE (t-SNARE) to facilitate membrane fusion and cargo release in the Golgi lumen.18,19 One study showed that the ARF guanine-nucleotide exchange factor (GEF) Golgi Brefeldin A-resistant GEF 1 (GBF1) modulates vesicle fission at the ER and TGN, having an impact on WPB size by controlling anterograde VWF transport and WPB segregation from the TGN.15 Exocytic SNARE proteins play a key role in WPB exocytosis,20–24 and some have also been associated with VWF plasma levels and severity of VWD.25,26 The role of SNARE proteins in WPB biogenesis and VWF trafficking, however, remains elusive. We have recently characterized the vSNARE SEC22B as a novel WPB size regulator through its role in anterograde VWF transport and support of elongated Golgi morphology.16 In the current study, we aimed at identifying novel determinants of VWF trafficking through mapping ER-Golgi fusion machinery in EC by elucidating the SEC22B interactome. The proteomic screen revealed a plethora of potential interactors, including the SNARE protein syntaxin 5 (STX5). Knockdown of STX5 in EC resulted in extensive fragmentation of Golgi architecture, VWF retention in the ER, and significantly shorter and fewer WPB. In addition, both intracellular VWF levels and regulated WPB exocytosis were significantly suppressed, highlighting STX5 as an essential component of the machinery driving WPB biogenesis and release.

Methods Mass spectrometry analysis Sample preparation, data acquisition and data analysis were performed as previously described.27 Detailed experimental procedures are described in the Online Supplementary Material. von Willebrand factor string assay VWF string assays were essentially performed as previously described using 75,000 shCTRL or shSTX5 transduced cells per channel in gelatin-coated six-channel IBIDI µ-slides VI 0.4.28 Strings were visualized by supplementing perfusion mix with AF488-conjugated anti-VWF antibody (DAKO) at 2 µM concentration. Further experimental details can be found in the Online Supplementary Material. Secretion assay EC were transduced and grown in six-well plates and cultured for 7 days prior to the experiment with regular replacement of medium. Basal VWF release was determined as unstimulated secretion over 48 h. For histamine-induced secretion cells were starved for 30 min in M-199 with 1% bovine serum albumin and subsequently stimulated with 100 µM histamine (Sigma-Aldrich; H7125) for 1 h. Lysates were obtained in lysis buffer (phosphatebuffered saline, 1% Triton X-100) supplemented with Halt protease and a phosphatase inhibitor cocktail. VWF levels were determined by enzyme-linked immunosorbent assay as described previously.21 Antibodies and DNA constructs are listed in Online Supplementary Tables S1 and S2, respectively. Additional methods can be found in the Online Supplementary Material.

Results The SEC22B interactome in endothelial cells contains SNARE proteins involved in anterograde, retrograde and intra-Golgi protein trafficking. To map the composition of the ER-to-Golgi SNARE networks that control VWF trafficking and WPB biogenesis, we employed unbiased affinity purification-mass spectrometry in EC using mEGFP-tagged SEC22B as bait. A total of 841 proteins were significantly enriched in mEGFP-SEC22B compared to the mEGFP control (Figure 1A, B, Online Supplementary Table S3). Gene ontology enrichment analysis revealed that the most prominent complexes within the cellular components ontology were ‘membrane protein’ ‘inner mitochondrial membrane protein’, and ‘SNARE’ (Figure 1C). Furthermore, SEC22B and its potential interacting partners collaborated in 235 en-

Haematologica | 107 August 2022

1828

ARTICLE - Syntaxin 5 controls Weibel-Palade body size

M. Kat et al.

riched biological processes, including anterograde and retrograde ER-Golgi trafficking, intra-Golgi trafficking as well as Golgi organization (Online Supplementary Table S4). To visualize protein interactions a STRING network was generated based on high confidence interactions, showing two major clusters representing proteins that are part of the inner MT membrane protein complex or a SNARE complex, with SEC22B as part of the latter (Figure 1D). Among the significant hits there were many SNARE, including STX5 and GOSR2 (Golgi SNAP Receptor Complex Member 2), which are part of the anterograde SNARE complex, and STX18, which facilitates retrograde trafficking.29 Another protein complex represented in the STRING analysis comprises components of the vesicle-tethering NBAS/RINT1/ZW10 (NRZ) complex (i.e., NAPA, NBAS, SCFD1, SCFD2, ZW10 and C19orf25), which regulates SNARE complex formation of incoming vesicles on the ER membrane.30 In addition, members of the conserved oligomeric Golgi (COG) complex were identified (i.e., COG16 and COG8), suggesting a link between SEC22B and the

intra-Golgi vesicle membrane tethering complex.31,32 A last notable hit in the SEC22B interactome screen is GBF1, which was previously implicated in ER-Golgi transport of VWF.15 The SEC22B interactome analysis uncovered a large protein network containing both known and novel candidates of protein trafficking in the endothelial biosynthetic pathway. Based on interaction score and the number of edges directed to SEC22B in the STRING analysis, we selected one of the most prominent hits within the SNARE cluster, STX5, for further follow-up. STX5 takes part in both anterograde ER-to-Golgi trafficking and in retrograde intraGolgi trafficking.29 We validated the interaction between STX5 and GFP-SEC22B by immunoblotting for co-precipitated endogenous STX5 (Figure 1E), which is expressed as long (~40 kDa; STX5L) and short (~34 kDa; STX5S) isoforms, resulting from an alternative start codon.33 Furthermore, immunofluorescent STX5 staining in human umbilical vein EC showed a perinuclear localization, overlapping with the trans-Golgi network marker TGN46 (Online Supplementary Figure S1A). SEC22B also concentrated

D E

Figure 1. STX5 is part of the SEC22B interactome in endothelial cells. (A) Label-free quantification (LFQ) of SEC22B protein levels in pulldown samples determined in triplicate by mass spectrometry analysis. (B) Volcano plot of significantly enriched proteins in the mEGFP and mEGFP-SEC22B pulldown samples. Red dot represents SEC22B used as bait. Orange dots represent SNARE complex proteins based on GO:0031201 annotation. Dotted lines indicate significance thresholds (P.adjust <0.05 and |LFC|>1). (C) Bar plot of enriched GO:CC protein complexes. (D) STRING-DB analysis of the enriched proteins in mEGFP-SEC22B showing high confidence interactors (combined STRING-DB scores >0.9). Colors represent enriched complexes: SNARE complex (orange), endoplasmic reticulum (ER) protein-containing complex (dark green), Golgi transport complex (pink), Vesicle tethering complex (blue) and inner MT membrane protein complex (green). Red node indicates SEC22B. Disconnected nodes are not shown. The panel below shows a zoom of the area with high SNARE complex protein density. (E) Western blot analysis of the short (S) and long (L) isoforms of STX5 in input and pulldown (PD) samples from human umbilical vein endothelial cells expressing mEGFP or mEGFP-SEC22B. B: beads. Haematologica | 107 August 2022

1829

ARTICLE - Syntaxin 5 controls Weibel-Palade body size

M. Kat et al.

in the Golgi area, but was additionally observed in small punctae localized in a wider area around the nucleus (Online Supplementary Figure S1B), most likely representing ER exit sites and trafficking vesicles.34 Collectively, these observations suggest that STX5 and SEC22B interact at the Golgi apparatus. STX5 and SEC22B depletion in endothelial cells induces unique and shared whole proteome alterations To study the role of STX5 in EC, we silenced its expression by stable expression of short hairpin (sh)RNA targeting STX5. We determined knockdown efficiency of five shRNA targeting STX5 in comparison to a non-targeting shRNA control (shCTRL). Two shRNA (shSTX5_59826 and shSTX5_59827) efficiently reduced expression of both STX5 isoforms (Online Supplementary Figure S2A, B). To assess the impact of STX5 silencing in an unbiased manner, we explored differences on the proteomic level between shSTX5 and shSEC22B compared to shCTRL and untransduced cells, to determine the similarities and differences between silencing either of the two interacting partners in EC. Samples from the same condition clustered together in the principal component analysis, showing minimal variability between replicates (Figure 2A). After confirming knockdown of SEC22B and STX5 (Figure 2B, Online Supplementary Figure S3), we analyzed the sig-

nificantly changed proteins (Online Supplementary Table S5). Analysis of proteomic alterations between the shSTX5 and shSEC22B conditions revealed 48 overlapping proteins, which included VWF and angiopoietin-2 (Ang-2), suggesting that STX5- and SEC22B-dependent intracellular trafficking regulates the transport of multiple WPB cargo proteins (Figure 2C). Moreover, SEC22B depletion rendered a total of 176 unique significantly changed proteins, whereas in shSTX5 cells only 59 proteins were altered. To further examine these hits and assess protein co-regulation, we generated a co-expression heatmap based on Pearson correlations, which visualized four differentially regulated protein clusters (Figure 2D). These clusters revealed proteins that were mainly driven by STX5 (clusters 1 and 4) or shared between SEC22B and STX5 (cluster 2). Further inspection of cluster 3 highlighted the presence of WPB proteins, as illustrated by the decrease in intracellular VWF levels in both shSTX5 and shSEC22B, and in addition a reduction in Ang-2 and multimerin 1 (MMRN1) (Figure 2E). Rab3D and VAMP3, which have previously been implicated in WPB formation and exocytosis,23,35 were only significantly downregulated in shSEC22B cells, but their expression levels appeared to be lower in shSTX5_59826 as well. In summary, STX5 and SEC22B depletion causes overlap-

D Figure 2. Whole proteome analysis reveals clusters of differentially regulated proteins in shSTX5 and shSEC22B endothelial cells. (A) Principal component (PC) analysis of analyzed samples in triplicate. (B) Label-free quantification (LFQ) plot showing mass spectrometry-based SEC22B and STX5 protein levels. (C) Venn diagrams showing overlap between individual shRNA and number of shared proteins between shSEC22B and shSTX5 within differentially regulated proteins compared to shCTRL and untransduced conditions. Red boxes show the amount of shared proteins for which the direction of regulation was opposed (up vs. down). (D) Correlation heatmap showing Pearson coefficients of regulated proteins separated into four clusters. Row annotation indicates if a protein was regulated by SEC22B (orange), STX5 (blue) or both. (E) LFQ plots of cluster members. N.D.: not detected.

Haematologica | 107 August 2022

1830

ARTICLE - Syntaxin 5 controls Weibel-Palade body size

M. Kat et al.

ping proteomic changes, but STX5 knockdown also induces unique alterations in the proteome, highlighting the importance of STX5-mediated protein transport. STX5 silencing results in altered Weibel-Palade body length and loss of the Golgi architecture To address a potential role of STX5 in WPB biogenesis we investigated WPB and Golgi morphology in shSTX5 and shCTRL human umbilical vein EC. STX5 knockdown was validated by strongly reduced expression on western blot and (nearly) absent immunofluorescent staining upon STX5 knockdown (Figure 3A, B). Since shSTX5_59826 yields the most efficient and consistent knockdown, we selected this shRNA for further experiments. To confirm that the effects were specific for STX5 knockdown, key experiments were also performed with shSTX5_59827. VWF staining revealed that upon STX5 silencing the characteristic elongated shape of WPB was lost, while the VWF that was present concentrated in spherical granules

mostly found in the perinuclear area (Figure 3C). Assessment of immunogold-labeled VWF by electron microscopy confirmed that these VWF-positive structures were indeed short WPB (Online Supplementary Figure S4). Quantitative analysis of the size of VWF-positive structures revealed that WPB length was drastically decreased after depletion of STX5 (Figure 3D). We also observed a concomitant disintegration of the Golgi; while TGN46 staining showed a compact, continuous ribbon structure in shCTRL cells, shSTX5 cells showed TGN46 staining on dispersed, unlinked structures (Figure 3C). Quantification revealed that in the shCTRL condition approximately 70% of the cells contained a continuous, compact Golgi, but the vast majority (~90%) of Golgi structures in shSTX5 cells appeared fragmented and dispersed (Figure 3E; examples of Golgi states shown below). Co-staining of TGN46 with the cis-Golgi marker GM130 revealed a similar dispersed phenotype for the cis-Golgi in shSTX5 cells, suggesting that the entire Golgi architecture is affected by STX5 de-

Figure 3. STX5 downregulation results in decreased Weibel-Palade body length and extensive Golgi fragmentation. (A) Western blot analysis of STX5 in shCTRL and shSTX5 transduced human umbilical vein endothelial cells (α-tubulin as loading control) and quantification of STX5 knockdown efficiency (mean ± standard error of the mean (SEM), n=4, ttest, **P<0.01, ***P<0.001). Molecular weights of protein marker bands are indicated in kDa. L and S indicate long and short STX5 isoforms, respectively. (B) Immunofluorescent staining of STX5 (gray) and nuclei (blue) in shCTRL and shSTX5 cells (scale bar: 10 µm). (C) Representative immunofluorescence microscopy images (n=7 biological replicates) of staining for VWF (green), TGN46 (magenta), STX5 (gray), and nuclei (Hoechst; blue) in shCTRL and shSTX5 cells (scale bar: 40 µm). Boxed areas are magnified below (individual channels in gray scale). (D) Weibel-Palade body (WPB) length in shCTRL and shSTX5 cells measured in micrometers (µm). Each transparent dot represents a WPB; the average length of each biological replicate is shown in filled color (median ± interquartile range [IR], n=4, t-test, ****P<0.0001). (E) The percentage of cells containing a continuous or fragmented Golgi in shCTRL and shSTX5 cells (median ± IR, shCTRL n=5, shSTX5 n=7, t-test, ****P<0.0001). Examples of continuous and fragmented Golgis stained with TGN46 (gray) are shown. (F) WPB length in shSTX5 cells with continuous versus fragmented Golgi (median ± IR, n=6, t-test, ***P<0.001).

Haematologica | 107 August 2022

1831

ARTICLE - Syntaxin 5 controls Weibel-Palade body size

M. Kat et al.

pletion (Online Supplementary Figure S5). These dispersed Golgi structure in shSTX5 cells produced shorter WPB, indicating that WPB length is dependent on extended Golgi ribbon organization (Figure 3F), in agreement with previous literature.13,16 Interestingly, SEC22B was no longer apparent on the dispersed Golgi structured, and instead prominent staining surrounding the nucleus was observed (Online Supplementary Figure S6). Together these results indicate that STX5 is needed for the formation of elongated WPB, by controlling the maintenance of extended Golgi stacks that allow for loading of multiple VWF quanta into newly forming WPB. No rough endoplasmic reticulum dilation due to von Willebrand factor retention upon STX5 silencing We used transmission electron microscopy to further evaluate the morphology of WPB, Golgi, and the ER (Figure 4A). Similar as earlier noted by light microscopy, the length of the WPB (Figure 4A; WPB indicated by cyan overlay) was reduced upon STX5 silencing (Figure 4B). Golgi ribbon structures arranged in compact stacks were clearly

identifiable in shCTRL cells, whereas shSTX5 cells contained widely dispersed Golgi fragments (Figure 4A; Golgi indicated by green overlay), which were a challenge to locate despite using the centriole as a reference point (Figure 4A; centrioles indicated by magenta overlay). As a consequence, the majority of shSTX5 cells that were analyzed by transmission electron microscopy seemingly lacked elongated Golgi stacks entirely (Figure 4C). These observations, along with immunofluorescence microscopy images showing an altered morphology of the Golgi apparatus in STX5-deficient cells (Figure 3C, Online Supplementary Figure S5), support the emerging concept that formation of elongated WPB depends on the presence of a TGN with extended ribbons to package multiple VWF quanta in nascent WPB.13,14 We previously showed that SEC22B depletion causes retention of VWF inside the ER lumen and massive dilation of the rough ER as the SEC22B-dependent anterograde trafficking pathway to the Golgi is blocked.16 STX5 depletion also caused some retention of VWF in the ER as judged by the overlap of a perinuclear pool of VWF with

Figure 4. STX5 depletion induces morphological changes in Golgi and rough endoplasmic reticulum. (A) Transmission electron microscopy images of shCTRL and shSTX5 cells. Boxed regions are magnified below outlined with corresponding colors. WeibelPalade bodies (WPB) and rough endoplasmic reticulum (ER) sheets are highlighted with cyan and yellow overlays, respectively. In the image on the right, Golgi segments are highlighted in green and centrioles in magenta. Scale bars represent 3 µm and 1 µm as indicated. (B) WPB length measured in transmission electron microscopy images (median ± interquartile range [IR], t-test, ***P<0.001). (C) The percentage of cells with Golgi ribbons versus no Golgi ribbons. (D) The rough ER sheet width measured in shCTRL and shSTX5 cells (median ± IR, n=3, t-test, ns: not significant). Haematologica | 107 August 2022

1832

ARTICLE - Syntaxin 5 controls Weibel-Palade body size

M. Kat et al.

the rough ER marker Protein disulfide isomerase A3 (PDI) (Online Supplementary Figure S7). Morphological analysis of the ER revealed that STX5 silencing did not cause VWF retention to the same extent as observed previously in SEC22B-depleted cells,16 as no large round dense (VWFpositive) structures were detected inside the ER lumen (Figure 4A; rough ER sheets indicated with yellow overlay). Modest dilation of the rough ER was occasionally observed upon STX5 knockdown (Online Supplementary Figure S8; rough ER sheets indicated with yellow overlay), but on average the luminal width of rough ER cisternae was not significantly altered in the shSTX5 cells (Figure 4D). Overall, these results indicate that while STX5 and SEC22B depletion have a comparable impact on WPB and Golgi morphology, the increase of ER volume to accommodate accumulating (secretory) proteins is unique to SEC22B knockdown.

Characteristic Weibel-Palade body markers co-localize with von Willebrand factor-positive structures in shSTX5 cells Since WPB contain an array of proteins besides VWF,8 we examined whether the localization of these proteins to WPB depends on STX5 by analyzing several WPB cargo and WPB membrane-associated proteins using immunofluorescence microscopy. The soluble angiogenic mediator Ang-2 and transmembrane adhesion receptor P-selectin (also referred to as CD62P) are sorted to the WPB during biogenesis at the TGN.36–38 Confocal imaging showed that Ang-2 and P-selectin both localized at WPB in shCTRL as well as shSTX5 cells, implying that trafficking of these proteins from the Golgi to their storage compartment continues despite the loss of elongated Golgi architecture upon STX5 silencing (Figure 5A, B). We also investigated localization of two late-stage WPB markers: Rab27A and CD63, each representative of a separate route for postGolgi protein arrival to WPB. Rab27A, an established matu-

Figure 5. Weibel-Palade body cargo proteins Angiopoietin-2 and P-selectin localize to small Weibel-Palade bodies in STX5-depleted cells. (A) Immunofluorescent staining of Angiopoietin 2 (Ang2, cyan), von Willebrand factor (VWF, magenta), and STX5 (green) and (B) P-selectin (CD62P, cyan), VWF (magenta), and TGN46 (green) in shCTRL and shSTX5 human umbilical vein endothelial cells (scale bar: 10 µm). Individual channels are shown below in gray scale. Boxed areas are magnified on the right. Yellow arrowheads indicate Ang-2-positive and P-selectin-positive WPB in (A) and (B), respectively.

Haematologica | 107 August 2022

1833

ARTICLE - Syntaxin 5 controls Weibel-Palade body size

M. Kat et al.

Figure 6. Normal recruitment of maturation marker Rab27A and endosome-derived CD63 to small Weibel-Palade bodies upon STX5 knockdown. (A) Immunofluorescent staining of Rab27A (cyan), VWF (magenta), and STX5 (green) and (B) CD63 (cyan), VWF (magenta), and TGN46 (green) in shCTRL and shSTX5 human umbilical vein endothelial cells (scale bar: 10 µm). Individual channels are shown below in gray scale. Boxed areas are magnified on the right. Yellow arrowheads indicate Rab27A-positive and CD63-positive Weibel-Palade bodies in (A) and (B), respectively.

ration marker that is mobilized from the cytosol and recruits essential exocytosis machinery to the WPB membrane,39–41 was present on the small WPB in STX5-depleted EC (Figure 6A). CD63, an integral membrane protein that reaches the WPB from late endosomes via an AP-3-dependent mechanism,20,42,43 was observed on spherical compartments, likely endosomes, and elongated WPB in shCTRL cells as well as rounded WPB in shSTX5 cells, indicating that its recruitment from endosomes was not impaired by STX5 depletion (Figure 6B). Thus, localization of a selection of established WPB markers remains unchanged after STX5 silencing, indicating that trafficking of cytosolic-, Golgi- and endosome-derived cargo is not impaired.

Short Weibel-Palade bodies in STX5-depleted cells are secretion-incompetent A previous report described that WPB size is correlated to the length and adhesive properties of VWF strings.11 Since loss of STX5 expression results in shorter WPB, we hypothesized that stimulated release of these WPB results in the formation of shorter VWF strings. To test this we perfused EC under 2.5 dynes/cm2 flow conditions with histamine to induce WPB exocytosis and a fluorescent VWF antibody to detect VWF strings. A large number of VWF strings of varying lengths were formed on the surface of shCTRL cells, but strikingly no VWF string formation was seen for shSTX5 cells (Figure 7A). Confocal analysis of unstimulated conditions revealed that small WPB were present in shSTX5 cells, predominantly localized in the perinuclear area (Online Supplementary Figure S9). However, the absence of VWF string formation indicated that

Haematologica | 107 August 2022

1834

M. Kat et al.

ARTICLE - Syntaxin 5 controls Weibel-Palade body size

F E

Figure 7. STX5 silencing impairs von Willebrand factor string formation. (A) von Willebrand factor (VWF) string assay with shCTRL and shSTX5-transduced human umbilical vein endothelial cells following stimulation with 100 µM histamine. Extracellular VWF is shown in green in differential interference contrast and fluorescence overlay images and in black in inverted fluorescence images below (scale bar: 100 µm). (B) The average number of Weibel-Palade bodies (WPB) per cell in shCTRL and shSTX5 cells (mean ± standard error of mean [SEM], n=3, t-test, *P<0.05). (C) Western blot analysis of VWF and STX5 expression in shCTRL and shSTX5 cells (α-tubulin as loading control). L and S indicate long and short STX5 isoforms, respectively. Molecular weights of protein marker bands are indicated in kDa. (D) VWF multimer analysis on media secreted over 24 h by shCTRL and shSTX5 cells. D: dimer, LMW: low molecular weight, HMW: high molecular weight, UL: ultra large. N.R.: non-reducing conditions, R.: reducing conditions. (E) Band intensity profile plot of representative VWF multimer patterns in arbitrary units (A.U.). (F) Quantification of the area under the curve (AUC) of D, LMW, HMW, and UL multimers as a percentage of total AUC per curve (mean ± SEM, n=3, two-way analysis of variance, ****P<0.0001). Haematologica | 107 August 2022

1835

ARTICLE - Syntaxin 5 controls Weibel-Palade body size

M. Kat et al.

these WPB are secretion-incompetent upon stimulation. Immunofluorescent analysis of WPB in shSTX5 and shCTRL cells showed that in addition to size, the average number of WPB per cell was significantly reduced upon STX5 knockdown (Figure 7B). To examine whether the observed decrease in WPB numbers was due to downregulated VWF expression in STX5-depleted cells, we quantified intracellular VWF levels by western blot and enzyme-linked immunosorbent assay (Figure 7C, Online Supplementary Figure S10A). In line with the reduction of WPB, intracellular VWF levels were significantly reduced upon STX5 silencing, which could indicate a decreased expression, or an increased degradation or secretion. Consistent with the VWF string assay data histamine-stimulated secretion of VWF was sharply reduced following STX5 silencing, emphasizing that the small WPB in STX5 knockdown cells are stimulus-insensitive (Online Supplementary Figure S10B, C). Unstimulated VWF secretion, which is a composite of basal, WPB-derived secretion and constitutive secretion, was decreased, but not proportionate to stored VWF (Online Supplementary Figure S10D, E). Subsequent VWF multimer analysis of the unstimulated releasates from shCTRL and shSTX5 cells revealed a remarkable change in multimer size composition, with relatively more VWF dimers and low molecular weight multimers, but less high molecular weight and ultra-large multimers secreted by STX5-depleted cells (Figure 7D-F). Notably, shSTX5-derived VWF dimers and low molecular weight multimers showed a slightly reduced mobility in gel electrophoresis, possibly indicating differences in post-translational modification. Collectively, these data indicate that loss of STX5 has critical consequences for WPB biogenesis, VWF string formation, secretion, and multimerization, indicating that this SNARE protein is a crucial player in the endothelial secretory pathway.

Discussion SNARE proteins constitute the key machinery that promotes membrane fusion during vesicle transport.19 We have previously identified several SNARE proteins involved in WPB maturation and exocytosis, while recently we have shown the critical role of SEC22B, an ER-to-Golgi SNARE, in VWF trafficking and WPB formation.16,20,21 In this study, a mass spectrometry-based approach elucidated the SEC22B interactome in EC and identified potential novel regulators of WPB biogenesis. We focused on STX5, a cognate SNARE protein that is primarily found in the Golgi membrane and has been shown to facilitate ER-to-Golgi and intra-Golgi protein trafficking.29 By silencing STX5 expression, we discovered a severe defect in WPB biogenesis, accompanied by fragmentation of the Golgi and

abrogation of secretagogue-induced VWF release. Our data point to a crucial role for STX5-containing SNARE complexes in the ability of EC to efficiently store and secrete VWF. The importance of VWF transport in WPB biogenesis is underlined by previous studies showing that WPB size depends on anterograde VWF delivery and Golgi morphology, and ultimately determines their functionality in hemostasis.11-13,44 Recently, novel players have been identified that modulate biosynthetic pathways crucially involved in WPB biogenesis.15,16 The Arf GTPases Arf1 and Arf4 and their GEF GBF1 are essential for the formation of WPB through regulation of membrane fission.15 GBF1 deficiency was accompanied by the formation of giant, secretion-incompetent WPB. Contrarily, v-SNARE SEC22B regulates ER-to-Golgi transport by facilitating vesicle fusion and depletion of SEC22B in EC blocks anterograde VWF transport and results in fragmented Golgi structures and shorter WPB.16 To this we can now add STX5, the depletion of which shows considerable phenotypic overlap with that of SEC22B-deficient cells. These studies indicate that regulators of vesicle fission and fusion are essential components of the trafficking pathways utilized for VWF multimer assembly as well as WPB biogenesis. Interestingly, GBF1 was also a hit in our SEC22B interactome screen, suggesting that there are opportunities for crosstalk between these two pathways. STX5 is an integral member of the Golgi apparatus and can form complexes with SEC22B, which is localized on ER-derived vesicles, facilitating their fusion with the Golgi (Figure 8).18,29 While STX5L localizes predominantly to the ER, STX5S has been shown to be crucial for intra-Golgi traffic, but not Golgi morphology.45 We observed that STX5 was mainly present in the Golgi, whereas SEC22B exhibited a broader localization. Knockdown of both STX5 isoforms caused fragmentation of Golgi stacks into dispersed mini-stacks, similar to the effect of SEC22B depletion.16 Therefore, we hypothesize that Golgi fragmentation is caused by the reduced delivery of membrane or structural proteins that maintain Golgi architecture from anterograde ER-derived vesicle transport. Alternatively, as STX5 has been reported to participate in Golgi reassembly after mitosis,46 reduced STX5 levels in EC could potentially interfere with the post-mitotic Golgi restoration, leading to permanent Golgi dispersal. Golgi architecture has a large impact on WPB biogenesis and elongation. According to the model first proposed by Ferraro et al.,16 continuous, extended Golgi ribbons allow for the integration of more VWF quanta, resulting in elongated WPB, while Golgi ministacks give rise to shorter WPB.13 In keeping with the VWF quanta model, WPB length is significantly reduced in STX5-depleted cells with dispersed Golgi (Figure 8). VWF release can be separated into three modes: consti-

Haematologica | 107 August 2022

1836

ARTICLE - Syntaxin 5 controls Weibel-Palade body size

M. Kat et al.

Figure 8. Proposed role of STX5 in the endothelial secretory pathway. Schematic overview of the endothelial secretory pathway (top left panel) showing von Willebrand factor (VWF, green) monomers form dimers in the endoplasmic reticulum (ER); the dimers are transported in vesicles to the Golgi (boxed area). The lower left panel depicts the proposed roles of SEC22B (red) and STX5 (blue) in this process. SEC22B, which is incorporated into the membrane of a vesicle budding off from the ER, forms a SNARE complex with STX5 present on the Golgi, facilitating membrane fusing and releasing the cargo (i.e., VWF dimers) into the Golgi lumen. Once arrived at the Golgi, VWF dimers multimerize and form quanta that arrive in the trans-Golgi network (TGN) to be packaged into a Weibel-Palade body (WPB). The WPB are transported to the plasma membrane where they can be secreted. Upon stimulation the WPB undergoes regulated exocytosis and releases its content into the vasculature, while VWF multimers form long strings. If STX5 is silenced (top right panel), more VWF is observed in the ER and the entire Golgi apparatus becomes fragmented and dispersed. Upon STX5 depletion, ER-derived vesicles can no longer fuse efficiently with the Golgi membrane (lower right panel) and only a limited amount of VWF reaches the TGN. WPB size is strongly reduced and regulated VWF secretion is diminished due to inefficient VWF anterograde transport and fragmented TGN morphology.

tutive, basal and regulated (stimulus-induced) VWF secretion.3 The WPB compartment ensures a rapid delivery of VWF to form strings and initiate hemostasis.7 Interestingly, histamine-stimulated VWF secretion was severely affected in STX5-depleted cells, as evidenced by the (nearly) absent string formation and secretion. This could be explained by the strongly depleted stimulus-sensitive compartment, i.e., reduced number of peripheral WPB, decrease in WPB length and reduced overall VWF expression levels. Recently, it has been shown that EC control functional responses by size selection of WPB during exocytosis, with a preference for longer WPB to undergo exocytosis following stimulation with a subset of secretagogues, in order to control functional responses.44 Thus, the relative unresponsiveness to histamine could also be a consequence of the reduced length of WPB in shSTX5 cells. Unstimulated VWF release was also decreased upon STX5 knockdown; however, the amount of secreted VWF relative to stored VWF was proportionally increased compared to the control, suggesting that basal secretion continues and is even slightly increased. Moreover, the

secreted VWF contained fewer larger multimers, indicating a multimerization defect. The observed shift in size of VWF dimers could perhaps be explained by incomplete glycosylation of VWF as a result of mislocalized glycosylation enzymes that normally reside in the different Golgi stacks of the Golgi.47 Therefore, we can hypothesize that an essential role of STX5 is to maintain a continuous Golgi from which properly multimerized VWF with post-translational modifications can be stored in elongated organelles, which have the capacity to mature into a stimulus-sensitive WPB. Partial quantitative VWF deficiency, known as VWD type 1 (VWD1), is the most prevalent subtype of VWD.2 A sizeable number of pathogenic VWF mutations in VWD1 lead to reduced plasma levels due to retention within EC,48–50 which is often accompanied by lower WPB numbers and loss of their characteristic elongated morphology.51,52 In about 25% of VWD1 patients and about 60% of individuals with “low VWF” no causal mutation can be detected in their VWF gene. It has been postulated that the genetic regulation of VWF levels largely depends on an interplay of several quantitative trait loci that influence VWF synthesis

Haematologica | 107 August 2022

1837

ARTICLE - Syntaxin 5 controls Weibel-Palade body size

M. Kat et al.

and clearance.53 While a large portion of the heritable variation in VWF is still unaccounted for, some of these modifiers include SNARE proteins that can influence WPB exocytosis.25,26,53,54 Given that a substantial portion of quantitative VWF deficiencies in the group with VWF mutations are characterized by defects in intracellular trafficking of VWF, it is plausible that VWD and low VWF pathogenesis in cases without damaging VWF mutations may also be driven by genetic variations in cellular components that control correct progression of VWF through the endothelial secretory pathway. The SEC22B interactome determined in this study uncovered an extensive network consisting of SNARE and their regulators and components of a number of vesicle tethering complexes that operate at the ER (NRZ complex), Golgi (COG complex) and post-Golgi (exocyst). Further cellular studies are needed to elucidate their importance in VWF trafficking. Potentially, data from such “trafficome” studies can be used to inform genetic and genomic studies that focus on identifying modifiers of VWF plasma levels in patients who suffer from bleeding or thrombosis. Disclosures No conflicts of interest to disclose.

Contributions MK, EK, AJH, PEB, AAM, FPJvA, and JO performed research and analyzed data; DG provided vital reagents; MvdB provided vital expertise; MK, EK, CM, JV, and RB designed the research and wrote the paper. Acknowledgments We would like to thank Dr. R.I. Koning and Dr. C.R. Jost for their critical reading of the manuscript. Funding Work in our laboratory was funded by grants from the Landsteiner Stichting voor Bloedtransfusie Research (LSBR-1707) and the Dutch Thrombosis Foundation (TSN 2017-01). Data-sharing statement The mass spectrometry proteomics data, including the .raw mass spectrometry files and search/identification files obtained with MaxQuant have been deposited with the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the PRIDE partner repository with the dataset identifier PXD027516.

References 1. Sadler JE. von Willebrand factor assembly and secretion. J Thromb Haemost. 2009;7(Suppl 1):24-27. 2. Leebeek FWG, Eikenboom JCJ. von Willebrand’s disease. N Engl J Med. 2016;375(21):2067-2080. 3. Lopes da Silva M, Cutler DF. von Willebrand factor multimerization and the polarity of secretory pathways in endothelial cells. Blood. 2016;128(2):277-285. 4. Wagner DD, Marder VJ. Biosynthesis of von Willebrand protein by human endothelial cells. J Biol Chem. 1983;258(4):2065-2067. 5. Voorberg J, Fontijn R, Calafat J, Janssen H, van Mourik JA, Pannekoek H. Biogenesis of von Willebrand factor-containing organelles in heterologous transfected CV-1 cells. EMBO J. 1993;12(2):749-758. 6. Schillemans M, Kat M, Westeneng J, et al. Alternative trafficking of Weibel-Palade body proteins in CRISPR/Cas9-engineered von Willebrand factor-deficient blood outgrowth endothelial cells. Res Pract Thromb Haemost. 2019;3(4):718-732. 7. Schillemans M, Karampini E, Kat M, Bierings R. Exocytosis of Weibel–Palade bodies: how to unpack a vascular emergency kit. J Thromb Haemost. 2019;17(1):6-18. 8. McCormack JJ, Lopes da Silva M, Ferraro F, Patella F, Cutler DF. Weibel-Palade bodies at a glance. J Cell Sci. 2017;130(21):3611-3617. 9. Valentijn KM, Valentijn JA, Jansen KA, Koster AJ. A new look at Weibel-Palade body structure in endothelial cells using electron tomography. J Struct Biol. 2008;161(3):447-458. 10. Berriman JA, Li S, Hewlett LJ, et al. Structural organization of Weibel-Palade bodies revealed by cryo-EM of vitrified endothelial cells. Proc Natl Acad Sci U S A. 2009;106(41):17407-17412.

11. Ferraro F, Lopes-da-Silva M, Grimes W, et al. Weibel-Palade body size modulates the adhesive activity of its von Willebrand factor cargo in cultured endothelial cells. Sci Rep. 2016;6(1):32473. 12. Nightingale TD, McCormack JJ, Grimes W, et al. Tuning the endothelial response: differential release of exocytic cargos from Weibel-Palade bodies. J Thromb Haemost. 2018;16(9):1873-1886. 13. Ferraro F, Kriston-Vizi J, Metcalf DJ, et al. A two-tier Golgibased control of organelle size underpins the functional plasticity of endothelial cells. Dev Cell. 2014;29(3):292-304. 14. Mourik MJ, Faas FGA, Zimmermann H, Voorberg J, Koster AJ, Eikenboom J. Content delivery to newly forming Weibel-Palade bodies is facilitated by multiple connections with the Golgi apparatus. Blood. 2015;125(22):3509-3516. 15. Lopes-da-Silva M, McCormack JJ, Burden JJ, Harrison-Lavoie KJ, Ferraro F, Cutler DF. A GBF1-dependent mechanism for environmentally responsive regulation of ER-Golgi transport. Dev Cell. 2019;49(5):786-801. 16. Karampini E, Bürgisser PE, Olins J, et al. Sec22b determines Weibel-Palade body length by controlling anterograde ER-Golgi transport. Haematologica. 2021;106(4):1138-1147. 17. Karampini E, Bierings R, Voorberg J. Orchestration of primary hemostasis by platelet and endothelial lysosome-related organelles. Arterioscler Thromb Vasc Biol. 2020;40(6):1441-1453. 18. Brandizzi F, Barlowe C. Organization of the ER–Golgi interface for membrane traffic control. Nat Rev Mol Cell Biol. 2013;14(6):382-392. 19. Jahn R, Scheller RH. SNAREs--engines for membrane fusion. Nat Rev Mol Cell Biol. 2006;7(9):631-643.

Haematologica | 107 August 2022

1838

ARTICLE - Syntaxin 5 controls Weibel-Palade body size

M. Kat et al.

20. Karampini E, Schillemans M, Hofman M, et al. Defective AP-3dependent VAMP8 trafficking impairs Weibel-Palade body exocytosis in Hermansky-Pudlak syndrome type 2 blood outgrowth endothelial cells. Haematologica. 2019;104(10):2091-2099. 21. Schillemans M, Karampini E, van den Eshof BL, et al. WeibelPalade body localized Syntaxin-3 modulates von Willebrand factor secretion from endothelial cells. Arterioscler Thromb Vasc Biol. 2018;38(7):1549-1561. 22. Zhu Q, Yamakuchi M, Lowenstein CJ. SNAP23 regulates endothelial exocytosis of von Willebrand factor. PLoS One. 2015;10(8):e0118737. 23. Pulido IR, Jahn R, Gerke V. VAMP3 is associated with endothelial Weibel-Palade bodies and participates in their Ca(2+)dependent exocytosis. Biochim Biophys Acta. 2011;1813(5):1038-1044. 24. van Breevoort D, Snijders AP, Hellen N, et al. STXBP1 promotes Weibel-Palade body exocytosis through its interaction with the Rab27A effector Slp4-a. Blood. 2014;123(20):3185-3194. 25. van Loon JE, Sanders YV, de Wee EM, Kruip MJ, de Maat MP, Leebeek FW. Effect of genetic variation in STXBP5 and STX2 on von Willebrand factor and bleeding phenotype in type 1 von Willebrand disease patients. PLoS One. 2012;7(7):e40624. 26. Van Loon J, Dehghan A, Weihong T, et al. Genome-wide association studies identify genetic loci for low von Willebrand factor levels. Eur J Hum Genet. 2016;24(7):1035-1040. 27. Schillemans M, Karampini E, Hoogendijk AJ, et al. Interaction networks of Weibel-Palade body regulators syntaxin-3 and syntaxin binding protein 5 in endothelial cells. J Proteomics. 2019;205:103417. 28. Ercig B, Graça NAG, Kangro K, et al. N-glycan–mediated shielding of ADAMTS13 prevents binding of pathogenic autoantibodies in immune-mediated TTP. Blood. 2021;137(19):2694-2698. 29. Linders PTA, van der Horst C, Ter Beest M, van den Bogaart G. Stx5-mediated ER-Golgi transport in mammals and yeast. Cells. 2019;8(8):780. 30. Tagaya M, Arasaki K, Inoue H, Kimura H. Moonlighting functions of the NRZ (mammalian Dsl1) complex. Front Cell Dev Biol. 2014;2:25. 31. Willett R, Kudlyk T, Pokrovskaya I, et al. COG complexes form spatial landmarks for distinct SNARE complexes. Nat Commun. 2013;4(1):1553. 32. Kudlyk T, Willett R, Pokrovskaya ID, Lupashin V. COG6 interacts with a subset of the Golgi SNAREs and is Important for the Golgi complex integrity. Traffic. 2013;14(2):194-204. 33. Hui N, Nakamura N, Sönnichsen B, Shima DT, Nilsson T, Warren G. An isoform of the Golgi t-SNARE, syntaxin 5, with an endoplasmic reticulum retrieval signal. Mol Biol Cel.l 1997;8(9):1777-1787. 34. Mancias JD, Goldberg J. The transport signal on Sec22 for packaging into COPII-coated vesicles is a conformational epitope. Mol Cell. 2007;26(3):403-414. 35. Knop M, Aareskjold E, Bode G, Gerke V. Rab3D and annexin A2 play a role in regulated secretion of vWF, but not tPA, from endothelial cells. EMBO J. 2004;23(15):2982-2992. 36. Fiedler U, Scharpfenecker M, Koidl S, et al. The Tie-2 ligand Angiopoietin-2 is stored in and rapidly released upon stimulation from endothelial cell Weibel-Palade bodies. Blood. 2004;103(11):4150-4156. 37. Bonfanti R, Furie BC, Furie B, Wagner DD. PADGEM (GMP140) is a component of Weibel-Palade bodies of human endothelial cells.

Blood. 1989;73(5):1109-1112. 38. McEver RP, Beckstead JH, Moore KL, Marshall-Carlson L, Bainton DF. GMP-140, a platelet alpha-granule membrane protein, is also synthesized by vascular endothelial cells and is localized in Weibel-Palade bodies. J Clin Invest. 1989;84(1):92-99. 39. van Breevoort D, van Agtmaal EL, Dragt BS, et al. Proteomic screen identifies IGFBP7 as a novel component of endothelial cell-specific Weibel-Palade bodies. J Proteome. Res 2012;11(5):2925-2936. 40. Hannah MJ, Hume AN, Arribas M, et al. Weibel-Palade bodies recruit Rab27 by a content-driven, maturation-dependent mechanism that is independent of cell type. J Cell Sci. 2003;116(Pt 19):3939-3948. 41. Kat M, Bürgisser PE, Janssen H, et al. GDP/GTP exchange factor MADD drives activation and recruitment of secretory Rab GTPases to Weibel-Palade bodies. Blood Adv. 2021;5(23):5116-5127. 42. Kobayashi T, Vischer UM, Rosnoblet C, et al. The tetraspanin CD63/lamp3 cycles between endocytic and secretory compartments in human endothelial cells. Mol Biol Cell. 2000;11(5):1829-1843. 43. Vischer U, Wagner D. CD63 is a component of Weibel-Palade bodies of human endothelial cells. Blood. 1993;82(4):1184-1191. 44. McCormack JJ, Harrison-Lavoie KJ, Cutler DF. Human endothelial cells size-select their secretory granules for exocytosis to modulate their functional output. J Thromb Haemost. 2020;18(1):243-254. 45. Linders PTA, Gerretsen ECF, Ashikov A, et al. Congenital disorder of glycosylation caused by starting site-specific variant in syntaxin-5. Nat Commun. 2021;12(1):6227. 46. Rabouille C, Kondo H, Newman R, Hui N, Freemont P, Warren G. Syntaxin 5 is a common component of the NSF- and p97mediated reassembly pathways of Golgi cisternae from mitotic golgi fragments in vitro. Cell. 1998;92(5):603-610. 47. Linders PTA, Peters E, Ter Beest M, Lefeber DJ, van den Bogaart G. Sugary logistics gone wrong: membrane trafficking and congenital disorders of glycosylation. Int J Mol Sci. 2020;21(13):4654. 48. Goodeve A. Diagnosing von Willebrand disease: genetic analysis. Hematol Am Soc Hematol Educ Progr. 2016;2016(1):678-682. 49. de Jong A, Eikenboom J. Von Willebrand disease mutation spectrum and associated mutation mechanisms. Thromb Res. 2017;159:65-75. 50. Eikenboom J, Hilbert L, Ribba AS, et al. Expression of 14 von Willebrand factor mutations identified in patients with type 1 von Willebrand disease from the MCMDM-1VWD study. J Thromb Haemost. 2009;7(8):1304-1312. 51. Wang J-W, Bouwens EAM, Pintao MC, et al. Analysis of the storage and secretion of von Willebrand factor in blood outgrowth endothelial cells derived from patients with von Willebrand disease. Blood. 2013;121(14):2762-2772. 52. Starke RD, Paschalaki KE, Dyer CEF, et al. Cellular and molecular basis of von Willebrand disease: studies on blood outgrowth endothelial cells. Blood. 2013;121(14):2773-2784. 53. Swystun LL, Lillicrap D. Genetic regulation of plasma von Willebrand factor levels in health and disease. J Thromb Haemost. 2018;16(12):2375-2390. 54. Smith NL, Chen M-H, Dehghan A, et al. Novel associations of multiple genetic loci with plasma levels of factor VII, factor VIII, and von Willebrand factor. Circulation. 2010;121(12):1382-1392.

Haematologica | 107 August 2022

1839

ARTICLE - Immunology

Antibody response after vaccination against SARS-CoV-2 in adults with hematological malignancies: a systematic review and meta-analysis Nico Gagelmann,1 Francesco Passamonti,2 Christine Wolschke,1 Radwan Massoud,1 Christian Niederwieser,1 Raissa Adjallé,1 Barbara Mora,2 Francis Ayuk1 and Nicolaus Kröger1 Department of Stem Cell Transplantation, University Medical Center Hamburg-Eppendorf, Hamburg, Germany and 2University of Insubria, Varese, Italy

Correspondence: Nico Gagelmann n.gagelmann@uke.de Received: October 7, 2021. Accepted: December 7, 2021. Pre-published: December 16, 2021 https://doi.org/10.3324/haematol.2021.280163 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

Abstract Vaccines against SARS-CoV-2 have shown remarkable efficacy and thus constitute an important preventive option against coronavirus disease 2019 (COVID-19), especially in fragile patients. We aimed to systematically analyze the outcomes of patients with hematological malignancies who received vaccination and to identify specific groups with differences in outcomes. The primary end point was antibody response after full vaccination (2 doses of mRNA or one dose of vectorbased vaccines). We identified 49 studies comprising 11,086 individuals. Overall risk of bias was low. The pooled response for hematological malignancies was 64% (95% confidence interval [CI]: 59-69; I²=93%) versus 96% (95% CI: 92-97; I²=44%) for solid cancer and 98% (95% CI: 96-99; I²=55%) for healthy controls (P<0.001). Outcome was different across hematological malignancies (P<0.001). The pooled response was 50% (95% CI: 43-57; I²=84%) for chronic lymphocytic leukemia, 76% (95% CI: 67-83; I²=92%) for multiple myeloma, 83% (95% CI: 69-91; I²=85%) for myeloproliferative neoplasms, 91% (95% CI: 82-96; I²=12%) for Hodgkin lymphoma, and 58% (95% CI: 44-70; I²=84%) for aggressive and 61% (95% CI: 48-72; I²=85%) for indolent non-Hodgkin lymphoma. The pooled response for allogeneic and autologous hematopoietic cell transplantation was 82% and 83%, respectively. Being in remission and prior COVID-19 showed significantly higher responses. Low pooled response was identified for active treatment (35%), anti-CD20 therapy ≤1 year (15%), Bruton kinase inhibition (23%), venetoclax (26%), ruxolitinib (42%), and chimeric antigen receptor T-cell therapy (42%). Studies on timing, value of boosters, and long-term efficacy are needed. This study is registered with PROSPERO (clinicaltrials gov. Identifier: CRD42021279051).

Introduction For patients with hematological malignancies, risk of death among adult patients who had coronavirus disease 2019 (COVID-19) was estimated to be 34% in predominantly hospitalized patients, being even higher for patients at older age.1 Similarly, patients undergoing hematopoetic cell transplantation or cellular therapy were found to be at increased risk for lower respiratory tract disease, intensive care admission, and death.2 In addition, COVID-19 elicits an impaired antibody response against SARS-CoV2 in hematological malignancies.3 Overall, this emphasized the need for stringent surveillance and urgent identification of therapeutic and preventive options.4

Vaccines against SARS-CoV-2 constitute such an important preventive option against COVID-19 in fragile patients, in addition to other non-pharmaceutical measures such as wearing masks, hand-washing, or social distancing.5 Several randomized trials have established the safety and efficacy of several vaccines, using novel messenger RNA (mRNA) or vector-based vaccines.6 For patients undergoing hematopoietic cell transplantation or cellular therapy, current consensus recommends initiation of vaccination approximately between 3 and 6 months after treatment for allogeneic and 2 months for autologous transplants.7–9 However, the actual effect in patients with hematological malignancies and transplantation/cellular therapy is unclear.

Haematologica | 107 August 2022

1840

ARTICLE - SARS-CoV-2 vaccine response

N. Gagelmann et al.

As the COVID-19 pandemic, to date, does not seem to be overcome and the disease still remains a major risk factor for morbidity and mortality in high-risk patients, evidence syntheses are needed facilitating risk group identification and decision-making. Here, we aimed to perform a systematic review and meta-analysis to assess the antibody response, efficacy, and safety after vaccination against SARS-CoV-2 in patients with hematological malignancies.

Methods Search strategy and selection criteria We followed the updated Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guideline and the Meta-analysis of Observational Studies in Epidemiology (MOOSE) checklist. All studies published since 1 July 2020 on adults with hematological malignancies (myeloid, lymphoid, or plasma cell dyscrasias) after one or two doses of vaccine were considered for inclusion. Full vaccination was defined as two doses of mRNA vaccination or one dose of vectorbased vaccines. Case reports/series, or cohort studies with an overall population of less than ten were excluded. The last updated literature search of MEDLINE, the Cochrane Library, and LitCovid was done on 16 September 2021. Additionally, the reference lists of relevant reviews published in 2021 were reviewed. Two authors (NG and NK) independently conducted the search strategy. Differences in opinion were discussed and resolved by consensus with a third author (FP). The full search strategy is available in the Online Supplementary Table S1. The following key characteristics were extracted: authors, number of patients, type of vaccine, disease, number of patients receiving hematopoietic cell transplantation (allogeneic or autologous) and cellular therapy (chimeric antigen receptor [CAR] T-cell therapy), other therapies, remission status, COVID-19 prior to vaccination, control group, antibody response, safety, infection rate after vaccination, age, antibody assay, and follow-up. Hematological malignancy subtypes were divided as: multiple myeloma (excluding monoclonal gammopathy of unknown significance), chronic lymphocytic leukemia, myeloproliferative neoplasms (including chronic myeloid leukemia, polycythaemia vera, essential thrombocythaemia, and myelofibrosis), lymphoproliferative disorder (aggressive or indolent non-Hodgkin lymphoma, and Hodgkin lymphoma). For remission status, we categorized patients as in remission at time of COVID-19 vaccination or at stable/progressive disease. Outcomes The primary end point was antibody response (seroconversion rate) after full vaccination (1 or 2 doses, depending

on the vaccine), as assessed by anti-SARS-CoV-2 spike protein IgG antibody testing. Thresholds for positivity were in accordance with the respective assays used in the studies. The secondary endpoints were efficacy, response after first dose, and safety. Regarding efficacy, COVID-19 diagnosis was based solely on real-time polymerase chain reaction (RT-PCR), as reported by the studies. Data analysis The methodological quality of each study was assessed using a tool designed specifically to evaluate non-comparative studies. Domains for potential bias are the following: selection of participants, ascertainment, causality, and reporting.10 We used the Q test to assess between-study heterogeneity, and calculated the I² statistic, which expresses the percentage of the total observed variability caused by study heterogeneity. I² values were defined as low (≤50%), moderate (50-75%), or high heterogeneity (>75%). Publication bias was assessed by visual inspection of the funnel plots, coupled with the Egger’s test. Event rates and confidence intervals (CI) were pooled for each intervention in a meta-analysis using a random-effects model (DerSimonian and Laird). In order to examine the association of prespecified continuous moderator variables such as age with study effect size, a mixed-effects model was selected for meta-regression. All values showing P<0.05 were considered as statistically significant. All analyses were performed using R version 4.0.3. No informed consent or Institutional Review Boards were needed for this analysis. This study is registered with PROSPERO (clinicaltrials gov. Identifier: CRD42021279051).

Results A total of 714 citations were retrieved, after removal of duplicates. Out of those, 89 citations were assessed eligible for full-text screening. Citations were further selected after exclusion of studies reporting on less than ten patients, studies with no extractable antibody response (seroconversion) rate, survey reporting only on qualitative outcomes, and remaining reviews. Subsequently, 49 studies comprising 11,086 patients with hematological malignancies, solid cancer, or healthy controls were eligible (Figure 1).11–59 Most studies investigated response after second dose of mRNA vaccines (either BNT162b2 or mRNA-1273). Twentyfour studies included healthy controls, and seven studies included patients with solid cancer. Median follow-up of studies on full vaccination (second dose of mRNA or first dose of vector-based vaccine) was 52 days (range, 35-107 days) and the median age of hematological patients was

Haematologica | 107 August 2022

1841

ARTICLE - SARS-CoV-2 vaccine response

N. Gagelmann et al. Figure 1. PRISMA flow diagram of study selection process.

67 years (range of median age, 46-82 years). Main study characteristics are depicted in the Online Supplementary Table S2. The overall risk of bias of included studies was judged to be low, with ten studies showing moderate risk of bias (Online Supplementary Table S3), and no publication bias was identified (P=0.56; Online Supplementary Figure S1). Antibody response in hematological malignancies Thirty-nine studies reported on antibody response after full vaccination (mostly second dose of mRNA vaccines). A significant between-group difference in antibody response was identified between hematological malignancies, solid cancer, and healthy control (P<0.001). Four thousand three hundred and eleven of 6,516 patients with hematological malignancies showed an antibody response, and the overall pooled response was 64% (95% CI: 59-69), with high heterogeneity (I²=93%). Seven hundred seventy-five of 806 patients with solid cancer in seven studies showed an

antibody response, and the pooled response was 96% (95% CI: 0.92-0.97), with no heterogeneity (I²=44%). One thousand five hundred and fifty-one of 1,588 healthy controls in 21 studies showed an antibody response, and the pooled response was 98% (95% CI: 96-99), and moderate heterogeneity was identified (I²=55%). Detailed results with number of patients and events are depicted in the corresponding forest plot in Figure 2. Seventeen studies reporting on 1,550 patients with hematological malignancies included early antibody response rates after first inoculum (Online Supplementary Figure S2). A significant difference was identified between hematological malignancies, solid cancer, and healthy controls (P<0.001). The pooled response was 37% (95% CI: 29-45; I²=88%) for hematological malignancies, 62% (95% CI: 3783; I²=93%) for solid cancer, and 78% (95% CI: 63-89; I²=92%) for healthy controls.

Haematologica | 107 August 2022

1842

ARTICLE - SARS-CoV-2 vaccine response

N. Gagelmann et al.

Figure 2. Forest plot of pooled antibody response rates across included studies.

Sensitivity analysis In order to minimize potential selection bias, prespecified sensitivity analysis was done according to risk of bias of studies evaluating hematological malignancies (Table 1). The overall pooled response was 64% (95% CI: 58-71; I²=94%) and 63% (95% CI: 52-62; I²=82%) for studies at overall low or moderate risk of bias, respectively (P=0.91). Next, response according to type of mRNA vaccine was evaluated in patients with hematological malignancies. Two thousand five hundred and one of 3,864 patients and 757 of 1,058 who received either BNT162b2 or mRNA-1273 showed an antibody response, and the pooled responses were 63% (95% CI: 57-69) and 72% (95% CI: 60-72), with high heterogeneity (I²= 93% and 96%), respectively. No significant between-group difference according to vaccine was identified (P=0.34; Online Supplementary Figures S3 and S4). Diseases and treatments We then stratified the overall population with hematologi-

cal malignancies according to underlying diagnosis and found a significant difference between the groups (P<0.001), accounting for 25% of the overall heterogeneity (Figure 3). Nine hundred and thirty-one of 1,753 patients with chronic lymphocytic leukemia showed an antibody response, and the pooled response was 50% (95% CI: 4357), with high heterogeneity (I²=84%). Next, 1,163 of 1564 patients with multiple myeloma showed an antibody response, and the pooled response was 76% (95% CI: 6783), with high heterogeneity (I²=91%). Two hundred and ninety-five of 365 patients with myeloproliferative neoplasms showed an antibody response, and the pooled response was 83% (95% CI: 69-91), with high heterogeneity (I²=85%). For non-Hodgkin lymphoma, 214 of 386 with aggressive and 316 of 494 with indolent lymphoma showed an antibody response, and the pooled responses were 58% (95% CI: 44-70) and 61% (95% CI: 48-72), with high heterogeneity (I²= 84% and 85%), respectively. For Hodgkin lymphoma, 125 of 133 patients showed antibody response, and the pooled response was 91% (95% CI: 82-96), with

Haematologica | 107 August 2022

1843

ARTICLE - SARS-CoV-2 vaccine response

N. Gagelmann et al.

Table 1. Pooled antibody response rates across subgroups. Subgroups

Pooled response

95% CI

I²

Hematopoietic cell transplant Allogeneic Autologous

697 547

82% 83%

77-87 73-90

64% 83%

0.60

CAR-T therapy

42%

27-60

54%

1,228 1,034

35% 76%

25-47 68-82

93% 83%

<0.001

Anti-CD20 therapy >1 year ≤1 year

388 321

59% 15%

46-72 9-24

87% 59%

<0.001

Anti-CD38

351

55%

40-69

84%

Chemotherapy

443

69%

54-81

83%

Bruton kinase inhibition

636

23%

14-35

85%

Venetoclax

155

26%

20-34

Disease status Remission Stable or progressive disease

835 590

72% 48%

64-79 31-66

80% 93%

0.014

Prior COVID-19 Yes No

107 2,654

87% 66%

75-94 57-74

26% 94%

0.005

Risk of bias Low Moderate

5,904 612

64% 63%

58-70 52-72

94% 82%

0.91

mRNA vaccine BNT162b2 mRNA-1273

4,224 1,058

63% 72%

57-69 52-86

93% 96%

0.43

Treatment Active No

CI: confidence interval; CAR: chimeric antigen receptor; N: number; COVID-19: coronavirus disease 2019.

no heterogeneity (I²=12%). A significant difference was found for patients in remission compared with patients with stable or progressive disease at time of vaccination (P=0.014). Six hundred and five of 835 patients in remission showed an antibody response compared with 331 of 590 of patients with stable or progressive disease (Online Supplementary Figure S5), and the pooled responses were 72% (95% CI: 6479) compared with 48% (95% CI: 31-66), with high heterogeneity (I²=80% and 93%), respectively. Antibody response was furthermore affected by history of COVID-19 prior to vaccination (P=0.005). Ninety-seven of 107 patients with hematological malignancies and prior COVID19 showed an antibody response, and the pooled response was 87% (95% CI: 75-94), with no heterogeneity (I²=26%; Online Supplementary Figure S6). Thirteen studies evaluated patients who underwent hematopoietic cell transplantation comprising a total of 1,324 patients. No between-group difference for allogeneic and autologous transplants was identified (P=0.60; Online Supplementary Figure S7). For allogeneic transplants, 577 of 697 patients achieved antibody response, and the pooled response was 82% (95% CI: 77-87), with moderate

heterogeneity (I²=64%). Most transplantations were received >1 year prior to vaccination and limited data suggested reduced response rates particularly for those receiving allogeneic transplantation <6 or 12 months prior to vaccination.22,41,47,59 For autologous transplants, 466 of 547 patients achieved an antibody response, and the pooled response was 83% (95% CI: 73-90), with high heterogeneity (I²=83%). For patients who received CAR-T therapy, only 35 of 72 patients showed a response, and the pooled response was 42% (95% CI: 27-60), with moderate heterogeneity (I²=54%; Online Supplementary Figure S8). Next, a significant difference in antibody response was found for active treatment at time of vaccination in comparison with no treatment (P<0.001). Five hundred and fifty-two of 1,228 patients under active treatment and 744 of 1,034 under no treatment showed antibody response (Online Supplementary Figure S9), and the pooled responses were 35% (95% CI: 25-47; I²=93%) compared with 76% (95% CI: 68-82; I²=83%). Furthermore, targeted treatments were evaluated. The response was significantly affected by the timing of antiCD20 therapy (P<0.001). Forty-six of 321 patients who

Haematologica | 107 August 2022

1844

ARTICLE - SARS-CoV-2 vaccine response

N. Gagelmann et al.

Figure 3. Forest plot of pooled antibody response rates across hematological malignancies. Note: aggressive lymphoma was defined as diffuse large B-cell lymphoma, Burkitt lymphoma, peripheral T-cell lymphoma, mantle-cell lymphoma, angioimmunoblastic lymphoma, and central nervous system lymphoma. CLL: chronic lymphocytic leukemia; MM: multiple myeloma; MPN: myeloproliferative neoplasms; NHL: non-Hodgkin lymphoma; HL: Hodgkin lymphoma.

received anti-CD20 therapy ≤1 year prior to vaccination showed an antibody response compared with 213 of 388 who received anti-CD20 therapy >1 year prior to vaccination (Online Supplementary Figure S10). The pooled responses were 15% (95% CI: 9-24; I²=59%) compared with 59% (95% CI: 46-72; I²=85%). Next, 11 studies comprising 636 patients evaluated outcome for Bruton kinase inhibitor therapy, and the pooled response was 23% (95% CI: 14-35), with high heterogeneity (I²=85%; Online Supplementary Figure S11). Seven studies evaluated the response for patients who received venetoclax therapy, and the pooled response was 26% (95% CI: 14-37), with no heterogeneity (I²=0%; Online Supplementary Figure S12). Four studies included 50 patients with myelofibrosis and ruxolitinib therapy, and the pooled response was 42% (95% CI: 25-61), with no heterogeneity (I²=36%). Seven studies in multiple myeloma patients evaluated response

for patients who received anti-CD38 therapy (Online Supplementary Figure S13). Of 351 patients analyzed 211 showed an antibody response, and the pooled response was 55% (95% CI: 40-69), with high heterogeneity (I²=84%). Patients who received chemotherapy showed a pooled response of 69% (Online Supplementary Figure S14). Efficacy Fifteen studies comprising 2,719 hematological malignancy patients assessed efficacy of vaccination during follow-up. There were 28 reported COVID-19 cases, with ten and nine reported cases in two studies,14,22 while the remaining nine cases were reported in six studies.15,37,51,53,57,58 The pooled proportion of evaluable patients without COVID-19 during follow-up was 99% (95% CI: 98-99), and no heterogeneity was observed (I²=23%; Online Supple-

Haematologica | 107 August 2022

1845

ARTICLE - SARS-CoV-2 vaccine response

N. Gagelmann et al.

mentary Figure S15). Five COVID-19 deaths were observed among fully vaccinated patients with hematological malignancies. Safety One-third of patients reported local adverse events such as injection side pain and sore arm. Most frequent systemic adverse events were weakness/fatigue (6-30%) and generalized muscle pain (4-30%). The Online Supplementary Table S4 summarizes the reported safety profiles after first and second dose of vaccination.

Discussion This meta-analysis of 49 studies and >11,000 patients with hematological malignancies, solid cancer, or healthy controls regarding antibody responses after vaccination against SARS-CoV-2 found markedly reduced response rates for patients with hematological malignancies. The humoral response to vaccination was further affected by type of disease, remission status, history of COVID-19, and treatment. In terms of disease subgroups, patients with chronic lymphocytic leukemia showed lowest pooled response rates of 50%, followed by patients with aggressive (58%) or indolent non-Hodgkin lymphoma (61%), whereas patients with multiple myeloma and myeloproliferative neoplasms showed responses of 76% and 83%, respectively. Only patients with Hodgkin lymphoma appeared to exhibit comparable responses to healthy controls, showing pooled response of 91%. The underlying causes for lower humoral response to vaccination may be multifactorial, with attributions to disease-related immune dysregulation, treatment-related immunosuppression, as well as to patient-specific factors.60 One identified treatment-related factor across disease was active treatment, suggesting markedly reduced humoral response rates (pooled, 35%) in comparison with patients with no treatment at time of vaccination. Particularly for patients with chronic lymphocytic leukemia and lymphoma, another important treatment-related factor identified in this analysis may be anti-CD20 antibody (rituximab or obinutuzumab) within 1 year prior to vaccination. This subgroup showed significantly reduced anti–SARS-CoV-2 antibodies in comparison with patients who received anti-CD20 therapy >1 year prior to vaccination who, however, still exhibited overall reduced response. These results are in line with previous findings of reduced response to other vaccines after exposure to B cell–depleting agents.61,62 Furthermore, patients actively treated with Bruton kinase inhibitors exhibited very low response rates. This result is in line with previous studies which found an association between blockaded B-cell re-

ceptor signaling and impaired responses to vaccines against influenza and hepatitis B.63,64 Our data synthesis may encourage active discourse and implementation of personalized immunization strategies for patients with respect to their individual treatment. Regarding weakened responses in multiple myeloma, although we could not further dissect patients into smoldering myeloma, one included study suggested suboptimal response irrespective of smoldering or active multiple myeloma, indicating that the disease itself may have a crucial immunosuppressive role, suppressing normal B-cell expansion and immunoglobulin production and characterized by dysfunctional antigen presentation.65 Myeloma patients were previously found to present suboptimal seroconversion rates after vaccination against other viruses and, usually, booster doses are needed.66 In terms of treatment-related factors, we found reduced pooled response rates of 55% for myeloma patients receiving anti-CD38 therapy. Mechanistically, depletion of antibody-producing B cells using anti-CD38 monoclonal antibodies directly diminishes immunogenicity.67 Previous findings suggested no interfering effect of this therapy in the setting of other usual vaccination programs.68 However, at least three studies reported ≤50% response rates for patients undergoing anti-CD38 treatment in the present analysis. Other treatment modalities such as immunomodulatory drugs appeared to be associated with higher response rates compared with antibody treatments, showing 81% (95% CI: 71-88; I2=51%).14,19,27,43,46 Whether these results are further affected by other concurrent treatments such as proteasome inhibition cannot be further dissected in this pooled analysis. Another relevant subgroup consists of patients who underwent hematopoietic cell transplantation or CAR-T therapy. Here, we found slightly higher response rates of 83% for autologous and 82% for allogeneic hematopoietic cell transplantation recipients. Our analysis was unable to differentiate results according to timing of transplantation and thus on the potential effect of ongoing immunosuppression, due to the limited number of studies, while first limited evidence showed higher responses for patients receiving vaccination at least 1 year after transplantation.22,41,47 Another study in myeloma suggested low serological responses for patients who received autologous transplantation within 6 months prior to vaccination, and responses improved significantly afterwards.22 With respect to allografts and although not systematically assessable in our data synthesis, exacerbation of graftversus-host disease was noted in approximately 3-5% of the patients without serious complications.33,38 Two studies documented development of transient grade 3 or 4 cytopenia.69 For CAR-T patients, we found markedly reduced pooled response of 45%. Whether this is associated with underlying disease, immunosuppression, disease

Haematologica | 107 August 2022

1846

ARTICLE - SARS-CoV-2 vaccine response

N. Gagelmann et al.

characteristic, previous exposition to B-cell depleting therapy, or cytokine release syndrome needs to be evaluated in larger studies. In terms of vaccines, mRNA-1273 appeared to induce higher seropositive rates in comparison with BNT162b2, which is in line with previous findings in the setting of solid organ transplantation,70 whereas nearly identical response and efficacy has been shown for healthy participants. Median antibody titers appeared to be higher for mRNA-1273, and whether this can be attributed to different amounts of spike mRNA, or whether this is affected by different compositions altering penetration into host cells remains to be determined. We acknowledge several shortcomings of our analysis. Considerable heterogeneity was observed in hematological malignancies, while no substantial heterogeneity was observed in solid cancer and healthy controls, suggesting inherent effect of diseases and treatments. In that regard, results in some patient subgroups may be interpreted with caution do due to small sample size, for instance patients who received CAR-T therapy or patients with different myeloproliferative neoplasms (including myelofibrosis or polycythemia vera). Patients with acute leukemia or myelodysplastic syndromes were only reported in three studies comprising a total of 170 patients. The pooled response after full vaccination was 91% (95% CI: 86-95; I²=0%).23,29,47 However, detailed reports on how many of them had acute lymphoblastic leukemia or acute myeloid leukemia, or which treatment (including transplantation) was received by either group were not extractable from all studies, challenging the interpretation of these results. In contrast, when interpreting results of the total cohort of hematological malignancies, some cohorts may confound the results by relative overrepresentation (including chronic lymphocytic leukemia and multiple myeloma), which may have been due to the prevalence as well as the need of rapid recruitment necessary for studies in this evolving pandemic. Therefore, we aimed to stratify patients as adequately as possible to dissect current evidence for each subgroup. Furthermore, we analyzed other patient-specific factors that may have influenced results such as age, which did not seem to affect response between the studies. Meta-regression showed that none of the observed between-study heterogeneity may be accounted to patient age, with a trend toward different outcome for patients >75 years (Online Supplementary Figure S16). Another limitation might have been introduced through the reliance on anti-spike protein IgG levels as a surrogate for immunity to COVID-19, with a risk of between-study difference due to different assays. Another potential limi-

tation may result from under-estimation of dynamics over time of anti-spike antibodies because cutoffs and followup needed to be taken from each study. As with any meta-analysis, the present work depended on time point evaluations, dynamics of outcomes can only be addressed using patient level data. The anti-spike IgG antibody used in each study might still not necessarily correlate with neutralizing activity against SARS-CoV-2. Not all studies reported homogenously on neutralizing activity which was therefore not included as the main objective to minimize selection bias. However, an explorative pooled analysis of neutralizing antibody response of five studies comprising 856 patients across diseases showed a pooled response of 52% (95% CI: 44-60) after full vaccination, with high heterogeneity (I²=82%).11,23,46,47,52 More studies on the relevance of measuring neutralization assays are needed. Last, some studies reported on the level of SARS-CoV-2specific T-cell responses. Recent research indicated that CD8+ T-cell responses may be protective in patients with hematological malignancies in the setting of limited antibody responses after vaccination.71 Analysis of T-cell responses was out of the scope of the present work and will need further evaluation. Despite all limitations identified, during the COVID-19 pandemic, gathering, analyzing, and reporting outcome data is particularly important for specific risk populations. This large meta-analysis of patients with hematological malignancies who underwent vaccination against SARSCoV-2 according to current guidance showed significantly reduced response in this cohort. The findings of the presented evidence synthesis may inform decision-making with regards to patient selection and highlights the need for further studies on the best timing of vaccinations as well as the added value of boosters. Disclosures No conflicts of interest to disclose. Contributions NG conceived the study, performed literature search, article selection, analysis, and manuscript writing; FP performed article selection, manuscript writing, and revision; NK performed article selection, analysis, manuscript writing, review, and revision; CW, RM, CN, RA, BM, and FA provided data clarifications, manuscript review, and revision. Data-sharing statement Extracted study characteristics and outcome data will be made available upon email request to the corresponding author.

Haematologica | 107 August 2022

1847

ARTICLE - SARS-CoV-2 vaccine response

N. Gagelmann et al.

References 1. Vijenthira A, Gong IY, Fox TA, et al. Outcomes of patients with hematologic malignancies and COVID-19: a systematic review and meta-analysis of 3377 patients. Blood. 2020;136(25):2881-2892. 2. Ljungman P, La Camara R de, Mikulska M, et al. COVID-19 and stem cell transplantation; results from an EBMT and GETH multicenter prospective survey. Leukemia. 2021;35(10):2885-2894. 3. Passamonti F, Romano A, Salvini M, et al. COVID-19 elicits an impaired antibody response against SARS-CoV-2 in patients with haematological malignancies. Br J Haematol. 2021;195(3):371-377. 4. Ljungman P, Mikulska M, La Camara R de, et al. The challenge of COVID-19 and hematopoietic cell transplantation; EBMT recommendations for management of hematopoietic cell transplant recipients, their donors, and patients undergoing CAR Tcell therapy. Bone Marrow Transplant. 2020;55(11):2071-2076. 5. Keehner J, Horton LE, Binkin NJ, et al. Resurgence of SARS-CoV-2 Infection in a highly vaccinated health system workforce. N Engl J Med. 2021;385(14):1330-1332. 6. Creech CB, Walker SC, Samuels RJ. SARS-CoV-2 Vaccines. JAMA. 2021;325(13):1318-1320. 7. Cordonnier C, Einarsdottir S, Cesaro S, et al. Vaccination of haemopoietic stem cell transplant recipients: guidelines of the 2017 European Conference on Infections in Leukaemia (ECIL 7). Lancet Infect Dis. 2019;19(6):e200-e212. 8. https://www.hematology.org/covid-19/ash-astct-covid-19vaccination-for-hct-and-car-t-cell-recipients, Accessed 8 November 2021. 9. https://www.ebmt.org/covid-19-and-bmt, Accessed 8 November 2021. 10. Murad MH, Sultan S, Haffar S, Bazerbachi F. Methodological quality and synthesis of case series and case reports. BMJ Evid Based Med. 2018;23(2):60-63. 11. Terpos E, Gavriatopoulou M, Ntanasis-Stathopoulos I, et al. The neutralizing antibody response post COVID-19 vaccination in patients with myeloma is highly dependent on the type of antimyeloma treatment. Blood Cancer J. 2021;11(8):138. 12. Terpos E, Trougakos IP, Gavriatopoulou M, et al. Low neutralizing antibody responses against SARS-CoV-2 in older patients with myeloma after the first BNT162b2 vaccine dose. Blood. 2021;137(26):3674-3676. 13. Thakkar A, Gonzalez-Lugo JD, Goradia N, et al. Seroconversion rates following COVID-19 vaccination among patients with cancer. Cancer Cell. 2021;39(8):1081-1090. 14. van Oekelen O, Gleason CR, Agte S, et al. Highly variable SARSCoV-2 spike antibody responses to two doses of COVID-19 RNA vaccination in patients with multiple myeloma. Cancer Cell. 2021;39(8):1028-1030. 15. Stampfer SD, Goldwater M-S, Jew S, et al. Response to mRNA vaccination for COVID-19 among patients with multiple myeloma. Leukemia. 2021;35(12):3534-3541. 16. Roeker LE, Knorr DA, Thompson MC, et al. COVID-19 vaccine efficacy in patients with chronic lymphocytic leukemia. Leukemia. 2021;35(9):2703-2705. 17. Ram R, Hagin D, Kikozashvilli N, et al. Safety and immunogenicity of the BNT162b2 mRNA COVID-19 vaccine in patients after allogeneic HCT or CD19-based CART therapy - a single-center prospective cohort study. Transplant Cell Ther. 2021;27(9):788-794. 18. Pimpinelli F, Marchesi F, Piaggio G, et al. Lower response to BNT162b2 vaccine in patients with myelofibrosis compared to polycythemia vera and essential thrombocythemia. J Hematol Oncol. 2021;14(1):119.

19. Pimpinelli F, Marchesi F, Piaggio G, et al. Fifth-week immunogenicity and safety of anti-SARS-CoV-2 BNT162b2 vaccine in patients with multiple myeloma and myeloproliferative malignancies on active treatment: preliminary data from a single institution. J Hematol Oncol. 2021;14(1):81. 20. Parry H, McIlroy G, Bruton R, et al. Antibody responses after first and second Covid-19 vaccination in patients with chronic lymphocytic leukaemia. Blood Cancer J. 2021;11(7):136. 21. Monin L, Laing AG, Muñoz-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. 22. Maneikis K, Šablauskas K, Ringelevičiūtė U, et al. Immunogenicity of the BNT162b2 COVID-19 mRNA vaccine and early clinical outcomes in patients with haematological malignancies in Lithuania: a national prospective cohort study. Lancet Haematol. 2021;8(8):e583-e592. 23. Malard F, Gaugler B, Gozlan J, et al. Weak immunogenicity of SARSCoV-2 vaccine in patients with hematologic malignancies. Blood Cancer J. 2021;11(8):142. 24. Lim SH, Campbell N, Johnson M, et al. Antibody responses after SARS-CoV-2 vaccination in patients with lymphoma. Lancet Haematol. 2021;8(8):e542-e544. 25. Kastritis E, Terpos E, Sklirou A, et al. Antibody response after initial vaccination for SARS-CoV-2 in patients with amyloidosis. Hemasphere. 2021;5(8):e614. 26. Iacono D, Cerbone L, Palombi L, et al. Serological response to COVID-19 vaccination in patients with cancer older than 80 years. J Geriatr Oncol. 2021;12(8):1253-1255. 27. Herzog Tzarfati K, Gutwein O, Apel A, et al. BNT162b2 COVID-19 vaccine is significantly less effective in patients with hematologic malignancies. Am J Hematol. 2021;96(10):1195-1203. 28. Gurion R, Rozovski U, Itchaki G, et al. Humoral serologic response to the BNT162b2 vaccine is abrogated in lymphoma patients within the first 12 months following treatment with anti-CD2O antibodies. Haematologica. 2022;107(3):715-720. 29. Greenberger LM, Saltzman LA, Senefeld JW, Johnson PW, DeGennaro LJ, Nichols GL. Antibody response to SARS-CoV-2 vaccines in patients with hematologic malignancies. Cancer Cell. 2021;39(8):1031-1033. 30. Ghione P, Gu JJ, Attwood K, et al. Impaired humoral responses to COVID-19 vaccination in patients with lymphoma receiving B-cell directed therapies. Blood. 2021;138(9):811-814. 31. Gavriatopoulou M, Terpos E, Kastritis E, et al. Low neutralizing antibody responses in WM, CLL and NHL patients after the first dose of the BNT162b2 and AZD1222 vaccine. Clin Exp Med. 2021;15. 32. 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. 33. Easdale S, Shea R, Ellis L, et al. Serologic responses following a single dose of SARS-Cov-2 vaccination in allogeneic stem cell transplantation recipients. Transplant Cell Ther. 2021;27(10):880.e1-880.e4. 34. Caocci G, Mulas O, Mantovani D, et al. Ruxolitinib does not impair humoral immune response to COVID-19 vaccination with BNT162b2 mRNA COVID-19 vaccine in patients with myelofibrosis. Ann Hematol. 2022;101(4):929-931. 35. Bird S, Panopoulou A, Shea RL, et al. Response to first vaccination against SARS-CoV-2 in patients with multiple myeloma. Lancet

Haematologica | 107 August 2022

1848

ARTICLE - SARS-CoV-2 vaccine response

N. Gagelmann et al.

Haematol. 2021;8(6):e389-e392. 36. Benjamini O, Rokach L, Itchaki G, et al. Safety and efficacy of BNT162b mRNA Covid19 vaccine in patients with chronic lymphocytic leukemia. Haematologica. 2022;107(3):625-634. 37. Benda M, Mutschlechner B, Ulmer H, et al. Serological SARS-CoV-2 antibody response, potential predictive markers and safety of BNT162b2 mRNA COVID-19 vaccine in haematological and oncological patients. Br J Haematol. 2021;195(4):523-531. 38. Ali H, Ngo D, Aribi A, et al. Safety and tolerability of SARS-CoV2 emergency-use authorized vaccines for allogeneic hematopoietic stem cell transplant recipients. Transplant Cell Ther. 2021;27(11):938.e1-938. 39. Addeo A, Shah PK, Bordry N, et al. Immunogenicity of SARS-CoV-2 messenger RNA vaccines in patients with cancer. Cancer Cell. 2021;39(8):1091-1098. 40. Agha ME, Blake M, Chilleo C, Wells A, Haidar G. Suboptimal response to coronavirus disease 2019 messenger RNA vaccines in patients with hematologic malignancies: a need for vigilance in the postmasking era. Open Forum Infect Dis. 2021;8(7):ofab353. 41. Dhakal B, Abedin SM, Fenske TS, et al. Response to SARS-CoV-2 vaccination in patients after hematopoietic cell transplantation and CAR-T cell therapy. Blood. 2021;138(14):1278-1281. 42. 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. 43. Avivi I, Balaban R, Shragai T, et al. Humoral response rate and predictors of response to BNT162b2 mRNA COVID19 vaccine in patients with multiple myeloma. Br J Haematol. 2021;195(2):186-193. 44. Ghandili S, Schönlein M, Lütgehetmann M, et al. Post-vaccination anti-SARS-CoV-2-antibody response in patients with multiple myeloma correlates with low CD19+ B-lymphocyte count and antiCD38 treatment. Cancers. (Basel) 2021;13(15):3800. 45. Redjoul R, Le Bouter A, Beckerich F, Fourati S, Maury S. Antibody response after second BNT162b2 dose in allogeneic HSCT recipients. Lancet. 2021;398(10297):298-299. 46. Chung DJ, Shah GL, Devlin SM, Lakshmi VR, Doddi S, Pessin MS. Disease and therapy-specific impact on humoral immune responses to COVID-19 vaccination in hematologic malignancies. Blood Cancer Discov. 2021;2(6):568-576. 47. Tamari R, Politikos I, Knorr DA, et al. Predictors of humoral response to SARS-CoV-2 vaccination after hematopoietic cell transplantation and CAR T cell therapy. Blood Cancer Discov. 2021;2(6):577-585. 48. Del Poeta G, Bomben R, Polesel J, et al. COVID-19 vaccination: evaluation of risk for protection failure in chronic lymphocytic leukemia patients. Hematol Oncol. 2021;39(5):712-714. 49. Jurgens EM, Ketas TJ, Zhao Z, et al. Serologic response to mRNA COVID-19 vaccination in lymphoma patients. Am J Hematol. 2021;96(11):E410-E413. 50. Mairhofer M, Kausche L, Kaltenbrunner S, et al. Humoral and cellular immune responses in SARS-CoV-2 mRNA-vaccinated patients with cancer. Cancer Cell. 2021;39(9):1171-1172. 51. Perry C, Luttwak E, Balaban R, et al. Efficacy of the BNT162b2 mRNA COVID-19 vaccine in patients with B-cell non-Hodgkin lymphoma. Blood Adv. 2021;5(16):3053-3061. 52. Terpos E, Gavriatopoulou M, Fotiou D, et al. Poor neutralizing antibody responses in 132 patients with CLL, NHL and HL after vaccination against SARS-CoV-2: a prospective study. Cancers. (Basel) 2021;13(17):4480. 53. Ollila TA, Lu S, Masel R, et al. Antibody response to COVID-19 vaccination in adults with hematologic malignant disease. JAMA Oncol. 2021;7(11):1714-1716.

54. Harrington P, Lavallade H de, Doores KJ, et al. Single dose of BNT162b2 mRNA vaccine against SARS-CoV-2 induces high frequency of neutralising antibody and polyfunctional T-cell responses in patients with myeloproliferative neoplasms. Leukemia. 2021;35(12):3573-3577. 55. Harrington P, Doores KJ, Radia D, et al. Single dose of BNT162b2 mRNA vaccine against severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) induces neutralising antibody and polyfunctional T-cell responses in patients with chronic myeloid leukaemia. Br J Haematol. 2021;194(6):999-1006. 56. Gastinne T, Le Bourgeois A, Coste-Burel M, et al. Safety and antibody response after one and/or two doses of BNT162b2 AntiSARS-CoV-2 mRNA vaccine in patients treated by CAR T cells therapy. Br J Haematol. 2022;196(2):360-362. 57. Peeters M, Verbruggen L, Teuwen L, et al. Reduced humoral immune response after BNT162b2 coronavirus disease 2019 messenger RNA vaccination in cancer patients under antineoplastic treatment. ESMO Open. 2021;6(5):100274. 58. Marchesi F, Pimpinelli F, Sperandio E, et al. The 12-week kinetics of anti-SARS-CoV-2 antibodies in different haematological cancers after vaccination with BNT162b2. Br J Haematol. 2022;196(2):362-367. 59. Le Bourgeois A, Coste-Burel M, Guillaume T, et al. Safety and antibody response after 1 and 2 doses of BNT162b2 mRNA vaccine in recipients of allogeneic hematopoietic stem cell transplant. JAMA Netw Open. 2021;4(9):e2126344. 60. Forconi F, Moss P. Perturbation of the normal immune system in patients with CLL. Blood. 2015;126(5):573-581. 61. Nazi I, Kelton JG, Larché M, et al. The effect of rituximab on vaccine responses in patients with immune thrombocytopenia. Blood. 2013;122(11):1946-1953. 62. Marasco V, Carniti C, Guidetti A, et al. T-cell immune response after mRNA SARS-CoV-2 vaccines is frequently detected also in the absence of seroconversion in patients with lymphoid malignancies. Br J Haematol. 2022;196(3):548-558. 63. Sun C, Gao J, Couzens L, et al. Seasonal influenza vaccination in patients with chronic lymphocytic leukemia treated with ibrutinib. JAMA Oncol. 2016;2(12):1656-1657. 64. Pleyer C, Ali MA, Cohen JI, et al. Effect of Bruton tyrosine kinase inhibitor on efficacy of adjuvanted recombinant hepatitis B and zoster vaccines. Blood. 2021;137(2):185-189. 65. Leone P, Solimando AG, Malerba E, et al. Actors on the scene: immune cells in the myeloma niche. Front Oncol. 2020;10:599098. 66. Ludwig H, Boccadoro M, Moreau P, et al. Recommendations for vaccination in multiple myeloma: a consensus of the European Myeloma Network. Leukemia. 2021;35(1):31-44. 67. Moreno L, Perez C, Zabaleta A, et al. The mechanism of action of the anti-CD38 monoclonal antibody Isatuximab in multiple myeloma. Clin Cancer Res. 2019;25(10):3176-3187. 68. Frerichs KA, Bosman PWC, van Velzen JF, et al. Effect of daratumumab on normal plasma cells, polyclonal immunoglobulin levels, and vaccination responses in extensively pre-treated multiple myeloma patients. Haematologica. 2020;105(6):e302-e306. 69. Lee E-J, Cines DB, Gernsheimer T, et al. Thrombocytopenia following Pfizer and Moderna SARS-CoV-2 vaccination. Am J Hematol. 2021;96(5):534-537. 70. Boyarsky BJ, Werbel WA, Avery RK, et al. Immunogenicity of a single dose of SARS-CoV-2 messenger RNA vaccine in solid organ transplant recipients. JAMA. 2021;325(17):1784-1786. 71. Bange EM, Han NA, Wileyto P, et al. CD8+ T cells contribute to survival in patients with COVID-19 and hematologic cancer. Nat Med. 2021;27(7):1280-1289.

Haematologica | 107 August 2022

1849

ARTICLE - Non-Hodgkin Lymphoma

Mutational landscape of high-grade B-cell lymphoma with MYC-, BCL2 and/or BCL6 rearrangements characterized by whole-exome sequencing Axel Künstner,1,2,3* Hanno M. Witte,4,5* Jörg Riedl,4,6* Veronica Bernard,6 Stephanie Stölting,6 Hartmut Merz,6 Vito Olschewski,4 Wolfgang Peter,7,8 Julius Ketzer,9 Yannik Busch,7 Peter Trojok,7 Nikolas von Bubnoff,3,4 Hauke Busch,1,2,3# Alfred C. Feller6# and Niklas Gebauer3,4# Medical Systems Biology Group, University of Lübeck, Lübeck; 2Institute for Cardiogenetics, University of Lübeck, Lübeck; 3University Cancer Center Schleswig-Holstein, University Hospital of Schleswig- Holstein, Campus Lübeck, Lübeck; 4Department of Hematology and Oncology, University Hospital of Schleswig-Holstein, Campus Lübeck, Lübeck; 5Department of Hematology and Oncology, Federal Armed Forces Hospital Ulm, Ulm; 6Hämatopathologie Lübeck, Reference Center for Lymph Node Pathology and Hematopathology, Lübeck; 7HLA Typing Laboratory of the Stefan-Morsch-Foundation; 8University of Cologne, Faculty of Medicine and University Hospital Cologne, Institute for Transfusion Medicine, Cologne and 9 Department of Pediatrics, University Hospital of Schleswig-Holstein, Campus Lübeck, Lübeck, Germany 1

Correspondence: Niklas Gebauer Niklas.Gebauer@uksh.de Received: July 15, 2021. Accepted: November 9, 2021. Prepublished: November 18, 2021. https://doi.org/10.3324/haematol.2021.279631 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

*AK, HMW and JR contributed equally as co-first authors. HB, ACF and NG contributed equally as co-senior authors.

Abstract High-grade B-cell lymphoma accompanied with double/triple-hit MYC and BCL2 and/or BCL6 rearrangements (HGBLDH/TH) poses a cytogenetically-defined provisional entity among aggressive B-cell lymphomas that is traditionally associated with unfavorable prognosis. In order to better understand the mutational and molecular landscape of HGBLDH/TH we here performed whole-exome sequencing and deep panel next-generation sequencing of 47 clinically annotated cases. Oncogenic drivers, mutational signatures and perturbed pathways were compared with data from follicular lymphoma (FL), diffuse large B-cell lymphoma (DLBCL) and Burkitt lymphoma (BL). We find an accumulation of oncogenic mutations in NOTCH, IL6/JAK/STAT and NFκB signaling pathways and delineate the mutational relationship within the continuum between FL/DLBCL, HGBL-DH/TH and BL. Further, we provide evidence of a molecular divergence between BCL2 and BCL6 rearranged HGBL-DH. Beyond a significant congruency with the C3/EZB DLBCL cluster in BCL2 rearranged cases on an exome-wide level, we observe an enrichment of the SBS6 mutation signature in BCL6 rearranged cases. Differential gene set enrichment and subsequent network propagation analysis according to cytogenetically defined subgroups revealed an impairment of TP53 and MYC pathway signaling in BCL2 rearranged cases, whereas BCL6 rearranged cases lacked this enrichment, but instead showed impairment of E2F targets. Intriguingly, HGBL-TH displayed intermediate mutational features considering all three aspects. This study elucidates a recurrent pattern of mutational events driving FL into MYC-driven BCL2-rearranged HGBL, unveiling the mutational pathogenesis of this provisional entity. Through this refinement of the molecular taxonomy for aggressive, germinal center-derived B-cell lymphomas, this calls into question the current World Health Organization classification system, especially regarding the status of MYC/BCL6rearranged HGBL.

Introduction High-grade B-cell lymphoma (HGBL) with MYC-, BCL2 and/or BCL6 rearrangements poses a novel, yet provisional, cytogenetically-defined entity within the current World Health Organization classification of lymphoid tumors. It is allocated in the pathobiological continuum be-

tween diffuse large B-cell (DLBCL) and Burkitt lymphoma (BL).1 The t(8;14)(q24;q32) IgH/MYC rearrangement constitutes the molecular hallmark of BL. This or further derivative chromosomal rearrangements that juxtapose MYC to a genomic enhancer, occur in approximately 10% of DLBCL and have been shown to correlate with inferior clinical outcome.2 The rearrangement is a driver of oncogenesis

Haematologica | 107 August 2022

1850

ARTICLE - HGBL-DH/TH mutational landscape

A. Künstner et al.

that is accompanied in approximately 50% of cases by additional rearrangements involving BCL2 and/or BCL6 and referred to as double-hit (DH) or triple-hit (TH) lymphomas.3-11 While the clinical outcome in double/triple-hit HGBL (HGBL-DH/TH) patients is generally poor, recent studies have hinted at a significant impact of MYC translocation partners and defined MYC/Ig rearrangements to be the most reliable predictors of adverse outcome.2,7 In a prior study, we discovered an elevated frequency of TP53 impairment in MYC-driven DH/TH, whose presence was subsequently demonstrated for a subset of patients with a single-hit MYC translocation as well, indicating inferior outcome.12,13 By conventional cytogenetics HGBL-DH/TH were shown to recurrently harbor a complex karyotype.4 Data on the genetic basis of this entity, however, remains elusive. Several preliminary studies, predominantly focusing on HGBLDH/TH with DLBCL morphology, have employed a panelbased next-generation sequencing (NGS) approach.14-16 The insights from these studies were all restricted by gene panel design and the associated clinicopathological data, even though their central assertions included a significant enrichment in mutations affecting CREBBP, BCL2 and KMT2D alongside an overall reflection of the phenotypical gray zone between DLBCL and BL. Most recently, Cucco et al. elucidated significant aspects of the molecular signature of HGBL in a panel-based sequencing and gene expression study, employing a 70-gene HaloPlex panel and an array-based gene expression approach. The authors restricted their study to samples with DLBCL morphology that stemmed from a clinical trial and the UK's populationbased Hematological Malignancy Research Network.17 A comprehensive, exome-wide assessment of oncogenic driver mutations in HGBL-DH/TH, including cases with BLlike morphology is, however, still warranted and of vital importance to the refinement of the pathogenetic understanding of this clinically challenging entity ultimately enabling targeted therapeutic approaches. We therefore conducted a whole-exome sequencing (WES) study on a large cohort of HGBL-DH/TH, validated by panel-based NGS and supplemented these data with a comprehensive clinicopathological assessment of the study group. Here, we report on oncogenic drivers, somatic copy number alterations (SCNA) and putative pathway perturbations, thus refining the molecular taxonomy of MYCdriven germinal center-derived aggressive lymphomas.

Methods Case selection and clinicopathological characteristics In a retrospective approach, we reviewed our institutional database to identify HGBL patients whose primary diag-

nostic biopsy specimen had been referred to the Reference Center for Hematopathology, University Hospital Schleswig Holstein, Campus Lübeck and Hämatopathologie Lübeck for centralized histopathological panel evaluation between January 2007 and December 2019. For additional Information on clinicopathological work-up, please see the Online Supplementary Appendix and the Online Supplementary Table S1. This retrospective study was approved by the ethics committee of the University of Lübeck (reference number 18356) and conducted in accordance with the Declaration of Helsinki. Patients had given written informed consent regarding routine diagnostic and academic assessment of their biopsy specimen including molecular studies at the Reference Center for Hematopathology and transfer of their clinical data. Whole-exome and targeted amplicon-based sequencing WES of n=47 HGBL-DH/TH samples was performed by a hybrid capture approach with the Agilent SureSelect Human All Exon V6 library preparation kit (Agilent Technologies) followed by Illumina short read sequencing on a NovaSeq platform (Illumina) to an average depth of 304x (standard deviation ±195x; median 234x; sequencing depth was estimated using mosdepth v0.3.2)18 by Novogene (UK) Co., Ltd (Online Supplementary Table S2). Seeking to validate the initial delineation of the exome sequencing-derived mutational landscape in HGBL we employed our in-house custom AmpliSeq panel (Thermo Fisher Scientific, Waltham, MA, USA) for targeted amplicon sequencing (tNGS), encompassing all coding exons of 43 genes (see Online Supplementary Table S3) in 21 cases. DNA preparation for validation experiments was extracted from the same sample but in an independent approach from deeper tissue sections. Raw paired-end data (fastq format) was trimmed and quality filtered using FASTP (v0.20.0; minimum length 50 bp, maximum unqualified bases 30%, trim tail set to 1)19 and trimmed reads were mapped to GRCh37/hg19 using BWA MEM (v0.7.15).20 Resulting alignment files in SAM format were cleaned, sorted, and converted into BAM format using PICARD TOOLS (v2.18.4). Single nucleotide variants (SNV), as well as short insertions and deletions (InDels) were identified following the best practices for somatic mutations calling provided by GATK.21 Somatic copy number aberrations (SCNA) were identified by CONTROL-FREEC (v11.4).22 For further details on nucleic acid extraction, panel sequencing, single nucleotide and copy number variant calling please see the Online Supplementary Appendix. Mutational deleteriousness and significance, network propagation, gene set enrichment and mutational cluster analysis The MUTSIGCV algorithm was employed on WES data to

Haematologica | 107 August 2022

1851

ARTICLE - HGBL-DH/TH mutational landscape

A. Künstner et al.

delineate significantly mutated genes within the study cohort while deleteriousness was assessed via the CADD v1.3. The acquired genomic data were then processed through the LymphGen algorithm and underwent manual screening for an enrichment in overlapping aberrations with the molecular clusters proposed by Chapuy et al. followed by validation through a logistic regression framework.23,24 Cytogenetically defined subgroups (HGBL with MYC and BCL2 aberrations, HGBL with MYC and BCL6 aberrations, and HGBL-TH) underwent differential downstream analysis by a network propagation approach simulating a protein-protein interaction network. Subsequently gene set variation analysis was performed against HALLMARK gene sets. For details including statistical approaches correlating molecular and clinicopathological findings please see the Online Supplementary Appendix.

Results Clinicopathological characteristics of the study group We collected 47 cases of HGBL-DH/TH at diagnosis with sufficient formalin-fixed, paraffin-embedded (FFPE) tissue samples for molecular studies (median age 71 years; range, 35–89 years) all of which were included in the final analysis, following successful library preparation for WES. There was insufficient clinical follow-up in nine of 47 (19%) cases. An underlying HIV infection was clinically excluded in all cases. The majority of patients in our study were male (25/47; 53%) and presented with advanced stage disease (24/38 stage III/IV; 63%) and an adverse prognostic constellation (24/38 [63%], revised International Prognostic Index [R-IPI] >2). Most patients received an intensive cyclophosphamide, doxorubicin hydrochloride, vincristine sulfate, and prednisone (CHOP)-like therapeutic frontline approach (25/38; 66%). The overall response rate after first line (immuno-) chemotherapy was 76% resembling a general therapeutic response in 29 of 38 cases. Table 1 summarizes the baseline characteristics of all HGBL-DH/TH cases included in the current study. Histologically, the predominant morphology was that of DLBCL (NOS) (32/47), however, Burkitt-like morphology and immunophenotype was present in 15 of 47 patients. Cytogenetically 21 of 47 cases presented with MYC/BCL2, 17 of 47 presented with MYC/BCL6 DH constellation and nine TH lymphomas were included. MYC translocation partner revealed MYC-Ig rearrangement in eight of 17 cases. The treatment outcome in our cohort was unfavorable yet in keeping with previous data reported by Rosenwald and colleagues.2 For confirmatory purposes, we included four cases, which were assessed for TP53 mutation status in a previous study and were able to validate both cytogenetic

as well as molecular observations.12 The mutational landscape of double-/triple-hit highgrade B-cell lymphoma identified by whole-exome sequencing In order to characterize the mutational landscape in an extensive cohort of HGBL-DH/TH cases, we successfully performed WES in 47 patient-derived tumor biopsies and matched constitutional DNA in seven cases. We further applied the analytical framework outlined above to analyze WES data in the absence of paired germline DNA in the majority of cases. Following the primary identification of SNV and Indels in individual samples and subsequent filtering to correct for FFPE-derived artefacts and spurious mutations, we applied the MutSig2CV algorithm and thereby identified 22 significant candidate driver genes (P<0.001; 13 genes with q<0.1; Online Supplementary Table S4).25 All HGBL-DH/TH cases carried mutations in genes of oncogenic relevance according to our bioinformatic annotations. In total, we described 10,092 presumably harmful somatic mutations (cut-off see materials and methods) involving 5,521 genes after variant filtering. Of these, SNV and InDels represented 74.1% of the mutations (7,479 SNV). Among them, missense mutations were the most frequent alterations (85.2%), followed by nonsense (5.7%) and InDels (5.6%), while splice site mutations posed 3.3% of somatic mutations (Figure 1A). Displaying an overall intermediate tumor mutational burden (median 3.974; range, 1.065–18.234 mutations/Mbase; Figure 1B), HGBLDH/TH revealed no evidence of MSI-related hypermutations, which is in keeping with observations in DLBCL (0.3%), even though it differs from other aggressive lymphomas (e.g., primary mediastinal B-cell lymphoma).26 Upon comparative analysis of WES and targeted resequencing data, we were able to demonstrate a concordance rate of 92.0% (46/50 in 18 matched samples) of mutational calls, prompting high confidence in mutational calls derived from WES, even in non-germline-paired cases. A comprehensive description of all variants described by WES as well as panel based NGS is provided in the Online Supplementary Tables S5 and S6. Nevertheless, we observed a significant enrichment of non-germline matched samples in non-synonymous SNV, which prompted us to include significantly mutated genes according to the MUTSIGCV analysis, only. As an exception to this rule, we also included MYC mutations below the statistical significance level due to their previously established clinical and functional relevance. Recurrent copy number alterations in double-/triple-hit high-grade B-cell lymphoma We investigated our HGBL-DH/TH cohort for SCNA employing the CONTROL-FREEC 22 algorithm in tumor-normal

Haematologica | 107 August 2022

1852

ARTICLE - HGBL-DH/TH mutational landscape

A. Künstner et al.

Table 1. Clinical characteristics of the study group. HGBL (N=47)

DHL-BCL2 (N=21)

DHL-BCL6 (N=17)

THL (N=9)

Insufficient FU, N (%)

9 (19.1%)

4 (19.0%)

2 (11.8%)

3 (33.3%)

Age in years, median (range)

71 (35–89)

73 (35–88)

72 (35–89)

66 (42–76)

Sex, N (%) Female Male

22 (46.8%) 25 (53.2%)

8 (38.1%) 13 (61.9%)

11 (64.7%) 6 (35.3%)

3 (33.3%) 6 (66.7%)

R-IPI, N (%) 0 1-2 >2

1 (2.6%) 13 (34.2%) 24 (63.2%)

1 (5.9%) 5 (29.4%) 11 (64.7%)

6 (40.0%) 9 (60.0%)

2 (33.3%) 4 (66.7%)

Stage (Ann Arbor), N (%) I II III IV

5 (13.2%) 9 (23.7%) 5 (13.2%) 19 (50.0%)

2 (11.8%) 4 (23.5%) 1 (5.9%) 10 (58.8%)

3 (20.0%) 3 (20.0%) 2 (13.3%) 7 (46.7%)

2 (33.3%) 2 (33.3%) 2 (33.3%)

B-symptoms, N (%) Yes No

20 (52.6%) 18 (47.4%)

10 (58.8%) 7 (41.2%)

7 (46.7%) 8 (53.3%)

3 (50.0%) 3 (50.0%)

Extranodal sites, N (%) 0 1-2 >2

9 (23.7%) 28 (73.7%) 1 (2.6%)

3 (17.6%) 14 (82.4%) -

4 (26.7%) 10 (66.7%) 1 (6.7%)

2 (33.3%) 4 (66.7%) -

ECOG PS, N (%) 0-1 ≥2

20 (52.6%) 18 (47.4%)

8 (47.1%) 9 (52.9%)

8 (53.3%) 7 (46.7%)

4 (66.7%) 2 (33.3%)

LDH, N (%) Normal Elevated

7 (18.4%) 31 (81.6%)

3 (17.6%) 14 (82.4%)

3 (20.0%) 12 (80.0%)

1 (16.7%) 5 (83.3%)

CNS involvement at diagnosis, N (%) Yes No

2 (5.3%) 36 (94.7%)

17 (100.0%)

2 (13.3%) 13 (86.7%)

6 (100.0%)

Morphology, N (%) DLBCL-like Burkitt-like

32 (68.1%) 15 (31.9%)

17 (80.9%) 4 (19.1%)

12 (70.6%) 5 (29.4%)

3 (33.3%) 6 (66.7%)

Frontline therapy regimen, N (%) CHOP-like R-based Intensified* Less intensive** Refusal

25 (65.8%) 31 (81.6%) 13 (34.2%) 9 (23.7%) 1 (2.6%)

11 (64.7%) 14 (82.4%) 7 (41.1%) 4 (23.5%) -

10 (66.7%) 12 (80.0%) 3 (20.0%) 4 (26.7%) 1 (6.7%)

4 (66.7%) 5 (83.3%) 3 (50.0%) 1 (16.7%) -

Characteristics

CHOP: cyclophosphamide/hydroxadaunorubicin/vincristine/predinislone; CNS: central nervous system; DHL: double-hit lymphoma; DLBCL: diffuse large B-Cell Lymphoma; ECOG: Eastern cooperative oncology group; FU: follow-up; HGBL: high grade B-cell lymphoma; LDH: lactate dehydrogenase; PS: performance status; R: rituximab; R-IPI: revised International Prognostic Index; THL: triple-hit lymphoma; yrs.,: years. *Intensified regimens: B-ALL, GMALL, CHOEP (additional etoposide), EPOCH.**Less intensive regimens: Bendamustin, mini-CHOP (50% dose reduction), rituximab mono.

and tumor-only mode, respectively, followed by GISTIC2.0 27 analysis. The analysis excluded chromosomes X and Y as well as common benign copy number variants defined by the Broad Institute’s panel of normals. Upon cross-referencing our findings with genomic loci of known oncogenes, tumor-suppressors and elements of significant signaling pathways we identified recurrent copy number gains in oncogenes such as MEF2B and CSF1R, which have previously been implicated in the pathogenesis of malig-

nant lymphomas.28,29 Further, copy number losses in tumor suppressors like NPM1 were recurrently identified (Figure 2A and B). No significant differences were detected for genes affected by copy number alterations between the three cytogenetically defined subgroups (Fisher exact test P>0.05 after Bonferroni correction for multiple testing). Common CNA, as defined by the above referenced panel of normal were encountered at the expected frequencies.30,31

Haematologica | 107 August 2022

1853

ARTICLE - HGBL-DH/TH mutational landscape

A. Künstner et al.

Significantly mutated candidate driver genes and mutational signatures Putative candidate driver genes comprised several genes previously implicated in HGBL-DH/TH pathogenesis, such as KMT2D, CREBBP and TP53 alongside several further mutated genes such as CDKN2A, LNP1 or SI.14 Established mutational drivers known from other B-cell lymphoproliferative disorders (e.g., FL, DLBCL, BL) were recurrently encountered (e.g., CCND3, ARID1A) (Figures 2C and 3A).32,33 Following SNV and InDel evaluation with MUTSIGCV, a network propagation approach (Figure 3B and C) was employed on significantly mutated genes to delineate the functional implications of significant genetic events on neighboring genes. Additionally, we investigated MUTSIGCV genes using our HGBL-DH/TH (D/THL) cohort, the cytogenetical subgroups (BCL2, BCL6, THL), as well as cohorts of ABC-type DLBCL (n= 67),33 germinal center B-cell like DLBCL (GCB-type DLBCL) (n=45)22 BL (n=108)38 and FL (n=199)38 (all retrieved via cBioPortal) and overlapping genes between the five lymphoma subtypes and the three cytogenetical subtypes of HGBL-DH/TH (Figure 3D and E). In our limited cohort, distinctions regarding subtype-specific mutational signatures were found to be marginal among BCL2/BCL6 status or Burkitt-like versus non-Burkitt-like morphology. However, we found CCND3 mutations, previously reported as driver mutations in BL pathogenesis, to be significantly enriched in HGBL-DH/TH patients with Burkitt-like morphology (9/15 vs. 3/28). This

observation hints at partially similar molecular paths of pathogenesis between BL and HGBL with Burkitt-like morphology. Further, an enrichment of mutations affecting CREBBP in HGBL-DH/TH patients with BCL2 rearrangement was observed, which is well in keeping with its proposed fundamental role in FL pathogenesis. Distribution of mutations within selected, significantly mutated genes is depicted in the Online Supplementary Figure S1. Additional profiling of mutational signatures driving HGBLDH/TH revealed a predominance of the SBS5 (implicated in aging, potential FFPE artifacts and tobacco exposure) signature across all subtypes alongside the emphasized occurrence of the SBS6 signature (implicated in defective DNA mismatch repair [MMR]) in patients with BCL6 rearrangements (Online Supplementary Figure S2; Online Supplementary Table S7). Comparative analysis of mutational landscape in double-/triple-hit high-grade B-cell lymphoma, related entities and molecular clusters in diffuse large B-cell lymphoma Next, we sought to refine the genomic taxonomy of aggressive GC-derived BCL and to investigate the mutational commonalities and differences between HGBL-DT/TH and other related pathological entities. We subsequently selected cBioPortal cohorts from several entities for their similar or divergent features of B-cell differentiation (FL, GCB-type DLBCL and BL vs. ABC-type DLBCL). A com-

Figure 1. Variant classification and mutations per sample. Panel (A) shows the number of variants stratified by variant classification while panel (B) delineates the number of mutations per sample with a median of 153 mutations per sample. Ins: insertion. Haematologica | 107 August 2022

1854

ARTICLE - HGBL-DH/TH mutational landscape

A. Künstner et al.

Figure 2. Genomic and mutational landscape in double-/triple-hit high-grade B-cell lymphoma. The location of somatic copy number alterations (SCNA) along the genome is shown in (A) (red bars denote gains; blue bars denote losses; gene names refer to affected oncogenes according to OncoKB). (B) Display of oncogenes from (A) and SCNA status (red refers to gain, blue refers to loss). Additionally, BCL2 and BCL6 status are shown for each case. Co-oncoplot for genes identified as significant driver genes by MUTSIGCV (P<0.001; n=22) in our cohort stratified by cytogenetical subtypes is shown in (C); different types of mutations are colour coded and additional covariates are shown below the plot for each sample. DLBCL: diffuse large B-cell lymphoma; Ins: insertion; Fol: folicular.

Haematologica | 107 August 2022

1855

ARTICLE - HGBL-DH/TH mutational landscape

A. Künstner et al.

parative analysis of candidate mutational drivers in HGBLDH/TH (as described earlier) and cohorts of ABC-type DLCBL (n=67), GCB-type DLBCL (n=45)23, BL (n=108)34 and FL (n=199)34 was conducted, screening for shared as well as mutually exclusive putative driver mutations. Interestingly, we identified one overlapping candidate driver common to all entities (KMT2D). Additionally, CREBBP was found in all entities except ABC-type DLCBL. Mutations affecting EZH2, IRF8 and TNFRSF14 were, however, specifically occured in HGBL-DT/TH and FL/GCB-type DLBCL, while CCND3 mutations appeared to be a pathogenetic feature shared between BL and HGBL-DT/TH. Additionally, TP53 mutations posed a predominant feature of aggressive lymphomas present in all HGBL-DH/TH, GCB-type DLBCL and BL types and therefore most likely acquired during high-grade transformation (Figure 3D and E). In basic accordance with previous studies, our data suggest a common origin especially for BCL2-rearranged HGBLDH/TH and FL/GCB-type DLBCL.14,17 Upon comparative investigation of our current data and mutational clusters, previously described in DLBCL, we observed a striking predominance of C3/EZB cluster cases in the BCL2-rearranged subgroup according to the integrative molecular classification proposed by Chapuy et al. and Wright et al., respectively. This is in keeping with a significant enrichment of these cases with DLBCL morphology in terms of MYC rearrangement status (Figure 4; Online Supplementary Figure S3).23,24 Complementary to our analysis, employing the LymphGen algorithm (cf. Online Supplementary Table S8), a logistic regression indicated a significantly different number of mutated genes in C3 between the HGBL subtypes. HGBL harboring only BCL2 were shown to exhibit the highest number of mutated C3 genes, while HGBL with BCL6 alterations had the lowest number of mutated C3 genes (BCL2/6 cases: P=6.059*10-5, adjusted R2=0.3108). In contrast to TH cases, HGBL with MYC and an isolated additional BCL6 rearrangement showed a significant decrease in the number of mutated C3 genes (P=2.00*10-5, estimate: -1.7589) (Online Supplementary Figure S4; Online Supplementary Table S9). Within the subgroup of BCL6 rearranged cases, the BN2 cluster was more prominent than the EZB cluster. In keeping with their strong affinity towards the C3/EZB cluster, BCL2-rearranged cases exhibited an enrichment for mutations in CREBBP and KMT2D, while BCL6-rearranged cases were, in contrast, enriched for mutations in ARID1A. The vast majority of TH cases was also classified within the EZB cluster. Mutational impairment of NOTCH, RTK-RAS and TP53 signaling in double-/triple-hit high-grade B-cell lymphoma Cumulatively, we detected genetic lesions, putatively impairing NOTCH signaling in 74% of HGBL-DH/TH patients.

Expanding on previously reported CREBBP, EP300 and DTX1 mutations in HGBL we further identified recurrent mutations affecting NCOR1 and others (Online Supplementary Figure S5).14,17 NOTCH signaling was thereby the predominant target of somatic mutation in HGBL-DH/TH, albeit with a quite heterogeneous mutational pattern affecting 35 of 47 patients with lesions in 28 of 71 genes (Online Supplementary Figure S6A). Most of these genomic aberrations had been previously reported to be gain-offunction mutations putatively resulting in constitutive NOTCH pathway activation in various types of predominantly GCB-type DLBCL. Several of these mutational hits including NCOR1 and DTX1 have been shown to herald adverse clinical outcome.35,36 This remained the case when undertaking a differential downstream analysis within the cytogenetically defined subgroups (HGBL with MYC and BCL2 aberrations, HGBL with MYC and BCL6 aberrations and HGBL-TH), which was prompted by their significantly divergent distribution onto molecular clusters. Through this analysis a mutational signature became apparent that is additionally dominated by impairment of TP53 and MYC signaling in BCL2 rearranged cases. BCL6 rearranged cases lacked this enrichment, while HGBL-TH cases revealed intermediate mutational features. As another recurrent feature across all subgroups we observed alterations, putatively affecting IL6/JAK/STAT signaling in 74% of patients (Online Supplementary Figure S5). Mutations in PIM1 and SOCS1 were most frequently encountered in our case series and have been previously implicated in HGBL-DH/TH pathogenesis. These genes failed, however, to reach the predefined level of statistical significance within the scope of our MUTSIGCV analysis and required further investigation in a more comprehensive dataset.17 These candidate driver genes are supplemented by mutations in LTB and STAT3 (both previously identified in HIV-associated plasmablastic lymphoma) among others.37 Beyond this, we observed a relatively dispersed mutational pattern with putative driver events affecting 28 genes within the NOTCH-pathway (Online Supplementary Figure S6A). In accordance with previous studies, we found mutations directly impacting NF-kB signaling in 62% of cases (Online Supplementary Figure S5).38 While this was, among the predominant pathways, identified through our network propagation approach, mutations affecting the pathway were narrowly detectable with only BCL2 harboring mutations in more than three patients (34%) followed by recurrent SNV and Indels in PARP1 (6%) and BIRC3 (4%) and CARD11 (4%). The network propagation approach further underscored the mutational impairment of the aforementioned pathways alongside WNT and PI3K signaling. These observations are in accordance with preliminary impressions derived from targeted sequencing studies, employing

Haematologica | 107 August 2022

1856

ARTICLE - HGBL-DH/TH mutational landscape

A. Künstner et al.

Continued on following page. Haematologica | 107 August 2022

1857

ARTICLE - HGBL-DH/TH mutational landscape

A. Künstner et al.

Figure 3. Mutational analysis for significantly mutated genes, network propagation and mutational overlap with related entities. (A) Significance levels for significantly mutated genes in our double-/triple-hit high-grade B-cell lymphoma (HGBL-DH/TH) cohort, regardless of subgroup (MUTSIGCV P<0.001; gene names in orange indicate q<0.1); (B) and (C) show pathway enrichment analysis results for network propagation analysis (see the Online Supplementary Appendix for details) of significant MUTSIGCV genes (MUTSIGCV P<0.001) for MYC/BCL2 subgroup (MYC/BCL2 genes included: POU2AF1, HVCN1, B2M, TP53, AP3S1, LNP1, CREBBP, GULP1, TNFRSF14) and MYC/BCL6 subgroup (MYC/BCL6 genes included: CDKN2A, CD78B, LNP1, KRTAP13-1, UBE2A, CCND3) against HALLMARK gene sets and NFκB pathway. UpSet plot (D) showing the overlap of MUTSIGCV genes using our HGBL-DH/TH (D/THL) cohort, the cytogenetical subgroups (BCL2, BCL6, THL), as well as cohorts of ABC-type diffuse large B-cell lymphoma (DLBCL) (n=67)34, germinal-center B-cell lymphoma (GCB-type DLBCL) (n=45)23, Burkitt lymphoma (BL) (n=108)34 and follicular lymphoma (FL) (n=199)34 (all retrieved via cBioPortal); set size refers to the number of genes per cohort and intersection size shows the number of overlapping genes per comparison. Comparisons are denoted by black points and black connecting lines; (E) shows the overlapping genes between the 5 lymphoma subtypes and the 3 cytogenetical subtypes of HGBL-DH/TH; grey denotes mutated genes.

panel-based approaches.14,17 In addition to the divergent results from our mutational pathway analysis, we identified an enrichment of E2F targets impacted by significantly mutated genes in both DH and TH cases affected by BCL6 rearrangements. Of further interest, we report on highly recurrent mutations in known activation-induced cytidine deaminase (AID-) and somatic hypermutation (SHM) targets such as PIM1, SOCS1 and others. Survival analysis Following integrated analysis of molecular and clinical data we investigated genomic alterations present in >15% of patients for their impact on overall survival (OS) and progression-free survival (PFS). Hereby we identified ARID1A mutations to be predictive of worse clinical outcome in our cohort (OS: P=0.0049; PFS: P=0.025). Subsequent Bonferroni correction for multiple testing was performed. Thus, we identified a significant impact of mutations affecting ARID1A which was maintained regarding OS when correction for multiple testing was applied while its primarily significant effect on PFS was reduced to a trend of borderline statistical significance (Figure 5). A Cox proportional hazard model revealed this effect to be independent of the established clinical International Prognostic Index (IPI) prognosticators (age, lactate dehydrogenase [LDH], extra nodal manifestations, stage, and performance status; OS: P<0.001; hazard ratio [HR]: 13.989; 95% confidence interval [CI]: 3.362–58.205; PFS: P=0.001; HR: 6.648; 95% CI: 2.098–21.061). Within the cytogenetically defined subgroups, we identified no alterations with independent impact on clinical outcome. However, a trend of borderline statistical significance towards inferior outcome in MYC/BCL2 rearranged cases harboring FOXO1 mutations was observed (Online Supplementary Figure S7).

Discussion Here we report on WES data from an extensive cohort of HGBL-DH/TH tumors, which is to the best of our knowledge the hitherto largest cohort and most extensive molecular data set for this entity. Previous reports on

HGBL-DH/TH were limited by low sample numbers and/or targeted sequencing approaches. Contrary to this, WES here allowed to systematically define recurrent mutations, predominant mutational signatures and SCNV in their respective clinicopathological context from which we report three central observations. Firstly, being the first exome-wide mutational investigation for this rare subtype of lymphoma, we identify a significant overlap of mutational drivers between HGBL-DH/TH and FL as well as GCB-type DLBCL (e.g., TNFRSF14, EZH2 and IRF4) as its high-grade counterpart. Aggressive transformation was associated with the acquisition of mutations in TP53. Moreover, shared features, including CCND3 and CDKN2A mutations underscore a close molecular relation between HGBL-DH/TH and BL.34,39 This is additionally reflected in the enrichment of HGBLDH/TH patients with Burkitt-like morphology for CCND3 mutations. Further, we identify a number of significant mutational drivers not captured by previous, panel-based sequencing studies. Most frequently among these, we find SI mutations that have been previously implicated in CLL progression aa well as mutations in POU2AF1, which has been recently found to be an augmented target of mutations during aggressive transformation of FL to DLBCL.40,41 Although MYC did not meet the predefined MUTSIGCV significance level in our study, we still observed mutations in 19% of cohort samples (Online Supplementary Figure S8), which is in agreement with previous panel-based studies.17 Secondly, upon screening the mutational landscape in HGBL-DH/TH in comparison to the molecular clusters of DLBCL, proposed by Chapuy et al. and the LymphGen algorithm proposed by Wright et al., we unveil a striking overlap of BCL2-rearranged cases with the C3/EZB cluster, which was previously shown to be enriched for MYC rearrangements and oncogenic drivers implicated in FL pathogenesis.23,24 We argue that a predominant subset of HGBL-DH/TH most likely corresponds to these transformed FL. This offers a potential explanation for the inferior clinical outcome of C3 DLBCL patients, despite their GCB-phenotype, through an enrichment for MYC-rearranged HGBL-DH/TH cases. Of note, we find the pre-

Haematologica | 107 August 2022

1858

ARTICLE - HGBL-DH/TH mutational landscape

A. Künstner et al.

dominant impairment of TP53 and to a lesser extent MYC signaling in BCL2 rearranged cases to be in keeping with a previous study on an independent set of HGBL-DH/TH, in which we found TP53 mutations to be a recurrent feature of HGBL-DH with BCL2, but not BCL6 rearrangements.12 Intriguingly, we further observed two

molecular subtypes in MYC/BCL6 only rearranged cases. While selected cases were categorized within the EZB cluster, several cases revealed an association with the BN2 cluster, potentially hinting at a MYC-driven high-grade transformation of a precursor lesion with an origin within the marginal zone, as previously described.24 In addition

Figure 4. Allocation of double-/triple-hit high-grade B-cell lymphoma samples unto the molecular subgroups/clusters of diffuse large B-cell lymphoma, according to LymphGen based on their mutational signature. Additionally, covariates are shown above the plot for each sample. Row names refer to chromosomal alterations, genes and fusions and are background coloured by their specific subtype (light blue = BN2, light orange = EZB).

Haematologica | 107 August 2022

1859

ARTICLE - HGBL-DH/TH mutational landscape

A. Künstner et al.

Figure 5. Survival curves according to ARID1A mutational status. (A) Overall survival (OS) and (B) progression-free survival (PFS) according to ARID1A mutational status. Numbers at risk alongside hazard ratios and P-values according to log-rank testing are provided. Subsequent Bonferroni correction for multiple testing (all genes with a mutational frequency >15% were investigated) identified a significant impact of ARID1A mutations regarding overall survival OS while its primarily significant effect on PFS was reduced to a trend of borderline statistical significance. Fol: folicular; DLBCL: diffuse large B-cell lymphoma.

Haematologica | 107 August 2022

1860

ARTICLE - HGBL-DH/TH mutational landscape

A. Künstner et al.

to these observations, we found TH cases to reflect EZB lymphomas in the vast majority of cases, potentially hinting at BCL6 rearrangements as late and non-defining events in HGBL-TH lymphomagenesis. Supporting this assumption, Pedrosa et al. have shown DLBCL with BCL2 and BCL6, but without MYC rearrangements to be exclusively associated with the EZB cluster.42 Considering the significantly mutated genes, our observations underscore previous assumptions regarding a molecular divergence between BCL2- and BCL6- rearranged HGBL-DH.14,17,43 The predominant mutational distinction between these groups was the presumably FL-derived enrichment for CREBBP mutations in the BCL2-rearranged subgroup. On an exome-wide level we observed an enrichment of the SBS6 signature (implicated in defective DNA mismatch repair) and a significantly diminished congruency with the C3/EZB DLBCL cluster in the BCL6-rearranged subgroup. A pronounced SBS6 signature in BCL6-rearranged cases is in keeping with previous reports by Gu et al. who described genomic instability as a result of defective MMR and thereby a shorter latency to the development of BCL6-driven DLBCL in a murine model.44 Of note, these findings fundamentally dispute the combined characterization in the current World Health Organization classification, despite several shared clinical aspects common to all subtypes of HGBL-DH/TH.1,2 Beyond de novo DLBCL with BCL6 rearrangement, potential alternative explanations for this phenomenon include both clonal evolution and subsequent aggressive transformation from rare cases of BCL6-rearranged marginal zone lymphomas alongside BCL2 non-rearranged/BCL6-rearranged FL, which were previously shown to be characterized by a heterogenous mutational landscape.45,46 From our data, we further deduce an intermediate role for HGBL-TH, which may indicate two divergent paths of clonal evolution originating from a BCL2- or a BCL6-driven disease with subsequent acquisition of the alternative rearrangement. The predominance of the SBS5 signature across all cytogenetic subtypes is most likely attributable to none-filtered FFPE-artifacts and advanced patient age, as was recently described.47,48 Lastly, we describe a pronounced mutational impairment of NOTCH, IL6/JAK/STAT and NFκB signaling pathways and recurrent oncogenetically relevant genes affected by SCNV (including MEF2B, which was previously shown to be enriched in mutations/aberrations within the C3 DLBCL cluster) thereby systematically characterize the oncogenetic footprint of this subgroup of lymphoma. This is further combined with the identification of novel putative mutational drivers (e.g., NCOR1, DTX1, LTB and STAT3) alongside several previously established mutational hotspots in HGBL-DH/TH. Of note, and in keeping with previous observations by Zhang et al., who described an increased AID activity in DH lymphomas, we observe a sig-

nificant accumulation of mutations in known AID and SHM-targets such as PIM1, SOCS1 and others.49-51 Moreover, among these significantly mutated genes we describe ARID1A which emerges as a potential prognosticator of treatment response and outcome from our correlative assessment of clinical and molecular features of our present cohort, which was found to be independent from previously established clinical prognostic factors. We acknowledge the shortcomings inherent to the retrospective design of the study alongside the limited availability of germline DNA for matched pair analysis. The latter aspect is reflected in a significantly elevated number of mutations in non-matched samples and an uneven distribution of controlled cases unto the cytogenetic subtypes. This prompted us to limit our subsequent analysis to significantly mutated genes (except for MYC and BCL2 mutations, which were additionally included based on their proven relevance in prior studies)14,17,23,52 and thereby equalizing the above-mentioned effect. A minor divergence in mutational calls between WES and amplicon sequencing may be attributable to a diverse clonal architecture with mutationally different subclones as DNA samples for WES and Panel-NGS were isolated from the same biopsies but different tissue sections. Additionally, on average between 78.35% and 99.17% (first quantile 91.67%, third quantile 96.59%) of the exome targets were covered with at least 40x coverage per sample, while only variants with a minimum coverage of 40x were considered present, which might have led to the exclusion of variants on a low percentage of occaisons due to too low WES sequencing coverage. However, this trade-off reduces the number of false positive variant calls and thereby enhances confidence in our calls. Pairing of our WES-results with RNA sequencing data, preferably in an extended, clinically annotated cohort, which was beyond the scope of the present study, would further deepen our molecular understanding of HGBL-DH/TH, especially regarding cases with prominent Burkitt or Burkittlike morphology. In summary, our identification of distinct mutational landscapes among HGBL-DH/TH, derived from an exomewide sequencing approach shows both overlapping and distinctive features compared with GC-derived lymphomas such as GCB-type DLBCL and low-grade FL as well as BL. Our work further underscores the developing notion of a recurrent pattern of mutational events driving a potentially unidentified preexisting FL into MYC-driven HGBL-DH/TH, offering insight into the molecular pathogenesis of this provisional entity. By refining the molecular taxonomy for aggressive, GC-derived B-cell lymphomas, these results call into question the current World Health Organization classification system, especially regarding the status of MYC/BCL6-rearranged HGBL.

Haematologica | 107 August 2022

1861

ARTICLE - HGBL-DH/TH mutational landscape

A. Künstner et al.

Disclosure No conflicts of interest to disclose. Acknowledgments AK and HB acknowledge computational support from the OMICS compute cluster at the University of Lübeck. Funding The research was supported by a grant to NG by the Stefan-Morsch-Foundation alongside infrastructural sup-

port. HB acknowledges funding by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany’s Excellence Strategy— EXC 22167-390884018). Data-sharing statement Sequencing data in bam format from WES and panel sequencing have been deposited in the European genomephenome archive (EGA) under the accession number EGAS00001005420.

References 1. Swerdlow SH, Campo E, Harris NL, et al. WHO Classification of Tumors of Haematopoietic and Lymphoid Tissues. Vol. 2. Lyon: IARC; 2017. 586 p. ISBN: 9789283244943. 2. Rosenwald A, Bens S, Advani R, et al. Prognostic significance of MYC rearrangement and translocation partner in diffuse large B-cell lymphoma: a study by the Lunenburg Lymphoma Biomarker Consortium. J Clin Oncol. 2019;37(35):3359-3368. 3. Aukema SM, Kreuz M, Kohler CW, et al. Biological characterization of adult MYC-translocation-positive mature Bcell lymphomas other than molecular Burkitt lymphoma. Haematologica. 2014;99(4):726-735. 4. Aukema SM, Siebert R, Schuuring E, et al. Double-hit B-cell lymphomas. Blood. 2011;117(8):2319-2331. 5. Copie-Bergman C, Cuilliere-Dartigues P, Baia M, et al. MYC-IG rearrangements are negative predictors of survival in DLBCL patients treated with immunochemotherapy: a GELA/LYSA study. Blood. 2015;126(22):2466-2474. 6. Oki Y, Noorani M, Lin P, et al. Double hit lymphoma: the MD Anderson Cancer Center clinical experience. Br J Haematol. 2014;166(6):891-901. 7. Pedersen MO, Gang AO, Poulsen TS, et al. MYC translocation partner gene determines survival of patients with large B-cell lymphoma with MYC- or double-hit MYC/BCL2 translocations. Eur J Haematol. 2014;92(1):42-48. 8. Petrich AM, Nabhan C, Smith SM. MYC-associated and doublehit lymphomas: a review of pathobiology, prognosis, and therapeutic approaches. Cancer. 2014;120(24):3884-3895. 9. Pillai RK, Sathanoori M, Van Oss SB, Swerdlow SH. Double-hit B-cell lymphomas with BCL6 and MYC translocations are aggressive, frequently extranodal lymphomas distinct from BCL2 double-hit B-cell lymphomas. Am J Surg Pathol. 2013;37(3):323-332. 10. Sarkozy C, Traverse-Glehen A, Coiffier B. Double-hit and double-protein-expression lymphomas: aggressive and refractory lymphomas. Lancet Oncol. 2015;16(15):e555-567. 11. Wang XJ, Medeiros LJ, Lin P, et al. MYC cytogenetic status correlates with expression and has prognostic significance in patients with MYC/BCL2 protein double-positive diffuse large B-cell lymphoma. Am J Surg Pathol. 2015;39(9):1250-1258. 12. Gebauer N, Bernard V, Gebauer W, et al. TP53 mutations are frequent events in double-hit B-cell lymphomas with MYC and BCL2 but not MYC and BCL6 translocations. Leuk Lymphoma. 2015;56(1):179-185. 13. Schiefer AI, Kornauth C, Simonitsch-Klupp I, et al. Impact of single or combined genomic alterations of TP53, MYC, and BCL2 on survival of patients with diffuse large B-cell lymphomas: a retrospective cohort study. Medicine (Baltimore).

2015;94(52):e2388. 14. Evrard SM, Pericart S, Grand D, et al. Targeted next generation sequencing reveals high mutation frequency of CREBBP, BCL2 and KMT2D in high-grade B-cell lymphoma with MYC and BCL2 and/or BCL6 rearrangements. Haematologica. 2019;104(4):e154-e157. 15. Stengel A, Kern W, Meggendorfer M, Haferlach T, Haferlach C. Detailed molecular analysis and evaluation of prognosis in cases with high grade B-cell lymphoma with MYC and BCL2 and/or BCL6 rearrangements. Br J Haematol. 2019;185(5):951-954. 16. Momose S, Weissbach S, Pischimarov J, et al. The diagnostic gray zone between Burkitt lymphoma and diffuse large B-cell lymphoma is also a gray zone of the mutational spectrum. Leukemia. 2015;29(8):1789-1791. 17. Cucco F, Barrans S, Sha C, et al. Distinct genetic changes reveal evolutionary history and heterogeneous molecular grade of DLBCL with MYC/BCL2 double-hit. Leukemia. 2020;34(5):1329-1341. 18. Pedersen BS, Quinlan AR. Mosdepth: quick coverage calculation for genomes and exomes. Bioinformatics. 2018;34(5):867-868. 19. Chen S, Zhou Y, Chen Y, Gu J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics. 2018;34(17):i884-i890. 20. Li H. Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM. 2013. [cited 2020 07.06.2020]. Available from: https://arxiv.org/abs/1303.3997. 21. McKenna A, Hanna M, Banks E, et al. The genome analysis toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res. 2010;20(9):1297-1303. 22. Boeva V, Popova T, Bleakley K, et al. Control-FREEC: a tool for assessing copy number and allelic content using nextgeneration sequencing data. Bioinformatics. 2012;28(3):423-425. 23. Chapuy B, Stewart C, Dunford AJ, et al. Molecular subtypes of diffuse large B cell lymphoma are associated with distinct pathogenic mechanisms and outcomes. Nat Med. 2018;24(5):679-690. 24. Wright GW, Huang DW, Phelan JD, et al. A probabilistic classification tool for genetic subtypes of diffuse large B cell lymphoma with therapeutic implications. Cancer Cell. 2020;37(4):551-568.e14. 25. Lawrence MS, Stojanov P, Polak P, et al. Mutational heterogeneity in cancer and the search for new cancerassociated genes. Nature. 2013;499(7457):214-218. 26. Chapuy B, Stewart C, Dunford AJ, et al. Genomic analyses of PMBL reveal new drivers and mechanisms of sensitivity to PD-1 blockade. Blood. 2019;134(26):2369-2382. 27. Mermel CH, Schumacher SE, Hill B, et al. GISTIC2.0 facilitates

Haematologica | 107 August 2022

1862

ARTICLE - HGBL-DH/TH mutational landscape

A. Künstner et al.

sensitive and confident localization of the targets of focal somatic copy-number alteration in human cancers. Genome Biol. 2011;12(4):R41. 28. Brescia P, Schneider C, Holmes AB, et al. MEF2B instructs germinal center development and acts as an oncogene in B cell lymphomagenesis. Cancer Cell. 2018;34(3):453-465.e9. 29. Edginton-White B, Cauchy P, Assi SA, et al. Global long terminal repeat activation participates in establishing the unique gene expression programme of classical Hodgkin lymphoma. Leukemia. 2019;33(6):1463-1474. 30. Lamprecht B, Walter K, Kreher S, et al. Derepression of an endogenous long terminal repeat activates the CSF1R protooncogene in human lymphoma. Nat Med. 2010;16(5):571-9, 1p following 579. 31. Pon JR, Wong J, Saberi S, et al. MEF2B mutations in nonHodgkin lymphoma dysregulate cell migration by decreasing MEF2B target gene activation. Nat Commun. 2015;6:7953. 32. Love C, Sun Z, Jima D, et al. The genetic landscape of mutations in Burkitt lymphoma. Nat Genet. 2012;44(12):1321-1325. 33. Dunleavy K, Little RF, Wilson WH. Update on Burkitt lymphoma. Hematol Oncol Clin North Am. 2016;30(6):1333-1343. 34. Ma MCJ, Tadros S, Bouska A, et al. Subtype-specific and cooccurring genetic alterations in B-cell non-Hodgkin lymphoma. Haematologica. 2022;107(3):690-701. 35. Reddy A, Zhang J, Davis NS, et al. Genetic and functional drivers of diffuse large B cell lymphoma. Cell. 2017;171(2):481-494.e15. 36. Meriranta L, Pasanen A, Louhimo R, et al. Deltex-1 mutations predict poor survival in diffuse large B-cell lymphoma. Haematologica. 2017;102(5):e195-e198. 37. Liu Z, Filip I, Gomez K, et al. Genomic characterization of HIVassociated plasmablastic lymphoma identifies pervasive mutations in the JAK-STAT pathway. Blood Cancer Discov. 2020;1(1):112-125. 38. Cinar M, Rong HR, Chineke I, et al. Genetic analysis of plasmablastic lymphomas in HIV (+) patients reveals novel driver regulators of the noncanonical NF-κB pathway. Blood. 2018;132(Suppl 1):S1565. 39. Schmitz R, Young RM, Ceribelli M, et al. Burkitt lymphoma pathogenesis and therapeutic targets from structural and functional genomics. Nature. 2012;490(7418):116-120. 40. Landau DA, Tausch E, Taylor-Weiner AN, et al. Mutations driving

CLL and their evolution in progression and relapse. Nature. 2015;526(7574):525-530. 41. Gonzalez-Rincon J, Mendez M, Gomez S, et al. Unraveling transformation of follicular lymphoma to diffuse large B-cell lymphoma. PLoS One. 2019;14(2):e0212813. 42. Pedrosa L, Fernandez-Miranda I, Perez-Callejo D, et al. Proposal and validation of a method to classify genetic subtypes of diffuse large B cell lymphoma. Sci Rep. 2021;11(1):1886. 43. Clipson A, Barrans S, Zeng N, et al. The prognosis of MYC translocation positive diffuse large B-cell lymphoma depends on the second hit. J Pathol Clin Res. 2015;1(3):125-133. 44. Gu X, Booth CJ, Liu Z, Strout MP. AID-associated DNA repair pathways regulate malignant transformation in a murine model of BCL6-driven diffuse large B-cell lymphoma. Blood. 2016;127(1):102-112. 45. Ye H, Remstein ED, Bacon CM, et al. Chromosomal translocations involving BCL6 in MALT lymphoma. Haematologica. 2008;93(1):145-146. 46. Nann D, Ramis-Zaldivar JE, Muller I, et al. Follicular lymphoma t(14;18)-negative is genetically a heterogeneous disease. Blood Adv. 2020;4(22):5652-5665. 47. Petljak M, Alexandrov LB, Brammeld JS, et al. Characterizing mutational signatures in human cancer cell lines reveals episodic APOBEC mutagenesis. Cell. 2019;176(6):1282-1294.e20. 48. Alexandrov LB, Nik-Zainal S, Wedge DC, et al. Signatures of mutational processes in human cancer. Nature. 2013;500(7463):415-421. 49. Zhang J, Shi Y, Zhao M, Hu H, Huang H. Activation-induced cytidine deaminase overexpression in double-hit lymphoma: potential target for novel anticancer therapy. Sci Rep. 2020;10(1):14164. 50. Kotani A, Kakazu N, Tsuruyama T, et al. Activation-induced cytidine deaminase (AID) promotes B cell lymphomagenesis in Emu-cmyc transgenic mice. Proc Natl Acad Sci U S A. 2007;104(5):1616-1620. 51. Schuhmacher B, Bein J, Rausch T, et al. JUNB, DUSP2, SGK1, SOCS1 and CREBBP are frequently mutated in T-cell/histiocyterich large B-cell lymphoma. Haematologica. 2019;104(2):330-337. 52. Kridel R, Chan FC, Mottok A, et al. Histological transformation and progression in follicular lymphoma: a clonal evolution study. PLoS Med. 2016;13(12):e1002197.

Haematologica | 107 August 2022

1863

ARTICLE - Non-Hodgkin Lymphoma

Immune pathway upregulation and lower genomic instability distinguish EBV-positive nodal T/NK-cell lymphoma from ENKTL and PTCL-NOS Cho Mar Myint Wai,1,* Shangying Chen,2,* The Phyu,1 Shuangyi Fan,1 Sai Mun Leong,1 Wenning Zheng,3 Louis Ching Yi Low,1 Shoa-Nian Choo,1 Chi-Kuen Lee,1 Tae-Hoon Chung,3 Kenneth Hon Kim Ban,2 Soumita Ghosh,3 Stefanus Lie,3 Seiichi Kato,4,5 Shigeo Nakamura,4 Emiko Takahashi,6 Young-Hyeh Ko,7 Joseph D. Khoury,8 Shih-Sung Chuang,9 Rex K.H. Au-Yeung,10 Soo-Yong Tan,1,11 Soon-Thye Lim,12 Choon-Kiat Ong,13-15 Yong-Howe Ho,16 Li Mei Poon,17 Sanjay de Mel,17 Anand D. Jeyasekharan,3 Wee-Joo Chng,3,17,18 Franziska Otto,19 Leticia Quintanilla-Martinez,19 Federica Zanardi,20 Fabio Iannelli,20 Claudio Tripodo,21 Jason J. Pitt3 and Siok-Bian Ng1,3,11 Department of Pathology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore; 2Department of Biochemistry, Yong Loo Lin School of Medicine, National University of Singapore, Singapore; 3Cancer Science Institute of Singapore, National University of Singapore, Singapore; 4Department of Pathology and Laboratory Medicine, Nagoya University Hospital, Nagoya, Japan; 5Department of Pathology and Molecular Diagnostics, Aichi Cancer Center Hospital, Nagoya, Japan; 6Department of Pathology, Aichi Medical University Hospital, Nagakute, Japan; 7Department of Pathology, Samsung Medical Center, Sungkyunkwan University, Seoul, Korea; 8Department of Hematopathology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA; 9Department of Pathology, Chi-Mei Medical Center, Tainan, Taiwan; 10Department of Pathology, Queen Mary Hospital, The University of Hong Kong, Hong Kong SAR, China; 11Department of Pathology, National University Hospital, National University Health System, Singapore; 12Lymphoma Genomic Translational Research Laboratory, National Cancer Center Singapore, Singapore; Division of Medical Oncology, National Cancer Center Singapore, Singapore; 13Lymphoma Genomic Translational Research Laboratory, Division of Medical Oncology, National Cancer Centre Singapore, Singapore; 14Duke-NUS Medical School, Singapore; 15Genome Institute of Singapore, A*STAR (Agency for Science, Technology and Research), Singapore; 16Department of Pathology, Tan Tock Seng Hospital, Singapore; 17Department of Hematology-Oncology, National University Cancer Institute Singapore, National University Hospital, National University Health System, Singapore;18Department of Medicine, Yong Loo Lin School of Medicine, National University of Singapore, Singapore; 19Institute of Pathology and Neuropathology, Eberhard Karls University of Tübingen and Comprehensive Cancer Center, Tübingen University Hospital, Tübingen, Germany; 20Bioinformatics Unit, IFOM - the FIRC Institute of Molecular Oncology, Milan, Italy and 21Tumor Immunology Unit, University of Palermo School of Medicine, Palermo, Italy 1

Correspondence: Siok-Bian Ng (lead contact) patnsb@nus.edu.sg Jason J. Pitt jason.j.pitt@nus.edu.sg Received: September 9, 2021. Accepted: January 4, 2022. Prepublished: January 13, 2022. https://doi.org/10.3324/haematol.2021.280003 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

CMMW and SC contributed equally as co-first authors.

Abstract Primary Epstein-Barr virus (EBV)-positive nodal T/NK-cell lymphoma (PTCL-EBV) is a poorly understood disease which shows features resembling extranodal NK/T-cell lymphoma (ENKTL) and is currently not recognized as a distinct entity but categorized as a variant of primary T-cell lymphoma not otherwise specified (PTCL-NOS). Herein, we analyzed copynumber aberrations (n=77) with a focus on global measures of genomic instability and homologous recombination deficiency and performed gene expression (n=84) and EBV miRNA expression (n=24) profiling as well as targeted mutational analysis (n=16) to further characterize PTCL-EBV in relation to ENKTL and PTCL-NOS. Multivariate analysis revealed that patients with PTCL-EBV had a significantly worse outcome compared to patients with PTCL-NOS (P=0.002) but not to those with ENKTL. Remarkably, PTCL-EBV exhibited significantly lower genomic instability and homologous recombination deficiency scores compared to ENKTL and PTCL-NOS. Gene set enrichment analysis revealed that many immune-related pathways, interferon α/γ response, and IL6_JAK_STAT3 signaling were significantly upregulated in PTCLEBV and correlated with lower genomic instability scores. We also identified that NFκB-associated genes, BIRC3, NFKB1 (P50) and CD27, and their proteins are upregulated in PTCL-EBV. Most PTCL-EBV demonstrated a type 2 EBV latency pattern and, strikingly, exhibited downregulated expression of most EBV miRNA compared to ENKTL and their target Haematologica | 107 August 2022

1864

ARTICLE - Nodal EBV+ T/NK-cell lymphoma

C.M.M. Wai et al.

genes were also enriched in immune-related pathways. PTCL-EBV also showed frequent mutations of TET2, PIK3CD and STAT3, and are characterized by microsatellite stability. Overall, poor outcome, low genomic instability, upregulation of immune pathways and downregulation of EBV miRNA are distinctive features of PTCL-EBV. Our data support the concept that PTCL-EBV could be considered as a distinct entity, provide novel insights into the pathogenesis of the disease and offer potential new therapeutic targets for this tumor.

Introduction Primary nodal Epstein-Barr virus (EBV)-positive T/NK-cell lymphoma is an uncommon group of peripheral T-cell lymphomas (PTCL) that presents primarily with lymph node disease but may involve a limited number of extranodal organs.1 While the majority of PTCL-EBV are derived from T cells, a minority are bona fide NK-cell lymphomas. They show significant overlap with extranodal NK/T-cell lymphoma, nasal type (ENKTL) as both tumors are associated with EBV and characterized by cytotoxic T- or NKcell proliferation. EBV-positive nodal T/NK-cell lymphomas are more common in the elderly, usually demonstrate a monomorphic growth pattern and lack angiodestruction and prominent necrosis.2,3 A few reports have described clinicopathological features distinct from those of ENKTL, including the lack of nasal involvement, frequent T-cell origin, and a CD8+/CD56– phenotype.2,4 The 2017 World Health Organization (WHO) lymphoma classification recommends that EBV-positive nodal T/NKcell lymphoma be considered as an EBV-positive variant of PTCL, not otherwise specified (PTCL-NOS) as data on this disease are limited, and it is currently unclear whether this group of lymphomas represents a distinct entity.5 The molecular biology of EBV-positive nodal T/NKcell lymphoma, henceforth referred to as an EBV-positive variant of PTCL (PTCL-EBV), and its relationship with ENKTL and PTCL-NOS remains poorly understood - mainly due to the rarity of these tumors and lack of tissue availability. Genomic instability (GI) is a hallmark of cancer and refers to the propensity of cells to accumulate a variety of DNA alterations. These alterations, a subset of which provides a selective growth advantage instrumental for tumorigenesis and progression, may delineate different profiles in entities sharing phenotypic or transcriptional features.6 Importantly, the broad characteristics of GI, as measured by scores, have prognostic and management implications, specifically with regard to the choice of therapeutic agents.7 Given the rarity of PTCL-EBV, scarcity of frozen samples and lack of available cell lines, we leveraged formalinfixed paraffin-embedded tissues. Using gene expression profiling and testing of copy number aberrations, we previously compared PTCL-EBV to ENKTL and showed that PTCL-EBV is characterized by PD-L1 upregulation, expression of T-cell related genes and frequent loss of

14q11.2, which correlates with loss of TCRA loci and Tcell origin.4,8 For the current study, we expanded the disease types and included PTCL-NOS in the comparison, and performed a comprehensive suite of analyses including EBV miRNA expression, custom gene panel sequencing and copy number aberration analysis focusing on global measures of GI. We further re-profiled the gene expression signatures using an improved microarray on our cases, including those studied in the previous work.4 Interestingly, we demonstrated that PTCL-EBV displays a remarkably lower degree of GI despite its more aggressive outcome among the three diseases and is characterized by upregulation of immune-related pathways and lower expression of EBV miRNA compared to ENKTL and PTCL-NOS. These novel findings not only offer insights into the pathogenesis of PTCL-EBV and raise considerations regarding its treatment, but also support the proposal that PTCL-EBV is a distinct entity in the WHO classification.

Methods Study cohort Cases from multiple institutions were reviewed by two hematopathologists to confirm the diagnosis of ENKTL (n=89), PTCL-EBV (n=25) and PTCL-NOS (n=36) based on the 2017 WHO lymphoma classification.5 PTCL-EBV shows a similar phenotype as ENKTL but, unlike ENKTL, (i) patients present primarily with nodal disease where the bulk of tumor is localized; (ii) nasal involvement is lacking; (iii) it often shows a CD8+/CD56– phenotype; and (iv) it is often of T-cell origin (Online Supplementary Table S1). Systemic and cutaneous EBV-positive T/NK lymphoproliferative diseases occurring in children were excluded. The diagnosis of PTCL-NOS was established when other specific subtypes of PTCL had been excluded. Cytotoxicity was defined as the expression of at least one cytotoxic marker (TIA1, granzyme B). Clinical data including age, sex, disease type, stage, International Prognostic Index (IPI) score, expression of CD4, CD8 and CD56, T or NK lineage, treatment and overall survival were obtained (Online Supplementary Table S2A, B). This study was approved by the National Healthcare Group Domain Specific Review Board B (2009/00212). Copy number aberration analysis An OncoScan® FFPE assay was performed on 34 ENKTL,

Haematologica | 107 August 2022

1865

ARTICLE - Nodal EBV+ T/NK-cell lymphoma

C.M.M. Wai et al.

14 PTCL-EBV and 29 PTCL-NOS cases as previously published.9 Copy number aberrations were analyzed using OncoScan® Console (v1.3) software (ThermoFisher Scientific, Waltham, MA, USA). Segmentation results were analyzed using GISTIC (v2.0.22)10 with the GENCODE hg19 build. GI and homologous recombination deficiency (HRD) scores were calculated according to published methods.11 Gene expression profiling Gene expression profiling was performed on 35 ENKTL, 23 PTCL-EBV, and 26 PTCL-NOS cases using a GeneChip® Clariom D Assay (Human) array and the data were analyzed as described previously.12 Gene expression profiling (GSE160119) and copy number aberration (GSE160118) raw data were deposited in the Gene Expression Omnibus (GEO). Polymerase chain reaction analysis of EBV miRNA and EBV latency Quantitative reverse transcription polymerase chain reaction (RT-qPCR) analysis of miRNA expression was performed using IDEAL miRNA qPCR assays (MiRXES, Singapore) (n=24) according to the manufacturer’s instructions on a QuantStudio™ 5 System (ThermoFisher Scientific, Waltham, MA, USA). RT-qPCR of EBV genes was performed to confirm EBV latency in PTCL-EBV and ENKTL. Fluorescence in situ hybridization and multiplex immunofluorescence Fluorescence in situ hybridization (FISH), multiplex immunofluorescence and multispectral imaging were performed as previously described.4 Mutational analysis targeted next-generation sequencing Targeted mutation analysis was performed by next-generation sequencing (Ion GeneStudio S5 prime, Thermo Fisher Scientific, Waltham, MA, USA) using an AmpliSeq customized T/NK-lymphoid panel comprising 35 genes recurrently mutated in ENKTL and PTCL-NOS (Online Supplementary Table S3A) and the 484-gene NovoPMTM 2.0 panel (Online Supplementary Table S3B) (total 500 genes including 19 common genes). Statistical methods Overall survival was investigated using Kaplan-Meier nonparametric survival analysis and log-rank tests. The effects of disease type, age, sex and disease stage were analyzed using unadjusted univariate and multivariate Cox proportional hazard models. Analyses were performed in R with the “survival” (v0.1.2) and “survminer” (v0.4.6) packages. Gene expression profiling analysis was performed in R with the “limma” package (v3.40.6). Unsupervised hier-

archical clustering was performed using Spearman distance and Ward.D2 linkage. All additional statistical tests were performed in R. The diagnostic criteria for the diseases, T/NK lineage assignment, detailed descriptions of study cohorts, experimental methods, data analysis and data accession are provided in the Online Supplementary Materials.

Results Differences in clinical and survival variables among the disease groups The salient clinicopathological and immunohistochemical features of ENKTL, PTCL-EBV and PTCL-NOS are summarized in Table 1A and Online Supplementary Table S2A, B. All 25 patients with PTCL-EBV, comprising 14 Japanese, eight Chinese, two Korean and one Bangladeshi person, presented with nodal disease and the bulk of disease involved lymph nodes. Phenotypically, PTCL-EBV was often positive for CD8 (17/25, 68%), cytotoxic markers (100%) and TCRB (12/25, 48%). CD56 was commonly negative (18/24, 75%) and none of the cases tested expressed TCRG. Four of 23 (17%) cases tested expressed CD4, of which one was positive for CXCL13 and CD10 (focal). Most of the cases of PTCL-EBV were of T-cell origin (20/23, 87%), in accordance with previous reports3,13–15 (Figure 1A). With regards to treatment, the patients in all three groups from different institutions were treated with a heterogeneous combination of chemotherapy regimens (Online Supplementary Table S2A). This is understandable as these are rare lymphomas lacking standardized and effective treatment. Two out of 15 PTCL-EBV patients with known treatment data were treated with SMILE therapy. One died 3.5 months after diagnosis and the other died 12.6 months after diagnosis. Despite the treatment heterogeneity, our results revealed that patients with PTCL-EBV had a significantly shorter median overall survival (4.6 months) compared to those with ENKTL (14.7 months, P=0.001) and PTCL-NOS (26 months, P=0.007). (Figure 1B; Table 1). Univariate analysis identified advanced disease stage (stage 4) (P<0.001) and older age (P=0.001) as significantly associated with poor prognosis. Compared to patients with PTCL-EBV, patients with ENKTL had a 61% lower risk of death (HR=0.39, 95% CI: 0.22-0.69, P=0.001), while those with PTCL-NOS had a 59% lower risk (HR=0.41, 95% CI: 0.21-0.80, P=0.009). After adjusting for disease stage, sex and age, PTCL-NOS remained significantly less aggressive compared to PTCLEBV (HR=0.3, 95% CI: 0.14-0.65, P=0.002) (Table 2). A Cox proportional-hazards model was also performed and after controlling for lineage, patients with PTCL-EBV showed significantly worse outcome than those with ENKTL (HR=2.34, 95% CI: 1.05-5.24, P=0.038), indicating that the

Haematologica | 107 August 2022

1866

ARTICLE - Nodal EBV+ T/NK-cell lymphoma

C.M.M. Wai et al.

Figure 1. Morphological features and survival of PTCL-EBV patients compared to those with ENKTL and PTCL-NOS. (A) Representative images of PTCL-EBV. The tumor cells are large and pleomorphic (a, Hematoxylin & eosin, original magnification x400). They are positive for CD3 (b, original magnification x400), CD8 (c, original magnification x600), T-cell receptor, beta (TCRb) (d, original magnification x600), EBER (e, original magnification x400) and granzyme B (f, original magnification x400). The positive expression for TCRβ indicates a T-cell origin. (B) Kaplan-Meier survival curve depicting overall survival of three disease groups. Patients in the PTCL-EBV group had significantly shorter overall survival compared to those in the ENKTL and PTCL-NOS groups.

Table 1. Clinicopathological features and gene expression profiling in patients with ENKTL, PTCL-EBV and PTCL-NOS. Parameters

ENKTL (N=89)

PTCL-EBV (N=25)

PTCL-NOS (N=36)

Median age in years (range, SD)

49 (17-82, 16.20)

59 (32-89, 14.43)

62 (11-95, 18.22)

0.002*

27/62

5/20

13/23

0.399†

16/57/16

20/3/2

36/0/0

<0.001†

52/29

3/20

4/22

<0.001†

37/24/8

3/11/6

1/6/5

0.001†

Median survival‡ in months

14.7

4.6

26.0

0.002†

CD56‡ positive/negative, N

66/20

6/18

1/25

<0.001†

CD8‡ positive/negative, N

11/64

17/8

11/25

<0.001†

TCRb‡ positive/negative, N

9/51

12/13

12/6

<0.001†

TCRγ‡ positive/negative, N

2/54

0/14

6/12

0.002†

Female/male, N COO, T/NK/indeterminate, N Stage‡, 1&2/3&4, N IPI score‡, low/intermediate/high, N

Kruskal-Wallis test, †chi-square test, ‡only cases with available data were included for analysis, §log-rank test; SD: standard deviation; COO, cell of origin; IPI: International Prognostic Index.

Haematologica | 107 August 2022

1867

ARTICLE - Nodal EBV+ T/NK-cell lymphoma

C.M.M. Wai et al.

Table 2. Univariate and multivariate analyses for overall survival in the ENKTL, PTCL-EBV and PTCL-NOS groups. Variable Disease type

Stage

Sex

Category

Multivariate

CI (95%)

ENKTL

0.39

0.22-0.69

0.001

0.65

0.34-1.26

0.2

PTLC-EBV

Ref

PTCL-NOS

0.41

0.21-0.80

0.3

0.14-0.65

Ref

1.33

0.54-3.22

0.50

1.5

0.61-3.67

0.4

2.46

1.00-6.07

0.051

1.56

0.51-4.81

0.4

4.18

2.31-7.57

<0.001

4.56

2.43-8.54

<0.001

Female

Ref

Male

1.23

0.76-1.99

0.40

1.38

0.82-2.32

0.2

1.02

1.01-1.04

0.001

1.02

1.01-1.04

0.006

Low

Ref

Intermediate

5.16

2.32-11.5

<0.001

High

8.33

3.48-19.9

<0.001

Age IPI score

Univariate

0.009

0.002

HR: hazard ratio; CI: confidence interval; Ref: reference; IPI: International Prognostic Index.

worse outcome of PTCL-EBV patients compared to ENKTL patients is independent of lineage. Interestingly, there was no longer a significant difference in the survival outcome between ENKTL and PTCL-EBV patients, suggesting that the worse outcome of those with PTCL-EBV could be ascribed to older age and more advanced disease stage at diagnosis. Differences in focal copy number aberrations among disease groups In order to identify differences in copy number profiles of the three diseases, we analyzed copy number aberrations of patients with ENKTL (n=34), PTCL-EBV (n=14) and PTCLNOS (n=29) and identified recurrent aberrations (q<0.25) with their putative target genes across all samples (Figure 2A). Previously,4 comparing ENKTL (n=29) and PTCL-EBV (n=12), we noted differences in 14q11.2 and were able to reproduce those here. However, with the addition of five and two new ENKTL and PTCL-EBV samples, respectively – as well as a cohort of PTCL-NOS – we were able to identify multiple unreported differences in focal copy number aberration rates across these groups. Of the 11 recurrently gained regions, 3p14.1, 6p22.3, 6p22.1 and 17q21.33 occurred at significantly different frequencies across disease groups (P<0.05, c2 test) (Online Supplementary Table S4). Gain of 3p14.1 was found in 14.3% of PTCLEBV cases compared to 5.9% and 76.0% of ENKTL and PTCL-NOS cases, respectively (P<0.001). Gain of 6p22.3 was observed more frequently in ENKTL (20.6%) and PTCL-NOS (58.6%) than in PTCL-EBV (7.1%) (P=0.005). Gain of 6p22.1 was observed in 21.4% of PTCL-EBV cases compared to 8.8% of ENKTL and 58.6% of PTCL-NOS cases

(P=0.001). Of the nine recurrent losses, only 14q11.2 occurred at significantly different frequencies among the disease groups (Online Supplementary Table S5). As expected, loss of 14q11.2 was observed in 100% of PTCL-EBV cases,4 whereas it was only found in 20.6% and 58.6% of ENKTL and PTCL-NOS, respectively (P=0.001). It is unclear whether the high frequency of 14q11.2 loss in PTCL-EBV is a result of additional critical driver events occurring at the TCRA locus, on top of monoclonal TCRA gene rearrangement, or related to preferential TCR usage in PTCL-EBV due to possible underlying immune perturbations and antigen selection that predispose to malignancy.16 The 14q11.2 loss in PTCL-NOS is slightly low, which may be a result of coverage bias related to the probe design of the Oncoscan assay or a lack of TCRA rearrangement in some PTCL-NOS cases.8 The top two recurrent copy-number gains (3p14.1 and 6p22.1) were subsequently validated by FISH (Online Supplementary Figure S1 and Online Supplementary Table S6). PTCL-EBV exhibits lower genomic complexity than other disease groups We further compared the genome-wide levels of copy number aberrations across the three disease types (Figure 2B, Online Supplementary Figure S2). Analysis of copy number burden (segment counts) revealed that PTCL-EBV exhibited significantly fewer segments than ENKTL (P=0.016, Mann-Whitney U test) and PTCL-NOS (P<0.001) (Figure 2C, left panel). While copy number segment differences between ENKTL and PTCL-EBV appeared primarily driven by gains (Figure 2C, middle panel), PTCL-NOS demonstrated increased losses compared to other groups

Haematologica | 107 August 2022

1868

ARTICLE - Nodal EBV+ T/NK-cell lymphoma

C.M.M. Wai et al.

Continued on following page. Haematologica | 107 August 2022

1869

ARTICLE - Nodal EBV+ T/NK-cell lymphoma

C.M.M. Wai et al.

Figure 2. Composite copy number alteration profiles of three disease groups. (A) Composite map showing the focal copy number alteration spectrum in three disease groups. The red and blue represent copy number gain and loss, respectively. Each row represents a genomic locus while each column represents a case. Bars on the right represent the proportion of each disease type in copy number aberrations identified. PTCL-EBV patients had fewer focal copy number aberrations compared to patients with ENKTL and PTCL-NOS. (B) Penetrance plots showing the frequency of gains and losses of genomic loci in ENKTL, PTCL-EBV and PTCL-NOS groups. The X-axis represents chromosome number and the Y-axis indicates the proportion of gain or loss of the corresponding genomic loci within the corresponding population. Red bars denote copy number gains and blue bars denote copy number losses. PTCL-EBV exhibited less frequent genomic alterations compared to other disease groups. (C) Boxplot depicting total copy number segment counts (left), gains only (middle) and losses only (right) across the three diseases. Differences among the three groups were determined using the Kruskal-Wallis test while pairwise comparisons were assessed by the MannWhitney U test (P values shown). PTCL-EBV displayed lower segment counts compared to ENKTL and PTCL-NOS. (D, E) Copy number segment size distribution of gains (D) and losses (E) in the three disease groups. Statistical significance was determined using a two-sample Kolmogorov-Smirnov test with P values indicated in the table. PTCL-EBV gain and loss distributions were enriched for smaller copy number segments compared to the other disease groups.

(Figure 2C, right panel). The sizes of gains and losses are known to vary across cancer types and these differences may be attributable to different mutagenic processes.17 Pairwise comparisons of copy number aberration size distributions indicated that all three diseases exhibited distinct patterns of gains and losses (P<0.05, Kolmogorov-Smirnov test) (Figure 2D, E) with PTCL-EBV showing a propensity to smaller gains (~300 kb in size) and a unimodal distribution for losses peaking around 300 kb. To further examine the genomic complexities of the three disease groups, GI and HRD scores were calculated; the former based on the ratio of total length of regions with aberrant copy number and the latter on the loss of heterozygosity, telomere allelic imbalance and large-scale state transitions.11 We observed that PTCL-EBV had significantly lower GI and HRD scores compared to ENKTL (P<0.001 [GI], P=0.004 [HRD], Mann-Whitney U) and PTCLNOS (P=0.0012 [GI], P=0.025 [HRD]) (Figure 3A, B). Despite these lower scores, ploidy levels were similar across all three groups, indicating that these results were not attributable to differential rates of whole genome duplication (Figure 3C). The GI score remained significantly lower in PTCL-EBV T-lineage (n=13) than in ENKTL T-lineage (n=9) cases, indicating that the lower GI score in PTCL-EBV compared to ENKTL is not related to lineage. Nevertheless, future studies with larger sample sizes will be necessary for a comparison of lineage effects across the disease groups. In line with previous reports describing TP53 alterations co-occurring with increased GI or HRD,18 we also observed that TP53 losses were associated with higher GI score (Online Supplementary Figure S3) and were less frequent in PTCL-EBV (7.1%) than in ENKTL (26.5%; n.s., Fisher exact test) and PTCL-NOS ( 31.4%; P=0.03) (Online Supplementary Table S5). Using 14 publicly available Oncoscan datasets (Online Supplementary Table S7), we compared GI and HRD scores of other hematolymphoid neoplasms and solid cancers with those of PTCL-EBV, ENKTL and PTCL-NOS. Oncoscan datasets for T-cell lymphomas were unavailable. GI and HRD scores varied widely across cancer types, with the highest scores in solid cancers and the lowest in chronic myeloid

leukemia and pediatric-type follicular lymphoma (Figure 3D). Within hematolymphoid malignancies, high-grade lymphomas, such as Burkitt-like lymphomas, large B-cell lymphomas, ENKTL and PTCL-NOS, had higher GI than low-grade malignancies. Compared to other aggressive Bcell lymphomas, PTCL-EBV demonstrated a remarkably low GI score (all P<0.003). NF-κB and immune pathways are overexpressed in PTCL-EBV To explore the biological pathways associated with PTCLEBV, we investigated genes that were differentially expressed between PTCL-EBV and ENKTL (EBVvsENKTL) or PTCL-NOS (EBVvsNOS). After filtering non-consensus coding sequence probes, 244 and 95 differentially expressed genes (P<0.01, adjusted P<0.05) were identified in EBVvsENKTL and EBVvsNOS, respectively (Online Supplementary Table S8). Interestingly, all 95 EBVvsNOS differentially expressed genes overlapped with the 244 from EBVvsENKTL. As expected, unsupervised hierarchical clustering of these 244 genes revealed three distinct clusters with PTCL-EBV separated from ENKTL and PTCL-NOS (Online Supplementary Figure S4). To assess differentially expressed genes for over-representation of gene ontology terms and protein-protein interactions, genes differentially expressed in EBVvsENKTL and EBVvsNOS were independently submitted to STRING.19 Both sets of differentially expressed genes were enriched for numerous immunity-related processes (Online Supplementary Tables S9 and S10) and known protein-protein interactions (P<0.005) (Online Supplementary Figure S5). Subsequent analyses of known gene-gene interactions identified NFκB-associated genes BIRC3, NFKB1, TLR8 and CD27 – which are upregulated in PTCL-EBV – as central nodes in the differentially expressed gene networks (Figure 4). To corroborate differential expression data at the protein level, multiplexed immunofluorescence was performed (Figure 5, Online Supplementary Figure S6A) and revealed significantly upregulated expression of BIRC3 and p50 (NFKB1) in tumor (Online Supplementary Figure S6B, C) and non-tumor (Online Supplementary Figure S6E, F) cells of

Haematologica | 107 August 2022

1870

ARTICLE - Nodal EBV+ T/NK-cell lymphoma

C.M.M. Wai et al.

Continued on following page. Haematologica | 107 August 2022

1871

ARTICLE - Nodal EBV+ T/NK-cell lymphoma

C.M.M. Wai et al.

Figure 3. Differences in genomic instability score, homologous recombination deficiency score and ploidy across different disease groups. Segmentation output data from OncoScan microarray (n=77 cases; ENKTL=34, PTCL-EBV=14, PTCL-NOS=29) was analyzed and quantified for scores of (A) GI, (B) LOH-HRD, LST-HRD, AIL-HRD and scaled HRD and (C) ploidy. (D) Comparison of GI- and HRD- scores across different cancer groups also profiled via OncoScan. Oncoscan datasets on T-cell lymphomas were unavailable. Wide variation of GI- and HRD- scores was observed across cancer types with HNSCC having highest scores while CML and pFL had lowest. High-grade lymphomas, such as Burkitt-like lymphomas, large B-cell lymphomas, ENKTL and PTCLNOS had higher GI- and HRD-scores than low-grade lymphomas. Our results showed that PTCL-EBV exhibited significantly lower GI- and HRD- scores among aggressive lymphomas and various solid tumors. Statistical significance was determined using Kruskal-Wallis tests for differences among the three disease groups while Mann-Whitney U tests were used for pairwise comparisons. LOH: loss of heterozygosity; LST: large-scale state transitions; AIL: telomere allelic imbalance; BL: Burkitt-like; CML: chronic myeloid leukemia; ESCC: esophageal squamous cell carcinoma; HNSCC: head and neck squamous cell carcinoma; LBC: large B-cell; LNMCC: lymph node metastases in colon cancer; OTC: oral tongue carcinoma; pFL: pediatric-type follicular lymphoma; RCC: renal cell carcinoma; SCC: synchronous colorectal cancer; HRD: homologous recombination deficiency.

PTCL-EBV compared to ENKTL and PTCL-NOS. CD27 expression was significantly higher in tumor and non-tumor cells (Online Supplementary Figure S6D, G) of PTCL-EBV compared with ENKTL (both P<0.001) but not with PTCLNOS. Interestingly, the proportions of non-tumor cells which were double positive for CD27/BIRC3, CD27/p50, and triple positive for CD27/p50/BIRC3 within a single nontumor cell, were also significantly higher in PTCL-EBV than in ENKTL and PTCL-NOS (Online Supplementary Figure S6A, H-K). To further explore NFκB across EBVvsENKTL and EBVvsNOS, we performed gene set enrichment analysis using five curated sets of NFκB target genes (Online Supplementary Methods). We observed consistent and significant upregulation of NFκB target gene expression in PTCL-EBV compared to ENKTL and PTCL-NOS (Figure 6A). In addition, we found no significant difference in tumor content (Online Supplementary Figure S7A) and observed a rather homogeneous composition of the main immune components of the TME according to computational transcriptome deconvolution across the disease groups (Online Supplementary Figure S7B) using CIBERSORTx. This suggests that the upregulation of immune-related pathways in PTCL-EBV is unlikely to be related to tumor content or TME composition difference between the disease groups. Overall, these results indicate prominent immune pathway upregulation and NFκB activation in PTCL-EBV, suggesting the potential role of persistent NFkB signaling in the disease pathogenesis. IFNγ, JAK-STAT and NFκB is upregulation in PTCL-EBV and correlation with PD-L1 To further elucidate the biological pathways associated with GI, we correlated GI score to the expression of each gene across all samples, which resulted in a list of genes ranked by the Spearman rho. This list was submitted to gene set enrichment analysis to identify hallmark gene sets whose expression was associated with GI scores (Online Supplementary Table S11). Our top three gene sets – interferon_alpha_response (false discovery rate <0.001), interferon_gamma_response (false discovery rate <0.001)

and IL6_JAK_STAT3_signaling (false discovery rate =0.001) (Online Supplementary Figure S8A-C) – displayed inverse correlations, indicating that these immune-related pathways are upregulated in PTCL-EBV and coincide with lower GI scores. We previously reported that the expression of PD-L1 is higher in PTCL-EBV than in ENKTL.4 Given that both interferon_gamma_response (IFNγ) and STAT3 are able to induce PD-L1 expression at both gene and protein levels in cancers, including ENKTL,20,21 we correlated the gene expression of IFNγ and the IL6_JAK_STAT3 pathway with PDL1 (CD274) to understand mechanisms driving PD-L1 upregulation in our disease groups. We observed a significant correlation between the gene expression of IFNγ (R=0.55, P<0.001) and IL6_JAK_STAT3 genes (R=0.79, P<0.001) with CD274 (Figure 6B, C). Similarly, we also assessed the association between NFκB activity and CD274 since NFκB can transcriptionally upregulate CD274 expression.22,23 A significant correlation between the median expression of NFκB transcriptional target genes and CD274 was also observed (R=0.69, P<0.001) (Figure 6D). Overall, it is possible that the upregulation of PD-L1 in PTCL-EBV may be related to activation of IFNγ, IL6_JAK_STAT3 and NFκB. EBV miRNA are downregulated in PTCL-EBV compared to ENTKL Since ENKTL and PTCL-EBV are both associated with EBV, we compared EBER expression, tumor content (via Oncoscan), and EBV miRNA (via qPCR) in both diseases. There was no significant difference in EBER positivity (Online Supplementary Figure S9A) or tumor content (Online Supplementary Figure S9B) between the two diseases. Based on the gene expression of EBNA1, EBNA2, LMP1 and LMP2A by RT-PCR, the majority (9/13, 69%) of PTCL-EBV showed a type 2 EBV latency pattern, while four cases (31%) demonstrated a type 3 latency pattern (Online Supplementary Table S12). Interestingly, PTCL-EBV (n=9) displayed a widespread lower EBV miRNA expression compared to ENKTL (n=15) and clustered separately from it (Online Supplementary Figure S10). The expression of 32 of 42 (76%) EBV

Haematologica | 107 August 2022

1872

ARTICLE - Nodal EBV+ T/NK-cell lymphoma

C.M.M. Wai et al.

miRNA was significantly lower in PTCL-EBV (adjusted P<0.05, t-test) (Online Supplementary Table S13). To better understand the potential transcriptional impact of this differential miRNA regulation, we correlated the expression of each differentially expressed EBV miRNA and its predicted targets. Given that miRNA negatively regulate target mRNA,24 all target genes that were negatively correlated (adjusted P<0.05) to their corresponding EBV miRNA were analyzed (n=172) (Online Supplementary Table S14). Strikingly, the pathways most

enriched (adjusted P<0.05) within this set of genes represented either immunity or interferon signaling (Online Supplementary Table S15). After calculating a gene expression index (median expression) for these target genes, we observed significantly higher expression in PTCL-EBV (P=0.03, Mann-Whitney U) (Online Supplementary Figure S11), which is consistent with lower EBV miRNA expression in this group. Overall, these results suggest that downregulation of EBV miRNA is unlikely to be related to a difference in tumor or EBER content and

Figure 4. Integrated network analysis of differentially expressed genes in the three disease groups. (A) STRING-based network (149 nodes; 209 edges) for differentially expressed genes between PTCL-EBV and ENKTL. NFKB1 and TLR8 are network hubs based on betweenness centrality calculations. (B) STRING-based network (45 nodes; 52 edges) for differentially expressed genes between PTCL-EBV and PTCL-NOS. BIRC3 and TLR8 are network hubs based on betweenness centrality calculations. Nodes are sized based on their degree (i.e., number of incoming edges). Genes that fall within the most enriched gene ontology process for each network — which are (A) “Regulation of Immune System Process” and (B) “immune response” — are indicated in yellow. All other genes are colored light purple.

Haematologica | 107 August 2022

1873

ARTICLE - Nodal EBV+ T/NK-cell lymphoma

C.M.M. Wai et al.

Figure 5. Multiplex immunofluorescence analysis of BIRC3, p50 (NFκB1) and CD27 across all three diseases. (A) Protein expression of CD27, p50 (NFκB1) and BIRC3 in ENKTL (left panel), PTCL-EBV (middle panel) and PTCL-NOS (right panel) using multiplexed immunofluorescence. For each panel, the left column represents the multiplexed immunofluorescence staining and the right column shows the corresponding multispectral analysis masks. PTCL-EBV showed higher expression of CD27 (membrane, green), p50 (NFκB1) (nuclear, yellow) and BIRC3 (nuclear, cyan), compared to ENKTL and PTCL-NOS. CD27+CD3+ cells are white while CD27+CD3- are green in CD3/CD27 masks. P50+CD3+ cells are white while P50+CD3- are yellow in CD3/p50 masks. BIRC3+CD3+ cells are white while BIRC3+CD3- cells are cyan in CD3/BIRC3 masks. The scale bars indicate 100 µm.

could, in part, contribute to the distinctive pattern of im- purity of our samples and the high frequency of TET2 mumune gene transcription in PTCL-EBV. tations and their associated variant allele frequencies, it is unlikely that these mutations are attributable to clonal PTCL-EBV showed frequent mutations of TET2, PIK3CD hematopoiesis of indeterminate potential (CHIP).25 Howand STAT3 ever, in the absence of blood samples from these patients, Based on the evidence that GI, gene expression, as well we cannot entirely rule out CHIP as a possible source of as EBV-miRNA patterns can delineate a distinct profile for TET2 mutation in some of our PTCL-EBV cases. Eight out PTCL-EBV, we further investigated whether this disease of nine cases positive for TET2 mutation were positive for harbors mutations in known driver genes of PTCL-NOS, CD8, indicating that they were not PTCL with T-follicular ENKTL and solid cancers using a 35-gene T/NK lymphoid helper phenotype. Interestingly, TP53 mutations, companel and 484-gene NovoPMTM 2.0 assay (total 500 genes monly present in ENKTL, were not detected in our PTCLwith 19 common genes covered in both panels). EBV cases. The median number of mutations detected The most commonly mutated gene was TET2 (9/14, 64%) was 2.5 per sample (range, 1 to 11). followed by PIK3CD (3/9, 33%), STAT3 (3/16, 19%), DDX3X The tumor mutational burden score ranges from 0 to 7.86 (2/10, 20%) and PTPRD (2/11, 18%) (Online Supplementary mutations/MB (median 4.285 mutations/MB) (Online SupFigure S12). The variant allele frequency ranges for TET2 plementary Figure S13A). A tumor mutational burden of in the NovoPMTM 2.0 panel and T/NK lymphoid panel were less than 5 is regarded as a low mutational burden in 22.2%-40.7% and 22%-76%, respectively. Given the tumor some studies.26,27 Eleven samples tested had a microsatelHaematologica | 107 August 2022

1874

ARTICLE - Nodal EBV+ T/NK-cell lymphoma

C.M.M. Wai et al.

Figure 6. PTCL-EBV demonstrates NFκB transcriptional target gene upregulation and PD-L1 (CD274) expression associates with immune pathway activation across all three diseases. (A) Gene set enrichment analysis (GSEA) comparing PTCL-EBV to ENKTL (EBVvsENKTL) and PTCL-NOS (EBVvsNOS) across five sets of NFκB transcriptional target genes. Genes were ranked by their relative expression differences in EBVvsENKTL and EBVvsNOS then submitted to GSEA. All enrichment scores were positive indicating target gene upregulation in PTCL-EBV compared to ENKTL and PTCL-NOS. The vertical dashed line represents a 0.05 P value threshold. Correlation of PD-L1 (CD274) expression with (B) IFNγ, (C) the IL6_JAK_STAT pathway and (D) NFκB target gene expression across all three diseases. Our results showed that expression of IFNγ and IL6_JAK_STAT genes (median) correlated with PD-L1 gene expression. Taking the union of the five aforementioned gene sets, there was also a positive correlation between Continued on following page. Haematologica | 107 August 2022

1875

ARTICLE - Nodal EBV+ T/NK-cell lymphoma

C.M.M. Wai et al.

NFκB transcriptional target gene expression (median) and PD-L1 expression. Correlations were assessed using the Spearman method. Rho and P values are shown. (E) Possible model of PTCL-EBV pathogenesis involving the activation of the NFκB pathway and upregulation of PD-L1, BIRC3, and CD27. BIRC3 plays key roles in the regulation of NFκB signaling and apoptosis. CD27 contributes to anti-tumor cytotoxic T-cell lymphocyte response in the host, T-cell exhaustion, compromise in antitumor immunity. In addition, EBV LMP1 and upregulation of IFNγ and IL6_JAK_STAT3 could also contribute to PDL1 overexpression in PTCL-EBV. Activation of these signaling pathways eventually contributes to inflammation, T-cell and immune activation, thereby promoting proliferation and survival, metastasis, immune evasion and oncogenesis. Some of these genes and signaling pathways may serve as potential therapeutic targets for PTCL-EBV and are indicated in red boxes. Dotted lines indicate hypothetical postulations which have not been experimentally validated in PTCL-EBV. Figure created with BioRender.com.

lite instability score below the threshold score of 0.4 (median 0.1882) and are regarded as being microsatellite stable (Online Supplementary Figure S13B). PTCL-EBV is more aggressive and shows lower genomic instability than cytotoxic PTCL-NOS Given that PTCL-EBV is characterized by a cytotoxic phenotype,2,5 we attempted to determine whether PTCL-EBV shows similarities with cytotoxic PTCL-NOS. We performed multivariate survival analysis and compared the GI and HRD scores, between PTCL-EBV, PTCL-NOS cytotoxic (n=15) and non-cytotoxic (n=11) cases. Interestingly, we observed that PTCL-EBV was significantly more aggressive than cytotoxic PTCL-NOS (HR=0.22, 95% CI: 0.08-0.58, P=0.002) but not non-cytotoxic PTCL-NOS (Online Supplementary Table S16). PTCL-EBV exhibited significantly lower GI scores compared to both cytotoxic (P=0.041) and non-cytotoxic PTCLNOS (P<0.001) (Online Supplementary Figure S14). HRD scores were lower in PTCL-EBV than in non-cytotoxic PTCL-NOS (P=0.021) but not cytotoxic PTCL-NOS. No significant difference was detected in survival, ploidy, and GI and HRD scores between cytotoxic and non-cytotoxic PTCL-NOS and there were no genes significantly differently expressed between the two. Overall, our results show a significant difference in overall survival and GI score between PTCL-EBV and cytotoxic PTCL-NOS, suggesting biological differences between these two tumors.

Discussion PTCL-EBV is a rare and aggressive tumor that occurs mostly in East Asia. It is a poorly understood lymphoma that is not recognized yet as a distinct entity but is considered as a variant of PTCL-NOS in the current WHO classification.5 In line with published literature, all cases of PTCL-EBV in this study are from South-East/East Asia. Herein, we performed an integrative analysis to compare PTCL-EBV with ENKTL and PTCL-NOS, and our findings demonstrated distinctive features in PTCL-EBV which set it apart from the other two tumors. Our results support the consideration of PTCL-EBV as a distinct entity in the WHO classification. Copy-number analysis of PTCL-EBV demonstrated a strik-

ing profile characterized by not only significantly fewer copy number aberrations, but also a different size distribution compared to ENKTL and PTCL-NOS. Remarkably, PTCL-EBV was more genomically stable than the other two diseases, as reflected by lower GI and HRD scores. High-grade lymphomas had higher GI and HRD scores than low-grade malignancies, compatible with reports documenting that GI is associated with aggressive lymphomas and high-grade transformation.28,29 Similarly, within and across a variety of solid tumors, increased copy numberbased GI is associated with poor outcomes.30 In this regard, the uniformly low GI and HRD scores in PTCL-EBV belies its aggressive behavior. The discrepancy between low GI and aggressive behavior is novel but difficult to explain. Given the prominent immune-related gene expression profile, it is reasonable to hypothesize that the aggressiveness of PTCL-EBV may be related to a cancer-promoting inflammation which blocks anti-tumor immunity and directs the tumor microenvironment toward a tumor-permissive state.31 Interestingly, our findings showed that PTCL-EBV has few TP53 mutations and has a paucity of TP53 losses compared to ENKTL and PTCL-NOS. Given the paucity of TP53 alterations in PTCLEBV, it is possible that low GI in PTCL-EBV may be related to an unperturbed p53 tumor suppressor function, which is necessary to preserve genomic stability and integrity through cell-cycle arrest, senescence and apoptosis.32 Given that chromosome instability in diffuse large B-cell lymphomas can be suppressed by NFκB activation,33 it is conceivable that NFκB activation in PTCL-EBV may also contribute towards the low GI but high aggressiveness of this tumor.34,35 Alternatively, there may be epigenetic deregulation driving oncogenesis in PTCL-EBV, which was not investigated in this study.36 It is now known that cancer cells interact with surrounding stromal and immune cells to form a pro-tumorigenic inflammatory tumor microenvironment.31 Compared to ENKTL and PTCL-NOS, PTCL-EBV is characterized by striking activation of immune-related pathways, in particular, upregulation of NFκB and its associated genes, BIRC3, NFKB1 and CD27, in both tumor and non-tumor cells. The upregulation of these three markers, either singly or colocalized with each other, suggests that there are different subsets of immune cells with varied combinations of BIRC3, p50 and CD27 expression in PTCL-EBV, and that the

Haematologica | 107 August 2022

1876

ARTICLE - Nodal EBV+ T/NK-cell lymphoma

C.M.M. Wai et al.

NF-κB signaling pathway likely plays a role in promoting an active inflammatory tumor microenvironment in PTCLEBV. NFκB plays a key role in linking inflammation to cancer31 and targets different immune cells to modulate inflammation, tumorigenesis and metastasis.37 The pronounced immune activation and NFκB upregulation in PTCL-EBV may augment the local inflammatory state and suppress cytotoxic T-lymphocyte effector function, resulting in an immunosuppressive tumor microenvironment that enhances immune evasion and aggressive behavior.38,39 While CD27 normally activates NFκB, promotes cell survival and enhances T- and B-cell receptor-mediated proliferative signals, dysregulation of CD27 signaling can result in T-cell exhaustion and dysfunction.40,41 Interestingly, we found that EBV-miRNA were predominantly downregulated in PTCL-EBV compared to ENKTL, which may further contribute to the inflammatory state with overexpression of immune-related genes.42 Based on our correlative data, we propose a model of PTCL-EBV pathogenesis involving NFκB activation and upregulation of PDL1, BIRC3 and CD27 (Figure 6E). Some of these pathways may serve as potential therapeutic targets for PTCL-EBV as drugs targeting JAK-STAT, NFκB, STAT3, IFNγ, PD1/PD-L1 are either approved by FDA for different cancers or are being evaluated in clinical trials for lymphomas.43-45 The upregulation of the immune checkpoint protein PDL1 in PTCL-EBV, inducible by IFNγ, may block the activation of the cytotoxic T-cell lymphocyte antitumor response, lead to T-cell exhaustion and promote immune evasion.23,46 In contrast to other lymphomas, this upregulation is unrelated to 9q24.1 gain.47 Interestingly, we identified a significant correlation between the gene expression of PD-L1 and its known transcriptional-regulators: IFNγ, IL6_JAK_STAT3 and NFκB, The overexpression of PD-L1 in PTCL-EBV, possibly a result of the upregulation of IFNγ, IL6_JAK_STAT3 and NFκB pathways,20,23,37 suggests that targeting the PD-1/PD-L1 axis43 may be a potentially effective therapeutic approach. Nevertheless, understanding the mechanisms causing PD-L1 upregulation, including structural variations in the 3’ untranslated region of PDL1,48 is essential for the design of more effective treatment strategies.23 Our data demonstrate that the majority of PTCL-EBV cases show a type 2 EBV latency program,15 similar to that of ENKTL and other EBV-associated T/NK lymphoproliferative diseases.49,50 A minority revealed a type 3 latency pattern. It remains unclear whether this may reflect underlying immunosuppression as it is known that, in a subset of patients, PTCL-EBV is associated with autoimmune conditions, viral infections or diabetes mellitus which may impair the host’s immune responses.1,51 The downregulation of EBV miRNA in PTCL-EBV compared to ENKTL suggests a difference in the EBV biology between the two diseases, although factors contributing to downregulation

of EBV miRNA have yet to be elucidated. Given that the predicted targets of the differentially expressed EBV miRNA are significantly enriched for immune-related pathways, it is tempting to postulate that the lower expression of EBV miRNA may contribute to the persistent expression of the many immune pathways in PTCL-EBV compared to ENKTL since miRNA are known to negatively regulate transcriptional gene expression.24 Nevertheless, the complex interplay between EBV miRNA, viral and cellular target genes in EBV-associated T/NK-cell lymphomas requires further investigation. In summary, PTCL-EBV is an aggressive lymphoma characterized by minimal GI, immune-related gene expression as well as activation of NFκB and its associated genes. While further studies are needed to corroborate our proposed model, these findings highlight the importance of the crosstalk between tumor and microenvironment, provide new insights hinting at the disease pathogenesis and offer potential new therapeutic targets for this aggressive disease. Disclosures No conflicts of interest to disclose. Contributions CMMW and SC performed research, analyzed data and wrote the manuscript. TP, SML, WZ, LCYL, ADJ, W-JC and CT contributed to data interpretation and analysis. S-NC, C-KL and SF performed histological experiments. T-HC, KHKB, SG, SL, FZ and FI contributed to bioinformatics analysis. Y-HH, SK, SN, ET, Y-HK, JDK, S-SC, RKHA-Y, S-YT, S-TL, LMP, SDM and CKO selected cases and acquired data. LQM and FO performed the mutational analysis. JJP analyzed and interpreted data and wrote the manuscript. SBN designed the project, analyzed data, and wrote and finalized the manuscript. Acknowledgments SBN is supported by the National Medical Research Council, Clinician Scientist Award, (CSAINV17nov016, WBS R‐179‐000‐063‐213), National Medical Research Council Open Fund Large Collaborative Grant, Singapore IYMPHoma translational study (SYMPHONY) (NMRC OFLCG18May-0028), and NUSMed Post-Doctoral Fellowship (PDF) (NUHSRO/2019/036/PDF/09). JJP is supported by the National Research Foundation Singapore and the Singapore Ministry of Education under its Research Centers of Excellence initiative. The computational work for this article was partially performed on resources of the National Supercomputing Center, Singapore (https://www.nscc.sg). Data-sharing statement De-identified data used in the preparation of this manuscript are available upon request.

Haematologica | 107 August 2022

1877

ARTICLE - Nodal EBV+ T/NK-cell lymphoma

C.M.M. Wai et al.

References 1. Ko YH, Chan JKC, Quintanilla-Martinez L. Virally associated Tcell and NK-cell neoplasms. In: Jaffe ES, Arber DA, Campo E, Harris NL, Quintanilla-Martinez L, editors. Haematopathology. Elsevier, Philadelphia 2017;565-598. 2. Kato S, Asano N, Miyata-Takata T, et al. T-cell receptor (TCR) phenotype of nodal Epstein-Barr virus (EBV)-positive cytotoxic T-cell lymphoma (CTL): a clinicopathologic study of 39 cases. Am J Surg Pathol. 2015;39(4):462-471. 3. Kato S, Takahashi E, Asano N, et al. Nodal cytotoxic molecule (CM)-positive Epstein-Barr virus (EBV)-associated peripheral T cell lymphoma (PTCL): a clinicopathological study of 26 cases. Histopathology. 2012;61(2):186-199. 4. Ng SB, Chung TH, Kato S, et al. Epstein-Barr virus-associated primary nodal T/NK-cell lymphoma shows a distinct molecular signature and copy number changes. Haematologica. 2018;103(2):278-287. 5. Pileri SA, Weisenburger DD, Sng I, et al. Peripheral T-cell lymphoma, NOS. In: Swerdlow SH, Campo E, Harris NL, et al., eds. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. International Agency for Research on Cancer; Lyon. 2017:403-407. 6. Pikor L, Thu K, Vucic E, Lam W. The detection and implication of genome instability in cancer. Cancer Metastasis Rev. 2013;32(3-4):341-352. 7. Ahmad SS, Ahmed K, Venkitaraman AR. Science in focus: genomic instability and its implications for clinical cancer care. Clin Oncol. 2018;30(12):751-755. 8. Oon ML, Lim JQ, Lee B, et al. T-cell lymphoma clonality by copy number variation analysis of T-cell receptor genes. Cancers (Basel). 2021;13(2):340. 9. Lee CS, Bhaduri A, Mah A, et al. Recurrent point mutations in the kinetochore gene KNSTRN in cutaneous squamous cell carcinoma. Nat Genet. 2014;46(10):1060-1062. 10. Mermel CH, Schumacher SE, Hill B, Meyerson ML, Beroukhim R, Getz G. GISTIC2.0 facilitates sensitive and confident localization of the targets of focal somatic copy-number alteration in human cancers. Genome Biol. 2011;12(4):R41. 11. Sinha S, Mitchell KA, Zingone A, et al. Higher prevalence of homologous recombination deficiency in tumors from African Americans versus European Americans. Nat Cancer. 2020;1(1):112-121. 12. Cheadle C, Vawter MP, Freed WJ, Becker KG. Analysis of microarray data using Z score transformation. J Mol Diagn. 2003;5(2):73-81. 13. Kato S, Yamashita D, Nakamura S. Nodal EBV+ cytotoxic T-cell lymphoma: a literature review based on the 2017 WHO classification. J Clin Exp Hematop. 2020;60(2):30-36. 14. Ha SY, Sung J, Ju H, et al. Epstein-Barr virus-positive nodal peripheral T cell lymphomas: clinicopathologic and gene expression profiling study. Pathol Res Pract. 2013;209(7):448-454. 15. Takahashi E, Asano N, Li C, et al. Nodal T/NK-cell lymphoma of nasal type: a clinicopathological study of six cases. Histopathology. 2008;52(5):585-596. 16. Gong Q, Wang C, Zhang W, et al. Assessment of T-cell receptor repertoire and clonal expansion in peripheral T-cell lymphoma using RNA-seq data. Sci Rep. 2017;7(1):11301. 17. Li Y, Roberts ND, Wala JA, et al. Patterns of somatic structural variation in human cancer genomes. Nature. 2020;578(7793):112-121. 18. Pitt JJ, Riester M, Zheng Y, et al. Characterization of Nigerian breast cancer reveals prevalent homologous recombination

deficiency and aggressive molecular features. Nat Commun. 2018;9(1):4181. 19. Szklarczyk D, Gable AL, Lyon D, et al. STRING v11: proteinprotein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res. 2019;47(D1):D607-D613. 20. Song TL, Nairismägi M-L, Laurensia Y, et al. Oncogenic activation of the STAT3 pathway drives PD-L1 expression in natural killer/Tcell lymphoma. Blood. 2018;132(11):1146-1158. 21. Garcia-Diaz A, Shin DS, Moreno BH, et al. Interferon receptor signaling pathways regulating PD-L1 and PD-L2 expression. Cell Rep. 2017;19(6):1189-1201. 22. Asgarova A, Asgarov K, Godet Y, et al. PD-L1 expression is regulated by both DNA methylation and NF-kB during EMT signaling in non-small cell lung carcinoma. Oncoimmunology. 2018;7(5):e1423170. 23. Shklovskaya E, Rizos H. Spatial and temporal changes in PD-L1 expression in cancer: the role of genetic drivers, tumor microenvironment and resistance to therapy. Int J Mol Sci. 2020;21(19):7139 24. Ebert MS, Sharp PA. Roles for microRNAs in conferring robustness to biological processes. Cell. 2012;149(3):515-524. 25. Ferrone CK, Blydt-Hansen M, Rauh MJ. Age-associated TET2 mutations: common drivers of myeloid dysfunction, cancer and cardiovascular disease. Int J Mol Sci. 2020;21(2):626 26. Riviere P, Goodman AM, Okamura R, et al. High tumor mutational burden correlates with longer survival in immunotherapy-naïve patients with diverse cancers. Mol Cancer Ther. 2020;19(10):2139-2145. 27. Liang WS, Vergilio J-A, Salhia B, et al. Comprehensive genomic profiling of Hodgkin lymphoma reveals recurrently mutated genes and Increased mutation burden. Oncologist. 2019;24(2):219-228. 28. Kamranvar SA, Gruhne B, Szeles A, Masucci MG. Epstein-Barr virus promotes genomic instability in Burkitt’s lymphoma. Oncogene. 2007;26(35):5115-5123. 29. Nagy M, Balázs M, Adám Z, et al. Genetic instability is associated with histological transformation of follicle center lymphoma. Leukemia. 2000;14(12):2142-2148. 30. Hieronymus H, Murali R, Tin A, et al. Tumor copy number alteration burden is a pan-cancer prognostic factor associated with recurrence and death. Elife. 2018;7:e37294. 31. Taniguchi K, Karin M. NF-κB, inflammation, immunity and cancer: coming of age. Nat Rev Immunol. 2018;18(5):309-324. 32. Yeo CQX, Alexander I, Lin Z, et al. p53 maintains genomic stability by preventing interference between transcription and replication. Cell Rep. 2016;15(1):132-146. 33. Ramachandiran S, Adon A, Guo X, et al. Chromosome instability in diffuse large B cell lymphomas is suppressed by activation of the noncanonical NF-κB pathway. Int J Cancer. 2015;136(10):2341-2351. 34. Crawley CD, Kang S, Bernal GM, et al. S-phase-dependent p50/NF-кB1 phosphorylation in response to ATR and replication stress acts to maintain genomic stability. Cell Cycle. 2015;14(4):566-576. 35. Wang J, Jacob NK, Ladner KJ, et al. RelA/p65 functions to maintain cellular senescence by regulating genomic stability and DNA repair. EMBO Rep. 2009;10(11):1272-1278. 36. Fennell KA, Bell CC, Dawson MA. Epigenetic therapies in acute myeloid leukemia: where to from here? Blood. 2019;134(22):1891-1901. 37. Betzler AC, Theodoraki M-N, Schuler PJ, et al. NF-κB and Its role

Haematologica | 107 August 2022

1878

ARTICLE - Nodal EBV+ T/NK-cell lymphoma

C.M.M. Wai et al.

in checkpoint control. Int J Mol Sci. 2020;21(11):3949 38. Xia Y, Shen S, Verma IM. NF-κB, an active player in human cancers. Cancer Immunol Res. 2014;2(9):823-830. 39. Lu C, Klement JD, Smith AD, et al. p50 suppresses cytotoxic T lymphocyte effector function to regulate tumor immune escape and response to immunotherapy. J Immunother Cancer. 2020;8(2):e001365. 40. Borst J, Hendriks J, Xiao Y. CD27 and CD70 in T cell and B cell activation. Curr Opin Immunol. 2005;17(3):275-281. 41. Riether C, Schürch C, Ochsenbein AF. Modulating CD27 signaling to treat cancer. Oncoimmunology. 2012;1(9):1604-1606. 42. Židovec Lepej S, Matulić M, Gršković P, Pavlica M, Radmanić L, Korać P. miRNAs: EBV mechanism for escaping host’s immune response and supporting tumorigenesis. Pathogens. 2020;9(5.):353 43. Wang L, Qin W, Huo Y-J, et al. Advances in targeted therapy for malignant lymphoma. Signal Transduct Target Ther. 2020;5(1):15. 44. Godwin P, Baird AM, Heavey S, Barr MP, O’Byrne KJ, Gately K. Targeting nuclear factor-kappa B to overcome resistance to chemotherapy. Front Oncol. 2013;3:120. 45. Spaccarelli N, Rook AH. The use of interferons in the treatment of cutaneous T-cell lymphoma. Dermatol Clin. 2015;33(4):731-745.

46. Chihara N, Madi A, Kondo T, et al. Induction and transcriptional regulation of the co-inhibitory gene module in T cells. Nature. 2018;558(7710):454-459. 47. Green MR, Monti S, Rodig SJ, et al. Integrative analysis reveals selective 9p24.1 amplification, increased PD-1 ligand expression, and further induction via JAK2 in nodular sclerosing Hodgkin lymphoma and primary mediastinal large B-cell lymphoma. Blood. 2010;116(17):3268-3277. 48. Lim JQ, Huang D, Tang T, et al. Whole-genome sequencing identifies responders to pembrolizumab in relapse/refractory natural-killer/T cell lymphoma. Leukemia. 2020;34(12):3413-3419. 49. Xu ZG, Iwatsuki K, Oyama N, et al. The latency pattern of Epstein-Barr virus infection and viral IL-10 expression in cutaneous natural killer/T-cell lymphomas. Br J Cancer. 2001;84(7):920-925. 50. Chiang AK, Tao Q, Srivastava G, Ho FC. Nasal NK- and T-cell lymphomas share the same type of Epstein-Barr virus latency as nasopharyngeal carcinoma and Hodgkin’s disease. Int J Cancer. 1996;68(3):285-290. 51. Yamashita D, Shimada K, Takata K, et al. Reappraisal of nodal Epstein-Barr virus-negative cytotoxic T-cell lymphoma: Identification of indolent CD5 diseases. Cancer Sci. 2018;109(8):2599-2610.

Haematologica | 107 August 2022

1879

ARTICLE - Non-Hodgkin Lymphoma

Programmed cell death ligand 1 expression in aggressive pediatric non-Hodgkin lymphomas: frequency, genetic mechanisms, and clinical significance Kevin E. Fisher,1,2 Lizmery S. Ferguson,2 Amy M. Coffey,1,2,3 Brian Y. Merritt,1 Jonathan L. Curry,4 Andrea N. Marcogliese,1,2 Angela M. Major,1 Kala Y. Kamdar,5 Dolores H. Lopez-Terrada1,2,5 and Choladda V. Curry1,2 Department of Pathology and Immunology, Baylor College of Medicine, Houston; Department of Pathology, Texas Children's Hospital, Houston; 3Clinical Pathology Associates, Austin; 4Department of Pathology and Translational Molecular Pathology, Division of Pathology and Laboratory Medicine, The University of Texas M.D. Anderson Cancer Center, Houston and 5Department of Pediatrics, Baylor College of Medicine, Texas Children’s Cancer and Hematology Centers, Houston, TX, USA 1

Correspondence: Kevin E. Fisher kevin.fisher@bcm.edu Choladda V. Curry ccurry@bcm.edu Received: November 12, 2021. Accepted: January 19, 2022. Prepublished: January 27, 2022. https://doi.org/10.3324/haematol.2021.280342 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

Abstract Programmed cell death 1 (PD-1) and programmed cell death ligand 1 (PD-L1) are immunomodulatory molecules overexpressed in lymphomas and are promising immunotherapy targets for hematologic malignancies. However, studies of PD-1/PD-L1 overexpression and their clinical significance in aggressive pediatric non-Hodgkin lymphomas (NHL) are limited. We assessed PD-1/PD-L1 overexpression using immunohistochemistry in 68 aggressive pediatric NHL: ALK-positive anaplastic large cell lymphoma (ALK+ ALCL, n=8), Burkitt lymphoma (BL, n=27), and large B-cell lymphoma (LBCL) de novo LBCL, n=22 and diffuse LBCL arising as monomorphic post-transplant lymphoproliferative disorder [PTLD-DLBCL], n=11. In LBCL, correlations between PD-L1 overexpression and Epstein-Barr virus (EBV) status, cell of origin, stage, nodal status, overall survival (OS), and event-free survival (EFS) were examined. The genetic mechanisms of PD-L1 overexpression were investigated using targeted next-generation sequencing (NGS) and cytogenetic data. All ALK+ ALCL samples, 50.0% of de novo LBCL (11/22), 72.7% of PTLD-DLBCL (8/11), and no BL overexpressed PD-L1. Overexpressed PD-L1 correlated with EBV positivity (P=0.033) in LBCL and lower EFS in de novo LBCL (P=0.017). NGS of select LBCL revealed distinct somatic mutations and an ultra-hypermutated PTLD-DLBCL. Most cases with 9p24.1 copy gains overexpressed PD-L1 although some cases had no discernible genetic drivers of PD-L1 overexpression. Overexpressed PD-L1 is common in pediatric LBCL, associated with EBV positivity and 9p24.1 gains, and may have prognostic significance in de novo LBCL. Furthermore, diverse molecular mechanisms for PD-L1 overexpression in aggressive pediatric NHL can occur. Thus, additional studies exploring the therapeutic and prognostic significance and molecular mechanisms of PD-L1 overexpression in aggressive pediatric NHL are warranted.

Introduction Immune checkpoint blockade has introduced a paradigm shift in cancer therapy for a variety of solid tumors and hematologic malignancies.1-3 Programmed cell death 1 (PD1) is a cell surface receptor in the B7/CD28 family of T-cell regulators that serves as an inhibitory co-receptor through binding of its ligands programmed cell death ligand 1 (PD-L1) and programmed cell death ligand 2 (PDL2). PD-L1 is broadly expressed across hematopoietic cells and human tissues, PD-1 is expressed on the cell surface

of B cells, T cells, and natural killer T cells, whereas PDL2 is only expressed on dendritic cells following cytokine induction.4,5 The PD-1/PD-L1/PD-L2 signaling axis promotes recognition of self and non-self antigens to maintain immune tolerance.5-7 However, tumor cells and tumor microenvironment cells can aberrantly overexpress PD-L1 (e.g., macrophages) and PD-1 (e.g., lymphocytes) to induce “T-cell exhaustion” and evade immune recognition and destruction.8-10 In adult hematologic malignancies, PD-L1 overexpression has been reported in classical Hodgkin lymphoma (cHL),

Haematologica | 107 August 2022

1880

ARTICLE - PD-L1 expression in aggressive pediatric lymphomas primary mediastinal large B-cell lymphoma (PMBCL), Tcell/histiocyte-rich large B-cell lymphoma (THRLBCL), ALK-positive anaplastic large cell lymphoma (ALK+ ALCL), subsets of diffuse large B-cell lymphoma (DLBCL), and follicular lymphomas.3,11,12 PD-1/PD-L1 overexpression is observed in approximately two thirds of adults with posttransplant lymphoproliferative disorder (PTLD) following solid organ transplantation as well as a subset of EpsteinBarr virus (EBV)-driven lymphomas (e.g., EBV-positive cHL). PD-1/PD-L1 expression in Burkitt lymphoma (BL) is typically negative.13 Notably, a vast majority of these data were established in adult studies and may not be representative of pediatric populations. Diverse molecular mechanisms drive PD-1/PD-L1 protein overexpression. For example, some tumors harbor high tumor mutation burden (TMB). This introduces numerous coding mutations that give rise to specific neoantigens that are presented to the cell surface and drive immune recognition and PD-1/PD-L1 overexpression.14,15 In adult LBCL, PD-L1 overexpression results from gene rearrangements involving the 9p24.1 gene locus which contains the CD274 (PD-L1), PDCD1LG2 (PD-L2), and JAK2 genes16 or 9p24.1 copy gains and amplifications similar to cHL.17,18 Structural variations of the CD274 gene 3’ untranslated region (3’-UTR) leading to increased transcript stabilization are also reported.19 In EBV-associated aggressive lymphomas, PD-L1 overexpression typically occurs via latent membrane protein-1 (LMP-1)-mediated activation of the CD274 JAK/STAT-dependent promoter or AP-1 associated enhancer activity.20,21 In ALK+ ALCL, PD-L1 overexpression is a result of STAT3 phosphorylation and IRF4/BATF3-mediated CD274 transcriptional upregulation.12,22 The anti-PD-1 immune checkpoint inhibitor pembrolizumab is a Food and Drug Admininistration (FDA)-approved therapeutic agent for relapsed/refractory adult and pediatric cHL. Another anti-PD-1 inhibitor, nivolumab, has shown promising therapeutic efficacy in early clinical trials for pediatric cHL and other PD-L1-expressing tumors.23,24 A factor contributing to the therapeutic efficacy of immune checkpoint inhibitors is the amount of target protein on the tumor tissue and/or microenvironment cells. Yet, recent evidence suggests that additional elements may also influence the use of PD-L1 expression as a predictive and prognostic marker for immune checkpoint inhibitor therapy. These may include the immunohistochemistry (IHC) methods employed, the time point at which the assessment of this dynamic biomarker is performed, TMB status,25,26 and other intrinsic biologic attributes such as DUSP22 rearrangements in ALK-negative ALCL,27 or the presence or absence of tumor infiltrating lymphocytes (TIL).28 Studies and clinical trials assessing PD-1/PD-L1 expression have focused primarily on adult lymphomas and pediatric cHL. The prognostic significance of PD-L1 expression in

K. Fisher et al.

adult DLBCL cohorts have shown conflicting results.29-31 Studies investigating PD-1/PD-L1 expression in aggressive pediatric non-Hodgkin lymphomas (NHL) such as ALCL, BL, and de novo pediatric DLBCL are limited. Thus, we investigated PD-1/PD-L1 expression in aggressive pediatric NHL including ALK+ ALCL, BL, a pediatric cohort LBCL (consisting of de novo LBCL and DLBCL arising as monomorphic PTLD [PTLD-DLBCL]). In the pediatric LBCL cohort, we reviewed available cytogenetic data to assess for rearrangements involving the 9p24.1 locus and performed targeted next-generation sequencing (NGS) on selected cases to detect DNA mutations, estimate TMB, and investigate copy-number changes of the 9p24.1 locus. The impact of PD-L1 status on clinicopathologic factors and treatment outcomes in the pediatric LBCL cohort was also investigated.

Methods Case selection and clinical data All authors conducted the study per Institutional-approved research protocols to Drs. Curry and Fisher. Sixtyeight aggressive pediatric NHL with sufficient tissue for IHC including ALK+ ALCL (n=8), BL (n=27), and LBCL (consisting of de novo LBCL, n=22; PTLD-DLBCL, n=11) were identified via a pathology database search from 1997-2012. The term de novo LBCL was used to describe pediatric LBCL cases in our cohort arising without previous transplants or previous low-grade B-cell lymphoma. Diagnoses and further subclassification of de novo LBCL were rendered using the revised 4th edition 2016 World Health Organization Classification of Tumors Hematopoietic and Lymphoid Tissues.32 Clinical and outcome data for the LBCL cohort were acquired from the electronic medical record. Staging was classified according to revised International Pediatric Non-Hodgkin Lymphoma Staging System (IPNHLSS).33 Immunohistochemistry Tissue microarrays were constructed for 68 cases. PD-L1 IHC was performed using the FDA-approved Dako PD-L1 IHC 28-8 pharmDx companion diagnostic assay per manufacturer established recommendations. PD-1 IHC staining (clone MRQ-22, Cell Marque) was performed at a 1:50 dilution on an automated Leica Bond III immunostainer. All PD-1/PD-L1 IHC results were independently reviewed and scored by two hematopathologists as previously described34 with a few modifications (Online Supplementary Table S1). Membranous PD-L1 staining of tumor cells and macrophages, and membranous PD-1 staining of tumor infiltrating lymphocytes (TIL) were scored for staining intensity: 0 (no staining), 1+ (weak or equivocal), 2+ (moderate), 3+ (strong). Percent positivity was recorded

Haematologica | 107 August 2022

1881

ARTICLE - PD-L1 expression in aggressive pediatric lymphomas as follows: the number of PD-L1-positive tumor cells out of total number of viable tumor cells (PD-L1-Tum); the area of PD-L1-positive macrophages out of total tumor area (PD-L1-Mac); number of PD-1-positive tumor infiltrating lymphocytes out of total number of tumor infiltrating lymphocytes (PD-1-TIL). The cut-offs for PD-1/PD-L1 positivity were 2+ or 3+ staining in ≥5% of tumor cells (PD-L1Tum) or ≥20% of macrophages (PD-L1-Mac) or TIL (PD-1-TIL). LBCL cases were classified based on cell of origin into germinal center B-cell (GCB) or non-GCB phenotypes according to the Hans IHC algorithm using CD10, BCL6, and MUM1 with a 30% cut-off.35 Epstein-Barr virus (EBV) encoded RNA (EBER) in situ hybridization and LMP1 (latent membrane protein 1) staining were used to determine EBV status. Targeted next-generation sequencing and cytogenetics Targeted NGS mutation analysis using a custom-designed 152 gene (1.25 megabase [Mb]) panel for pediatric hematologic malignancies was performed on 13 cases of aggressive pediatric NHL (5 de novo LBCL not otherwise specified, 2 primary-immunodeficiency associated LBCL, 5 PTLDDLBCL, and 1 BL) as previously reported.36 Variants were classified as established or potential clinical significance (Tier I/II) or variants of uncertain significance (Tier III).37 Gains of 9p24.1 were determined from copy number analysis of the targeted sequencing data. 9p24.1 rearrangements were assessed for seven LBCL cases using available cytogenetic data (Online Supplementary Table S2). Statistical analysis LBCL cases were classified as PD-L1 positive if PD-L1-Tum positive. Statistical analyses were performed using GraphPad Prism version 6.0.0. For the de novo LBCL and PTLDDLBCL cohorts, correlation between PD-L1-positivity and EBV status, cell of origin, stage, and nodal status was calculated using Fisher Exact test. Overall survival (OS) was defined as the time from diagnosis to death of any cause, and event-free survival (EFS) as the time to disease progression, relapse, or death of any cause, respectively. For survival analyses, Kaplan-Meier method with log-rank test was performed. Two-sided P-values <0.05 were considered statistically significant.

Results Diagnoses and clinicopathologic features A summary of the relevant clinicopathologic features, follow-up periods, and outcomes in the study cohort is provided in Table 1. Complete details are provided in the Online Supplementary Table S2. A total of 68 cases were studied; eight ALK+ALCL, 27 BL, and 33 LBCL (consisting of de novo LBCL, n=22; PTLD-DLBCL, n=11). Patients

K. Fisher et al.

ranged in age from 2.6 to 21.4 years (mean 10.6, median 10.8 years) with 45 males and 23 females. More than half of the cases were stage III/IV (n=41, 61.2%) and extranodal involvement was documented in a majority of cases (n=50, 73.5%). Of the patients with LBCL, the de novo LBCL cohort age ranged from 4.2 to 18.0 years (mean 11.6, median 11.3 years), with an equal number of males (n=11) and females (n=11). Diagnoses included DLBCL not otherwise specified (DLBCL-NOS, n=14), PMBCL (n=3), DLBCL with IGHα-MYC rearrangement (n=2), primary immunodeficiency-associated LBCL (PIA-LBCL) arising in Ataxia-Telangiectasia (n=2), and T-cell/histiocyte-rich LBCL (THRLBCL, n=1); one PIA-LBCL was subclassified further as THRLBCL. Two de novo LBCL (9.1%) were EBV+ and 15 (68.2%) were classified as germinal center B-cell (GCB) phenotype including 10 DLBCL-NOS (71.4%). De novo LBCL patients were treated according to various COG and POG protocols for pediatric B-cell NHL without Rituximab and three patients received bone marrow transplant (BMT) for relapsed/refractory disease (Online Supplementary Table S2). The follow-up period ranged from 4.2 to 253.9 months (mean 91.4, median 88.4 months). Of the patients with LBCL, the PTLD-DLBCL cohort age ranged from 3.4 to 21.4 years (mean 12.9, median 15.9 years). Disease occurred after lung (n=5), bone marrow (n=2), renal (n=2), liver (n=1), or heart (n=1) transplantation. Ten PTLD-DLBCL (90.9%) were EBV+ and two (22.2%) were classified as GCB. PTLD-DLBCL patients were treated according to various regimens that all included Rituximab (Online Supplementary Table S2). The follow up period ranged from 0.6 to 265.9 months (mean 61.1, median 50.0 months). PD-1/PD-L1 immunohistochemistry PD-L1 immunoreactivity in tumor cells (PD-L1-Tum) and macrophages (PD-L1-Mac), and PD-1 immunoreactivity in tumor infiltrating lymphocytes (PD-1-TIL) were assessed in the ALK+ ALCL, BL, and LBCL cohorts. A summary of the PD-1/PD-L1 IHC results is provided in Table 2. Representative IHC staining is shown in Figure 1. PD-L1-Mac and PD-1-TIL

No ALK+ ALCL cases showed PD-L1 staining in macrophages. In the BL cohort, PD-L1-Mac positivity was observed in five cases (range, 30-60%), but none of the PD-L1-Mac-positive BL demonstrated positive PD-L1-Tum staining. PD-L1-Mac positivity was seen in two DLBCLNOS (20% each) and one PIA-LBCL that was classified as TCHRBCL (70%) although neither PD-L1-Mac-positive DLBCL-NOS demonstrated positive PD-L1-Tum staining. PD-1-TIL positivity was observed in four cases (2 ALK+ ALCL, 1 BL, and 1 THRLBCL), and all cases occurred with concomitant PD-L1-Tum positivity.

Haematologica | 107 August 2022

1882

ARTICLE - PD-L1 expression in aggressive pediatric lymphomas

K. Fisher et al.

Table 1. Summary of study cohort (N=68). Site of involvement, N (%) Nodal Extranodal

Median age in years (range)

M | F, N

Stage III/IV, N

ALK+ ALCL (8)

12.7 (6.7-15.8)

6|2

5 (62.5%)

6 (75.0%)

2 (25.0%)

BL (27)

8.4 (2.6-18.5)

24 | 3

16 (59.3%)

2 (7.4%)

25 (92.6%)

LBCL (33) De novo (22) PTLD-DLBCL (11)

11.7 (3.4-21.4) 11.6 (4.2-18.0) 12.9 (3.4-21.4)

15 | 18 11 | 11 4|7

20 (62.5%)* 13 (61.9%)* 7 (63.6%)

10 (30.3%) 6 (27.3%) 4 (36.4%)

23 (69.7%) 16 (72.7%) 7 (36.6%)

Totals (68)

10.6 (2.6-21.4)

45 | 23

41 (61.2%)*

18 (26.5%)

50 (73.5%)

Diagnoses (count)

LBCL classification

Count (%)

GCB, N (%)

EBV+, N (%)

De novo LBCL DLBCL-NOS PMBCL THRLBCL DLBCL with IGHα-MYC PIA-LBCL^

22 (66.7%) 14 3 1 2 2

15/22 (68.2%) 10/14 (71.4%) 1 1 2 1

2/22 (9.1%) 2/14 (14.3%) 0 0 0 0

PTLD-DLBCL

11 (33.3%)

2/9 (22.2%)†

10/11 (90.9%)$

33 (100.0%)

17/31 (54.8%)†

12/33 (36.4%)$

All LBCL (33)

De novo (22)

PTLD-DLBCL (11)

80.7 (0.6-265.9)

91.4 (4.2-253.9)

61.1 (0.6-265.9)

21 (63.6%) 4 (12.1%) 2 (6.1%) 1 (3.0%) 4 (12.1%) 1 (3.0%)

15 (68.2%) 3 (13.6%) 2 (9.1%) 1 (4.5%) 1 (4.5%) -

6 (54.5%) 1 (9.1%) 3 (27.3%) 1 (9.1%)

Totals LBCL events Months follow-up (range) Outcome, N (% cohort) NED REL, NED REL, DOD REF, DOD DOD DOC (transplant)

*Staging was scored using International Pediatric Non-Hodgkin Lymphoma Staging System (IPNHLSS) criteria32 and data were unavailable for 1 de novo PIA-LBCL. ^Both PIA-LBCL occurred in Ataxia-Telangiectasia patients. One PIA-LBCL was further subclassified as THRLBCL. †Cell of origin (GCB vs. non-GCB) classification for 2 cases of PTLD-DLBCL was unsuccessful due to insufficient tissue. $One EBV+ PTLD-DLBCL was positive for EBER (Epstein-Barr virus encoded RNA) only and negative for LMP-1 (latent membrane protein 1). ALK+ ALCL: ALK-positive anaplastic large cell lymphoma; BL: Burkitt lymphoma; DOC: died of complications; DOD: died of disease; DLBCL: diffuse large B-cell lymphoma; EBV+: Epstein-Barr virus positive; F: female; GCB: germinal center B-cell phenotype; LBCL: large B-cell lymphoma; M: male; NED: no evidence of disease; NOS: not otherwise specified; PIA: primary immunodeficiency-associated; PTLD: post-transplant lymphoproliferative disorder; PMBCL: primary mediastinal large B-cell lymphoma; REF: refractory; REL: relapsed; THRLBCL: T-cell/histocyte-rich large B-cell lymphoma.

Table 2. Summary of PD-1/PD-L1 immunohistochemistry results. Diagnoses (count) ALK+ ALCL (8) BL (27) LBCL (33) De novo LBCL (22) DLBCL-NOS (14) PMBCL (3) THRLBCL (1) DLBCL with IGHα-MYC (2) PIA-LBCL (2) PTLD-DLBCL (11)

PD-L1-Tum ≥5%* N (%)

Range %

PD-L1-Mac ≥20%* N (%)

Range %

PD-1-TIL ≥20%* N (%)

Range %

8 (100.0%)

30-100%

0 0.0%

2 25.0%

20%

0 (0.0%)

5 18.5%

30-60%

0 0.0%

19 (57.6%)

30-100%

3 9.1%

20-70%

1 3.0%

20%

11 (50.0%) 4/14 (28.6%) 3 1 1 2

50-60% 50-90% 100% 30% 30-100%

3 13.6% 2 0 0 0 1

20% . . . 70%

1 4.5% 0 0 1 0 0

. . 20% . .

8 (72.7%)

40-100%

0 0.0%

0 0.0%^

*Cases were considered PD-L1-Tum positive if PD-L1 membranous expression on tumor cells was ≥5%, PD-L1-Mac positive if PD-L1 membranous expression on macrophages was ≥20%, and PD-1-TIL positive if ≥20% tumor infiltrating lymphocytes showed membranous PD-1 staining (Online Supplementary Table S1). ^Scoring PD-1 expression on tumor infiltrating lymphocytes (PD-1-TIL) was unsuccessful for one PTLD-DLBCL case due to insufficient tissue. ALK+ ALCL: ALK-positive anaplastic large cell lymphoma; BL: Burkitt lymphoma; DLBCL: diffuse large B-cell lymphoma; LBCL: large B-cell lymphoma; Mac: macrophages; NOS: not otherwise specified; PIA: primary immunodeficiency-associated; PDL1-Mac: PD-L1 expression on macrophages; PMBCL: primary mediastinal large B-cell lymphoma; PTLD: post-transplant lymphoproliferative disorder; THRLBCL: T-cell/histocyte-rich large B-cell lymphoma; TIL: tumor-infiltrating lymphocyte. Haematologica | 107 August 2022

1883

ARTICLE - PD-L1 expression in aggressive pediatric lymphomas PD-L1-Tum

As expected, all eight (100.0%) of the ALK+ ALCL cases showed positive PD-L1-Tum staining (range, 30-100%) and none of the 27 BL cases (0.0%). Overall, PD-L1-Tum positivity ranging from 30-100% was detected in 19 LBCL (57.6%) and 11 de novo LBCL (50.0%) including four DLBCLNOS (28.6%). Eight PTLD-DLBCL (72.7%) showed PD-L1Tum-positive staining ranging from 40-100%. In our cohort, none of the PD-L1-Tum-positive cases demonstrated less than 30% membranous positivity despite a 5% threshold for positivity. Three de novo LBCL cases showed a cytoplasmic PD-L1 staining pattern that was

K. Fisher et al.

considered PD-L1-negative, and three LBCL showed essentially similar PD-L1 staining intensity and percent positivity on initial and relapsed/refractory lesions (data not shown). PD-L1 expression in pediatric large B-cell lymphomas LBCL cases were recorded as positive for PD-L1 expression only if PD-L1-Tum positivity was observed and subsequent statistical analyses focused on PD-L1-Tum staining status only (Table 3). PD-L1-Tum positivity significantly correlated with EBV+ LBCL (P=0.033). There were no statistically significant differences between PD-

Figure 1. PD-L1 immunohistochemical stains in representative case of pediatric aggressive non-Hodgkin lymphomas. (A and B) ALK-positive anaplastic large cell lymphoma (ALK+ ALCL) showed tumor cells admixed with tumor microenvironment cells (lymphocytes and macrophages). Nearly all tumor cells show PD-L1 expression with moderate to strong staining intensity (2+ to 3+). (C and D) Burkitt lymphoma showed predominantly intermediate-sized lymphoma cells admixed with macrophages. PD-L1 showed no staining. (E and F) Epstein-Barr virus-postive diffuse large B-cell lymphoma not otherwise specified (EBV+ DLBCL-NOS) showed large lymphoma cells and PD-L1 overexpression in 50% of tumor cells with moderate staining intensity (2+). (A to F magnification 400x). Haematologica | 107 August 2022

1884

ARTICLE - PD-L1 expression in aggressive pediatric lymphomas L1-Tum staining and cell of origin, stage, or site of involvement in either the total cohort or within the de novo or PTLD-DLBCL subsets. The correlation between PD-L1Tum and OS and EFS was also examined (Figure 2). PDL1-Tum did not correlate with OS in the total LBCL cohort or within the de novo or PTLD-DLBCL subsets. Likewise, PD-L1-Tum was not significantly associated with EFS in the total LBCL cohort or in the PTLD-DLBCL subset. However, PD-L1 overexpression significantly correlated with lower EFS in the de novo LBCL subset (P=0.017). Neither OS nor EFS could be calculated in the DLBCLNOS subset due to small sample size (n=14) and number of events.

K. Fisher et al.

Targeted next-generation sequencing In order to investigate the molecular mechanisms of PDL1 positivity in a subset of aggressive pediatric NHL, targeted NGS to assess for 9p24.1 copy number and TMB was performed on 12 LBCL (7 de novo LBCL [5 DLBCL-NOS, 2 PIA-LBCL] and 5 PTLD-DLBCL) and one BL. Cytogenetic data to assess for the presence of 9p24.1 gene rearrangements were reviewed for seven LBCL (5 de novo LBCL [4 DLBCL-NOS, 1 DLBCL with IGH@-MYC] and 2 PTLDDLBCL). Characteristic somatic ARID1A and TP53 and germline ATM mutations were detected in the BL38 and PIA-LBCL samples, respectively. One PTLD-DLBCL harbored an ARID1A mutation, another PTLD-DLBCL harbored

Table 3. Summary of PD-L1 statistical analyses for correlation study in large B-cell lymphoma cohort. LBCL cohort

Count

EBV+ / EBV-

GCB / non-GCB*

Stage I-II / III*

Nodal / Extranodal

De novo LBCL cohort PD-L1-Tum POS PD-L1-Tum NEG P-value

11 11 -

2/9 0 / 11 0.476

7/4 8/3 1.000

3/7 5/6 0.659

2/9 4/7 0.635

PTLD-DLBCL cohort PD-L1-Tum POS PD-L1-Tum NEG P-value

8 3 -

8/0 2/1 0.273

1/6 1/1 0.417

4/4 0/3 0.236

Total cohort PD-L1-Tum POS PD-L1-Tum NEG P-value

19 14 -

10 / 9 2 / 12 0.033

8 / 10 9/4 0.275

7 / 11 5/9 1.000

6 / 13 4 / 10 1.000

*Cell of origin classification for 2 cases of PTLD-DLBCL was unsuccessful due to insufficient tissue. Staging was unavailable for 1 PIA-LBCL. LBCL: large B-cell lymphoma; EBV: Epstein-Barr virus; GCB: germinal center B-cell phenotype; PD-L1-Tum: PD-L1 expression on tumor cells; PTLD-DLBCL: diffuse large B-cell lymphoma arising as post-transplant lymphoproliferative disorder; POS: positive; NEG: negative.

Figure 2. Kaplan-Meier survival curves. Overall survival (OS) and event-free survival (EFS) curves for PD-L1 staining in the total large B-cell lymphoma (LBCL), de novo LBCL, and diffuse LBCL arising as monomorphic post-transplant lymphoproliferative disorder (PTLD-DLBCL) cohorts. Haematologica | 107 August 2022

1885

ARTICLE - PD-L1 expression in aggressive pediatric lymphomas DDX3X, ID3, and two TP53 co-mutations, and two DLBCLNOS cases showed mutations in CCND3 alone, and BCOR, CCND3, and U2AF1 co-mutations, respectively. The results are summarized in Table 4 and the Online Supplementary Table S3. PD-L1-positive LBCL showed EBV positivity (n=2), copy gains of the 9p24.1 locus (n=1), or both (n=1). One PD-L1positive PIA-LBCL had no discernible mechanism for PDL1 overexpression and no 9p24.1 rearrangements were observed in any case. EBV positivity was observed in one PTLD-DLBCL and DLBCL-NOS without PD-L1 overexpression. Interestingly, one PTLD-DLBCL (LBCL-29) harbored an estimated TMB of 180 mutations per megabase (Mb), which is considered an “ultra-hypermutated” tumor (>100 mutations/Mb).39 Pathogenic mutations in ASXL1, CDKN2A, CREBBP, FBXW7 (n=2), KMT2C, KMT2D, KRAS, TP53 (n=2), and ZFHX3 as well as 104 coding or splicing variants of uncertain significance were detected in this tumor (Online Supplementary Table S3). Notably, this case

K. Fisher et al.

did not show PD-L1-Tum positivity.

Discussion Pembrolizumab is an FDA-approved PD-1 inhibitor for refractory childhood cHL. Likewise, the anti-PD-1 immune checkpoint inhibitor nivolumab has shown promising therapeutic efficacy in early phase clinical trials for pediatric lymphomas that overexpress PD-L1.23,24 Thus, defining PD-L1 staining patterns and cut-off positivity in aggressive pediatric NHLs using the 28-8 clone antibody, the complementary IHC diagnostic for nivolumab, is of potential clinical significance. In this study, we defined positive PD-L1 staining as membranous staining of 5% of total viable tumor cells (PD-L1Tum). We observed positive PD-L1-Tum staining in all eight ALK+ ALCL and none of the 27 BL consistent with previous reports.20,40 PD-L1-Tum positivity was observed in a ma-

Table 4. Summary of next-generation sequencing and 9p24.1 locus results. Diagnosis

PD-L1

EBV+

GCB versus non-GCB

9p24.1 rearrangements

9p24.1 gains

Tier I/II Mutations

VAF

Neg

N/A

Not Detected

ARID1A p.Y1679X TP53 p.R175H

0.32 0.75

LBCL-14*

PIA-LBCL^

Pos

Neg

non-GCB

N/A

Detected

ATM p.S1924X$

0.52

LBCL-9*

PIA-LBCL

Pos

Neg

non-GCB

Not Detected

ATM p.W2638X$ ATM p.Q1970X$

0.43 0.52

LBCL-27

PTLD-DLBCL

Pos

non-GCB

N/A

Detected

ARID1A p.Q172X

0.40

LBCL-30

PTLD-DLBCL

Pos

non-GCB

N/A

Not Detected

None Detected

LBCL-32

PTLD-DLBCL

Pos

non-GCB

Not Detected

None Detected

Sample BL-21

DDX3X c.1171-2A>G ID3 p.L70Q Not Detected TP53 p.I50fs TP53 p.R175H

0.71 0.43 0.26 0.35

LBCL-26

PTLD-DLBCL

Neg

Pos

N/A

LBCL-17*

DLBCL-NOS

Neg

non-GCB

Not Detected

None Detected

LBCL-16*

DLBCL-NOS

Neg

GCB

Not Detected

CCND3 p.R271fs

0.17

LBCL-12*

DLBCL-NOS

Neg

GCB

N/A

Not Detected

None Detected

LBCL-13*

DLBCL-NOS

Neg

GCB

N/A

Not Detected

BCOR p.N1459S CCND3 p.R271fs U2AF1 p.S34F

0.89 0.41 0.45

LBCL-3*

DLBCL-NOS

Neg

Pos

GCB

N/A

Not Detected

None Detected

LBCL-29† PTLD-DLBCL

Neg

GCB

Not Detected

11 total, high TMB†

0.43

*Denotes de novo LBCL. ^Further subclassified as T-cell/histiocyte-rich large B-cell lymphoma: †Ultra-hypermutated tumor with ~180 mutations per megabase (Mb). A complete list of mutations is provided in the Online Supplementary Table S3. $Known germline. BL: Burkitt lymphoma; DLBCL: diffuse large B-cell lymphoma; GCB: germinal center B-cell phenotype; LBCL: large B-cell lymphoma; N/A: not available; Neg: negative; Pos: positive; PTLD: post-transplant lymphoproliferative disorder; TMB: tumor mutation burden; VAF: variant allele fraction.

Haematologica | 107 August 2022

1886

ARTICLE - PD-L1 expression in aggressive pediatric lymphomas jority of LBCL including 50.0% of de novo LBCL. The percentage of PD-L1 positive DLBCL-NOS cases in our pediatric cohort (28.6%) was slightly higher than two recent adult studies in which positive PD-L1 staining was reported in approximately 10-15% of DLBCL-NOS cases.29,31 The frequency of PD-L1 positive pediatric PTLD-DLBCL cases (72.7%) was similar to other pediatric and adult PTLD-DLBCL cohorts that reported approximately 60-80% positivity.13,34,41 Despite defining 5% as the threshold for PD-L1-positivity, all PD-L1-positive cases in our cohort showed at least 30% positive membranous staining. Kiyasu et al.29 employed a similar cut-off in adult DLBCL and reported PDL1 positivity ranges from 30-100% while observing a marked decrease in staining below the 30% threshold. Additionally, we detected minimal PD-L1-Mac staining or PD-1-TIL staining consistent with other pediatric studies that reported low levels of PD-1-positivity.34 Together, the data support a potential role for PD-L1 testing in pediatric aggressive NHL, including ALK+ ALCL, de novo LBCL, and PTLD-DLCBL. Furthermore, our data suggest that a positivity threshold of 30% is suitable when assessing pediatric aggressive NHL for PD-L1 staining of tumor cells. We correlated LBCL PD-L1-Tum staining with multiple clinicopathologic features although the small number of cases precluded comprehensive statistical analyses. There was no statistically significant correlation between stage, nodal status, cell of origin, or OS in the total LBCL cohort, or when separated into de novo LBCL and PTLDDLBCL. PD-L1-Tum staining correlated with EBV positivity in our study for the entire LBCL cohort, but not with PTLD-DLBCL cohort alone. Although PTLD-DLBCL comprised a large portion of the EBV+ LBCL samples in our cohort, the grouping of EBV+ de novo LBCL and PTLDLBCL is not necessarily representative of clinical practice. Nonetheless, correlation between EBV positivity and PDL1 overexpression in pediatric LBCL was previously reported; Nicolae et al.42 reported a strong correlation in patients younger than 45 years of age between EBV+ and PD-L1 expression in LBCL, particularly in THRBCL and nonGCB DLBCL, which was recapitulated here. Furthermore, LBCL classified as intrinsic (oncogenic) induction of TME correlated with EBV positivity suggesting that EBV is a potent regulator of PD-L1 in pediatric LBCL. In two previous adult LBCL studies in which a higher threshold for tumor cell positivity was utilized (30%), PDL1 positivity correlated with inferior OS.29,31 Using a de facto cut-off of 30% in our study, PD-L1 staining correlated with reduced EFS in de novo LBCL, although no statistical significance in OS was observed. However, studies investigating the correlation between PD-L1 positivity and survival in adult DLBCL are confounded by multiple variables such as the antibody clone employed, IHC scoring methodology adopted, and/or non-uniform

K. Fisher et al.

positivity thresholds.29-31,43 We acknowledge that conclusions regarding OS and EFS are limited due to the retrospective nature of our study, the small numbers of LBCL subtypes, and non-uniform treatment regimens, particularly pediatric intensive treatment regimens without rituximab for de novo LBCL treatment. Also, the grouping of diagnostic entities that have distinct clinicopathologic and therapeutic considerations (e.g., PMBCL) restricts the generalization of the statistical findings. Nonetheless, the negative correlation between PD-L1 and EFS in pediatric de novo LBCL was observed, and additional studies are warranted to establish PD-L1-Tum as a negative prognostic marker in clinical pediatric practice. CD274 gene overexpression and increased PD-L1 protein translation can occur through a variety of mechanisms such as increased CD274 transcription via promoter activation or CD274 copy gains, amplifications, and gene rearrangements.17,18,20,21 In our study, PD-L1 staining was observed most frequently in cases with EBV positivity and 9p24.1 copy gains. However, in two EBV+ cases, one de novo LBCL and one PTLD-DLBCL, no PD-L1 staining was observed. This is not unexpected, as EBV-associated CD274 gene regulation requires complex interplay between transcriptional elements, microRNA, and post-transcriptional modifications.44,45 Also, in one PD-L1-positive PIALBCL case no 9p24.1 copy gains or rearrangements were observed. Thus, it is possible that this case represented a false-negative 9p24.1 result; comprehensive molecular assessments of CD274 3’-UTR structural variations and gene rearrangements46 not detectable by karyotype analysis were not performed. Interestingly, we encountered an EBV-negative PTLDDLBCL with no PD-L1 positivity with an estimated TMB of 180 mutations/Mb consistent with an “ultra-hypermutated” tumor (>100 mutations/Mb). The high TMB was not attributable to somatic hypermutation; all mutations were detected in coding exons and not in immunoglobulin heavy chain and light chain variable region genes.47 The accuracy of TMB assessment is proportional to the number of bases sequenced so smaller targeted panels can lead to imprecise TMB scores. However, recent studies suggest that highly accurate TMB assessments can be obtained with targeted panels larger than 1.0 Mb.48,49 In this study, we utilized a targeted sequencing panel that captures and sequences approximately 1.25 million individual bases (1.25 Mb) so we conclude that our TMB estimates are reasonably accurate. Hypermutation is rare in pediatric tumors; however when detected, it often occurs in the setting of congenital mismatch repair deficiency (CMMRD) syndromes and/or acquisition of somatic POLE/POLD1 mutations39 but genetic testing for CMMRD or POLE/POLD1 somatic mutation testing was not performed in this case. Despite the lack of PD-L1 positivity in this case, emerging data suggest that either PD-L1

Haematologica | 107 August 2022

1887

ARTICLE - PD-L1 expression in aggressive pediatric lymphomas staining or high TMB should be considered positive biomarkers for response to immune checkpoint inhibitor therapy.50 Collectively, our data indicate that EBV and 9p24.1 copy gains are key regulators of PD-L1 overexpression in a subset of pediatric LBCL, particularly in PTLD-DLBCL, which is consistent with other studies in adult patients.40 Assessment of TMB may identify some patients that may benefit from immune checkpoint inhibitor therapy in the absence of PD-L1-Tum staining. The genomic landscape of pediatric LBCL has not been extensively studied, but in adults, mutations in DLCBL tend to cluster in pathways that regulate cell signaling and growth, B-cell development, and nucleic acid transcription and translation.51 Our sequencing analysis revealed similar somatic pathogenic mutations regulators of the cell cycle, transcription, and DNA repair (CCND3, DDX3X, ID3, TP53), and mRNA splicing (U2AF1), but also identified mutations in epigenetic modifiers (ARID1A, BCOR). There did not appear to be significant differences in the types of mutations encountered in GCB versus nonGCB cases or differences in mutational profiles between de novo LBCL or PTLD-DLBCL. However, only 152 genes were sequenced in our study and the panel did not include several genes that were recently reported to cluster in discreet LBCL subtypes (e.g., B2M, CD79B, CIITA, GNAI2, and SGK1).52, 53 Thus, the limited gene content and small number of cases hinders a definitive differential assessment of mutational signatures.

K. Fisher et al.

In conclusion, as one of the largest pediatric series to date, we showed that 30% cut-off for membranous PDL1 staining in the tumor cells (PD-L1-Tum) is an appropriate threshold to assess for PD-L1 positivity and correlated with EBV positivity in LCBL and lower EFS in de novo LBCL. Also, EBV and copy gains of 9p24.1 appear to regulate PD-L1 overexpression in a subset of cases and rare cases of pediatric NHL may demonstrate high TMB highlighting a potential new biomarker in this disease. Additional studies exploring the therapeutic and prognostic significance, molecular mechanisms, and host environments that affect PD-1/PD-L1 staining in aggressive pediatric NHL are warranted. Disclosures No conflicts of interest to disclose. Contributions KEF and CVC designed the study, wrote, edited the manuscript, and supervised the project. LSF, AMC, AMM, and BYM performed experiments and collected and analyzed data. JLC, ANM, DLT, and KYK assisted in data collection and provided clinical expertise. Data sharing statement Data sharing is subject to Institutional policies governing clinical research data.

References 1. Weber J. Immune checkpoint proteins: a new therapeutic paradigm for cancer -preclinical background: CTLA-4 and PD-1 blockade. Semin Oncol. 2010;37(5):430-439. 2. Xia Y, Medeiros LJ, Young KH. Immune checkpoint blockade: releasing the brake towards hematological malignancies. Blood Rev. 2016;30(3):189-200. 3. Xie W, Medeiros LJ, Li S, Yin CC, Khoury JD, Xu J. PD-1/PD-L1 pathway and its blockade in patients with classic Hodgkin lymphoma and non-Hodgkin large-cell lymphomas. Curr Hematol Malig Rep. 2020;15(4):372-381. 4. Jin HT, Ahmed R, Okazaki T. Role of PD-1 in regulating T-cell immunity. Curr Top Microbiol Immunol. 2011;350:17-37. 5. Keir ME, Butte MJ, Freeman GJ, Sharpe AH. PD-1 and its ligands in tolerance and immunity. Annu Rev Immunol. 2008;26:677-704. 6. Pardoll DM. The blockade of immune checkpoints in cancer immunotherapy. Nat Rev Cancer. 2012;12(4):252-264. 7. Topalian SL, Drake CG, Pardoll DM. Immune checkpoint blockade: a common denominator approach to cancer therapy. Cancer Cell. 2015;27(4):450-461. 8. Bryan LJ, Gordon LI. Blocking tumor escape in hematologic malignancies: the anti-PD-1 strategy. Blood Rev. 2015;29(1):25-32. 9. Afreen S, Dermime S. The immunoinhibitory B7-H1 molecule as

a potential target in cancer: killing many birds with one stone. Hematol Oncol Stem Cell Ther. 2014;7(1):1-17. 10. Sznol M, Chen L. Antagonist antibodies to PD-1 and B7-H1 (PDL1) in the treatment of advanced human cancer. Clin Cancer Res. 2013;19(5):1021-1034. 11. Xia Y, Jeffrey Medeiros L, Young KH. Signaling pathway and dysregulation of PD1 and its ligands in lymphoid malignancies. Biochim Biophys Acta. 2016;1865(1):58-71. 12. Shen J, Li S, Medeiros LJ, et al. PD-L1 expression is associated with ALK positivity and STAT3 activation, but not outcome in patients with systemic anaplastic large cell lymphoma. Mod Pathol. 2020;33(3):324-333. 13. Kinch A, Sundström C, Baecklund E, Backlin C, Molin D, Enblad G. Expression of PD-1, PD-L1, and PD-L2 in posttransplant lymphoproliferative disorder after solid organ transplantation. Leuk Lymphoma. 2019;60(2):376-384. 14. Chan TA, Yarchoan M, Jaffee E, et al. Development of tumor mutation burden as an immunotherapy biomarker: utility for the oncology clinic. Ann Oncol. 2019;30(1):44-56. 15. Rizvi NA, Hellmann MD, Snyder A, et al. Cancer immunology. Mutational landscape determines sensitivity to PD-1 blockade in non-small cell lung cancer. Science. 2015;348(6230):124-128. 16. Georgiou K, Chen L, Berglund M, et al. Genetic basis of PD-L1 overexpression in diffuse large B-cell lymphomas. Blood.

Haematologica | 107 August 2022

1888

ARTICLE - PD-L1 expression in aggressive pediatric lymphomas 2016;127(24):3026-3034. 17. Green MR, Monti S, Rodig SJ, et al. Integrative analysis reveals selective 9p24.1 amplification, increased PD-1 ligand expression, and further induction via JAK2 in nodular sclerosing Hodgkin lymphoma and primary mediastinal large Bcell lymphoma. Blood. 2010;116(17):3268-3277. 18. Roemer MG, Advani RH, Ligon AH, et al. PD-L1 and PD-L2 genetic alterations define classical Hodgkin lymphoma and predict outcome. J Clin Oncol. 2016;34(23):2690-2697. 19. Kataoka K, Shiraishi Y, Takeda Y, et al. Aberrant PD-L1 expression through 3'-UTR disruption in multiple cancers. Nature. 2016;534(7607):402-406. 20. Chen BJ, Chapuy B, Ouyang J, et al. PD-L1 expression is characteristic of a subset of aggressive B-cell lymphomas and virus-associated malignancies. Clin Cancer Res. 2013;19(13):3462-3473. 21. Green MR, Rodig S, Juszczynski P, et al. Constitutive AP-1 activity and EBV infection induce PD-L1 in Hodgkin lymphomas and posttransplant lymphoproliferative disorders: implications for targeted therapy. Clin Cancer Res. 2012;18(6):1611-1618. 22. Atsaves V, Tsesmetzis N, Chioureas D, et al. PD-L1 is commonly expressed and transcriptionally regulated by STAT3 and MYC in ALK-negative anaplastic large-cell lymphoma. Leukemia. 2017;31(7):1633-1637. 23. Davis KL, Fox E, Merchant MS, et al. Nivolumab in children and young adults with relapsed or refractory solid tumours or lymphoma (ADVL1412): a multicentre, open-label, single-arm, phase 1-2 trial. Lancet Oncol. 2020;21(4):541-550. 24. Geoerger B, Kang HJ, Yalon-Oren M, et al. Pembrolizumab in paediatric patients with advanced melanoma or a PD-L1positive, advanced, relapsed, or refractory solid tumour or lymphoma (KEYNOTE-051): interim analysis of an open-label, single-arm, phase 1-2 trial. Lancet Oncol. 2020;21(1):121-133. 25. Hamanishi J, Mandai M, Matsumura N, Abiko K, Baba T, Konishi I. PD-1/PD-L1 blockade in cancer treatment: perspectives and issues. Int J Clin Oncol. 2016;21(3):462-473. 26. Davis KL, Agarwal AM, Verma AR. Checkpoint inhibition in pediatric hematologic malignancies. Pediatr Hematol Oncol. 2017;34(6-7):379-394. 27. Luchtel RA, Dasari S, Oishi N, et al. Molecular profiling reveals immunogenic cues in anaplastic large cell lymphomas with DUSP22 rearrangements. Blood. 2018;132(13):1386-1398. 28. Teng MW, Ngiow SF, Ribas A, Smyth MJ. Classifying cancers based on T-cell infiltration and PD-L1. Cancer Res. 2015;75(11):2139-2145. 29. Kiyasu J, Miyoshi H, Hirata A, et al. Expression of programmed cell death ligand 1 is associated with poor overall survival in patients with diffuse large B-cell lymphoma. Blood. 2015;126(19):2193-2201. 30. Kwon D, Kim S, Kim PJ, et al. Clinicopathological analysis of programmed cell death 1 and programmed cell death ligand 1 expression in the tumour microenvironments of diffuse large B cell lymphomas. Histopathology. 2016;68(7):1079-1089. 31. Xing W, Dresser K, Zhang R, et al. PD-L1 expression in EBVnegative diffuse large B-cell lymphoma: clinicopathologic features and prognostic implications. Oncotarget. 2016;7(37):59976-59986. 32. Swerdlow SH, Campo E, Pileri SA, et al. The 2016 revision of the World Health Organization classification of lymphoid neoplasms. Blood. 2016;127(20):2375-2390. 33. Angelo Rosolen A, Perkins SK, Pinkerton CR, et al. Revised International Pediatric Non-Hodgkin Lymphoma Staging System. J Clin Oncol. 2015:33(18):2112-2118. 34. Schiefer AI, Salzer E, Füreder A, et al. PD-L1 and PD1 expression

K. Fisher et al.

in post-transplantation lymphoproliferative disease (PTLD) of childhood and adolescence: an inter- and intra-individual descriptive study covering the whole spectrum of PTLD categories. Cancer Med. 2019;8(10):4656-4668. 35. Hans CP, Weisenburger DD, Greiner TC, et al. Confirmation of the molecular classification of diffuse large B-cell lymphoma by immunohistochemistry using a tissue microarray. Blood. 2004;103(1):275-282. 36. Zhou T, Bloomquist MS, Ferguson LS, et al. Pediatric myeloid sarcoma: a single institution clinicopathologic and molecular analysis. Pediatr Hematol Oncol. 2020;37(1):76-89. 37. Li MM, Datto M, Duncavage EJ, et al. Standards and guidelines for the interpretation and reporting of sequence variants in cancer: a joint consensus recommendation of the Association for Molecular Pathology, American Society of Clinical Oncology, and College of American Pathologists. J Mol Diagn. 2017;19(1):4-23. 38. Giulino-Roth L, Wang K, MacDonald TY, et al. Targeted genomic sequencing of pediatric Burkitt lymphoma identifies recurrent alterations in antiapoptotic and chromatin-remodeling genes. Blood. 2012;120(26):5181-5184. 39. Shlien A, Campbell BB, de Borja R, et al. Combined hereditary and somatic mutations of replication error repair genes result in rapid onset of ultra-hypermutated cancers. Nat Genet. 2015;47(3):257-262. 40. Marzec M, Zhang Q, Goradia A, et al. Oncogenic kinase NPM/ALK induces through STAT3 expression of immunosuppressive protein CD274 (PD-L1, B7-H1). Proc Natl Acad Sci U S A. 2008;105(52):20852-20857. 41. Veloza L, Teixido C, Castrejon N, et al. Clinicopathological evaluation of the programmed cell death 1 (PD1)/programmed cell death-ligand 1 (PD-L1) axis in post-transplant lymphoproliferative disorders: association with Epstein-Barr virus, PD-L1 copy number alterations, and outcome. Histopathology. 2019;75(6):799-812. 42. Nicolae A, Pittaluga S, Abdullah S, et al. EBV-positive large Bcell lymphomas in young patients: a nodal lymphoma with evidence for a tolerogenic immune environment. Blood. 2015;126(7):863-872. 43. McCord R, Bolen CR, Koeppen H, et al. PD-L1 and tumorassociated macrophages in de novo DLBCL. Blood Adv. 2019;3(4):531-540. 44. Garcia-Lacarte M, Grijalba SC, Melchor J, Arnaiz-Leché A, Roa S. The PD-1/PD-L1 checkpoint in normal germinal centers and diffuse large B-cell lymphomas. Cancers (Basel). 2021;13(18):4683. 45. Li X, Zhang W. Expression of PD-L1 in EBV-associated malignancies. Int Immunopharmacol. 2021;95:107553. 46. Chong LC, Twa DD, Mottok A, et al. Comprehensive characterization of programmed death ligand structural rearrangements in B-cell non-Hodgkin lymphomas. Blood. 2016;128(9):1206-1213. 47. Xu-Monette ZY, Li J, Xia Y, et al. Immunoglobulin somatic hypermutation has clinical impact in DLBCL and potential implications for immune checkpoint blockade and neoantigenbased immunotherapies. J Immunother Cancer. 2019;7(1):272. 48. Buchhalter I, Rempel E, Endris V, et al. Size matters: dissecting key parameters for panel-based tumor mutational burden analysis. Int J Cancer. 2019;144(4):848-858. 49. Endris V, Buchhalter I, Allgäuer M, et al. Measurement of tumor mutational burden (TMB) in routine molecular diagnostics: in silico and real-life analysis of three larger gene panels. Int J Cancer. 2019;144(9):2303-2312. 50. Yarchoan M, Albacker LA, Hopkins AC, et al. PD-L1 expression

Haematologica | 107 August 2022

1889

ARTICLE - PD-L1 expression in aggressive pediatric lymphomas and tumor mutational burden are independent biomarkers in most cancers. JCI Insight. 2019;4(6):e126908. 51. Reddy A, Zhang J, Davis NS, et al. Genetic and functional drivers of diffuse large B cell lymphoma. Cell. 2017;171(2):481494. 52. Kotlov N, Bagaev A, Revuelta MV, et al. Clinical and biological

K. Fisher et al.

subtypes of B-cell lymphoma revealed by microenvironmental signatures. Cancer Discov. 2021;11(6):1468-1489. 53. Wright GW, Huang DW, Phelan JD, et al. A probabilistic classification tool for genetic subtypes of diffuse large B cell lymphoma with therapeutic implications. Cancer Cell. 2020;37(4):551-568.

Haematologica | 107 August 2022

1890

ARTICLE - Plasma Cell Disorders

Comprehensive genomic analysis of refractory multiple myeloma reveals a complex mutational landscape associated with drug resistance and novel therapeutic vulnerabilities Nicola Giesen,1,2* Nagarajan Paramasivam,3,4* Umut H. Toprak,3,5* Daniel Huebschmann,4,6,7,8* Jing Xu,1,2,9 Sebastian Uhrig,4,9 Mehmet Samur,10,11 Stella Bähr,4 Martina Fröhlich,4,9 Sadaf S. Mughal,9 Elias K. Mai,1 Anna Jauch,12 Carsten Müller-Tidow,1,13 Benedikt Brors,8,9,13 Nikhil Munshi,14 Hartmut Goldschmidt,1,13 Niels Weinhold,1# Matthias Schlesner3# and Marc S. Raab1,2# Department of Internal Medicine V, Heidelberg University Hospital, Heidelberg, Germany; Clinical Cooperation Unit Molecular Hematology/Oncology, Department of Internal Medicine V, Heidelberg University Hospital, and German Cancer Research Center (DKFZ), Heidelberg, Germany; 3Bioinformatics and Omics Data Analytics, German Cancer Research Center (DKFZ), Heidelberg, Germany; 4Computational Oncology, Molecular Diagnostics Program, National Center for Tumor Diseases (NCT) Heidelberg and German Cancer Research Center (DKFZ) Heidelberg, Germany; 5Division of Neuroblastoma Genomics, German Cancer Research Center (DKFZ), Heidelberg, Germany; 6Heidelberg Institute for Stem Cell Technology and Experimental Medicine (HI-STEM), Heidelberg, Germany; 7Department of Pediatric Immunology, Hematology and Oncology, Heidelberg University Hospital, Heidelberg, Germany; 8German Cancer Consortium (DKTK), Core Center Heidelberg, Germany; 9Division of Applied Bioinformatics, German Cancer Research Center (DKFZ), Heidelberg, Germany; 10Department of Data Sciences, Dana-Farber Cancer Institute, Boston, MA, USA; 11Department of Biostatistics, Harvard TH Chan School of Public Health, Boston, MA, USA; 12Institute for Human Genetics, Heidelberg University Hospital, Heidelberg, Germany; 13National Center for Tumor Diseases (NCT), Heidelberg, Germany and 14Jerome Lipper Multiple Myeloma Center, Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA, USA 1

Correspondence: Marc S. Raab m.raab@dkfz-heidelberg.de Nicola Giesen nicola.giesen@med.uni-heidelberg.de Received: June 7, 2021. Accepted: January 7, 2022. Prepublished: January 20, 2022. https://doi.org/10.3324/haematol.2021.279360 Haematologica material is published under a CC-BY license

*NG, NP, UHT, and DH contributed equally as co-first authors. NW, MS, and MSR contributed equally as co-senior authors.

Abstract The outcomes of patients with multiple myeloma (MM) refractory to immunomodulatory agents (IMiDs) and proteasome inhibitors (PIs) remain poor. In this study, we performed whole genome and transcriptome sequencing of 39 heavily pretreated relapsed/refractory MM (RRMM) patients to identify mechanisms of resistance and potential therapeutic targets. We observed a high mutational load and indications of increased genomic instability. Recurrently mutated genes in RRMM, which had not been previously reported or only observed at a lower frequency in newly diagnosed MM, included NRAS, BRAF, TP53, SLC4A7, MLLT4, EWSR1, HCFC2, and COPS3. We found multiple genomic regions with bi-allelic events affecting tumor suppressor genes and demonstrated a significant adverse impact of bi-allelic TP53 alterations on survival. With regard to potentially resistance conferring mutations, recurrently mutated gene networks included genes with relevance for PI and IMiD activity; the latter particularly affecting members of the Cereblon and the COP9 signalosome complex. We observed a major impact of signatures associated with exposure to melphalan or impaired DNA double-strand break homologous recombination repair in RRMM. The latter coincided with mutations in genes associated with PARP inhibitor sensitivity in 49% of RRMM patients; a finding with potential therapeutic implications. In conclusion, this comprehensive genomic characterization revealed a complex mutational and structural landscape in RRMM and highlights potential implications for therapeutic strategies.

Haematologica | 107 August 2022

1891

ARTICLE - Genomic analysis of refractory multiple myeloma

Introduction The introduction of novel therapies such as immunomodulatory agents (IMiDs) and proteasome inhibitors (PIs) has improved the outcomes of patients with multiple myeloma (MM), including those with relapsed MM (RMM) following second or third lines of therapy.1 However, survival remains short if the disease becomes resistant to the major drug classes despite the advent of monoclonal antibodies.2,3 Relapsed/refractory MM (RRMM) therefore represents a patient population of particularly high unmet medical need.4 Thus, a better understanding of the pathophysiology of RRMM is key to improving outcome of these patients. In recent years, advances have been made in elucidating the genomic landscape of newly diagnosed MM (NDMM). These studies have revealed marked clonal heterogeneity with recurrently mutated genes each only affecting a minority of patients.5-7 Clonal evolution over the course of the disease caused both by therapeutic interventions and ongoing genetic instability leads to more resistant clones in RMM, and ultimately to RRMM.8-11 Accordingly, RRMM represents, in many respects, a fundamentally different biological disease entity. However, in contrast to NDMM, genomic data on RRMM is still limited. Data from targeted sequencing in RRMM has identified an evolved set of mutated genes with enrichment for certain oncogenic drivers such as KRAS, NRAS, and TP53 mutations and the development of what are presumed to be resistance-associated mutations.11 Mutations conferring resistance to PIs and IMiDs have been described in proteasomal subunits and in the IMiD target gene Cereblon (CRBN), respectively.11,12 It is noteworthy that most of these mutations only occur in a minority of patients and at low allele frequency.13 Additional mechanisms of resistance may, therefore, be important in RRMM. An unbiased and comprehensive molecular study is, consequently, required to fully dissect the biology underlying RRMM. Here, we used whole genome sequencing (WGS) and RNA sequencing (RNA-Seq) to comprehensively analyze a carefully selected cohort of 39 heavily pretreated RRMM patients with at least double refractory disease, revealing a complex mutational and structural landscape and highlighting potential implications for personalized therapeutic strategies.

Methods Patient characteristics WGS and RNA-Seq was performed on samples from 39 highly refractory MM patients. Nine of these samples have previously been analyzed and reported by our group using targeted sequencing.11 The median number of prior therapy lines was five (2-13), all had relapsed following IMiDs and

N. Giesen et al.

PIs, and 34 (87%) had had an autologous transplant. All patients were at least double-refractory, 62% were at least triple-refractory, and 38% were quadruple-refractory to IMiDs and PIs (Online Supplementary Table S1). In addition, 8% were also refractory to anti-CD38 monoclonal antibodies. Median progression-free survival (PFS) calculated from the time of sampling was 3.5 months and median overall survival (OS) was 7.4 months. Using fluorescence in situ hybridization (FISH)-based methods, 21 (54%) RRMM patients harbored high-risk cytogenetic aberrations; specifically, 16 (41%) had deletion 17p. For a comparison with NDMM, we further analyzed a WGS data set of 21 NDMM patients. For comparative analyses of mutational signatures, we included WGS data of 15 RMM patients with a median of two (1-4) prior therapy lines, thus less heavily pretreated than our RRMM cohort. Both additional WGS data sets are publicly available and had a median coverage of 40x.14,15 Sample acquisition and preparation Between March 2014 and October 2017, tumor samples from 39 RRMM patients were collected at Heidelberg University Hospital. Written informed consent was obtained prior to sampling in accordance with the Declaration of Helsinki. This study was approved by the Institutional Ethics Committee. CD138+ plasma cells were isolated, as described previously.11 DNA and RNA were extracted using the AllPrep kit (QIAGEN, Hilden, Germany). Saliva, buccal swabs or bone marrow stroma cells obtained from cultured CD138- cells were used as germline controls. Saliva was collected in Oragene-Dx tubes and DNA was extracted using prepIT-L2P (DNA Genotek, Ottawa, Canada). DNA from buccal swabs was extracted using the blackPREP Swab DNA kit (Analytik Jena, Jena, Germany). DNA from stroma cells was extracted using the QIAamp DNA Mini kit (QIAGEN, Hilden, Germany). Whole genome sequencing As described previously,16 DNA libraries were prepared following the Illumina TruSeq Nano DNA Library protocol using the TruSeq DNA Nano kit (Illumina, Hayward, CA) and then sequenced on two lanes on the HiSeq X (2×151 bp) using the HiSeq X Ten Reagent Kit v2.5 (Illumina, Hayward, CA) to a median coverage of 77x. Alignment and small variant calling The raw reads were mapped to the human reference genome (build 37, version hs37d5), using BWA mem17 (version 0.7.8). To assess the effect of differing sequencing depths between the NDMM (median coverage 40x), RMM (40x), and RRMM (77x) samples on variant calling, subsampling of RRMM samples was performed for comparative analyses using Sambamba18 (version 0.6.6) to achieve a 50% lower coverage. Small nucleotide variants (SNVs) were called using SAMtools mpileup (version 0.1.19) and bcftools view. Indels were

Haematologica | 107 August 2022

1892

ARTICLE - Genomic analysis of refractory multiple myeloma called using Platypus19 (version 0.8.1). Variants were annotated with Gencode20 (version 19) and ANNOVAR.21 Splicing SNVs or SNVs resulting in nonsynonymous coding were called ‘functional SNVs’. For the prediction of the functional relevance of SNVs, we calculated Combined Annotation Dependent Depletion (CADD) scores (version 1.3) and used a cut-off of 20. Driver genes were identified using IntOGen (version 3.0.5).22 Analysis of mutational signatures and Kataegis clusters A supervised analysis of mutational signatures was performed with the R package YAPSA,23 based on the mutational catalogue of the 30 known signatures from COSMIC v2 (https://cancer.sanger.ac.uk/signatures/signatures_v2), supplemented with the MM1 signature which was recently linked to melphalan exposure.15,24 We defined Kataegis-like clusters to be regions of increased SNV density with at least five SNVs with, at most, a 1000 bp inter-mutational distance in one sample.25 Calculation of HRDetect scores HRDetect is a weighted model used to predict BRCA1/2 deficient tumors.26 We used an implementation of HRDetect available at https://github.com/eyzhao/hrdetect-pipeline.27 Detection of copy number variation and structural variants Copy number states were called, as described previously,16,25 and estimation of tumor purity and ploidy was performed using ACEseq (allele-specific copy number estimation from sequencing; https://www.biorxiv.org/content/early/2017/10/29/210807). Structural variants (SVs) were detected using the DKFZ SOPHIA workflow (version 2.0.2, https://github.com/DKFZODCF/SophiaWorkflow).16,25 SV candidate detection is a process of split-read and discordant mate evidence collection across each breakpoint as precursors for an SV. SV candidates (pairs of breakpoints) are filtered by a complex decision tree trained by expert assessment of orthogonal FISH data. RNA sequencing RNA-Seq libraries were prepared using the Illumina TruSeq stranded mRNA kit and were sequenced on the Illumina HiSeq 2000 V4 platform. The paired-end reads were mapped to the STAR index-generated reference genome (build 37, version hs37d5) with gencode (version 19) using STAR28 (version 2.5.2b). The gene expressions were quantified using featureCounts (Subread version 1.5.1). Gene fusions were detected using Arriba version 1.0.0 (https://github.com/suhrig/arriba), as described previously.29 A detailed description of the bioinformatics workflow and subsequent analyses is provided in the Online Supplementary Methods.

N. Giesen et al.

Results High mutational load and genomic instability in RRMM RRMM displayed a complex mutational landscape and an increase of both chromosomal and nucleotide aberrations compared to NDMM (Figure 1A, Online Supplementary Figure S1). With a median of 67 (range 7-496), the overall load of SVs was significantly higher in RRMM (P=0.002) (Figure 1B, Online Supplementary Figure S1). Complex structural rearrangements and catastrophic events were a frequent finding in RRMM patients, notably chromoplexy (n=10) and chromothripsis (n=5) (Figure 1C, D). Numerical chromosomal aberrations, which occurred significantly more frequently in RRMM compared to NDMM, included gain(1q) and deletions of 1p, 13q, and 17p (all P<0.05, Online Supplementary Figure S1). The overall mutational load in RRMM was also significantly higher than in NDMM (P<10-5) (Figure 1E, Online Supplementary Figure S1). We observed a median number of 116 (range 42-237) functional SNVs and five (range 1-15) functional indels per patient and, overall, a median of 3.94 somatic small variants per megabase in RRMM. RRMM patients showed a much higher prevalence of SNVs outside of Kataegis-like clusters than NDMM patients (P<10-5), indicating activity of mutational mechanisms on a broader scale in RRMM (Online Supplementary Figure S1). Genomic instability, assessed by calculation of the unbiased sum of homologous recombination deficiency (HRD), large-scale transition (LST) and telomeric allelic imbalance (TAI) scores, was significantly increased in RRMM compared to NDMM (P=0.004, Figure 1F, Online Supplementary Figure S1). Driver gene aberrations in RRMM Significantly mutated driver genes in RRMM featured prominently members of the mitogen-activated protein kinase (MAPK) pathway and TP53 (Figure 2, Online Supplementary Table S2) and showed a large overlap with known drivers in NDMM. However, the prevalence of BRAF (P<0.001) and TP53 (P<0.0001) mutations was significantly higher in RRMM compared to data from NDMM patients published by Walker et al.;7 a trend was also seen for NRAS mutations (P=0.05). In addition, we identified the following genes as significantly mutated drivers in RRMM: the sodium bicarbonate co-transporter SLC4A7 (13%); the Ras target MLLT4 (5%); the RNA binding protein EWSR1 (5%); the MLL complex member HCFC2 (5%); the COP9 signalosome subunit COPS3 (5%) (Figure 2). RNA-Seq data confirmed expression of the vast majority of variants and virtually all variants were predicted to be functionally relevant by CADD score (Online Supplementary Figure S2). We found multiple bi-allelic events or ‘double hits’ in RRMM affecting known tumor suppressor genes (TSGs) (Online Supplementary Figure S3). In total, 25/39 RRMM

Haematologica | 107 August 2022

1893

ARTICLE - Genomic analysis of refractory multiple myeloma

N. Giesen et al.

Continued on following page. Haematologica | 107 August 2022

1894

ARTICLE - Genomic analysis of refractory multiple myeloma

N. Giesen et al.

Figure 1. High genomic complexity and mutational load in RRMM vs. NDMM. (A) SV and SNV load per patient in RRMM vs. NDMM. SNV and SV counts are plotted for each patient in both cohorts showing a higher overall mutational load in RRMM (blue) vs. NDMM (red). Each dot represents an individual patient. The example cases shown in panels C and D are annotated as RRMM_16 and RRMM_15, respectively. (B) Differences in SV types in RRMM vs. NDMM. Median and range of number of overall SVs per patient in RRMM (blue) vs. NDMM (red) are shown as well as deletions (DEL), duplications (DUP), inversions (INV), and translocations (TRA). (C) Example case of RRMM displaying chromoplexy. Green lines represent translocations, blue lines deletions, red lines duplications, and black lines inversions. Transparency of lines is based on estimated SV clonality. Variant existence is represented by bar plots. The outer layer represents copy number variations and displays the copy-neutral nature of the chromoplexy event. (D) Example case of RRMM displaying chromothripsis. Green lines represent translocations, blue lines deletions, red lines duplications, and black lines inversions. Transparency of lines is based on estimated SV clonality. Variant existence is represented by bar plots. The outer layer represents copy number variations. (E) Genome-wide small variant mutational load in RRMM vs. NDMM. The number of mutations per patient and length of genome in megabases (MB) is shown in RRMM (blue) vs. NDMM (red). (F) Genomic instability scores in RRMM vs. NDMM. The unbiased sum of HRD, LST, and TAI scores is shown for RRMM (blue) vs. NDMM (red), illustrating a higher genomic instability in RRMM. To compensate for differing sequencing depths in both cohorts, the RRMM dataset was subsampled for these analyses.

patients presented with at least one double hit in TSGs, significantly increased compared to NDMM (3/21, P<0.001). The most frequently affected TSGs in RRMM were TP53 (n=8), RB1 (n=7), and TRAF3 (n=6). RRMM patients with a double hit of TP53 had significantly inferior PFS (P=0.004) and a trend for inferior OS (P=0.07); both calculated from time of sampling in univariate log-rank tests compared to patients with no or a single hit of TP53 (Figure 3). While treatment of patients before and after sampling was heterogeneous, the adverse impact of bi-allelic TP53 alterations remained significant in multivariate analysis including age, number of prior therapies and ISS as possible confounders with regard to both PFS (hazard ratio (HR) 4.02; P=0.01) and OS (HR 4.77; P=0.02). ‘Double hits’ in other single TSGs, or a combination thereof, did not show a significant impact on survival which would have been independent of TP53 events.

Oncogenic networks in RRMM Next, we analyzed whether gene groups, resistance mechanisms or signaling networks were recurrently affected by small variants in our RRMM cohort (Online Supplementary Table S4). We first addressed genes linked to the mechanisms of action of PIs and IMiDs (Online Supplementary Figure S3-5, Online Supplementary Table S5). We found that, overall, 21% of RRMM patients as opposed to only 5% of NDMM patients harbored mutations in genes considered relevant to PI activity. These were mainly proteasome subunits (PSMB5, PSMC2, PSMC6, PSMD2, PSMD11, and PSME3). Recurrent mutations were also detected in TJP1, previously found to modulate PI sensitivity in MM.30 However, statistical significance was not reached (P=0.14), likely due to the limited sample size. Regarding IMiD resistance, candidate mutations were found to be significantly more frequent in RRMM patients compared to

Figure 2. Significantly mutated driver genes in RRMM. Significantly mutated drivers and their prevalence in the RRMM cohort are shown as well as copy number aberrations (CNAs) of chromosome arms 13q, 1q, 17p, 1p, presence or absence of hyperdiploid karyotype, and the genomic instability score. Haematologica | 107 August 2022

1895

ARTICLE - Genomic analysis of refractory multiple myeloma

N. Giesen et al.

Figure 3. Impact of TP53 alteration on PFS (A) and OS (B) in RRMM. PFS and OS are shown for RRMM patients with a bi-allelic event involving the TP53 locus (red), a mono-allelic event (blue) or no event (black), illustrating the inferior outcome of patients with bi-allelic TP53 aberrations. PFS and OS were calculated from time of sampling.

NDMM (31% vs. 5%, P=0.02). We observed mutations in all four members of the CRBN complex (CRBN, CUL4B, RBX1, DDB1), IZKF1 as downstream target, as well as in three members of the COP9 signalosome complex (COPS3, COPS4, COPS8) and in CAND1, which are regulators of cullin-RING ligase neddylation and which were recently identified in a CRISPR-Cas9 screen as being essential for the IMiD mechanism of action.31 The MAPK pathway harbored mutations in 77% of RRMM patients (52% in NDMM), mostly due to mutations in NRAS, KRAS and BRAF. NFκB signaling was affected in 23% of such patients (5% in NDMM) with recurrent mutations in NFKB1, NFKB2, TRAF3 and CHUK. Both pathways showed a trend to being more frequently affected in RRMM, as did the large functional group of epigenetic modifiers which was affected in 77% of RRMM patients (52% in NDMM, all P=0.08). Of potential therapeutic interest, mutations in genes associated with sensitivity to PARP inhibitors were found in 49% of RRMM compared to 29% of NDMM patients, however, not reaching statistical significance (P=0.17), with recurrent mutations in ATM, NBN, and TOP3A (Online Supplementary Figure S3, S6, Online Supplementary Table S5). Enrichment of mutational signatures of impaired DNA damage repair in RRMM To identify mechanisms contributing to high mutational load and high genomic instability observed in RRMM as compared to NDMM, we analyzed mutational signatures (Figure 4A, Online Supplementary Figure S7). For further

evaluation of different disease stages, we also included a cohort of relapsed but less heavily pretreated RMM patients as an intermediate stage. This cohort had previously been analyzed by Maura et al.15 In supervised fitting with signature-specific cutoffs, we found a significantly higher contribution from COSMIC signatures AC3 (associated with deficiency in homologous recombination repair) and the melphalan signature MM1 (P=0.006 and P<0.001, respectively) in RRMM compared to NDMM (Figure 4B) at the cost of the clock-like signatures AC1 (spontaneous deamination) and AC5 (clock-like but unknown) with significantly lower contributions (P<0.001 and P=0.005, respectively). In RMM, we detected signature MM1 but to a lesser extent when compared to our RRMM cohort (P=0.004), and signature AC3 was found with exposures in-between the NDMM and RRMM cohorts (Figure 4A, Online Supplementary Figure S7). A comparison of the clinical information on the published RMM cohort with our RRMM patients confirmed a significantly higher overall number of prior therapies as well as more extensive exposure to both novel agents and to high-dose melphalan therapy in our RRMM patients. Information on all mutational signatures detected in this manuscript, including asserted mutational mechanisms, is summarized in Online Supplementary Table S6. To further assess the finding of enriched deficiency in homologous recombination repair as indicated by signature AC3, we applied HRDetect, which is a weighted model used to predict BRCA1/2 deficient tumors.26 HRDetect scores were significantly higher in RRMM than in both NDMM and

Haematologica | 107 August 2022

1896

ARTICLE - Genomic analysis of refractory multiple myeloma RMM (P<0.0001 and P=0.01, respectively, Figure 4C, Online Supplementary Figure S7). This was in line with the observation of higher exposure to mutational signature AC3 in RRMM than in NDMM and RMM. In 6/39 RRMM samples, the HRDetect scores exceeded the value of 0.7, a cutoff which was recently established to identify tumors with a high level

N. Giesen et al.

of BRCA1/BRCA2 deficiency.26 Taken together, these observations indicate a shift in the activities of different mutational mechanisms during the course of the disease. This might be a consequence of mechanisms intrinsic to the tumor cells (e.g., acquisition of DNA repair deficiencies) or may result from exposure to therapeutic agents.

Figure 4. Exposure to mutational signatures in RRMM vs. RMM vs. NDMM patients. (A) Absolute exposure to mutational signatures in RRMM vs. RMM vs. NDMM patients. Exposure to mutational signatures based on the Alexandrov COSMIC (AC) catalogue with the addition of the MM1 signature recently linked to melphalan exposure is shown for RRMM (blue) vs. RMM (green) vs. NDMM (red) patients. Most notable is an increased impact of signatures AC3 (light brown) and MM1 (black) in RRMM. (B) Relative exposure to mutational signatures in RRMM vs. RMM vs. NDMM patients. Exposure to mutational signatures is shown for RRMM (blue) vs. RMM (green) vs. NDMM (red) patient cohorts. Significant differences in exposure are indicated with the following P-values: * <0.05, ** <0.01, *** < 0.001. Most notable is a significant increase of signature AC3 in RRMM compared to NDMM and RMM as well as a significant step-wise increase of signature MM1 in RRMM vs. RMM vs. NDMM. (C) HRDetect scores in RRMM vs. RMM vs. NDMM patients. HRDetect scores in RRMM (blue) vs. RMM (green) vs. NDMM (red) indicate a significant increase in impaired homologous recombination repair features in RRMM. To compensate for differing sequencing depths in both cohorts, the RRMM dataset was subsampled for this analysis. Haematologica | 107 August 2022

1897

ARTICLE - Genomic analysis of refractory multiple myeloma

N. Giesen et al.

Enriched chromosomal translocations in RRMM Next, we sought to identify translocations, which were enriched or unique to RRMM (Figure 5A, Online Supplementary Figure S8). While MYC rearrangements per se were not enriched in RRMM, FAM46C-MYC translocations (n=4) and local rearrangements of chr8q24.21 (n=4) were both found exclusively in RRMM patients (Figure 5B). The super-enhancer of FAM46C showed rearrangements involving a number of target genes such as the transcription factor and presumed oncogene LMO4. An analysis of expression data confirmed enhancer hijacking (Online Supplementary Figure S9). IgH translocations involving MYCN were observed in two cases as part of composite t(4;14)-t(2;4) translocations also involving the MMSET locus, co-activating MMSET and MYCN. In these cases, MYCN was highly expressed, while MYC expression was completely suppressed (Online Supplementary Figure S10). Both cases showed extramedullary disease, secondary plasma cell leukemia and a distinct plasmablastic cytological appearance. In one case, retrospective analysis of earlier samples revealed the absence of the secondary t(2;4) translocation and an absence of

concomitant MYCN activation and MYC suppression, suggesting the emergence of MYCN overexpression with disease progression.

Discussion This comprehensive study of extensively pretreated, highly refractory MM by WGS and RNA-Seq suggests that the pathophysiology of NDMM and the biology of refractory disease are strikingly different. In RRMM, we observed marked genomic instability with impaired DNA repair mechanisms, in particular homologous recombination repair (HRR), which we confirmed using the well-established HRDetect algorithm. Mutational signatures indicative of impaired DNA repair, such as AC3, have been reported in NDMM although not to the extent observed here in this highly refractory setting.32,33 A recent study has reported the absence of mutational signature AC3 and HRR deficiency in NDMM and has classified the appearance of signature AC3 in NDMM as a false positive effect of the applied deconvolution method.15 A study on the genomic

Figure 5. Immunoglobulin translocations and MYC rearrangements in RRMM. (A) Immunoglobulin translocations in RRMM. Translocations involving the immunoglobulin heavy chain (IGH) locus, the lambda light chain (IGL), and the kappa light chain (IGK) locus are shown. The number of patients with involvement of the respective partner genes are given in brackets. Patient RRMM_34, harboring a very complex IGL translocation, was excluded from this graph for reasons of readability and is shown separately in Online Supplementary Figure S8. Cytobands, chromosome arms and chromosomes were also stretched and compressed to emphasize targets of immunoglobulin translocations and to improve readability. (B) MYC rearrangements in RRMM. Most notable are local rearrangements as well as those involving the IGL locus or FAM46C. Orange lines represent immunoglobulin locus translocations; purple lines secondary immunoglobulin locus related translocations. Secondary translocations of the immunoglobulin loci were defined as secondary events with one of the breakpoints of the SV not further away than 2MB from the target breakpoint of a given primary immunoglobulin translocation (i.e., IG —PrimarySV—> PrimaryTarget —SecondarySV—>SecondaryTarget). Green lines indicate non-IG locus translocations, blue lines deletions, red lines duplications, and black lines inversions. Haematologica | 107 August 2022

1898

ARTICLE - Genomic analysis of refractory multiple myeloma make-up of relapsed but less heavily pretreated RMM using whole-exome sequencing did not report mutational signature AC3, but rather a novel signature associated with alkylator therapies.13 In our work, we find both mutational signatures and thereby provide evidence that they are not reflective of the same underlying mutational process and can be deconvoluted with sufficient statistical power in the underlying data. Walker et al. recently used the term double hit MM to indicate a subset of NDMM with a very poor prognosis, including patients with bi-allelic TP53 inactivation.34 In our study, patients with a TP53 double hit experienced the worst outcomes in RRMM, confirming recent observations in the relapsed setting.10,35 Further genes with recurrent biallelic aberrations in RRMM include RB1 and TRAF3, though they did not independently affect the prognosis in our cohort. This was likely attributable to the limited number of patients who had ‘double hits’ in these TSGs but not concomitant bi-allelic TP53 events. BRAF mutations occur at a higher prevalence in RRMM. The frequency seen in our cohort seems to exceed even those previously reported in RRMM,36 a finding which may reflect the heavily pretreated nature of the patient cohort presented here. However, a possible selection bias has to be borne in mind as a strong driver mutation such as BRAF V600E might be associated with higher tumor load, thus potentially resulting in a higher success rate of plasma cell purification for WGS. Further, the limited sample size might have contributed to this finding. Nonetheless, the eight cases of the druggable mutation BRAF V600E are of particular therapeutic interest.37,38 Among potential resistance-conferring SNVs, individual genes were only affected at low frequencies. However, at the level of functional networks, recurrence was actually seen: we found mutations in several proteasomal subunits as well as in TJP1 which modulates PI sensitivity in MM.12,30,39 While the functional impact of most of these mutations remains to be proven, their enrichment in RRMM supports an association with PI resistance. The same holds true for mutations in genes presumably associated with IMiD resistance, such as CRBN, CUL4B, and IZKF1. Furthermore, we detected mutations in three members of the COP9 signalosome complex (COPS3, COPS4, and COPS8) and CAND1, further supporting their functional impact on IMiD activity.31 The increased mutational load might explain both the higher capacity of MM cells to adapt to treatment and facilitate the emergence of resistance. One of the mechanisms contributing to the higher mutational load in RRMM appears to be impaired DNA double-strand break HRR with the potential therapeutic implication of synthetic lethality to pharmacological inhibitors of DNA damage response, such as ATR inhibitors. At the same time, we found mutations in genes associated with sensitivity to PARP inhibitors in 49%

N. Giesen et al.

of RRMM patients. This further strengthens the rationale for assessing the therapeutic efficacy of ATR and PARP inhibition in RRMM as has been shown, for example, in solid cancers with BRCAness characteristics.40 In fact, there is plenty of pre-clinical evidence for synthetic lethality conferred by such inhibitors in MM cell lines with high rates of ongoing DNA damage.41-43 As PI treatment has been suggested to induce a BRCAness-like state in MM cells via impairment of DNA repair pathways,44 there may also be a rationale for combining PI and PARP inhibitors in MM. However, these concepts need to be confirmed in clinical trials. One major limitation of our study is that comparative analyses were performed between independent NDMM, RMM, and RRMM patient cohorts, limiting the ability to draw conclusions regarding the tumor evolution under treatment. Thus, one can only speculate as to whether genomic instability was pre-existing in these cases, e.g., restricted to certain focal lesions,45 and was then selected for or whether it was newly acquired following treatment with e.g., DNA damaging drugs. Longitudinal analyses that also address spatial heterogeneity will, therefore, be of particular interest. While this real-world RRMM cohort represents similarly ultra-refractory MM patients, the route taken to this end-stage disease differed greatly between individual patients. It is well conceivable that important biological differences exist between RRMM patients reaching refractoriness after multiple lines of treatment and those progressing quickly through a limited number of therapies. Our RRMM cohort size might, therefore, be too small to overcome the heterogeneity of this patient population with sufficient statistical power. Further analyses on larger or more homogeneous RRMM patient cohorts will help elucidate these issues. In conclusion, based on our observation that RRMM is characterized by marked genomic instability, which enables MM cells to rapidly adapt to selective therapeutic pressure, treatment strategies focused on exploiting impaired HRR should be evaluated within prospective clinical trials. Such strategies might be particularly useful in the current era of novel immunotherapies in MM as recent reports suggest genomic instability as a mechanism of resistance to CAR T cell treatment.46,47 Targeting impaired DNA repair mechanisms may, therefore, help to improve the outcomes of patients with RRMM. Disclosures NG received honoraria from MSD and advisory board from Pfizer; EKM received honoraria from Janssen, Celgene, Takeda, has a consulting or advisory role with Janssen, Celgene, Takeda, received research funding from Takeda and travel support from Janssen, Takeda, Celgene and Mundipharma; HG received grants and/or a provision of Investigational Medicinal Products from Amgen, BMS, Celgene, Chugai, Dietmar-Hopp-Stiftung, Janssen, John Hopkins

Haematologica | 107 August 2022

1899

ARTICLE - Genomic analysis of refractory multiple myeloma

N. Giesen et al.

University, Sanofi and research support from Amgen, BMS, Celgene, Chugai, Janssen, Molecular Partners, MSD, Sanofi, Mundipharma, Takeda, Novartis, advisory boards from Adaptive Biotechnology, Amgen, BMS, Celgene, Janssen, Sanofi, Takeda and honoraria from ArtTempi, BMS, Celgene, Chugai, Janssen, Novartis and Sanofi; MSR received honoraria, consultancy, research grants and travel support from Novartis and Amgen. The other authors declare no competing interests.

Acknowledgments The authors would like to thank Anja Baumann, Katrin Pfütze and Bettina Meißburger, the DKFZ Omics IT and Data Management Core Facility, and the DKFZ Genomics and Proteomics Core Facility for their excellent technical support, and the Heidelberg Center for Personalized Oncology (DKFZ-HIPO), the Dietmar-Hopp Foundation, the Olympia-Morata program, Novartis and Amgen for funding support; BB received funding from the ERC under the European Union's Horizon 2020 research and innovation program (grant agreement N°825835); Contributions MS and NM received funding from NIH (grants P50-CA100707 NG, NW, MS, and MSR designed the study, collected and and P01-CA155258). We especially thank all the patients and analyzed data, and wrote the paper; NP, UHT, and DH ana- their families for their participation in this study. lyzed the data and wrote the paper; JX, SU, SB, MF, SSM, AJ, and BB analyzed data; EKM, CMT, NM, and HG collected Data-Sharing statement data. All authors critically reviewed and approved the final Sequence data has been made available at the European version of the manuscript. Genome-phenome Archive (EGAS00001004363).

References 1. Kumar SK, Rajkumar SV, Dispenzieri A, et al. Improved survival in multiple myeloma and the impact of novel therapies. Blood. 2008;111(5):2516-2520. 2. Kumar SK, Lee JH, Lahuerta JJ, et al. Risk of progression and survival in multiple myeloma relapsing after therapy with IMiDs and bortezomib: a multicenter international myeloma working group study. Leukemia. 2012;26(1):149-157. 3. Kumar SK, Dimopoulos MA, Kastritis E, et al. Natural history of relapsed myeloma, refractory to immunomodulatory drugs and proteasome inhibitors: a multicenter IMWG study. Leukemia. 2017;31(11):2443-2448. 4. Usmani SZ, Weiss BM, Plesner T, et al. Clinical efficacy of daratumumab monotherapy in patients with heavily pretreated relapsed or refractory multiple myeloma. Blood. 2016;128(1):37-44. 5. Chapman MA, Lawrence MS, Keats JJ, et al. Initial genome sequencing and analysis of multiple myeloma. Nature. 2011;471(7339):467-472. 6. Walker BA, Boyle EM, Wardell CP, et al. Mutational spectrum, copy number changes, and outcome: results of a sequencing study of patients with newly diagnosed myeloma. J Clin Oncol. 2015;33(33):3911-3920. 7. 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. 8. Bolli N, Avet-Loiseau H, Wedge DC, et al. Heterogeneity of genomic evolution and mutational profiles in multiple myeloma. Nat Commun. 2014;5:2997. 9. 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. 10. Weinhold N, Ashby C, Rasche L, et al. Clonal selection and double-hit events involving tumor suppressor genes underlie relapse in myeloma. Blood. 2016;128(13):1735-1744. 11. Kortum KM, Mai EK, Hanafiah NH, et al. Targeted sequencing of refractory myeloma reveals a high incidence of mutations in CRBN and Ras pathway genes. Blood. 2016;128(9):1226-1233. 12. Barrio S, Stuhmer T, Da-Via M, et al. Spectrum and functional validation of PSMB5 mutations in multiple myeloma. Leukemia.

2019;33(2):447-456. 13. Ziccheddu B, Biancon G, Bagnoli F, et al. Integrative analysis of the genomic and transcriptomic landscape of double-refractory multiple myeloma. Blood Adv. 2020;4(5):830-844. 14. Maura F, Bolli N, Angelopoulos N, et al. Genomic landscape and chronological reconstruction of driver events in multiple myeloma. Nat Commun. 2019;10(1):3835. 15. Maura F, Degasperi A, Nadeu F, et al. A practical guide for mutational signature analysis in hematological malignancies. Nat Commun. 2019;10(1):2969. 16. Paramasivam N, Hubschmann D, Toprak UH, et al. Mutational patterns and regulatory networks in epigenetic subgroups of meningioma. Acta Neuropathol. 2019;138(2):295-308. 17. Li H. Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM. arXiv. 2013 May 26. https://doi.org/10.48550/arXiv.1303.3997 [preprint, not peer-reviewed]. 18. Tarasov A, Vilella AJ, Cuppen E, Nijman IJ, Prins P. Sambamba: fast processing of NGS alignment formats. Bioinformatics. 2015;31(12):2032-2034. 19. Rimmer A, Phan H, Mathieson I, et al. Integrating mapping-, assembly- and haplotype-based approaches for calling variants in clinical sequencing applications. Nat Genet. 2014;46(8):912-918. 20. Harrow J, Frankish A, Gonzalez JM, et al. GENCODE: the reference human genome annotation for The ENCODE Project. Genome Res. 2012;22(9):1760-1774. 21. Wang K, Li M, Hakonarson H. ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Res. 2010;38(16):e164. 22. Gonzalez-Perez A, Perez-Llamas C, Deu-Pons J, et al. IntOGenmutations identifies cancer drivers across tumor types. Nat Methods. 2013;10(11):1081-1082. 23. Hubschmann D, Jopp-Saile L, Andresen C, et al. Analysis of mutational signatures with yet another package for signature analysis. Genes Chromosomes Cancer. 2021;60(5):314-331. 24. Rustad EH, Yellapantula V, Leongamornlert D, et al. Timing the initiation of multiple myeloma. Nat Commun. 2020;11(1):1917. 25. Lopez C, Kleinheinz K, Aukema SM, et al. Genomic and

Haematologica | 107 August 2022

1900

ARTICLE - Genomic analysis of refractory multiple myeloma transcriptomic changes complement each other in the pathogenesis of sporadic Burkitt lymphoma. Nat Commun. 2019;10(1):1459. 26. Davies H, Glodzik D, Morganella S, et al. HRDetect is a predictor of BRCA1 and BRCA2 deficiency based on mutational signatures. Nat Med. 2017;23(4):517-525. 27. Zhao EY, Shen Y, Pleasance E, et al. Homologous recombination deficiency and platinum-based therapy outcomes in advanced breast cancer. Clin Cancer Res. 2017;23(24):7521-7530. 28. Dobin A, Davis CA, Schlesinger F, et al. STAR: ultrafast universal RNA-S eq aligner. Bioinformatics. 2013;29(1):15-21. 29. Heining C, Horak P, Uhrig S, et al. NRG1 Fusions in KRAS wildtype pancreatic cancer. Cancer Discov. 2018;8(9):1087-1095. 30. Zhang XD, Baladandayuthapani V, Lin H, et al. Tight junction protein 1 modulates proteasome capacity and proteasome inhibitor sensitivity in multiple myeloma via EGFR/JAK1/STAT3 signaling. Cancer Cell. 2016;29(5):639-652. 31. Sievers QL, Gasser JA, Cowley GS, Fischer ES, Ebert BL. Genome-wide screen identifies cullin-RING ligase machinery required for lenalidomide-dependent CRL4(CRBN) activity. Blood. 2018;132(12):1293-1303. 32. Hoang PH, Cornish AJ, Dobbins SE, Kaiser M, Houlston RS. Mutational processes contributing to the development of multiple myeloma. Blood Cancer J. 2019;9(8):60. 33. Hoang PH, Dobbins SE, Cornish AJ, et al. Whole-genome sequencing of multiple myeloma reveals oncogenic pathways are targeted somatically through multiple mechanisms. Leukemia. 2018;32(11):2459-2470. 34. 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. 35. Lakshman A, Painuly U, Rajkumar SV, et al. Impact of acquired del(17p) in multiple myeloma. Blood Adv. 2019;3(13):1930-1938. 36. Xu J, Pfarr N, Endris V, et al. Molecular signaling in multiple

N. Giesen et al.

myeloma: association of RAS/RAF mutations and MEK/ERK pathway activation. Oncogenesis. 2017;6(5):e337. 37. Andrulis M, Lehners N, Capper D, et al. Targeting the BRAF V600E mutation in multiple myeloma. Cancer Discov. 2013;3(8):862-869. 38. Raab MS, Lehners N, Xu J, et al. Spatially divergent clonal evolution in multiple myeloma: overcoming resistance to BRAF inhibition. Blood. 2016;127(17):2155-2157. 39. Shi CX, Kortum KM, Zhu YX, et al. CRISPR Genome-wide screening identifies dependence on the proteasome subunit PSMC6 for bortezomib sensitivity in multiple myeloma. Mol Cancer Ther. 2017;16(12):2862-2870. 40. Lord CJ, Ashworth A. BRCAness revisited. Nat Rev Cancer. 2016;16(2):110-120. 41. Cottini F, Hideshima T, Suzuki R, et al. Synthetic lethal approaches exploiting DNA damage in aggressive myeloma. Cancer Discov. 2015;5(9):972-987. 42. Herrero AB, Gutierrez NC. Targeting ongoing DNA damage in multiple myeloma: effects of DNA damage response inhibitors on plasma cell survival. Front Oncol. 2017;7:98. 43. Botrugno OA, Bianchessi S, Zambroni D, et al. ATR addiction in multiple myeloma: synthetic lethal approaches exploiting established therapies. Haematologica. 2020;105(10):2440-2447. 44. Neri P, Ren L, Gratton K, et al. Bortezomib-induced "BRCAness" sensitizes multiple myeloma cells to PARP inhibitors. Blood. 2011;118(24):6368-6379. 45. Rasche L, Chavan SS, Stephens OW, et al. Spatial genomic heterogeneity in multiple myeloma revealed by multi-region sequencing. Nat Commun. 2017;8(1):268. 46. Orlando EJ, Han X, Tribouley C, et al. Genetic mechanisms of target antigen loss in CAR19 therapy of acute lymphoblastic leukemia. Nat Med. 2018;24(10):1504-1506. 47. Da Via MC, Dietrich O, Truger M, et al. Homozygous BCMA gene deletion in response to anti-BCMA CAR T cells in a patient with multiple myeloma. Nat Med. 2021;27(4):616-619.

Haematologica | 107 August 2022

1901

ARTICLE - Platelet Biology & its Disorders

Sorting nexin 24 is required for α-granule biogenesis and cargo delivery in megakaryocytes Joanne Lacey,1 Simon J. Webster,1 Paul R. Heath,2 Chris J. Hill,3 Lucinda Nicholson-Goult,4 Bart E. Wagner,4 Abdullah O. Khan,5 Neil V. Morgan,5 Michael Makris1 and Martina E. Daly1 Department of Infection, Immunity and Cardiovascular Disease, University of Sheffield, Sheffield; 2Sheffield Institute for Translational Neuroscience (SITraN), Department of Neuroscience, University of Sheffield, Sheffield; 3Department of Molecular Biology and Biotechnology, University of Sheffield, Sheffield; 4Histopathology Department, Royal Hallamshire Hospital, Sheffield and 5Institute of Cardiovascular Sciences, College of Medical and Dental Sciences, University of Birmingham, Birmingham, UK. 1

Correspondence: Martina E. Daly M.Daly@sheffield.ac.uk Received: July 14, 2021 Accepted: January 3, 2022. Prepublished: January 13, 2022. https://doi.org/10.3324/haematol.2021.279636 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

Abstract Germline defects affecting the DNA-binding domain of the transcription factor FLI1 are associated with a bleeding disorder that is characterized by the presence of large, fused α-granules in platelets. We investigated whether the genes showing abnormal expression in FLI1-deficient platelets could be involved in platelet α-granule biogenesis by undertaking transcriptome analysis of control platelets and platelets harboring a DNA-binding variant of FLI1. Our analysis identified 2,276 transcripts that were differentially expressed in FLI1-deficient platelets. Functional annotation clustering of the coding transcripts revealed significant enrichment for gene annotations relating to protein transport, and identified Sorting nexin 24 (SNX24) as a candidate for further investigation. Using an induced pluripotent stem cell-derived megakaryocyte model, SNX24 expression was found to be increased during the early stages of megakaryocyte differentiation and downregulated during proplatelet formation, indicating tight regulatory control during megakaryopoiesis. CRISPR-Cas9 mediated knockout (KO) of SNX24 led to decreased expression of immature megakaryocyte markers, CD41 and CD61, and increased expression of the mature megakaryocyte marker CD42b (P=0.0001), without affecting megakaryocyte polyploidisation, or proplatelet formation. Electron microscopic analysis revealed an increase in empty membrane-bound organelles in SNX24 KO megakaryocytes, a reduction in α-granules and an absence of immature and mature multivesicular bodies, consistent with a defect in the intermediate stage of α-granule maturation. Co-localization studies showed that SNX24 associates with each compartment of α-granule maturation. Reduced expression of CD62P and VWF was observed in SNX24 KO megakaryocytes. We conclude that SNX24 is required for α-granule biogenesis and intracellular trafficking of α-granule cargo within megakaryocytes.

Introduction The ETS transcription factor Friend leukemia virus integration 1, or FLI1, plays a fundamental role in megakaryopoiesis by cooperating with the ETS factor GA binding protein, alpha subunit (GABPA), to regulate the expression of multiple megakaryocyte-specific genes expressed during the early and late stages of megakaryopoiesis.1,2 Partial deletion of chromosome 11, in a region that includes the gene encoding FLI1 (11q23.3-24), is associated with Paris Trousseau syndrome, which is characterized by dysmegakaryopoiesis in the bone marrow.3,4 Affected individuals have an increased tendency to bleed, as well as thrombocytopenia characterized by the presence of large platelets containing large fused α-granules in the

circulation.5-7 Germline defects in FLI1 have also been described.8-10 Indeed, we previously reported an enrichment of heterozygous FLI1 defects among patients with excessive bleeding in association with a significant reduction in δ-granule secretion, and mild thrombocytopenia.8 Further characterization of two FLI1 alterations predicting amino acid substitutions in the ETS DNA-binding domain of FLI1 showed that they disrupted transcriptional activity and would therefore cause a reduction in the expression of megakaryocyte-specific genes, providing an explanation for the bleeding tendency observed in the patients.8 The identification of a homozygous FLI1 defect, which predicted a substitution in the ETS domain and resulted in a defect in transcription among members of a family affected by a bleed-

Haematologica | 107 August 2022

1902

ARTICLE - SNX24 and α-granule biogenesis

J. Lacey et al.

ing disorder that resembled Paris Trousseau syndrome and was characterized by thrombocytopenia and the presence of abnormally large α-granules in a subpopulation of platelets, suggested that the abnormal platelet granules were due to loss of FLI1 activity.9 The identification of two further FLI1 defects affecting residues in the ETS domain of FLI1 which were associated with the presence of giant α-granules and depletion of δ-granules supports this hypothesis.10 While the role of platelet degranulation in maintaining vascular integrity has long been recogniszd, the pathways leading to the assembly and exocytosis of platelet α- and δ-granules, and the genes that regulate them, remain to be fully established. The biogenesis of α- and δ-granules shares some common features, but the pathways involve distinct protein trafficking machinery.11,12 Both granules are derived through a process of budding from the trans-Golgi network to form vesicles that merge with early endosomes and mature into multivesicular bodies.13-15 The α-granules can also be derived from the platelet membrane through clathrin-coated pit-mediated endocytosis to form vesicles that traffic to the early endosomes.15,16 Studies in patients with inherited granule storage disorders,17-19 combined with extensive analyses of platelet lysates, and studies in mice carrying mutations in different secretion genes, have allowed characterization of the cargo of platelet storage granules, and provided essential information on the secretory machinery of platelets.20-22 The abnormally large α-granules and reduced number of δgranules, and the defect in δ-granule secretion that have been described in platelets from patients harboring FLI1 defects likely reflect the disruptions in gene expression that occur either directly or indirectly as a result of the abnormal transcriptional activity of FLI1.8-10 Furthermore, some of the genes showing disrupted expression in FLI1-deficient platelets displaying granule abnormalities could potentially be involved in platelet granule biogenesis under physiological conditions. We explored this hypothesis by undertaking transcriptome analysis of platelets from subjects harboring a deleterious defect in FLI1 (c.1028A>G; p.Tyr343Cys) to identify differentially expressed genes encoding proteins which may be involved in platelet granule biogenesis and secretion, and identified a role for Sorting nexin 24 (SNX24) in platelet granule biogenesis using wild-type and SNX24 knockout (KO) induced pluripotent stem cell (iPSC)-derived megakaryocytes to model maturation and platelet production.

Methods Subjects and platelet transcriptome analysis Total platelet RNA was isolated from 50 mL samples of peripheral blood from two heterozygous carriers of the c.1028A>G transition in FLI1 and from three sex-matched healthy individuals. All subjects were studied in parallel on two separate occasions. The study was approved by the Na-

tional Research Ethics Service Committee (REC reference: 06/MRE07/36). Following differential centrifugation to obtain platelet-rich plasma, total platelet RNA was isolated using Trizol (see Online Supplementary Methods). Two hundred nanograms of RNA (RIN >7 as measured on the Agilent Bioanalyzer) were converted to double-stranded cDNA and transcriptome analysis was carried out using human Clariom D Assay chips (Thermo Fisher), which were washed and stained according to the manufacture’s standard protocols. The arrays were scanned using the Affymetrix 3000 7G scanner, and the .CEL files analyzed using Transcriptome Analysis Console (TAC) 4.0 software (Thermo Fisher) to identify genes showing >2 and <-2 fold log change and P<0.05. Functional annotation clustering was carried out for those genes which were differentially expressed using the Database for Annotation, Visualization and Integrated Discovery (DAVID; http://david.abcc.ncifcrf.gov/) with the default settings and a low stringency setting.23,24 Induced pluripotent stem cell differentiation to megakaryocytes The Gibco episomal human iPSC line was cultivated feederfree on Geltrex-coated flasks and maintained in StemFlex medium (Thermo Scientific). iPSCs were differentiated to mature megakaryocytes and proplatelets as described previously.25,26 Details of the cell culture conditions and differentiation of iPSC are included in the Online Supplementary Methods. Generation of the SNX24 knockout cell line The SNX24 KO cell line was generated using the Alt-R RNP system (Integrated DNA Technologies; IDT). SNX24 crRNA and Atto-555 labeled tracrRNA were annealed and the complex incubated with HiFi Cas9 V3 (IDT) to form stable RNP complexes, which were introduced into iPSC using Lipofectamine Stem (Life Technologies). For single-cell cloning, StemFlex medium was supplemented with CloneR (Stemcell Technologies) and the manufacturer’s workflow was followed (see Online Supplementary Methods for further details). Gene expression analysis of SNX24 knockout induced pluripotent stem cells Gene expression of lineage-specific and cellular markers was measured during SNX24 KO iPSC differentiation by quantitative polymerase chain reaction (qPCR) using an ABI 7900 HT analyzer (Applied Biosystems). Further details are included in the Online Supplementary Methods, and all oligonucleotide sequences are listed in Online Supplementary Table S1. Characterization of megakaryocytes derived from SNX24 knockout induced pluripotent stem cells Differentiation of SNX24 KO cells to proplatelet-forming

Haematologica | 107 August 2022

1903

ARTICLE - SNX24 and α-granule biogenesis

J. Lacey et al.

megakaryocytes was assessed by immunofluorescent staining of cells, and imaging using a Zeiss A1 confocal microscope. Further details of the antibodies used, and the methodology are provided in the Online Supplementary Methods. Transmission electron microscopy of platelets and megakaryocytes FLI1-deficient platelets and iPSC-derived megakaryocytes were examined by transmission electron microscopy (TEM) using a Philips 400 and a FEI Tecnai transmission electron microscope, respectively. Further details are provided in the Online Supplementary Methods. Statistical analysis Results are expressed as means ± standard deviation. Unless otherwise specified, statistical significance was determined using a Student t-test with P<0.05 considered to be statistically significant. Analyses were performed using GraphPad Prism software.

Results Transcriptome analysis of FLI1-deficient platelets Transcriptome analysis was undertaken using Human Clariom D Assay chips for platelets from a father (P1) and son (P2), both of whom were heterozygous for the c.1028A>G FLI1 variant predicting a p.Tyr343Cys substitution in the DNA-binding domain of FLI1. Both subjects were recruited to the UK GAPP study with a history of excessive bleeding and a suspected inherited platelet disorder, which was characterized by mild thrombocytopenia and a profound reduction in platelet ATP secretion in response to thrombin. The clinical features, and genotypic and phenotypic characteristics of both subjects have been reported previously.8 Electron microscopic examination of platelets from both P1 and P2 revealed the presence of giant and fused α-granules similar to those previously described in platelets from subjects with FLI1 defects (Figure 1A).9,10 Platelet transcriptomes were also analyzed from three healthy male subjects (C1, C2 and C3). Transcripts were considered to be significantly differentially expressed if they had a >2 or <-2 fold log change, and a P<0.05. Hierarchical clustering and principal component analysis showed clear differences between the FLI1-deficient platelets from P1 and P2 and normal platelets from C1, C2 and C3 (Figure 1B and Online Supplementary Figure S1A). Comparison of gene expression in platelets from the two cases with the c.1028A>G FLI1 variant with that in normal platelets (C1, C2 and C3) identified 2,276 significantly differentially expressed transcripts (926 downregulated and 1350 upregulated) in the FLI1-deficient platelets (Figure 1C and Online Supplementary Tables S2 and S3). The 30

coding transcripts displaying the greatest up- and downregulation in expression in FLI1-deficient platelets relative to normal platelets are indicated in Figure 1D, while Online Supplementary Tables S4 and S5 list all coding transcripts showing significant down- and up-regulation in FLI1-deficient platelets. To identify genes for further investigation, functional annotation clustering of the 1,487 differentially expressed coding transcripts was undertaken using the DAVID tool (Online Supplementary Table S6). This identified 234 clusters, which were enriched for multiple classes of annotation categories. Given our interest in platelet granule biogenesis and secretion, we focused on the gene ontology (GO) terms associated within the top four clusters: protein transport, cell-cell adhesion, endoplasmic reticulum, and late endosome (Figure 1E and Online Supplementary Table S7). Together these clusters included three of the 30 most differentially expressed genes; SNX24, HBE1 and TESPA1. Of these, SNX24 was the most downregulated, having a 45.91-fold reduction in expression in FLI1-deficient platelets compared with normal platelets (false discovery rate P=0.0034), and was the only one of the nine genes represented in a subgroup of 53 genes from the cluster that was enriched for the GO term ‘protein transport’ (P=0.024) (Figure 1E, F and Online Supplementary Table S8). The 232 genes associated with GO terms in the top four annotation clusters are listed by their gene symbols and descriptions, and expression levels in Online Supplementary Table S9. Further work focused on SNX24. qPCR of SNX24 expression using independent probes confirmed its downregulation in platelet RNA isolated from both P1 and P2 (Online Supplementary Figure S1B). SNX24 is required during early megakaryopoiesis SNX24 is a member of the Sorting nexin family of proteins which are defined by the presence of a Phox homology (PX) phosphoinositide-binding domain and play essential roles in regulating protein trafficking, through all stages of the endocytic pathway. While a specific role for SNX24 in platelets has not been demonstrated, variants of SNX24 have been associated with platelet-crit and mean platelet volume,27,28 and more recently, SNX24 was shown to be upregulated in megakaryocytes with ploidy.29 These findings, and our observation that SNX24 was downregulated in FLI1-deficient platelets displaying abnormal α-granules, suggest a role for SNX24 in platelet formation which we explored further in iPSC-derived proplatelet-forming megakaryocytes. We generated an SNX24 KO in iPSC using CRISPR guides targeting the first exon of SNX24 and confirmed the absence of SNX24 expression in two clones by qPCR (Online Supplementary Figure S2A). Further analysis confirmed that both clones were homozygous KO, and that both the wild-type and KO cells retained normal karyotypes (Online

Haematologica | 107 August 2022

1904

ARTICLE - SNX24 and α-granule biogenesis

J. Lacey et al.

Figure 1. Transcriptome analysis of FLI1-deficient platelets identified SNX24 for further investigation. ( A) Representative electron micrographs of platelets from two subjects, P1 and P2, displaying large and fused -granules (indicated by yellow arrows). (B) Hierarchical clustering of differentially expressed transcripts in platelets from controls (C1, C2, C3) and FLI1-deficient subjects (P1, P2). All samples were analyzed in parallel on two occasions. (C) Volcano plot showing the transcripts that are downregulated (green) or upregulated (red) in FLI1-deficient platelets relative to control platelets. (D) Log fold change of the 30 most downregulated and upregulated coding transcripts observed in FLI1-deficient platelets relative to control platelets (<2 or ≥2 fold change, false discovery rate ≤0.05). (E) Functional annotation clustering of differentially expressed coding transcripts using DAVID identified 234 annotation clusters. Gene ontology (GO) terms from the top four annotation clusters which were significantly enriched are shown and the log10 P-value for each GO term is displayed on the X-axis. The numbers of upregulated and downregulated genes with each GO term are indicated. (F) Log fold change of the 15 most downregulated (green) and upregulated (red) genes showing enrichment for the GO terms protein transport, cell adhesion, endoplasmic reticulum and late endosome (<2 or ≥2 fold change, false discovery rate ≤0.05). Haematologica | 107 August 2022

1905

ARTICLE - SNX24 and α-granule biogenesis

J. Lacey et al.

Supplementary Figure S2B-E). We used established procedures to differentiate iPSC (OCT4+) to hematopoietic progenitors derived from hemogenic endothelium (CD34+; day 0-6), then to immature megakaryocytes (CD41+/CD42–; day 6-12) before terminal differentiation to mature proplatelet-forming megakaryocytes (CD41+/CD42+; day 12-17) (Figure 2A). qPCR to assess expression of specific markers of stem cells (OCT4), hematopoietic stem cells (CD34) and megakaryocytes (CD61/GpIIIa, CD41/GPIIb and CD42b/GPIbα) confirmed the expected expression for the different stages of differentiation of the wild-type iPSC (Online Supplementary Figure S3A). Similarly, assessment of FLI1 expression by qPCR showed the expected increase in expression during megakaryocyte maturation as FLI1 regulates the expression of both early and late megakaryocyte-specific genes (Online Supplementary Figure S3B).2 Examination of SNX24 expression during differenti-

ation of wild-type iPSC showed that it steadily increased during the early stages of megakaryocyte differentiation and was subsequently downregulated during proplatelet formation (Figure 2B). This suggests that expression of SNX24 is tightly controlled during megakaryopoiesis. We observed a decrease in expression of the immature megakaryocyte markers (CD41 and CD61) and an increase in expression of the mature megakaryocyte marker (CD42b) at day 12 of differentiation of the SNX24 KO iPSC, suggesting the presence of more mature megakaryocyte progenitors (Figure 2C, D, and Online Supplementary Figure S3C). SNX24 KO iPSC generated megakaryocytes with similar ploidy numbers to the wild-type iPSC, suggesting that depletion of SNX24 does not inhibit polyploidization (Figure 2E). SNX24 KO cells also retained the ability to generate proplatelet-forming megakaryocytes after terminal differentiation (Figure 2F). These data suggest a po-

Figure 2. Characterization of megakaryocytes derived from SNX24 knockout induced pluripotent stem cells (A) Schematic of differentiation of induced pluripotent stem cells (iPSC) to megakaryocytes. (B) Quantitative polymerase chain reaction (qPCR) analysis of SNX24 expression during iPSC differentiation to megakaryocytes. n=3 experiments. ****P<0.0001, **P<0.01, Student t-test. (C) Immunofluorescence staining for CD41 (green) and CD42 (red) in wild-type (WT) cells at day 12 of differentiation. Nuclei are counterstained with Hoechst 33342. Scale bar 50 µm. (D) qPCR analysis of CD41, CD61 and CD42 gene expression in day 12 WT and SNX24 knockout (KO) megakaryocyte progenitors. ****P<0.0001, ***P<0.001. Student t-test, n=2. (E) Representative confocal microscopy fluorescence images of CD41 (green) and CD42 (red) in WT and SNX24 KO megakaryocytes. Nuclei are counterstained with Hoechst 33342. Scale bar 20 µm. (F) Brightfield images of proplatelets in WT and SNX24 KO cells. Scale bar 20 µm. MK: megakaryocytes; PP: proplatelets.

Haematologica | 107 August 2022

1906

ARTICLE - SNX24 and α-granule biogenesis

J. Lacey et al.

tential role for SNX24 in the early stages of megakaryopoiesis during megakaryocyte progenitor formation. SNX24 depletion causes loss of granule content To assess whether SNX24 plays a role in granule formation, we examined the ultrastructure of SNX24 KO megakaryocytes by TEM. We observed several organelles in wild-type iPSC-derived megakaryocytes including α-granules, δ-granules, endosomal intermediates (multivesicular bodies [MVB] types I and II) and mitochondria. We defined an α-granule as a single membrane enclosing a matrix and a δ-granule as a round organelle with a high-density core surrounded by a white rim. There was a dramatic increase in the presence of empty membrane-bound organelles in the SNX24 KO cells, although the mitochondria appeared

unaffected (Figure 3A). The empty spherical organelles, multivesicular subclasses and tubular shaped compartments were suggestive of morphologically distinct stages of α-granules. δ-granules could not be accurately evaluated in iPSC-derived megakaryocytes, for which unstained electron microscopy would be more informative. Wildtype megakaryocytes contained early endosomal compartments packed with vesicles and intraluminal contents, multivesicular bodies (MVB) with intraluminal vesicles and late endosomes with intraluminal contents. Loss of SNX24 resulted in empty intermediate endosomal compartments during granule biogenesis (Figure 3B). These findings support the participation of SNX24 in the biogenesis and maturation of MVB as well as in the development of α-granules.

B Figure 3. SNX24 knockout megakaryocytes lack α-granules (A) Representative transmission electron micrographs of wild-type (WT) and SNX24 knockout (KO) megakaryocytes. Spherical and tubular α-granules (arrow heads) are observed in WT cells and absent in SNX24 KO cells. The upper panels are low magnification with a scale bar of 2 µm. The lower panels are high magnification images with a scale bar of 0.5 µm. (B) Electron micrographs of WT cells and SNX24 KO cells. Multivesicular bodies (arrow heads) containing intraluminal vesicles were observed in WT cells but not in SNX24 KO cells. The upper panels are low magnification images with a scale bar of 1 µm. The lower panels are high magnification images with a scale bar of 0.5 µm. Mag.: magnification; MK: megakaryocytes. Haematologica | 107 August 2022

1907

ARTICLE - SNX24 and α-granule biogenesis

J. Lacey et al.

SNX24 localizes to α-granules and endosomal compartments We explored whether SNX24 is required for the biogenesis of α-granules and endosomal precursors using confocal fluorescence microscopy to examine its co-localization with markers for α-granules (CD62P), early endosomes (EEA1), late endosomes and MVB (Rab7a) in megakaryocytes. We observed SNX24 localization in iPSC-derived megakaryocytes, which we identified by multiple nuclei (N). In megakaryocytes, SNX24 partially co-localized with CD62P in punctate structures in 2N and 6N megakaryocytes (Figure 4A). We observed that SNX24 was localized to both larger α-granules and smaller punctate structures, which we speculate could be transport vesicles. SNX24 was also associated with the plasma membrane where it was distributed as small puncta and aggregated clusters, some of which co-localized with CD62P. SNX24 expression was low in the proplatelets but still co-localized to CD62P positive punctate structures (Figure 4B). SNX24 was partially co-localized with early endosomes in megakaryocytes around the plasma membrane, and associated with EEA1-positive compartments (Figure 4C). SNX24 also associated with Rab7a-positive late endosomes and MVB (Figure 4D). These observations indicate that SNX24 is associated with each stage of α-granule maturation, and suggest that it may traffic from early endosomes to mature α-granules.

We also examined gene expression of markers specific for early endosomes (Rab5a) and MVB (Rab7a) in SNX24 KO megakaryocytes. Interestingly, Rab5a and Rab7a gene expression was significantly increased in SNX24 KO megakaryocytes compared to the wild-type cells, which was consistent with disruption to the endosomal trafficking pathway (Online Supplementary Figure S4). In contrast, NBEAL2 expression was significantly decreased in the SNX24 KO megakaryocytes, supporting the involvement of SNX24 at an earlier stage of α-granule development than NBEAL2 (Online Supplementary Figure S4). SNX24 knockout cells lack α-granule cargo The appearance of empty α-granules and intermediate endosomal compartments in SNX24 KO cells is suggestive of a depletion in α-granule cargo. Furthermore, the distribution pattern of SNX24 around the cell periphery and in α-granule compartments suggests it may be involved in the trafficking of α-granule cargo. We therefore examined the subcellular distribution of α-granule cargo (CD62P, von Willebrand factor [VWF]) in SNX24 KO megakaryocytes. We observed a reduction in CD62P and VWF staining in megakaryocytes, indicating an α-granule trafficking defect (Figure 5A, B). Reverse transcriptase qPCR analysis confirmed that CD62P and VWF gene expression was significantly reduced in SNX24 KO cells (Figure 5C, D). These findings indicate a requirement for SNX24 for traf-

Figure 4. SNX24α localizes to α-granules and endosomal intermediates. (A) Representative confocal microscopy images of wild-type early megakaryocytes (upper panels) and late megakaryocytes (lower panels) stained with anti-SNX24 (green) and anti-CD62 (red). Nuclei are counterstained with Hoechst 33342. Scale bar 20 µm. (B) Representative confocal microscopy images of wild-type proplatelets stained with anti-SNX24 (green) and anti-CD62 (red). Nuclei are stained with Hoechst 33342. Scale bar 20 µm. (C) Representative confocal microscopy images of wild-type megakaryocytes stained with anti-SNX24 (green) and anti-EEA1 (red). Nuclei are stained with Hoechst 33342. Scale bar 10 µm. (D) Representative confocal fluorescence microscopy images of wild-type megakaryocytes immunostained with anti-SNX24 (green) and anti-Rab7a (red). Nuclei are stained with Hoechst 33342. Scale bar 10 µm. MK: megakaryocytes. Haematologica | 107 August 2022

1908

ARTICLE - SNX24 and α-granule biogenesis

J. Lacey et al.

ficking of both soluble (VWF) and membrane (CD62P) cargo.

Discussion The pathways leading to platelet granule biogenesis and exocytosis, and the genes that regulate them, remain to be fully established. In this study, we sought to identify candidate genes that potentially contribute to platelet granule biogenesis by taking advantage of the changes in gene expression that occur in platelets from individuals with germline mutations that affect the transcriptional activity of FLI1, which also display abnormal α-granules. Thus, transcriptome analysis of platelets harboring a DNAbinding variant of FLI1 revealed significant differences in expression profiles of FLI1-deficient and wild-type platelets with 2,276 transcripts identified as being differentially expressed in the FLI1-deficient platelets. Functional annotation clustering of the differentially expressed coding transcripts revealed significant enrichment for annotations relating to protein transport, and led us to focus on SNX24, a protein not previously implicated in platelet granule biogenesis. We confirmed that SNX24 expression was significantly downregulated in the FLI1-deficient platelets and observed that SNX24 is localized throughout the pathway of α-granule biogenesis, associating with early endosomes, MVB and α-granules. Furthermore, loss of SNX24 disrupts the endosomal trafficking pathway in megakaryocytes, resulting in an α-granule defect and decreased expression of α-granule proteins. Our findings lead us to propose that SNX24 is a novel component of the protein sorting machinery during α-granule maturation. The Sorting nexins (SNX) are a diverse family of cytoplasmic and membrane-associated proteins that are involved in endocytosis, endosomal sorting and endosomal signaling. They are characterized by the presence of a conserved PX domain which binds specific phosphoinositides, facilitating targeting of proteins to distinct endosomal compartments.30 Sorting nexins play critical roles in many aspects of cellular function, and dysfunction of SNX proteins has been described in association with a variety of human disorders including neurodegenerative diseases, pathological infection, cancer and cardiovascular disease.31,32 SNX24 is one of a subgroup of the SNX protein family that is relatively poorly characterized, although genome-wide association studies have associated variants of SNX24 with platelet-crit and volume.27,28 An interesting correlation between the SNX24 single nucleotide polymorphism (rs28891) and complications due to coronary artery aneurysm in Kawasaki disease has also been reported, and in the same study siRNA knockdown of SNX24 expression was shown to significantly decrease expression

of the pro-inflammatory cytokines IL-1b, IL-6, and IL-8 in lipopolysaccharide-treated human umbilical vein endothelial cells.33 Interestingly, α-granules contain a wide range of chemokines including IL-8.21 Other Sorting nexins have been implicated in platelet granule formation. In particular, SNX17 was identified in a yeast two-hybrid screen as an interaction partner of P-selectin34 and it is thought that SNX17 may regulate the endocytosis of P-selectin from the plasma membrane and inhibit trafficking to lysosomes.35 The mechanism underlying the significant downregulation of SNX24 in FLI1-deficient platelets is unclear. Previous work which mapped the genome-wide FLI1-binding sites in primary human megakaryocytes indicated that FLI1 does not bind directly to the SNX24 promoter.36 The reduced expression is more likely to be an indirect effect of FLI1 loss, potentially through GABPA which acts in concert with FLI1 to regulate megakaryocyte gene expression, since GABPA gene expression was significantly upregulated in the FLI1-deficient platelets and it is predicted to bind to the SNX24 promoter.37 We used functional annotation clustering as a tool to prioritize candidate genes encoding proteins that may be involved in platelet granule biogenesis, focusing our attention on the cluster that was enriched for the GO annotation of ‘protein transport’ which included SNX24. Interestingly, this subgroup also included SEC22B, Homolog B Vesicle Trafficking Protein, a membrane-resident trafficking protein that was recently shown to be required for α-granule biogenesis in megakaryocytes and was upregulated in the FLI1-deficient platelets.38 Although further work would be required to determine whether any of the remaining genes in this subgroup are required for platelet granule formation, this observation nonetheless supports the use of transcriptome analysis of FLI1-deficient platelets as an approach to identify novel components of the granule biogenesis machinery. We used iPSC-derived megakaryocytes to study the role of SNX24 during megakaryopoiesis, comparing wild-type cells and SNX24 KO cells which were generated using CRISPR-Cas9 gene editing. SNX24 expression appears to be tightly regulated during megakaryopoiesis, being highest during early megakaryopoiesis at the megakaryocyte progenitor stage, and dropping significantly in mature proplatelet-forming megakaryocytes. We observed that SNX24 KO cells have a higher proportion of mature CD42b+ megakaryocyte progenitors. However, loss of SNX24 did not inhibit formation of polyploid megakaryocytes or proplatelet-forming megakaryocytes. TEM analysis of SNX24 KO megakaryocytes showed, in contrast to wild-type megakaryocytes, a reduction in α-granules and an increase in the presence of empty vacuoles, resembling the phenotype seen in platelets from patients with gray platelet syndrome, which is characterized by the presence

Haematologica | 107 August 2022

1909

ARTICLE - SNX24 and α-granule biogenesis

J. Lacey et al.

B D

Figure 5. Abnormal trafficking of α-granule proteins in SNX24 knockout cells (A) Representative confocal microscopy images of wild-type (WT) and SNX24 knockout (KO) megakaryocytes immunostained with anti-VWF (green) and anti-CD42 (red). Nuclei are stained with Hoechst 33342. Scale bar 20 µm. (B) Representative confocal microscopy images of WT and SNX24 KO megakaryocytes immunostained with anti-CD62P (green) and anti-CD42 (red). Nuclei are stained with Hoechst 33342. Scale bar 20 µm. (C) Quantitative polymerase chain reaction (qPCR) analysis of VWF gene expression in day 12 WT and SNX24 KO megakaryocytes. ****P<0.0001 Student t-test, n=2. (D) qPCR analysis of CD62P gene expression in day 12 WT and SNX24 KO megakaryocytes. **P<0.01, Student t-test, n=2. MK: megakaryocytes; iPSC. Induced pluripotent stem cells.

of empty α-granules.39 SNX24 KO cells were also devoid of immature type I MVB that contain internal vesicles, and type II MVB that contain internal vesicles and an electron dense core. The ultrastructural abnormalities observed in SNX24 KO megakaryocytes are consistent with defects in the intermediate stages of α-granule maturation and resemble those seen in VPS33B KO cells.40 Studies in patients with inherited defects that cause deficiencies in α-granule cargo and number have yielded valu-

able insights into the mechanisms underlying platelet granule biogenesis. Thus, defects in the genes encoding the VPS33B-VPS16B complex, and the BEACH-domain containing NBEAL2, have been shown to underlie the absence of α-granules in arthrogryposis, renal dysfunction and cholestasis (ARC) syndrome and in gray platelet syndrome, respectively.18,19,41-43 Loss of VPS33B results in the absence of α-granules, decreased levels of α-granule cargo and a reduction in MVB.40,44 Likewise, loss of VPS16B

Haematologica | 107 August 2022

1910

ARTICLE - SNX24 and α-granule biogenesis

J. Lacey et al.

results in reduced or undetectable α-granule proteins as well as a complete absence of α-granules in platelets.41 The VPS33B/16B complex localizes on endosomes and promotes protein trafficking between MVB and α-granules.45 Megakaryocytes deficient in NBEAL2 show a loss of α-granules and a reduction in α-granule cargo proteins such as VWF but retain some CD62P46 and Sec22b was recently identified as an interaction partner of NBEAL2, which facilitates α-granule cargo stability and granule development.38 Our TEM observations indicated an α-granule defect in SNX24 KO cells, which led us to assess whether SNX24 was distributed within α-granules and intermediate compartments. We have shown that SNX24 is co-localized with CD62P in α-granules during megakaryopoiesis. Vesicles carrying α-granule cargo bud off from either the trans-Golgi network or plasma membrane and are subsequently directed to MVB via endosomes.13 We show that SNX24 is co-localized with the early endosome marker EEA1 around the cell periphery, and partially associated with Rab7+ MVB. This suggests that SNX24 traffics between, or binds to α-granule components during their maturation. Likewise, known α-granule machinery such as VPS16B and VPS33B traffic between late endosomes and α-granules.42 We observed that loss of SNX24 leads to a reduction in expression of α-granule cargo including VWF and CD62P. P-selectin is a membrane protein that contains a signal peptide to direct it towards a developing granule.47 VWF is a soluble protein that self-assembles into large aggregates and eventually forms tubular structures within α-granules.48 The differences in expression of VWF and CD62P seen in SNX24 KO cells could reflect differential packaging or vesicle transport to distinct subcompartments within α-granules. Further studies will be required to define the role of SNX24 in intracellular trafficking. In particular, it will be important to investigate its expression throughout megakaryopoiesis and to determine, more precisely, the point at which loss of SNX24 disrupts the trafficking of α-granule cargo and leads to the appearance of empty vacuoles, although our findings suggest that this occurs early during maturation as 2N megakaryocytes derived from SNX24 KO cells lack α-granule cargo. Given its downregulation in mature proplatelet-forming megakaryocytes, and that loss of SNX24 did not inhibit formation of polyploid megakaryocytes or proplatelet-forming megakaryocytes, it

would be interesting to determine the role of SNX24, if any, in platelets. There are no disorders reported to be associated with the SNX24 gene, and there are no phenotyping or viability data available for the Snx24 knockout mouse49, but it would be interesting to assess the appearance and behavior of their platelets, and to compare them with platelets derived from SNX24 KO megakaryocytes. In conclusion, we have presented data that evidence a requirement for SNX24 in α-granule biogenesis and the intracellular trafficking of α-granule cargo within megakaryocytes. Future studies characterizing the molecular interactions of SNX24 will likely reveal further insights into the underlying molecular machinery and protein complexes required for α-granule formation in platelets. Disclosure No conflicts of interest to disclose. Contributions JL and SJW designed and performed the experiments; PRH performed the transcriptome analysis; CJH performed electron microscopy on iPSC-derived megakaryocytes; LNG and BEW performed electron microscopy on clinical samples; MM recruited the patients; AOK provided guidance on the iPSC differentiation; MED and NVM formulated the research idea and designed the study; MED supervised the study; JL analyzed the data, and drafted the first version of the manuscript, which was read and commented on by all authors. Acknowledgments The authors would like to thank the individuals who participated in this study by kindly donating blood samples for transcriptome analysis. Funding This work was supported by the British Heart Foundation through the award of a project grant (PG/15/61/31634). AOK is a Henry Wellcome fellow (218649/Z/19/Z). Data sharing statement Reasonable requests to view the original data will be considered on receipt of an email by Joanne Lacey (j.lacey@sheffield.ac.uk) or Martina Daly (m.daly@sheffield.ac.uk).

References 1. Li Y, Luo H, Liu T, Zacksenhaus E, Ben-David Y. The ets transcription factor Fli-1 in development, cancer and disease. Oncogene. 2014;34(16):2022-2031. 2. Pang L, Xue H-H, Szalai G, et al. Maturation stage-specific regulation of megakaryopoiesis by pointed-domain Ets proteins.

Blood. 2006;108(7):2198-2206. 3. Hromas R, May W, Denny C, et al. Human FLI-1 localizes to chromosome 11Q24 and has an aberrant transcript in neuroepithelioma. Biochim Biophys Acta. 1993;1172(1-2):155-158. 4. Breton-Gorius J, Favier R, Guichard J, et al. A new congenital

Haematologica | 107 August 2022

1911

ARTICLE - SNX24 and α-granule biogenesis

J. Lacey et al.

dysmegakaryopoietic thrombocytopenia (Paris-Trousseau) associated with giant platelet alpha-granules and chromosome 11 deletion at 11q23. Blood. 1995;85(7):1805-1814. 5. Hart A, Melet F, Grossfeld P, et al. Fli-1 is required for murine vascular and megakaryocytic development and is hemizygously deleted in patients with thrombocytopenia. Immunity. 2000;13(2):167-177. 6. Favier R, Jondeau K, Boutard P, et al. Paris-Trousseau syndrome: clinical, hematological, molecular data of ten new cases. Thromb Haemost. 2003;90(5):893-897. 7. Raslova H, Komura E, Le Couédic JP, et al. FLI1 monoallelic expression combined with its hemizygous loss underlies ParisTrousseau/Jacobsen thrombopenia. J Clin Invest. 2004;114(1):77-84. 8. Stockley J, Morgan NV, Bem D, et al. Enrichment of FLI1 and RUNX1 mutations in families with excessive bleeding and platelet dense granule secretion defects. Blood. 2013;122(25):4090-4093. 9. Stevenson WS, Rabbolini DJ, Beutler L, et al. Paris-Trousseau thrombocytopenia is phenocopied by the autosomal recessive inheritance of a DNA-binding domain mutation in FLI1. Blood. 2015;126(17):2027-2030. 10. Saultier P, Vidal L, Canault M, et al. Macrothrombocytopenia and dense granule deficiency associated with FLI1 variants: ultrastructural and pathogenic features. Haematologica. 2017;102(6):1006-1016. 11. Machlus KR, Italiano JE Jr. The incredible journey: from megakaryocyte development to platelet formation. J Cell Biol. 2013;201(6):785-796. 12. Sharda A, Flaumenhaft R. The life cycle of platelet granules. F1000Res. 2018;7:236. 13. Heijnen HF, Debili N, Vainchencker W, Breton-Gorius J, Geuze HJ, Sixma JJ. Multivesicular bodies are an intermediate stage in the formation of platelet alpha-granules. Blood. 1998;91(7):2313-2325. 14. Ambrosio AL, Boyle JA, Di Pietro SM. Mechanism of platelet dense granule biogenesis: study of cargo transport and function of Rab32 and Rab38 in a model system. Blood. 2012;120(19):4072-4081. 15. Chen Y, Yuan Y, Li W. Sorting machineries: how platelet-dense granules differ from α-granules. Biosci Rep. 2018;38(5):BSR20180458. 16. Behnke O. Coated pits and vesicles transfer plasma components to platelet granules. Thromb Haemost. 1989;62(2):718-722. 17. Wei AH, Li W. Hermansky-Pudlak syndrome: pigmentary and non-pigmentary defects and their pathogenesis. Pigment Cell Melanoma Res. 2013;26(2):176-192. 18. Gunay-Aygun M, Falik-Zaccai TC, Vilboux T, et al. NBEAL2 is mutated in gray platelet syndrome and is required for biogenesis of platelet α-granules. Nat Genet. 2011;43(8):732-734. 19. Gissen P, Johnson CA, Morgan NV, et al. Mutations in VPS33B, encoding a regulator of SNARE-dependent membrane fusion, cause arthrogryposis-renal dysfunction-cholestasis (ARC) syndrome. Nat Genet. 2004;36(4):400-404. 20. Kahr WH, Lo RW, Li L, et al. Abnormal megakaryocyte development and platelet function in Nbeal2(-/-) mice. Blood. 2013;122(19):3349-3358. 21. Blair P, Flaumenhaft R. Platelet alpha-granules: basic biology and clinical correlates. Blood Rev. 2009;23(4):177-189. 22. Nurden P, Stritt S, Favier R, Nurden AT. Inherited platelet diseases with normal platelet count: phenotypes, genotypes and diagnostic strategy. Haematologica. 2021;106(2):337-350.

23. Huang DW, Sherman BT, Lempicki RA. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat. Protoc. 2009;4(1):44-57. 24. Huang DW, Sherman BT, Lempicki RA. Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res. 2009;37(1):1-13. 25. Feng Q, Shabrani N, Thon JN, et al. Scalable generation of universal platelets from human induced pluripotent stem cells. Stem Cell Reports. 2014;3(5):817-831. 26. Khan AO, Slater A, Maclachlan A, et al. Post-translational polymodification of β1-tubulin regulates motor protein localisation in platelet production and function. Haematologica. 2022;107(1):243-259. 27. Astle WJ, Elding H, Jiang T, et al. The allelic landscape of human blood cell trait variation and links to common complex disease. Cell. 2016;167(5):1415-1429. 28. Vuckovic D, Bao EL, Akbari P, et al. The polygenic and monogenic basis of blood traits and diseases. Cell. 2020;182(5):1214-1231. 29. Choudry FA, Bagger FO, Macaulay IC, et al. Transcriptional characterization of human megakaryocyte polyploidization and lineage commitment. J Thromb Haemost. 2021;19(5):1236-1249. 30. Haft CR, de la Luz Sierra M, Barr VA, Haft DH, Taylor SI. Identification of a family of sorting nexin molecules and characterization of their association with receptors. Mol Cell Biol. 1998;18(12):7278-7287. 31. Hanley SE, Cooper KF. Sorting nexins in protein homeostasis. Cells. 2020;10(1):17. 32. Yang J, Villar VAM, Rozyyev S, Jose PA, Zeng C. The emerging role of sorting nexins in cardiovascular diseases. Clin Sci (Lond). 2019;133(5):723-737. 33. Lin YJ, Chang JS, Liu X, et al. Sorting nexin 24 genetic variation associates with coronary artery aneurysm severity in Kawasaki disease patients. Cell Biosci. 2013;3(1):44. 34. Florian V, Schlüter T, Bohnensack R. A new member of the sorting nexin family interacts with the C-terminus of P-selectin. Biochem Biophys Res Commun. 2001;281(4):1045-1050. 35. Williams R, Schlüter T, Roberts MS, Knauth P, Bohnensack R, Cutler DF. Sorting nexin 17 accelerates internalization yet retards degradation of P-selectin. Mol Biol Cell. 2004;15(7):3095-3105. 36. Tijssen MR, Cvejic A, Joshi A, et al. Genome-wide analysis of simultaneous GATA1/2, RUNX1, FLI1, and SCL binding in megakaryocytes identifies hematopoietic regulators. Dev Cell. 2011;20(5):597-609. 37. Fishilevich S, Nudel R, Rappaport N, et al. GeneHancer:genomewide integration of enhancers and target genes in GeneCards. Database (Oxford). 2017;2017:bax028. 38. Lo RW, Li L, Pluthero FG, Leung R, Eto K, Kahr WHA. The endoplasmic reticulum protein SEC22B interacts with NBEAL2 and is required for megakaryocyte α-granule biogenesis. Blood. 2020;136(6):715-725. 39. Maynard DM, Heijnen HF, Gahl WA, Gunay-Aygun M. The αgranule proteome: novel proteins in normal and ghost granules in gray platelet syndrome. J Thromb Haemost. 2010;8(8):1786-1796. 40. Bem D, Smith H, Banushi B, et al. VPS33B regulates protein sorting into and maturation of α-granule progenitor organelles in mouse megakaryocytes. Blood. 2015;126(2):133-143. 41. Urban D, Li L, Christensen H, et al. The VPS33B-binding protein VPS16B is required in megakaryocyte and platelet α-granule biogenesis. Blood. 2012;120(25):5032-5040. 42. Kahr WH, Hinckley J, Li L, et al. Mutations in NBEAL2, encoding a BEACH protein, cause gray platelet syndrome. Nat Genet.

Haematologica | 107 August 2022

1912

ARTICLE - SNX24 and α-granule biogenesis

J. Lacey et al.

2011;43(8):738-740. 43. Albers CA, Cvejic A, Favier R, et al. Exome sequencing identifies NBEAL2 as the causative gene for gray platelet syndrome. Nat Genet. 2011;43(8):735-737. 44. Lo B, Li L, Gissen P, Christensen H, et al. Requirement of VPS33B, a member of the Sec1/Munc18 protein family, in megakaryocyte and platelet alpha-granule biogenesis. Blood. 2005;106(13):4159-4166. 45. Ambrosio AL, Di Pietro SM. Mechanism of platelet α-granule biogenesis: study of cargo transport and the VPS33B-VPS16B complex in a model system. Blood Adv. 2019;3(17):2617-2626. 46. Lo RW, Li L, Leung R, Pluthero FG, Kahr WHA. NBEAL2 (NeurobeachinLike 2) Is required for retention of cargo proteins

by α-granules during their production by megakaryocytes. Arterioscler Thromb Vasc Biol. 2018;38(10):2435-2447. 47. Disdier M, Morrissey JH, Fugate RD, Bainton DF, McEver RP. Cytoplasmic domain of P-selectin (CD62) contains the signal for sorting into the regulated secretory pathway. Mol Biol Cell. 1992;3(3):309-321. 48. Huang RH, Wang Y, Roth R, et al. Assembly of Weibel-Palade body-like tubules from N-terminal domains of von Willebrand factor. Proc Natl Acad Sci U S A. 2008;105(2):482-487. 49. Dickinson, M, Flenniken A, Ji X, et al. High-throughput discovery of novel developmental phenotypes. Nature. 2016;537(7621):508-514.

Haematologica | 107 August 2022

1913

ARTICLE - Red Cell Biology & its Disorders

Effects of corticosteroids in patients with sickle cell disease and acute complications: a systematic review and meta-analysis Julien Lopinto,1,2* Ségolène Gendreau,1,2* Enora Berti,1,2 Pablo Bartolucci,3,4 Anoosha Habibi3,4 and Armand Mekontso Dessap 1,2,3 AP-HP, Hôpitaux Universitaires Henri-Mondor, Service de Médecine Intensive Réanimation, Créteil; 2Université Paris Est Créteil, CARMAS; 3Université Paris Est Créteil, INSERM, IMRB, FHU SENEC and 4Université Paris Est Créteil, Centre de Référence des Syndromes Drépanocytaires Majeurs, Unité des Maladies Génétiques du Globule Rouge (UMGGR), Créteil, France 1

Correspondence: Julien Lopinto julien.lopinto@aphp.fr Received: September 28, 2021. Accepted: January 4, 2022. Prepublished: January 13, 2022. https://doi.org/10.3324/haematol.2021.280105 ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

*JL and SG contributed equally as co-first authors.

Abstract Whether corticosteroids improve outcome in patients with acute complications of sickle cell disease (SCD) is still debated. We performed a systematic review of the literature with the aim of estimating effects of corticosteroids on the clinical course of vaso-occlusive crisis (VOC) or acute chest syndrome (ACS) in patients with SCD. The primary outcome was transfusion requirement during hospitalization. Studies were identified by search of MEDLINE and CENTRAL database. Three randomized clinical trials (RCT) and three retrospective cohort studies (RCS) were included, involving 3,304 participants and 5,562 VOC or ACS episodes. There was no difference between corticosteroids and standard treatment regarding transfusion requirement overall (odds ratio [OR]=0.98, 95% confidence interval [CI]: 0.38-2.53) but there was a significant interaction of the study type (P<0.0001): corticosteroid therapy was associated with a lower risk of transfusion in RCT (OR=0.13, 95% CI: 0.04-0.45) and a higher risk of transfusion in RCS (OR=2.12, 95% CI: 1.33-3.40. In RCT, the length of hospital stay was lower with corticosteroids as compared with standard treatment: mean difference - 24 hours (95% CI: -35 to -14). Corticosteroids were associated with an increased risk of hospital readmission as compared with standard treatment, in RCT, RCS, and the entire cohort: OR=5.91, 95% CI: 1.40-24.83; OR=3.28, 95% CI: 1.46-7.36 and OR=3.21, 95% CI: 1.97-5.24, respectively. Corticosteroids were associated with reduced number of transfusions and length of stay in RCT but not in RCS, with more rehospitalizations overall. Additional RCT should be conducted while minimizing the risk of rehospitalizations.

Introduction Sickle cell disease (SCD) is one of the most common monogenic disorders in the world. Acute painful episodes (vaso-occlusive crisis, VOC) and acute chest syndrome (ACS) represent the two most common acute events of this disease. Indeed, VOC is the primary cause of emergency department admissions and hospitalization1 while ACS is the main cause of death in adults.2,3 Erythrocyte abnormalities in SCD lead to microvascular occlusion and intravascular hemolysis, producing free hemoglobin and resulting in ischemia-reperfusion organ injury and infarction. These vaso-occlusive events promote inflammation and expression of adhesion receptor on endothelium cells.4,5

Corticosteroids reduce inflammation, by inhibiting cytokines and endothelial cell activation, reducing expression of P-selectin and vascular cell adhesion molecule-1 on endothelium.5–7 These effects could be useful to mitigate VOC and/or ACS, considering the potential benefit of inhibiting inflammatory response and intravascular hemolysis. Whether corticosteroids improve outcome (e.g., transfusion requirement or hospital length of stay) in patients with VOC or ACS is still debated. Many SCD specialists discourage the use of corticosteroids and report potential severe adverse events following corticosteroids therapy. Using a systematic review including a metaanalysis of randomized clinical trials (RCT) and retrospective cohort studies (RCS) may provide an accurate estimation of the potential benefits of corticosteroids in

Haematologica | 107 August 2022

1914

ARTICLE - Corticosteroids in sickle cell disease

J. Lopinto et al.

the course of acute complications in patients with SCD. The main objective of this study was to review the literature and provide a meta-analysis on the effects of corticosteroids on the clinical course of VOC or ACS in children and adults with SCD.

Methods Search strategy and selection criteria This systematic review was performed according to the PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analysis) guidelines.8 We searched CENTRAL and MEDLINE, from inception to 04/28/2021 using the search items “sickle cell disease,” “acute chest syndrome,” “vasoocclusive crisis,” “steroids,” “corticosteroids,” “methylprednisolone,” “prednisone,” “dexamethasone,” “prednisolone.” Type of studies, participants and outcomes We restricted our research to RCT and RCS comparing the effects of systemic corticosteroids with standard treatment. Inclusion criteria were as follows: population of children or adults, hospitalized for VOC or ACS, intervention involving systemic corticosteroids therapy during hospitalization (prednisone, methylprednisolone, prednisolone or dexamethasone). Exclusion criteria included the following: studies without a control group, studies assessing inhaled or topical corticosteroids therapy, comments and studies not written in English. The primary outcome of the review was transfusion requirement during hospitalization. References of all selected articles were scanned for additional relevant manuscripts. Ethical approval was not required. Data extraction See the Online Supplementary Appendix. Assessment of risk of bias (quality assessment) Two authors (JL, SG) independently assessed the study method quality by using the RoB 2.0 tool released through the Cochrane, to assess risk of bias in randomized studies.9,10 In non-randomized trials, we assessed the risk of bias using the ROBINS-I tool released through the Cochrane.11 In case of disagreement, a third researcher (AMD) adjudicated the assessment of risk of bias. The risk of bias was assessed according to preliminary considerations, the randomization process, effect of assignment to intervention, effect of starting and adhering to intervention, missing outcome data, and selection of the reported result. Data were presented using the robvis R package.12 Data analysis We estimated the effective sample size for each study based on their respective design effect. Data were sum-

marized using medians and interquartile ranges (IQR) or mean (standard deviation, SD) where appropriate.13 We adopted the inverse variance method for developing weights for individual study effects. We quantified heterogeneity using I² and Q statistics, with values greater than 50% regarded as being indicative of moderate-to-high heterogeneity.14 We used a random effect model to assess the population average mean difference and 95% IC of intervention. We did subgroup analyses of the type of trial included (RCT or RCS) and a sensitivity analysis focusing on ACS. Treatment effects in subgroups were compared by a test of interaction (using Cochran’s Q test and Higgins’ I2).15 In order to evaluate an outcome, we needed at least three studies that analyzed it with complete data. Individual study effects and pooled effects were visualized through forest plots. Publication biases were assessed graphically through traffic-light plot. Data were pooled using Review manager with a 2-sided significance of 5%. Study registration This study was submitted to the International Prospective Register of Systematic Reviews (PROSPERO) database (clinicaltrials gov. Identifier: CRD42021265528).

Results Selection of studies We identified 2,577 references from our searches (Figure 1). After removing 1,112 duplicates, we screened 1,465 titles of whom 1,437 were excluded. We then screened 28 abstracts, among which we identified nine potentially eligible articles. Among the 19 studies excluded by abstract, there were seven systematic reviews not eligible for analysis because they did not provide original data and/or did not address the question of systemic corticosteroids therapy in the field of VOC and/or ACS. Full texts of nine studies were reviewed and three were excluded because of the absence of a control group (n=2)16,17 or absence of comparative analysis between corticosteroids and a control group (n=1).18 Finally, six studies were included in the meta-analysis, involving 3,304 participants and 5,562 VOC or ACS episodes. Clinical definitions VOC was defined as pain or tenderness affecting at least one part of the body, requiring opioids, and not explained by other causes.19 ACS was defined as an acute illness with fever and/or respiratory symptoms, accompanied by a new pulmonary infiltrate on chest X-ray.20–24 Severe ACS was defined by the need for intensive care unit admission,24 or the presence of one of the following severity signs, involving the neurologic system (e.g., lethargy) or the respiratory system (e.g., extensive pulmonary infiltrates [bilateral, one

Haematologica | 107 August 2022

1915

ARTICLE - Corticosteroids in sickle cell disease

J. Lopinto et al.

Figure 1. Flow chart. ACS: acute chest syndrome; VOC: vaso-occlusive crisis; RCT: randomized controlled trial; RCS: retrospective cohort studies.

complete lung or three lobes], marked arterial hypoxemia [transcutaneous oxygen saturation (SpO2) <85-90% despite supplemental oxygen, tachypnea and SpO2 <95% in room air], need for invasive or non-invasive respiratory support).20–23 Study design Among the included studies, three were RCT19–21 and three were RCS22–24. Studies were held in United State of America; they were multi-center (n=3)21,23,24 or took place in a single-center of a department of pediatric hematology (n=3).19,20,22. Studies evaluated mostly the pediatric population (n=5);19,20,22–24 one assessed both the adult and pediatric population.21 All studies reported corticosteroids impact in the field of ACS (n=5)20–24 or VOC (n=1)19 (Table 1). In the three RCT, corticosteroid therapy was evaluated versus placebo in a controlled double-blinded fashion; corticosteroids and placebo groups were comparable in terms of demographics, clinical and biological characteristics at baseline.19–21 On the contrary, in each of the three RCS included, several baseline clinical characteristics were significantly different between patients receiving steroids and their counterparts. There was no severe ACS

included in RCT, while the rate of severe ACS in RCS was higher in patients receiving steroids as compared with their counterparts (Online Supplementary Table S1). Quality Included studies differed in their methodological quality (Online Supplementary Figures S1). High risk of bias was related to confounding factors and classification of interventions in the RCS.22–24 Intervention All studies compared corticosteroids with a control group of standard of care. Standard of care was comparable among all studies and was in accordance with current guidelines:25 bed rest, supplemental oxygen, intravenous hydration, folate supplementation, analgesics and blood products transfusion, as needed.19–24 The intervention protocols were heterogeneous regarding corticosteroids class, dosing and duration of therapy (Table 2). Corticosteroid therapy consisted of dexamethasone (n=2),20,21 methylprednisolone (n=1),19 prednisone (n=1)23 or several classes of corticosteroids (including dexamethasone, prednisone, prednisolone and methylprednisolone) (n=2).22,24 Duration of corticosteroid therapy

Haematologica | 107 August 2022

1916

ARTICLE - Corticosteroids in sickle cell disease

J. Lopinto et al.

Table 1. Characteristics of included trials.

Trial, year

Griffin, 199419

Bernini, 199820 Strouse, 200822 Moderate acute chest syndrome

Inclusion criteria

Vaso-occlusive crises

Type of study

Randomized Randomized controlled double-blind controlled doubleprotocol blind protocol

Sobota, 200924

Kumar, 201023

Quinn, 201121

Acute chest syndrome

Retrospective cohort

Randomized Retrospective controlled doublecohort blind protocol

Recruitment period

1990-1991

1992-1995

1998-2004

2004-2008

2005-2007

2006-2008

Participants (number of episodes)

36 (56)

38 (43)

65 (126)

3,090 (5,247)

63 (78)

12 (12)

Readmission rate

Duration of acute chest syndome

Primary outcome

Doses of intravenousmorphine Readmission rate Length of hospital Length of hospital Continuous intraLength of hospital stay stay venous morphine stay Readmission rate Duration of analgesia, Clinical complications

d: day.

Table 2. Characteristics of corticosteroid protocol in included studies.

Study, year

Corticosteroid class (dose)

Equivalent dose of Prednisone, mg/kg/d

Griffin, 199419

Bernini, 199820

Strouse, 200822

Sobota, 200924

Kumar, 201023

Quinn, 201121

Methylprednisolone (15 mg/kg/d)

Dexamethasone (0.6 mg/kg/d)

Dexamethasone (0.6 mg/kg/d) or Prednisone (1 to 2 mg/kg/d) or Prednisolone (2 mg/kg/d)

Dexamethasone (NI) or Prednisone (NI) or Prednisolone (NI) or Methylprednisolone (NI)

Prednisone (2 mg/kg/d)

Dexamethasone (0.6 mg/kg/d)

18.75

1 to 4

1 to 6

Treatment duration, days NI: no information; d: day.

varied from 1 to 8 days. Details on the intervention protocol was not available in one study.24 Outcomes Variables reported for each study as primary outcomes included hospital readmission rate (n=1),23 length of hospital stay (n=1),20 duration of ACS (n=1)21 and composite outcomes (n=3)19,22,24 (Table 1). Outcome data were completely available for transfusion requirement (n=6), and readmission rate (72 hours, n=220,24 or 2 weeks, n=419,21–23 after hospital discharge), but not for length of hospital stay (n= 3).19–21 Data on volumes transfused and mortality

were never reported in included studies. We could not evaluate other outcomes (opioids doses and duration of analgesic therapy, incidence rate and duration of oxygen therapy, delay for hospital readmission) because less than three studies reported each of these outcomes with complete data. Effect of intervention Transfusion requirement (Figure 2)

In the subgroup of RCT, corticosteroid therapy was associated with a lower risk of transfusion as compared with standard treatment: OR=0.13 (95% CI: 0.04-0.45;

Haematologica | 107 August 2022

1917

ARTICLE - Corticosteroids in sickle cell disease

J. Lopinto et al.

I²=0%).19–21 On the contrary, in the subgroup of RCS, corticosteroids therapy was associated with a higher risk of transfusion: OR=2.12 (95% CI: 1.33-3.40; I²= 33%).22–24 When we pooled all included trials, there was no difference between corticosteroid therapy and standard treatment regarding transfusion: OR=0.98 (95% CI 0.382.53; I²= 75%).19–24 We found a significant interaction between the type of study (RCT or RCS) and the risk of transfusion (Cochran’s Q test: c2= 17.10, df=1, P<0.0001, Higgins’s I²=94.2%, Figure 2). Length of hospital stay (Figure 3)

In the three RCT with available data, the length of hospital stay was lower with corticosteroids as compared with standard treatment: mean difference - 24 hours (95% CI: -35 to -14; I²= 0%).19–21

spectively. Tests of interaction between the study type (RCT or RCS) and the effect on hospital readmission did not evidence a significant interaction between subgroups (Cochran’s Q test: c2= 0.49, df=1, P=0.48, Higgins’s I²=0%, Figure 4). The main reason for readmission was painful crisis recurrence (Online Supplementary Table S2). Sensitivity analysis A sensitivity analysis was performed including only studies on ACS.20–24 Results were similar to those observed in the main analysis (Online Supplementary Figure S2; Online Supplementary Table S3). We also performed a sensitivity analysis including only the pediatric population.19,20,22–24 The results were also similar to those observed in the main analysis (Online Supplementary Figure S3; Online Supplementary Table S4).

Hospital readmission (Figure 4)

Corticosteroids therapy was associated with a significantly increased risk of hospital readmission as compared with standard treatment, in RCT, RCS, and the entire cohort: OR=5.91 (95% CI: 1.40-24.83; I²=0%)19–21; OR=3.28 (95% CI: 1.467.36; I²=56%)22–24; and 3.21 (95% CI: 1.97-5.24, I²=13%)19–24, re-

Discussion Several systematic reviews have been published over the past years in the field of SCD to assess analgesic intervention26–28 or transfusion therapy efficacy29 in the setting of

Figure 2. Effect of corticosteroids on transfusion requirement. CI: confidence interval.

Figure 3. Effect of corticosteroid therapy on the length of hospital stay. SD: standard devaition; CI: confidence interval. Haematologica | 107 August 2022

1918

ARTICLE - Corticosteroids in sickle cell disease

J. Lopinto et al.

VOC or ACS. Two recent studies reported a descriptive review, assessing available evidence in favor and against the use of corticosteroids therapy, but without metaanalysis.30,31 Our work is, to the best of our knowledge, the first meta-analysis on the impact of corticosteroids on the clinical course of acute complications in patients with SCD. The use of corticosteroids was associated with i) a decrease in the need for transfusion in RCT, but not in RCS, with a significant interaction between the outcome and the type of study and ii) an increase in hospital readmission rates both in RCT and RCS. Conflicting effects on transfusion requirement Our analysis on transfusion requirement reported conflicting results between RCT and RCS. Indeed, it appears that corticosteroids reduce transfusion requirement in RCT whereas they were associated with more transfusion requirement in RCS, with a significant interaction between the outcome and study subgroups. One hypothesis for this discrepancy could be the absence of comparable groups in retrospective cohorts. Indeed, differences in baseline severity are potential confounders in RCS (Online Supplementary Table S1). Thus, results from retrospective cohorts could reflect the worse baseline severity of patients having received corticosteroids therapy (inasmuch as exchange transfusion is recommended in case of VOC or ACS with severity criteria in some guidelines25). Another hypothesis about conflicting results between RCT and RCS could be the timing between corticosteroids and transfusion: whereas steroids where administered before transfusion in RCT, they were given after transfusion in some patients in RCS.22 The association between steroid use and increased

transfusion requirement in RCS should be interpreted with caution and may reflect the rescue use of steroids in more severe patients. By contrast, the subgroup analysis of the three RCT showed a significant reduction of transfusion requirement in the corticosteroids group. This result is in accordance with the reduction in length of hospital stay observed in the corticosteroid group in these trials. By inhibiting cytokine production and endothelial activation, corticosteroids could impede the inflammatory cascade, reducing the need for exchange transfusion. The beneficial effects of corticosteroids is reported in other forms of vascular and lung injury.32 Rebound effect and hospital readmission A risk of rebound pain was described in previous reports of corticosteroids in SCD.18–24 This legitimate concern led physicians over the past years to restrict corticosteroids administration to hospitalized patients with SCD and comorbid asthma.18,24 Despite the heterogeneous design of available studies, our pooled analysis showed that corticosteroids administered during VOC or ACS increased the risk for readmission within 72 hours to 2 weeks after hospital discharge.19–24 Interestingly, results from RCT and RCS are concordant for this outcome, with no interaction between subgroups. Several case reports described the poor tolerance of corticosteroids in SCD patients, with a high frequency of pain recurrence or relapse after withdrawal.33,34 One hypothesis could be a rebound upregulation of vascular cell adhesion molecule-1 on endothelium and delayed leukocytosis after corticosteroids withdrawal that can lead to VOC recurrence.5,35 Whether specific dose de-escalation protocols or adjuvant anti-inflammatory

Figure 4. Effect of corticosteroid therapy on readmission rate. CI: confidence interval.

Haematologica | 107 August 2022

1919

ARTICLE - Corticosteroids in sickle cell disease

J. Lopinto et al.

therapies might reduce the rebound upregulation induced by corticosteroids warrants further research.16,17 Exchange transfusion alters inflammation during SCD crisis,36 and its systematic association with corticosteroids might reduce the upregulation of inflammatory cascade observed after corticosteroids therapy withdrawal,5,35 and therefore the associated risk of rebound pain and readmission. Indeed, some authors observed a lack of rebound pain when dexamethasone was systematically associated with exchange transfusion in patients presenting severe ACS16 or mild VOC;17 the decrease of sickle red blood cell percentage, as a consequence of red cell exchange transfusion, might reduce the risk of rebound pain. Of note, three episodes of cerebral complication were reported in included studies (one stroke20 and two intracranial bleedings22), all in patients having received corticosteroids. The use of corticosteroids has already been associated with the occurrence of intracranial bleeding.37 Although the basis for this potential association remains unclear (e.g., increase of systolic blood pressure induced by corticosteroids might be poorly tolerated in SCD patients), physicians should consider this possible complication, notably in patients with other risk factors for intracranial bleeding.37 Strengths and limitations We conducted a comprehensive and exhaustive literature search for comparative trials on corticosteroids use in the field of ACS or VOC. However, our study has several limitations. First, our systematic review found that trials assessing corticosteroids in the field of the two main acute complications of SCD are rare. Indeed, our research revealed only three RCT, with heterogeneous populations and intervention protocols,19–21 and three RCS with several inherent confounding bias.22–24 Second, results were divergent concerning the efficacy of corticosteroids, with steroids associated with reduced transfusion in RCT and more transfusions in RCS. However, we could scrutinize these conflicting results (subgroup interaction) and suggest a likely role for baseline severity in RCS. The association of steroid use with increased transfusions in RCS is complex to interpret because steroid use did not always precede transfusion in all patients,22 excluding a causal effect. Third, the populations and corticosteroids protocols were different among included studies: only one study assessed SCD patients presenting VOC19 and only one study assessed adult population.21 Nonetheless, our sensitivity analyses in patients with ACS and in the pediatric population were similar to those observed in the main analysis. Although they share some pathophysiological features, VOC and ACS are different primary diseases and these acute complications of SCD may have specific features in adult versus pediatric patients. Therefore, further studies are needed especially in patients with VOC and in the adult population. Fourth, we could not evalu-

ate length of hospital stay in all included trials due to the lack of complete data in the retrospective cohorts. We nonetheless observed that corticosteroids reduced the length of hospital stay in RCT, along with reduced transfusion requirement. Last, we did not evaluate corticosteroid effects on the onset of pain crisis in patients free of VOC/ACS; indeed the demargination process induced by corticosteroids in non-hyperleukocytic patients could play a specific role. Similarly, although none of the included studies mentioned delayed hemolysis transfusion reaction, we cannot formally exclude an indication bias in patients with a medical history of delayed hemolysis transfusion reaction. Conclusion and future research In conclusion, as compared with standard care, corticosteroids administered to patients with VOC or ACS reduced the length of hospital stay and the need for transfusion in RCT but were associated with more transfusions in RCS. Corticosteroids increased the risk of readmission, both in RCT and RCS. Given the small number of included studies, the lack of data on volumes transfused, the presence of confounding bias in retrospective cohorts and the high heterogeneity of our analysis, we could not give any recommendation for the use of corticosteroids to treat VOC or ACS in patients with SCD. Considering the potential benefit of corticosteroids, in particular in reducing the length of hospital stay, further prospective studies should be conducted. However, the risk of readmission associated with corticosteroids withdrawal must be carefully considered and anticipated when using corticosteroids in these trials. Disclosures No conflicts of interest to disclose. Contributions JL, SG, EB and AMD collected the data; JL, SG and AMD analyzed and interpreted the data; JL, SG and AMD drafted the manuscript; JL and AMD contributed to the study conception and design; AMD, AH and PB critically revised the manuscript. All authors read and approved the final manuscript. Acknowledgments We thank Timothy C. Griffin, author of one of the primary studies, who kindly provided supplementary information and data included in our systematic review and metaanalysis. Data-sharing statement The datasets and materials used and analyzed during the current study are available from the corresponding author on reasonable request.

Haematologica | 107 August 2022

1920

ARTICLE - Corticosteroids in sickle cell disease

J. Lopinto et al.

References 1. Ballas SK, Lusardi M. Hospital readmission for adult acute sickle cell painful episodes: frequency, etiology, and prognostic significance. Am J Hematol. 2005;79(1):17-25. 2. Booth C, Inusa B, Obaro SK. Infection in sickle cell disease: a review. Int J Infect Dis IJID Off Publ Int Soc Infect Dis. 2010;14(1):e2-e12. 3. Vichinsky EP, Neumayr LD, Earles AN, et al. Causes and outcomes of the acute chest syndrome in sickle cell disease. National Acute Chest Syndrome Study Group. N Engl J Med. 2000;342(25):1855-1865. 4. Kato GJ, Steinberg MH, Gladwin MT. Intravascular hemolysis and the pathophysiology of sickle cell disease. J Clin Invest. 2017;127(3):750-760. 5. Belcher JD, Mahaseth H, Welch TE, et al. Critical role of endothelial cell activation in hypoxia-induced vasoocclusion in transgenic sickle mice. Am J Physiol-Heart Circ Physiol. 2005;288(6):H2715-H2725. 6. Okpala I. Leukocyte adhesion and the pathophysiology of sickle cell disease. Curr Opin Hematol. 2006;13(1):40-44. 7. Stuart MJ, Setty BN. Sickle cell acute chest syndrome: pathogenesis and rationale for treatment. Blood. 1999;94(5):1555-1560. 8. Moher D, Liberati A, Tetzlaff J, Altman DG, Group TP. Preferred reporting items for systematic reviews and meta-analyses: The PRISMA Statement. PLoS Med. 2009;6(7):e1000097. 9. Savović J, Weeks L, Sterne JAC, et al. Evaluation of the Cochrane Collaboration’s tool for assessing the risk of bias in randomized trials: focus groups, online survey, proposed recommendations and their implementation. Syst Rev. 2014;3:37. 10. Page MJ, McKenzie JE, Higgins JPT. Tools for assessing risk of reporting biases in studies and syntheses of studies: a systematic review. BMJ Open. 2018;8(3):e019703. 11. Sterne JA, Hernán MA, Reeves BC, et al. ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions. BMJ. 2016;355:i4919. 12. McGuinness LA, Higgins JPT. Risk-of-bias VISualization (robvis): An R package and Shiny web app for visualizing risk-of-bias assessments. Res Synth Methods. 2021;12(1):55-61. 13. Wan X, Wang W, Liu J, Tong T. Estimating the sample mean and standard deviation from the sample size, median, range and/or interquartile range. BMC Med Res Methodol. 2014;14:135. 14. Higgins JPT, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med. 2002;21(11):1539-1558. 15. Sedgwick P. Meta-analyses: heterogeneity and subgroup analysis. BMJ. 2013;346:f4040. 16. Isakoff MS, Lillo JA, Hagstrom JN. A single-institution experience with treatment of severe acute chest syndrome: lack of rebound pain with dexamethasone plus transfusion therapy. J Pediatr Hematol Oncol. 2008;30(4):322-325. 17. Yeral M, Boga C, Aytan P, Ozdogu H. Corticosteroid-induced vaso-occlusive events may be prevented by lowering hemoglobin S levels in adults with sickle cell disease. Transfus Apher Sci. 2017;56(5):717-718. 18. Sobota A, Graham DA, Neufeld EJ, Heeney MM. Thirty-day readmission rates following hospitalization for pediatric sickle cell crisis at freestanding children’s hospitals: Risk factors and hospital variation. Pediatr Blood Cancer. 2012;58(1):61-65. 19. Griffin TC, McIntire D, Buchanan GR. High-dose intravenous methylprednisolone therapy for pain in children and adolescents with sickle cell disease. N Engl J Med.

1994;330(11):733-737. 20. Bernini JC, Rogers ZR, Sandler ES, Reisch JS, Quinn CT, Buchanan GR. Beneficial effect of intravenous dexamethasone in children with mild to moderately severe acute chest syndrome complicating sickle cell disease. Blood. 1998;92(9):3082-3089. 21. Quinn CT, Stuart MJ, Kesler K, et al. Tapered oral dexamethasone for the acute chest syndrome of sickle cell disease: Short Report. Br J Haematol. 2011;155(2):263-267. 22. Strouse JJ, Takemoto CM, Keefer JR, Kato GJ, Casella JF. Corticosteroids and increased risk of readmission after acute chest syndrome in children with sickle cell disease: acute chest syndrome in sickle cell disease. Pediatr Blood Cancer. 2008;50(5):1006-1012. 23. Kumar R, Qureshi S, Mohanty P, Rao SP, Miller ST. A short course of prednisone in the management of acute chest syndrome of sickle cell disease. J Pediatr Hematol Oncol. 2010;32(3):e91-94. 24. Sobota A, Graham DA, Heeney MM, Neufeld EJ. Corticosteroids for acute chest syndrome in children with sickle cell disease: variation in use and association with length of stay and readmission. Am J Hematol. 2010;85(1):24-28. 25. Habibi A, Arlet J-B, Stankovic K, et al. [French guidelines for the management of adult sickle cell disease: 2015 update]. Rev Med Interne. 2015;36(5 Suppl 1):5S3-84. 26. Cooper TE, Hambleton IR, Ballas SK, Johnston BA, Wiffen PJ. Pharmacological interventions for painful sickle cell vasoocclusive crises in adults. Cochrane Database Syst Rev. 2019;2019(11):CD012187. 27. Dunlop RJ, Bennett KCLB. Pain management for sickle cell disease. Cochrane Database Syst Rev. 2006;(2):CD003350. 28. Meremikwu MM, Okomo U. Sickle cell disease. BMJ Clin Evid. 2016;2016:2402. 29. Dolatkhah R, Dastgiri S. Blood transfusions for treating acute chest syndrome in people with sickle cell disease. Cochrane Database Syst Rev. 2020;1(1):CD007843. 30. Ogunlesi F, Heeney MM, Koumbourlis AC. Systemic corticosteroids in acute chest syndrome: friend or foe? Paediatr Respir Rev. 2014;15(1):24-27. 31. Vandy Black L, Smith WR. Evidence-based mini-review: are systemic corticosteroids an effective treatment for acute pain in sickle cell disease? Hematol Am Soc Hematol Educ Program. 2010;2010(1):416-417. 32. RECOVERY Collaborative Group, Horby P, Lim WS, et al. Dexamethasone in hospitalized patients with Covid-19 preliminary report. N Engl J Med. 2021;384(8):693-704. 33. Couillard S, Benkerrou M, Girot R, Brousse V, Ferster A, BaderMeunier B. Steroid treatment in children with sickle-cell disease. Haematologica. 2007;92(3):425-426. 34. Darbari DS, Fasano R s., Minniti CP, et al. Severe Vaso-occlusive episodes associated with use of systemic corticosteroids in patients with sickle cell disease. J Natl Med Assoc. 2008;100(8):948-951. 35. Mishler JM, Emerson PM. Development of neutrophilia by serially increasing doses of dexamethasone. Br J Haematol. 1977;36(2):249-257. 36. Liem RI, O’Gorman MR, Brown DL. Effect of red cell exchange transfusion on plasma levels of inflammatory mediators in sickle cell patients with acute chest syndrome. Am J Hematol. 2004;76(1):19-25. 37. Strouse JJ, Hulbert ML, DeBaun MR, Jordan LC, Casella JF. Primary hemorrhagic stroke in children with sickle cell disease is associated with recent transfusion and use of corticosteroids. Pediatrics. 2006;118(5):1916-1924.

Haematologica | 107 August 2022

1921

LETTER TO THE EDITOR

m6A reader Ythdf3 protects hematopoietic stem cell integrity under stress by promoting the translation of Foxm1 and Asxl1 transcripts N6-methyladenosine (m6A), which is the most abundant internal modification of eukaryotic mRNA, is dynamically regulated by RNA methyltransferases (writers) and demethylases (erasers).1 Many RNA binding proteins, including the YT521-B homology (YTH) domain family of proteins (YTHDF1, 2, 3, YTHDC1, 2) and insulin-like growth factor 2 mRNA-binding proteins (IGF2BP1, 2, 3), which can recognize unique m6A-modified RNA and mediate its function, are characterized as “readers”.1 The fate of m6A-marked RNA is determined by different readers through binding to alternative locations of m6A modifications. m6A methylation adds another layer of post-transcriptional regulation of gene expression during normal and malignant hematopoiesis.2 YTHDF family proteins display high similarity in protein structure with a YTH domain near the N-terminus and contain a low complexity structure that includes a prion-like domain.3 Previous studies showed that YTHDF proteins have unique functions in RNA metabolism with some overlapping functions. YTHDF1 facilitates the translation of target transcripts by interacting with translation initiation factor complex3 (elF3), while YTHDF2 modulates mRNA decay by recruiting the CCR4-NOT deadenylase complex, and YTHDF3 cooperates with YTHDF1 and YTHDF2 to facilitate the translation and decay of target mRNA.1,3 However, there are conflicting data. Two recent reports described a different model of YTHDF function, in which YTHDF1, 2, and 3 bind the same m6A-modified mRNA and function together to regulate mRNA stability in HeLa and mouse embryonic stem cells.4,5 Analysis of publicly available RNA-sequencing data of human CD34+ stem/progenitor cells6 revealed that YTHDF2 has a relatively higher level of expression compared with YTHDF1 and YTHDF3, which have comparable levels of expression (Figure 1A). Knockout of Ythdf2 results in the increase of functional hematopoietic stem cells (HSC) in young mice,7,8 but reduces HSC self-renewal upon aging by stimulating pro-inflammatory pathways.9 However, the role of Ythdf3 in hematopoiesis remains undetermined. To elucidate the function of Ythdf3 in vivo, we generated and characterized a Ythdf3 knockout mouse model, in which exon 3 of Ythdf3 was deleted using CRISPR-Cas9 mediated genome editing technology (Figure 1B). The deletion of the Ythdf3 allele was validated by polymerase chain reaction (PCR) analysis of genomic DNA from mouse tails (Figure 1C). Ythdf3 deletion led to the absence of Ythdf3 expression but did not affect the expression of

Ythdf1 and Ythdf2 in c-Kit+ mouse bone marrow cells (Online Supplementary Figure S1A). Ythdf3-deficient mice were viable and displayed normal hematopoiesis, as evidenced by normal white blood cell, platelet, and red blood cell counts and hemoglobin level in the peripheral blood compared with those of control littermates at 2-3 months of age (Online Supplementary Figure S1B). Both Ythdf3 wildtype and knockout mice had comparable bone marrow cellularity (Online Supplementary Figure S1C). As determined by flow cytometric analysis, the lineage distribution of mature myeloid cells, B cells, T cells, and red cells was comparable in bone marrow, spleen, or thymus in a cohort of Ythdf3-knockout mice and control mice (Online Supplementary Figure S1D, E). We further characterized the hematopoietic stem and progenitor cell compartments in these mice. As shown in Online Supplementary Figure S1F and Figure 1D, there were no differences in the frequency and the absolute number of the HSC-enriched populations LSK (Lin–Sca1+c-Kit+), multipotent progenitor (MPP), short-term HSC (ST-HSC), and long-term HSC (LT-HSC, CD150+CD48–Lin–Sca1+c-Kit+) in bone marrow between 2- to 3-month-old Ythdf3 knockout and control mice. Additionally, we found that apoptosis and quiescence of HSC were not disturbed by Ythdf3 loss (Figure 1E, F). The frequency and number of hematopoietic progenitor cells (HPC) and subpopulations of myeloid progenitors, including common myeloid progenitors, megakaryocyte and erythroid progenitors, and granulocyte and macrophage progenitors, were not changed as a consequence of Ythdf3 depletion (Online Supplementary Figure S1G). These results suggest that Ythdf3 is dispensable for normal hematopoiesis and the maintenance of HSC/HPC during steady-state conditions. To determine the role of Ythdf3 in HSC under stress, we performed a bone marrow transplantation assay. More than 90% of donor-derived peripheral blood cells and bone marrow cells were detected in both Ythdf3-deficient and control recipient mice in the first and fourth months after transplantation, respectively (Online Supplementary Figure S1H), suggesting a comparable engraftment ability of Ythdf3-deficient and control HSC/HPC. Further characterization of Ythdf3-deficient and control recipient mice revealed there was a decrease of bone marrow cellularity and frequency of LSK (Online Supplementary Figure S1I) while the frequencies of MPP, ST-HSC and LT-HSC in the LSK population (Online Supplementary Figure S1J) were

Haematologica | 107 August 2022

1922

LETTER TO THE EDITOR comparable in Ythdf3-deficient recipient mice and control recipients. The total numbers of LSK, MPP, and ST-HSC but not LT-HSC were decreased in Ythdf3-deficient recipient mice compared with the numbers in control recipients (Figure 1G). However, there were no differences in mature cell lineage distribution (Online Supplementary Figure S1K,

L) between Ythdf3-deficient and control recipient mice. The frequency of apoptosis was significantly increased in Ythdf3-deficient HSC and LSK compared with control cells (Figure 1H). To determine the effects of Ythdf3 depletion on hematopoietic reconstitution in situ, a cohort of Ythdf3-deficient and control recipient mice were injected

Figure 1. Ythdf3 knockout reduces hematopoietic stem cell self-renewal and survival under regenerative stress. (A) Expression of YTHDF1, YTHDF2, and YTHDF3 mRNA in CD34+ cells from healthy individuals. (B) Diagram showing the strategy for generating a mouse with the knockout (KO) of Ythdf3. (C) Polymerase chain reaction genotyping of tail genomic DNA from mice generated by heterozygote (Het) crosses. (D) The absolute number of Lin–Sca1+c-Kit+ cells (LSK), multipotent progenitors (MPP), short-term (ST)-hematopoietic stem cells (HSC), and long-term (LT)-HSC in bone marrow (BM) from wild-type (WT) and Ythdf3 KO mice. (E) Flow cytometric analysis of the frequency of apoptotic cells in the HSC population from 8-week-old WT and Ythdf3 KO mice (n=4-5). (F) Flow cytometric analysis of cell cycle in the HSC population from 8-week-old WT and Ythdf3 KO mice (n=4-5). (G) Total LSK, MPP, ST-HSC, and LT-HSC in BM from WT and Ythdf3 KO recipient mice. The mice were analyzed 4 months after the first round of transplantation (mean ± standard deviation). (H) Frequency of apoptosis in gated Lin- cells, HPC, LSK, ST-HSC and LT-HSC stained with annexin V and DAPI from Ythdf3 WT or KO recipient mice. (I) Kaplan-Meier curve representing percent survival of Ythdf3 WT (n=7) and KO (n=8) mice over time after weekly injections of 5-fluorouracil (5-FU) for three times. *P<0.05, **P<0.01, ***P<0.001. Haematologica | 107 August 2022

1923

LETTER TO THE EDITOR

Figure 2. Ythdf3 loss reduces the reconstitution ability of hematopoietic stem and progenitor cells. (A) Schematic diagram of the experimental strategy. (B) Ratio of donor-derived CD45.2+CD45.1- peripheral blood cells (Donor/Competitor) in chimeric CD45.1+ C57BL/6 wild-type (WT) mice reconstituted with bone marrow cells from Ythdf3 WT or knockout (KO) mice, assessed every month for 4 months after primary transplantation (1-1, 1-2, 1-2, 1-4) or secondary transplantation (2-1, 2-2, 2-3, 2-4). (C-E) Evaluation of the ratio of donor-derived vs. competitor-derived mature lineage cells in peripheral blood after transplantation. (F, G) Ratio of donor-derived Lin- cells, hematopoietic progenitor cells (HPC), Lin–Sca1+c-Kit+ (LSK), and hematopoietic stem cells (HSC) as well as myeloid and B cells in bone marrow from indicated mice. *P<0.05, **P<0.01, ***P<0.001.

with 5-fluorouracil, which kills proliferating hematopoietic progenitors, thus promoting HSC to enter the cell cycle and reconstitute the hematopoietic system. As shown in Figure 1I, all Ythdf3-deficient mice died within 20 days, while 25% of control mice survived after three weekly treatments of 5-fluorouracil, indicating that Ythdf3-deficient HSC lost the ability to replenish the hematopoietic system more quickly than wildtype HSC did. To further evaluate the long-term self-renewal capacity of Ythdf3deficient HSC in a competitive situation, we performed competitive serial transplantation assays (Figure 2A). An equal number of bone marrow cells (CD45.2) from Ythdf3deficient or control mice along with competitor wildtype bone marrow cells (CD45.2/CD45.1) were transplanted into

lethally irradiated syngeneic recipients (CD45.1). Thereafter, we performed secondary transplants with mixed bone marrow cells from two or three mice in each group. The frequency of donor-derived peripheral blood cells was monitored monthly after transplantation. The ratio of Ythdf3-/-derived vs. competitor-derived total peripheral blood cells was gradually decreased significantly by approximately 2-fold compared with the ratio of Ythdf3+/+ derived vs. competitor-derived peripheral blood cells after secondary transplantation (Figure 2B, Online Supplementary Figure S2A). Of note, Ythdf3-/- hematopoietic stem/progenitor cells (HSPC) produced a lower number of all mature cells, including myeloid cells, B cells, and T cells, compared with Ythdf3+/+ HSPC (Figure 2C-E, Online

Haematologica | 107 August 2022

1924

LETTER TO THE EDITOR Supplementary Figure S2B-D). Consistent with the results from peripheral blood, Ythdf3-/- donor-derived Lin- cells, HPC, LSK, and HSC as well as myeloid and B cells in bone

marrow were significantly lower compared with Ythdf3+/+ donor-derived cells 4 months after secondary transplantation (Figure 2F, G). Collectively, these results indicate

Figure 3. Ythdf3 facilitates mRNA translation of Foxm1 and Asxl1 in hematopoietic stem and progenitor cells. (A) m6A RNA immunoprecipitation (MeRIP) analysis of mRNA m6A methylation of Foxm1 and Asxl1 in mouse hematopoietic precursor cell-7 (HPC-7) cells. (B) Ythdf3 RIP analysis suggesting that Ythdf3 binds directly to mRNA of Foxm1 and Asxl1 in HPC-7 cells. (C) Reverse transcriptase quantitative polymerase chain reaction (RT-qPCR) analysis showing the knockdown efficiency of Ythdf3 in HPC-7 cells. (D) RT-qPCR analysis indicating that levels of Foxm1 and Asxl1 transcription are not affected by Ythdf3 knockdown in HPC7 cells. (E) RT-qPCR analysis showing transcription levels of Foxm1 and Asxl1 in wild-type and Ythdf3 knockout c-Kit+ mouse HSPC. (F, G) Western blot analysis showing the effect of Ythdf3 knockdown (F), or knockout (G) on protein levels of Foxm1 and Asxl1 in HPC-7 cells (F), or c-Kit+ mouse HSPC (G), respectively. (H, I) Polysome profiling analysis indicating that Ythdf3 depletion significantly inhibits mRNA translation of Foxm1 and Asxl1. (J) eIF3A RIP analysis showing the effect of Ythdf3 knockdown on eIF3A enrichment at mRNA of Foxm1 and Asxl1 in HPC-7 cells. (K) Western blot analysis showing protein levels of Foxm1 and Asxl1 in wild-type or Ythdf3 depleted HPC-7 cells with or without 4EGI-1 treatment. All the statistical analyses were performed by GraphPad Prism 5. *P<0.05; **P<0.01; ***P<0.001; ns: no significant difference. Haematologica | 107 August 2022

1925

LETTER TO THE EDITOR that Ythdf3-deficient HSC had decreased self-renewal capacity and survival under stress conditions. To investigate the molecular mechanisms underlying the function of Ythdf3 in HSPC, we analyzed the publicly available YTHDF3 cross-linking immunoprecipitation-sequencing data in human cell lines.10 Both Foxm1 and Asxl1, which play important roles in HSC maintenance,11,12 are potential direct binding targets of Ythdf3. We performed m6A RNA immunoprecipitation (Me-RIP)-PCR analysis and Ythdf3 RIP-PCR analysis of Foxm1 and Asxl1 in HPC-7 cells, multipotent hematopoietic precursors.13 The results showed significant enrichment of m6A antibody and Ythdf3 binding sites in both Foxm1 and Asxl1 transcripts (Figure 3A, B). More importantly, only the wildtype Ythdf3 but not the m6A-binding-deficient Ythdf3-W495A binds to transcripts of Foxm1 and Asxl1, suggesting that Ythdf3 regulates Foxm1 and Asxl1 expression by direct binding in an m6A-dependent manner (Online Supplementary Figure S3A, B). We employed a recently reported FAMSi system,14 in which siRNA oligos against the target genes are expressed in a retroviral backbone vector, to express Ythdf3-specific siRNA and scramble siRNA in HPC-7 cells. Quantitative real-time PCR analysis revealed that both Ythdf3-specific siRNA significantly knocked down Ythdf3 expression (Figure 3C) but did not affect the expression of Foxm1 and Asxl1 transcripts in HPC-7 cells (Figure 3D). Consistently, Ythdf3 knockout did not affect Foxm1 and Asxl1 mRNA expression in primary cKit+ HSPC from mice (Figure 3E). Of note, as determined by western blot analysis, Ythdf3 knockdown or knockout significantly inhibited Foxm1 and Asxl1 protein expression in HPC-7 cells and primary c-Kit+ HSPC from mice, respectively (Figure 3F, G). Notably, Ythdf3-depletion-mediated downregulation of Foxm1 and Asxl1 was rescued by wild-type but not the m6A-binding-deficient Ythdf3W495A, suggesting that Ythdf3-mediated regulation of Foxm1 and Asxl1 expression is dependent on its m6A RNA binding activity (Online Supplementary Figure S3C). Protein abundance is determined by a balance between protein turnover and protein synthesis, namely mRNA translation. The half-lives of Foxm1 and Asxl1 proteins were not affected by Ythdf3 depletion (Online Supplementary Figure S3D-I), indicating that Ythdf3 depletionmediated downregulation of Foxm1 and Asxl1 protein is not attributed to protein turnover. Meanwhile, we performed polysome profiling analysis and observed that Ythdf3 deletion led to a significant decrease of Foxm1 and Asxl1 transcripts in polysome fractions and an increase of these transcripts in the non-ribosome fraction (Figure 3H, I), suggesting that Ythdf3 regulates Foxm1 and Asxl1 mRNA translation. elF3a facilitates the initiation of translation by directly binding to m6A in the 5’-untranslated region.15 elF3A RIP-PCR analysis showed that Ythdf3 knockdown markedly reduced the binding of elf3a to Foxm1 and Asxl1 transcripts (Figure 3J). We treated HPC-

7 cells stably expressing vector, or Ythdf3-specific siRNA with or without 4EGI-1, an inhibitor of Cap-dependent translation initiation by disruption of the eIF4E/eIF4G complex formation. Both Ythdf3 deletion and 4EGI-1 treatment led to downregulation of Foxm1 and Asxl1 (Figure 3K). More importantly, protein levels of Foxm1 and Asxl1 were not further decreased after 4EGI-1 treatment in Ythdf3-depleted cells (Figure 3K). Thus, we concluded that Ythdf3 regulates the Cap-dependent translation initiation of Foxm1 and Asxl1. In contrast, Ythdf2 knockdown did not affect either transcriptional or translational regulation of Foxm1 and Asxl1 expression in HPC-7 cells (Online Supplementary Figure S3J-N). In conclusion, our study provides the first evidence that Ythdf3 is required for HSC maintenance under stress, but not normal hematopoiesis in homeostatic conditions. We show that Ythdf3 but not Ythdf2 promotes translation of Foxm1 and Asxl1, critical regulators of HSC maintenance in HSPC. However, whether Ythdf1 has a role in HSC maintenance and is involved in Ythdf3-mediated regulation of translation of Foxm1 and Asxl1 remains to be determined. Our data suggest that the role of Ythdf1, 2, and 3 in gene regulation is likely cell context-dependent.

Authors Qinglin Dang,1,2,* Qiong Wu,1,2,* Fang Yu,2,* Yue Sheng,2,* Chunjie Yu,2 Guangzhong Song,3 Kimberly Paulsen,2 Jianxin Lyu1,3 and Zhijian Qian2 1

School of Laboratory Medicine and Life Sciences, Wenzhou Medical

University, Wenzhou, Zhejiang, China; 2Department of Medicine and Department of Biochemistry and Molecular Biology, UF Health Cancer Center, University of Florida, Gainesville, FL, USA and 3

Zhejiang Provincial People’s Hospital, Affiliated People’s Hospital of

Hangzhou Medical College, and Laboratory Medicine of Hangzhou Medical College, Zhejiang, China *QD, QW, FY and YS contributed equally as co-first authors. Correspondence: J. - jxlu313@163.com Z. - zhijian.qian@medicine.ufl.edu https://doi.org/10.3324/haematol.2021.279300 Received: May 27, 2021. Accepted: October 15, 2021. Prepublished: November 25, 2021. ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license Disclosures No conflicts of interest to disclose.

Haematologica | 107 August 2022

1926

LETTER TO THE EDITOR

Contributions ZQ: study design; QD, QW, FY, YS, CY, GS, KP, JL, and ZQ: acquisition, analysis, and interpretation of data; ZQ, QD, QW, FY, and YS: manuscript preparation.

Acknowledgments ZQ is a Leukemia & Lymphoma Society (LLS) scholar. Funding This study was supported by funding from the UF start-up (to ZQ).

References 1. Shi H, Wei J, He C. Where, when, and how: context-dependent functions of RNA methylation writers, readers, and erasers. Mol Cell. 2019;74(4):640-650. 2. Vu LP, Cheng Y, Kharas MG. The biology of m(6)A RNA methylation in normal and malignant hematopoiesis. Cancer Discov. 2019;9(1):25-33. 3. Zaccara S, Ries RJ, Jaffrey SR. Reading, writing and erasing mRNA methylation. Nat Rev Mol Cell Biol. 2019;20(10):608-624. 4. Lasman L, Krupalnik V, Viukov S, et al. Context-dependent functional compensation between Ythdf m(6)A reader proteins. Genes Dev. 2020;34(19-20):1373-1391. 5. Zaccara S, Jaffrey SR. A unified model for the function of YTHDF proteins in regulating m(6)A-modified mRNA. Cell. 2020;181(7): 1582-1595. 6. Pellagatti A, Armstrong RN, Steeples V, et al. Impact of spliceosome mutations on RNA splicing in myelodysplasia: dysregulated genes/pathways and clinical associations. Blood. 2018;132(12):1225-1240. 7. Li Z, Qian P, Shao W, et al. Suppression of m(6)A reader Ythdf2 promotes hematopoietic stem cell expansion. Cell Res. 2018;28(9):904-917. 8. Wang H, Zuo H, Liu J, et al. Loss of YTHDF2-mediated m(6)Adependent mRNA clearance facilitates hematopoietic stem cell

regeneration. Cell Res. 2018;28(10):1035-1038. 9. Mapperley C, van de Lagemaat LN, Lawson H, et al. The mRNA m6A reader YTHDF2 suppresses proinflammatory pathways and sustains hematopoietic stem cell function. J Exp Med. 2021;218(3):e20200829. 10. Shi H, Wang X, Lu Z, et al. YTHDF3 facilitates translation and decay of N(6)-methyladenosine-modified RNA. Cell Res. 2017;27(3):315-328. 11. Hou Y, Li W, Sheng Y, et al. The transcription factor Foxm1 is essential for the quiescence and maintenance of hematopoietic stem cells. Nat Immunol. 2015;16(8):810-818. 12. Fujino T, Kitamura T. ASXL1 mutation in clonal hematopoiesis. Exp Hematol. 2020;83:74-84. 13. Pinto do O P, Kolterud A, Carlsson L. Expression of the LIMhomeobox gene LH2 generates immortalized steel factor-dependent multipotent hematopoietic precursors. EMBO J. 1998;17 (19):5744-5756. 14. He F, Ni N, Zeng Z, et al. FAMSi: a synthetic biology approach to the fast assembly of multiplex siRNAs for silencing gene expression in mammalian cells. Mol Ther Nucleic Acids. 2020;22:885-899. 15. Bos PD, Zhang XH, Nadal C, et al. Genes that mediate breast cancer metastasis to the brain. Nature. 2009;459(7249):1005-1009.

Haematologica | 107 August 2022

1927

LETTER TO THE EDITOR

Cell-free DNA sequencing as a potential screening tool for phase I targeted treatment in refractory/relapse diffuse large B-cell lymphoma Diffuse large B-cell lymphoma (DLBCL) represents the most common subtype of non-Hodgkin lymphoma (NHL).1 With standard of care first line treatment, namely rituximab, cyclophosphamide, doxorubicin, vincristine, prednisone (R-CHOP), 35% of patients will relapse or present a refractory disease (rrDLBCL).2 For these patients, salvage options include high-dose chemotherapy, chimeric antigen receptor T cells and at advanced stage (≥3 lines), novel approaches with targeted agents. Given the low number of treatment options approved for such patients,3 clinical trials evaluating new drugs can provide access to more treatment options. The need for molecular orientation before inclusion in clinical trials and targeted treatment strategies is increasingly highlighted in the solid tumor and lymphoma fields.4 In order to do so, the ability to provide molecular characterization at the time of relapse, and during a period compatible with the aggressiveness of the disease, is mandatory. We previously reported on the feasibility of a “real-time” targeted screening platform based on tissue mutations analysis.5 However, this screening process might be associated with failure due to biopsy-related technical limits and/or legitimate patient refusal to undergo another biopsy. Using cell-free tumoral DNA (ctDNA) to provide mutation profiling would increase the rate of molecularly-oriented inclusion.6 In this study, we show that using a targeted panel to characterize the potential oncogenic driver in rrDLBCL for a molecularly-oriented treatment, ctDNA-based sequencing could identify 80% of the tumor variants, with a high sensitivity and an excellent coefficient of concordance per gene, providing here a proof-of-concept for molecular orientation based on ctDNA only, in rrDLBCL. In this study, a real-life series of 53 rrDLBCL patients for which a targeted molecular characterization was performed at time of relapse, both in the tumor and in ctDNA is presented. All patients signed a written informed consent prior to tumor biopsy, which was consistently performed before ctDNA sampling. Eight to 10 mL10 mL of plasma per patient were isolated from blood sampled into EDTA tubes (centrifuged within 3 hours after drawing) and frozen before subsequent ctDNA extraction (using Maxwell-RSC Instrument and RSC-LV kit, Promega). Multiplex polymerase chain reaction was performed on 5-10 ng DNA using a customized 152 kb-panel (IAD137284) covering exonic regions of 44 genes.5 The 114 bp libraries were prepared according to the manufacturer’s recommendation

for Ion AmpliSeq deep-targeted sequencing on Ion-S5 System (LifeTechnology). A mean sequencing depth of 3,900× was achieved for tumoral biopsies and 3,500× for ctDNA samples. Variant calling of GRCh37/hg19-aligned reads were performed using an institutional pipeline aggregating results from three variant callers (IonTorrent_Suite, GATK HaplotypeCaller, FreeBayes), run with low-stringency default setting parameters: minimum base_calling_score=4, min relative_reads_quality=11, maximum strand-bias=0.95, min position_depth=40×, min either_strand_mutated_reads=2×, minimum variant allele frequency (VAF) =0.001. For subsequent analysis, ctDNA and tumor biopsy sequencing results were then filtered out based on VAF threshold (0.1% for ctDNA variants, and 1% for tumor biopsy variants [TB variants]); a minimum number of mutated reads of 4× for ctDNA variants, and 6× for TB variants; a minimum depth of 300× for ctDNA variants and 100× for TB variants; a strand-bias >80% (or >70% considering insertion/deletion polymorphism [indel]). Remaining called variants were validated thanks to integrative genomics viewer (IGV) visualization to (i) exclude artifacts according to misestimated strand-bias, contextdependent errors (homopolymers/repeated sequences) or read-relative positional errors (i.e., focusing on amplicon’s end artifact), and to (ii) evaluate background error noise (considering occurrence of other single nucleotide variants [SNV]/indel than the variant call) allowing true low allele frequency (AF) ctDNA mutations to be considered. CD3+sorted lymphocytes germline DNA were used as a control to filter out patient-specific polymorphisms for n=25 patients. For the other patients, known polymorphisms and variants described in general population at frequency >0.5% (according to 1000Genomes Project or NHLBI-GOExome Sequencing Project public databases) were excluded, as well as remaining variants classified as “benign” or “likely benign” using VarSome online resource. The TB variants detection was considered the gold standard to perform a comparison with ctDNA variants detection. Cohen's κ coefficient7 was calculated per gene to analyze the concordance between both methods and interpreted with the Landis-Koch scale.8 A total of 53 rrDLBCL patients were included in the study. The median age at the time of tumor biopsy was 68 years. The vast majority of the patients, at the time of tumor biopsy, were found to have a disseminated disease (70% of stage 3-4), lactate dehydrogenase (LDH) upper the normal

Haematologica | 107 August 2022

1928

LETTER TO THE EDITOR

Figure 1. Detection of tumor mutations in cell-free tumoral DNA. (A) Correlation between variant allele fraction (VAF) as assessed in cell-free tumoral DNA (ctDNA) sequencing or genomic DNA sequencing extracted from tumor biopsy (TB), and (B) level of similarity within most recurrently mutated genes, calculated for each gene as the percentage of TB variants found in ctDNA.

limit (60%) and received in median two prior lines of therapies (range, 1-10) (Online Supplementary Table S1). The median time between tumor biopsy and ctDNA sampling was 24 days (range, 0-60). The turnaround time for ctDNA analysis was 7 to 10 days. Of note, in this real-life study, among the 53 patients, 28 received a short course of treatment between TB and ctDNA analysis (13 patients received a single short course/cycle of chemotherapy-based salvage and 15 an oral treatment). These patients had higher LDH levels than the others, however similar clinical characteristics overall. A total of 300 TB variants, within 34 genes, were found in the tumor, mean six per patient (median 6; range, 0-15), 241 (80%) were also present in ctDNA (ctDNA variants) (mean 4.8; median 5; range, 0-13), and eight variants were present in ctDNA only. Three patients were not mutated for any 44 targeted genes in both tumor biopsy and cell-free DNA (cfDNA). The incidence of genomic abnormalities is in line with previous reports in rrDLBCL9,10. Importantly, the TB variants VAF and ctDNA variants VAF were correlated (Figure1A; R=0.41, P<0.001; Online Supplementary Figure S1A patient level). Furthermore, the ability to detect the TB variant in ctDNA was significantly correlated with TB variant VAF (Student t-test: P<0.001) (Online Supplementary Figure S1B). Overall, of the 50 patients with at least one mutation detected in the tumor, mutation profiles were mostly similar between the two methods (mean level of similarity: 81% of TB variants per patients detected in ctDNA). Importantly, there was no significant distinction between the percentage of TB variants found in ctDNA in patients that received a short course of treatment and those that did not. Indeed, of the 25 untreated patients, 84% (0-100%)

of TB variants were detected versus 78% patients (0-100%) of the treated patients (P=0.54). There was also no significant correlation between the number of days elapsed between tumor biopsy and ctDNA sampling and the percentage of TB variants detected in ctDNA (P=0.12). Bulky disease was the only factor significantly associated with a higher percentage of TB variant detected in ctDNA for each patient (P=0.005) whereas there was only a nonsignificant trend for higher LDH and Ann Arbor stage (P=0.26, P=0.21, respectively) and no correlation with the number of lines of previous therapies (P=0.8). In mean, for each mutated gene, 78% of the TB variants were found in ctDNA (median 81%, range, 0-100%). When looking at the most recurrently mutated genes (in more than 8% of the patients, or top-20 recurrently mutated genes in this cohort), the percentage of TB variants found in ctDNA was always greater than 50% (mean 81%, range, 50-100%) (Figure 1B). Within this list of recurrently mutated genes, BCL7A had the lowest similarity level. Furthermore, the incidence of top-20 mutations was fully comparable between tumor biopsy and ctDNA (Fisher’s test: P>0.3) (as shown in the co-oncoplot, Online Supplementary Figure S1C). In order to further validate ctDNA sequencing as a good alternative to tumor biopsy sequencing, we assessed, within the 53 patients, the κ coefficient of concordance between the two techniques and the sensitivity/specificity (per gene) of variants detection in ctDNA, taking the tumor sequencing as reference. For this analysis, 308 variants (241 present in both compartments, 59 in tumor biopsy only, 8 in ctDNA only) were considered. As defined by the Landis-Koch8 scale, the concordance was excellent or good in 30 of 34 mutated genes (88%), medium/weak in

Haematologica | 107 August 2022

1929

LETTER TO THE EDITOR Table 1. Concordance between tumor biopsy and cell-free tumoral DNA variants detection per gene.

False positive N

False negative N

True positive N

Cohen’s κ coefficient7

Interpretation according to the Landis-Koch scale8

ARID1A

73%

good

ATM

66%

good

B2M

82%

excellent

BCL2

34%

weak

BCL7A

65%

good

CARD11

100%

excellent

CCND3

100%

excellent

CD58

100%

excellent

CD79B

88%

excellent

CDKN2A

73%

good

CIITA

100%

excellent

CREBBP

77%

good

CXCR4

100%

excellent

EED

100%

excellent

EP300

65%

good

EZH2

85%

excellent

FOXO1

100%

excellent

GNA13

100%

excellent

IRF4

69%

good

KMT2D

91%

excellent

MEF2B

100%

excellent

MYC

77%

good

MYD88

78%

good

NOTCH1

100%

excellent

PIM1

81%

excellent

PRDM1

bad

PTPN1

100%

excellent

SF3B1

bad

SOCS1

85%

excellent

TCF3

100%

excellent

TNFAIP3

85%

excellent

TNFRSF14

93%

excellent

TP53

70%

good

XPO1

49%

medium

Gene

True positive: number of tumor biopsy tumor biopsy (TB) variants also found in cell-free tumoral DNA (ctDNA); false negative: number of TB variants not found in ctDNA; false positive: number of variants found in ctDNA but not reported in biopsy. The Cohens’s κ concordance coefficient was interpreted with the Landis-Koch scale8: concordance is excellent if coefficient is in between: 1-0.81; concordance is good if coefficient is in between: 0.81-0.61; concordance is medium if coefficient is in between: 0.60-0.41; concordance is weak if coefficient is in between: 0.40-0.21; concordance is bad if coefficient is in between: 0.20-0.00.

two (6%) genes (XPO1-E571K and BCL2) and bad in two genes being mutated only once (PRDM1, SF3B1), potentially explaining this result (Table 1). The mean and median sensitivity of ctDNA detection per gene was 83% and 84% respectively (range, 33-100%), and specificity 99% and 100% (range, 98-100%) (Table2). The sensitivity of ctDNA detec-

tion reported here is in line or even slightly better compared to the literature data published so far,10,11 reporting on smaller cohort of patients at time of diagnosis.11 Finally, in this real-life cohort, 18% of patients with a molecular alteration could be molecularly oriented for a targeted treatment.5 This is in line with other studies in solid

Haematologica | 107 August 2022

1930

LETTER TO THE EDITOR Table 2. Sensitivity and specificity assessment per gene.

Gene

Sensitivity, %

Specificity, %

PPV, %

NPV, %

ARID1A

ATM

100

B2M

100

BCL2

100

BCL7A

100

CARD11

100

CCND3

100

CD58

100

CD79B

100

CDKN2A

100

CIITA

100

CREBBP

100

CXCR4

100

EED

100

EP300

EZH2

100

FOXO1

100

GNA13

100

IRF4

100

KMT2D

100

MEF2B

100

MYC

MYD88

100

NOTCH1

100

PIM1

100

PRDM1

100

PTPN1

100

SF3B1

100

SOCS1

100

TCF3

100

TNFAIP3

100

TNFRSF14

100

TP53

XPO1

100

Range

33-100

98-100

PPV: positive predictive value; NPV: negative predictive value; na: non-calculable values (no positive test).

tumors,4 suggesting that molecular orientation remains a work in progress and is highly dependent on availability of clinical trials at time of inclusion. Our analysis has some limitations, including the time difference between tissue and plasma collection, and receipt of intervening treatment for some patients, even though these two parameters were not correlated with the ability to detect ctDNA mutations. In conclusion, our study shows that using a targeted panel to characterize the potential oncogenic driver in rrDLBCL for a molecularly-oriented treatment, ctDNA-based se-

quencing could identify 80% of the tumor variants, with a high sensitivity and an excellent-to-good κ coefficient of concordance per gene. Importantly, the number of lines of previous therapy as well as the number of days between tumor biopsy and ctDNA sampling had no impact on the level of similarity between the two techniques, whereas the VAF or tumor bulk were associated with a greater similarity, by patient. All in all, this technology seems appropriate in routine to screen rrDLBCL patients for inclusion in clinical trials with a molecular orientation process.

Haematologica | 107 August 2022

1931

LETTER TO THE EDITOR

Authors

https://doi.org/10.3324/haematol.2021.280464

Cyril Quivoron,1,2* Anthony Tarabay,3* Jean-Marie Michot,3,4* Arnaud

Received: December 15, 2021.

Pagès,

Accepted: March 11, 2022.

5,6

Hélène Lecourt, Anne Aupérin, Alina Danu, Julien 1

Prepublished: March 24, 2022.

Lazarovici, Julien Rossignol, David Ghez, Peggy Dartigues, 3

Véronique Vergé, Christophe Massard, Valérie Camara-Clayette, 7

1,8

Vincent Ribrag1,2,3,4 and Clémentine Sarkozy2,4 1

Hematology Translational Research Laboratory, AMMICa, INSERM

US23 / CNRS UAR 3655, Gustave Roussy; 2INSERM U1170, Gustave

Disclosures

Roussy; Hematology Department, Gustave Roussy; DITEP, Gustave

No conflicts of interest to disclose.

Roussy; Department of Biostatistics and Epidemiology, Gustave 5

Roussy; 6INSERM U1018, Gustave Roussy; 7Department of Medical

Contributions

Biology and Pathology, Gustave Roussy and Plateforme de

CQ, AT, J-MM, PD, VC-C, VR and CS developed the concept and

Recherche Gustave Roussy, Centre de Ressources Biologiques, UMS

designed the study; CQ, AT, J.-MM, HL, AD, JL, JR, DG, PD, VV, CM, VR

AMMICa, INSERM US23 / CNRS UMS 3655, Villejuif, France

and CS acquired data; CQ, AT, AP, HL, AA, VR and CS analyzed and

*CQ, AT and JMM contributed equally as co-first authors.

interpreted data. All authors reviewed and approved the manuscript.

Correspondence: C. SARKOZY - clementine.sarkozy@gustaveroussy.fr

References 1. Swerdlow SH, Campo E, Pileri SA, et al. The 2016 revision of the World Health Organization classification of lymphoid neoplasms. Blood. 2016;127(20):2375-2390. 2. Maurer MJ, Ghesquières H, Jais JP, et al. Event-free survival at 24 months is a robust end point for disease-related outcome in diffuse large B-cell lymphoma treated with immunochemotherapy. J Clin Oncol. 2014;32(10):1066-1073. 3. Sarkozy C, Sehn LH. New drugs for the management of relapsed or refractory diffuse large B-cell lymphoma. Ann Lymphoma. 2019;3:10. 4. Massard C, Michiels S, Ferté C, et al. High-throughput genomics and clinical outcome in hard-to-treat advanced cancers: results of the MOSCATO 01 trial. Cancer Discov. 2017;7(6):586-595. 5. Sarkozy C, Michot J, Quivoron C, et al. Innovative therapies based on molecular orientation in patients with relapse and refractory diffuse large B cell lymphoma, results of LNH‐EP1 study. Am J Hematol. 2021;96(10):E376-379. 6. Huet S, Salles G. Potential of circulating tumor DNA for the management of patients with lymphoma. JCO Oncol Pract.

2020;16(9):561-568. 7. Cohen J. A Coefficient of agreement for nominal scales. Educ Psychol Meas. 1960;20(1):37-46. 8. Koch GG, Landis JR, Freeman JL, Freeman DH, Lehnen RG. A general methodology for the analysis of experiments with repeated measurement of categorical data. Biometrics. 1977;33(1):133. 9. Rushton CK, Arthur SE, Alcaide M, et al. Genetic and evolutionary patterns of treatment resistance in relapsed B-cell lymphoma. Blood Adv. 2020;4(13):2886-2898. 10. Liu H, Yang C, Zhao X, et al. Genotyping on ctDNA identifies shifts in mutation spectrum between newly diagnosed and relapse/refractory DLBCL. Onco Targets Ther. 2020;13:10797-10806. 11. Rivas-Delgado A, Nadeu F, Enjuanes A, et al. Mutational landscape and tumor burden assessed by cell-free DNA in diffuse large B-cell lymphoma in a population-based study. Clin Cancer Res. 2021;27(2):513-521.

Haematologica | 107 August 2022

1932

LETTER TO THE EDITOR

Full versus prophylactic-intermediate doses of anticoagulants in COVID-19: a meta-analysis Coronavirus disease-19 (COVID-19) is a serious pandemic associated with an elevated risk of venous and arterial thrombosis.1 There is no consensus on the anticoagulant approach to be implemented to reduce the thrombotic risk; thus, most international guidelines recommend standard thromboprophylaxis with use of intermediate doses restricted to high-risk and critically ill patients.2 Therapeutic dosage superiority of anticoagulants in reducing thrombosis in moderately but not severely ill patients or uncertain cost-benefit of full-intermediate anticoagulation versus prophylactic doses were reported by previous meta-analyses.3,4 Therefore, we decided to perform a new, updated meta-analysis of randomized clinical trials comparing the effects of full anticoagulation (FA) versus prophylactic-intermediate anticoagulation (PIA) on death and thrombotic-related events in COVID-19 patients. Eligibility criteria Types of studies: randomized clinical trial studies that assessed the effect of therapeutic versus prophylactic anticoagulant therapy in COVID-19 hospitalized patients. No language, publication date, or publication status restrictions were imposed. We conducted all analyses according to the intention-to-treat principle. For trials with a factorial design, we based main results on 2-way analyses, that is, all trial participants receiving FA were compared with all those treated with PIA dose. Information sources The studies were identified by searching electronic databases. This search was applied to Pubmed, ISI Web of Science, SCOPUS and Cochrane database. The last search was run on November 14, 2021. Reference lists of all studies included in the present meta-analysis were screened for potential additional eligible studies. Search Two investigators independently searched in the electronic databases combining the following text terms and MeSH terms: "COVID-19"[All Fields] OR "COVID-19"[MeSH Terms] OR "SARS-CoV-2"[All Fields] OR "sars-cov-2"[MeSH Terms] OR "Severe Acute Respiratory Syndrome Coronavirus 2"[All Fields] OR "NCOV"[All Fields] OR "2019 NCOV"[All Fields] AND ("thrombosis"[MeSH Terms] OR "thrombosis"[All Fields])) AND ("anticoagulants"[Pharmacological Action] OR "anticoagulants"[MeSH Terms] OR "anticoagulants"[All Fields] OR "anticoagulant"[All Fields]) AND "humans"[MeSH Terms]Studies. We limited our search to human studies.

Study selection Two authors independently reviewed the titles and abstracts generated by the search. Studies were excluded if the title and/or abstract showed that the papers did not meet the selection criteria of our meta-analysis. Studies not including a control group and animal studies were excluded. Case reports, editorials, commentaries, letters, review articles, guidelines were also excluded from the analysis. We defined the following exclusion criteria: (i) studies that included only the intermediate anticoagulation; (ii) studies unrelated to our topic; (iii) studies without randomization assignment of the treatment. A flowchart of the selection studies is reported in the Online Supplementary Figure S1. Main analysis We evaluated the effect of FA versus PIA dose in COVID19 hospitalized patients. Our primary outcome was to compare the effect of these treatments on death and thrombotic events. Our secondary endpoints were the comparison of these treatments on death, arterial and venous thrombotic events, arterial thrombotic events and VTE; the safety endpoint was to evaluate the effect on major bleedings. This meta-analysis was conducted and reported according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analysis) Statement issued in 2009. Statistical analysis We allocated the results of each trial as dichotomous frequency data. Risk ratios (RR) and 95% confidence intervals (CI) were calculated. Data were pooled and compared with a random-effect model. We calculated the number needed to treat (NNT) and the number needed to harmful (NNH) as the reciprocal of the absolute risk reduction (ARR). Statistical heterogeneity was calculated by the I2. Presence of publication bias was explored using funnel plots of effect size against standard error and Egger's test. The software Comprehensive Meta Analysis (version 2.2.064, USA, 2011) and R (version 3.1.2, Vienna, 2014) supported the analysis. Seven studies5-11 evaluated the effect of FA versus PIA. The relevant parameters of each study are reported in the Table 1. Antiplatelet drugs were more frequently used in patients treated with FA (11.9%) versus PIA (10.6%). A heterogeneous definition of the severity of COVID-19 infection was found among the studies. Briefly, we considered

Haematologica | 107 August 2022

1933

-311 Patients treated with therapeutic doses (Mild: 10%, Moderate: 83%, Severe: 8%). -304 Patients treated with prophylactic doses (Mild: 13%, Moderate: 82%, Severe: 5%).

Moderate disease was defined by an oxygen saturation <94%, pulmonary infiltrates >50%, or a partial pressure of oxygen to fractional concentration of oxygen in inspired air ratio <300. Severe disease was defined as respiratory failure, hemodynamic instability, or multiple organ dysfunction. Mild disease includes cases not meeting the criteria for moderate or severe disease.

Study

Action5

Death,duration of hospitalization or of supplemental oxygen support

The study included patients with Ddimer criterion greater than 4 times the upper limit of normal or a sepsis-induced coagulopathy score ≥4. The study considered severe patients those hospitalized in ICU and not severe the “Non-ICU stratum”.

Moderate illness was defined as admission to hospital ward level of care (not to ICU), not already mechanically ventilated, and not imminently requiring mechanical ventilation or Not Severe (N=465) critical care. D-dimer levels were required to be above the upper limit of normal with an oxygen saturation ≤93% on room air, or ≥2 times the upper limit of normal irrespective of oxygen saturation.

HEPCOVID6

Haematologica | 107 August 2022

1934

RAPID7

Composite of ICU admission, non-invasive or invasive mechanical ventilation.

Males Primary Endpoints %

Severe(N=84) and Not Severe(N=45) patients were treated with theraVTE, arterial thrompeutic doses. Severe(N=86) and Not botic events or death Severe(N=38) patients were treated with prophylactic doses.

Disease state at baseline

Definition of disease state at baseline

Table 1. Characteristics of the studies.

Antiplatelets: Therapeutic: 25/310 (8.1%) Prophylactic 30/304 Standard in-hospital (9.9%) enoxaparin or UFH Corticosteroids: Therapeutic: 257/310 (83%), Prophylactic 253/304 (83%)

Patients allocated to therapeutic heparin receive therapeutic doses of LMWH or UFH as used for the treatment of venous thromboembolism.

Antiplatelets: Therapeutic 24/228 (10%), Prophylactic 29/237 (12%) Corticosteroids: Therapeutic 161 (70.6 %), Prophylactic 162 (68.4%)

28 days

Defined according to the ISTH criteria.

Defined according to the ISTH criteria

Defined according to ISTH* criteria

Major Bleeding

Continued on following page.

Patients allocated to prophylactic heparin received dose capped prophylactic s.c. heparin (LMWH or UFH) adjusted for body mass index and creatinine clearance.

30 days

Antiplatelets Prophylactic/ Duration and intermediate dose corticosteroid

Heparin regimens could include UFH, Antiplatelets: Enoxaparin at a up to 22 500 IU s.c. Therapeutic: dose of 1 mg/kg (twice or thrice daily); 40/129 (31.0%), s.c. twice daily if enoxaparin, 30 mg or Prophylactic 24/124 CrCl was 30 40 mg s.c. once or (19.4) mL/min/1.73m2 or twice daily (weightCorticosteroids: greater or 0.5 based enoxaparin Therapeutic: mg/kg twice daily 0.5mg/kg subcu111/127 (87.4), if CrCl was 15-29 taneously twice daily Prophylactic 93/123 mL/min/1.73m2 was permitted; or dal(75.6) teparin, 2500 IU or 5000 IU s.c. daily

Rivaroxaban (20 mg once daily) for patients with a stable condition or enoxaparin (1 mg/kg twice daily) for patients with an unstable condition

Age Anticoagulant in therapeutic years dose

LETTER TO THE EDITOR

Definition of disease state at baseline

Disease state at baseline

1935

Haematologica | 107 August 2022

BEMICOP11

The variation in gas exchange over time

Organ support–free days

10 days

21 days

Antiplatelet drug was an exclusion criterium Bamiparin 115 IU/Kg Bamiparin 3,500 IU once Corticosteroids: once daily adjusted for Therapeutic: 30/33 (90) daily body weight. Prophylactic: 32/32 (100)

Usual-care pharmacologic thromboprophylaxis

Antiplatelets: Therapeutic 37/485 (7.6), Prophylactic 38/494 Therapeutic-dose anti(7.7) coagulation with UFH Corticosteroids: or LMWH Therapeutic: 426/522 (81.6), Prophylactic 458/555 (82.5)

21 days

14 days

Usual-care pharmacologic thromboprophylaxis

Antiplatelets: Therapeutic: 148/1140 (13.0), Prophylactic 111/1013 Therapeutic dose anti(11.0) coagulation with UFH Corticosteroids: or LMWH Therapeutic:479/791 (60.6) Prophylactic 415/656 (63.3)

Duration

The standard thromboprophylaxis group Antiplatelets: was allocated to receive s.c., Enoxaparin s.c. with Therapeutic 0/10 (0), UFH at a dose of 5000 IU the dose according to Prophylactic 0/10 (0). TID (if weight<120 kg) and age and adjusted daily Corticosteroids: 7500 IU TID (if weight>120 by the creatinine Therapeutic: 7/70 (70), kg) or enoxaparin at a dose clearance Prophylactic: 7/70 (70) of 40 mg OD (if weight<120 kg) and 40 mg BID (if weight>120 Kg).

Prophylactic/ intermediate dose

Antiplatelets and corticosteroid

Age Anticoagulant in therapeutic dose years

Defined according to the ISTH criteria.

Defined according to the TIMI** criteria

Defined according to the ISTH criteria.

Major Bleeding

* International Society on Thrombosis and Hemostasis (ISTH) criteria: fatal bleeding; symptomatic bleeding in a critical area or organ; hemoglobin level <2 g/dL or more; bleeding leading to transfusion of 2 or more units of whole blood or red cells.** TIMI criteria: defined as: any intracranial bleeding (excluding microhemorrhages <10 mm evident only on gradient-echo MRI), clinically overt signs of hemorrhage associated with a drop in hemoglobin of ≥5 g/dL or a ≥15% absolute decrease in hematocrit, fatal bleeding. s.c.: subcutaneously; LMWH: low molecular weight heparin; UFH: unfractionated heparin. ICU: intensive care unit.

Not severe Thrombotic (N=65) Events

positive COVID-19 diagnosis CURB≤2, baseline blood oxygen saturation ≥90%

Severe (N=1,103)

HESACOVID10

Severe Covid-19 was defined as Covid-19 that led to receipt of ICU-level respiratory or cardiovascular organ support in an ICU.

Severe clinical presentation with respiratory failure Severe requiring mechanical (N=20) ventilation, D-dimer levels greater than 1000 μg/L

ATTACC, ACTIV-4a, and REMAP-CAP Investigators 9

Primary Males Endpoints %

Moderate disease was defined as hospitalization for ATTACC, Covid-19 without the need ACTIV-4a, for ICU-level care. Not Severe Organ supand ICU-level care was de(N=2,231) port–free days REMAP-CAP fined as the use of Investigators 8 respiratory or cardiovascular organ support in an ICU.

Study

LETTER TO THE EDITOR

LETTER TO THE EDITOR as severe patients those: hospitalized at admission in an intensive care unit (ICU), or with respiratory failure requiring mechanical ventilation or organ support as high flow oxygen, extracorporeal life support, vasopressors, or inotropes. Comparison between subgroups with severe and non-severe disease was prespecified analysis. Funnel plots are reported in the Online Supplementary Figure S2.

No difference was observed between the rates of death and thrombotic events (as composite outcome) in patients treated with FA (16.5%) versus PIA (19.9%) in seven studies including 4,734 patients (Figure 1A). Subgroup analysis according to severity showed a lower rate of this composite outcome in patients treated with FA as compared to those treated with PIA in non-severe (RR=0.53,

Figure 1. Forrest plots of death and thrombotic events. (A) Forest plots for death and thrombotic events according to the subgroups for full anticoagulation (FA) and prophylactic-intermediate anticoagulation (PIA) doses. (B) Forest plots for death according to the subgroups for FA and PIA doses. (C) Forest plots for arterial and venous and thrombotic events according to the subgroups for FA and PIA doses. Haematologica | 107 August 2022

1936

LETTER TO THE EDITOR 95% CI: 0.30-0.94), but not in severe patients (Figure 1A). No difference was observed between the rates of death in patients treated with FA (15.4%) and PIA (14.8%) in seven studies including 4,741 patients (Figure 1B). In seven studies including 4,732 patients a difference was observed between the rates of arterial and venous thrombotic events in patients treated with FA (4.0%) compared to those treated with PIA (7.2%, RR=0.58, 95% CI: 0.45-

0.76), both in severe and non-severe patients (Figure 1C). In six studies (n=4,667 patients) a difference was observed between the rate of venous thrombotic events (VTE) in FA (2.7%) versus PIA patients (5.9%, RR=0.47, 95% CI: 0.350.63), both in severe and non-severe patients (Figure 2A). No significant difference was observed between the rates of arterial thrombotic events in three studies including 1,332 patients treated with FA or PIA (Figure 2B).

Figure 2. Forrest Plots of thrombotic events and bleeding. (A) Forest plots for venous thrombotic events according to the subgroups for full anticoagulation (FA) and prophylactic-intermediate anticoagulation (PIA) doses. (B) Forest plots for arterial thrombotic events. (C) Forest plots for major bleeding events according to the subgroups for FA and PIA doses. Haematologica | 107 August 2022

1937

LETTER TO THE EDITOR Six studies, including 4,650 patients, reported major bleeding rates. Major bleeding was observed with an higher rate in FA (2.5%) versus PIA patients (1.4%, Figure 2C). Number needed to treat (NNT) and number needed to harm (NNH) values for major bleedings and several outcomes are reported in the Online Supplementary Figure S3, in all studies and according to severity. The study suggests a beneficial effects of FA towards venous thrombosis compared to PIA COVID-19 patients. Even if the risk of major bleeding is higher in FA-treated patients, the overall cost benefit of treatments is in favor of FA. It is, so far, unclear whether FA is superior to lower doses of anticoagulants such as PIA on the incidence of death, arterial and venous thrombosis, and major bleeding in COVID-19 patients. The results of the present study show that, in comparison with PIA, FA did not affect the rate of death or arterial thrombosis but reduced the incidence of venous thrombosis interdependently of the clinical presentation, ie., either severe or non-severe disease. As expected, the rate of major bleeding was higher in patients on FA. Concomitant treatment with antiplatelet drugs may be an important confounder as we would expect a higher risk of bleeding in patients under full anticoagulation12 but lack of clinical information on this specific subgroup precludes definite conclusion. The study has implications and limitations. Even if the full anticoagulation did not reduce the risk of death, the positive impact on VTE is of clinical relevance for the management and the potentially harmful long term effects of VTE; in accordance with this, for VTE and major bleeding the NTT and NNH are 31 and 90 respectively and NTT/NNH 0.34. This finding adds more to the previous meta-analyses3,4 as the bleeding risk seems to be lower and the cost-benefit in favor of full anticoagulation in either moderate or severely ill patients. Clinical outcomes were restricted to 30-day followup, thereby it is unclear the impact of several anticoagulant regimens on clinical outcomes, death included, during long term followup. In conclusion, the results of this meta-analysis suggest the potential usefulness of FA to reduce VTE in COVID19 patients with either severe or non-severe disease.

Further RCT with large sample size are needed to support this finding.

Authors Lorenzo Loffredo,1 Augusto Di Castelnuovo,2 Giovanni Alfonso Chiariello,3,4 Pasquale Pignatelli1 and Francesco Violi2,5 1

Department of Clinical Internal Medicine, Anesthesiologic and

Cardiovascular Sciences, Sapienza University of Rome, Rome; 2

Mediterranea Cardiocentro, Napoli; 3Cardiovascular Sciences

Department, Agostino Gemelli Foundation Polyclinic IRCCS, Rome; 4

Catholic University of The Sacred Heart, Rome and 5Sapienza

University of Rome, Rome, Italy. Correspondence: F. VIOLI - francesco.violi@uniroma1.it L. LOFFREDO - lorenzo.loffredo@uniroma1.it https://doi.org/10.3324/haematol.2022.280652 Received: January 7, 2022. Accepted: March 22, 2022. Prepublished: March 31, 2022. ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

Disclosures No conflicts of interest to disclose. Contributions FV and LL developed the study concept and design; data: PP, ADC, LL and FV analyzed and interpreted data; FV and LL wrote the initial draft; PP, ADC and GAC critically revised the manuscript for important intellectual content; ADC and LL performed the statistical analysis. Data-sharing statement All the data used to support the results of this study are available from the corresponding author upon request.

References 1. Violi F, Pastori D, Cangemi R, Pignatelli P, Loffredo L. Hypercoagulation and antithrombotic treatment in Coronavirus 2019: a new challenge. Thromb Haemost. 2020;120(6):949-956. 2. Spyropoulos AC, Levy JH, Ageno W, et al. Scientific and standardization committee communication: clinical guidance on the diagnosis, prevention, and treatment of venous thromboembolism in hospitalized patients with COVID-19. J Thromb Haemost. 2020;18(8):1859-1865.

3. Ortega-Paz L, Galli M, Capodanno D, et al. Safety and efficacy of different prophylactic anticoagulation dosing regimens in critically and non-critically ill patients with COVID-19: a systematic review and meta-analysis of randomized controlled trials. Eur Heart J Cardiovasc Pharmacother. 2021 Sep 14. [Epub ahead of print] 4. Sholzberg M, da Costa BR, Tang GH, et al. Randomized trials of therapeutic heparin for COVID-19: a meta-analysis. Res Pract Thromb Haemost. 2021;5(8):e12638.

Haematologica | 107 August 2022

1938

LETTER TO THE EDITOR 5. Lopes RD, de Barros ESPGM, Furtado RHM, et al. Therapeutic versus prophylactic anticoagulation for patients admitted to hospital with COVID-19 and elevated D-dimer concentration (ACTION): an open-label, multicentre, randomised, controlled trial. Lancet. 2021;397(10291):2253-2263. 6. Spyropoulos AC, Goldin M, Giannis D, et al. Efficacy and safety of therapeutic-dose heparin vs standard prophylactic or intermediate-dose heparins for thromboprophylaxis in highrisk hospitalized patients with COVID-19: the HEP-COVID randomized clinical trial. JAMA Intern Med. 2021;181(12):1612-1620. 7. Sholzberg M, Tang GH, Rahhal H, et al. Effectiveness of therapeutic heparin versus prophylactic heparin on death, mechanical ventilation, or intensive care unit admission in moderately ill patients with covid-19 admitted to hospital: RAPID randomised clinical trial. BMJ. 2021;375:n2400.

8. Investigators A, Investigators AC-a, Investigators R-C, et al. Therapeutic anticoagulation with heparin in noncritically ill patients with Covid-19. N Engl J Med. 2021;385(9):790-802. 9. Investigators R-C, Investigators AC-a, Investigators A, et al. Therapeutic anticoagulation with heparin in critically ill patients with Covid-19. N Engl J Med. 2021;385(9):777-789. 10. Lemos ACB, do Espirito Santo DA, Salvetti MC, et al. Therapeutic versus prophylactic anticoagulation for severe COVID-19: a randomized phase II clinical trial (HESACOVID). Thromb Res. 2020;196:359-366. 11. Marcos M, Carmona-Torre F, Vidal Laso R, et al. Therapeutic vs. prophylactic bemiparin in hospitalized patients with nonsevere COVID-19 (BEMICOP): an open-label, multicenter, randomized trial. Thromb Haemost. 2021;122(2):295-299. 12. Schaefer JK, Errickson J, Li Y, et al. Adverse events associated with the addition of aspirin to direct oral anticoagulant

Haematologica | 107 August 2022

1939

LETTER TO THE EDITOR

The use of hydroxyurea pretreatment in chronic myeloid leukemia in the current tyrosine kinase inhibitor era The added value of hydroxyurea (hydrea) as a treatment modality in chronic myeloid leukemia (CML) is limited since the availability of tyrosine kinase inhibitors (TKI). Nonetheless, many clinicians still use it as a temporary treatment while awaiting a definite diagnosis to achieve some cytoreduction and alleviate symptoms due to hyperleukocytosis.1 However, it is uncertain if early cytoreduction with hydrea confers any benefit for patients, particularly in the absence of symptomatic hyperleukocytosis. Several reports even hinted that hydrea had an antagonistic effect on subsequent imatinib responses, but this was only observed with extended use of hydrea (>6 months,2 >12 months3). Current European LeukemiaNet (ELN) recommendations advise that a short course of hydrea may be given in symptomatic patients with high leukocyte or platelet counts while molecular and cytogenetic confirmation of the CML diagnosis is pending.4 However, the impact of hydrea pretreatment has not been specifically studied so far and it remains unclear how this recommendation is applied in real-world clinical practice. The objectives of this study were to evaluate the use of hydrea as a pretreatment modality in a population-based nationwide CML-cohort treated in the TKI era, to assess its influence on achieving the ELN response milestones and the occurrence of hematotoxicity during subsequent TKI therapy. We performed a retrospective analysis on data from a real-world population-based Dutch patient cohort from the PHAROS CML Registry (445 patients diagnosed between 2008 and 2014) and HemoBase (35 patients diagnosed between 1999 and 2015). Additional information on both registries can be found in Geelen et al.5 Patients were included if diagnosed in chronic phase (CP) according to the ELN criteria, treated with a TKI in first line and if sufficient treatment information was available. Patients were grouped based on having received pretreatment with hydrea. Hematologic, cytogenetic, molecular responses and response milestones were defined as described in the ELN recommendations.4 In the registry, BCR-ABL1 levels were reported as a percentage on the International Scale (%IS) or as a log reduction of BCR-ABL1 transcripts from baseline in case molecular laboratories were unable to report on the IS at that time. The achievement of ELN response milestones was assessed at 3, 6 or 12 months with a margin of 1 month for each point in time. Differences between proportions were tested using the Chi-square method with a Bonferroni correction for multiple testing. For the time-to-response analysis, the Fine & Gray cumu-

lative incidence competing risk (CICR) method was used with the start of TKI treatment as time point zero and death or progression to acceleration phase or blast crisis as a competing event. The achievement of a complete cytogenetic response and/or a BCR-ABL1-value <1%IS or a two log reduction (CCyR/MR2.0) were pooled in the timeto-response analysis since they represent an equivalent disease burden.6 However, in the main CICR analysis, patients without standardized (IS) molecular results were excluded. A Fine & Gray competing risks (CR) regression model was used for multivariable analysis including sex, age, EUTOS long term survival score (ELTS), leukocyte count at diagnosis and first-line TKI generation as covariates. Missing values in covariates were handled using substantive model compatible fully conditional specification (SMC-FCS) multiple imputation. A second CICR analysis for responses was performed on propensity score matched cohorts. A P-value lower than 0.05 was considered significant. All statistical analyses were performed in SPSS (version 24) and R (version 1.3.1093). The Medical Ethics Committee of the Erasmus Medical Center in Rotterdam approved this study and the exemption from informed consent. The study was conducted in accordance with the Declaration of Helsinki. The PHAROS CML Registry combined with HemoBase comprises a total of 480 CML patients. Four patients underwent leukapheresis for hyperleukocytosis, 15 patients were diagnosed in advanced phase disease according to the ELN criteria, for 54 patients we had insufficient treatment information available and five patients were not treated with a TKI in first line. These patients were excluded from further analyses. Out of 402 included patients, 175 (44%) patients received hydrea pretreatment. Hydrea pretreatment was given more frequently in the first decade after the introduction of imatinib (2003-2010) than after 2010 (2011-2015; 48% vs. 35%, P=0,012). Hydrea-treated patients had a less favorable risk profile based on Sokal score or ELTS, and more often reported constitutional symptoms (36% vs. 25%, P=0.027) or symptomatic splenomegaly (25% vs. 12%, P=0.001) at diagnosis (Table 1). Median leukocyte count was higher in hydrea-pretreated patients (Figure 1A), of whom 21% had a relatively low leukocyte count below 100x109/L. Symptomatic hyperleukocytosis with clinical signs of hyperviscosity was more prevalent in hydrea-pretreated patients, but constituted a minority of patients (10% vs. 3%, P=0.005). Also, approximately 40% of hydreatreated patients were asymptomatic.

Haematologica | 107 August 2022

1940

LETTER TO THE EDITOR Table 1. Baseline characteristics, treatment characteristics and adverse events in patients with or without hydrea pretreatment.

No hydrea (N=227) Baseline characteristics Sex, male

Hydrea pretreatment (N=175)

116

108

Age at diagnosis (years), mean (SD)

57.5 (16,8)

54.0 (16.6)

P-value

0.034 0.039

Sokal score, low/intermediate/high

59/102/45

29/50/22

31/65/61

20/41/39

0.002

ELTS, low/intermediate/high

111/72/23

54/35/11

60/64/33

38/41/21

0.004

Leukocytes, x109/L median (IQR)

70.6 (37.6-142.1)

Hyperleukocytosis*

136

<0.001

Symptomatic hyperleukocytosis

0.005

Symptoms, any

0.001

Constitutional symptoms

0.027

Symptomatic splenomegaly

0.001

195.0 (118.0-283.4)

<0.001

Treatment characteristics Time from diagnosis to start first line TKI (days), median (IQR)

11.0 (1.0-21.0)

17.0 (7.0-29.0)

0.001

14.0 (8.0-26.5)

Hydrea treatment duration (days), median (IQR) First line TKI, imatinib/2GTKI

178/49

78/22

135/40

77/23

0.885

Hematotoxicity on first line TKI, N (%)

<0.001

TKI reduction/interruption first line overall

0.167

TKI reduction/interruption first line due to hematotoxicity

0.018

Adverse events

SD: standard deviation; ELTS: EUTOS long term survival score; IQR: interquartile range; TKI: tyrosine kinase inhibitor; 2GTKI: second generation TKI. *Hyperleukocytosis was defined as a leukocyte count of >100x109/L.

The hydrea-treated group had a significantly longer interval between diagnosis and the start of first-line TKI treatment (21 vs. 15 days, P=0.002). After the start of TKI therapy, hematotoxicity occurred more frequently in hydrea-pretreated patients (25% vs. 11%, P<0.001) resulting in more hematotoxicity-related TKI dose reduction or interruption (14% vs. 8%, P=0.018). No apparent benefit was observed for hydrea pretreatment for patients achieving a complete hematologic response (CHR), a partial cytogenetic response (PCyR) or a BCR-ABL1 <10%IS by 3 months, a CCyR by 6 months or a MMR by 12 months (Figure 1B). Significantly less hydreapretreated patients achieved a BCR-ABL1 <1%IS by 6 months (65% vs. 80%, P=0.015). A total of 346 patients could be assessed for CCyR/MR2.0 and 304 patients for MMR. Molecular results were available as %IS (n=243) or as log-reduction (n=61). In these cohorts, 325 (94%) and 260 (86%) patients achieved CCyR/MR2.0 and MMR, respectively. For the following main CICR analysis, only patients with standardized molecular results (%IS) were included. The cumulative inci-

dence of CCyR/MR2.0 by 6 months was 40% (95% confidence interval [CI]: 32-47) and 48% (95% CI: 41-56) for patients with or without hydrea pretreatment, respectively (Figure 1C). The cumulative incidence of MMR by 12 months was 44% (95% CI: 42-47) and 56% (95% CI: 47-64) for patients with or without hydrea pretreatment, respectively (Figure 1D). Similar results were found in the propensity score matched cohorts with a cumulative incidence of CCyR/MR2.0 of 42% and 56% (P=0.056) by 6 months and a cumulative incidence of MMR of 43% and 52% (P=0.233) by 12 months for patients with or without hydrea pretreatment, respectively. Propensity scores were calculated with the following baseline parameters: age, sex, ELTS, leukocyte count, symptoms at diagnosis and time from diagnosis to start first-line TKI. The matched cohorts consisted of 262 and 242 patients for CCyR/MR2.0 and MMR, respectively. In a univariable CR regression model, the subdistribution hazard ratio (SHR) of hydrea pretreatment for CCYR and/or BCR-ABL1 <1%IS was 0.77 (95% CI: 0.62-0.96; P=0.022). The

Haematologica | 107 August 2022

1941

LETTER TO THE EDITOR

Figure 1. Characteristics of patients with versus without hydrea pretreatment. (A) leukocyte count (x109/L) at diagnosis. Red dots represent the median values in both groups; (B) proportion (%) achieving the ELN defined milestones by 3, 6 and 12 months; (C) cumulative of incidence of a complete cytogenetic response (CCyR) or BCR-ABL1 <1% on the International Scale (IS); (D) cumulative of incidence of a major molecular response (MMR). CHR: complete hematologic response; PcyR: partial cytogenetic response.

SHR of hydrea pretreatment in the multivariable model was 0.86 (95% CI: 0.67-1.12; P=0.274) for CCYR and/or BCRABL1 <1%IS and 0.95 (95% CI: 0.72-1.26; P=0.722) for MMR. Of note, in both models ELTS and first-line 2GTKI remained strong, independent predictors for respective responses. In two subgroup multivariable analyses of patients with hyperleukocytosis (leukocyte count >100x109/L) and with first-line imatinib, the SHR of hydrea pretreatment for CCyR/MR2.0 were 1.04 (95% CI: 0.75-1.44; P=0.826) and 0.916 (95% CI: 0.69-1.21; P=0.535), respectively. Hydrea pretreatment was used in nearly half of newly diagnosed CML patients in our cohort. Surprisingly, we observed a significantly longer delay in starting first-line TKI treatment in hydrea-pretreated patients. Furthermore, hydrea pretreatment was associated with a higher rate of hematotoxicity during first-line TKI treatment and hematotoxicity-related first-line TKI dose reduction or interruption. These patients had a longer median hydrea exposure (19 days) compared to patients without hematotoxicity (14 days), however this difference was not significant (P=0.251). Hematotoxicity did not occur significantly more in patients treated with a first-line 2GTKI.

Patients receiving hydrea pretreatment had no apparent benefit in achieving the ELN-defined responses both in the descriptive analysis by each milestone and in the time-to-response analysis, even while using the start of TKI therapy as time point zero. As expected, patients receiving hydrea pretreatment had a less favorable baseline profile with a higher ELTS and higher leukocyte count at diagnosis, but even after correction for these confounders hydrea pretreatment did not exert a beneficial effect on achieving treatment responses. In line with this, no differences in response rates were observed when assessing the subgroup of patients with hyperleukocytosis. Furthermore, response analysis in propensity score matched cohorts demonstrated a trend towards an antagonistic effect of hydrea pretreatment for the achievement of CCyR/MR2.0. A limitation of this study is the retrospective character, assessing historical response data only until 2015. However, the time frame of our study corresponds with the implementation of first and second generation TKI in CML treatment recommendations and in clinical practice. Our population-based data indeed demonstrate a gradual decrease in the use of hydrea pretreatment. Still, even after 2010 at least 35% of patients were receiving hydrea pre-

Haematologica | 107 August 2022

1942

LETTER TO THE EDITOR treatment, including patients without a high leukemic load or symptom burden. To our knowledge, this is the first study comparing clinical outcome in patients in relation to hydrea pretreatment in the TKI era. Our results indicate that early cytoreduction with hydrea has no added value in the treatment of CML for achieving the ELN response milestones and support current recommendation on restricting the use of hydrea pretreatment to patients with a symptomatic hyperleukocytosis or symptomatic splenomegaly while waiting for confirmatory diagnostic testing. Hydrea should not delay the start of TKI treatment as soon as the diagnosis is confirmed. More restrictive use of hydrea might shorten this delay and hematologists should be aware of the additional hematotoxicity of hydrea pretreatment on subsequent TKI therapy.

of Hematology, Medical Center Leeuwarden, Leeuwarden, the Netherlands. Correspondence: C.C.B. KOCKEROLS - c.c.b.kockerols@asz.nl https://doi.org/10.3324/haematol.2021.280501 Received: December 22, 2021. Accepted: March 24, 2022. Prepublished: March 31, 2022. ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

Disclosures The Pharos-CML database was financially supported by grants from Novartis and BMS to the Netherlands Comprehensive Cancer

Authors

Organisation (IKNL). The authors have no other conflicts of interest to disclose.

Camille C.B. Kockerols,1 Inge Geelen,2 Mark-David Levin,1 Jeroen J.W.M. Janssen, Avinash G. Dinmohamed, 3

4,5

Contributions

Mels Hoogendoorn, Jan 6

J. Cornelissen and Peter E. Westerweel 2

CK curated, analyzed and visualized the data, and wrote the first draft of the manuscript. PW conceptualized the research idea,

provided supervision and wrote the final draft of the manuscript. IG Department of Internal Medicine, Hematology, Albert Schweitzer

provided an initial curation of the database and revised the

Hospital, Dordrecht; Department of Hematology, Erasmus

manuscript together with the other co-authors. All authors revised

University Medical Center, Rotterdam; Department of

and approved the final version of the manuscript.

Hematology, Amsterdam University Medical Center, location VU Medical Center, Amsterdam; 4Department of Research &

Data-sharing statement

Development, the Netherlands Comprehensive Cancer

Data can be made available on request to other researchers, when

Organisation (IKNL), Utrecht; Department of Public Health,

in collaboration with the Dutch Cancer Registry, which is the owner

Erasmus University Medical Center, Rotterdam and 6Department

of the data.

References 1. Hochhaus A, Saussele S, Rosti G, et al. Chronic myeloid leukaemia: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2018;29(Suppl 4):Siv261. 2. Rinaldi I, Reksodiputro AH, Jusman SW, et al. Longer hydroxyurea administration prior to imatinib mesylate is risk factor for unsuccessful major molecular response in chronicphase chronic myeloid leukemia: possibility of P-glycoprotein role. Asian Pac J Cancer Prev. 2019;20(12):3689-3695. 3. Zhaleiko IO, Perekhrestenko TP, Bilko DI, Dyagil IS, Bilko NM. Determination of the optimal chemotherapy drugs pretreatment time through cultivation of hemopoietic cells in CML-patients treated with tyrosine kinase inhibitors. Exp Oncol.

2014;36(2):112-116. 4. Hochhaus A, Baccarani M, Silver RT, et al. European LeukemiaNet 2020 recommendations for treating chronic myeloid leukemia. Leukemia. 2020;34(4):966-984. 5. Geelen IGP, Thielen N, Janssen JJWM, et al. Treatment outcome in a population-based, 'real-world' cohort of patients with chronic myeloid leukemia. Haematologica. 2017;102(11):1842-1849. 6. Geelen IGP, Thielen N, Janssen JJWM, et al. Omitting cytogenetic assessment from routine treatment response monitoring in chronic myeloid leukemia is safe. Eur J Haematol. 2018;100(4):367-371.

Haematologica | 107 August 2022

1943

LETTER TO THE EDITOR

Preclinical evaluation of the preservation of red blood cell concentrates by hypoxic storage technology for transfusion in sickle cell disease Sickle cell disease (SCD) is a congenital hemolytic anemia characterized by vaso-occlusive crises. Red cell exchange transfusion is commonly used to treat both acute and chronic complications.1,2 Improving the quality of stored red blood cells (RBC) is essential to reduce transfusion frequency and to limit their pathological interactions with hemolysis damaged endothelial cells (EC).3,4 Several groups demonstrated an improvement of RBC properties when stored under hypoxia,5–9 mainly due to a reduction of oxidative damage. However, the use of hypoxic red blood cell concentrates (RBCC) for transfusion of SCD patients has not been specifically explored yet. To that purpose, we evaluated hypoxic RBC properties that could be relevant for transfusion of SCD patients throughout storage (up to 42 days at 4°C) compared to conventionallystored normoxic RBCC. We demonstrated a non-inferiority of hypoxic storage on i) RBC senescence parameters (intracellular calcium entry, reactive oxygen species [ROS] and phosphatidylserine exposure [PS]), ii) activation and interaction of EC with sickle RBC and an improvement in i) hemolysis, ii) RBC adhesion to thrombospondin 1 (TSP1), compared to normoxic storage. These results suggest that RBCC hypoxic storage could meet the safety criteria for transfusion of SCD patients and may provide clinical benefits, specifically by reducing free hemoglobin generation through storage (which could damage EC during transfusion) and RBC interaction with TSP-1 (involved in acute chest syndrome [ACS]10 and RBC recruitment on EC in SCD11). For each experiment, two whole blood units were collected from two compatible ABO healthy volunteers, leukofiltred, pooled together and split into two RBCC, one of them being deoxygenated. Both hypoxic and normoxic bags were stored for 42 days at 4°C (Figure 1A). RBCC were successfully deoxygenated with an oxygen saturation remaining inferior to 21% throughout storage (Online Supplementary Table S1). Hemolysis was measured in RBCC supernatants by the quantification of free oxyhemoglobin (HbO2) (Figure 1B). Despite a higher HbO2 concentration at the beginning of the storage in the hypoxic bags, accumulation of free HbO2 over time was significantly reduced under hypoxia. However, this free hemoglobin (Hb) was more oxidized under hypoxia, as suggested by the free methemoglobin (MetHb) concentration (Online Supplementary Figure S1A), while intra-RBC MetHb showed no significant difference between both storage conditions (Online Supplementary Figure S1B).

RBC senescence was evaluated by the measurement of intracellular ROS, calcium influx into RBC and extracellular membrane exposure of phosphatidylserine (PS). While no difference in ROS concentrations was found on day 0, a significant reduction (16%) was observed in RBC stored under hypoxia for 21 days (Figure 1C, when both groups were compared at one time point). Calcium influx (Figure 1D) was significantly decreased in hypoxic RBC on day 21 and 42 (reduction of 12% and 18%, respectively, when both groups were compared at one time point). PS exposure at the RBC surface was similar between both storage methods (Online Supplementary Figure S1C). Evaluation of intracellular ROS and calcium influx over time during the entire storage period only showed a trend in favor of hypoxic storage, probably due to the small sample size. Increased adhesion to TSP-1 is one of the senescence features that appears throughout RBC storage. RBC adhesion to TSP-1 was measured under physiologic flow with a shear stress of 0.5 dyn/cm2, on RBC incubated for 24 hours (h) in plasma from healthy donors (HD) or SCD patients at steady state or during an ACS (Figure 2). At the beginning of the storage (day 0, Figure 2A) hypoxia reduced RBC adhesion to TSP-1, irrespective of plasma origin. Evaluation of adhesion over time during the entire study period showed a significant increase with storage time and hypoxic conditions consistently trended lower than normoxic but the differences were not significant (Figure 2B to D), probably due to the low sample size. In order to assess RBC binding force to TSP-1, shear stress was increased to 2, then 5 dyn/cm2 after the initial adhesion step at 0.5 dyn/cm2 (Online Supplementary Figure S2 and data not shown for 5 dyn/cm2). Similar results were obtained with no significant difference between groups on day 21 and 42 of storage. At 5 dyn/cm2 hypoxic RBC incubated with HD plasma showed a significantly reduced adhesion over time (Figure 2E). No significant difference was observed between non-incubated hypoxic and normoxic RBC (Online Supplementary Figure S2E). In order to evaluate the effect of hypoxic RBC perfusion on EC, human umbilical vein endothelial cells (HUVEC) were perfused for 4 h with hypoxic or normoxic RBC (sampled on day 0, 21, 42) diluted in HD serum and supplemented with or without 10% hemolysate (conditioning step) to reproduce acute hemolysis that can be observed in SCD patients (Figure 3A). HUVEC activation marker levels (E-selectin and Intercellular Adhesion Molecule 1) were assessed by flow cytometry and showed no signifi-

Haematologica | 107 August 2022

1944

LETTER TO THE EDITOR cant differences between hypoxic and normoxic RBC per- on interactions between sickle RBC and EC (Figure 3B and fusion (data not shown). Adhesion flow experiments with C). No significant difference was observed between norSCD blood were also performed after HUVEC conditioning, moxic and hypoxic RBC perfusions except a significant in order to evaluate the effect of hypoxic RBC perfusion 20% decrease of RBC adhesion on EC previously perfused

Figure 1. Hemolysis and senescence measurements during red blood cell concentrates storage. (A) Diagram of blood bag preparation and storage. Whole blood was collected from two ABO compatible healthy donors for each experiment (n=6). After white blood cell (WBC) removal, red blood cells (RBC) were pooled and split into two red blood cell concentrate (RBCC) bags. One of them was conventionally stored (normoxic) while the other underwent deoxygenation through Hemanext Inc. technology. Both bags were stored at 4°C for 42 days. Sampling was performed on day 0, 21 and 42 of storage for blood gas, hemolysis, senescence and adhesion measurements. (B) Free oxyhemoglobin (HbO2) measured by spectrophotometry in normoxic (black) or hypoxic (grey) bag supernatants throughout the 42 days of storage. Mean +/- standard error of the mean (SEM) of 6 independent experiments. *Storage method effect, #interaction time/storage method, •comparison between hypoxic and normoxic storage at one time point. (C and D) Senescence parameters measured by flow cytometry in RBC from normoxic (black) or hypoxic (grey) bags throughout 42 days of storage, (C) intracellular reactive oxygen species (ROS) concentration measured using CM-H2DCFDA fluorescent probe, (D) calcium influx measured using Fluo-3/AM fluorescent probe. Mean +/- SEM of 6 independent experiments. *Storage method effect, # interaction time/storage method, •comparison between hypoxic and normoxic storage at one time point. ns: not significant. MFI: mean fluorescence intensity, A.U.: arbitrary units. Haematologica | 107 August 2022

1945

LETTER TO THE EDITOR with hemolysate and hypoxic RBC collected at day 42, perfusion with normoxic RBC and could even be beneficial compared to normoxic RBC (Figure 3B). in specific contexts (at the end of storage and in the presAltogether these data suggest that hypoxic RBC exposure ence of acute hemolysis). under flow does not excessively damage EC compared to Hypoxic storage of RBCC was demonstrated to reduce

Figure 2. Red blood cell adhesion to TSP-1 at 0.5 and 5 dyn/cm2. Red blood cells (RBC) from normoxic or hypoxic bags were sampled on day 0, 21 and 42 of storage and incubated at 37°C for 24 hours (h) with plasma from healthy donors (HD), steady state sickle cell disease patients (SCD) or sickle cell disease patients with an acute chest syndrome (ACS SCD). RBC were washed and resuspended at 0.66% hematocrit and perfused into TSP-1 coated channels at a shear stress of 0.5 dyn/cm2 for 5 minutes (min) for adhesion, then washed for 5 min before quantification of adherent RBC on a light microscope. RBC adhesion strength was evaluated by increasing shear stress to 2 dyn/cm2 for 2 min 30 seconds (sec) and at 5 dyn/cm2 for 2 min 30 sec (adherent RBC quantification at the end of each period). (A) RBC adhesion to TSP-1 (0.5 dyn/cm2) on day 0 of storage after 24 h incubation with HD, SCD or ACS SCD plasma. RBC adhesion to TSP-1 (0.5 dyn/cm2) on day 0, 21 and 42 of storage after 24 h incubation with (B) HD plasma , (C) SCD plasma, (D) ACS SCD plasma or (E) HD plasma at 5 dyn/cm2. Mean +/- standard error of the mean of 6 independent experiments, *storage method effect, #interaction time/storage method, •comparison between hypoxic and normoxic storage at one time point. ns: not significant. RBCC: red blood cell concentrate. Haematologica | 107 August 2022

1946

LETTER TO THE EDITOR

Figure 3. Red blood cell adhesion to endothelial cells. (A) Microfluidic model diagram. Human umbilical endothelial cells (HUVEC) were perfused for 16-40 hours (h) with culture medium in microslides coated with fibronectin (step 1), then perfused for 4 h at 1 dyn/cm2 with red blood cells (RBC) from normoxic or hypoxic bags, diluted at 25% in compatible serum (from healthy donor, HD) and with the addition (or not) of 10% hemolysate (step 2). After the conditioning step (step 2), HUVEC were washed with culture medium and perfused with whole blood from sickle cell disease (SCD) patients at 1 dyn/cm2 for 10 minutes before a second wash with culture medium to assess RBC adhesion to endothelial cells. All steps (except washes) were performed at 37°C. (B and C) Adhesion of RBC from steady state SCD patients to HUVEC perfused for 4 h with normoxic or hypoxic RBC sampled on day 0, 21 and 42 of storage, diluted at 25% in HD serum and with (B) 10% hemolysate or (C) 0% hemolysate. Adherent RBC were quantified on a light microscope. Mean +/- standard error of the mean of 6 independent experiments, *storage method effect. #interaction time/storage method. •Comparison between hypoxic and normoxic storage at one time point. ns: not significant. Haematologica | 107 August 2022

1947

LETTER TO THE EDITOR storage lesion and improve RBC quality in vitro,5,6,8,9 as well as RBC survival after transfusion in healthy volunteers.7,12 However, the use of hypoxic RBC for the transfusion of SCD patients had not yet been explored. In this study, we demonstrated a non-inferiority of RBCC hypoxic storage compared to standard normoxic RBCC conservation. Moreover, several parameters were shown to be improved with hypoxic storage, which suggests a potential beneficial impact for transfusion of SCD patients. In accordance with previous studies performed with different hemolysis measurement methods,7,9 we observed a reduced hemolysis in hypoxic bags over time, except at the beginning of storage. The observed increase in hemolysis on day 0 may be due to the extra preparation steps required for deoxygenation. However, although a hypoxic environment should reduce hemoglobin oxidation to methemoglobin, we still observed a higher methemoglobin concentration in the supernatant under hypoxia. This result is probably linked to the increased Hb autoxidation rate at low O2 pressure.13 However, the relatively low quantity of free Methemoglobin (23 mM on day 42 vs. 15 mM in normoxic supernatant), once diluted in patient circulation, does not seem to be a concern. In addition, we did not observe any difference in intra-RBC methemoglobin content. We also evaluated the effect of hypoxia on RBC senescence markers, but the improvement observed was limited. Indeed, senescence marker appearance is also modest with conventional normoxic storage, which is consistent with previous studies and could have limited hypoxic storage impact. However, these improvements in the measured in vitro metrics of hypoxic RBC still support the improved post-transfusion recovery observed in healthy volunteers and could be of interest in reducing the frequency of transfusion in SCD patients. This study also showed a significant decrease in the adhesion of hypoxic RBC to TSP-1, but only at the beginning of the storage. Such a result suggests that the reduced RBC adhesion to TSP-1 is not linked to a reduction of oxidative damage but rather to other mechanisms induced by hypoxia such as metabolic changes for example. However, a decreased interaction with TSP-1 may be beneficial to patients undergoing ACS, who present high TSP-1 plasma level. Finally, this study did not identify any detrimental impact of hypoxic RBC on EC, only a slight reduction of their adhesion properties induced by hemolysate, towards RBC from SCD patients. Hypoxic RBC were successfully transfused in healthy volunteers and were shown to have higher 24 h post-transfusion recovery after 42 days of storage7,12. However, SCD patients present specific features which could require an improved monitoring during clinical trials. The first issue is the risk of oxygen capture by transfused

RBC, at the expense of recipient’s RBC, which could entail their sickling. In this respect, in vitro experiments were performed with a mixture of hypoxic RBC and RBC from SCD patients, which did not show any significant sickling increase.14 The second issue is the risk of pulmonary arterial vasoconstriction during transfusion. As a matter of fact, the transfusion of large amounts of hypoxic RBC could reduce vascular pO2, which could lead to vasoconstriction. However, if donor RBC are infused slowly enough into the body, a sudden drop in vascular pO2 may not be likely. Particular attention should be paid to patients with pulmonary hypertension during clinical trials since they could be more prone to vascular constriction.

Authors Laura Bencheikh,1,2 Kim-Anh Nguyen,2,3 Philippe Chadebech,1,2 Laurent Kiger,2 Gwellaouen Bodivit,1,2 Alicia Jouard,1,2 Sadaf Pakdaman,1 Sandia Adypagavane,1,2 Etienne Audureau,4 Khouloud Tebbakha,1 Thibaut Bocquet,1 Blandine Mignen,1 Nicolas Hebert,1,2 Marion Seguin,2 France Pirenne,1,2 Samuel Sowemimo-Coker,5 Andrew Dunham5 and Pablo Bartolucci2,4 1

Etablissement Français du Sang (EFS) Ile-de-France, France;

Université Paris Est Creteil, INSERM, IMRB, Creteil, France; 3Institut

Imagine, Paris, France; 4CHU Henri Mondor, Assistance Publique– Hôpitaux de Paris, Creteil, France and 5Hemanext Inc, Lexington, MA, USA Correspondence: P. BARTOLUCCI - pablo.bartolucci@aphp.fr https://doi.org/10.3324/haematol.2021.279721 Received: August 4, 2021. Accepted: March 24, 2022. Prepublished: March 31, 2022. ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

Disclosures Hemanext Inc supplied the funding for hiring LB. AD and SSC are employees of Hemanext Inc. PB declares being member on a standing advisory council or committee consultancy for Addmedica, Roche, Bluebird bio, Emmaus, Agios, Global Blood Therapeutics, Novartis and Hemanext Inc. PB declares being co‐founder of INNOVHEM. The other authors declare no competing financial interests. Contributions LB, PC, LK, GB, AJ, SA, NH and MS performed research. KAN, FP and PB designed the study. LB, LK, EA, and PB performed data analysis.

Haematologica | 107 August 2022

1948

LETTER TO THE EDITOR

SP and KT selected and enrolled patients. TB and BM selected healthy donor samples and prepared red blood cell concentrates. LB, SSC and PB wrote the manuscript. FP, SSC, AD and PB supervised the study.

Français du Sang”, the Université Paris Est Créteil (UPEC) and INSERM. Data-sharing statement The data that support the findings of this study are available on request from the corresponding author, PB.

Funding This study was supported by Hemanext Inc, the “Etablissement

References 1. 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. 2. Vichinsky EP, Neumayr LD, Earles AN, et al. Causes and outcomes of the acute chest syndrome in sickle cell disease. National Acute Chest Syndrome Study Group. N Engl J Med. 2000;342(25):1855-1865. 3. Kucukal E, Ilich A, Key NS, Little JA, Gurkan UA. Red blood cell adhesion to heme‐activated endothelial cells reflects clinical phenotype in sickle cell disease. Am J Hematol. 2018;93(8):1050-1060. 4. Belcher JD, Mahaseth H, Welch TE, et al. Critical role of endothelial cell activation in hypoxia-induced vasoocclusion in transgenic sickle mice. Am J Physiol Heart Circ Physiol. 2005;288(6):H2715-H2725. 5. D’Alessandro A, Nemkov T, Blair A, et al. Anaerobic storage condition enhances GSHlLevels while maintaining pentose phosphate pathway activity. Transfusion. 2016;56(S4):51A. 6. Burns JM, Yoshida T, Dumont LJ, Yang X, Piety NZ, Shevkoplyas SS. Deterioration of red blood cell mechanical properties is reduced in anaerobic storage. Blood Transfus. 2016;14(1):80-88. 7. DʼAlessandro A, Yoshida T, Nestheide S, et al. Hypoxic storage of red blood cells improves metabolism and post-transfusion recovery. Transfusion. 2020;60(4):786-798.

8. Dumont LJ, D’Alessandro A, Szczepiorkowski ZM, Yoshida T. CO2 -dependent metabolic modulation in red blood cells stored under anaerobic conditions. Transfusion. 2016;56(2):392-403. 9. Yoshida T, Blair A, D’alessandro A, et al. Enhancing uniformity and overall quality of red cell concentrate with anaerobic storage. Blood Transfus. 2017;15(2):172-181. 10. Novelli EM, Kato GJ, Ragni MV, et al. Plasma thrombospondin-1 is increased during acute sickle cell vaso-occlusive events and associated with acute chest syndrome, hydroxyurea therapy, and lower hemolytic rates. Am J Hematol. 2012;87(3):326-330. 11. Brittain HA, Eckman JR, Swerlick RA, Howard RJ, Wick TM. Thrombospondin from activated platelets promotes sickle erythrocyte adherence to human microvascular endothelium under physiologic flow: a potential role for platelet activation in sickle cell vaso-occlusion. Blood. 1993;81(8):2137-2143. 12. Dumont LJ, Yoshida T, AuBuchon JP. Anaerobic storage of red blood cells in a novel additive solution improves in vivo recovery. Transfusion. 2009;49(3):458-464. 13. Bonaventura C, Henkens R, Alayash AI, Banerjee S, Crumbliss AL. Molecular controls of the oxygenation and redox reactions of hemoglobin. Antioxid Redox Signal. 2013;18(17):2298-2313. 14. Sowemimo-Coker SO, Kuypers FA, Larkin S, Tung G, Dunham A. Effects of hypoxic red blood cells on sickling kinetics of red blood cells from patients with sickle cell disease. Transfusion. 2019;59(Suppl 3):S157A.

Haematologica | 107 August 2022

1949

LETTER TO THE EDITOR

Hemostatic and protein C pathway dysfunction in the pathogenesis of experimental cerebral malaria Accumulating data suggest that hemostatic dysfunction contributes to Plasmodium falciparum malaria pathogenesis.1 In addition, specific mechanisms through which the protein C pathway modulates P. falciparum pathogenesis have been described.1 We hypothesized that the anticoagulant and anti-inflammatory activities of recombinant activated protein C (APC) may possess therapeutic value in the setting of cerebral malaria (CM). In order to address this hypothesis, we assessed hemostatic parameters in an established murine model of experimental cerebral malaria (ECM), and using the same model, investigated the ability of recombinant APC to ameliorate ECM. In keeping with findings in patients with severe P. falciparum malaria, we observed that dysregulated thrombin generation and protein C pathway dysfunction were both late features of ECM. Furthermore, pretreatment with a monoclonal antiEPCR antibody that blocks protein C/APC binding prior to P. berghei inoculation significantly reduced overall survival. Conversely, mice treated with recombinant APC exhibited a marked attenuation in clinical ECM progression and parasitemia, in parallel with a significant increase in overall survival. All together, these findings confirm that hemostatic and protein C pathway dysfunction are both consistent features in human and ECM, and demonstrate for the first time a role for recombinant APC in reducing clinical progression and mortality in ECM. CM is a life-threatening complication of P. falciparum infection characterized by ataxia, seizures, altered consciousness and coma. Although CM is associated with significant mortality, the biological mechanisms underlying its pathogenesis remain poorly defined. Significant coagulation cascade activation including elevations in levels of fibrin degradation products (FDP) and thrombin-antithrombin (TAT) complexes are common in patients with P. falciparum infection.2,3 Furthermore, Moxon et al. recently described overt DIC in 19% of children with retinopathypositive CM.3 Together, these findings suggest that hemostatic dysfunction may contribute to malaria pathogenesis. This hypothesis is supported by the observation that coagulation activation correlates with peripheral blood parasitemia levels, and is inversely-related to overall survival.4,5 Moreover, microvascular fibrin deposition has also been reported in postmortem studies of patients with fatal CM.6 A number of molecular mechanisms through which P. falciparum infection triggers coagulation activation have been described.1 Recent studies have also highlighted specific mechanisms through which the protein C pathway influences P. falciparum pathogenesis.6,7 Importantly, both

thrombomodulin (TM) and the endothelial protein C receptor (EPCR) can bind to PfEMP1 expressed on P. falciparum to infected erythrocyte (IE) surfaces,7,8 thereby facilitating IE cytoadhesion to endothelial cells (EC). PfEMP1 binding to EPCR also limits generation of anticoagulant activated protein C (APC), and inhibits EPCR-dependent PAR1-mediated protection of EC barrier integrity.9,10 The typical late clinical presentation of patients with CM makes it difficult to determine whether hemostatic dysfunction directly contributes to the pathogenesis of CM, or whether coagulation activation merely constitutes a secondary epiphenomenon. Therefore, in this study we sought to further investigate the role of coagulation activation and the protein C pathway in malaria pathogenesis using an established murine model of ECM, in which C57BL/6J mice were infected with P. berghei ANKA.11 We have previously described in detail the murine experimental cerebral malaria (ECM) model. All mouse experiments were performed in compliance with the Irish Medicines Board and approved by the Trinity College Dublin BioResource Ethics Committee. Mice aged 8 to 10 weeks were infected by intraperitoneal injection of 2x106 P. berghei ANKA. Following inoculation, malaria progression was monitored using a previously validated ECM clinical scoring system.11 P. berghei parasitemia levels were monitored by Giemsa-stained thin blood smears. Platelet counts were measured using a Sysmex haematology analyser (KX-21N). In order to prepare platelet-poor plasma, blood samples were centrifuged at 1,500 g for 15 minutes at 20°C, aliquoted and stored at -80°C until use. Murine plasma activated partial thromboplastin time (APTT) was determined using a commercial kit (C.K. Prest, Stago) and time for clot formation was determined using Amelung KC4 Micro Clinical Coagulation Analyzer (Amelung, Trinity Biotech, Ireland). Similarly, PT was determined using Hemosil recombiplastin 2G according to the manufacturer’s guidelines. Plasma levels of thrombin-antithrombin (TAT) complexes, soluble thrombomodulin (sTM) and soluble endothelial cell protein C receptor (EPCR) were quantified using specific commercial enzyme-linked immunsorbant assays (ELISA) (Abcam, Cambridge, UK and R&D Systems, UK) according to the manufacturer’s instructions. In order to study the role of EPCR, ECM progression was assessed in mice pretreated with 50 µg of an EPCR blocking antibody RCR-16,12 or isotype control antibody, immediately prior to P. berghei infection. Recombinant murine APC was generated, expressed, and purified as previously described.13 Consistent with previous studies

Haematologica | 107 August 2022

1950

LETTER TO THE EDITOR assessing the anti-inflammatory properties of recombinant APC in murine endotoxemia models,14 mice were treated with 10 µg mAPC administered intravenously immediately prior to P. berghei infection. Subsequently another 10 µg of mAPC was administered 4 hours post-infection. Alternatively, mice were treated with two 10 µg mAPC boluses administered 4 hours apart on day +3 following P. berghei inoculation. Progressive clinical features similar to those observed in human CM were evident in mice from day +3 following infection (Figure 1A). Peripheral P. berghei parasitemia levels increased progressively (Figure 1B), and mice typically died within 7-10 days. In keeping with previous findings in patients with severe P. falciparum malaria, a significant increase in APTT was observed by day +5 following P. berghei inoculation (Figure 1C).2 Again in keeping with human studies,3 no significant changes in either PT or murine fibrinogen levels were seen through the course of P. berghei infection (data not shown). Nevertheless, by day +5 after inoculation, median plasma TAT levels were increased approximately 2.5 fold (2.0 ng/mL at day 0 vs. 5.5 ng/mL at day +5; P<0.01) (Figure 1D), which was similar in magnitude to the increase in plasma TAT levels observed in patients

with CM.2,4,6 In contrast to the early increase in plasma VWF levels that occurs within 24 hours following P. berghei infection,11 plasma TAT levels remained normal for 72 hours post-infection. Murine plasma protein C and antithrombin levels were not significantly reduced during P. berghei infection (data not shown). However, again as reported in previous findings in children infected with P. falciparum,3,6 plasma levels of both sTM and sEPCR were significantly elevated in C57BL/6J mice following P. berghei infection (sTM 6.5 ng/mL at day 0 vs. 21.6 ng/mL at day +5; P<0.001 and sEPCR 0.99 ng/mL at day 0 vs. 3.50 ng/mL at day +5; P<0.001) (Figure 1E and F). In keeping with the hypothesis that hemostastic dysfunction is a relatively late feature in ECM, the increases in plasma sTM and sEPCR were not observed until day +5 following innoculation. Collectively, these findings demonstrate that dysregulated thrombin generation represents a consistent feature of both human and experimental murine CM. Critically however, unlike the acute EC activation that represents an early hallmark in both murine and human malaria, hemostatic and protein C pathway dysfunction both develop at a much later stage. In order to investigate whether protein C pathway dysfunction contributes to ECM pathogenesis, mice were pre-

Figure 1. Hemostatic and protein C pathway dysfunction constitute late features of experimental cerebral malaria. (A) Following intraperitoneal inoculation with 2x106 Plasmodium berghei ANKA parasites, experimental cerebral malaria (ECM) progression was followed in wild-type (WT) C57BL/6J mice (n=12) using a validated clinical scoring algorithm. Results presented represent the mean values ± standard error of the mean (SEM) unless otherwise stated. (B) Following P. berghei infection, peripheral blood parasitemia levels were determined each day using Giemsa-stained smears (n=12). By day +5 following P. berghei infection, murine activated partial thromboplastin time (APTT) levels (C) and plasma thrombin-antithrombin (TAT) complex levels (D) were both significantly increased (*P<0.05, **P<0.01, ***P<0.0001 respectively). Similarly, plasma levels of soluble endothelial cell protein C receptor (sEPCR) (E) and soluble thrombomodulin (sTM) (F) were also both significantly elevated by day +5. Data are expressed as mean values ± SEM. In order to assess statistical differences, data were analyzed using Student’s unpaired 2-tailed t-test. ECM clinical scoring data were assessed by two-way ANOVA analysis. Haematologica | 107 August 2022

1951

LETTER TO THE EDITOR treated with a monoclonal anti-EPCR antibody (RCR-16) previously shown to block protein C/APC binding,12 or an isotype control antibody, immediately prior to P. berghei inoculation. We observed significantly reduced overall survival (P<0.05) in the cohort of mice treated with RCR-16 (Figure 2A), suggesting that EPCR-dependent APC generation and/or signaling is important for controlling ECM development. Case studies involving a small number of patients with severe P. falciparum malaria treated with APC have reported variable effects.15 Consequently, we further investigated whether administration of recombinant mAPC influenced ECM pathogenesis. Mice were treated with 10 µg mAPC immediately prior to P. berghei infection, and a subsequent second 10 µg mAPC dose was administered 4

hour later. Mice treated with recombinant mAPC exhibited a mild but significant reduction in parasitemia at day +4 (Figure 2B; P<0.001). In contrast however, the APC-treated mice demonstrated significantly attenuated clinical ECM progression (Figure 2C; P<0.001) and weight loss (Figure 2D; P<0.001). Furthermore, mAPC administration also caused a significant increase in overall survival (Figure 2E; P<0.001). In order to investigate whether mAPC administered later in the course of ECM can still influence clinical progression, the effect of administering mAPC on day +3 was investigated. Once again, clinical progression (Figure 3A) and overall survival (Figure 3B) were both marginally improved in mice treated with mAPC compared to controls (P<0.05). Interestingly, however, the magnitude of this ef-

Figure 2. Recombinant anticoagulant activated protein C significantly attenuates clinical progression and markedly improves overall survival in experimental cerebral malaria. (A) In order to determine if endothelial protein C receptor (EPCR) plays a role in experimental cerebral malaria (ECM) pathogenesis, mice were pretreated with the EPCR blocking antibody RCR-16 or an isotype control antibody prior to infection with 2x106 Plasmodium berghei ANKA parasites. Twelve mice were treated in the RCR-16 and 11 in the isotype control groups. Overall survival was significantly reduced in mice treated with RCR-16. In order to investigate whether recombinant murine anticoagulant activated protein C (mAPC) administration influences ECM progression, mice were pretreated with 10 µg mAPC immediately prior to P. berghei infection and a second 10 µg mAPC dose was administered 4 hour later. Twelve mice were treated in the mAPC and control groups. Mice treated with recombinant mAPC exhibited (B) mildly reduced parasitemia at day +4, (C) attenuated clinical ECM progression (D) reduced weight loss and (E) significantly increased overall survival (*P<0.05, **P<0.01, ***P<0.001). Haematologica | 107 August 2022

1952

LETTER TO THE EDITOR

Figure 3. Recombinant anticoagulant activated protein C later in the course of experimental cerebral malaria has only mild effects on clinical progression and survival. In order to investigate whether murine anticoagulant activated protein C (mAPC) later in the course of experimental cerebral malaria (ECM) can still influence clinical progression, 2 doses of mAPC were administered on day +3 following Plasmodium berghei innoculation. Once again, (A) clinical progression and (B) overall survival were both mildly improved in mice treated with mAPC compared to controls. Mouse survival data were compared using a log-rank (Mantel-Cox) Test.

fect was markedly less than that observed in mice treated with mAPC on day 0. This attenuated efficacy of APC administered later in the disease course is in keeping with the concept that CM is associated with progressive shedding of EPCR and TM from EC surfaces.10 In conclusion, our findings demonstrate that hemostatic and protein C pathway dysfunction are both consistent features in human and experimental murine CM. We also show for the first time that recombinant mAPC markedly reduces clinical progression and overall mortality in ECM. Further studies will be required to elucidate the molecular mechanisms through which APC modulates ECM pathogenesis, together with the optimal APC dosing and treatment regimen. Nevertheless, given the significant morbidity and mortality that are still associated with CM, novel adjunctive therapies to limit vascular dysfunction and slow disease progression are urgently required.

Correspondence: J. O’DONNELL - jamesodonnell@rcsi.ie https://doi.org/10.3324/haematol.2021.280450 Received: December 2, 2021. Accepted: April 8, 2022. Prepublished: April 21, 2022. ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

Disclosures RP has received research grant funding awards from Novo Nordisk Daiichi Sankyo and Bayer. JOD has served on the speaker’s bureau for Baxter, Bayer, Novo Nordisk, Boehringer Ingelheim, Leo Pharma and Octapharma; he has served on the advisory boards of Baxter, Bayer, Octapharma CSL Behring, Daiichi Sankyo, Boehringer Ingelheim and Pfizer; he has also received research grant funding

Authors

awards from Baxter, Bayer, Pfizer and Novo Nordisk. Contributions

Niamh O’Regan, Kristina Gegenbauer, Eimear M. Gleeson, Kenji

NO’R, KG, EMG, JMOS, CD, ND, AC, and TMB performed experiments;

Fukudome, Jamie M. O’Sullivan, Clive Drakeford, Niall Dalton, Alain

NO’R, KG, EMG, JMOS, KJ, TMB, OPS, RJSP and JSO’D designed the

Chion, Teresa M. Brophy, Owen P. Smith, Roger J.S. Preston

research and analyzed the data. All authors were involved in writing

1,2

1,4

James S. O’Donnell

1,2

2,4

and

and reviewing the paper.

2,4,6

National Children's Cancer Service, Children's Health Ireland at

Funding

Crumlin, Dublin, Ireland; 2Systems Biology Ireland, School of

This work was supported by funds from the NIH for the Zimmerman

Medicine, University College Dublin, Dublin, Ireland; 3Department of

Program (HL081588); a Science Foundation Ireland Principal

Life Science, Saga University Organization for General Education,

Frontiers for the Future (FFP) award (20/FFP-A/8952), a Health

Saga, Japan; Irish Center for Vascular Biology, School of Pharmacy

Research Board Investigator Lead Project Award (ILP-POR-2017-008)

and Biomolecular Sciences, Royal College of Surgeons in Ireland,

and a National Children’s Research Center Project Award (C/18/1).

Dublin, Ireland; Hematology Department, Our Lady's Children's 5

Hospital, Dublin, Ireland and 6National Coagulation Center, St

Data-sharing statement

James’s Hospital, Dublin, Ireland

All original data and protocols can be made available to other investigators upon request.

Haematologica | 107 August 2022

1953

LETTER TO THE EDITOR

References 1. O'Sullivan JM, Preston RJ, O'Regan N, O'Donnell JS. Emerging roles for hemostatic dysfunction in malaria pathogenesis. Blood. 2016;127(19):2281-2288. 2. Clemens R, Pramoolsinsap C, Lorenz R, Pukrittayakamee S, Bock HL, White NJ. Activation of the coagulation cascade in severe falciparum malaria through the intrinsic pathway. Br J Haematol. 1994;87(1):100-105. 3. Moxon CA, Chisala NV, Mzikamanda R, et al. Laboratory evidence of disseminated intravascular coagulation is associated with a fatal outcome in children with cerebral malaria despite an absence of clinically evident thrombosis or bleeding. J Thromb Haemost. 2015;13(9):1653-1664. 4. Holst FG, Hemmer CJ, Foth C, Seitz R, Egbring R, Dietrich M. Low levels of fibrin-stabilizing factor (factor XIII) in human Plasmodium falciparum malaria: correlation with clinical severity. Am J Trop Med Hyg. 1999;60(1):99-104. 5. Horstmann RD, Dietrich M. Haemostatic alterations in malaria correlate to parasitaemia. Blut. 1985;51(5):329-335. 6. Moxon CA, Wassmer SC, Milner DA, Jr., et al. Loss of endothelial protein C receptors links coagulation and inflammation to parasite sequestration in cerebral malaria in African children. Blood. 2013;122(5):842-851. 7. Turner L, Lavstsen T, Berger SS, et al. Severe malaria is associated with parasite binding to endothelial protein C receptor. Nature. 2013;498(7455):502-505. 8. Gysin J, Pouvelle B, Le Tonqueze M, Edelman L, Boffa MC. Chondroitin sulfate of thrombomodulin is an adhesion receptor

for Plasmodium falciparum-infected erythrocytes. Mol Biochem Parasitol. 1997;88(1-2):267-271. 9. Gillrie MR, Avril M, Brazier AJ, et al. Diverse functional outcomes of Plasmodium falciparum ligation of EPCR: potential implications for malarial pathogenesis. Cell Microbiol. 2015;17(12):1883-1899. 10. Gleeson EM, O'Donnell JS, Preston RJ. The endothelial cell protein C receptor: cell surface conductor of cytoprotective coagulation factor signaling. Cell Mol Life Sci. 2012;69(5):717-726. 11. O'Regan N, Gegenbauer K, O'Sullivan JM, et al. A novel role for von Willebrand factor in the pathogenesis of experimental cerebral malaria. Blood. 2016;127(9):1192-1201. 12. Ye X, Fukudome K, Tsuneyoshi N, et al. The endothelial cell protein C receptor (EPCR) functions as a primary receptor for protein C activation on endothelial cells in arteries, veins, and capillaries. Biochem Biophys Res Commun. 1999;259(3):671-677. 13. Harmon S, Preston RJ, Ni Ainle F, et al. Dissociation of activated protein C functions by elimination of protein S cofactor enhancement. J Biol Chem. 2008;283(45):30531-30539. 14. Kerschen EJ, Fernandez JA, Cooley BC, et al. Endotoxemia and sepsis mortality reduction by non-anticoagulant activated protein C. J Exp Med. 2007;204(10):2439-2448. 15. Rankin LG, Austin DL. The use of activated protein C in severe Plasmodium falciparum malaria. Anaesth Intensive Care. 2007;35(3):428-432.

Haematologica | 107 August 2022

1954

LETTER TO THE EDITOR

COVID-19 infection in acute lymphoblastic leukemia over 15 months of the pandemic. A Campus ALL report The spread of coronavirus disease 2019 (COVID-19), with the two peaks of infection documented worldwide (February 2020-June 2020 and September 2020-April 2021), represented a challenge in the management of patients with hematologic malignancies,1-3 typically immunosuppressed either because of their primary disease and/or because of treatment. This is particularly true among patients with acute lymphoblastic leukemia (ALL). However, given the rarity of this disease in adulthood, information is limited and based, so far, mainly on case reports,4-6 with only two larger series having been published.7,8 In order to define the clinico-biological features of the COVID-19-infected ALL population, and their ALL management and ALL outcome, as well as COVID-19-related variables, e.g. geographical distribution, source of infection, COVID-19-related support and sequelae, we conducted a cross-sectional, observational study in 34 Italian hematology centers within the nationwide Campus ALL network. The protocol (ref. 2694CESC on 30-04-2020) was approved by the Ethics Committee of the coordinating center and by the local institutional review boards of participating centers. All patients gave written informed consent to participation in the study. With regard to the geographic distribution of the centers that participated in the study, 17 were located in the north of Italy, 11 in the center and six in southern Italy. The period covered by the survey spanned from February 2020, the start of the first wave of the pandemic, to April 2021, to include the second peak. Out of 756 adults with ALL (237 with Philadelphia chromosome [Ph]-positive ALL, 363 with Ph-negative B-ALL and 156 with T-ALL) actively followed during this 15month period, 63 (8.3%, 95% confidence interval [95% CI]: 6.5-10.5) developed an infection with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), detected by molecular testing in all but one case. No patients had received any anti-SARS-CoV-2 vaccine at the time of the survey. All patients were monitored on a regular basis (in most centers every 3 days) if hospitalized, at hospital admission if outside, whereas for the few patients who were receiving a tyrosine kinase inhibitor (TKI), maintenance and/or were off-therapy (n=6), molecular testing was carried out if they had symptoms or they had a close contact with an infected individual. There was no preferential distribution among the various regions: 30 infections were documented in northern Italy, 19 in central Italy and 14 in southern Italy. Thus, the incidence of the infection in the ALL population in the geographical areas was 7%, 8% and

11.5%, respectively. Most cases were recorded during the second wave of the pandemic. All five COVID-19-positive cases (7.9%) that occurred between February 2020 and April 2020 were recorded in northern Italy, in line with the geographical spread of the first peak of the pandemic. One case (1.6%) was documented between May and August 2020, while 57 cases (90.5%) occurred between September 2020 and April 2021 (Figure 1A-C). The source of infection was nosocomial in 26 cases (41.3%), familial in 23 (36.5%), unknown in 13 (20.6%) and work-related in one (1.6%). The features of the 63 infected patients are summarized in Table 1. There was no preferential distribution according to ALL subtype, with COVID-19 occurring in 7.6% of patients with Ph-positive ALL, 7.7% of patients with Phnegative B-ALL and 10.9% of patients with T-ALL (P=n.s.). Of the infected patients, 36 (57.1%) had no comorbidities, 11 (17.5%) had one comorbidity and 16 (25.4%) had more than one comorbid condition (Table 1). COVID-19 infection was documented at the onset of the leukemia and/or during the induction phase in 14 cases (22.2.%), during consolidation in 13, which was TKI-based in two (20.6%), during chemotherapy maintenance in six (9.5%), after allogeneic transplantation in 15 (23.8%), during maintenance with a TKI in seven (11.1%) and off-treatment in six (9.5.%), while in two cases (3.2%) it occurred at relapse. Of the 63 COVID-19-positive patients with ALL, nine were asymptomatic, ten had only isolated fever, 36 had respiratory symptoms and eight presented other symptoms, including - but not limited to - ageusia and anosmia; notably, the Ph-positive ALL cases were rarely symptomatic (n=3). As a consequence, management of the infection was variable: 29 patients (46%) did not require hospitalization, 28 (44.4%) were hospitalized in a COVID ward and 13 of them required respiratory assistance; finally, six (9.6%) were transferred directly to an intensive care unit (Table 1). Seven patients succumbed to the infection, with an overall mortality rate of 11.1% (0.9% of the whole cohort); two were male and five were female; three were less than 65 years old. Three out of seven had T- ALL and all were infected at the onset of disease; three had Ph-positive BALL, two developed the infection during consolidation and the other at relapse. Finally, among Ph-positive ALL patients, only one death was recorded in an allografted case. In 54 patients (85.7%) there were no sequelae. In one patient pulmonary fibrosis was documented and in another the delay in treatment led to a relapse. In seven cases

Haematologica | 107 August 2022

1955

LETTER TO THE EDITOR

Figure 1. Incidence and distribution of COVID-19, and treatment interruption in patients with acute lymphoblastic leukemia. (A) Percentages of patients with acute lymphoblastic leukemia (ALL) with or without COVID-19. (B) Geographical distribution of infected cases, together with percentages in relation to the geographical area. (C) Distribution of infected cases according to period. (D) Treatment interruption according to the type of ALL.

(9.5%), the infection was still ongoing at the time of the survey and at the latest update (July 2021) it had resolved in all. Since a key aspect in ALL management is the adherence to the timing of treatment, we also investigated whether COVID-19-positive patients interrupted their ALL treatment during the infection (Figure 1D). Among 43 evaluable patients (patients who underwent hematopoietic cell transplantation - unless receiving therapy for graft-versus-host disease - or were off-treatment were excluded), ALL treatment was stopped in 28 (66.6%). Importantly, among 11 Ph-positive ALL patients only three (27.3%) suspended treatment: of these, one was receiving posttransplant treatment (ruxolitinib for graft-versus-host disease) and two were on maintenance therapy, whereas of the patients who continued treatment one was in the induction phase with a TKI, two were in consolidation and five were receiving maintenance. In contrast, among 23 cases of Ph-negative B-ALL, 19 (82.6%) stopped treatment. More in detail, six patients were receiving induction therapy, eight consolidation, two immunotherapy for relapse and three were on maintenance. Of the four patients who continued treatment, one was receiving induction, one consolidation and two were on maintenance. Likewise, also in T-ALL, seven of nine

patients (77.8%) stopped treatment: of them, four were in induction, two in consolidation and one in maintenance, whereas the two who continued therapy were in induction. Thus, treatment was interrupted significantly more frequently in T-ALL and in Ph-negative ALL than in Ph-positive ALL (P=0.002), possibly because of the less immunosuppressive treatment received by patients with Ph-positive ALL, and because treatment for Ph-positive ALL does not require hospitalization. Finally, the median time to obtain viral clearance was 34 days (range, 7-91); this is remarkably longer than the time previously reported in non-hematologic patients, in whom the median time to clearance is 12-14 days, but similar to that reported in other onco-hematologic patients.9 Taken together, the results obtained show that the incidence of SARS-CoV-2 infection in ALL patients during 15 months of the pandemic in Italy was similar to that in the general population, and was recorded mostly in the second wave. The largest fraction of infections was nosocomial, despite the use of personal protective equipment by both healthcare personnel and patients, stringent limitation to access for visiting relatives, and constant monitoring of nasopharyngeal swabs. These considerations are particularly important in the light of the new waves of in-

Haematologica | 107 August 2022

1956

LETTER TO THE EDITOR Table 1. Features and in-hospital management of patients with acute lymphoblastic leukemia positive for COVID-19. Patients (N=63) Gender, N (%) Male Female Age, N (%) 18-35 years 35-50 years 50-65 years >65 years Type of ALL, N (%; % of whole cohort) Ph+ B-ALL Ph- B-ALL T-ALL Comorbidities present; N (%) None One More than one Phase of disease, N (%) Onset/induction Consolidation Consolidation with TKI Chemotherapy maintenance TKI maintenance Allogeneic SCT Recurrence Off-therapy Most common comorbidity reported, N Hypertension Diabetes Obesity Dyslipidemia Metabolic syndrome Parossistic atrial fibrillation Ischemic cardiomyopathy

43 (68) 20 (32) 21 (33) 17 (27) 15 (24) 10 (16) 18 (28.6; 7.6) 28 (44.4; 7.7) 17 (27; 10.9) 36 (57) 11 (17) 16 (25) 14 (22.2) 11 (17.4) 2 (3.2) 6 (9.5) 7 (11.1) 15 (24) 2 (3) 6 (9) 10 2 2 1 1 1 1

Treatment administered to patients admitted to a COVID ward without need for respiratory assistance (N=15)*, N (%) Steroids Remdesivir Enoxaparin Antibiotics Hydroxychloroquine§ Darunavir/ritonavir§

8 (53) 4 (27) 4 (27) 4 (27) 1 (7) 1 (7)

Treatment administred to patients admitted to COVID ward with need for respiratory assistance (N=13)*, N (%) High flow oxygen Non-invasive ventilation Antibiotics Steroids Enoxaparin Remdesivir Convalescent plasma Hydroxychloroquine§ Lopinavir/ritonavir§ Ruxolitinib

8 (61) 5 (39) 8 (61) 8 (61) 7 (54) 5 (38) 2 (15) 1 (8) 1 (8) 1 (8)

Treatment administred to patientsa dmitted to an Intensive Care Unit (N=6)*, N (%) Invasive mechanical ventilation Antibiotics Ruxolitinib

6 (100) 4 (67) 1 (17)

*The sum for each category exceeds 100% since patients received more than one type of treatment; §These drugs were used only during the first wave of the pandemic. COVID-19: coronavirus disease 2019; ALL: acute lymphoblastic leukemia; Ph: Philadelphia chromosome; TKI: tyrosine kinase inhibitor; SCT: stem cell transplantation.

fections caused by novel virus variants, and suggest that outpatient and home care should be pursued whenever possible. No enrichment for age, subtype and comorbidities was identified. The infection was manageable, with 46% of patients not requiring any medical intervention. In line with a previous report by our group,7 it appears that Ph-positive ALL patients were more manageable, requiring fewer treatment interruptions (72.7%). Likewise, in a recent report by the Campus chronic myeloid leukemia (CML) group, the infection and mortality rates in CML in Italy during the pandemic were low, with only 2.5% positive patients (217 COVID-19-positive patients among among 8,665 patients with CML), and a mortality rate of 5.5% and 0.5% in the infected population and entire cohort, respectively.10 Bonifacio et al.11 reported the serological prevalence of infection in 564 CML patients in different phases of the disease. Eleven patients resulted IgG positive, with only three having a diagnosis of COVID19. Overall, these observations, although derived from a different setting, suggest a possible protective role of TKI treatment against the severity of COVID-19, as suggested by other studies,12 although it must be remembered that imatinib does not exert a direct anti COVID-19 effect.13 Interestingly, also in chronic lymphocytic leukemia, a disorder that occurs predominantly in the elderly and is associated with immune impairment,14 in which targeted treatment is frequently used front-line (reducing hospital admissions), the incidence of COVID-19 was only 3.3% (494/15,039 cases). These findings support the role of targeted treatment rather than chemotherapy. In this respect, beyond TKI plus monoclonal antibodies in Ph-positive ALL, the use of monoclonal antibodies also in Ph-negative B-ALL is increasing. Finally, the death rate in our series was 11.1% (<0.9% of the whole cohort), which is better than that recently published by Ribera and colleagues,8 who reported a mortality rate of 33% in 52 COVID-19-positive ALL patients stratified according to the period of the pandemic (first or second wave). Several reasons might contribute to the lower mortality rate we report. First, in our cohort the majority of patients were infected during the second wave of the pandemic, when the knowledge on the management of COVID-19 was superior. Second, the prevalence of Ph-positive ALL patients was higher in our population (28.6% vs. 15% in the study by Ribera et al.), as in the elderly population15, for whom the Italian strategy over the years has been based on TKI-based chemofree induction without systemic chemotherapy. During the pandemic, the ongoing front-line GIMEMA protocol was based on a chemotherapy-free induction/consolidation with dasatinib and blinatumomab, designed for patients of all ages, with no upper age limit.16 This approach reduces hospitalization considerably and patients are largely managed as outpatients. Third, our cohort in-

Haematologica | 107 August 2022

1957

LETTER TO THE EDITOR cluded a higher number of patients who contracted the infection during maintenance (20.6% vs. 4% in the study by Ribera et al.). Future research will address other crucial issues in ALL patients, such as the efficacy of vaccination and the efficiency and duration of antibody production according to the treatments received and the disease stage, as well as a more thorough investigation of the long-term consequences of the infection.

Sciences, University of Modena and Reggio Emilia, Modena; Hematology, Department of Medical Area, St. Eugenio Hospital,

Rome and

Azienda Ospedaliera Universitaria Careggi, SOD

Ematologia, Florence, Italy *SC and MB contributed equally as co-first authors.

Correspondence: S. CHIARETTI - chiaretti@bce.uniroma1.it https://doi.org/10.3324/haematol.2021.280289

Authors

Received: November 18, 2021. Accepted: April 8, 2022.

Sabina Chiaretti, Massimiliano Bonifacio, 1*

Roberta Agrippino, Fabio 1

Prepublished: April 21, 2022.

Giglio, Mario Annunziata, Antonio Curti, Maria Ilaria Del Principe, 3

Prassede Salutari,7 Mariarita Sciumè,8 Mario Delia,9 Marco Armenio,10 Valentina Mancini,11 Antonino Mulè,12 Francesco Grimaldi,13 Giovanna

Rege-Cambrin,14 Lidia Santoro,15 Federico Lussana,16 Patrizia Chiusolo,17,18 Crescenza Pasciolla,19 Anna Maria Scattolin,20 Marco

Disclosures

Cerrano, Maria Ciccone,

No conflicts of interest to disclose.

Mazzone,

Matteo Piccini,

Marzia Defina,

Fabio Forghieri,

Carla

Felicetto Ferrara, Giovanni Pizzolo and 4

Robin Foà

Contributions

SC, MB and RA analyzed the data, and wrote the manuscript. FG, Department of Translational and Precision Medicine, Sapienza

MA, AC, MIDP, PS, MS, MD, MA, VM, AM, FG, GRC, LS, FL, PC, CP,

University, Rome; 2Department of Medicine, Section of Hematology,

AMS, MC, MC, MD, FF, CM and MP provided clinical data. FF and GP

University of Verona, Verona; Haematology and Bone Marrow

revised the manuscript. RF planned and designed the survey,

Transplantation Unit, San Raffaele Scientific Institute, Milan;

analyzed the data, and wrote the manuscript. A complete list of the

Division of Hematology, AORN Cardarelli, Naples; IRCCS Azienda

members of the Campus ALL working group who completed the

Ospedaliero-Universitaria di Bologna, Istituto di Ematologia

survey appears in the Appendix.

“Seràgnoli”, Bologna; Department of Biomedicine and Prevention, 6

University of Rome "Tor Vergata", Rome; 7UOC Ematologia Ospedale

Funding

Santo Spirito, Pescara; 8Hematology Unit, Fondazione IRCCS Ca'

This work was partly supported by the Associazione Italiana per la

Granda Ospedale Maggiore Policlinico, Milan; 9Hematology and Stem

Ricerca sul Cancro (AIRC), Metastases Special Program, N. 21198,

Cell Transplantation Unit, AOUC Policlinico, Bari; Department of

Milan, Italy (RF); Progetti di Rilevante Interesse Nazionale (PRIN)

Clinical and Biological Sciences, Università di Torino, Turin;

Italia, 2017PPS2X4 project

Dipartimento di Ematologia e Oncologia, Ospedale Niguarda, Milan;

Data-sharing statement

Azienda Villa Sofia-Cervello, Palermo; 13Dipartimento di Medicina

Contact Sabina Chiaretti via mail: chiaretti@bce.uniroma1.it

Clinica e Chirurgia, AOU Federico II di Napoli, Naples; Medicina 14

Interna e Ematologia Ospedale San Luigi Gonzaga, Università di Torino, Turin; 15U.O.C. Ematologia e Trapianto di Midollo Osseo,

Appendix

Avellino; 16Hematology and Bone Marrow Transplantation Unit,

The Campus ALL chairs are R.F., F.F. and G.P. and the working group

Azienda Socio-Sanitaria Territoriale Papa Giovanni XXIII, Bergamo;

coordinators are S.C., C.F., M.I.D.P, M.B., A.Cu. and A.Ca. Other

Dipartimento di Diagnostica per Immagini, Radioterapia Oncologica

members who completed the survey include: Ilaria Tanasi

ed Ematologia, Fondazione Policlinico Universitario A. Gemelli

(Department of Medicine, Section of Hematology, University of

IRCCS, Rome; Sezione di Ematologia, Dipartimento di Scienze

Verona), Alessandro Bruno (Haematology and Bone Marrow

Radiologiche ed Ematologiche, Università Cattolica del Sacro Cuore,

Transplantation Unit, San Raffaele Scientific Institute, Milan),

Rome; Hematology Unit, IRCCS Istituto Tumori "Giovanni Paolo II",

Gianluca Cristiano (IRCCS Azienda Ospedaliero-Universitaria di

Bari;

Bologna, Istituto di Ematologia “Seràgnoli”, Bologna), Giovangiacinto

UOC di Ematologia Ospedale dell'Angelo, Mestre; 21Division of

Hematology, Department of Oncology, A.O.U. Città della Salute e

Paterno (Department of Biomedicine and Prevention, University of

della Scienza di Torino, Turin;

Rome "Tor Vergata", Rome), Nicola Fracchiolla (Hematology Unit,

UO Ematologia, Dipartimento di

Medicine Specialistiche, Azienda Ospedaliero-Universitaria

Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan),

Arcispedale S Anna, Ferrara;

Francesca Caciolli (Medicina Interna e Ematologia Osp. San Luigi

Universitaria Senese, Siena;

UOC Ematologia, Azienda Ospedaliero

Department of Medical and Surgical

Gonzaga, Università di Torino, Vito Pier Gagliardi (Hematology and

Haematologica | 107 August 2022

1958

LETTER TO THE EDITOR

Stem Cell Transplantation Unit, AOUC Policlinico; Bari), Matteo Molica (Haematology, Department of Medical Area, St. Eugenio Hospital, Rome). The members who completed the survey and reported no COVID- 19-positive patients include: Anna Candoni,

Michele Cedrone, Michelina Dargenio, Fabio Guolo, Michela Lamanda, Monia Lunghi, Federico Lussana, Alessandro Fiorentini, Endri Mauro and Elisabetta Todisco.

References 1. Lee LYW, Cazier JB, Starkey T, et al. Coronavirus Cancer Monitoring Project Team. COVID-19 prevalence and mortality in patients with cancer and the effect of primary tumour subtype and patient demographics: a prospective cohort study. Lancet Oncol. 2020;21(10):1309-1316. 2. Passamonti F, Cattaneo C, Arcaini L, et al; ITA-HEMA-COV Investigators. Clinical characteristics and risk factors associated with COVID-19 severity in patients with haematological malignancies in Italy: a retrospective, multicentre, cohort study. Lancet Haematol. 2020;7(10):e737-e745. 3. Girmenia C, Gentile G, Micozzi A, et al. COVID-19 in patients with hematologic disorders undergoing therapy: perspective of a large referral hematology center in Rome. Acta Haematol. 2020;143(6):574-582. 4. Wu Y, Lin H, Xie Q, Chen Q, Huang Y, Zhu Y, Chen L. COVID-19 in a patient with pre-existing acute lymphoblastic leukaemia. Br J Haematol. 2020;190(1):e13-e15. 5. Butt A, Ali N. COVID-19 and adult acute lymphoblastic leukemia: presentation and management. Hematol Transfus Cell Ther. 2021;43(2):219-221. 6. Sneha LM, Tatapudi V, Scott JX, et al. SARS-CoV-2 at diagnosis of acute lymphoblastic leukemia-case series. To worry about COVID or leukemia? A developing country's perspective. Pediatr Blood Cancer. 2021;68(9):e29124. 7. Foà R, Bonifacio M, Chiaretti S, et al. Philadelphia-positive acute lymphoblastic leukaemia (ALL) in Italy during the COVID19 pandemic: a Campus ALL study. Br J Haematol. 2020;190(1):e3-e5. 8. Ribera JM, Morgades M, Coll R, et al. Frequency, clinical characteristics and outcome of adults with acute

lymphoblastic leukemia and COVID 19 infection in the first vs. second pandemic wave in Spain. Clin Lymphoma Myeloma Leuk. 2021;21(10): e801-e809. 9. Cogliati Dezza F, Oliva A, Cancelli F, et al. Determinants of prolonged viral RNA shedding in hospitalized patients with SARS-CoV-2 infection. Diagn Microbiol Infect Dis. 2021;100(2):115347. 10. Breccia M, Abruzzese E, Accurso V, et al. COVID-19 infection in chronic myeloid leukemia after 1 year of the pandemic in Italy. A Campus CML report. Br J Haematol. 2022; 196(3):559-565. 11. Bonifacio M, Tiribelli M, Miggiano MC, et al. The serological prevalence of SARS-CoV-2 infection in patients with chronic myeloid leukemia is similar to that in the general population. Cancer Med. 2021;10(18):6310-6316. 12. Galimberti S, Petrini M, Baratè C, et al. Tyrosine kinase inhibitors play an antiviral action in patients affected by chronic myeloid leukemia: a possible model supporting their use in the fight against SARS-CoV-2. Front Oncol. 2020;10:1428. 13. Zhao H, Mendenhall M, Deininger MW. Imatinib is not a potent anti-SARS-CoV-2 drug. Leukemia. 2020;34(11):3085-3087. 14. Cuneo A, Rigolin GM, Coscia M, et al. Management of CLL in Italy during a one year of the COVID-19 pandemic and at the start of the vaccination program. A CAMPUS CLL report. Hematol Oncol. 2021;39(4):570-574. 15. Chiaretti S, Vitale A, Cazzaniga G, et al. Clinico-biological features of 5202 patients with acute lymphoblastic leukemia enrolled in the Italian AIEOP and GIMEMA protocols and stratified in age cohorts. Haematologica. 2013;98(11):1702-1710. 16. Foà R, Bassan R, Vitale A, et al. Dasatinib-blinatumomab for Ph-positive acute lymphoblastic leukemia in adults. N Engl J Med. 2020;383(17):1613-1623.

Haematologica | 107 August 2022

1959

LETTER TO THE EDITOR

A novel BCMA CAR-T-cell therapy with optimized human scFv for treatment of relapsed/refractory multiple myeloma: results from phase I clinical trials Chimeric antigen receptor (CAR) T cells targeting B-cell maturation antigen (BCMA) demonstrate appealing antitumor activity in patients with relapsed/refractory multiple myeloma (RRMM), but toxicity and short-term efficacy limit their clinical usage.1-4 For recently Food and Drug Administration-licensed idecabtagene vicleucel (murine-derived single-chain variable fragment [scFv]), 128 patients with RRMM achieved 73.4% objective response rate (ORR) and 10.7-month median duration of response (DOR). The study also reported severe toxicities including ≥grade 3 cytokine release syndrome (CRS), neurotoxicity, and treatment-related deaths that occurred within 8 weeks after infusion.3 In order to improve safety, overcome limited efficacy, and reduce immunogenicity from non-human–derived components, we developed autologous CAR-BCMA T cells (CT053) expressing the fully-human BCMA-specific scFv (25C2). Three investigator-initiated phase I studies investigated CT053’s safety, pharmacokinetics, and preliminary efficacy in patients with RRMM (clinicaltrials gov. Identifiers: NCT03302403, NCT03380039, NCT03716856). We found that CT053 demonstrated an acceptable safety, pharmacokinetic, and efficacy profile. We hypothesized that an optimized, human scFv would avoid immunogenicity and improve safety and efficacy. We identified 25C2 through naïve human scFv phage library screening. Results showed that 25C2 could bind to human and mouse BCMA but not other TNF receptor family members, human TACI and BAFFR, indicating its BCMA-specific binding capacity (Online Supplementary Figure S1A). Binding analysis indicated that 25C2 had 603.9 pM binding affinity against recombinant human BCMA and 87% monomer ratio (Online Supplementary Figure S1B). CT053 was generated by transducing T cells with lentivirus encoding a CAR comprising the 25C2 scFv, human CD8α hinge domain, CD8α transmembrane domain, 4-1BB co-stimulatory domain and CD3ζ activation domain. In preclinical studies, CT053 showed low tonic signaling and potent in vitro MM cell killing (Online Supplementary Figure S1C-E). Accordingly, we launched three phase I studies to evaluate CT053 in patients with RRMM. Thirty patients consented, 27 underwent leukapheresis, and 24 received CT053 infusion (0.5-1.8×108 cells) in the three trials (Table 1). Patients met International Myeloma Working Group diagnostic criteria for RRMM and enrollment eligibility criteria.5 Median age was 60 years (range, 39-70 years) and 37.5% had International Staging System (ISS) stage III dis-

ease. Median number of prior systemic regimens was five (range, 2–11). Patients included 41.7% with extramedullary disease (EMD), 50% with high-risk cytogenetics, and 33.3% with Eastern Cooperative Oncology Group (ECOG) scores 23. Seven patients received bridging therapy before lymphodepletion (Table 1). All autologous CT053 products were successfully manufactured at CARsgen’s GMP facility. Patients received lymphodepletion comprising fludarabine (median dose, 21 mg/m2/day [range, 19–27 mg/m2/day] for 2–4 days) and cyclophosphamide (median dose, 467 mg/m2/day [range, 192–543 mg/m2/day] for 1–5 days). Subsequently, patients received one CT053 infusion: P1 received 0.5×108 cells, P2 received 1.8×108 cells due to their weight (91 kg), and P24’s poor clinical condition prompted the 1.0×108-cell modified dose. Remaining patients received 1.5×108 cells. All patients experienced ≥grade 3 treatment-related hematological adverse events (AE). Treatment-related hematological toxicities of ≥grade 3, expected lymphodepletion effects, were leukopenia (83.3%), lymphocytopenia (79.2%), neutropenia (75.0%), anemia (33.3%), and thrombocytopenia (25.0%) (Table 2). The median durations for grade 3 or grade 4 neutropenia and thrombocytopenia to recover to ≤grade 2 were 9 days (95% confidence interval [CI]: 5.0-44.0) and 55 days (95% CI: 8.0-not evaluable), respectively. Thirteen SAE were reported in seven patients: eight events were infections, and four events were hematological toxicities. One death occurred: P15 died on day 25 due to bone marrow failure and neutropenic infection related to lymphodepletion and disease progression. Fifteen patients (62.5%) experienced CRS; however, no events were ≥grade 3 (4 grade 1, 11 grade 2). Generally, CRS occurred a median 3 days (range, 1–9 days) after infusion and resolved in a median 6 days (range, 3–9 days). There were no significant differences in peak levels of ferritin, Creactive protein, and IL-6 within 28 days after infusion between patients with or without CRS (data not shown). Four patients received two tocilizumab doses, and five patients received one dose (4–6 mg/kg). Tocilizumab had no impact on CAR-BCMA copy numbers (data not shown). P11, with no prior convulsive history, experienced grade 3 neurotoxicity with grade 2 CRS, presenting as epilepsy. This event started 6 days after infusion, and resolved within 3 days after treatment with methylprednisolone, diazepam, and sodium valproate. As of the cutoff date of June 30, 2021, median follow-up

Haematologica | 107 August 2022

1960

LETTER TO THE EDITOR Table 1. Characteristics in patients with or without extramedullary disease. Characteristics

All patients (N=24)

Without EMD (N=14)

With EMD (N=10)

Age, years 18 to < 65 years, N (%) ≥ 65 years, N (%)

60 (39-70) 18 (75.0%) 6 (25.0%)

58 (39-67) 12 (85.7%) 2 (14.3%)

63 (39-70) 6 (60.0%) 4 (40.0%)

Males, N (%)

13 (54.2%)

7 (50.0%)

6 (60.0%)

BSA (m2)a

1.7 (1.3-2.1)

1.7 (1.4-1.9)

1.6 (1.3-2.1)

Time since diagnosis, yearsb, a

3.5 (0.4-10.8)

4.0 (0.4-10.8)

3.2 (0.8-5.9)

22/2

13/1

9/1

12 (50%)

4 (29%)

8 (80%)

ECOG,d N (%) 0-1 2 3

16 (66.7%) 6 (25.0%) 2 (8.3%)

11 (78.6%) 3 (21.4%) 0 (0.0%)

5 (50.0%) 3 (30.0%) 2 (20.0%)

ISS, N (%) I & II III

15 (62.5%) 9 (37.5%)

7 (50.0%) 7 (50.0%)

8 (80.0%) 2 (20.0%)

91.3 (30.4-99.8)

90.8 (58.5-99.8)

93.5 (30.4-99.5)

N of prior anti-MM regimens,a N (%) Proteasome inhibitors Bortezomib Ixazomib Carfilzomib Immunomodulatory drugs Lenalidomide Pomalidomide Thalidomide Anti-CD38 monoclonal antibody (daratumumab) Stem cell transplantation Refractory to last regimen Yes Unknown

5 (2-11) 24 (100%) 24 (100%) 2 (8.3%) 3 (12.5%) 22 (91.7%) 18 (75%) 3 (12.5%) 11 (45.8%) 5 (20.8%) 10 (41.7%)

6 (2-11) 14 (100%) 14 (100%) 1 (7.1%) 2 (14.3%) 13 (92.9%) 11 (78.6%) 3 (21.4%) 7 (50%) 4 (28.6%) 7 (50.0%)

4 (2-8) 10 (100%) 10 (100%) 1 (10.0%) 1 (10.0%) 9 (90.0%) 7 (70.0%) 0 (0.0%) 4 (40.0%) 1 (10.0%) 3 (30.0%)

22 (91.7%) 2 (8.3%)

13 (92.9%) 1 (7.1%)

9 (90.0%) 1 (10.0%)

Bridging therapy,e N (%)

7 (29.2%)

2 (14.3%)

5 (50%)

Heavy chain /light chains, N High-risk cytogenetics,c N (%)

BCMA expression in BM, N (%)a

BCMA: B-cell maturation antigen; BM: bone marrow; BSA: body surface area; EMD: extramedullary disease; ISS: International Staging System; MM: multiple myeloma. Thirty patients consented for the 3 studies. Three patients who consented were ineligible to receive CT053 because of their blood test results (creatinine increased beyond range of eligibility). Twenty-seven patients underwent leukapheresis. Of these, patients were excluded because of intracranial hemorrhage (1), intracranial infiltration (1), and low platelet count (1). The 24 patients who received CT053 infusion are included in the table. aMedian (range), bthe time between the initial diagnosis and the study screening visit. cResults of cytogenetics obtained by fluorescence in situ hybridization from bone marrow aspirate performed at any point before treatment with CT053. High-risk cytogenetics were defined as gain(1q), del(13), del(17p), t(4;14), t(14;16), t(14;20). dEastern Cooperative Oncology Group (ECOG) performance status scores range from 0 to 5, with higher scores indicating greater disability; a score of 5 indicates death. eBridging therapy was administered in seven patients after leukapheresis and before lymphodepletion. Five patients received a bortezomib-based regimen, 1 patient received a lenalidomide-based regimen, and 1 patient received etoposide-, cyclophosphamide- and cisplatin-based combination chemotherapy.

time was 17.4 months (range, 0.9–38.7 months), and ORR (partial response [PR] or better) was 87.5%, with 79.2% patients experiencing complete reponse (CR) (12.5%) or stringent complete response (sCR, 66.7%) (Figure 1A). Seven patients died due to disease progression, including four who relapsed from sCR, in addition to P15 who died due to SAE. Responses occurred early, with median 4.1 weeks (range, 1.9–12.7 weeks) to first PR or better after infusion. Median time to best response was 8.3 months (range, 1.0–16.5 months). Nine patients (37.5%) had persistent CR/sCR and completed 24-month follow-up, including seven who had minimal residual disease (MRD)-negative status through the

last follow-up visit at 24 months.5 The CR/sCR rate was 70% for patients with EMD and 86% in patients without EMD. Notably, P21 presented with thoracic cutaneous plasmacytomas that significantly shrank after infusion, and computed tomography (CT) showed 80% reduction at day 12 after infusion (Figure 1B and C). The lesions were confirmed eliminated by CT at day 64. P21 remained in sustained remission at data cutoff. Median progression-free survival (PFS) was 18.8 months (95% CI: 10.1–not evaluable [NE]) in all patients, and there was no statistical difference in patients with or without EMD (Online Supplementary Figure S2). Median overall sur-

Haematologica | 107 August 2022

1961

LETTER TO THE EDITOR Table 2. Incidence of treatment-related and treatment-emergent adverse events in ≥2 patients (N=24). Preferred terma, N (%)

Treatment-related adverse eventsb

Treatment-emergent adverse eventsc

All grades

Grade ≥3

All grades

Grade ≥3

15 (62.5)

Neutrophil count decreased

23 (95.8)

18 (75.0)

22 (91.7)

16 (66.7)

White blood cell count decreased

21 (87.5)

20 (83.3)

14 (58.3)

Platelet count decreased

14 (58.3)

6 (25.0)

14 (58.3)

8 (33.3)

Lymphocyte count decreased

19 (79.2)

13 (54.2)

6 (25.0)

Anemia

10 (41.7)

8 (33.3)

12 (50.0)

7 (29.2)

Pyrexia

18 (75.0)

6 (25.0)

18 (75.0)

6 (25.0)

Hypokalemia

6 (25.0)

2 (8.3)

14 (58.3)

3 (12.5)

Hypocalcemia

4 (16.7)

10 (41.7)

Aspartate aminotransferase increased

4 (16.7)

1 (4.2)

8 (33.3)

1 (4.2)

Alanine aminotransferase increased

2 (8.3)

8 (33.3)

Upper respiratory tract infection

2 (8.3)

1 (4.2)

7 (29.2)

2 (8.3)

Immunoglobulins decreased

5 (20.8)

6 (25.0)

Decreased appetite

6 (25.0)

5 (20.8)

Asthenia

5 (20.8)

Diarrhea

3 (12.5)

5 (20.8)

Pneumonia

3 (12.5)

4 (16.7)

Cytokine release syndrome Hematologic adverse events

Non-hematologic adverse events

Mouth ulceration

Medical Dictionary for Regulatory Activities (version 24.0), graded according to National Cancer Institute Common Terminology Criteria for Adverse Event version 4.01. bTreatment-related adverse event indicates lymphodepletion-related and/or CT053-related adverse event. cTreatment-emergent adverse event is defined as any adverse event starting from CT053 infusion to 24 months after infusion. a

vival (OS) was not reached. Median DOR was 21.8 months (95% CI: 9.2–NE) in all patients. Numerically higher median DOR was observed in patients achieving MRD-negativity than those with MRD-positivity, though not statistically significant (24.0 months, 95%CI: 10.3–NE vs. 8.5 months, 95%CI: 7.6–NE, respectively). After CT053 infusion, CAR-BCMA transgene copies became detectable at days 1–7 in all patients. Median peak value of transgene copies was 92,621 copies/µg genomic DNA (15,047–449,369 copies/µg genomic DNA), and median time to peak value was 13.5 days (range, 7–21 days). CT053 was detectable in nine of 20 patients at 6 months, and three of seven patients had detectable CT053 at 12 months (Online Supplementary Figure S3A). CT053 expansion correlated with tumor antigen exposure. Peak transgene copy numbers significantly correlated with the burden of BCMA-positive plasma cells (r=0.7684, P<0.001) (Online Supplementary Figure S3B). However, IL-6 levels stayed relatively low (median 17.23 pg/mL) regardless of transgene copy numbers (Online Supplementary Figure S3C). Median peak values of transgene copy numbers were significantly higher in patients with very good partial re-

sponse (VGPR) or better versus those who had not reached VGPR at month 4 (164,380 copies/µg genomic DNA [gDNA] vs. 60,547 copies/µg gDNA, respectively) and across the study (111,214 copies/µg gDNA vs. 17,301 copies/µg gDNA, respectively) (Online Supplementary Figure S3D). Results indicated that early tumor responses and best responses were associated with CT053 expansion levels. Nevertheless, we observed little difference in CAR-T-cell expansion and persistence between CT053 and non-human BCMA CAR-T cells.3 Anti-drug antibody (ADA) was not detected in patients after infusion, demonstrating no obvious immunogenicity for CT053. ADA is a risk factor for suppressed CAR-T-cell expansion.4 Given fully-human CT053’s lack of immunogenicity, we plan to explore repeat dose efficacy in ongoing trials. Our results may reflect CT053’s scFv optimization. We selected CT053’s novel fully-human scFv for its high affinity, stability, and favorable preclinical efficacy with reduced toxicity. Higher affinity scFv improve tumor recognition and enhance antitumor efficacy in vitro and in vivo.6 However, the highest affinities can inversely correlate with efficacy and cause on-target off-tumor toxicity. Fine-tuning scFv af-

Haematologica | 107 August 2022

1962

LETTER TO THE EDITOR

Figure 1. Patient-level responses to CT053. (A) Duration of patient response after CT053 infusion. Numbers to the right of the lanes represent duration of response (days). CR: complete response; MR: minimal response; MRD: minimum residual disease; NE: not evaluable; PD: partial response; sCR: stringent complete response; SD: stable disease; VGPR: very good partial response; # patients with extramedullary disease (EMD) at baseline. (B) Computed tomography images showing rapid regression of subcutaneous plasmacytomas (extramedullary myeloma) in patient 21 at baseline, day 12 and day 64 after CT053 infusion. The size of two lesions significantly reduced at day 12 and completely disappeared at day 64. (C) The expansion of patient 21’s CT053 related to the reduction of the sum of the products of maximal perpendicular diameter (SPD). The lesion SPD changes are presented as percentage of baseline value. Blue line with dots represents the transgene vector copy numbers (left y-axis), and green line with triangles represents the percentage change of SPD (right y-axis). Haematologica | 107 August 2022

1963

LETTER TO THE EDITOR finity could increase CAR-T cells’ ability to distinguish tumors from normal tissues with low-level target antigen expression while retaining robust anti-tumor efficacy.6-8 High binding affinity may have helped CT053 recognize MM cells with low-level BCMA expression and resulted in a high CR/sCR rate. CT053’s 25C2 scFv was 87% monomeric, suggesting that its high stability could have limited CD3 autophosphorylation and subsequent IL-6 secretion. Lower-grade CRS events in this study may have resulted from lower IL-6 levels, ~10 pg/mL, compared to reported grade 3-5 neurotoxicity events with IL-6 levels ≥100 pg/mL.9 In patients, CT053 showed a better safety profile with lower CRS severity versus other BCMA CAR-T-cell programs reporting 6-41% of ≥grade 3 CRS.1,2,4,10-12 Despite enrolling 41.7% patients with EMD, ECOG scores ≥2 (33.3%), and/or high-risk cytogenetics (50%), we obtained 79.2% CR/sCR rate with sustained 21.8-month DOR compared to 33% CR/sCR rate and 10.7-month DOR reported in a trial with non-human scFv.3 Also, this study reports significant improvement in clinical outcomes for patients with EMD compared to previous BCMA CAR-T-cell therapies.4,13 Four of ten patients with EMD at baseline were still in sustained CR/sCR (range, 27.3–34.5 months). However, the study’s non-Western population mostly lacked exposure to anti-CD38 antibody. In order to address this limitation, patients with RRMM and prior anti-CD38 treatment are actively enrolling in the North American pivotal phase II LUMMICAR STUDY 2. Taken together, we generated CT053 with an optimized, fully-human scFv, and we demonstrated that CT053 had strong efficacy and a good safety profile when administered to RRMM patients. Our study indicated CT053’s promise for treatment of RRMM patients and showed that scFv selection in CAR-T cells is critical to achieving better clinical outcomes.

*MY, WZ, KY, PW and HJ contributed equally as co-first authors. Correspondence: S. HAO - haosiguo@xinhuamed.com.cn J. JIN - jiej0503@163.com Z. LI - zonghaili@shsmu.edu.cn S. JIANG - jiangsongfu@189.cn https://doi.org/10.3324/haematol.2022.280629 Received: January 6, 2022. Accepted: April 8, 2022. Prepublished: April 21, 2022. ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

Disclosures ZL, PW, and HW have submitted patent applications related to this work. The other authors declare no potential conflicts of interest. Preliminary results from this report were presented in a virtual oral session at the 62nd annual meeting of the American Society of Hematology (ASH) held December 5-8, 2020. A related high-risk group integrated analysis of this study combined with LUMMICAR STUDY 1 was presented in a poster at the ASH annual meeting, December 11-14, 2021. Contributions ZL and JJ designed the overall project; PW and HJ performed preclinical studies and analysis; JJ, SJ and SH were responsible for the clinical design, supervision, data analysis and interpretation; Patient care: MY, WZ, KY, LC, HM, YW, RT, XH, CX, JW, SW, LD, SH, JJ, and SJ took care of patients; HW manufactured CT053; ZL, HM, AYH, WW, JX, SH, SJ, and JJ were responsible for medical oversight, data analysis, drafting or revision of the manuscript. All authors reviewed and approved the manuscript. Acknowledgments

Authors

We would like to thank the patients and their families for participating in the study. We also thank the staff in the clinical units for patient care, Daijing Yuan for data management, Xiaochen Dong for project

Min Yang,1* Wenhao Zhang,2* Kang Yu,3* Peng Wang,4,5* Hua Jiang,5*

management, Jinan Qi for cell manufacturing, and Liu Zhen for

Linjun Chen, Haitao Meng, Yiqin Weng, Rong Tao, Xin Huang,

patient follow-up sample support.

Chongyun Xing, Huamao Wang, Jiangbo Wan, Shasha Wang, Lihui 3

Dai,3 Amanda Y. Hendrix,6 Jun Xiao,5 Wei Wang,5 Hong Ma,6 Siguo Hao,2

Funding

Jie Jin, Zonghai Li

This work was supported by a grant from the Shanghai Strategic

5,7

and Songfu Jiang

Emerging Industry Development Foundation and by CARsgen 1

Department of Hematology, The First Affiliated Hospital, Zhejiang

Therapeutics.

University School of Medicine, Hangzhou; 2Department of Hematology, Xinhua Hospital Affiliated to Shanghai Jiao Tong

Data-sharing statement

University School of Medicine, Shanghai and Department of

All data generated or analyzed during this study are included in this

Hematology, The First Affiliated Hospital of Wenzhou Medical

published article and its Online Supplementary Appendix. Further

University, Wenzhou, China.

information is available from the corresponding author on reasonable

request. Haematologica | 107 August 2022

1964

LETTER TO THE EDITOR

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. Munshi NC, Anderson LD, Jr., Shah N, et al. Idecabtagene vicleucel in relapsed and refractory multiple myeloma. N Engl J Med. 2021;384(8):705-716. 4. Xu J, Chen LJ, Yang SS, et al. Exploratory trial of a biepitopic CAR T-targeting B cell maturation antigen in relapsed/refractory multiple myeloma. Proc Natl Acad Sci U S A. 2019;116(19):9543-9551. 5. Kumar S, Paiva B, Anderson KC, et al. International Myeloma Working Group consensus criteria for response and minimal residual disease assessment in multiple myeloma. Lancet Oncol. 2016;17(8):e328-e346. 6. Hudecek M, Lupo-Stanghellini MT, Kosasih PL, et al. Receptor affinity and extracellular domain modifications affect tumor recognition by ROR1-specific chimeric antigen receptor T cells. Clin Cancer Res. 2013;19(12):3153-3164. 7. Liu X, Jiang S, Fang C, et al. Affinity-tuned ErbB2 or EGFR chimeric antigen receptor T cells exhibit an increased

therapeutic index against tumors in mice. Cancer Res. 2015;75(17):3596-3607. 8. Smith EL, Staehr M, Masakayan R, et al. Development and evaluation of an optimal human single-chain variable fragmentderived BCMA-targeted CAR T cell vector. Mol Ther. 2018;26(6):1447-1456. 9. Gust J, Hay KA, Hanafi LA, et al. Endothelial activation and blood-brain barrier disruption in neurotoxicity after adoptive immunotherapy with CD19 CAR-T cells. Cancer Discov. 2017;7(12):1404-1419. 10. Usmani SZ, Rodriguez-Otero P, Bhutani M, Mateos MV, Miguel JS. Defining and treating high-risk multiple myeloma. Leukemia. 2015;29(11):2119-2125. 11. Zhao WH, Liu J, Wang BY, et al. A phase 1, open-label study of LCAR-B38M, a chimeric antigen receptor T cell therapy directed against B cell maturation antigen, in patients with relapsed or refractory multiple myeloma. J Hematol Oncol. 2018;11(1):141. 12. Yan Z, Cao J, Cheng H, et al. A combination of humanised antiCD19 and anti-BCMA CAR T cells in patients with relapsed or refractory multiple myeloma: a single-arm, phase 2 trial. Lancet Haematol. 2019;6(10):e521-e529. 13. Topp MS, Duell J, Zugmaier G, et al. Anti-B-cell maturation antigen BiTE molecule AMG 420 induces responses in multiple myeloma. J Clin Oncol. 2020;38(8):775-783.

Haematologica | 107 August 2022

1965

LETTER TO THE EDITOR

Low-dose tyrosine kinase inhibitors in patients with chronic myeloid leukemia: a retrospective study in China The advent of tyrosine kinase inhibitors (TKI) has significantly improved the treatment and prognosis of chronic myeloid leukemia (CML) patients who can nowadays expect a near-normal life expectancy.1,2 However, the adverse events accompanying lifelong treatment dramatically diminish patients’ quality of life and dependence on TKI, eventually leading to poor treatment outcomes.3,4 In addition, long-term TKI treatment increases the financial burden of patients and society.5 Several clinical studies have investigated the possibility of TKI discontinuation in patients with sustained deep molecular response (DMR). Only approximately 50% of patients with DMR remained in treatment-free remission (TFR) at least 1 year after discontinuing TKI.6 In clinical practice, TKI dose reduction is an important measure for alleviating adverse events, improving quality of life, and adherence. Interestingly, Fassoni et al. developed a patient data-based mathematical model which showed that a reduction in TKI dose of at least 50% did not exacerbate long-term treatment outcomes.7 Recent evidence suggests that low-dose TKI can effectively maintain molecular response without impairing the achievement of TFR.8 Nonetheless, limited data are available on the effect of reduced TKI in China. In this multi-center, retrospective trial, we sought to ascertain whether dose reduction of TKI is appropriate for patients with discontinuation requirements, financial stress, and adverse effects. We retrospectively analyzed the efficacy of TKI dose reduction of CML-CP patients from September 2011 to October 2021 from three hospitals in China (Union Hospital, Tongji Medical College, Huazhong University of Science and Technology; Huazhong University of Science and Technology Union Shenzhen Hospital (Nanshan Hospital); The Affiliated Cancer Hospital of Zhengzhou University). All included patients were 14 years of age or older, with positive BCR-ABL1 transcripts and no history of accelerated or blast phase. Patients who had previously received hematopoietic stem cell transplantation or immune cell therapy were excluded. In this study, we included 108 patients (54 females and 54 males, median age 43.5 years (range, 14-70 years) who underwent low-dose TKI treatment for financial burden (21.3%), attempt for TFR (23.1%), and adverse events (55.6%) (Table 1). TKI at dose reduction were imatinib (n=51), dasatinib (n=47), and nilotinib (n=10) (Figure 1). Patients were treated at doses of imatinib of 300 mg/d (n=2), 200 mg/d (n=48), and 400 mg every other day (qod) (n=1). Dasatinib at a dose of 50 mg/d (n=45), 70 mg/d (n=1), and 50 mg qod (n=1) was prescribed. The doses of nilotinib were 300 mg/d (n=1), 400 mg/d (n=9).

28 (25.9%) patients displayed resistance to at least one TKI before dose reduction according to European LeukemiaNet recommendations.2 Ninety eight of 108 (90.7%) patients achieved major molecular remission (MMR) at the time of dose reduction with a median duration of 49.8 months (range, 1-146 months), among which 92 patients achieved MR4 (BCR-ABL IS≤0.01%) with a median duration of 39.8 months (range, 1-146 months). Ten non-MMR patients received low-dose TKI due to adverse events. Molecular relapse-free survival (MRFS) in MMR and MR4 were Table 1. Patient characteristics.

Patient characteristics (N=108)

N (%)

Median age at diagnosis, years (range)

43.5 (14-70)

Female, sex

54 (50)

TKI at dose reduction Imatinib

51 (47.2)

Dasatinib

47 (43.5)

Nilotinib

10 (9.3)

Resistance to at least one TKI before dose reduction

28 (25.9)

CP at diagnosis

108 (100)

Median duration of TKI before dose reduction, months, (range)

69.9 (1-153)

Reason for dose reduction Financial burden

23 (21.3)

Attempt for TFR

25 (23.1)

Adverse events

60 (55.6)

In ≥ MR4 at dose reduction, n (%) Yes

92 (85.2)

16 (14.8)

In ≥ MMR at dose reduction, n (%) Yes

98 (90.7)

10 (8.7)

Median duration of MR4 before dose reduction, months, (range) Median duration to achieve MR4 before dose reduction, months, (range)

31.6 (0-146) 15 (2-97)

The median duration of MMR before dose reduction, months, (range)

41.7 (0-146)

Median duration to achieve MMR before dose reduction, months, (range)

10.5 (2-88)

Median follow-up of dose reduction, months, (range)

15 (2-66)

TKI: tyrosine kinase inhibitor; CP: chronic phase, TFR: treatment-free remission; n: number; MMR: major molecular remission; MR4: BCRABL IS≤0.01%.

Haematologica | 107 August 2022

1966

LETTER TO THE EDITOR defined as the probability of survival in remaining in MMR and MR4 on low-dose TKI treatment. Of the 98 patients who achieved MMR at the time of dose reduction, 94 patients (95.9%) experienced maintenance or deeper molecular response, including 94% (47/50) of the patients on imatinib, 97.4% (37/38) of the patients on dasatinib, and 100% of the patients (10/10) on nilotinib. The

1-year and 2-year MRFS in MMR were 96.7% (95% confidence interval [CI]: 90.1-98.9) and 95.1% (95% CI: 87.398.2). Moreover, seven of ten patients not in MMR initially, achieved MMR or deeper molecular response with lowdose treatment, with one patient achieving MMR, six achieving MR4 or deeper response. Of 92 patients with MR4, 81 (88.0%) patients maintained MR4 or reached a

Figure 1. Study population flowchart. TKI:tyrosine kinase inhibitor; MMR: major molecular remission; MR4: BCR-ABL IS≤0.01%. d: day; qod: every other day. Haematologica | 107 August 2022

1967

LETTER TO THE EDITOR deeper molecular response during dose-reduction. The 1year and 2-year MRFS in MR4 were 87.8% (95% CI: 78.493.3) and 86.1% (95% CI: 76.0-92.1) (Figure 2). Univariate analysis showed that the type of TKI at the time of dose reduction was the only clinical variable significantly associated with MRFS in MR4 (Online Supplementary Table S1). The 2-year MRFS in MR4 of patients on imatinib at the time of dose reduction was significantly lower than patients on 2G-TKI, 79.1% (95% CI: 63.2-88.6) versus 93.9% (95% CI: 77.2-98.5, log-rank P=0.041) (Figure 2). Four patients who lost MMR returned to a standard dose (n=2) or half standard dose (n=2) of the same TKI, three regained MMR after a median time of 3 months (range, 1.5-10 months), and two regained MR4 after a median time of 7.5 months (range, 3-15 months). A 70-year-old patient who continued to take 50 mg dasatinib due to pleural effusion did not obtain MMR. Seven patients only underwent MR4 loss, all patients regained MR4 after further treatment for 7 months (range, 1-18 months) on the same low-dose TKI. In this study, a total of 90 patients were in MR4 or deeper response during dose reduction. Sixty six patients were eligible to discontinued TKI therapy according to the eligi-

bility criteria in the EURO-SKI study.6 Eighteen of the patients (imatinib, n= 13; dasatinib, n=4; nilotinib, n=1) further discontinued low-dose TKI therapy based on their own needs, of which, with six patients fearing adverse effects of long-term treatment, four for financial burden, seven for the pursuit of TFR, and one for pancreatic cancer. The median duration of TKI treatment and MR4 was 87 months (range, 44-128 months) and 58 months (range, 28-125 months), respectively. The median duration of TKI reduction was 21 months (range, 3-66 months). At a median follow-up of 6 months (range, 1-42 months), eight patients lost MMR, of which six patients lost MMR within 6 months. The TFR at 6 months and 12 months was 59.6% (95% CI: 30.7-79.7) and 44.7% (95% CI: 14.3-71.6). Seven patients who restarted the same low-dose TKI achieved MMR after a median follow-up time of 4 months (range, 1-6 month), of which six patients achieved MR4. One patient did not restart TKI due to worsening pancreatic cancer and died of pancreatic cancer. In the present study, the 1- and 2-year MRFS in MMR for patients on low-dose imatinib were 93.5% (Online Supplementary Table S1). In the DESTINY study, where 174 pa-

Figure 2. Molecular relapse-free survival after tyrosine kinase inhibitor dose reduction. (A and B) Molecular relapse-free survival (MRFS) in major molecular remission (MMR) and MR4 (BCR-ABL IS≤0.01%) in patients after tyrosine kinase inhibitor (TKI) dose reduction. (C and D) MRFS in MMR and MR4 in patients after TKI dose reduction according to the type of TKI. 2G TKI: secondgeneration tyrosine kinase inhibitor (nilotinib, dasatinib). Haematologica | 107 August 2022

1968

LETTER TO THE EDITOR tients (85% patients were on imatinib) at least in MMR received half-dose treatment for 12 months, only 12 patients (7%) experienced molecular recurrence (loss of MMR).9 97.4% (37/38) of patients remained in MMR with 50 mg dasatinib or lower, consistent with another study that reported a 3-year MRFS in MMR of 94.5%.10 Importantly, of nine patients not in MMR, six (66.7%) achieved MMR or deeper molecular response with low dasatinib dose treatment. In our study, all patients (n=10) on half-dose nilotinib experienced maintenance of MMR. Similarly, the NILO-RED study reported a 12-months survival without unconfirmed MMR loss of 97% for 67 patients switching from a standard-dose nilotinib to a half-dose.11 Given that MR4 is regarded as the threshold for TKI withdrawal, increased emphasis should be placed on ascertaining whether dose optimization is effective in maintaining MR4. In our study, of 92 patients with MR4, 81 (88.0%) maintained MR4 or attained a deeper molecular response during dose reduction. Similarly, a prospective study reported that most patients (50/52) could maintain MR4 with a lower dose of TKI.12 In the DESTINY study, patients in MR4 had a higher probability of maintaining MMR than patients only in MMR after 12 months of half-dose therapy (98% vs. 71%, P=0.0007).9 Due to the small sample size of patients (n=6) who were only in MMR in this study, the difference in MMR maintenance between the MR4 and MMR groups was not analyzed. Interestingly, we found patients with second-generation tyrosine kinase inhibitor (2G TKI) at dose reduction had a higher probability of remaining in MR4. Consistently, in a multiple TKI discontinuation study, patients with 2-3G TKI experienced increased TFR compared to patients with imatinib.13 In our study, at the last follow-up, we observed an improved molecular response in three of the six (50.0%) patients who were only in MMR at dose reduction. It is well-established that in clinical practice, dose optimization is important to maintain efficacy and mitigate adverse effects related to TKI. A French study previously reported that 23% (196/853) of patients treated with dasatinib experienced different degrees of pleural effusion, of which 59.2% (116/196) patients required a reduced dose.14 In the present study, 71.8% (28/39) of patients received lower dasatinib doses due to pleural effusion (Online Supplementary Table S2). The pleural effusion in all cases alleviated or disappeared after low-dose treatment (50 mg/d, n=27; 50 mg qod, n=1), of which 12 received diuretic therapy. Eighteen patients underwent imatinib dose reductions (200 mg/d, n=16; 300 mg/d, n=2) mainly for leukopenia, anemia, fatigue edema, and rash itching. Nilotinib was reduced to 300 mg (n=1) and 400 mg (n=2) due to bone pain and elevated bilirubin (Online Supplementary Table S2). Similarly, all adverse events caused by imatinib and nilotinib were relieved or resolved with low-dose treatment.

The DESTINY study reported a recurrence-free survival at 24 months after stopping TKI of 72%,8 higher than the reported TFR (50%) at 24 months in the EURO-SKI study.6 Similarly, Claudiani et al. reported the 2-year TFRS of patients with low-dose treatment was 74.1%,10 and they considered that low-dose TKI might improve the chances of obtaining TFR by extending the duration of TKI and MR4. Eighteen of our patients with at least MR4 discontinued TKI after 21 months (range, 3-66 months) of low-dose treatment. A total of eight patients lost MMR, of which six patients lost MMR within 6 months. The TFR at 6 and 12 months was 59.6% and 44.7%, in agreement with the EURO-SKI study.6 Another study found that full-dose group and low-dose group displayed a similar TFR at 60 months after TKI cessation (47.5% vs. 58.8%, P=0.14), indicting low-dose TKI before discontinuation do not impair TFR.15 Nonetheless, it remains unclear whether TKI dose reduction prior to discontinuation can improve TFR. Consistent with the literature, our findings showed that lowdose TKI did not affect the achievement of TFR. In conclusion, the present study provides evidence that low-dose therapy can address the needs of the majority of patients, including alleviation of financial burden, preparation before discontinuation, and reduction of adverse effects. However, our study still has some shortcomings. Firstly, this is a retrospective study with a small group of patients that only obtained MMR. Thus, we did not compare the difference in molecular response after dose reduction between patients who obtained MMR only and those who obtained MR4. In addition, only 18 patients discontinued low-dose TKI in this study, and the data on discontinuation after dose reduction might be biased. Importantly, prospective clinical studies with large patient numbers are warranted to corroborate the effect of dose reduction.

Authors Yilin Chen,1* Zelin Liu,2# Jing Zou,1# Danyu Wang,2 Wenjuan He,1 Li Meng,3 Fanjun Cheng,1 Yanli Zhang4 and Weiming Li1 1

Department of Hematology, Union Hospital, Tongji Medical College,

Huazhong University of Science and Technology, Wuhan, Hubei; 2

Department of Hematology, Huazhong University of Science and

Technology Union Shenzhen Hospital (Nanshan Hospital), Shenzhen, Guangdong; 3Department of Hematology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei and 4Department of Hematology, The Affiliated Cancer Hospital of Zhengzhou University, Zhengzhou, Henan, China *

YC, ZL and JZ contributed equally a co-first authors.

Haematologica | 107 August 2022

1969

LETTER TO THE EDITOR Correspondence:

Disclosures

W. LI - lee937@126.com

No conflicts of interest to disclose.

Y. - 13203729690@163.com

Contributions

F. CHENG - chengfanjun001@sina.com

YLC, ZLL, JZ, FJC, YLZ and WML designed research, collected data, and wrote the manuscript; YLC, ZLL and JZ interpreted the data

https://doi.org/10.3324/haematol.2022.280637

and performed statistical analysis; YLC and WML designed and supervised the overall study; DYW, WJH and LM collected data and

Received: January 12, 2022.

reviewed the manuscript.

Accepted: April 14, 2022. Prepublished: April 28, 2022.

Data-sharing statement The data used to support the findings of this study are available

from the corresponding author upon reasonable request. The data are not publicly available due to privacy or ethical restrictions.

References 1. Jabbour E, Kantarjian H. Chronic myeloid leukemia: 2020 update on diagnosis, therapy and monitoring. Am J Hematol. 2020 Jun;95(6):691-709. 2. Hochhaus A, Baccarani M, Silver RT, et al. European LeukemiaNet 2020 recommendations for treating chronic myeloid leukemia. Leukemia. 2020;34(4):966-984. 3. Efficace F, Cannella L. The value of quality of life assessment in chronic myeloid leukemia patients receiving tyrosine kinase inhibitors. Hematol-Am Soc Hemat. 2016;2016(1):170-179. 4. Winn AN, Keating NL, Dusetzina SB. Factors associated with tyrosine kinase inhibitor initiation and adherence among medicare beneficiaries with chronic myeloid leukemia. J Clin Oncol. 2016;34(36):4323-4328. 5. Saussele S, Richter J, Hochhaus A, et al. The concept of treatment-free remission in chronic myeloid leukemia. Leukemia. 2016;30(8):1638-1647. 6. Saussele S, Richter J, Guilhot J, et al. Discontinuation of tyrosine kinase inhibitor therapy in chronic myeloid leukaemia (EURO-SKI): a prespecified interim analysis of a prospective, multicentre, non-randomised, trial. Lancet Oncol. 2018;19(6):747-757. 7. Fassoni AC, Baldow C, Roeder I, et al. Reduced tyrosine kinase inhibitor dose is predicted to be as effective as standard dose in chronic myeloid leukemia: a simulation study based on phase III trial data. Haematologica. 2018;103(11):1825-1834. 8. Clark RE, Polydoros F, Apperley JF, et al. De-escalation of tyrosine kinase inhibitor therapy before complete treatment discontinuation in patients with chronic myeloid leukaemia (DESTINY): a non-randomised, phase 2 trial. Lancet Haematol. 2019;6(7):e375-e383.

9. Clark RE, Polydoros F, Apperley JF, et al. De-escalation of tyrosine kinase inhibitor dose in patients with chronic myeloid leukaemia with stable major molecular response (DESTINY): an interim analysis of a non-randomised, phase 2 trial. Lancet Haematol. 2017;4(7):e310-e316. 10. Claudiani S, Apperley JF, Szydlo R, et al. TKI dose reduction can effectively maintain major molecular remission in patients with chronic myeloid leukaemia. Br J Haematol. 2021;193(2):346-355. 11. Rea D, Cayuela JM, Dulucq S, Etienne G. Molecular responses after switching from a standard-dose twice-daily nilotinib regimen to a reduced-dose once-daily schedule in patients with chronic myeloid leukemia: a real life observational study (NILO-RED). Blood. 2017;130(Suppl 1):S318. 12. Cayssials E, Tartarin F, Guilhot J, et al. Sustained molecular response in chronic myeloid leukemia deep responders treated with low dose tyrosine kinase inhibitors. Leuk Lymphoma. 2018;59(3):766-769. 13. Etienne G, Dulucq S, Bauduer F, et al. Incidences of deep molecular responses and treatment-free remission in de novo CP-CML patients. Cancers (Basel). 2020;12(9):2521. 14. Iurlo A, Galimberti S, Abruzzese E, et al. Pleural effusion and molecular response in dasatinib-treated chronic myeloid leukemia patients in a real-life Italian multicenter series. Ann Hematol. 2018;97(1):95-100. 15. Cayssials E, Torregrosa-Diaz J, Gallego-Hernanz P, et al. Low-dose tyrosine kinase inhibitors before treatment discontinuation do not impair treatment-free remission in chronic myeloid leukemia patients: results of a retrospective study. Cancer. 2020;126(15):3438-3447.

Haematologica | 107 August 2022

1970

LETTER TO THE EDITOR

Copy number alterations define outcome in Philadelphia chromosome-positive acute lymphoblastic leukemia The introduction of tyrosine kinase inhibitor (TKI) imatinib has improved outcome of Philadelphia chromosome-positive (Ph+) acute lymphoblastic leukemia (ALL), and the second- and third-generation TKI, such as dasatinib or ponatinib, may prove even more effective.1 Unfortunately, treatment failures remain frequent, often due to the emergence of BCR-ABL1 kinase domain (KD) mutations. The second-generation TKI were developed to overcome most imatinib-resistant KD mutations. However, several mutations, such as T315I and F317I/L, evade also the second-generation TKI, and highly-resistant compound mutations even the third.2 In addition, IKZF1 deletions, especially in combination with deletions in PAX5 and/or CDKN2A/B genes, may define a group with unfavorable outcome (“IKZF1 plus”).3 Considering that the first trials of chemotherapy-free treatment regimens in Ph+ ALL have shown promising results,4 and that TKI-based therapies induce durable remissions in some patients even without transplantation,5 identification of prognostic and predictive markers is of utmost importance for treatment stratification. Here, we investigated potential biomarkers for treatment outcome in a retrospective, nationwide Ph+ ALL adult population. In addition to assessing clinical parameters, we sequenced diagnostic and relapse-phase patient samples with a targeted next-generation sequencing (NGS) gene panel consisting of 75 leukemia-associated genes. We also analyzed copy-number alterations (CNA) in IKZF1, PAX5, and CDKN2A/B genes. As T315I kinase domain mutations cause broad resistance to TKI,2 we examined the prevalence of subclonal T315I with digital droplet polymerase chain reaction (ddPCR). All clinical data was obtained from the Finnish Hematology Registry (FHR), a population-based centralized database, which stores data on clinical variables, treatments, treatment outcomes, laboratory values, and results from cytogenetic and molecular analyses. Bone marrow samples were retrieved from the clinical laboratories and from the Finnish Hematology Registry and Clinical Biobank (FHRB; https://www.fhrb.fi/). All patients signed a written informed consent. The study was approved by the Helsinki University Hospital Ethical Committee, and it was conducted in accordance with the Declaration of Helsinki. FHR contained data of 141 Ph+ ALL adult patients (years 1984-2020). A total of 82 patients had received TKI-based therapies first-line and were selected for the biomarker analyses. The median overall survival (OS) of was 87.6 months (95% confidence interval [CI]: 51.3-169.8; n=82). Of the 82 TKI-era patients, 36 received CVAD, nine

CVAD+pegasparaginase, 18 MEA (mitoxantrone+etoposide+cytarabine) and 13 NOPHO ALL-2008 non-HR as an induction regimen. Six patients received other/customized induction treatments, and of them three were treated solely with steroids and TKI. Two of the patients who were treated with TKI and steroids succumbed to leukemia rapidly, but one patient is still alive, more than 6 years from the diagnosis. The outcome of imatinib (n=43) and dasatinib-treated (n=39) patients did not differ statistically. For imatinib-treated patients, 3-year and 5-year OS estimates were 67% and 58%, respectively, and for dasatinib-treated patients 64% and 51%. Allogeneic hematopoietic stem cell transplantation (alloHSCT) was associated with better outcome in the imatinib (Online Supplementary Figure S1A and B), but not in the dasatinib-treated patients (Online Supplementary Figure S1C and D), even though the dasatinib-treated, non-allotransplanted patients were significantly older than the transplanted patients (median age 58; range, 28-79 years vs. 41 years; range, 20-69 years; P=0.009, Wilcoxon signed-rank test). Presently, Finnish Leukemia Group recommends dasatinib as frontline TKI in adult Ph+ ALL. Therefore, the dasatinib-treated patients in this study reflect a more modern treatment era (median year of diagnosis 2015; range, 2009-2020 vs. 2006; range, 2001-2020; P<0.0001, Wilcoxon signed-rank test), which may partly explain the difference in survival. In a single prospective randomized study, dasatinib had significant survival benefit compared to imatinib.1 In addition, the benefit of alloHSCT was no longer evident in a dasatinib-treated population.3,5 Factors that favor dasatinib over imatinib include more potent and broader kinase inhibition, blood-brain barrier penetration, and possible anti-leukemic immunomodulatory effects.6 In total, 43 of the TKI-treated patients (52%) were allotransplanted. The non-allotransplanted patients were expectedly older (median age 64 years; range, 28-80 years vs. 42 years; range, 19-69 years; P<0.0001, Wilcoxon signed-rank test), but the groups did not otherwise differ significantly. AlloHSCT was associated with better outcome (Online Supplementary Figure S2A and B), but after excluding elderly (age >65 years) patients, the survival advantage disappeared (n=63; Online Supplementary Figure S2C and D). This exclusion was done to allow more realistic comparison, as over 65-year-olds are in most cases considered ineligible for alloHSCT. Especially in a retrospective, real-life cohort, allotransplanted and non-transplanted patients represent differing entities, with the tendency of the non-transplanted patients to be older and more often non-eligible for intensive treatment modal-

Haematologica | 107 August 2022

1971

LETTER TO THE EDITOR

Figure 1. Detected mutations in the analyzed samples. (A) The detected mutations in the relapse-phase samples and their relation to the given tyrosine kinase inhibitor treatment. Timeline starting from the diagnosis. For the T315I digital droplet polymerase chain reaction (ddPCR) assay, RNA was extracted using a QIAamp RNA Blood Mini kit (Qiagen, Hilden, Germany) and 2 mg was converted to cDNA using a SuperScript VILO cDNA Synthesis Kit (ThermoFisher, Waltham, MA) according to the manufacturer’s protocol. A 40 cycle PCR amplification was performed with a forward primer located in BCR exon 1 and reverse primer in ABL1 exon 10,2 using Q5 High Fidelity DNA polymerase (New England Biolabs, Ipswich, MA) according to the manufacturer’s protocol. ddPCR was performed on the 4 dilutions (105 to 108) using ddPCR Supermix for Probes on a QX200 ddPCR system (Bio-Rad, Hercules, CA) with forward primer: GGTCTGCACCCGGGAG, reverse primer: AGGTAGTCCAGGAGGTTC, wild-type probe: HEX-CCGTTCTATATCATCACTGAGTTCATGACCTAGAACG-BHQ1 and T315I probe: FAM-CCGTTCTATATCATCAtTGAGTTCATGACCTAGAACGG-BHQ1. Cycling conditions were 95°C for 10 minutes, followed by 40 cycles of 94°C for 30 seconds and 60⁰C for 60 seconds. (B) The detected mutations in the diagnosis-phase samples. Copy-number alterations in IKZF1, CDKN2A/B, PAX5, EBF1, ETV6, BTG1, and RB1 genes were detected with SALSA MLPA Probemix P335 ALL-IKZF1 kit (MRC Holland, Amsterdam, the Netherlands). The assay was performed according to the manufacturer’s protocol and the data were analyzed with Coffalyser.Net software (MRC Holland, Amsterdam, the Netherlands). Both diagnosis and relapse-phase samples were analyzed with a targeted next-generation sequencing gene panel consisting of 75 leukemia-associated genes. 150 ng of genomic DNA was processed according to SeqCap EZ HyperCap Workflow User’s Guide, v2.1 Dec 2017 Enzymatic Fragmentation (Kapa Biosystems, Inc., Wilmington, MA, USA) using Continued on following page. Haematologica | 107 August 2022

1972

LETTER TO THE EDITOR Unique Dual Index adapters by IDT (Integrated DNA Technologies, Coralville, IA, USA). Library quality check was performed using LabChip GX Touch HT High Sensitivity assay (PerkinElmer, USA). 7 cycles were used for precapture amplification. SeqCap custom captures (170621_HG38_ALL-75G_EZ_HX3) were performed in 6-7 samples multiplexed DNA Sample Library Pools using 600 µg of each library. 10 cycles were used for post capture amplification. The captured library pools were quantified for sequencing using KAPA Library Quantification Kit (KAPA Biosystems, Wilmington, MA, USA) and 2100 Bioanalyzer High sensitivity kit. The samples were sequenced in 3 batches. The first batch was sequenced with Illumina HiSeq2500 system in HiSeq high output mode using v4 kits (Illumina, San Diego, CA, USA). Read length for the paired-end run was 2x101 bp. The following batches were sequenced with Illumina NovaSeq system using S4 flow cell with lane divider (Illumina, San Diego, CA, USA) and v1.0 chemistry. Read length for the paired-end run was 2x101.

ities, making direct comparison difficult. After the exclusion, the non-allotransplanted patients were still significantly older than the transplanted ones (median 54 years; range, 28-64 years vs. 42 years; range, 19-64 years; P=0.003, Wilcoxon signed-rank test). The patient characteristics between these two cohorts did not otherwise differ significantly. In the non-allotransplanted patients, 27% (7/26) of deaths

were caused by non-leukemia-related reasons (such as heart failure, gastric cancer, breast cancer), 50% (13/26) were due to relapse or primary refractory disease, and 23% (6/26) were treatment-related. In the allotransplanted cohort, 65% (11/17) of the deaths were due to transplantation-related causes and 29% (5/17) due to a relapse. One death was due to other causes. In the competing risks analysis, the proportion of both leukemia-re-

Figure 2. IKZF1 plus genotype predicts poor survival. (A) Overall survival and (B) relapse-free survival of patients according to the presence of IKZF1 deletion. (C) Overall survival and (D) relapse-free survival of patients according to the presence of IKZF1 plus (IKZF1 deletion with CDKN2A/B and/or PAX5 deletion). Events after 80 months are not shown. Kaplan-Meier estimate, log rank test. Haematologica | 107 August 2022

1973

LETTER TO THE EDITOR Table 1. Univariate Cox regression analysis of overall and relapse-free survival in first-line tyrosine kinase inhibitor-treated patients.

Univariate analysis for OS HR (95% CI)

Univariate analysis for RFS HR (95% CI)

N of observations

Age, continuous

1.04 (1.02-1.06)

0.0004

1.04 (1.01-1.06)

0.001

Age, categorical (>65 y vs. ≦ 65 y)

3.32 (1.75-6.26)

<0.0001

3.04 (1.63-5.67)

0.0002

AlloHSCT

0.45 (0.24-0.82)

0.008

0.51 (0.28-0.93)

0.03

Dasatinib vs. imatinib

1.16 (0.60-2.24)

0.7

1.21 (0.64-2.28)

0.6

1.003 (0.999-1.007)

0.2

1.003 (0.998-1.007)

0.2

WBC ≧30x109/L

1.81 (0.98-3.37)

0.06

1.76 (0.96-3.23)

0.06

WBC ≧50x109/L

1.77 (0.89-3.50)

0.1

1.64 (0.83-3.23)

0.1

Hemoglobin

1.002 (0.990-1.014)

0.7

1.003 (0.99-1.02)

0.6

Platelets

0.998 (0.993-1.002)

0.3

0.998 (0.993-1.002)

0.3

LDH

1.00 (0.9997-1.001)

0.5

1.00 (0.9998-1.001)

0.5

BM blast%

1.02 (0.99-1.06)

0.2

1.03 (0.99-1.06)

0.1

P210 vs. P190, transcript type

0.50 (0.17-1.46)

0.2

0.61 (0.23-1.64)

0.3

ACA (karyotype)

1.05 (0.48-2.27)

0.9

0.92 (0.43-1.93)

0.8

MRD-negative in 3 months, no vs. yes

0.65 (0.32-1.33)

0.2

0.61 (0.31-1.21)

0.2

IKZF1 deletion

1.59 (0.52-4.84)

0.4

1.27 (0.46-3.46)

0.6

IKZF1 plus

8.37 (2.71-25.83)

<0.0001

8.56 (3.04-24.07)

<0.0001

WBC

*OS: overall survival; RFS: relapse-free survival; HR: hazard ratio; CI: confidence interval; y: years; AlloHSCT: allogeneic hematopoietic stem cell transplantation; WBC: white blood cell count; LDH: lactate dehydrogenase; BM: bone marrow; ACA: additional chromosomal abnormalities; MRD: minimal residual disease.

lated deaths (P=0.02, Gray’s test) and other than leukemia or treatment-related causes of death (P=0.02, Gray’s test) were more common in the non-allotransplanted patients. A proportion of Ph+ ALL patients experience prolonged survival with TKI-based therapies only,7 and many die due to non-leukemia-related causes. TKI modulate the immune system, and part of their effect might stem from overcoming the immunosuppressive state in the leukemic bone marrow.6 The targeted NGS panel detected only a single missense PAX5 mutation (p.V26G) in the diagnostic samples (n=41) and no point mutations in the IKZF1 or CDKN2A/B genes. In the relapse-phase samples (n=11), ABL1 mutations causing TKI resistance were frequent (p.T315I, n=5; p.E355A, n=1; p.F317I, n=1; p.T315I+p.Y253F, n=1; p.E255K+p.Y253F, n=1; Figure 1A). In addition, one TP53 mutation (p.C176F) was detected in a relapse-phase sample. The variant calling was performed as previously described.8 As especially T315I mutations possess clinical significance, we designed a ddPCR assay capable of detecting T315I down to 0.04% variant allele frequency (VAF). Altogether 32 samples (26

pretreatment, 6 relapse-phase) from 25 patients were examined. Contrary to previous studies, where up to 25% of Ph+ ALL patients have been reported to harbor pretreatment T315I subclones,7 only one baseline sample tested positive for T315I with a VAF of 0.10%. The patient was treated with imatinib but relapsed 9 months later with an E255K mutation.2 Imatinib was switched to dasatinib, followed by a relapse with T315I 2 months later. In line with our results, Short et al. reported only a single pretreatment T315I mutation in 63 Ph+ ALL samples using highly accurate duplex-sequencing.9 Importantly, these baseline mutations did not correlate with treatment success. In view of these data, screening for pretreatment T315I seems unwarranted. Our ddPCR assay detected T315I mutations in five of six of the relapse phase samples, all with a VAF >25%. All these mutations were also detected by clinical routine mutation analysis and the NGS panel. The detected mutations from the relapse phase samples are collected to Figure 1A. Most Ph+ ALL patients harbor IKZF1 deletions, although their role in the risk stratification remains undecided.3,10

Haematologica | 107 August 2022

1974

LETTER TO THE EDITOR Using multiplex ligation-dependent probe amplification analysis we detected IKZF1 deletions in 75% (n=30), CDKN2A/B deletions in 42% (n=17), and PAX5 deletions in 28% (n=11) of the analyzed samples (n=40). Other candidate gene alterations detected by the assay were less frequent (EBF1, n=1; ETV6, n=1; BTG1, n=6; RB1, n=5). In 38% of cases, IKZF1 deletion was accompanied by a deletion in CDKN2A/B and/or PAX5 genes (IKZF1 plus). The detected mutations from the diagnosis phase samples are shown in Figure 1B. The presence of IKZF1 deletions alone had no effect on survival (Figure 2A and B), whereas IKZF1 plus genotype lead to inferior prognosis (Figure 2C and D). Within the IKZF1 plus group, alloHSCT did not improve survival, although the cohort size is limited (Online Supplementary Figure S3A and B). Cumulative incidence for relapse was 56% at 12 months after alloHSCT in IKZF1 plus patients and only 7% in nonplus patients (Online Supplementary Figure S3C). Non-relapse mortality did not differ between these two cohorts (Online Supplementary Figure S3D). IKZF1 plus patients had higher white blood cell count at diagnosis, but the cohorts did not otherwise differ significantly at baseline. When analyzing all first-line TKI-treated patients, in Cox regression univariate analyses IKZF1 plus genotype, age, and alloHSCT were the only significant predictors for both OS and relapse-free survival (Table 1). No consensus currently exists how treatment of IKZF1 plus patients should be modified. AlloHSCT may not improve survival in this group.3 Limited data indicate that a combination of dasatinib and blinatumomab might prove beneficial.4 In the younger patients, a modern MRD-driven intensive chemotherapy protocol seemed effective.10 IKZF1 plus patients may be primary resistant at the progenitor/stem cell level to TKI-based therapies and more detailed mechanistic studies may give insight to effective treatment alternatives.11 Retinoids, immunomodulatory drugs, crizotinib, and a combination of asciminib and ponatinib are currently being investigated and may provide an alternative also for the elderly patients and those non-eligible for intensive therapies.12–15 To conclude, testing for CNA should be implemented in the routine diagnostics of Ph+ ALL. IKZF1 plus genotype constitutes a high-risk group, which may benefit from immuno-oncological or intensified treatment approaches.

Research Program, University of Helsinki and Helsinki University Hospital Comprehensive Cancer Center, Helsinki, Finland; 2

Department of Pathology, Zealand University Hospital, Roskilde,

Denmark; 3Laboratory of Genetics, HUS Diagnostic Center, Hospital District of Helsinki and Uusimaa (HUS), Helsinki, Finland; 4

Medical and Clinical Genetics, University of Helsinki, Helsinki

University Hospital, Helsinki, Finland; 5iCAN Digital Precision Cancer Medicine Flagship, Helsinki, Finland; 6Institute for Molecular Medicine Finland, University of Helsinki, Helsinki, Finland; 7Division of Hematology, Oulu University Hospital, Oulu, Finland; 8Division of Hematology, Tampere University Hospital, Tampere, Finland; 9Division of Hematology, Kuopio University Hospital, Kuopio, Finland; 10Division of Hematology, Turku University Hospital, Turku, Finland; 11Division of Hematology, Helsinki University Hospital Comprehensive Cancer Center, Helsinki, Finland and 12Department of Clinical Chemistry, University of Helsinki, Helsinki, Finland. Correspondence: K. PORKKA - kimmo.porkka@helsinki.fi https://doi.org/10.3324/haematol.2021.280578 Received: December 23, 2021. Accepted: April 14, 2022. Prepublished: April 28, 2022. ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

Disclosures TS (not related to this study) is a member of the advisory board of Celgene and AbbVie; is a member of the advisory board of and received lecture fees from Pfizer and Janssen-Cilag; received lecture fees from Bristol Myers Squibb; received congress fees from and is a member of the advisory board of Novartis; received congress fees from Amgen. MP (not-related to this study) is a member of the advisory board of Pfizer and AbbVie; received lecture and congress fees from Novartis. OB received consultancy fees from Novartis and Sanofi. SM (not related to this study) received research funding from Novartis, BMS, Janpix, and Pfizer. All other authors have no conflicts of interest to disclose. Contributions HH, NP, PE, SM and KP delevoped the concept and design of the study; HH, TS, MSä, MSi, MP, MI-R, PK, EE and KP collected and

Authors

assembled data; HH, NP, MK, OB and KP analyzed and interpreted the data. All authors wrote the manuscript and gave their approval of the final version. 1

1,3,4,5

Helena Hohtari, Niels Pallisgaard, Matti Kankainen,

Pekka Ellonen,

Oscar Brück,1 Timo Siitonen,7 Marjaana Säily,7 Marjatta Sinisalo,8 Marja 9

Pyörälä, Maija Itälä-Remes, Perttu Koskenvesa, Erkki Elonen, Satu 1,5,12

Mustjoki

1,5,11

and Kimmo Porkka

Acknowledgments NGS library preparation, sequencing and sequence analysis were performed by the Institute for Molecular Medicine Finland (FIMM) Technology Center, University of Helsinki. We thank laboratory

Hematology Research Unit Helsinki, Translational Immunology

technicians Jay Klievink in Hematology Research Unit Helsinki (HRUH)

Haematologica | 107 August 2022

1975

LETTER TO THE EDITOR and Minna Suvela in FIMM for technical support with the DNA

Funding

extractions and laboratory coordinator Minna Tuominen in FIMM for

This study was supported by the Doctoral Program in Clinical

technical support with multiplex ligation-dependent probe

Research at the University of Helsinki and personal grants (to HH)

amplification. We are grateful to the members of the HRUH for

from Emil Aaltonen Foundation, Ida Montin Foundation, Blood

discussions and technical help. We thank research nurses Anne

Disease Research Foundation, Finnish Hematology Association,

Gesterberg, Jenni Raali and Susanna Helkkula for help with clinical

Finnish Medical Foundation, Biomedicum Helsinki Foundation, Paulo

data. We thank Dr Veli Kairisto in Tykslab, Dr Taru Kuittinen in Kuopio

Foundation, (to SM) Finnish Cancer Organizations, Sigrid Juselius

University Hospital and clinical laboratory geneticists Anne Juvonen

Foundation, Signe and Ane Gyllenberg Foundation, Relander

and Tarja Salonen in HUSLAB for help with clinical samples. The

Foundation, and state funding for university-level health research in

samples of this project were provided by Finnish University Hospital

Finland. The laboratory analytics costs of this study were funded by

clinical laboratories and the Finnish Hematology Registry and Clinical

Incyte.

Biobank (FHRB) with appropriate ethics approval (Dnro 202/06.01.00/2013). We thank all the patients for their generous

Data-sharing statement

participation. The FHRB Biobank is supported by the Finnish

The datasets generated and analyzed for the current study can be

Association of Hematology, the Finnish Red Cross Blood Service,

shared upon request to the corresponding author. All data will be

Institute for Molecular Medicine Finland, and the participating

shared in an anonymized form, as regulated by the General Data

hospitals in Finland.

Protection Regulation (GDPR) of the European Union.

References 1. Shen S, Chen X, Cai J, et al. Effect of dasatinib vs imatinib in the treatment of pediatric Philadelphia chromosome-positive acute lymphoblastic leukemia. JAMA Oncol. 2020;6(3):358-366. 2. Soverini S, Benedittis CD, Polakova KM, et al. Unraveling the complexity of tyrosine kinase inhibitor–resistant populations by ultra-deep sequencing of the BCR-ABL kinase domain. Blood. 2013;122(9):1634-1648. 3. Chiaretti S, Ansuinelli M, Vitale A, et al. A multicenter total therapy strategy for de novo adult Philadelphia chromosome positive acute lymphoblastic leukemia patients: final results of the GIMEMA LAL1509 protocol. Haematologica. 2021;106(7):1828-1838. 4. Foà R, Bassan R, Vitale A, et al. Dasatinib–blinatumomab for Ph-positive acute lymphoblastic leukemia in adults. N Engl J Med. 2020;383(17):1613-1623. 5. Chang J, Douer D, Aldoss I, et al. Combination chemotherapy plus dasatinib leads to comparable overall survival and relapse‐free survival rates as allogeneic hematopoietic stem cell transplantation in Philadelphia positive acute lymphoblastic leukemia. Cancer Med. 2019;8(6):2832-2839. 6. Kreutzman A, Porkka K, Mustojoki S. Immunomodulatory effects of tyrosine kinase inhibitors. Int Trends Immun. 2013;01:22-33. 7. Rousselot P, Coudé MM, Gökbuget N, et al. Dasatinib and lowintensity chemotherapy in elderly patients with Philadelphia chromosome-positive ALL. Blood. 2016;128(6):774-782. 8. Dufva O, Kankainen M, Kelkka T, et al. Aggressive natural killercell leukemia mutational landscape and drug profiling highlight JAK-STAT signaling as therapeutic target. Nat Commun. 2018;9(1):1567.

9. Short NJ, Kantarjian H, Kanagal-Shamanna R, et al. Ultraaccurate Duplex Sequencing for the assessment of pretreatment ABL1 kinase domain mutations in Ph+ ALL. Blood Cancer J. 2020;10(5):61. 10. Moorman AV, Barretta E, Butler ER, et al. Prognostic impact of chromosomal abnormalities and copy number alterations in adult B-cell precursor acute lymphoblastic leukaemia: a UKALL14 study. Leukemia. 2022;36(3):625-636. 11. Rogers JH, Gupta R, Reyes JM, et al. Modeling IKZF1 lesions in B-ALL reveals distinct chemosensitiviy patterns and potential therapeutic vulnerabilities. Blood Adv. 2021;5(19):3876-3890. 12. Churchman ML, Low J, Qu C, et al. Efficacy of retinoids in IKZF1-mutated BCR-ABL1 acute lymphoblastic leukemia. Cancer Cell. 2015;28(3):343-356. 13. Harama D, Yahata T, Kagami K, et al. IMiDs uniquely synergize with TKIs to upregulate apoptosis of Philadelphia chromosomepositive acute lymphoblastic leukemia cells expressing a dominant-negative IKZF1 isoform. Cell Death Discov. 2021;7(1):139. 14. Eide CA, Zabriskie MS, Stevens SLS, et al. Combining the allosteric inhibitor asciminib with ponatinib suppresses emergence of and restores efficacy against highly resistant BCR-ABL1 Mutants. Cancer Cell. 2019;36(4):431-443.e5. 15. Mian AA, Haberbosch I, Khamaisie H, et al. Crizotinib acts as ABL1 inhibitor combining ATP-binding with allosteric inhibition and is active against native BCR-ABL1 and its resistance and compound mutants BCR-ABL1T315I and BCR-ABL1T315I-E255K. Ann Hematol. 2021;100(8):2023-2029.

Haematologica | 107 August 2022

1976

LETTER TO THE EDITOR

PD-1/PD-L1 expression is frequent and correlated with lymphocyte density in Erdheim-Chester disease In patients with Erdheim-Chester disease (ECD), a rare histiocytosis of the L group of the 2016 revised classification, the accumulation of foamy histiocytes leads to multisystemic disease with the involvement of various organs.1 The detection of BRAFV600E mutation in up to 70% of ECD tissue samples tested has led to the reclassification of ECD as a myeloid neoplasm, which has already considerably improved the treatment of adults with histiocytoses, whether wild-type or carrying BRAFV600E mutations.2,3,4 In November 2017, the BRAF inhibitor vemurafenib was approved by the US Food and Drug Administration (FDA) for the treatment of BRAFV600E-mutant ECD. The MEK inhibitor cobimetinib will probably follow this year in the US. Vemurafenib has an orphan drug designation for BRAFV600E-mutant ECD in Europe, but the therapeutic options for multisystem and refractory ECD, and for other histiocytic neoplasms, may be limited in Europe and elsewhere due to the current lack of access to targeted therapies for such indications. Moreover, further improvements to ECD treatment are required, as targeted therapies can cause morbidity and late treatment effects, and patients almost always experience relapses when these therapies are stopped.5 Over the last 10 years, immune checkpoint inhibitors, such as programmed death-1 (PD-1) and programmed death ligand-1 (PD-L1) inhibitors, have proven remarkably effective for the treatment of several hematological and solid-organ cancers.6–8 This high efficacy has led to their approval for use in diverse indications being fast-tracked by the US FDA. In 2015, Gatalica et al. reported a high expression of PDL1 (≥2+/≥5%) in three of four ECD cases tested, all of which presented BRAFV600E mutations.9 Shortly after the publication of this article, we decided to analyze a larger case series of patients, to see if PD-L1 expression could provide a rationale for the addition of immune checkpoint inhibitors to treatment regimens for multisystemic and/or refractory histiocytoses. Goyal et al. recently reported conflicting results for an additional three cases of ECD, which displayed low levels of PD-L1 expression on IHC (1415%).10 This led us to extend our series analysis further. We included 54 ECD patients in our study and biopsy samples were reviewed for all patients (Table 1). Lymphocyte and plasma cell densities were evaluated and classified as low (+), intermediate (++), or high (+++) on hematoxylin and eosin (H&E) staining (Figure 1). Immunostaining was performed to detect PD-L1 (QR1Clone) in histiocytes and PD-1 (NAT105 clone) in lymphocytes. PD-L1 levels were assessed as the percentage of histiocytes

positive for this molecule. The combined positivity score (CPS), which is the number of PD-L1 staining cells (tumor cells, lymphocytes, macrophages) divided by the total number of viable tumor cells, multiplied by 100, is a predictive marker for response to the therapy with inhibitors of immune checkpoints in various types of cancer. In our study, the percentage of PD-L1+ histiocytes was used rather than the CPS because no distinction is possible between tumoral and inflammatory histiocytes in ECD. Patients were PD-L1-positive if ≥ 5% of the histiocytes expressed this molecule. PD-1 immunostaining was evaluated and classified as weak (+), moderate (++), or strong (+++). Patients were PD1-positive if PD-1 immunostaining was moderate or strong. PD-1/PD-L1 expression and the density of lymphocyte and plasma cell infiltration assessment was performed subjectively and classified in three categories by comparing slides with reference patterns as presented in Figure 1. C-reactive protein levels were assessed at diagnosis, at the time of the biopsy. Table 1. Clinical and biological characteristics of ECD patients.

Variables Age, yr (IQR) Male, N (%) Mutation status, N (%) WT BRAFV600E MAP2K1 NRAS Not determined Langerhans cell histiocytosis, N (%) PD-1+, N (%) -+ - ++ PD-L1+, N (%) - 5 - 50 % - >/= 50 % PD-L1+ / PD-1-, N (%) PD-L1- / PD-1+, N (%) PD-L1+ / PD-1+, N (%) PD-L1- / PD-1-, N (%) Lymphocyte/plasma cell density, N (%) Low (+) Intermediate (++) High (+++) High C-reactive protein (> 5 mg/L) at diagnosis *, N (%)

ECD patients (N=54) 62 (55-70) 42 (78) 15 (28) 27 (50) 5 (9) 2 (4) 5 (9) 4 (7) 31 (57) 21 (39) 10 (18) 22 (40) 13 (23) 9 (17) 4 (7) 13 (24) 18 (33) 19 (35) 34 (63) 12 (22) 8 (15) 23 (74)

*% calculated for a denominator of 31 patients. ECD: Erdheim-Chester disease; yr: years; IQR: interquartile range; WT: wild-type; PD-1: programmed death-1; PD-L1: programmed death ligand 1.

Haematologica | 107 August 2022

1977

LETTER TO THE EDITOR

Figure 1. Histopathological images displaying all types of staining over three cases (three BRAFV600E mutated patients). (A) In this case, there is only weak lymphocyte infiltration (+) (hematoxylin and eosin [H&E] x200 magnification). (B) the histiocytes are CD68 positive. (C and D) PD-1 and PD-L1 are negative (immunoperoxidase, x200). (E) In this case, the lymphocyte density is moderate (++) (H&E x200); F) the histiocytes are CD68 positive. (G) PD-1 expressed by lymphocytes was evaluated as mild (+), and H) PD-L1 is expressed by 30% of histiocytes (immunoperoxidase, x200). (I) In this case, the lymphocyte density is marked (+++) (H&E x200); J) the histiocytes are CD68-postive (immunoperoxidase, x200). (K) PD-1 expressed by lymphocytes was evaluated as moderate (++), and L) PD-L1 is expressed in 50% of histiocytes (immunoperoxidase, x200).

Continuous variables are expressed as the mean and standard deviation, and categorical variables are expressed as numbers and percentages. The significance of differences between groups of patients was evaluated in Student’s t-tests for continuous data and Pearson's chi-squared tests with Yates' continuity correction for categorical data. We used RStudio (Version 1.1.456) for analyses. The patients had a mean age of 62 years, and 42 (78%) patients were male. BRAFV600E mutation was detected in 27 patients (50%), MAP2K1 mutation in five (9%) and NRAS mutation in two (4%). Four of the 54 ECD patients also had Langerhans-cell histiocytosis (LCH) (Table 1; Figure 2). Overall, 22 patients were positive for PD-L1 (40%), 31 were positive for PD-1 (57%) and 18 were positive for both (33%). Lymphocyte/plasma cell infiltration density was low in 34 (63%) patients, moderate in 12 patients (22%) and of high in eight patients (15%) (Figure 2). We found a strong association between PD-1 status and lymphocyte/plasma cell density: density was intermediate-to-high in 18 (58%) PD-1-positive patients versus in only two PD-1-negative patients (9%) (P<0.001). A similar association was found concerning PD-L1 status:

lymphocyte/plasma cell density was intermediate-to-high in 15 (68%) PD-L1-positive patients whereas it was intermediate-to-high in only five (16%) PD-L1-negative patients (P<0.0003). PD-L1 positivity was negatively associated with BRAFV600E mutation status: five (25%) PD-L1-positive patients were BRAFV600E-mutated, whereas 22 (76%) PD-L1-negative patients had the mutation (P=0.001). We found no association between PD-1 positivity and BRAFV600E mutation (P=0.39). PD-1 status and PD-L1 status were significantly associated with one another: 37 (69%) patients were either PD-1-/PDL1- or PD-1+/PD-L1+, 19 (35%) were PD-1-/PD-L1-, and 18 (33%) PD-1+/PD-L1+ (P=0.006). Nine (75%) PD-L1-/PD-1+ patients had BRAFV600E mutations. By contrast, none of the three PD-L1+/PD-1- patients had BRAFV600E mutations. Patients with BRAFV600E mutations had significantly lower levels of lymphocyte/plasma cell infiltration, with intermediate-to-high cell density detected in only seven (26%) patients with mutations, versus 13 (59%) wild-type (WT) patients (P=0.04). We report the largest study to date exploring PD-1 status

Haematologica | 107 August 2022

1978

LETTER TO THE EDITOR

Figure 2. Mutation and PD-1/PD-L1 status. (A) Distribution of BRAFV600E, NRAS and MAP2K1 mutations. (B) Distribution of PD-1/PDL1 C). (D to F) Associations of BRAFV600E and lymphocyte infiltration with PD-1/PD-L1 status. PD-1: programmed death-1; PD-L1: programmed death ligand 1.

and PD-L1 status in ECD. We found that PD-1 and/or PDL1 were frequently expressed in ECD. Positivity for PD-L1 was significantly associated with an absence of BRAFV600E mutation, and intermediate-to-high lymphocyte/plasma cell density. Our data suggest that they may be two phenotypes, one combining WT BRAF status with intermediate-to-high lymphocyte density and positivity for PD-L1 (+/- PD-1), and the other combining a BRAFV600E mutated phenotype with a low lymphocyte/plasma cell density and negativity for PD-L1 (+/- PD-1). Sengal et al.11 previously performed a phenotypic analysis of LCH lesions and reported an association between BRAFV600E expression and PD-L1 expression that we do not find in our series of ECD samples. We evaluated lymphocyte and plasma cell density, but did neither analyze T cells nor dendritic cells (DC). Furthermore, there are subtle but profound differences between LCH and ECD. LCH cells belong to the DC lineage, whereas ECD cells have phenotype of macrophages. Regarding the mechanistic effects of the expression PD-1 and PD-L1 on histiocyte proliferation and

lymphocyte activity, it is still unknown whether ECD cells do proliferate or if a proliferation occurs in mutated monocytes seeding the tissues. Single cell transcriptomic data will probably help address these questions. The recent success of immune checkpoint blockade therapy for many different types of hematological and solid-organ cancers, and the demonstration of immune checkpoint antigen expression in the tissues of patients with ECD suggest that such therapies could be tested for the treatment of patients with multisystemic and refractory ECD, particularly those with a contraindication for MEK inhibitors.

Authors Fréderic Charlotte,1 Fleur Cohen-Aubart,2* Lévi-Dan Azoulay,2* Zahir Amoura,2 Jean-François Emile3,4 and Julien Haroche2 Sorbonne Université, Assistance Publique-Hôpitaux de Paris,

Haematologica | 107 August 2022

1979

LETTER TO THE EDITOR

Département d’Anatomo-Pathologie, Hôpital Pitié-Salpêtrière, Paris; 2 Sorbonne Université, Assistance Publique-Hôpitaux de Paris, Service de Médecine Interne 2, Centre National de Référence des Histiocytoses, Hôpital Pitié-Salpêtrière, Paris; 3Service de Pathologie, Hôpital Ambroise Paré, Boulogne and 4EA4340-BECCOH, Université de Versailles SQY, Université Paris-Saclay, Boulogne, France

©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license Disclosures No conflicts of interest to disclose. Contributions FC, FC-A, J-FE and JH designed the study; FC, L-DA and JH collected the data; L-DA and JH conducted the statistical analysis; L-DA, FC-A, L-DA, ZA, J-FE, and JH analyzed and interpreted the data; FC, FC-A, L-DA and JH wrote the manuscript. All the authors critically reviewed and approved the final version of the manuscript.

*FC-A and L-DA contributed equally to this work. Correspondence: J. HAROCHE - julien.haroche@aphp.fr https://doi.org/10.3324/haematol.2021.280312

Data-sharing statement The datasets used and/or analyzed during the current study are available from the corresponding authors (JH) on reasonable request.

Received: November 4, 2021. Accepted: April 14, 2022. Prepublished: April 28, 2022.

References 1. Emile JF, Abla O, Fraitag S, et al. Revised classification of histiocytoses and neoplasms of the macrophage-dendritic cell lineages. Blood. 2016;127(22):2672-2681. 2. Haroche J, Cohen-Aubart F, Emile JF, et al. Dramatic efficacy of vemurafenib in both multisystemic and refractory ErdheimChester disease and Langerhans cell histiocytosis harboring the BRAF V600E mutation. Blood. 2013;121(9):1495-1500. 3. Haroche J, Cohen-Aubart F, Emile JF, et al. Reproducible and sustained efficacy of targeted therapy with vemurafenib in patients with BRAF(V600E)-mutated Erdheim-Chester disease. J Clin Oncol. 2015;33(5):411-418. 4. Diamond EL, Durham BH, Ulaner GA, et al. Efficacy of MEK inhibition in patients with histiocytic neoplasms. Nature. 2019;567(7749):521-524. 5. Cohen-Aubart F, Emile JF, Carrat F, et al. Targeted therapies in 54 patients with Erdheim-Chester disease, including follow-up after interruption (the LOVE study). Blood. 2017;130(11):1377-1380. 6. Sun L, Zhang L, Yu J, et al. Clinical efficacy and safety of anti-

PD-1/PD-L1 inhibitors for the treatment of advanced or metastatic cancer: a systematic review and meta-analysis. Sci Rep. 2020;10(1):2083. 7. Allen PB, Savas H, Evens AM, et al. Pembrolizumab followed by AVD in untreated early unfavorable and advanced-stage classical Hodgkin lymphoma. Blood. 2021;137(10):1318-1326. 8. Pardoll DM. The blockade of immune checkpoints in cancer immunotherapy. Nat Rev Cancer. 2012;12(4):252-264. 9. Gatalica Z, Bilalovic N, Palazzo JP, et al. Disseminated histiocytoses biomarkers beyond BRAFV600E: frequent expression of PD-L1. Oncotarget. 2015;6(23):19819-19825. 10. Goyal G, Lau D, Nagle AM, et al. Tumor mutational burden and other predictive immunotherapy markers in histiocytic neoplasms. Blood. 2019;133(14):1607-1610. 11. Sengal A, Velazquez J, Hahne M, et al. Overcoming T-cell exhaustion in LCH: PD-1 blockade and targeted MAPK inhibition are synergistic in a mouse model of LCH. Blood. 2021;137(13):1777-1791.

Haematologica | 107 August 2022

1980

LETTER TO THE EDITOR

Blood cell and marrow changes in patients with Kikuchi disease Kikuchi disease (KD) is a self-limiting lymphadenitis,1,2 common in East Asia3 but rare in other countries.1 Although it most often affects young adults, it can occur at any age.1,2 Abnormal blood cell counts are the most wellknown laboratory abnormality.1,2 In this study, we reviewed blood cell counts and bone marrow studies in patients with KD. We found that the rates and recovery time of abnormal cell counts differ between age groups; children more commonly develop pancytopenia, and their anemia frequently persists for several months. Bone marrow and reticulocyte data suggest that myelosuppression is the mechanism responsible for cytopenia. Few patients with KD developed hemophagocytic lymphohistiocytosis (HLH). The blood and marrow changes were distinctly different between KD and HLH. We screened 367 patients with KD, and 282 (77%) had complete blood count (CBC) data: 101 were male, and 181 were female (M:F=0.56). The mean age was 26±10 (range, 4–66) years. On average, female patients were older than male patients (27± 9 vs. 24±10 years, P=0.009). Female patients predominantly developed KD in young adulthood, but the age distribution for male patients was relatively

even (Figure 1A). Children younger than 15 years with KD were predominantly male (M:F=2.67), and patients older than 15 were predominantly female (Figure 1B). Few researchers have pointed out the sex ratio difference between children and adults. However, all previous adult-including studies identified female patients as predominant, with a male–female ratio ranging from 0.28 to 0.91.4-6 By contrast, most pediatric studies have reported male patients as predominant or a male–female ratio close to one.7,8 Many experts consider the predominance of female patients with KD controversial.1,2 We believe the difference is related to the age of patients. The frequency of abnormalities is presented in Figure 1C. The definitions of abnormalities are listed in the Online Supplementary Table S1. Of the 282 patients, anemia (22%) was the most common abnormality, followed by lymphopenia (17%), neutropenia (11%), atypical lymphocytes (9%), and thrombocytopenia (8%). Increased cell counts were relatively rare. The results were similar to previous reports.5,9-11 We also found that most cytopenias occurred in a single cell lineage. Pancytopenia (2%) and bicytopenia (5%) were uncommon.

Continued on following page. Haematologica | 107 August 2022

1981

LETTER TO THE EDITOR

Figure 1. Frequency of abnormal blood cell counts. (A) Age distribution of all patients (upper), female patients (middle), and male patients (lower). (B) Percentage of male (black) and female (gray) patients by age group. (C) Frequency of abnormal cell counts.

We plotted the frequency of abnormal cell counts by age group, revealing that the abnormalities exhibited a similar bimodal distribution (Figure 2A). Young adults had the lowest rate, and children and middle-aged patients had higher abnormality rates. The rate of lymphopenia was significantly different between age groups (P<0.001). For all the abnormalities, pairwise comparisons revealed significant differences between age groups (Figure 2A). Likewise, the distribution of pancytopenia and bicytopenia also exhibited bimodal patterns (Figure 2B). The rates of pancytopenia differed between age groups (P=0.024). Children younger than 15 years had a higher rate of pancytopenia than patients in other age groups (adjusted P=0.008). The results suggest that patients of extreme ages, especially children, are more vulnerable to cytopenia. We reviewed the follow-up CBC in these patients. The recovery time, summarized in Figure 2C, varied considerably, ranging from days to months. Cytopenia can persist for months in patients with KD—considerably longer than lymphadenopathy, which resolves within weeks.1,2 Anemia (34%) was the cytopenia that most frequently persisted for more than 6 months, and lymphopenia (4%) was the least likely to persist that long. Like the abnormality rates, the late recovery rates varied by age. We plotted the case numbers by age group (Figure 2D). Anemia usually lasted more than 6 months in children, occasionally did so in middle-aged patients, but never lasted in young adults (P<0.001). Likewise, thrombocytopenia, neutropenia, and lymphopenia lasted more

than 6 months in some patients of extreme ages but were absent in 16–25-year-old patients (Figure 2D). In summary, children more frequently develop cytopenias, and their cytopenias are more protracted than those of young adults. Therefore, pediatricians should be aware of cytopenias in children with KD. We further explored the mechanism underlying the abnormal blood cell counts. First, all the available reticulocyte count data (n=8) indicated a low erythropoietic response to anemia. Second, we reviewed bone marrow studies. Sixteen (6%) patients had undergone bone marrow biopsy (Table 1) and 88% (14/16) had hypocellular marrow with little compensatory hematopoiesis. A previous study has demonstrated that the serum of patients with KD suppressed granulopoiesis in vitro.2 Our clinical observation is consistent with their results. We quantified CD68+ histiocytes and CD123+ plasmacytoid dendritic cells in the bone marrow biopsy. Compared with that of age-matched controls, the bone marrow in patients with KD did not display increased histiocytes or plasmacytoid dendritic cells (Online Supplementary Figure S1). In most patients with KD, the lymphadenopathy features should be absent from bone marrow. Two of the patients in our study developed HLH (cases 5 and 10 in Table 1). The prevalence of HLH was 0.71% (2/282) in patients with CBC data and 0.54% (2/367) in all the patients screened. The prevalence of HLH was 0–3% in the literature.4,5,11,12 HLH is the most frequently reported bone marrow finding

Haematologica | 107 August 2022

1982

LETTER TO THE EDITOR

Figure 2. Abnormal rates by age group and cytopenia recovery. (A) Frequency of abnormal blood cell counts by age group. (B) Cytopenia lineage number by age group. The adjusted P-value of the post hoc test and pairwise comparison are denoted by *P< 0.05; **P<0.01. (C) Recovery time by type of cytopenia. (D) Case number of patients with and without late recovery (red and gray bars, respectively) by age group. Haematologica | 107 August 2022

1983

LETTER TO THE EDITOR in patients with KD,13 but most patients with KD do not have HLH. We compared our cases with an independent cohort of patients with HLH to investigate the differences between HLH and KD. The HLH cohort included 133 patients: 80 were males and 53 were females, with a higher percentage of males than KD (P<0.001). The mean age was 50 (range, 2–91) years, older than KD (P<0.001). This HLH cohort has been partially published in a previous study.14 The patients with KD (n=282) had higher hemoglobin levels (13.0±1.6 vs. 8.5±1.4 g/dL, P<0.001) and higher platelet counts (240±76 vs. 67±84 × 103/µL, P<0.001) than those with HLH (n=133; Online Supplementary Figure S2). The patients with KD had lower rates of anemia (23% vs. 98%, P<0.001), thrombocytopenia (10% vs. 92%, P<0.001), and neutropenia (14% vs. 38%, P<0.001) than did the patients with HLH (Online Supplementary Figure S2). With regard to severity, patients with KD had lower rates of severe anemia (8% vs. 69%, P<0.001), thrombocytopenia (19% vs. 96%, P<0.001), and neutropenia (33% vs. 80%, P<0.001; Online Supplementary Figure S2). In terms of the

number of cytopenia lineages, only 1% (4/282) of patients with KD had more than two lineages of severe cytopenia in contrast to 72% (96/133) of patients with HLH (P<0.001). Moreover, patients with KD (n=16) had lower bone marrow cellularity (34±25% vs. 63±21%, P<0.001) and less histiocytic infiltrate (10± 7% vs. 29±28%, P<0.001) than those with HLH (n=133; Online Supplementary Figure S2). Cytopenias are present in both patients with KD and HLH, but the frequency and severity are much lower in those with KD. HLH typically exhibits increased cellularity, histiocyte infiltrates, and hemophagocytosis in bone marrow.14,15 By contrast, KD exhibits decreased cellularity and no increase in histiocytes. In summary, we report comprehensive blood changes in patients with KD. Patients of extreme ages are more susceptible to cytopenias, which can persist for several months. The mechanism underlying cytopenias is probably mild myelosuppression. Patients with HLH exhibit severe cytopenias and compensatory hematopoiesis, but those with KD exhibit mild cytopenia and hypocellular marrow.

Table 1. Patients with bone marrow studies. Peripheral blood No

Age yr Sex

WBC x109/L

HB g/dL

Bone marrow biopsy

PLT x109/L

Mye

Ery

Cell

Mye

Ery

HPh

↓

(−)

↓↓

(−)

↓↓

(−)

↓

↑

(−)

3.32

↓

9.6

2.32

↓

11.7

3.33

↓

11.0

↓

160

2.23

↓

8.9

↓↓

2.47

↓

13.7

179

2.10

↓↓

↓

↓↓

↓

↓↓

(+)

4.89

13.1

273

1.55

↓↓

↓

↓↓

(−)

1.64

↓

12.2

↓↓

↓

↓↓

(−)

3.40

↓

11.1

↓

↓↓

(−)

1.22

↓

15.1

↓↓

0.43

(−)

1.83

↓

14.5

↓↓

1.37

↓

↓↓↓

(−)

2.22

↓

11.5

241

1.25

↓

↓↓↓

(−)

2.80

↓

13.5

265

1.83

↓

(−)

2.39

↓

12.2

255

1.15

↓

↓↓

↑

(−)

1.89

↓

8.2

↓↓

211

1.13

↓

↑↑

↑

(−)

2.81

↓

10.6

↓

221

1.85

↓↓

↓

↓↓

(−)

19.89

10.4

↓

398

18.14

↑↑

↑

↓

(−)

136

↓

2.00

Cell

↓

250

ANC x109/L

Bone marrow aspirate smear

↓

0.68

↓↓

1.50 ↓↓

↓↓

344

1.33

0.62

↓

↓↓ ↓↓↓

2.67

↓↓ ↓↓ ↓↓↓

↓↓↓ ↓↓↓ ↓↓↓

ANC: absolute neutrophil count; Cell: cellularity; Ery: erythroid; F: female; HB: hemoglobin; HPh: hemophagocytosis; M: male; MK: megakaryocyte; Mye: myeloid; N: no specific change; No: number; PLT: platelet; WBC: white blood cell; yr: year. Haematologica | 107 August 2022

1984

LETTER TO THE EDITOR

Authors

Shan-Chi Yu,1,2 Huai-Hsuan Huang,3 Chun-Nan Chen,4 Tseng-Cheng Chen4 and Tsung-Lin Yang4

Disclosures No conflicts of interest to disclose.

Department of Pathology and Graduate Institute of Pathology,

College of Medicine, National Taiwan University; 2Department of

Contributions

Pathology, National Taiwan University Hospital; Department of

S-CY designed the study, reviewed bone marrow trephine biopsies,

Internal Medicine, National Taiwan University Hospital; 4Department

and drafted the manuscript; H-HH reviewed the laboratory data and

of Otolaryngology, National Taiwan University Hospital, Taipei,

bone marrow aspirate smears and critically revised the manuscript;

Taiwan

C-NC, T-CC, T-LY collected clinical data.

Correspondence:

Funding

S-C. YU - b88401002@ntu.edu.tw

This study was supported by National Taiwan University Hospital

(grant 110-2) and Ministry of Science and Technology, Taiwan (grant https://doi.org/10.3324/haematol.2022.280746

110-2635-B-002-001).

Received: January 27, 2022.

Data-sharing statement

Accepted: April 15, 2022.

The data underlying this article will be shared on reasonable

Prepublished: April 28, 2022.

request to the corresponding author.

References 1. Perry AM, Choi SM. Kikuchi-Fujimoto disease: a review. Arch Pathol Lab Med. 2018;142(11):1341-1346. 2. Bosch X, Guilabert A, Miquel R, Campo E. Enigmatic KikuchiFujimoto disease: a comprehensive review. Am J Clin Pathol. 2004;122(1):141-152. 3. Yu SC, Chang KC, Wang H, et al. Distinguishing lupus lymphadenitis from Kikuchi disease based on clinicopathological features and C4d immunohistochemistry. Rheumatology (Oxford). 2021;60(3):1543-1552. 4. Song JY, Lee J, Park DW, et al. Clinical outcome and predictive factors of recurrence among patients with Kikuchi's disease. Int J Infect Dis. 2009;13(3):322-326. 5. Cheng CY, Sheng WH, Lo YC, Chung CS, Chen YC, Chang SC. Clinical presentations, laboratory results and outcomes of patients with Kikuchi's disease: emphasis on the association between recurrent Kikuchi's disease and autoimmune diseases. J Microbiol Immunol Infect. 2010;43(5):366-371. 6. Kuo TT. Kikuchi's disease (histiocytic necrotizing lymphadenitis). A clinicopathologic study of 79 cases with an analysis of histologic subtypes, immunohistology, and DNA ploidy. Am J Surg Pathol. 1995;19(7):798-809. 7. Han HJ, Lim GY, Yeo DM, Chung NG. Kikuchi's disease in children: clinical manifestations and imaging features. J Korean Med Sci. 2009;24(6):1105-1109. 8. Chuang CH, Yan DC, Chiu CH, et al. Clinical and laboratory manifestations of Kikuchi's disease in children and differences

between patients with and without prolonged fever. Pediatr Infect Dis J. 2005;24(6):551-554. 9. Fu JF, Wang CL, Liang L, Dayan C, Dong GP, Hong F. KikuchiFujimoto disease manifesting as recurrent thrombocytopenia and Mobitz type II atrioventricular block in a 7-year-old girl: a case report and analysis of 138 Chinese childhood Kikuchi-Fujimoto cases with 10 years of follow-up in 97 patients. Acta Paediatr. 2007;96(12):1844-1847. 10. Kucukardali Y, Solmazgul E, Kunter E, Oncul O, Yildirim S, Kaplan M. Kikuchi-Fujimoto disease: analysis of 244 cases. Clin Rheumatol. 2007;26(1):50-54. 11. Jung IY, Ann HW, Kim JJ, et al. The incidence and clinical characteristics by gender differences in patients with KikuchiFujimoto disease. Medicine (Baltimore). 2017;96(11):e6332. 12. Dumas G, Prendki V, Haroche J, et al. Kikuchi-Fujimoto disease: retrospective study of 91 cases and review of the literature. Medicine (Baltimore). 2014;93(24):372-382. 13. Yang Y, Lian H, Ma H, et al. Hemophagocytic lymphohistiocytosis associated with histiocytic necrotizing lymphadenitis: a clinical study of 13 children and literature review. J Pediatr. 2021;229:267-274. 14. Yu SC, Cheng CL, Huang HH, et al. Bone marrow histology in hemophagocytic lymphohistiocytosis. Arch Pathol Lab Med. In press. 15. Florena AM, Iannitto E, Quintini G, Franco V. Bone marrow biopsy in hemophagocytic syndrome. Virchows Arch. 2002;441(4):335-344.

Haematologica | 107 August 2022

1985

COMMENT

Alloimmunization against Fy3 is a serious threat in the era of cell therapy In a recent article of Haematologica, Stone et al. described a case of a severe delayed hemolytic transfusion reaction due to an anti-Fy3 alloimmunization of a patient suffering from sickle cell disease and expecting gene therapy.1 This case report comes at the right time in the field of sickle cell disease management for two reasons: numerous protocols of cell therapy (i.e. gene therapy or allogeneic hematopoietic stem cell transplantation) are forthcoming and all advocate intensive transfusions before and during the procedure. It also reminds us of the real immunological transfusion issues related to the genetic distance between donors and recipients. We currently treat about 100 sickle cell patients in our center, and we have been particularly concerned about this complication of alloimmunization. Three young women episodically transfused for sickle cell disease developed anti-Fy3 alloimmunization, between 2017 and 2021. All of them were homozygous for the allele FY*02N.01 (GATA-1 mutation). Their red cell concentrates were selected according to the international guidelines.2,3 Consequently, before this alloimmunization, they received between four and 33 red cell concentrates, crossmatched and Rh (D, C/c, E/e), K and, if possible, S, s, Fya, Fyb, Jka, Jkb matched. In particular, special attention was paid to antigens of the FY protein and the majority of red cell concentrates were Fya negative. Interestingly, prior development of anti-Fya antibody was identified for only one of the three, as this antibody is considered a risk factor for developing anti-Fy3 by experts.1 Two of these patients had prior alloimmunizations directed against other red cell antigens: one patient had anti-S, anti-D, anti-C, anti-E, anti-Jka; another one had anti-M, anti-S, anti-Lea, anti-Fya, anti-Jkb, and anti-Doa. Due to the presence of the anti-Fy3 antibody, these two patients were in a transfusion deadlock in Switzerland. The three cases suggest that immunogenicity of Fy3 antigen might be more important than previously thought. An alloimmunization directed against such a common antigen among Caucasian donors is a matter of great concern in small countries like ours, because of the limited availability of rare blood products. Another interesting fact is that all of these antibodies directed against the Fy3 antigen were evanescent. This complicates the situation since this antibody is wellknown for an acute or delayed hemolytic transfusion reaction and, in a less severe manner, hemolytic disease

of the newborn.4 In our center, two of the three patients developed a severe and delayed hemolytic transfusion reaction. It is worth adding that anti-Fy3 alloantibody was still undetectable at the moment of these two transfusion reactions. This clinical presentation, without detection of the causative antibody just after the triggering transfusion, can occur in more than one third of cases, as shown by the multi-center study of Habibi et al. in 2016.5 In conclusion, the well-known discrepancy in FY protein antigens between donors and recipients must be seriously taken into account because it can definitively harm a promising project of cell therapy for young patients. Moreover, as suggested by the authors, cell therapy providers must be aware of the risk of alloimmunizations. Indeed, it increases inevitably with the number of transfusions.

Authors Baptiste Lemaire1,2 and Sophie Waldvogel Abramowski1,2 Department of Diagnostics, University Hospital of Geneva and

Department of Medicine, University Hospital of Geneva, Geneva,

Switzerland Correspondence: B. LEMAIRE - Baptiste.Lemaire@hcuge.ch https://doi.org/10.3324/haematol.2022.280632 Received: January 11, 2022. Accepted: January 19, 2022. Prepublished: March 3, 2022. ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

Disclosures No conflicts of interest to disclose. Contributions BL and SWA conbributed equally. Data-sharing statement All data can be obtained by email request to the corresponding author.

Haematologica | 107 August 2022

1986

COMMENT

References 1. Stone EF, Avecilla ST, Wuest DL, et al. Severe delayed hemolytic transfusion reaction due to anti-Fy3 in a patient with sickle cell disease undergoing red cell exchange prior to hematopoietic progenitor cell collection for gene therapy. Haematologica. 2021;106(1):310-312. 2. Chou ST, Alsawas M, Fasano RM, et al. American Society of Hematology 2020 guidelines for sickle cell disease: transfusion support. Blood Adv. 2020;4(2):327-355. 3. Pirenne F, Yazdanbakhsh K. How I safely transfuse patients with

sickle-cell disease and manage delayed hemolytic transfusion reactions. Blood. 2018;131(25):2773-2781. 4. Tormey CA, Hendrickson JE. Transfusion-related red blood cell alloantibodies: induction and consequences. Blood. 2019;133(17):1821-1830. 5. Habibi A, Mekontso-Dessap A, Guillaud C, et al. Delayed hemolytic transfusion reaction in adult sickle-cell disease: presentations, outcomes, and treatments of 99 referral center episodes. Am J Hematol. 2016;91(10):989-994.

Haematologica | 107 August 2022

1987

RESPONSE TO COMMENT

Intricacies of GATA-ca, continued We thank Lemaire et al. for continuing the call for attention to the issue of evanescence of red cell alloantibodies in sickle cell disease (SCD). Evanescence of anti-Fy3 is particularly relevant because approximately 68% of people of African descent have a Duffy red cell phenotype of Fy(a–b–) due to the GATA-1 mutation, which is otherwise present in < 1% of most ethnic populations. The risk of developing anti-Fy3 might be reduced by transfusing units that are Fy(a–), a phenotype present in only 34% of Caucasian but in 90% of Black donors. The Lemaire et al. cases demonstrate, however, that although production of anti-Fya is a risk factor for the development of anti-Fy3, it cannot necessarily be used as a predictive risk factor, since anti-Fy3 may develop in those in whom anti-Fya was never detected, possibly due to evanescence or timing of antibody screening. The development of multiple antibodies appears to be the greatest risk for making anti-Fy3. Review of unpublished cases from our laboratory over 10 years, and those in the literature (including the Lemaire et al. cohort) reveal a total of 80 cases of patients with anti-Fy3, of which 14 (18%) had only anti-Fy3 reactivity and 66 (83%) had additional red cell alloantibodies; 32 (48%) had no current anti-Fya detected. One might suggest that Fy(a–b–) red cells should be used for transfusion of all Fy(a–b–) patients, however this limits the transfusion options to units from minority donors, always in short supply, and G6PD deficiency, which is associated with decreased red cell survival, may be present in up to 20% of such units.1,2 Unfortunately, it remains unexplained why some patients make anti-Fy3, as these patients have the GATA mutation associated with absence of risk for anti-Fyb due to loss of expression on red cells, but Fy protein would be expressed in tissues. No genetic difference has been found in the coding region of the gene to explain production of anti-Fy3 (our unpublished observations). The cases of Lemaire et al. also bring attention to measures not yet fully implemented that may improve blood availability and transfusion-related outcomes in patients with SCD. Should SCD patients undergo antibody screening approximately 1 month following transfusion in order to detect all clinically relevant alloantibodies (including anti-Fy3) that may evanesce? Can a universally accessible red cell alloantibody patient registry that maintains privacy compliance be estab-

lished? Are there alternatives to transfusion for gene therapy support? What can be done to increase recruitment of minority donors, particularly donors of Black African descent? Do blood centers need to find a way to have extended antigen typing on all donors? In order to better preserve transfusion service budget allotment for antigen matching in SCD, can we reduce the use of phenotypically matched units in lower-risk patient populations, such as those on monoclonal antibody treatments like daratumumab?3,4 Also key is the recruitment and training of technical staff to recognize complex alloantibodies in patients with SCD. We hope for more progress to be made in these areas for our patients undergoing transfusion for SCD in the years ahead.

Authors Christine Lomas-Francis,1 Elizabeth F. Stone,2 Connie M. Westhoff1 and Patricia A. Shi1,3 New York Blood Center, New York; 2Division of Transfusion Medicine,

Columbia University Medical Center, New York and 3Sickle Cell Program, Division of Hematology, Albert Einstein College of Medicine, Bronx, NY, USA Correspondence: P.A. SHI - pshi@nybc.org https://doi.org/10.3324/haematol.2022.280876 Received: February 17, 2022. Accepted: February 21, 2022. Prepublished: March 3, 2022. ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

Disclosures No conflicts of interest to disclose. Contributions CLF collated immunohematology studies; all authors contributed to writing the manuscript.

References 1. Yee ME, Francis RO, Luban NLC, et al. Glucose-6-phosphate dehydrogenase deficiency is more prevalent in Duffy-null red blood cell transfusion in sickle cell disease. Transfusion. 2022;62(3):551-555. 2. Roubinian NH, Reese SE, Qiao H, et al. Donor genetic and nongenetic factors affecting red blood cell transfusion effectiveness. JCI Insight. 2022;7(1):e152598.

3. Bullock T, Foster A, Clinkard B. Alloimmunisation rate of patients on daratumumab: a retrospective cohort study of patients in England. Transfus Med. 2021;31(6):474-480. 4. Tauscher C, Moldenhauer S, Bryant S, DiGuardo M, Jacob EK. Antibody incidence and red blood cell transfusions in patients on daratumumab. Transfusion. 2021;61(12):3468-3472.

Haematologica | 107 August 2022

1988

CASE REPORT

Occurrence of a paroxysmal nocturnal hemoglobinuria clone in an essential thrombocythemia: a link between PIGV and MPL Paroxysmal nocturnal hemoglobinuria (PNH) is a rare acquired hemopathy (about 1.3 individuals per million incidence) characterized by hemolytic anemia and venous thrombosis.1 The molecular defect involved is glycosylphosphatidylinositol (GPI) anchor loss related to X-linked PIGA gene mutations. Other genes responsible for the GPI anchor biosynthesis pathway could also be involved,2–5 and, research on their involvement in the pathophysiogenesis of PNH and its borderline forms is being uncovered. Here, we report a case of essential thrombocythemia (ET) caused by a somatic mutation in MPL (c.1544G>T:p.W515L) shortly preceded by copy-neutral loss of heterozygosity (CN-LOH) in cis on chromosome 1, leading to homozygosity of both the acquired MPL missense mutation and an inherited heterozygous stop-gain mutation in PIGV (c.1405C>T:p.R469X) causing PNH. Clinical presentation In July 2017, a 73-year-old Caucasian patient was referred to the cardiology intensive care untit due to pulmonary embolism associated with cardiogenic shock. Thrombolysis was performed, followed by anticoagulant therapy. Active neoplasm and the antiphospholipid syndrome were excluded. The persistence of thrombocytosis after 3 months (platelets >600 109/L) without any sign of infection or iron deficiency, justified a myelogram revealing ET with a clone size of 74% (Figure 1). Concomitantly, glomerulopathy was identified (urinary protein to creatinine ratio [uPCr] 5.3 g/g; serum albumin 33 g/L; 39 urinary red cells/mL). Renal biopsy showed interstitial fibrosis with focal segmental glomerulosclerosis (FSGS). Causes of secondary FSGS were ruled out. Increased level of the nephrotic syndrome-associated soluble urokinase-type plasminogen activated receptor (suPAR) was identified (731 pg/mL). Blood tests showed hemolysis (lactate dehydrogenase 1,270 IU/L, undetectable haptoglobin) without schistocytes or positive Coombs test (Figure 1). PNH was suspected then confirmed by flow cytometry (80.4% CD14- CD55FLAER- monocytes and 88.8% CD24- CD16- CD55- FLAER- granulocytes) with an 88% clone size at Q4 2017 (Figure 1). Thus, this patient developed a massive pulmonary embolism, with two underlying clonal myeloid disorders (ET and PNH) and a nephrotic syndrome. After 24 months of treatment, combining hydroxycarbamide (6,500 mg/week) and the C5 complement inhibitor eculizumab (900 mg/2 weeks), the clones size dramatically decreased, cell counts normalized, hemolysis stopped, and uPCr decreased below 2 g/g

Figure 1. Follow-up over time of biochemical (LDH, creatininemia, proteinuria [urinary protein to creatinine ratio, uPCr], soluble urokinase-type plasminogen receptor [uPAR]) and hematological paroxysmal nocturnal hemoglobinuria (PNH) clone (flow cytometry), variant allele frequency (VAF) of the essential thrombocythemia (ET) clone, platelets (PLT), and leukocytes (white blood cells, WBC) parameters. Eculizumab treatment was first started in quarter 4 (Q4) of 2017, leading to a decrease in LDH Hydroxycarbamide was administered in Q1 of 2018. Apart from creatininemia, which did not vary, all other parameters followed the evolution of the PNH/ET clone undergoing cytoreductive treatment. The concentration of serum suPAR was assessed using the human suPAR enzyme-linked immunosorbant assay kit (Fine Test, EMELCA Biosciences) following the manufacturer’s instructions.

Haematologica | 107 August 2022

1989

CASE REPORT

Figure 2. About 2 years of essential thrombocythemia/paroxysmal nocturnal hemoglobinuria (ET/PNH) clone follow-up. (A) Fish diagram representation of mutation events in hematopoietic stem and progenitor cells (HSPC) and clonal expansion of the essential thrombocythemia/paroxysmal nocturnal hemoglobinuria (ET/PNH) clone over time. Heterozygous mutations are depicted as colored triangles in squares, homozygous mutations as colored squares: PIGV p.R469X red, MPL p.W515L blue, DNMT3A p.R882H purple; the GPI-anchor is depicted as green circles. ET and PNH clone size and timeline of mutation events were determined by variant allele frequency (VAF) of MPL p.W515L and DNMT3A p.R882H mutations in FLAER-positive and FLAER-negative sorted cells collected in quarter 1 (Q1) of 2018. Upon hydroxycarbamide treatment, the PNH/ET clone reduced in size reaching a low of 21% in Q4 of 2018. For cell sorting analysis, whole blood was labeled with FLAER-Alexa 488; after erythrocyte lysis, cells were resuspended in phosphate-buffered saline (PBS) and sorted with BD InfluxTM cell sorter (BD Biosciences, Software software) (pressure 200 PSI, nozzle 100 µm) according to the gating strategy. (B) Flow cytometry follow up at Q4 of 2017, Q4 of 2018, and Q1 of 2020 of the PNH/ET clone, monitoring of neutrophil and monocyte populations. After counting, white blood cells (WBC) were incubated with pretitrated FLAER-Alexa 488 (Cedarlane Laboratories), CD16-PE (clone 3G8, Beckman Coulter), CD33-PC5.5 (clone D3HL60.251, Beckman Coulter), CD55-PE-Cy7 (clone JS11, Biolegend), CD14-APC (clone M5E2, Biolegend), CD24-APC-Vio770 (clone REA832, Miltenyi), CD15-V450 (clone HI98, BD Biosciences) and CD45-BV510 (clone HI30, BD Biosciences). After erythrocyte lysis (BD FACS Lysing Solution, BD Biosciences), cells were analyzed (Navios flow cytometer, Beckman Coulter, >200,000 cells/analysis). Percentages of FLAER- and GPI-anchored protein expressions (CD14, CD55 on monocytes [mono] and CD24, CD16, CD55 on polymorphonuclear neutrophils [PMN]) were analyzed using Kaluza software (monocytes gated as side scatterlow, CD45bright, and CD33bright events; PMN gated as side scatterhigh, CD45mid, and CD33mid events). Haematologica | 107 August 2022

1990

CASE REPORT

Figure 3. Illustration of the different molecular genetic explorations of the identified mutations. (A) Integrative genomic view of 41 bp windows showing read coverage of aligned short reads from whole exome sequencing (WES) on DNA from quarter 1 (Q1) 2018 (paroxysmal nocturnal hemoglobinuria [PNH] clone sized of 94%). Varainat allele frequency (VAF) is depicted in extra window for PIGV (R469X; chr1: 27.124.258C>T), MPL (W515L; chr1:43.815.009G>T), and DNMT3A (R882H; chr2:25.457.242C>T), hg19 reference sequence and amino acid sequence at the bottom. WES was performed using a SureSelect Human All Exon XT V6 (Agilent) enrichment kit on DNA extracted from blood samples obtained on Q1 of 2018 (PNH clone size 94%) followed by sequencing on a HiSeq 2500 (Illumina). Reads were mapped to GRCh37 and variants were called according to GATK14 best practice guidelines. Variants were filtered and assessed based on ACMG guidelines. All mutations were confirmed by Sanger sequencing. (B) Sanger sequencing trays of PIGV gene on DNA from saliva, and DNA from two blood samples (PNH clone size 94% and 57% at Q1 and Q3 of 2018, respectively). VAF is depicted by reduced chromatogram intensity of the C allele on position c.1405. (C) HomSI15 analysis of WES data on DNA from Q1 of 2018 (PNH clone size of 94%). From position 0 to 105,000,000 homozygous variants followed consecutively within a window size of 6 Mbp. Copy number abnormalities and copy-neutral loss of heterozygosity (CN-LOH) on the p-arm of chromosome 1. CN-LOH was confirmed by chromosomal microarray (CMA) according to manufacturer instructions using a SurePrint G3 Human CGH 4x180K oligonucleotide array (Agilent). The estimated average distance between euchromatic oligonucleotide markers was 16.1 kb. Copy number variation (CNV) calling was performed using Cytogenomics software (vs 4.0.3.12, Agilent) using aberration algorithm ADM-2. The mosaic aberration filter was used with the following settings for deletions/losses: minimum size (kb) of the region for deletion 0.0, the minimum number of probes for deletion ≥5, minimum absolute average Log-ratio for deletion ≥0.15. Haematologica | 107 August 2022

1991

CASE REPORT (Figure 1). He has subsequently been seen twice at 6-month autoinflammation in this patient, contrary to the reports intervals by teleconsultation due to the pandemic. All his of PIGT and PIGB-PNH.4,5 Thus, we speculate that a GPI hematological, cytological, and biochemical parameters precursor with only one mannosyl group and lacking were stable and proteinuria remained constant at <1 g/g ethanolamine residues does not trigger auto-activation of (data not shown). the inflammasome. The ET-type myeloproliferative syndrome due to the MPL (p.W515L) mutation gave the proResults and discussion liferative advantage to the PNH clone within a mixed In order to identify the molecular cause underlying the pa- ET-PNH clone that explains the response to cytoreductive tient’s condition targeted sequencing was performed on treatment and retrogression of PNH in the patient. InterDNA extracted in Q4 2017 (clone size 74%) and mutations in estingly, a correlation between the clone size, the suPAR MPL (c.1544G>T:p.W515L) and DNMT3A (c.2645G>A:p.R882H) serum levels, the renal disease course, and the adminiswere discovered with a variant allele frequency (VAF) of 59% tration of hydroxycarbamide was observed. Although and 29%, respectively. As mutations in PIGA were not found, nephrotic syndrome and FSGS are not common manifeswhole exome sequencing (WES) was performed on DNA tations of PNH and ET, excessive concentrations of suPAR from blood extracted in Q1 2018 revealing the cause of the secreted by myeloid cells were associated with FSGS and increased PNH clone (size 94%): a mutation in PIGV kidney disease in human and animal models.8–10 Physio(c.1405C>T:p.R469X) with a VAF of 93%; concomitantly, the logically, the binding of uPAR to podocyte αvb3 integrins VAF of MPL (p.W515L) and DNMT3A (p.R882H) increased to ensures the structural stability of the podocyte cyto94% and 41% (Figure 2). Subsequent sequencing of FLAER skeleton and its filtration properties.11 We hypothesized sorted cells isolated at Q1 2018 revealed a VAF of the MPL that the massive expansion of the GPI-deficient myeloid (p.W515L) mutations of 86% and DNMT3A (p.R882H) 45% in clones may have led to the production of the high conFLAER-negative cells, while the MPL mutation p.W515L was centration of suPAR, which in turn may have competed absent in FLAER-positive cells, DNMT3A p.R882H had a VAF with its functional GPI-anchored form, uPAR, affecting the of 10% (Figure 2, VAF box). podocyte integrity in the glomeruli, potentially causing Congenital heterozygosity of PIGV p.R469X was confirmed FSGS.12 by Sanger sequencing of DNA from saliva. Homozygosity Although PIGV functions downstream of PIGA in the GPI of p.R469X in PIGV and p.W515L in MPL was due to a CN- anchor synthesis pathway,13 such a clinical course with LOH in cis of almost the entire chromosome 1 p-arm in major thrombotic events, hemolysis, and kidney failure has cells derived from the mutated GPI-deficient hemato- neither been described in a patient nor has it been aspoietic stem and progenitor cells (HSPC), as confirmed by sociated with a GPI anchor deficiency before. exome data and chromosomal microarray analyses (CMA) In summary, we described the first case of PIGV-PNH that (Figure 3). emerged with an ET clone. We concluded that driving muBased on the VAF of the identified mutations in cells tations of the myeloproliferative syndrome in combination sorted by FLAER, we can conclude on the temporal oc- with mutations in genes of the GPI anchor pathway may currence of the observed somatic mutations (Figure 2). lead to a mixed ET-PNH clone. Finally, we showed that First, an HSPC acquired a somatic mutation in DNMT3A cytoreductive treatment in combination with administra(p.R882H) on one allele of chromosome 2, possibly a long tion of eculizumab led to improvement of renal and hematime ago, in a stable manner and without malignant con- tological outcomes. sequences, as discussed by Sun and Babushok.6 Hypomethylation of DNA due to mutated DNMT3A7 probably Authors facilitated further somatic events in this HSPC: i) homozygosity of p.R469X in PIGV giving rise to GPI-defiAlexej Knaus,1* François Vergez,2* Cédric Garcia,3* Hartmut Engels,4 cient cells (deriving from about 14% of GPI-deficient cells Hela Hundertmark,4 David Ribes,5 Laetitia Largeaud,2 Suzanne without MPL p.W515L mutation), ii) acquisition of the Tavitian,6 Bernard Payrastre,3 Peter Krawitz,1 Stanislas Faguer5 and p.W515L mutation in MPL and iii) finally, the triple mutant Agnes Ribes3 HSPC with heterozygous mutations in DNTM3A (p.R882H), PIGV (p.R469X ), and MPL (p.W515L) acquired CN-LOH of 1Institute for Genomic Statistics and Bioinformatics, University of almost the entire p-arm on chromosome 1, causing Bonn, School of Medicine and University Hospital of Bonn, Bonn, homozygosity of the PIGV (p.R469X) and MPL (p.W515L) Germany.; 2Department of Biological Hematology, Toulousemutations leading to GPI deficiency. Immune evasion, Oncopole University Cancer Institute, Toulouse, France; clonal expansion, and dominance of the double mutant 3Hematology Laboratory, Toulouse University Hospital Center, and HSPCs led to the emergence of PIGV-PNH and ET. Inter- National Referral Center for Platelets Diseases, INSERM U1297, and estingly, hemolysis, thrombosis, and the glomerular syn- Paul Sabatier University, Institute of Cardiovascular and Metabolic drome were observed without any signs of Diseases, Toulouse, France; 4Institute of Human Genetics, University Haematologica | 107 August 2022

1992

CASE REPORT

of Bonn, School of Medicine and University Hospital of Bonn, Bonn, Germany; 5Department of Nephrology and Organ Transplantation, Toulouse University Hospital Center, and Referral Center for Rare Renal Diseases, Toulouse University Hospital Center, Toulouse, France and 6Hematology Department, Toulouse-Oncopole University Cancer Institute, Toulouse, France. *AK, FV and CG contributed equally as co-first authors. Correspondence: A. RIBES - ribes.a@chu-toulouse.fr

Disclosures No conflicts of interest to disclose. Contributions AK performed sequencing studies and analyzed the data; FV performed the cytometry analysis for PNH clone and analyzed the data; DR and ST provided patient samples and characterized the patient; CG and AR performed research and analyzed the data; AR designed the study, HE, HH, BP and PK participated in the discussion; AK, FV, SF, and AR wrote the paper. Acknowledgments The authors would like to thank Alexia Zakaroff-Girard and Elodie Riant for helpful assistance in their field of expertise, Cytometry Platform, I2MC, and the Institute of Metabolic and Cardiovascular Diseases, Toulouse, France. The authors would also like to thank Dr. Sophie Kaltenbach (Hôpital Necker-Enfants Malades, Paris, France) and Prof. Dominique Helley (Hôpital Européen Georges Pompidou, Paris, France) for their technical assistance on the molecular biology of PIGA as well as for their expert advice and gracious availability.

https://doi.org/10.3324/haematol.2021.279804 Received: August 20, 2021. Accepted: January 12, 2022. Pre-published: January 27, 2022. ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

References 1. Brodsky RA. Paroxysmal nocturnal hemoglobinuria. Blood. 2014;124(18):2804-2811. 2. Höchsmann B, Murakami Y, Osato M, et al. Complement and inflammasome overactivation mediates paroxysmal nocturnal hemoglobinuria with autoinflammation. J Clin Invest. 2019;129(12):5123-5136. 3. Krawitz PM, Höchsmann B, Murakami Y, et al. A case of paroxysmal nocturnal hemoglobinuria caused by a germline mutation and a somatic mutation in PIGT. Blood. 2013;122(7):1312-1315. 4. Kawamoto M, Murakami Y, Kinoshita T, Kohara N. Recurrent aseptic meningitis with PIGT mutations: a novel pathogenesis of recurrent meningitis successfully treated by eculizumab. BMJ Case Rep. 2018;2018:bcr2018225910. 5. Langemeijer S, Schaap C, Preijers F, et al. Paroxysmal nocturnal hemoglobinuria caused by CN-LOH of constitutional PIGB mutation and 70-kbp microdeletion on 15q. Blood Adv. 2020;4(22):5755-5761. 6. Sun L, Babushok DV. Secondary myelodysplastic syndrome and leukemia in acquired aplastic anemia and paroxysmal nocturnal hemoglobinuria. Blood. 2020;136(1):36-49. 7. Liu X, Kamatani Y, Terao C. Genetics of autosomal mosaic chromosomal alteration (mCA). J Hum Genet. 2021;66(9):879-885. 8. Hayek SS, Leaf DE, Samman Tahhan A, et al. Soluble Urokinase

receptor and acute kidney injury. N Engl J Med. 2020;382(5):416-426. 9. Shankland SJ, Pollak MR. A suPAR circulating factor causes kidney disease. Nat Med. 2011;17(8):926-927. 10. Wei C, El Hindi S, Li J, et al. Circulating urokinase receptor as a cause of focal segmental glomerulosclerosis. Nat Med. 2011;17(8):952-960. 11. Wei C, Möller CC, Altintas MM, et al. Modification of kidney barrier function by the urokinase receptor. Nat Med. 2008;14(1):55-63. 12. Sloand EM, Pfannes L, Scheinberg P, et al. Increased soluble urokinase plasminogen activator receptor (suPAR) is associated with thrombosis and inhibition of plasmin generation in paroxysmal nocturnal hemoglobinuria (PNH) patients. Exp Hematol. 2008;36(12):1616-1624. 13. Kang JY, Hong Y, Ashida H, et al. PIG-V involved in transferring the second mannose in glycosylphosphatidylinositol. J Biol Chem. 2005;280(10):9489-9497. 14. McKenna A, Hanna M, Banks E, et al. The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res. 2010;20(9):1297-1303. 15. Görmez Z, Bakir-Gungor B, Sagiroglu MS. HomSI: a homozygous stretch identifier from next-generation sequencing data. Bioinformatics. 2014;30(3):445-447.

Haematologica | 107 August 2022

1993

CASE REPORT

Functional testing of relapsed chronic lymphocytic leukemia guides precision medicine and maps response and resistance mechanisms. An index case Relapsed chronic lymphocytic leukemia (CLL) after sequential treatment with targeted therapies has a dismal prognosis and represents an increasing unmet medical need.1,2 Immunotherapies such as chimeric antigen receptor T-cell (CAR-T) therapy or bispecific antibodies may be efficacious in this setting, but are not readily available to patients outside clinical trials.3 Direct drug testing on tumor cells can indicate treatment vulnerabilities,4 and implementation of this approach in treatment decisions for aggressive refractory hematological malignancies led to improved treatment.5 Elucidation of treatment sensitivities in multi-drug refractory CLL may thus inform novel therapeutic concepts for this patient group. Indeed, our demonstration of ex vivo sensitivity to proteasome inhibition provided the basis for the use of off-label ixazomib citrate in an index case of relapsed CLL after treatment with ibrutinib, idelalisib, alemtuzumab, and venetoclax/rituximab. We report a high-resolution cellular and functional analysis using mass cytometry, flow cytometry, ex vivo killing assays, and drug sensitivity testing on peripheral blood mononuclear cells (PBMC) collected from the index patient at seven time points before, during, and after treatments. Our findings may indicate the molecular and cellular determinants of the treatment-responding and non-responding states of the disease and highlight the clinical value of direct drug testing to identify effective, personalized therapies for relapsed CLL. Written informed consent was obtained before sample collection. The study was approved by the Regional Committee for Medical and Health Research Ethics of SouthEast Norway. The index patient was diagnosed with CLL at the age of 70. The disease presented with unmutated IGVH, mutated TP53 and homozygous del(13q14). His treatment history is presented in Figure 1A. The patient was intolerant to ibrutinib and idelalisib, then received treatment with alemtuzumab experiencing stable disease (Figure 1A). Upon disease progression, CLL was treated with venetoclax/rituximab, and the patient obtained complete remission (CR) with undetectable minimal residual disease (uMRD) (Figure 1A). At this point, the therapy was stopped (Figure 1A). After almost 2.5 years off therapy, the disease relapsed with severe bone marrow failure. Retreatment with venetoclax failed (Figure 1A). Serial peripheral blood samples were collected from the patient (Figure 1A). PBMC collected at T1 (after ibrutinib

and idelalisib) and T6 (after venetoclax retreatment) (Figure 1A), and from three treatment-naïve CLL patients, were subjected to direct drug sensitivity screening with 93 single agents at five concentrations.6 The drug sensitivity score (DSS) was calculated based on the area under the concentration-response curve (Figure 1B).7 The drug testing confirmed statistically significant reduced drug sensitivity at T6 relative to T1 and treatment-naïve CLL (P<0.0001 using 2-way ANOVA with Dunnett’s multiple comparisons test). The concentration-response curves for venetoclax are shown in Figure 1C. The solid, vertical line indicates the in vitro venetoclax concentration (2,000 nM) which corresponds to the peak plasma concentration (1.75 µg/mL) obtained when venetoclax is administered at 400 mg/day.8 As shown, venetoclax was effective at clinically achievable concentrations in PBMC collected at T1, while the sensitivity was lower in PBMC collected at T6 (Figure 1D). Indeed, the DSS was reduced with 67% at T6 relative to T1. Interestingly, the CLL cells collected at T6 remained highly sensitive to the proteasome inhibitors bortezomib and ixazomib citrate (Figure 1B and C). Ixazomib citrate is an orally administered second-generation proteasome inhibitor approved for treatment of multiple myeloma. Preclinical effects on CLL have been observed,9-11 and phase I/II trials in non-Hodgkin lymphoma are active (www.clinicaltrials.gov). Ixazomib citrate is administered at 4 mg/day, which gives a maximum observed plasma concentration of 65.3 ng/mL,12 corresponding to an in vitro concentration of about 130 nM (Figure 1C; dashed, vertical line). The patient was started on treatment with 4 mg ixazomib citrate on day 1 of each 7-day cycle combined with 20 mg (instead of 40 mg to reduce side effects) dexamethasone on days 1 and 2 of each 7-day cycle. This is according to the summary of product characteristics (SPC) for ixazomib citrate and the approved dosing according to the European Medicines Agency. The patient was transfusiondependent with very severe thrombocytopenia (<10x109/L, Figure 1D) which made an oral proteasome inhibitor preferable. Therapy resulted in increased numbers of reticulocytes, thrombocytes, and hemoglobin, indicating that the therapy was effective (Figure 1D). The bone marrow response to treatment (T7) is illustrated in Figure 1E. At present, >120 days after treatment initiation, the patient is transfusion-independent with no bleedings, he is able to exercise and has an active life.

Haematologica | 107 August 2022

1994

CASE REPORT

Figure 1. Ixazomib citrate is effective against chronic lymphocytic leukemia ex vivo and in vivo. (A) Graphical illustration of the index patient’s treatment history, shown as years since diagnosis. Peripheral blood mononuclear cells (PBMC) were collected at the indicated time points T1-T7. The figure was created with BioRender.com. (B) PBMC collected from 3 treatment-naïve chronic lymphocytic leukemia (CLL) patients, as well as at T1 and T6 from the index patient, were co-cultured with CD40L+, BAFF+, and APRIL+ L cells (ratio 1:1:1) for 24 hours (h) prior to initiation of the experiment to mimic the tumor microenvironment. The L cells were then removed and the CLL cells were treated with the indicated 93 single agents at 5 different concentrations (1–10,000 nM) for 72 h. Cell viability was assessed with the CellTiter-Glo luminescent assay. The response readouts were normalized to the negative (0.1% DMSO) and positive (100 µM benzethonium chloride) controls. The heatmap was created using ClustVis (https://biit.cs.us.ee/clustvis/) and illustrates the calculated DSS on a scale from 0-100 (see key, right). Rows are clustered using Manhattan distance and Ward (unsquared distances) linkage. (C) Relative cell viability of PBMC collected at T1 or T6 in response to venetoclax, bortezomib or ixazomib citrate exposure. The experiment is described in (B). The solid, vertical line indicates the maximum plasma concentration of venetoclax reported for patients treated with 400 mg/day. The dashed, vertical line indicates the maximum plasma concentration of ixazomib citrate for patients treated with 4 mg/day. (D) Blood counts of the index patient in response to treatment with ixazomib citrate + dexamethasone. (E) Anti-Pax-5- and hematoxylin and eosin (H&E)-stained sections of bone marrow from T6 and T7, at 2x magnification. Pax-5-positive cells are shown in brown. CLL cells were reduced from 90% (T6) to 65% (T7) of the bone marrow cellularity. AE: adverse event; CR: complete remission; NR: no response; SD: stable disease; UM-CLL: IGVH unmutated CLL; uMRD: undetectable minimal residual disease. Haematologica | 107 August 2022

1995

CASE REPORT In order to map the cellular responses to venetoclax treatment, PBMC collected at T1-T3 (Figure 1A) were subjected to immune cell phenotyping by single-cell mass cytometry (Figure 2A). As expected, the IgM+CD19+ CLL Bcell population dominated at T1-T2 (Figure 2A). However, after 7 months on venetoclax, this population was almost eradicated (T3, Figure 2A). This demonstrated the efficacy of venetoclax and aligned with the subsequently achieved CR with uMRD (Figure 1A). Interestingly, as a result of treatment, the CD3+ T-cell and CD14+ monocyte populations were restored (Figure 2A). Reshaping of the immune cell composition in response to venetoclax treatment has been reported.13 Of particular interest, we also observed that the CD56+

natural killer (NK) cell population had significantly expanded at T3 (41% of PBMC at T3 vs. 0.4% at T1, Figure 2A). This population included the standard CD56dimCD16hi cells and expanded, immature CD56brightCD16low cells. Both cell types were similarly activated (HLA-DR+, Ki67+, Figure 2B). NK cell numbers continued to increase after venetoclax therapy was held, and were still high (>20%) more than 1 year prior to disease relapse (T5, Figure 2C), but had fallen below 10% when the disease was progressing (T6, Figure 2C). The expanded NK cells (T3-T5) contained a large fraction of CD56bright cells (34-54%) (Figure 2D), possibly reflecting immature cells recruited from the bone marrow. We next investigated the lytic activity of purified, bulk NK cells at T4-T7 against autologous CLL or K562 cells. NK

Figure 2. Evolution of natural killer cells in response to venetoclax and ixazomib citrate treatments. (A) Opt-SNE representation of immune cells (CyTOF) based on protein expression of lineage-defining surface markers in peripheral blood mononuclear cells (PBMC) collected at T1-T3 and from a healthy age-matched blood donor (HD). The top row shows density of cells. The numbers indicate percentage of cells in the respective gate. The bottom row shows back-gated chronic lymphocytic leukemia (CLL) cells and naïve B cells (IgM+CD19+), T cells (CD3+, including natural killer [NK] T cells CD3+CD56+, lower regions), monocytes (CD14+), and NK cells (CD56+) (see color key). (B) Gated NK cells shown in overlayed contour plots for a healthy age-matched blood donor (HD; black) and CLL T3 (blue). CD56 vs. CD16 (upper panel) and Ki67 vs. HLA-DR (lower panel) are shown. Quadrant percentages are shown in black (HD) and blue (T3) font. (C) Frequencies of CD3-CD56+ NK cells among CD19- lymphocytes at T1-T7, detected by flow cytometry. (D) Frequencies of CD56bright and CD56dim NK cell subsets within CD3-CD56+ NK cells at T1-T7, detected by flow cytometry. (E) Specific lysis of K562 cells or CLL blasts (from T1 or T6) by CD56+-positively selected NK cells isolated at indicated time points or from a healthy donor (HD). Cell death was monitored as PI+ events within CFSE prestained CLL blasts. (F) Frequencies of indicated subsets within CD3-CD56dim NK cells at T1-T7, detected by flow cytometry. Haematologica | 107 August 2022

1996

CASE REPORT cells from T5 exerted increased cytotoxicity against the autologous T1 CLL cells, relative to NK cells from T4 (Figure 2E). Killing of K562 cells was lowest at T6, suggesting reversal of the augmented NK cell-mediated cytotoxicity at relapse (Figure 2E). Cytotoxicity was higher at T7, but no activity against T6 CLL cells was observed (Figure 2E). Further analysis showed a temporal decreased expression of CD16 and an increased expression of the activation marker CD69 on CD56dim NK cells between T3-T5 (Figure 2F). This matched the expression of the exhaustion marker TIGIT that initially was high, then normalized, except for a temporal increase at T6 (Figure 2F). Further, a compensatory temporal increase in the less mature NKG2A+CD57- subset was observed after venetoclax treatment (T3-T5), with concomitant reduction of terminally differentiated CD57+ NK cells (Figure 2F). Collectively, these data show an activated NK cell compartment with signs of exhaustion and enhanced killing efficacy after venetoclax treatment (>T3) with normalization at relapse (T6) in this patient. In order to further evaluate the mechanism of venetoclax response and resistance in the index patient and of ixazomib citrate response, we next profiled the expression and activation status of 30 intracellular proteins in the serial CLL samples14 (Figure 3A). Interestingly, PBMC col-

lected when the patient was responding to either venetoclax or ixazomib citrate (T4 and T7, respectively), showed a more similar profile than PBMC collected when the patient had active disease (T1 and T6; Figure 3A). Notably, expression of BCL-2 was significantly lower at T4 than at the other time points (Figure 3B), while Bim dropped at T6 (Figure 3B). Proteins downstream of the Bcell receptor, including BTK, MEK1, and S6-ribosomal protein, displayed enhanced phosphorylation levels at time of relapse (T6 vs. T4; Figure 3A and B). Upregulation of the MEK pathway in combination with decreased Bim has also previously been associated with drug resistance in CLL.15 In general, treatment with ixazomib citrate restored the protein expression and activation levels to a similar level as when the patient was in remission after venetoclax treatment (T7 vs. T4; Figure 3B). Taken together, our study provides mechanistic insight to clinical response and resistance to targeted therapies (Figure 3C), as well as proof-of-concept for direct drug testing as a method to guide effective personalized therapy for relapsed CLL. Since drug sensitivity screens can be performed and analyzed in only 5 days, it is possible that this method can be used as a companion diagnostic for CLL patients in need of therapy. Clinical trials are needed to test this approach to functional precision medicine.

Figure 3. Evolution of protein profile in response to venetoclax and ixazomib citrate treatments. (A) Peripheral blood mononuclear cells (PBMC) collected at T1, T4, T6, and T7 were fixed, permeabilized and stained with the indicated antibodies. Signals in CD19+ cells were analyzed by flow cytometry. Raw data were analyzed in Cytobank (https:/cellmass.cytobank.org/cytobank/) and transformed to an arcsinh ratio relative to the signal in isotype control-stained cells, which was set to zero. The heatmap was created using ClustVis (https://biit.cs.us.ee/clustvis/). Both rows and columns are clustered using Manhattan distance and Ward linkage. (B) Protein expression and phosphorylation levels detected in (A). (C) Graphical summary of some of the cellular characteristics of the PBMC collected at T1-T7. The size of the symbols reflects the relative detected level at each sampling time. Refer to Figure 1A for details regarding treatment history. n.e: not established. The figure was created with BioRender.com. Haematologica | 107 August 2022

1997

CASE REPORT

Authors

Disclosures No conflicts of interest to disclose.

Sigrid S. Skånland, Marit Inngjerdingen, Henrik Bendiksen, Jamie York,2,4 Signe Spetalen,5 Ludvig A. Munthe2,4 and Geir E. Tjønnfjord2,6 1,2

1,2

Contributions SSS designed the research with MI and LAM; SSS, MI, HB and JY performed experiments and analyzed data with LAM; SS

Department of Cancer Immunology, Institute for Cancer Research,

performed bone marrow histopathology; GET contributed with

Oslo University Hospital; K. G. Jebsen Center for B-Cell

patient samples and provided clinical care; SSS wrote the

Malignancies, Institute of Clinical Medicine, University of Oslo;

manuscript. All authors read and commented on draft versions of

the manuscript and approved the final version.

Department of Pharmacology, Institute of Clinical Medicine,

University of Oslo; Department of Immunology, Oslo University 4

Acknowledgements

Hospital; 5Department of Pathology, Oslo University Hospital and 6

We are thankful to all study participants. We thank the High

Department of Hematology, Oslo University Hospital, Oslo,

Throughput Biomedicine Unit at Institute for Molecular Medicine

Norway

Finland (FIMM) for assistance with drug sensitivity screens. Correspondence: Funding

S.S. SKÅNLAND - sigrid.skanland@ous-research.no

The work was supported by the Research Council of Norway under the frame of ERA PerMed (to SSS, project number 322898), Lilly

https://doi.org/10.3324/haematol.2021.280393

Constance og Karl Ingolf Larssons stiftelse (to SSS), and Stiftelsen Kristian Gerhard Jebsen (to GET and LAM, grant 19).

Received: November 22, 2021. Accepted: February 23, 2022.

Data-sharing statement

Prepublished: March 3, 2022.

The data that support the findings of this study are available from

the corresponding author (sigrid.skanland@ous-research.no) upon reasonable request.

References 1. Lew TE, Lin VS, Cliff ER, et al. Outcomes of patients with CLL sequentially resistant to both BCL2 and BTK inhibition. Blood Adv. 2021;5(20):4054-4058. 2. Mato AR, Davids MS, Sharman J, et al. Recognizing unmet need in the era of targeted therapy for CLL/SLL: "What's past is prologue" (Shakespeare). Clin Cancer Res. 2022 Feb 15; 28(4):603-608. 3. Skånland SS, Mato AR. Overcoming resistance to targeted therapies in chronic lymphocytic leukemia. Blood Adv. 2021;5(1):334-343. 4. Letai A, Bhola P, Welm AL. Functional precision oncology: testing tumors with drugs to identify vulnerabilities and novel combinations. Cancer Cell. 2022;40(1):26-35. 5. Kornauth C, Pemovska T, Vladimer GI, et al. Functional precision medicine provides clinical benefit in advanced aggressive hematologic cancers and identifies exceptional responders. Cancer Discov. 2022;12(2):372-387. 6. Skånland SS, Cremaschi A, Bendiksen H, et al. An in vitro assay for biomarker discovery and dose prediction applied to ibrutinib plus venetoclax treatment of CLL. Leukemia. 2020;34(2):478-487. 7. Yadav B, Pemovska T, Szwajda A, et al. Quantitative scoring of differential drug sensitivity for individually optimized anticancer therapies. Sci Rep. 2014;4:5193. 8. Salem AH, Agarwal SK, Dunbar M, et al. Pharmacokinetics of venetoclax, a novel BCL-2 inhibitor, in patients with relapsed or refractory chronic lymphocytic leukemia or non-Hodgkin

lymphoma. J Clin Pharmacol. 2017;57(4):484-492. 9. Duechler M, Shehata M, Schwarzmeier JD, et al. Induction of apoptosis by proteasome inhibitors in B-CLL cells is associated with downregulation of CD23 and inactivation of Notch2. Leukemia. 2005;19(2):260-267. 10. Ruiz S, Krupnik Y, Keating M, et al. The proteasome inhibitor NPI-0052 is a more effective inducer of apoptosis than bortezomib in lymphocytes from patients with chronic lymphocytic leukemia. Mol Cancer Ther. 2006;5(7):1836-1843. 11. Billard C. Apoptosis inducers in chronic lymphocytic leukemia. Oncotarget. 2014;5(2):309-325. 12. Suzuki K, Handa H, Chou T, et al. Phase 1 study of ixazomib alone or combined with lenalidomide-dexamethasone in Japanese patients with relapsed/refractory multiple myeloma. Int J Hematol. 2017;105(4):445-452. 13. de Weerdt I, Hofland T, de Boer R, et al. Distinct immune composition in lymph node and peripheral blood of CLL patients is reshaped during venetoclax treatment. Blood Adv. 2019;3(17):2642-2652. 14. Skånland SS. Phospho flow cytometry with fluorescent cell barcoding for single cell signaling analysis and biomarker discovery. J Vis Exp. 2018;(140):e58386. 15. Hallaert DY, Jaspers A, van Noesel CJ, et al. c-Abl kinase inhibitors overcome CD40-mediated drug resistance in CLL: implications for therapeutic targeting of chemoresistant niches. Blood. 2008;112(13):5141-5149.

Haematologica | 107 August 2022

1998

CASE REPORT

Distinct genetic alterations in Burkitt-like lymphoma with 11q aberration and Burkitt lymphoma: a novel case report of composite lymphoma Composite lymphoma (CL) is defined as two or more morphologically and immunophenotypically distinct lymphomas within the same tissue site, most of each component has a different clonality accounting for 1-4.7% of all lymphomas. To date, variable components of CL have been reported, such as combination of non-Hodgkin lymphomas (NHL) and Hodgkin lymphoma (HL); multiple B-cell lymphomas, B-cell and T-cell NHL; and complex B-cell, T-cell, and HL. Most reported cases had the co-existence of a B-cell NHL and HL or two distinct B-cell NHL, usually low-grade such as mantle cell lymphoma with chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL) or follicular lymphoma (FL), or FL and CLL/SLL.1,2 However, CL consisting of large Bcell lymphomas has rarely been reported in the literature.2,3 Burkitt-like lymphoma with 11q aberration (BLL-11q) is a newly defined subset in the revised 4th edition of World

Health Organization (WHO) classification. BLL-11q has been described to resemble BL on morphological, immunophenotypic, and gene expression levels, but lacks MYC rearrangements and harbors a chromosome 11q aberration.4 Although they have overlapping features, co-occurrence of those two lymphomas has not been reported to date. Here, we present a novel case of composite BLL-11q and Burkitt lymphoma (BL) that have distinct morphologic, immunophenotypic, molecular, and genetic features. A 62-year-old male was referred to the emergency department with a 2-month history of epigastric pain and aggravation. A computed tomography (CT) scan revealed a segmental wall thickening of 5.7 cm in the terminal ileum with encasement of soft-tissue mass-like lesions along the ileocolic vessel, leading to small bowel obstruction. Multiple hypermetabolic lymph nodes in mesentery, paratracheal,

Figure 1. Gross image, histomorphology (haematoxylin and eosin stain) and immunohistochemical staining of composite lymphoma with Burkitt-like lymphoma with 11q aberration and Burkitt lymphoma. (A) Cut surface of a yellow-tan lesion involving the ileal wall along both sides of the muscularis propria. (B) The Burkitt-like lymphoma with 11q aberration (BLL-11q) involving mucosa and submucosa exhibits CD20-positive pleomorphic large lymphoid cells with negativity on BCL2 and MYC. The Burkitt lymphoma (BL) component in subserosa and serosa shows CD20-positive, BCL2-negative, and MYC-positive monomorphic medium size lymphoid cells with starry sky appearance, which are typical findings for BL (×400 magnification). Haematologica | 107 August 2022

1999

CASE REPORT and pulmonary hilar as well as hypermetabolic bone lesions in clavicle and ischium were detected using positron emission tomography (PET)-CT. The patient had Ann Arbor stage Ⅳ, but on a bone marrow biopsy, there was no evidence of involvement of lymphoma. The initial lactate dehydrogenase (LDH) level was increased with 297 IU/L (normal range, 120–250). A right hemicolectomy was performed as an elective surgery, and an intraluminal fungating mass was observed in the terminal ileum with severe luminal narrowing. The lesion was along both sides of the proper muscle layer consisting of two morphologically distinct components located in the serosal and mucosal layers of the ileum, separately (Figure 1A). The diagnosis of each component was highly suggestive of large cell lymphoma and BL, respectively, based on the histologic findings and immunostaining results. The component of large cell lymphoma showed positivity for CD20, CD10,

BCL6 and Ki-67 proliferation index of approximately 90% but negativity for CD5, BCL2, MYC, MUM1/IRF4 and EBER in situ hybridization (Figure 1B). Fluorescence in situ hybridisation (FISH) analysis was performed with BCL2, BCL6 and MYC break-apart rearrangement probes and showed neither rearrangement nor amplification in all three genes in the large cell lymphoma. These findings were suggestive of large Bcell lymphoma with a high probability of diagnosis of diffuse large B-cell lymphoma (DLBCL) at this level. The BL was positive for CD20, MYC, CD10, and BCL6 with Ki-67 proliferation index of over 95% but negative for BCL2, CD5, MUM1/IRF4 and EBER in situ hybridization (Figure 1B). 8q24 MYC rearrangement by FISH analysis was observed in the BL component (78.3% of all tumor cells) but no rearrangement or amplification was detected in BCL2 and BCL6 (Figure 2A). We performed IgH gene rearrangement tests on the CL and the two tumor components displayed distinct rearrange-

Continued on following page. Haematologica | 107 August 2022

2000

CASE REPORT

Figure 2. Molecular analyses performed on the composite lymphoma of Burkitt-like lymphoma with 11q aberration and Burkitt lymphoma. (A) On the MYC fluorescence in situ hybridization (FISH) analysis, the Burkitt-like lymphoma with 11q aberration (BLL11q) component are negative for MYC rearrangement whereas the Burkitt lymphoma (BL) shows MYC translocation (78.3% of the tumor cells). (B) Both components showed distinct monoclonality with different sizes of clonal peaks (arrows) on multiplex polymerase chain reaction assay with IGH-A (FR1, FR3, and DH7)) and IGH-B (FR2 and DH) primers. (C) BLL-11q shows gain of 11q23.3 and loss of 11q24.2–24.3 (arrows) but no significant copy number variation is observed in BL.

ment patterns (Figure 2B). The tumor-only targeted nextgeneration sequencing (NGS) test was performed separately for the two distinct tumor components. The test revealed a distinct pattern of copy number variation (CNV) with an 11q duplication region and terminal deletion on the large B-cell lymphoma component (gain of 11q23.3 and loss of 11q24.2– 24.3), which is consistent with that of BLL-11q. In contrast, no significant copy number aberration was noted in the conventional BL component (Figure 2C). The BLL-11q component also exhibited additional gains of 4q22.1 and 7q21.2 and loss of 2p24. The mutations identified through NGS tests are listed in Table 1. A missense mutation of MYC (p.P72S), which is known to be pathogenic, was detected in the BL component, whereas it was not identified in the BLL-11q component.5 Interestingly, these two components exhibited

common mutations in TP53 I195N gene, which is likely oncogenic, and KMT2D R5027Q of unknown biologic significance.6,7 After the surgery, the patient was treated with a dose-adjusted etoposide, prednisone, vincristine, cyclophosphamide, doxorubicin, and rituximab (EPOCH-R) regimen. After four cycles, the patient developed disease progression with extensive lymphomatous involvement of abdominal cavity and expired 5 months after initiation of chemotherapy. Four months later, additional excisional biopsy was performed, and the specimen showed Burkittoid morphologic appearance (monomorphic medium-size cells, scanty cytoplasm, fine chromatin with multiple small nucleoli, and “Starry-sky” appearance). It also showed MYC rearrangement on FISH analysis but no BCL2 and BCL6 which were the similar find-

Haematologica | 107 August 2022

2001

CASE REPORT ings with the BL component of the prior resection specimen. Horing et al. reported CL of DLBCL and BL in ileocecal area which was similar site with our case exhibiting two separate clonal population on IgH gene rearrangement analysis.8 Considering that genetic alteration of the case was not fully examined in detail and several cases of BLL-11q have been morphologically indistinguishable from DLBCL, the prior case of CL at the same site may provide a clue for understanding pathogenesis of large B-cell lymphoma.9,10 Although BLL-11q has been described to resemble BL on morphological, immunophenotypic and gene-expression levels, it lacks MYC rearrangements and has a chromosome 11q alteration characterized by proximal gains and telomeric losses: specifically, interstitial gains including a minimal region of gain in 11q23.2–23.3 and losses of 11q24.1–qter.9,11,12 Recently, Wagener et al. have demonstrated that apart from the absence of MYC rearrangement and the unique copy number aberration of 11q, BLL-11q lacks genetic mutations in genes of the ID3-TCF3 axis or the SWI/SNF complex such as ID3, MYC, or CCND3 that are frequently detected in BL.11 Gonzalez-Farre et al. have shown that BLL-11q has potential driver mutations particularly involving BTG2, DDX3X, ETS1, EP300, and GNA13 but ID3, TCF3, or CCND3 mutations are absent, suggesting that BLL-11q is a germinal center-derived and closer to highgrade B-cell lymphoma or diffuse large B-cell lymphoma (DLBCL) than to BL.10 In line with those studies, we demonstrated distinct histologic and genetic features between the BL and BLL-11q components, particularly as the form of CL. The BLL-11q

component had a gain of 11q23.3 and loss of 11q24.2–24.3 and there were no gains of 1q, 2p16.1 or 7p or loss of 1p36.32 that are usually observed in BL or germinal center B-cell (GCB) like-DLBCL. In addition, we could not detect any significant CNV in the BL component.10,13 Distinct mutation profiles were also observed between the two components in that BL harbored MYC missense mutation, but BLL-11q did not. Thus, it is likely that BLL-11q is a separate entity from BL; it is also noteworthy that those distinct entities can exhibit the development of a composite neoplasm. The BLL-11q component was CD10/BCL6-positive and MUM1-negative, highly suggestive of GCB derivation. Intriguingly, it showed several common mutations in TP53, PDCD11, KMT2D, and DDX3X with the BL component. Those mutations have been reported more frequently in BL or GCB B-cell lymphoma than in non-GCB type lymphomas, although the prevalence of KMT2D mutation is not significantly different between GCB-derived and non-GCB B-cell lymphomas.10,14,15 These findings suggest that both components might originate from a common precursor of GCB cell, then diverge before the clonal expansion. In this respect, it can be hypothesized that identically detected mutations of TP53, PDCD11, and KMT2D in both components might occur in early B-cell differentiation during lymphomagenesis before the divergence, and subsequently show distinct clonal evolution including IgH gene rearrangement, copy number aberration, and MYC gene rearrangement. Nevertheless, the possibility that the mutations shared with two components were germline alterations, or that the same mutations occurred

Table 1. Mutations identified in composite lymphoma of Burkitt-like lymphoma with 11q aberration and Burkitt lymphoma.

Type

Nucleotide Changes

Amino Acid Changes

Variant ID

BRCA2

c.623T>G

p.V208G

rs80358865

PDCD11

c.2875C>G

p.L959V

rs568839783

KMT2D

c.15080G>A

p.R5027Q

rs774403945

TP53

c.584T>A

p.I195N

DDX3X

c.676A>C

p.T226P

PHF6

c.374G>C

p.X125_splice

BRCA2

c.623T>G

p.V208G

rs80358865

MYC

c.214C>T

p.P72S

rs28933407

MYC

c.223C>G

p.P75A

PDCD11

c.2875C>G

p.L959V

rs568839783

KMT2D

c.15080G>A

p.R5027Q

rs774403945

c.584T>A

p.I195N

P2RY8

c.800_840del

p.Y267Sfs*62

DDX3X

c.488A>G

p.Y163C

BLL-11q

TP53

BLL-11q: Burkitt-like lymphoma with 11q aberration; BL: Burkitt lymphoma. Haematologica | 107 August 2022

2002

CASE REPORT

by chance in two different progenitors of the lymphoma components, cannot be completely ruled out. In conclusion, we reported a novel case of CL with BLL-11q and BL in which the components had distinct IgH rearrangement patterns and different features on morphologic, immunophenotypic, and genetic levels. Although multiple cases of CL with variable combinations have been reported, there have been few cases consisting of clonally unrelated and aggressive B-cell lymphomas. Considering that categorization of BLL-11q as a variant of BL, DLBCL or another distinct form of large B-cell lymphoma remains uncertain, the present case is significant in keeping with previous reports in terms of supporting the perspective that BLL-11q is a distinct entity from BL.8,12 This sheds light on the lymphomagenesis of CL that may originate from non-immunoglobulin gene-rearranged common progenitor cells.

Correspondence: H. GO - damul37@amc.seoul.kr https://doi.org/10.3324/haematol.2021.280543 Received: December 17, 2021. Accepted: March 18, 2022. Prepublished: March 31, 2022. ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

Disclosures No conflicts of interest to disclose. Contributions MK and HG performed the histological examination of the case and MK wrote the first draft of the manuscript with support from HSH. SC and DHY provided material or data of the case. MK, HSH and HG analyzed

Authors

and interpretated data. HG supervised the work. All authors approved the submission of the manuscript.

Meejeong Kim, Hee Sang Hwang, Dok Hyun Yoon, Sung-Min Chun and 1

Data-sharing statement

Heounjeong Go1

Data available on request from the authors. Department of Pathology and 2Department of Oncology, Asan Medical

Center, University of Ulsan College of Medicine, Seoul, Korea

References 1. Goyal G, Nguyen AH, Kendric K, Caponetti GC. Composite lymphoma with diffuse large B-cell lymphoma and classical Hodgkin lymphoma components: a case report and review of the literature. Pathol Res Pract. 2016;212(12):1179-1190. 2. Kuppers R, Duhrsen U, Hansmann ML. Pathogenesis, diagnosis, and treatment of composite lymphomas. Lancet Oncol. 2014;15(10):e435-446. 3. Miyaoka M, Kikuchi T, Carreras J, et al. Composite follicular lymphoma and CD5-positive nodal marginal zone lymphoma. J Clin Exp Hematop. 2016;56(1):55-58. 4. Swerdlow SH, Campo E, Harris NL, et al. WHO classification of tumours of haematopoietic and lymphoid tissues. Revised 4th edition. Lyon: International Agency for Research on Cancer, 2017. 5. Sati AOM, Osman WA, Ahmedon EAM, et al. Single nucleotide polymorphisms of the c-MYC gene’s relationship with formation of Burkitt’s lymphoma using bioinformatics analysis. bioRxiv. 2018 Oct 24. doi:10.1101/450783. [preprint, not peer-reviewed] 6. Testoni M, Zucca E, Young KH, Bertoni F. Genetic lesions in diffuse large B-cell lymphomas. Ann Oncol. 2015;26(6):1069-1080. 7. Zenz T, Kreuz M, Fuge M, et al. TP53 mutation and survival in aggressive B cell lymphoma. Int J Cancer. 2017;141(7):1381-1388. 8. Horing E, Staiger AM, Lenze D, et al. Burkitt lymphoma and diffuse large B-cell lymphoma: a unique case of a composite lymphoma of different clonal origin. Leuk Lymphoma. 2018;59(1):249-252. 9. Salaverria I, Martin-Guerrero I, Wagener R, et al. A recurrent 11q

aberration pattern characterizes a subset of MYC-negative highgrade B-cell lymphomas resembling Burkitt lymphoma. Blood. 2014;123(8):1187-1198. 10. Gonzalez-Farre B, Ramis-Zaldivar JE, Salmeron-Villalobos J, et al. Burkitt-like lymphoma with 11q aberration: a germinal centerderived lymphoma genetically unrelated to Burkitt lymphoma. Haematologica. 2019;104(9):1822. 11. Wagener R, Seufert J, Raimondi F, et al. The mutational landscape of Burkitt-like lymphoma with 11q aberration is distinct from that of Burkitt lymphoma. Blood. 2019;133(9):962-966. 12. Grygalewicz B, Woroniecka R, Rymkiewicz G, et al. The 11qgain/loss Aberration occurs recurrently in MYC-negative Burkitt-like lymphoma with 11q aberration, as well as MYCpositive Burkitt lymphoma and MYC-positive high-grade B-cell lymphoma, NOS. Am J Clin Pathol. 2017;149(1):17-28. 13. Boerma EG, Siebert R, Kluin PM, Baudis M. Translocations involving 8q24 in Burkitt lymphoma and other malignant lymphomas: a historical review of cytogenetics in the light of todays knowledge. Leukemia. 2009;23(2):225-234. 14. Schmitz R, Young RM, Ceribelli M, et al. Burkitt lymphoma pathogenesis and therapeutic targets from structural and functional genomics. Nature. 2012;490(7418):116-120. 15. Dubois S, Viailly P-J, Mareschal S, et al. Next-generation sequencing in diffuse large B-cell lymphoma highlights molecular divergence and therapeutic opportunities: a LYSA study. Clin Cancer Res. 2016;22(12):2919-2928.

Haematologica | 107 August 2022

2003

CASE REPORT

Treatment of leptomeningeal disease in blastic plasmacytoid dendritic cell neoplasm following tagraxofusp-erzs induction Treatment of blastic plasmacytoid dendritic cell neoplasm (BPDCN) has significantly improved following the introduction of tagraxofusp-erzs (SL-401), a cytotoxic immunotoxin targeting CD123/IL-3 receptor α.1 However, management of this rare hematologic malignancy remains demanding considering its aggressiveness with an often dismal outcome. Particularly leptomeningeal manifestation of BPDCN poses significant challenges. Here, we report diagnostics and successful hematopoietic stem cell transplantation of a first BPDCN patient initially treated with tagraxofusp-erzs and subsequently developing leptomeningeal disease. The significance of early disease detection in cerebrospinal fluid and the importance of central nervous system (CNS) directed chemotherapy are highlighted. A 52-year-old previously healthy house painter was referred to our department for evaluation of an allogeneic hematopoietic stem cell transplantation (HSCT) in April 2021. Approximately 5 months prior to referral, the patient presented with multiple circumscribed brownish, partly violaceous plaques up to 4 cm on his trunk and head (Figure 1A). Besides, the patient had enlarged cervical, axillary, and inguinal lymph nodes. A full body computerized tomography (CT) scan revealed additional lymphadenopathy in the mediastinum and portocaval region. Infectious diseases were excluded serologically and by polymerase chain reaction. Cytology and immunophenotyping (IPT) of peripheral blood and bone marrow samples did not show any signs of lymphoma, leukemia, or other hematologic diseases. Surgical biopsies of several skin lesions and an inguinal lymph node confirmed the diagnosis of CD4+CD56+CD123+ BPDCN. The patient was treated with three cycles of the anti-CD123 antibody tagraxofusp-erzs. Apart from an elevation of liver enzymes during the first treatment cycle, tagraxofusp-erzs was well tolerated. The patient showed complete remission (CR) of the skin lesions and lymphadenopathy. A brain magnetic resonance imaging (MRI) in April 2021 did not reveal any pathological findings. The patient’s brother was identified as an HLA-identical stem cell donor for consolidative allogeneic HSCT. A week before planned HSCT the patient was admitted to another hospital with acute onset dysarthria. Neurological examination showed slightly atactic upper extremities. Cerebral ischemia was excluded by CT including angiography. However, there were multiple cortical contrast medium enhancements on an MRI-scan which were indicative of lep-

tomeningeal disease. A lumbar puncture (LP) revealed a cell count of >7x109/L. Cerebrospinal fluid (CSF) was examined cytologically (Figure 1B) which demonstrated large blastic cells with lobulated nuclei. IPT showed positivity for CD4 and CD56, thus confirming the suspected diagnosis of leptomeningeal BPDCN recurrence (Figure 1C). At that time, additional extracerebral manifestations were excluded by positron emission tomography (PET)-CT scanning. The patient was treated with high-dose methotrexate (HDMTX; 4 g/m2) and ifosfamide (2 g/m2) as well as bi-weekly intrathecal therapy comprising methotrexate, cytarabine, and dexamethasone. During hematopoietic regeneration autologous peripheral blood stem cells were harvested. After two cycles of systemic treatment and four courses of intrathecal chemotherapy the patient showed full neurologic recovery and BPDCN cells were cleared from the CSF. Following consolidation with high-dose chemotherapy (carmustin/ BCNU 400 mg/m2 and thiotepa 20 mg/kg) and subsequent autologous HSCT the patient recovered without major complications (Figure 2). Over the next 2 months the patient did not show any clinical signs of BPDCN recurrence and a PET-CT scan as well as further CSF analysis confirmed a CR. Thus, the patient was admitted for consolidative allogeneic HSCT after conditioning with fludarabine (150 mg/m2), busulfan (6.4 mg/kg), and post-transplantation cyclophosphamide (100 mg/kg).2 Apart from a period of short bacteremia, he recovered without further complications. Immunosuppressives were tapered over the next 100 days without major graft-versus-host disease (GvHD). Ten months after CNS recurrence of BPDCN the patient is in CR and excellent clinical condition (Figure 2). BPDCN is a rare disease accounting for less than 0.5% of all hematological malignancies.3,4 Skin involvement is the most common feature, present in more than 85% of cases.5 Recently, Pemmaraju et al. retrospectively analyzed 103 patients, of which only 29 had received LP (57% performed routinely at diagnosis) over the course of their disease.6 Of these patients, 13 had frontline CNS disease and ten developed overt CNS disease over the course of time, respectively. In another recently published Italian cohort of 68 patients, four patients presented with frontline CNS disease and two developed CNS recurrence (70% routine CSF exams at diagnosis).7 These studies indicate the importance of CNS involvement in BPDCN, but also show, that routine CNS testing and prophylactic intrathecal treatment have

Haematologica | 107 August 2022

2004

CASE REPORT

Figure 1. Cutaneous manifestation and central nervous system relapse of blastic plasmacytoid dendritic cell neoplasm. (A) Skin lesions at diagnosis of blastic plasmacytoid dendritic cell neoplasm (BPDCN). (B) Cytology of cerebrospinal fluid (CSF) cells at BPDCN relapse. (C) Immunophenotyping of CSF cells at BPDCN relapse.

not been implemented so far, despite clinical practice guidelines (NCCN 20228) and some data suggesting a benefit of acute lymphoblastic (ALL)-like treatment protocols including intrathecal treatment.9,10 The anti-CD123 antibody-immunotoxin conjugate tagraxofusp-erzs (SL-401) received Food and Drug Administration and European Medicines Agency approval as upfront single agent after the results of a single arm phase I/II trial.11 In this study, patients with known CNS involvement were excluded. CNS relapses were not detected throughout the study. Notably, routine LP had not been performed. Even more recently, an analysis of intensive and CNS active upfront chemotherapy (HCVAD) versus SL-401 did not reveal any CNS relapses in the targeted therapy group.12 To the best of our knowledge, our case is the first fully published CNS relapse after successful induction treatment with tagraxofusp-erzs. Our case adds to recently published data on CNS involvement in BPDCN.7,8,12 Furthermore, it conveys three important messages: first, LP is an essential initial diagnostic procedure in BPDCN, also in the context of therapy with tagraxofusp-erzs. Due to their distinct morphology and aberrant immunophenotype BPDCN cells can be easily detected (Figure 1B and C). This has been highlighted in several studies and has been incorporated in current expert-guidelines.8 Since our patient had not received any CSF cytology prior to induction therapy, unrecognized CNS involvement at diagnosis cannot be

ruled out. Considering the clinical significance and the treatability, sensitive detection of CD4+CD56+CD123+ CSF cells, e.g., by next-generation flow cytometry,10 is strongly recommended in each BPDCN patient at diagnosis.8 Second, although the rates of CNS manifestations are currently unknown under CD123-targeted therapy with tagraxofusp-erzs, they obviously do occur. One explanation is the missing capability of the antibody tagraxofusp-erzs to cross the blood-brain barrier. Considering historical data utilizing CNS active ALL protocols for BPDCN treatment, intrathecal prophylaxis and/or HD-MTX might be options and should be evaluated in any further clinical trial. Indeed, three cases have been reported as abstract, although these patients were treated without hematopoietic stem cell transplantation.13 Our case underlines the cytotoxic effectiveness of the chosen treatment regimen against BPDCN cells and points to the importance of allogeneic HSCT. Finally, in case of CNS disease treatment in analogy to CNS lymphomas incorporating HD-MTX and other CNS- penetrating drugs in combination with intrathecal chemotherapy as well as high-dose chemotherapy including autologous HSCT offers a feasible option which led to CR in our patient.13,15 Nevertheless, allogeneic HSCT at the moment remains the consolidation treatment of choice for eligible patients.15,16 As CR prior to allogeneic HSCT is an important prognostic factor,17 we decided to first treat CNS relapse intensively including high-dose chemotherapy with

Haematologica | 107 August 2022

2005

CASE REPORT

Figure 2. Blastic plasmacytoid dendritic cell neoplasm disease course and treatment time-line. BPCDCN: blastic plasmacytoid dendritic cell neoplasm; HSCT: hematopoietic stem cell transplantation; MTX: methotrexate; MRI: magnetic resonance imaging; CSF: cerebrospinal fluid; CNS: central nervous system; PR: partial remission; CR: complete remission; PET: Positron emission tomography; GvHD: graft-versus-host disease: IT-triple: triple intrathecal treatment.

autologous HSCT and conduct allogeneic HSCT following a slightly modified conditioning regime previously reported in primary CNS lymphoma by our group.2 In order to achieve the maximum therapeutic effect, our treatment approach encompassed a broad spectrum of CNS effective drugs, including MTX, thiotepa, BCNU, and busulfan, as previously established within CNS protocols.14,15 Ifosfamide was administered to facilitate early blood stem-cell collection.15 Considering overall cumulative toxicity, we decided against myeloablative conditioning prior to allogeneic HSCT. In summary, CNS involvement should be suspected right from the start with any BPDCN diagnosis and CNS manifestation can be treated by protocols analogous to CNS lymphoma. Realizing the frequency of occult CNS disease, early addition of prophylactic treatment with e.g., MTX should be considered as part of the initial therapy with tagraxofusp-erzs in routine and within future clinical studies.

Department of Hematology and Oncology,

Knappschaftskrankenhaus, Ruhr-University Bochum and 2Skin Cancer Center, Department of Dermatology, St. Joseph Hospital, RuhrUniversity Bochum, Bochum, Germany Correspondence: R. SCHROERS - Roland.Schroers@rub.de https://doi.org/10.3324/haematol.2022.280843 Received: February 14, 2022. Accepted: March 25, 2022. Prepublished: April 7, 2022. ©2022 Ferrata Storti Foundation Haematologica material is published under a CC-BY-NC license

Disclosures No conflicts of interest to disclose.

Authors

Contributions DV, TG, and RS guided diagnosis and treatment of the patient and wrote the manuscript. DV, VN, TM, TG, and RS collected the data. All

Deepak B. Vangala,1 Verena Nilius-Eliliwi,1 Thomas Mika,1 Thilo

authors discussed the data and manuscript.

Gambichler,2 Rene Stranzenbach2 and Roland Schroers1

References 1. Pemmaraju N, Konopleva M. Approval of tagraxofusp-erzs for blastic plasmacytoid den-dritic cell neoplasm. Blood Adv. 2020;4(16):4020-4027. 2. Mika T, Ladigan S, Baraniskin A, et al. Allogeneic hematopoietic stem cell transplantation for primary central nervous system lymphoma. Haematologica. 2020;105(4):e160-e163. 3. Bueno C, Almeida J, Lucio P, et al. Incidence and characteristics

of CD4(+)/HLA DRhi den-dritic cell malignancies. Haematologica. 2004;89(1):58-69. 4. Pagano L, Valentini CG, Pulsoni A, et al. Blastic plasmacytoid dendritic cell neoplasm with leukemic presentation: an italian multicenter study. Haematologica. 2013;98(2):239-246. 5. Sullivan JM, Rizzieri DA. Treatment of blastic plasmacytoid dendritic cell neoplasm. Hematology. 2016;2016(1):16-23.

Haematologica | 107 August 2022

2006

CASE REPORT 6. Pemmaraju N, Wilson NR, Khoury JD, et al. Central nervous system involvement in blastic plasmacytoid dendritic cell neoplasm. Blood. 2021;138(15):1373-1377. 7. Valentini CG, Piciocchi A, Facchetti F, et al. Blastic plasmocitoid dendritic cell neoplasm with leukemic spread: a GIMEMA survey. Blood Adv. 2021;5(24):5608-5611. 8. National Cancer Network Guideline (NCCN). Version 1.2022. BPDCN-B. 9. Martín-Martín L, López A, Vidriales B, et al. Classification and clinical behavior of blastic plasmacytoid dendritic cell neoplasms according to their maturation-associated immunophenotypicprofile. Oncotarget. 2015;6(22):19204-19216. 10. Martín-Martín L, Almeida J, Pomares H, et al. Blastic plasmacytoid dendritic cell neoplasm frequently shows occult central nervous system involvement at diagnosis and benefits from intrathecal therapy. Oncotarget. 2016;7(9):10174-10181. 11. Pemmaraju N, Lane AA, Sweet KL, et al. Tagraxofusp in blastic plasmacytoid dendritic-cell neoplasm. N Engl J Med. 2019;380(17):1628-1637. 12. Pemmaraju N, Wilson NR, Garcia-Manero G, et al. Characteristics and outcomes of patients with blastic plasmacytoid dendritic cell neoplasm treated with frontline

HCVAD. Blood Adv. 2022;6(10):3027-3035. 13. Greenwell IB, Davis J, Li H, et al. Outcomes of CNS involvement in blastic plasmacytoid dendritic cell neoplasm (BPDCN). J Clin Oncol. 2021;39(Suppl 15):Se19043. 14. Kasenda B, Ihorst G, Schroers R, et al. High-dose chemotherapy with autologous haema-topoietic stem cell support for relapsed or refractory primary CNS lymphoma: a prospec-tive multicentre trial by the German Cooperative PCNSL study group. Leukemia. 2017;31(12):2623-2629. 15. Korfel A, Elter T, Thiel E, et al. Phase II study of central nervous system (CNS)-directed chemotherapy including high-dose chemotherapy with autologous stem cell transplanta-tion for CNS relapse of aggressive lymphomas. Haematologica. 2013;98(3):364-370. 16. Taylor J, Haddadin M, Upadhyay VA, et al. Multicenter analysis of outcomes in blastic plasmacytoid dendritic cell neoplasm offers a pretargeted therapy benchmark. Blood. 2019;134(8):678-687. 17. Bashir Q, Milton DR, Popat UR, et al. Allogeneic hematopoietic cell transplantation for patients with blastic plasmacytoid dendritic cell neoplasm (BPDCN). Bone Marrow Transplant. 2022;57(1):51-56.

Haematologica | 107 August 2022

2007