Page 1


haematologica Journal of the Ferrata Storti Foundation

Editor-in-Chief Luca Malcovati (Pavia)

Managing Director Antonio Majocchi (Pavia)

Associate Editors Omar I. Abdel-Wahab (New York), Hélène Cavé (Paris), Simon Mendez-Ferrer (Cambridge), Pavan Reddy (Ann Arbor), Andreas Rosenwald (Wuerzburg), Monika Engelhardt (Freiburg), Davide Rossi (Bellinzona), Jacob Rowe (Haifa, Jerusalem), Wyndham Wilson (Bethesda), Paul Kyrle (Vienna), Swee Lay Thein (Bethesda), Pieter Sonneveld (Rotterdam)

Assistant Editors Anne Freckleton (English Editor), Cristiana Pascutto (Statistical Consultant), Rachel Stenner (English Editor), Kate O’Donohoe (English Editor), Ziggy Kennell (English Editor)

Editorial Board Jeremy Abramson (Boston); Paolo Arosio (Brescia); Raphael Bejar (San Diego); Erik Berntorp (Malmö); Dominique Bonnet (London); Jean-Pierre Bourquin (Zurich); Suzanne Cannegieter (Leiden); Francisco Cervantes (Barcelona); Nicholas Chiorazzi (Manhasset); Oliver Cornely (Köln); Michel Delforge (Leuven); Ruud Delwel (Rotterdam); Meletios A. Dimopoulos (Athens); Inderjeet Dokal (London); Hervé Dombret (Paris); Peter Dreger (Hamburg); Martin Dreyling (München); Kieron Dunleavy (Bethesda); Dimitar Efremov (Rome); Sabine Eichinger (Vienna); Jean Feuillard (Limoges); Carlo Gambacorti-Passerini (Monza); Guillermo Garcia Manero (Houston); Christian Geisler (Copenhagen); Piero Giordano (Leiden); Christian Gisselbrecht (Paris); Andreas Greinacher (Greifswals); Hildegard Greinix (Vienna); Paolo Gresele (Perugia); Thomas M. Habermann (Rochester); Claudia Haferlach (München); Oliver Hantschel (Lausanne); Christine Harrison (Southampton); Brian Huntly (Cambridge); Ulrich Jaeger (Vienna); Elaine Jaffe (Bethesda); Arnon Kater (Amsterdam); Gregory Kato (Pittsburg); Christoph Klein (Munich); Steven Knapper (Cardiff); Seiji Kojima (Nagoya); John Koreth (Boston); Robert Kralovics (Vienna); Ralf Küppers (Essen); Ola Landgren (New York); Peter Lenting (Le Kremlin-Bicetre); Per Ljungman (Stockholm); Francesco Lo Coco (Rome); Henk M. Lokhorst (Utrecht); John Mascarenhas (New York); Maria-Victoria Mateos (Salamanca); Giampaolo Merlini (Pavia); Anna Rita Migliaccio (New York); Mohamad Mohty (Nantes); Martina Muckenthaler (Heidelberg); Ann Mullally (Boston); Stephen Mulligan (Sydney); German Ott (Stuttgart); Jakob Passweg (Basel); Melanie Percy (Ireland); Rob Pieters (Utrecht); Stefano Pileri (Milan); Miguel Piris (Madrid); Andreas Reiter (Mannheim); Jose-Maria Ribera (Barcelona); Stefano Rivella (New York); Francesco Rodeghiero (Vicenza); Richard Rosenquist (Uppsala); Simon Rule (Plymouth); Claudia Scholl (Heidelberg); Martin Schrappe (Kiel); Radek C. Skoda (Basel); Gérard Socié (Paris); Kostas Stamatopoulos (Thessaloniki); David P. Steensma (Rochester); Martin H. Steinberg (Boston); Ali Taher (Beirut); Evangelos Terpos (Athens); Takanori Teshima (Sapporo); Pieter Van Vlierberghe (Gent); Alessandro M. Vannucchi (Firenze); George Vassiliou (Cambridge); Edo Vellenga (Groningen); Umberto Vitolo (Torino); Guenter Weiss (Innsbruck).

Editorial Office Simona Giri (Production & Marketing Manager), Lorella Ripari (Peer Review Manager), Paola Cariati (Senior Graphic Designer), Igor Ebuli Poletti (Senior Graphic Designer), Marta Fossati (Peer Review), Diana Serena Ravera (Peer Review)

Affiliated Scientific Societies SIE (Italian Society of Hematology, www.siematologia.it) SIES (Italian Society of Experimental Hematology, www.siesonline.it)


haematologica Journal of the Ferrata Storti Foundation

Information for readers, authors and subscribers Haematologica (print edition, pISSN 0390-6078, eISSN 1592-8721) publishes peer-reviewed papers on all areas of experimental and clinical hematology. The journal is owned by a non-profit organization, the Ferrata Storti Foundation, and serves the scientific community following the recommendations of the World Association of Medical Editors (www.wame.org) and the International Committee of Medical Journal Editors (www.icmje.org). Haematologica publishes editorials, research articles, review articles, guideline articles and letters. Manuscripts should be prepared according to our guidelines (www.haematologica.org/information-for-authors), and the Uniform Requirements for Manuscripts Submitted to Biomedical Journals, prepared by the International Committee of Medical Journal Editors (www.icmje.org). Manuscripts should be submitted online at http://www.haematologica.org/. Conflict of interests. According to the International Committee of Medical Journal Editors (http://www.icmje.org/#conflicts), “Public trust in the peer review process and the credibility of published articles depend in part on how well conflict of interest is handled during writing, peer review, and editorial decision making”. The ad hoc journal’s policy is reported in detail online (www.haematologica.org/content/policies). Transfer of Copyright and Permission to Reproduce Parts of Published Papers. Authors will grant copyright of their articles to the Ferrata Storti Foundation. No formal permission will be required to reproduce parts (tables or illustrations) of published papers, provided the source is quoted appropriately and reproduction has no commercial intent. Reproductions with commercial intent will require written permission and payment of royalties. Detailed information about subscriptions is available online at www.haematologica.org. Haematologica is an open access journal. Access to the online journal is free. Use of the Haematologica App (available on the App Store and on Google Play) is free. For subscriptions to the printed issue of the journal, please contact: Haematologica Office, via Giuseppe Belli 4, 27100 Pavia, Italy (phone +39.0382.27129, fax +39.0382.394705, E-mail: info@haematologica.org). Rates of the International edition for the year 2019 are as following: Print edition

Institutional Euro 700

Personal Euro 170

Advertisements. Contact the Advertising Manager, Haematologica Office, via Giuseppe Belli 4, 27100 Pavia, Italy (phone +39.0382.27129, fax +39.0382.394705, e-mail: marketing@haematologica.org). Disclaimer. Whilst every effort is made by the publishers and the editorial board to see that no inaccurate or misleading data, opinion or statement appears in this journal, they wish to make it clear that the data and opinions appearing in the articles or advertisements herein are the responsibility of the contributor or advisor concerned. Accordingly, the publisher, the editorial board and their respective employees, officers and agents accept no liability whatsoever for the consequences of any inaccurate or misleading data, opinion or statement. Whilst all due care is taken to ensure that drug doses and other quantities are presented accurately, readers are advised that new methods and techniques involving drug usage, and described within this journal, should only be followed in conjunction with the drug manufacturer’s own published literature. Direttore responsabile: Prof. Edoardo Ascari; Autorizzazione del Tribunale di Pavia n. 63 del 5 marzo 1955. Printing: Press Up, zona Via Cassia Km 36, 300 Zona Ind.le Settevene - 01036 Nepi (VT)


haematologica Journal of the Ferrata Storti Foundation

Table of Contents Volume 104, Issue 3: March 2019 Cover Figure Histiocyte that has engulfed platelets in a case of familial hemophagocytic lymphohistiocytosis. Courtesy of Prof. Rosangela Invernizzi.

Editorials 417

Anemia and adverse outcomes in the elderly: a detrimental inflammatory loop? Domenico Girelli and Fabiana Busti

419

Targeting a major hub of cell fate decisions – the mitochondrial-associated membrane William L. Carroll and Nikki A. Evensen

422

Learning the next-generation sequencing alphabet of immune reconstitution: factors determining CD8+ T-cell receptor α-chain repertoire dynamics after hematopoietic stem cell transplantation Esteban Arrieta-Bolaños and Katharina Fleischhauer

425

Mosaicism by somatic non-functional mutations: one cell lineage at a time Willy Albert Flegel

Perspective Article 428

Endothelialized flow models for blood transfusion research Monica S.Y. Ng et al.

Review Articles 435

Reactivation of hepatitis B virus infection in patients with hematologic disorders Bo Wang et al.

444

IgG4-related disease: what a hematologist needs to know Luke Y.C. Chen et al.

Articles Hematopoiesis

456

Chronic sympathetic driven hypertension promotes atherosclerosis by enhancing hematopoiesis Annas Al-Sharea et al.

Red Cell Biology & its Disorders

468

Association of anemia with health-related quality of life and survival: a large population-based cohort study Hanneke J.C.M. Wouters et al.

477

Sotatercept, a novel transforming growth factor β ligand trap, improves anemia in β-thalassemia: a phase II, open-label, dose-finding study Maria Domenica Cappellini et al.

Myelodysplastic Syndromes

485

The high NRF2 expression confers chemotherapy resistance partly through up-regulated DUSP1 in myelodysplastic syndromes Peipei Lin et al.

497

Dyserythropoiesis evaluated by the RED score and hepcidin:ferritin ratio predicts response to erythropoietin in lower-risk myelodysplastic syndromes Sophie Park et al.

Haematologica 2019; vol. 104 no. 3 - March 2019 http://www.haematologica.org/


haematologica Journal of the Ferrata Storti Foundation Acute Myeloid Leukemia

505

Transglutaminase 2 programs differentiating acute promyelocytic leukemia cells in all-trans retinoic acid treatment to inflammatory stage through NF-kB activation Károly Jambrovics et al.

516

Clinical implications of subclonal TP53 mutations in acute myeloid leukemia Katharina T. Prochazka et al.

524

Haploidentical versus unrelated allogeneic stem cell transplantation for relapsed/refractory acute myeloid leukemia: a report on 1578 patients from the Acute Leukemia Working Party of the EBMT Eolia Brissot et al.

Acute Lymphoblastic Leukemia

533

ActivinA: a new leukemia-promoting factor conferring migratory advantage to B-cell precursor-acute lymphoblastic leukemic cells Federica Portale et al.

546

Targeting the endoplasmic reticulum-mitochondria interface sensitizes leukemia cells to cytostatics Fabian Koczian et al.

556

Trypsin-encoding PRSS1-PRSS2 variations influence the risk of asparaginase-associated pancreatitis in children with acute lymphoblastic leukemia: a Ponte di Legno toxicity working group report Benjamin O. Wolthers et al.

Hodgkin Lymphoma

564

CCR5 antagonism by maraviroc inhibits Hodgkin lymphoma microenvironment interactions and xenograft growth Naike Casagrande et al.

Chronic Lymphocytic Leukemia

576

Mutations in the RAS-BRAF-MAPK-ERK pathway define a specific subgroup of patients with adverse clinical features and provide new therapeutic options in chronic lymphocytic leukemia Neus Giménez et al.

Hemostasis

587

Unraveling the effect of silent, intronic and missense mutations on VWF splicing: contribution of next-generation sequencing in the study of mRNA Nina Borràs et al.

Coagulation & its Disorders

599

Factor VIII cross-matches to the human proteome reduce the predicted inhibitor risk in missense mutation hemophilia A Daniel P. Hart et al.

Immunodeficiencies

609

Selective loss of function variants in IL6ST cause Hyper-IgE syndrome with distinct impairments of T-cell phenotype and function Tala Shahin et al.

Stem Cell Transplantation

622

T-cell receptor-α repertoire of CD8+ T cells following allogeneic stem cell transplantation using next-generation sequencing Cornelia S. Link-Rachner et al.

Blood Transfusion

632

Somatic mosaicisms of chromosome 1 at two different stages of ontogenetic development detected by Rh blood group discrepancies Eva-Maria Daube et al.

Letters to the Editor Letters are available online only at www.haematologica.org/content/104/3.toc

e87

Transient decrease of serum iron after acute erythropoietin treatment contributes to hepcidin inhibition by ERFE in mice Irene Artuso et al. http://www.haematologica.org/content/104/3/e87

Haematologica 2019; vol. 104 no. 3 - March 2019 http://www.haematologica.org/


haematologica Journal of the Ferrata Storti Foundation

e91

A gain of function variant in PIEZO1 (E756del) and sickle cell disease Helen Rooks et al. http://www.haematologica.org/content/104/3/e91

e94

Significant hemolysis is not required for thrombosis in paroxysmal nocturnal hemoglobinuria Morag Griffin et al. http://www.haematologica.org/content/104/3/e94

e97

Highly sensitive methods are required to detect mutations in histiocytoses Sarah Melloul et al. http://www.haematologica.org/content/104/3/e97

e100

CD38 as a therapeutic target for adult acute myeloid leukemia and T-cell acute lymphoblastic leukemia Jyoti Naik et al. http://www.haematologica.org/content/104/3/e100

e104

DNA methylation profiling of hepatosplenic T-cell lymphoma Anke K. Bergmann et al. http://www.haematologica.org/content/104/3/e104

e108

T-cell large granular lymphocytic leukemia and plasma cell disorders M. Hasib Sidiqi et al. http://www.haematologica.org/content/104/3/e108

e111

Impaired factor XIII activation in patients with congenital afibrinogenemia Franรงoise Bridey et al. http://www.haematologica.org/content/104/3/e111

Case Reports Case reports are available online only at www.haematologica.org/content/104/3.toc

e114

Molecular analysis of a CD19-negative diffuse large B-cell lymphoma Lorric Delage et al. http://www.haematologica.org/content/104/3/e114

e117

T-cell large granular lymphocyte leukemia transfomation into aggressive T-cell lymphoma: a report of two cases with molecular characterization Maya Belhadj et al. http://www.haematologica.org/content/104/3/e117

e121

Resolution of celiac disease, IgA deficiency and platelet refractoriness after allogeneic bone marrow transplantation for acute leukemia Sasan Zandi et al. http://www.haematologica.org/content/104/3/e121

Comments Comments are available online only at www.haematologica.org/content/104/3.toc

e124

Rethinking the elusive boundaries of EBV-associated T/NK-cell lymphoproliferative disorders Zihang Chen et al. http://www.haematologica.org/content/104/3/e124

e126

Additional considerations related to the elusive boundaries of EBV-associated T/NK-cell lymphoproliferative disorders Sebastian Fernandez-Pol et al. http://www.haematologica.org/content/104/3/e126

Haematologica 2019; vol. 104 no. 3 - March 2019 http://www.haematologica.org/


haematologica Journal of the Ferrata Storti Foundation

Ancient Greek

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

Scientific Latin

aÂma [haima] = blood a·matow [haimatos] = of blood lÒgow [logos]= reasoning

Scientific Latin

haematologicus (adjective) = related to blood

Modern English

haematologica (adjective, plural and neuter, used as a noun) = hematological subjects The oldest hematology journal, publishing the newest research results. 2017 JCR impact factor = 9.090


EDITORIALS Anemia and adverse outcomes in the elderly: a detrimental inflammatory loop? Domenico Girelli and Fabiana Busti Department of Medicine, Section of Internal Medicine, University of Verona, EuroBloodNet Referral Center for Iron Metabolism Disorders, Azienda Ospedaliera Universitaria Integrata Verona, Italy. E-mail: DOMENICO GIRELLI - domenico.girelli@univr.it doi:10.3324/haematol.2018.208066

A

nemia is one of the easiest diagnoses in clinical practice, being based on a widely available and cheap laboratory parameter: hemoglobin concentration. In the near future, this diagnosis could become even simpler through a smartphone app.1 Nonetheless, anemia is frequently overlooked, particularly when it is mild (e.g. hemoglobin values >10 g/dL but <12 g/dL or <13 g/dL in females and males, respectively) and no obvious symptoms are apparently associated. This typically occurs in the elderly, in whom other co-morbidities are often present, distracting the attention of physicians and caregivers.2 Indeed, the prevalence of anemia in people aged >65 years is high, ranging from nearly 12% in those living in the community to more than 45% in institutionalized nursing-home residents.2 While a mild anemia has been traditionally considered as a “physiological” consequence of aging, more recent studies have shown that the decline in hemoglobin is minimal, if any, in the “wellderly”, namely people aging well without significant comorbidities.3 Moreover, growing evidence suggests that anemia in the elderly is not an innocent bystander, being strongly and independently associated with a number of adverse outcomes, including cognitive decline, reduced physical performance, increased risk of falls and fractures, and even increased mortality.2 In this issue of Haematologica, Wouters and colleagues report the results of the large Lifelines Cohort Study, confirming that anemia in the elderly is negatively associated with either quality of life or survival.4 A key point of anemia in the elderly lies in the difficulty of establishing the etiology, which in turn should drive its management. While anemia in young people is generally due to a single cause (e.g. iron deficiency in premenopausal women with heavy menstrual bleeding), anemia in the elderly is often multifactorial, reflecting the typical multimorbidity of aged people.5 This frequently makes it hard to dissect out the main mechanism leading to anemia in a given elderly individual. For example, iron deficiency in the elderly can be due to a mixture of malnutrition, malabsorption and bleeding, often aggravated by multiple medication use.6 Such difficulty is even more pronounced in large epidemiological surveys, such as the Lifelines Study, in which only three general subcategories of anemia can be distinguished, based on a few available laboratory parameters: nutritional deficiencies, anemia of inflammation (also named anemia of chronic diseases), and “unexplained” anemia,7 each of them accounting roughly for one third of cases. Notwithstanding these inherent limitations, the study by Wouters and colleagues points out anemia of inflammation as the most detrimental subcategory in the elderly, because of the strongest association with quality of life and mortality. This is somewhat at variance with the findings of other studies, in which subjects with anemia due to nutritional deficiency showed the worst prognosis.8 Such a discrepancy reflects the uncertainty in the correct etiological classificahaematologica | 2019; 104(3)

tion of anemia in the elderly, as well as our limited knowledge in the field. The broad category of “unexplained anemia” well illustrates this gap, although it likely represents a heterogeneous group of conditions that cannot be adequately addressed by large epidemiological surveys because of the need for second level laboratory tests. Such conditions include androgen deficiency, vitamin D deficiency, unrecognized iron deficiency with apparently normal traditional biomarkers, impaired bone marrow response to erythropoietin, clonal hematopoiesis, and “inflammaging”.2 Of note, recent advances suggest that the last two conditions may share a relevant role in the pathophysiology of anemia in the elderly, both by inducing a low-grade chronic inflammatory status (Figure 1). Clonal hematopoiesis refers to agerelated acquisition of somatic mutations in certain driver genes (e.g. TET2, DNMT3A, and JAK2) in hematopoietic stem cells, conferring them a competitive advantage and hence giving rise to a clonal progeny in the peripheral blood, at variance with normal polyclonal hematopoiesis.9 Clonal hematopoiesis is detectable through next-generation sequencing studies in nearly 10% of 70-year old individuals without abnormalities of peripheral blood cell counts and, in these people, is a risk factor for subsequent development of myeloid neoplasms. This is generally associated with accumulation of multiple mutations, but the rate of progression appears as low as 0.5% per year. Such a condition has been termed clonal hematopoiesis of indeterminate potential (CHIP), since, in fact, most carriers of clonal hematopoiesis will never develop myeloid neoplasms. Nevertheless, individuals with CHIP have been found at increased risk of mortality due to cardiovascular events, rather than to hematologic complications.10 Mounting evidence suggests that accelerated atherosclerosis in CHIP is associated with a systemic pro-inflammatory status driven by abnormal clonal leukocytes deriving from mutated hematopoietic stem cells.11 On the other hand, “inflammaging” also contributes to a chronic upregulation of proinflammatory cytokines in the elderly. This phenomenon is thought to be the result of activation of the nuclear factorκB/inflammasome pathway driven by the so-called damaged-associated molecular pattern, i.e., endogenous altered molecules and reactive oxygen species that accumulate with aging.12,13 Reciprocal influences between CHIP and “inflammaging” are likely, for example considering that reactive oxygen species can also cause genomic instability, and that subclinical inflammation by itself can prime a detrimental vicious circle in tumorigenesis.14 The exact proportions by which CHIP and “inflammaging” actually contribute to unexplained anemia in the elderly deserve further studies. Similarly, whether or not chronic subclinical inflammation, which is difficult to detect using classical biomarkers such as C-reactive protein, contributes to anemia through the same mechanisms as those associated with overt inflamma417


Editorials

Figure 1. The complex relationship among aging, inflammation, and anemia. Clonal hematopoiesis and “inflammaging” are increasingly recognized as age-related conditions (see the text). Both are able to induce a mild chronic pro-inflammatory status, and can influence each other in a vicious circle. For example, chronic inflammation in the bone marrow can favor the development of somatic mutations in hematopoietic stem cells (dotted arrow). Clonal hematopoiesis and “inflammaging” are plausible explanations for at least a fraction of unexplained anemia in the elderly, likely through mechanisms similar to those involved in the pathophysiology of the classical anemia of overt inflammation. It is worth noting that clonal hematopoiesis is also able to induce anemia because of ineffective erythropoiesis. When cytopenia occurs in association with clonal hematopoiesis, it is termed clonal cytopenia of undetermined significance. Anemia in the elderly is strongly and independently associated with adverse outcomes, but whether or not this relationship is actually causal remains to be demonstrated. CV: cardiovascular; CCUS: clonal cytopenia of undetermined significance.

tion (e.g. hepcidin-driven iron-restricted erythropoiesis and cytokine-mediated suppression of erythropoiesis) remains to be demonstrated. Of note, clonal hematopoiesis can also cause anemia because of ineffective erythropoiesis, since progenitors carrying certain mutations can have a proliferative advantage over normal hematopoietic stem cells, but are less able to differentiate adequately.15 Indeed, clonal hematopoiesis has been identified in a high proportion (64%) of elderly with unexplained cytopenia, for example with anemia not fulfilling all the morphological and cytogenetic criteria for the diagnosis of myelodysplastic syndrome.16 In such cases, the term “clonal cytopenia of undetermined significance” has been proposed.9 Finally, a key point to be kept in mind when thinking about anemia in the elderly is that its robust association with mortality, even independently of the many possible 418

confounders, does not prove a causal link, because of inherent limitations of observational epidemiology.2 The only way to prove causality definitively would be the reduction of mortality and/or other adverse outcomes after successful correction or amelioration of anemia. The increasing number of promising anti-anemic drugs that are entering the clinical arena, including hepcidin antagonists,17 novel iron formulations,18 activin receptor IIA ligand trap,19 and hypoxia inducible factor stabilizers,20 may represent an unprecedented opportunity to clarify this crucial point in the near future.

References 1 Mannino RG, Myers DR, Tyburski EA, et al. Smartphone app for non-invasive detection of anemia using only patient-sourced photos. Nat Commun. 2018;9(1):4924.

haematologica | 2019; 104(2)


Editorials 2. Girelli D, Marchi G, Camaschella C. Anemia in the elderly. HemaSphere. 2018;2(3):e40. 3. Erikson GA, Bodian DL, Rueda M, et al. Whole-genome sequencing of a healthy aging cohort. Cell. 2016;165(4):1002-1011. 4. Wouters H, van der Klauw MM, de Witte T, et al. Association of anemia with health-related quality of life and survival: a large population-based cohort study. Haematologica. 2019;104(3):468-476. 5. Barnett K, Mercer SW, Norbury M, Watt G, Wyke S, Guthrie B. Epidemiology of multimorbidity and implications for health care, research, and medical education: a cross-sectional study. Lancet. 2012;380(9836):37-43. 6. Busti F, Campostrini N, Martinelli N, Girelli D. Iron deficiency in the elderly population, revisited in the hepcidin era. Front Pharmacol. 2014;5:83. 7. Weiss G, Ganz T, Goodnough LT. Anemia of inflammation. Blood. 2018 Nov 6 [Epub ahead of print]. 8. Shavelle RM, MacKenzie R, Paculdo DR. Anemia and mortality in older persons: does the type of anemia affect survival? Int J Hematol. 2012;95(3):248-256. 9. Steensma DP. Clinical consequences of clonal hematopoiesis of indeterminate potential. Blood Adv. 2018;2(22):3404-3410. 10. 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. 11. Libby P, Ebert BL. CHIP (clonal hematopoiesis of indeterminate potential). Circulation. 2018;138(7):666-668. 12. Franceschi C, Bonafe M, Valensin S, et al. Inflamm-aging. An evolu-

13. 14. 15. 16. 17. 18. 19.

20.

tionary perspective on immunosenescence. Ann N Y Acad Sci. 2000;908:244-254. Youm YH, Grant RW, McCabe LR, et al. Canonical Nlrp3 inflammasome links systemic low-grade inflammation to functional decline in aging. Cell Metab. 2013;18(4):519-532. Grivennikov SI, Greten FR, Karin M. Immunity, inflammation, and cancer. Cell. 2010;140(6):883-899. Qu X, Zhang S, Wang S, et al. TET2 deficiency leads to stem cell factor-dependent clonal expansion of dysfunctional erythroid progenitors. Blood. 2018;132(22):2406-2417. Malcovati L, Galli A, Travaglino E, et al. Clinical significance of somatic mutation in unexplained blood cytopenia. Blood. 2017;129(25):3371-3378. Crielaard BJ, Lammers T, Rivella S. Targeting iron metabolism in drug discovery and delivery. Nat Rev Drug Discov. 2017;16(6):400423. Girelli D, Ugolini S, Busti F, Marchi G, Castagna A. Modern iron replacement therapy: clinical and pathophysiological insights. Int J Hematol. 2018;107(1):16-30. Platzbecker U, Germing U, Gotze KS, et al. Luspatercept for the treatment of anaemia in patients with lower-risk myelodysplastic syndromes (PACE-MDS): a multicentre, open-label phase 2 dosefinding study with long-term extension study. Lancet Oncol. 2017;18(10):1338-1347. Gupta N, Wish JB. Hypoxia-inducible factor prolyl hydroxylase inhibitors: a potential new treatment for anemia in patients with CKD. Am J Kidney Dis. 2017;69(6):815-826.

Targeting a major hub of cell fate decisions â&#x20AC;&#x201C; the mitochondrial-associated membrane William L. Carroll and Nikki A. Evensen Departments of Pediatrics and Pathology, Perlmutter Cancer Center, NYU-Langone Medical Center, New York, NY, USA E-mail: WILLIAM L. CARROLL - william.carroll@nyumc.org doi:10.3324/haematol.2018.208355

G

reat progress has been made in the treatment of human cancer but, unfortunately, there remains an abundance of treatment failures due to primary therapy resistance and/or the emergence of drug refractory clones. This is certainly true for acute leukemias, which are the most common malignancies in children and while they make up a smaller fraction of cancer in adults, their impact is substantial given the poorer outcome. Furthermore, the high doses of cytotoxic agents that are often used in therapy are associated with short- and long-term side effects. Thus, there is an urgent need to develop novel therapeutic approaches to improve outcome and decrease side effects. One such strategy, taken by Koczian and colleagues and described in this issue of Haematologica, is to augment the effectiveness of conventional agents.1 They report that the use of the small molecule inhibitor of protein disulfide isomerase (PDI), PS89, has a significant impact on the effectiveness of cytostatic agents used routinely in the therapy of acute leukemias. The model that emerges is that PS89 amplifies the apoptotic stimulus induced by cytotoxic therapy, thereby allowing for increased efficacy at lower doses, through modulation of proteins at the mitochondrial-endoplasmic reticulum (ER) interface. The agent itself has poor pharmacokinetic properties limiting in vivo examination, but the results indicate substantial benefit and a wide therapeutic index. Much work remains to be done but the results emphasize the opportunity to target haematologica | 2019; 104(3)

a unique intracellular sub-compartment that plays a key role in cell fate decisions: the interface between the ER and mitochondria. Mitochondria are multifaceted organelles responsible for an array of cell functions critical for energy production, redox balance, adaptation to cell stress, and activation of the intrinsic apoptotic pathway. They make up 20% of the cytoplasmic volume of a cell and are dynamic, motile structures constantly altering shape through fission and fusion. These alterations involve two lipid bilayers that make up the inner membrane forming cristae (containing membrane-bound enzymes involved in oxidative phosphorylation) which enclose the matrix, and the smooth outer membrane. Mitochondria make important contact with other organelles, particularly the ER, which is in direct contacts with 20% of the mitochondrial surface. Changes in energy metabolism related to cancer, the so-called Warburg effect, have received renewed interest, especially with the discovery of â&#x20AC;&#x153;oncometabolitesâ&#x20AC;?, but changes that occur at the mitochondrial-ER interface are also critical in controlling mitochondrial metabolism and cell fate decisions.2 The mitochondrial-ER interface, commonly referred to as the mitochondria-associated membrane (MAM), is a proteinaceous tether facilitating bidirectional communication between the two organelles controlling the balance between survival and death.3,4 The exchange of metabolites and contact at the interface controls energy produc419


Editorials

Figure 1. The mitochondrial associated membrane. Hundreds of proteins operate at the mitochodrial-associated membrane (MAM) but the location of proteins discussed in this editorial are illustrated here. A major route of communication between the endoplasmic reticulum (ER) and mitochondria occurs through the release of ER calcium via the inositol 1,4,5-trisphosphate receptor (IP3R), the volatage-gated anion channel (VDAC) and the chaperone GRP75. Calcium gains access to the matrix through the mitochondrial Ca2+ uniporter (MCU) leading to membrane depolarization and cytochrome C release as part of the apoptotic pathway. This exchange of calcium is triggered by the interaction of fission protein 1 (FIS1), B-cell receptor-associated protein 31 (BAP31), and caspase 8 within the MAM upon apoptotic stimuli, such as combination treatment with etoposide and PS89.

tion, mitochondrial shape and, importantly, apoptosis. One of the most important regulators of mitochondrial energy production and cell death is the release of ER calcium into the mitochondria thereby facilitating the opening of the permeability transition pore at the inner mitochondrial membrane, depolarization, and cytochrome C release. This process is orchestrated by a vast network of proteins that could serve as potential targets as evidenced by their alteration in cancer cells as a means of escape from chemotherapy-induced apoptosis that relies on Ca2+ signaling.5 A number of proteins appear to stabilize ER-mitochondria contact thereby prolonging calcium flux including mitofusin-2, phosphofurin acidic cluster sorting protein 2 (PACS-2), and double-stranded RNA-activated protein (PKR)-like ER kinase (PERK) among others. Modulation of such contacts and crosstalk may facilitate apoptosis in cancer cells particularly since the Bcl-2 family of proteins reside and interact at the MAM, highlighted by the increasing interest in BH3-mimetics that alter proteins of this interconnected network. Likewise, a number of oncoproteins and tumor suppressors such as p53, PTEN and AKT function locally. For example, p53 is enriched on the MAM where it interacts with the ER Ca2+ pump SERCA to boost ER-mitochondrial Ca2+ flux and apoptosis.6 420

In addition to Ca2+ storage/signaling, the ER also plays a major role in protein synthesis, post-translational modifications and folding. ER homeostasis is a delicate balance and when the folding machinery can no longer keep up with protein synthesis an adaptive response called the unfolded protein response (UPR) occurs. This response seeks to restore balance by attenuating protein translation, upregulating ER protein degradation and increasing the level of chaperone proteins to inflate protein folding capacity.7 When the UPR fails to restore ER proteostasis, the pathway shifts to promote cell death primarily though the PERK branch of the UPR. In this way, the ER functions as a sensor of protein stress and perturbations lead to the induction of a variety of survival/death pathways, many of which rely on crosstalk with the mitochondria. Correct folding of many proteins (e.g. 80% of secretory proteins) requires disulfide (S-S) bonds between cysteine residues. The PDI family of proteins is responsible for the formation and rearrangement of protein disulfide bonds and these ER-resident enzymes also function as chaperones independently of their role in disulfide bond formation.8 Therefore, these proteins are essential in maintaining protein homeostasis at baseline and during the UPR. Numerous studies have indicated that PDIA1 (gene name, haematologica | 2019; 104(2)


Editorials

aka PDI), encoding the archetype PDI protein, and other members of the family are upregulated in many human tumors, correlate with invasiveness (metastasis) and, in some cases, may confer therapy resistance.8,9 These findings, along with the fact that the interplay between the ER and mitochondria is critical for survival versus death, motivated the authors and many other investigators to examine the role of PDI inhibitors in cancer treatment. PS89 is a derivative of a lead compound discovered by the authors using a screen for chemosensitization of etoposide-induced apoptosis in a variety of cell lines.10 This initial work showed that PS89 is a reversible inhibitor of PDI and induced the UPR. However, in the current investigation, genetic silencing of PDI failed to recapitulate PS89 activity seen in leukemia cell lines and overexpression did not rescue the chemosensitizing effect. This launched the search for additional targets. It should be noted however that these experiments do not rule out an impact of the UPR in mediating PS89’s effect or its modulation of other PDI family members. Indeed, activity-based protein profiling performed by these investigators did show that other PDI family proteins are targets of PS89 as are other resident ER proteins. B-cell receptor-associated protein 31 (BAP31) was one of the most prominent proteins identified as a PS89 binding partner and is known to operate at the MAM. BAP31 is located at the ER membrane and is tethered to mitochondria through mitochondrial fission protein 1 (FIS1), which appears to serve as a platform for procaspase 8. Previous work has shown that apoptotic signals originating from the mitochondria lead to cleavage of BAP31 into the pro-death p20BAP31 fragment, thereby activating caspase 8 and launching/amplifying apoptosis.11 Kozcian and colleagues showed that etoposide or PS89 alone had only a modest impact on activation of caspase 8 and other downstream effector pathways, but the combination was highly effective at inducing activity. Moreover, they showed that the combination triggered increased ERassociated calcium influx, loss of mitochondria membrane integrity, cytochrome C release, and increased levels of reactive oxygen species. Other studies have shown that BAP31 plays a critical role in mediating ER stressrelated apoptosis through interaction with cell death inducing p53 target 1 (CDIP) resulting in sequestration of BCL-2 and activation of BAX and caspase 8.12 Thus, BAP31 emerges as a hub for integrating a variety of apoptotic signals from both organelles. Furthermore, their work supports continued efforts to target this interconnected network with the aim of amplifying apoptotic signals. The authors emphasize the concept of “network pharmacology” in developing potential strategies for the development of novel cancer therapeutics. Given the myriad roles of PDI family members in cell homeostasis, inhibitors of these enzymes are likely to operate through multiple targets in orchestrating their impact in a variety of model systems. For example, two members of the family have opposing effects on the activation of PERK which, as mentioned, plays a key role in the UPR.13 PERK

haematologica | 2019; 104(3)

inhibitors have been associated with unacceptable side effects and PDI inhibitors might provide an alternative, less toxic way to inhibit the ER’s ability to maintain homeostasis under stress. Likewise, a recent paper also published in Hematologica described that a thioredoxin inhibitor, SK053, promoted differentiation and apoptosis in acute myeloid leukemia cells.14 As in the experiments by Koczian et al., knock down of the target failed to replicate the phenotype also motivating the authors to search for other targets. In fact, they showed that SK053 binds to and inhibits PDI (i.e., the product of the PDIA1 gene). Previous work had shown that PDI interacts with a stem loop structure of the CCAAT enhancer-binding protein α (C/EBPα) and blocks its translation. Expression of C/EBPα is critical for myeloid differentiation and its activity is often corrupted in acute myeloid leukemia. PDI inhibition overcomes the translational block leading to increased C/EBPα levels, differentiation and apoptosis in acute myeloid leukemia. Thus, the observations by Koczian and others highlight the potential utility of targeting the MAM, a dynamic scaffold that plays a critical role in a variety of biological processes, using PDI inhibition or other approaches to augment a variety of cancer therapies.

References 1 Koczian F, Naglo O, Vomacka J, et al. Targeting the endoplasmic reticulum-mitochondria interface sensitizes leukemia cells to cytostatics. Haematologica. 2019;104(3):546-555. 2. Marchi S, Patergnani S, Pinton P. The endoplasmic reticulum-mitochondria connection: one touch, multiple functions. Biochim Biophys Acta. 2014;1837(4):461-469. 3. Herrera-Cruz MS, Simmen T. Cancer: untethering mitochondria from the endoplasmic reticulum? Front Oncol. 2017;7(105). 4. Prudent J, McBride HM. The mitochondria-endoplasmic reticulum contact sites: a signalling platform for cell death. Curr Opin Cell Biol. 2017;47:52-63. 5. Kerkhofs M, Bittremieux M, Morciano G, et al. Emerging molecular mechanisms in chemotherapy: Ca(2+) signaling at the mitochondriaassociated endoplasmic reticulum membranes. Cell Death Dis. 2018;9(3):334. 6. Giorgi C, Bonora M, Missiroli S, et al. Alterations in mitochondrial and endoplasmic reticulum signaling by p53 mutants. Front Oncol. 2016;6:42. 7. van Vliet AR, Agostinis P. Mitochondria-associated membranes and ER stress. Curr Top Microbiol Immunol. 2018;414:73-102. 8. Lee E, Lee DH. Emerging roles of protein disulfide isomerase in cancer. BMB Rep. 2017;50(8):401-410. 9. Xu S, Sankar S, Neamati N. Protein disulfide isomerase: a promising target for cancer therapy. Drug Discov Today. 2014;19(3):222-240. 10. Eirich J, Braig S, Schyschka L, et al. A small molecule inhibits protein disulfide isomerase and triggers the chemosensitization of cancer cells. Angew Chem Int Ed Engl. 2014;53(47):12960-12965. 11. Iwasawa R, Mahul-Mellier AL, Datler C, Pazarentzos E, Grimm S. Fis1 and Bap31 bridge the mitochondria-ER interface to establish a platform for apoptosis induction. EMBO J. 2011;30(3):556-568. 12. Namba T, Tian F, Chu K, et al. CDIP1-BAP31 complex transduces apoptotic signals from endoplasmic reticulum to mitochondria under endoplasmic reticulum stress. Cell Rep. 2013;5(2):331-339. 13. Kranz P, Neumann F, Wolf A, et al. PDI is an essential redox-sensitive activator of PERK during the unfolded protein response (UPR). Cell Death Dis. 2017;8(8):e2986. 14. Chlebowska-Tuz J, Sokolowska O, Gaj P, et al. Inhibition of protein disulfide isomerase induces differentiation of acute myeloid leukemia cells. Haematologica. 2018;103(11):1843-1852.

421


Editorials

Learning the next-generation sequencing alphabet of immune reconstitution: factors determining CD8+ T-cell receptor α-chain repertoire dynamics after hematopoietic stem cell transplantation Esteban Arrieta-Bolaños1 and Katharina Fleischhauer1,2 1

Institute for Experimental Cellular Therapy, University Hospital Essen and 2German Cancer Consortium, Heidelberg, Germany

E-mail: KATHARINA FLEISCHHAUER - katharina.fleischhauer@uk-essen.de doi:10.3324/haematol.2018.209130

I

n this issue of Haematologica, Link-Rachner et al.1 report their findings on CD8+ T-cell receptor-alpha (TRα) chain dynamics in patients after hematopoietic stem cell transplantation (HSCT) using next-generation sequencing (NGS) in relation to different treatment platforms. Their study aimed at unraveling the effect of posttransplant T-cell-depleting immunosuppressive therapy (namely anti-thymocyte globulin, ATG, and post-transplant cyclophosphamide, PTCy) and degree of HLA matching on the TRα diversity of naïve and memory CD8+ T-cell repertoires reconstituting the patient’s peripheral blood in the first six months after transplantation. Furthermore, the authors attempted to determine the extent to which the donor’s TRα repertoire influences the post-transplant repertoires in the respective patients. The TR and its huge variability is one of the pillars of adaptive immunity. Healthy, diverse TR repertoires in normal individuals contain millions of different clones with unique TRs,2 and provide the immune system with an arsenal of highly specific, yet also cross-reactive cells to fight off pathogens and malignant cells. Until recently, this extreme diversity limited a detailed and deep analysis of TR repertoires in healthy and pathological conditions, with most available techniques focusing on broad repertoire alterations, and extensive, cumbersome T-cell cloning required to investigate specific complementaritydetermining region 3 (CDR3) variants. The advent of NGS-based high-throughput analysis of TRs has revolutionized the field of immune repertoire analysis.3 TR NGS can now provide qualitative and quantitative information on hundreds of thousands of different T-cell clones directly from a single blood or tissue sample. Hematopoietic stem cell transplantation is a field in which TR analysis is of extreme interest, both for medical and biological reasons. Patients undergoing allogeneic HSCT see the partial or nearly total elimination of their own hematopoietic system with radio- and/or chemotherapy followed by its replacement with that of a donor. In most cases, HSCT is the only curative therapy for the underlying disease. However, HSCT poses several risks for the patient, many of which derive from the ablation of their bone marrow and the concomitant risk of infection and pathogen reactivation. In addition, the new hematopoietic system can induce graft-versus-host disease (GvHD) associated with tissue and organ damage and, in some cases, death.4 T-cell reconstitution dynamics is central to these post-transplant immune processes and represents an area of intense research in the HSCT field. High-throughput TR NGS has thus quickly attracted researchers eager to use its power to study post-HSCT T422

cell clonal dynamics, its relationship to transplant-related factors, and its role in transplant complications.5 The study by Link-Rachner et al. contains a number of noteworthy aspects. Contrary to most previous reports, the researchers focus on the TRα chain. Most published TR analyses have studied the TRb chain, probably because it is considered more diverse due to the added combinatorial potential conferred by the D segment, but also perhaps on account of the fact that TR NGS methodologies for this chain are more extended. However, both the α and the b chains contribute to TR specificity and, because of the order of the recombination events during T-cell maturation and development, one mature T cell can express two different functional α chains, both of which can pair with the cell’s b chain, forming two different TR heterodimers.6 Indeed, this happens in approximately 10% of the T cells in peripheral blood, and TRα CDR3 diversity has actually been observed to be 1.2-2.4 times greater than TRb in T-cell subsets from a single individual.7 Importantly, these ‘dual’ T cells have been associated with increased autoimmune and alloreactive capacities,8 as well as with acute9 and chronic10 GvHD. Hence, analysis of TRα chains in HSCT is of special interest, and this study should encourage researchers to pay more attention to this locus in future analyses. Furthermore, the authors designed their study in order to perform a comparison of TR dynamics between different clinical platforms. By including in their analysis patients transplanted from HLA-matched and -mismatched unrelated donors treated or not treated with ATG, haploidentical donors treated with PTCy, and related donors with no T-cell depletion (TD), the authors provide valuable insights into how these clinical platforms differ in terms of T-cell and TR dynamics during immune reconstitution at different time points in the first six months after HSCT. Despite the limited sample size, each of the groups was analyzed separately, leading to some interesting observations. First, the data suggest that TD has a stronger effect on the reconstitution of naïve than of memory T cells in terms of numbers and TRα diversity; both were significantly lower in patients treated with TD (Figure 1). Considering that diversity is deemed a hallmark of a ‘healthy’ T-cell repertoire,11 this raises questions as to ways in which one could improve TR diversity in patients undergoing HSCT platforms with in vivo TD, as already suggested in the context of TRαb-depleted grafts.12 In addition, the data from Link-Rachner et al. provide some intriguing new insights into the relative contribution of the donor’s memory and naïve T-cell repertoires to haematologica | 2019; 104(3)


Editorials

immune reconstitution after HLA-identical sibling HSCT in the absence of TD. The authors find that 60% and 11.5% of the TRα clonotypes in the patients’ memory and naïve T cells at day 180 post transplantation could be traced back to the donors’ memory repertoires, compared to only 30% and 15% to the donors’ naïve repertoires, respectively (Figure 2). These data suggest that the donor’s memory compartment is significantly reflected not only in the memory but also to a certain extent in the naïve CD8+ compartment after HLA-identical sibling HSCT. Given the high (>99%) purity of the FACS-sorted naïve and memory T-cell subsets used for the experiments, these intriguing findings are unlikely to have been impacted by sample spill-over. It should be noted that 2 of the 5 donors used for these analyses were cytomegalovirus (CMV) seropositive, while the remain-

ing 3 donors were seronegative. In transplants performed under PTCy regimen, donor CMV seropositivity has been shown to correlate significantly with a predominance of donor CD8+ memory T-cell reconstitution post transplantation.13 The data from Link-Rachner et al. suggest that this may hold true also for CMV seronegative donors in the non-TD setting, although a separate analysis of a larger number of seropositive and seronegative donors will be needed to verify this point. The data also have potential practical implications. If the donor’s memory repertoire plays a leading role in shaping the patient’s repertoires after transplantation, the ‘quality’ of that donor memory repertoire might be assessed before transplant as a factor to be considered in donor selection, or as a prognostic tool for post-transplant complications. This observation becomes relevant also in

Figure 1. Impact of T-cell depletion (TD) on the size and diversity of the reconstituting naïve and memory CD8+ T-cell repertoires at six months post hematopoietic stem cell transplantation (HSCT). A schematic representation of the donors’ (left) and the patients’ naïve (N) or memory (M) CD8+ T-cell repertoires at day 180 post transplantation (right), in the presence (TD) or absence (No TD) of TD by either anti-thymocyte globulin or post-transplant cyclophosphamide. The size of the bubbles indicates the approximate relative size of the repertoires. Within each bubble, CD8+ T cells of identical T-cell receptor-α clonotypes are indicated by identical colors. A significant shrinking of both size and diversity of the naïve CD8+ T-cell repertoire occurs in the patient with TD compared to the patient without TD.

haematologica | 2019; 104(3)

423


Editorials

view of the interest in the depletion of naïve T cells from stem cell grafts in the related donor setting.14 However, further research will be needed to understand if this occurs also on the other HSCT platforms, including cord blood transplantation,15 and what the desired qualities of a donor repertoire are.

Figure 2. Relative contribution of donor memory or naïve CD8+ T cells to the patients’ repertoires at six months post hematopoietic stem cell transplantation (HSCT). The average percentage of T-cell receptor-α clonotypes in memory (M; dark orange) or naïve (N; light orange) CD8+ T cells shared between 5 donors (left) and the respective patients at day 180 after HLA-identical sibling HSCT in the absence of T-cell-depleting immunosuppressive therapy. Shaded blue areas indicate shrinking or expansion of the relative percentage of shared clonotypes in the donor compared to the patient post transplantation. Non-shared clonotypes are indicated in gray. Note that an average of approximately 60% and 11.5% of the patients’ memory and naïve repertoires, respectively, could be traced back to the donors’ memory repertoires. In contrast, an average of only approximately 30% and 15% of the patients’ memory and naïve repertoires, respectively, could be traced back to the donors’ naïve repertoires. This figure uses data from Figure 2A of the study by Link-Rachner et al.1

424

The study by Link-Rachner et al. does leave some open questions that warrant further study. First, patients who relapsed were explicitly excluded from the study. However, relapse remains the main cause of treatment failure in HSCT,16 and, similar to GvHD, a central role for T cells in the therapeutic graft-versus-leukemia effect has been well established.17 Hence, NGS-based studies of T-cell repertoire characteristics that might associate with leukemia relapse after transplantation are of the foremost importance. Second, in this study, CD4+ T-cell reconstitution and repertoire dynamics were not analyzed, yet they are likely to play a central role in patients after HSCT, especially in clinical contexts where HLA-DPB1 mismatches are frequent (e.g. HSCT with unrelated donors).18 Of note, the well-established permissiveness of a proportion of these HLA-DPB1 mismatches and its relationship with TR repertoire characteristics is also of interest and this is currently under investigation.19 The delayed reconstitution of this Tcell compartment might pose methodological challenges for TR repertoire analyses, but that and its central role in the co-ordination of effective immune, as well as alloreactive responses, warrant special attention. Finally, while Link-Rachner et al. focus their attention on the TRα repertoire post HSCT, it would be interesting to assess whether the TRα and TRb repertoires follow similar, complementary, or independent dynamics after different transplant settings. Parallel analysis of both chains, and potentially even attempting to determine their pairing with recent novel high-throughput approaches,20 is likely to give an even more complete picture of TR immune reconstitution with clinical and translational relevance. Overall, the study by Link-Rachner et al. illustrates how the power of NGS has revolutionized the assessment of immune repertoires and opened a broad spectrum of possibilities for unprecedented analysis of T-cell dynamics in the field of HSCT. Future studies building on this and other earlier pioneering work in this still developing field are called for to fully embrace this potential to enrich our understanding of T-cell immune reconstitution after HSCT. This knowledge should ultimately serve the objective of promoting the regeneration of a healthy TR repertoire that contributes both to the eradication of the underlying disease and to the control of transplant-related morbidities, maximizing the therapeutic potential of allogeneic HSCT. Acknowledgments This contribution was supported by grants from the Deutsche José Carreras Leukämie Stiftung (DJCLS R 15/02 and DJCLS R/2017), Dr. Werner Jackstädt Stiftung, and the Josef Senker Stiftung.

References 1 Link-Rachner CS, Eugster A, Rucker-Braun E, et al. T cell receptor alpha repertoire of CD8+ T cells following allogeneic stem cell transplantation using next-generation sequencing. Haematologica. 2019;104(3):622-631. 2. Qi Q, Liu Y, Cheng Y, et al. Diversity and clonal selection in the human T-cell repertoire. Proc Natl Acad Sci U S A. 2014;111(36): 13139-13144. 3. Rosati E, Dowds CM, Liaskou E, Henriksen EKK, Karlsen TH, Franke A. Overview of methodologies for T-cell receptor repertoire analysis.

haematologica | 2019; 104(3)


Editorials BMC Biotechnol. 2017;17(1):61. 4. Ferrara JL, Levine JE, Reddy P, Holler E. Graft-versus-host disease. Lancet. 2009;373(9674):1550-1561. 5. Warren EH, Matsen FA 4th, Chou J. High-throughput sequencing of B- and T-lymphocyte antigen receptors in hematology. Blood. 2013;122(1):19-22. 6. Padovan E, Casorati G, Dellabona P, Meyer S, Brockhaus M, Lanzavecchia A. Expression of two T cell receptor alpha chains: dual receptor T cells. Science. 1993;262(5132):422-424. 7. Wang C, Sanders CM, Yang Q, et al. High throughput sequencing reveals a complex pattern of dynamic interrelationships among human T cell subsets. Proc Natl Acad Sci U S A. 2010;107(4):15181523. 8. Ni PP, Solomon B, Hsieh CS, Allen PM, Morris GP. The ability to rearrange dual TCRs enhances positive selection, leading to increased Allo- and Autoreactive T cell repertoires. J Immunol. 2014;193(4):1778-1786. 9. Morris GP, Uy GL, Donermeyer D, Dipersio JF, Allen PM. Dual receptor T cells mediate pathologic alloreactivity in patients with acute graft-versus-host disease. Sci Transl Med. 2013;5(188): 188ra174. 10. Balakrishnan A, Gloude N, Sasik R, Ball ED, Morris GP. Proinflammatory Dual Receptor T Cells in Chronic Graft-versusHost Disease. Biol Blood Marrow Transplant. 2017;23(11):18521860. 11. Attaf M, Huseby E, Sewell AK. alphabeta T cell receptors as predictors of health and disease. Cell Mol Immunol. 2015;12(4):391-399. 12. Zvyagin IV, Mamedov IZ, Tatarinova OV, et al. Tracking T-cell

13. 14. 15.

16. 17. 18. 19.

20.

immune reconstitution after TCRalphabeta/CD19-depleted hematopoietic cells transplantation in children. Leukemia. 2017;31 (5):1145-1153. Kanakry CG, Coffey DG, Towlerton AM, et al. Origin and evolution of the T cell repertoire after posttransplantation cyclophosphamide. JCI Insight. 2016;1(5):e86252. Bleakley M, Heimfeld S, Loeb KR, et al. Outcomes of acute leukemia patients transplanted with naive T cell-depleted stem cell grafts. J Clin Invest. 2015;125(7):2677-2689. Gkazi AS, Margetts BK, Attenborough T, et al. Clinical T Cell Receptor Repertoire Deep Sequencing and Analysis: An Application to Monitor Immune Reconstitution Following Cord Blood Transplantation. Front Immunol. 2018;9:2547. 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. Negrin RS. Graft-versus-host disease versus graft-versus-leukemia. Hematology Am Soc Hematol Educ Program. 2015;2015:225-230. Fleischhauer K, Shaw BE. HLA-DP in unrelated hematopoietic cell transplantation revisited: challenges and opportunities. Blood. 2017;130(9):1089-1096. Arrieta-Bolanos E, Crivello P, Metzing M, et al. Alloreactive T Cell Receptor Diversity against Structurally Similar or Dissimilar HLADP Antigens Assessed by Deep Sequencing. Front Immunol. 2018;9:280. Howie B, Sherwood AM, Berkebile AD, et al. High-throughput pairing of T cell receptor alpha and beta sequences. Sci Transl Med. 2015;7(301):301ra131.

Mosaicism by somatic non-functional mutations: one cell lineage at a time Willy Albert Flegel1,2 1

Department of Transfusion Medicine, NIH Clinical Center, National Institutes of Health, Bethesda, MD and 2Department of Pathology, Georgetown University Medical Center, Washington, DC, USA E-mail: WILLY ALBERT FLEGEL - waf@nih.gov

doi:10.3324/haematol.2018.208165

S

omatic mutations are abundant in most cells of our tissues.1,2 The impact of any somatic mutation may be small and temporary if it occurs in differentiated cells without giving rise to malignant growth by unlocking their terminal differentiation. We expect a much wider and lasting impact from the same somatic mutations if they occur during earlier steps of differentiation. If they arise in a stem or progenitor cell, a whole set of cells (the lineage downstream of the progenitor) will be affected. Any such novel somatic mutation may stay with the individual for life, when different somatic mutations can accumulate in cell lineages over time.3 The combination of somatic mutations becomes not only part of our lives, but contributes to our phenotype.4 The variety of somatic mosaicism that will develop differs randomly among individuals and forms the basis of our healthy constitutions as well as of some medical conditions. In this issue of the Journal, Dauber et al. present a technically advanced study5 describing the molecular causes of mosaicism in 2 patients. The antigens of the Rhesus blood group system (ISBT 004) served as markers allowing the authors to take advantage of routine serology and detect affected individuals. The patients with red cell mosaicism were identified by the loss of the c antigen (RH4) in a subset of their red cells. The causes of this serological phenotype were traced to distinct precursor stages of myeloid and pluripotent stem cells, respectively.5 As the antigens were only the markers and not the focus, this study, at the interhaematologica | 2019; 104(3)

section of ‘erythropoiesis gone wrong’ and red cell antigens, has relevance beyond blood groups and offers valuable information to the hematology community. Red cell mosaicism was documented in leukemia for ABO antigens in 1957,6,7 and for Rh a few years later;8 it has been documented for various blood groups many times since then.5,9,10 During routine blood group typing worldwide, immunohematologists repeatedly encounter the incidental finding of a ‘mixed field’ agglutination or a discrepancy with previous results in the patient’s health record. In both patients of the current study,5 ‘mixed field’ agglutination was observed and prompted further investigation. Barring technical issues with the serology, such findings are infrequent, although not rare, in the absence of transfusions. Many patients could be identified with mosaicism that cannot be explained simply by recent transfusion, and these patients could then be followed up in order to evaluate the clinical implications. Possible causes are loss of heterozygosity (LOH) associated with loss of tumor suppressor gene functions,11 or copy-neutral LOH associated with the gain of oncogenic mutations.12,13 Despite the discrepancy in blood group, the immediate transfusion support given to these patients is usually straightforward. Hence, further analysis is not considered necessary within the practical approach to clinical care. The clinical prognosis of LOH in hematologically asymptomatic patients is currently unknown, although the authors point out that this may now be changing. The diagnosis of red 425


Editorials

cell mosaicism could be important for the patients, who are not regularly informed of the incidental finding. Because the technologies required to confirm mosaicism lie far outside routine procedures, they are not available to clinical laboratories because of the limited resources available. This was not the case for these 2 patients5 whose pathophysiology was examined in unusual detail. The current study5 represents a step forward in this field, as the exact types of precursors, affected by distinct somatic mutations, have been determined. LOH was, indeed, the underlying cause in both patients, affecting the RH locus on the short arm of chromosome 1 and encompassing at least 26.7 and 42.4 Mb, respectively. Such large deletions of parts of a chromosome are considered rare in patients without malignancies or other hematologic pathologies. The RHCE gene spans only 60 kb and an RHCE deletion is still extremely rare in genomic DNA. As both large deletions encompassed the RHD/RHCE gene loci, they explained the serological phenotype with the loss of the c antigen in approximately 50% of the patientsâ&#x20AC;&#x2122; red cells (Figure 1).

Besides the deletion of a chromosomal segment or a whole gene, less intrusive mechanisms, all the way down to 1 single nucleotide polymorphism (SNP), can functionally mimic the same LOH for RHCE, which prompted the authors to study these 2 patients. Like any DNA variation, somatic mutations,14 if found in exons, can affect the protein by missense and non-functional mutations, and less so by silent mutations. A novel SNP in a somatic cell lineage occurs much more frequently than any huge alteration affecting long DNA segments,5 but can be as important for the blood group phenotype.15 A SNP may also have an impact on gene regulation when occurring in the introns, the 3â&#x20AC;&#x2122;- or 5â&#x20AC;&#x2122;-untranslated regions of a gene, or somewhere in the long DNA segments interspersed between genes. Genes differ in their propensity to lose their function,16 as do the mechanism causing changes, ranging from an SNP in an RH gene,17 to large deletions affecting the RH gene locus18,19 and adjacent genes.5,13 Somatic mosaicism is always acquired and occurs during mitosis in 1 cell lineage derived from 1 zygote, often after

Figure 1. Schematic representation of differentiation from early stem to red cells. A model for normal hematopoiesis (top) assumes equal cell division at all stages (lines with open circles). In the models for early (middle) and late (bottom) mosaicism, the red cells eventually reach 50% clonality. Some cell lineages are squeezed out, possibly experiencing apoptosis (closed circles) before differentiating into red cells. This 50% mosaicism represents the red cell populations actually observed in the peripheral blood of the 2 patients in the study by Dauber et al.5 reported in this issue of the Journal.

426

haematologica | 2019; 1043)


Editorials

many cell divisions during aging. Most acquired chimerism is iatrogenic and takes place after allogeneic solid organ or hematopoietic progenitor cell transplantations. Diagnostic laboratories routinely come across such patients when testing peripheral blood, and hematologists and other consultants should be willing to share this clinically quite important patient history with the laboratory. In comparison, congenital chimerism, as sometimes found by blood group typing,19 is much less common. The diagnostic hallmark for chimerism, whether congenital or acquired, are 2 entirely different genotypes derived from 2 zygotes and found throughout the genome in 2 sets of cells or tissues. In contrast, a typical somatic mosaicism is limited to only 1 chromosomal location. In depth genotyping at the single cell level may eventually reveal a host of mosaicisms in individuals and even in a set of cells that is presumed to belong to 1 cell lineage. Certain congenital diseases predispose to somatic mutations, such as in fragile X syndrome. Conversely, the highly individual accumulation of somatic mutations causes the great variety of phenotype in dyskeratosis congenita. The exact combination of mosaicism accrued throughout life in a variety of cell lineages constitutes the substrate of single cell medicine. Current progress in gene editing will allow somatic mutations in differentiating cells to be modified or corrected; this will raise few ethical concerns since the germline is not affected. It was possible to clarify more molecular details and the regulatory mechanism was addressed in both of these patients. The exact break points involved in the deletion events, as defined by nucleotide sequence, can be determined with current technology. This would also allow the claim that the proposed CDe haplotype is duplicated in the affected cell lineages to be corroborated. While the patterns may seem random, close inspection can reveal preferred sites. The nucleotide sequences in the break point regions often hint at distinct deletion and duplication events18,20,21 involving specific molecular repair mechanisms. Variants of the RhCE protein are not known to exert any regulatory effect or growth advantage, and one or more of the many other genes located in the deleted or duplicated long DNA segments could be evaluated as a cause for the clonal expansion. Red cells represent 84% of all human cells.22 Isnâ&#x20AC;&#x2122;t it an odd coincidence that both patients happen to have approximately half of their red cells affected despite 2 distinct LOH mutations (Figure 1)? While the single pluripotent stem cell and myeloid precursor, having incurred different LOH mutations,5 can develop some clonality,23 they are hardly expected to produce 50% of all red cells in each patient. It might be an observation bias that these reported findings are so similar, if many clonal events of a lesser degree are overlooked in routine clinical practice. With the advent of mass scale sequencing of whole genomes in cell lines and cell lineages, mosaicism is becoming more frequently observed. And it can only become more important as single cell genomics is used to study aging populations. As understanding of underlying mechanisms widens, we will observe similar mutation patterns and their prevalence. Each cell lineage will accumulate its distinct combination of mutations, and these are bound to vary among individuals; even between monozygotic, identical twins. As 99% of all inherited nucleotide sequence in haematologica | 2019; 104(3)

any individual is homozygous, the acquired heterozygous sequence, along with the common inherited status, may constitute a major cause for phenotypes, and ultimately diseases. Today, since systematic analysis of mosaicism remains challenging, its role as a critical mechanism for disease and aging may quite possibly be underappreciated.4 Acknowledgments The author thanks Harvey Gordon Klein for his review. Supported by the Intramural Research Program (project ID Z99 CL999999) of the NIH Clinical Center.

References 1 Yoshizato T, Dumitriu B, Hosokawa K, et al. Somatic mutations and clonal hematopoiesis in aplastic anemia. N Engl J Med. 2015;373(1):3547. 2. Babushok DV, Duke JL, Xie HM, et al. Somatic HLA mutations expose the role of class I-mediated autoimmunity in aplastic anemia and its clonal complications. Blood Adv. 2017;1(22):1900-1910. 3. Frumkin D, Wasserstrom A, Kaplan S, Feige U, Shapiro E. Genomic variability within an organism exposes its cell lineage tree. PLoS Comput Biol. 2005;1(5):e50. 4. Machiela MJ, Chanock SJ. The ageing genome, clonal mosaicism and chronic disease. Curr Opin Genet Dev. 2017;42:8-13. 5. Dauber EM, Mayr WR, Hustinx H, et al. Somatic mosaicisms of chromosome 1 at two different stages of ontogenetic development detected by Rh blood group discrepancies. Haematologica. 2019;104(3):632-638. 6. van Loghem JJ, Dorfmeier H, van der Hart M. Two A antigens with abnormal serologic properties. Vox Sang. 1957;2(1):16-24. 7. Gold ER, Tovey GH, Benney WE, Lewis FJ. Changes in the group A antigen in a case of leukaemia. Nature. 1959;183(4665):892-893. 8. Tovey GH, Lockyer JW, Tierney RB. Changes in Rh grouping reactions in a case of leukaemia. Vox Sang. 1961;6:628-631. 9. Kolins J, Holland PV, McGinniss MH. Multiple red cell antigen loss in acute granulocytic leukemia. Cancer. 1978;42(5):2248-2253. 10. Kolins J, Allgood JW, Burghardt DC, Klein HG, McGinniss MH. Modifications of B, I, i, and Lewisb antigens in a patient with DiGuglielmo's erythroleukemia. Transfusion. 1980;20(5):574-577. 11. Thiagalingam S, Foy RL, Cheng KH, Lee HJ, Thiagalingam A, Ponte JF. Loss of heterozygosity as a predictor to map tumor suppressor genes in cancer: molecular basis of its occurrence. Curr Opin Oncol. 2002;14(1):65-72. 12. O'Keefe C, McDevitt MA, Maciejewski JP. Copy neutral loss of heterozygosity: a novel chromosomal lesion in myeloid malignancies. Blood. 2010;115(14):2731-2739. 13. Montemayor-Garcia C, Coward R, Albitar M, et al. Acquired RhD mosaicism identifies fibrotic transformation of thrombopoietin receptor-mutated essential thrombocythemia. Transfusion. 2017;57(9):21362139. 14. Miao X, Li X, Wang L, Zheng C, Cai J. DSMNC: a database of somatic mutations in normal cells. Nucleic Acids Res. 2018 Oct 31. [Epub ahead of print] 15. Wagner FF, Flegel WA. Polymorphism of the h allele and the population frequency of sporadic nonfunctional alleles. Transfusion. 1997;37(3):284-290. 16. Schmid P, Flegel WA. Codon usage in vertebrates is associated with a low risk of acquiring nonsense mutations. J Transl Med. 2011;9:87. 17. Flegel WA, Eicher NI, Doescher A, et al. In-frame triplet deletions in RHD alter the D antigen phenotype. Transfusion. 2006;46(12):21562161. 18. Wagner FF, Flegel WA. RHD gene deletion occurred in the Rhesus box. Blood. 2000;95(12):3662-3668. 19. Wagner FF, Frohmajer A, Flegel WA. RHD positive haplotypes in D negative Europeans. BMC Genet. 2001;2:10. 20. Srivastava K, Stiles DA, Wagner FF, Flegel WA. Two large deletions extending beyond either end of the RHD gene and their red cell phenotypes. J Hum Genet. 2018;63(1):27-35. 21. Wagner FF, Flegel WA. RHCE represents the ancestral RH position, while RHD is the duplicated gene. Blood. 2002;99(6):2272-2273. 22. Sender R, Fuchs S, Milo R. Are we really vastly outnumbered? Revisiting the ratio of bacterial to host cells in humans. Cell. 2016;164(3):337-340. 23. Cooper JN, Young NS. Clonality in context: hematopoietic clones in their marrow environment. Blood. 2017;130(22):2363-2372.

427


PERSPECTIVE ARTICLE Ferrata Storti Foundation

Endothelialized flow models for blood transfusion research

Monica S.Y. Ng,1,2 Jacky Y. Suen,1 John-Paul Tung1,2 and John F. Fraser1

Critical Care Research Group, Faculty of Medicine, University of Queensland, Brisbane and 2Research and Development, Australian Red Cross Blood Service, Brisbane, Australia 1

Haematologica 2019 Volume 104(3):428-434

Introduction Recent advances in cell culture and microfabrication technologies have enabled the development of perfusable endothelialized channels in vitro. To date, these techniques have primarily been applied to tissue engineering research. However, this set-up provides the unique opportunity to simulate blood product transfusion in a cost-effective, robust and reproducible manner – incorporating blood, endothelial and flow components in one unit. This Perspective describes the value of vascular models in transfusion research and discusses key decision points in the design process.

What happens to blood after transfusion?

Correspondence: MONICA S.Y. NG monica.ng91@gmail.com Received: October 31, 2018. Accepted: January 15, 2019. Pre-published: February 14, 2019.

doi:10.3324/haematol.2018.205203 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/103/4/428 ©2019 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

428

On transfusion, blood products interact with blood cells and plasma components to alter platelet activation, leukocyte function and red blood cell (RBC) oxygen-carrying capacity. The pro-/anti-inflammatory balance hinges on whether neutrophil or macrophage responses dominate after blood product transfusion.1-3 These responses are dependent on the activation state of recipient neutrophils and macrophages which, in turn, is influenced by cytokines in the local microenvironment. RBCs and platelets modify immune system function by activating complement, releasing cytokines and participating in receptor-ligand interactions.4 Blood products and recipient blood are encased by endothelium in blood vessels, one of the largest organs in the body with a surface area of 350-1000 m2.5 The endothelium conveys blood to tissues, provides a surface that prevents improper clotting and cellular activation, acts as a selective barrier to macromolecule extravasation and regulates microvascular blood flow.6 Activated endothelium participates in inflammation by releasing chemotactic molecules (e.g., interleukin8 and monocyte chemoattractant protein-1), generating reactive oxygen species and expressing adhesion molecules (CD62, CD106, CD54, CD31) to attract leukocytes and facilitate leukocyte transmigration.7,8 Furthermore, activated endothelium also enhances thrombosis by elaborating procoagulant surface molecules (von Willebrand factor, tissue factor) and microparticles.8 Endothelial dysfunction has been implicated in transfusion-related acute lung injury, sepsis and multiple organ dysfunction.8,9 The mixture of blood product and recipient blood is constantly mixed and propelled by the cardiac cycle – maximizing cellular interactions and minimizing inappropriate endothelial adhesion.10 After leaving the heart, blood flows through arteries to reach capillaries in the tissues and then veins before returning to the heart. The three types of blood vessels differ in structure, diameter, flow patterns and shear stress.10 In arteries and veins, RBCs and leukocytes flow in the center of the flow stream and platelets are distributed to the periphery of the stream.11 RBCs exhibit a parabolic velocity profile with shear-dependent rotation which continuously mixes the blood components. The configuration of cells in the flow stream can be modified by RBC plasticity, shear rate and fluid viscosity.12,13 In the microvasculature, cells travel in single file with uniform distribution of platelets, RBCs and leukocytes in the flow stream. Additionally, blood flow exerts shear stress on endothelium thereby altering endothelial gene expression, apoptosis, migration, permeability and alignment.6,14 Endothelial cells cultured under flow conditions demonstrate enhanced barrier function in conjunction with minimal adhesion molecule activation.15,16 Physiological shear stress is protective against inappropriate endothelial cytokine release compared to low shear stress.9 Additionally, flow patterns (e.g. continuous versus pulsatile) influence endothelial adhesion protein expression, structure and haematologica | 2019; 104(3)


Perspective Article Table 1. Available models for testing blood-endothelial interactions in vitro.

Model Static conditions

Protocol

Outcome measures

Research output and clinical impact

Blood unit sampling

• Repeated sampling of blood product over shelf life (37-41)

• Markers of platelet activation + apoptosis (37) • Media change (38) • Microparticle accumulation (38) • Proteomics (42) • Procoagulant characteristics (39) • RBC anti-oxidant oxidation during storage (40) • Membrane deformability (41)

• PC undergo deterioration of mitochondrial integrity + apoptosis (37) during storage • PRBC storage leads to enhanced RMP formation, osmotic fragility, hemolysis, decreased deformability (38), prolonged activated clotting time (39), increased RBC anti-oxidant oxidation (40)

Specific component co-incubation

• Blood products or supernatant of blood product mixed with purified cell populations from healthy volunteers - Neutrophils (1, 17-21) - Macrophages (3) - Platelets (22) - HUVECs (23)

• Neutrophil apoptosis (17) • Neutrophil ROS formation (17, 19, 20) • Neutrophil activation markers (1, 18, 21) • Neutrophil phagocytosis (1) • Macrophage cytokine elaboration (3) • Platelet adhesion + aggregation (22) • RBC adhesion to endothelial cells (23)

• Supernatant from stored PRBCs delays neutrophil apoptosis and primes neutrophils (17) • PMP binding to neutrophils increases CD11b expression + phagocytic activity (1) • Non-polar lipids accumulated during ex vivo PRBC storage prime neutrophils (20) • PRBCs stored for prolonged periods are less capable of supporting platelet adhesion + aggregation (22)

• Cytokine production after adding fresh or stored PRBC supernatant to whole blood (2, 24) • Thromboelastometry, microparticle characterization, cytokine concentration (43)

• Incubation of microparticles from PRBCs with whole blood induced host production of TNF, IL-6, IL-8 (2) • RMPs from stored PRBCs trigger coagulation of whole blood through TF signaling (43)

Whole blood incubation • Blood product components incubated with whole blood (2, 24) • Storage-induced RMP incubated with whole blood (43)

Laminar flow conditions Acellular planar flow model

• Cells from blood products perfused over layer of subendothelial matrix or collagen (44)

• Influence of PC storage duration on adhesion capacity to subendothelial matrix + collagen (44)

• Platelet aggregation + adhesion capacity are improved in BC PC compared to PRP PC (44)

Cellular planar flow model

• Endothelial cells cultured on coverslips and subjected to specific shear stress - 0.3 dyn/cm2 (27) - 5 dyn/cm2 (26) - 10 dyn/cm2 (25) - Shear stress not reported (25) • Different mixtures of blood perfused through system - 10% whole blood suspension (25) - 1.5% PRBC suspension (26, 27) • Use of activated endothelium - E. coli 0127:B8 LPS (45) - TNF-α (25, 45)

• Adhesion of RBCs to endothelium from PRBCs stored for different periods (26, 27, 45) • Tissue factor expression + adhesion to endothelial cells (25)

• RBCs from stored PRBCs more adherent to HUVECs under laminar flow conditions compared to those from fresh PRBCs (26, 27) • Exposure to TNF-stimulated HUVECs under arterial flow conditions led to increased tissue factor expression on monocytes compared to quiescent HUVECs (25) • RBCs more adherent to E. coli-activated endothelium compared to quiescent endothelium (45)

Ex vivo artery model

• Human umbilical artery (up to 20 cm) and connected to circuit (28) • Whole blood diluted 1:10 circulated with shear pressure of 0.4 dyn/cm2 • Endothelium activated with TNF-α

• Microparticle formation, endothelial ROS formation (28)

• Whole blood perfused through TNF-stimulated umbilical arteries demonstrated increased microparticle formation. These microparticles enhanced HUVEC ROS formation in vitro (28)

• Confocal microscopy, biochemical extracellular matrix analysis, endothelial nitric oxide formation (29)

• Monocytes circulated through the TNF-stimulated model adhered to the endothelium and transmigrated into the intima (29)

Synthetic arterial model • Biodegradable tubular scaffold matrices (length 4 cm, inner diameter 6 mm) seeded with HUVECs (29)

continued on the next page

haematologica | 2019; 104(3)

429


Perspective Article continued from the previous page

Model Static conditions

Protocol

Outcome measures

Research output and clinical impact

Acellular synthetic microvascular model

• Soft lithography with vessel diameters of 5-70 mm in PDMS with no cell lining • RBCs perfused through system at different concentrations - 0.4 haematocrit (30)

• Perfusion rate for RBCs stored for different periods (30)

• Perfusion rate for stored RBCs was 19-30% lower than for fresh RBCs. Washing stored RBCs in saline improved perfusion rate by 41% (30)

Cellular synthetic microvascular model

• Soft lithography with vessel diameters • RBC adhesion, endothelial marker of 50-200 mm in PDMS (9) or collagen (31) expression, confocal microscopy (31) • Channels lined with HUVECs (9, 31) • Endothelial damage (9) • RBCs perfused at 10-20% (v/v) RBCs with shear stress up to 17 dyn/cm2 (9, 31) • Model can be stretched to simulate respiratory forces (9) • Perivascular cells can be added to collagen matrix (31)

In vivo microvascular model (large animal)

• Microvasculature assessed with • Microcirculatory flow index, microvascular • Isotonic or hypertonic colloidal fluids - Sublingual sidestream dark field blood flow, capillary-venous adequately restored sublingual imaging (46) hemoglobin oxygen saturation (46) microcirculatory blood flow and - Laser Doppler flowmeter (46) flow quality (46) - Tissue reflectance spectrophotometer (46) • Gelatin + hydroxyethyl starch improved • Animals used microvascular hemoglobin oxygen - Domestic pig (46) saturation (46)

In vivo microvascular model (small animal)

• Animals transfused with blood product and then assessed using intravital microscopy • Animals used - Rat: cremaster flap (47), extensor digitorum longus muscle (48) - Mice cremaster flap (49) - Hamster: dorsal skin flap (50)

• Low hemodynamic shear stress due to altered microcirculatory flow may predispose HUVECs to necroptosis (9) • Cyclic stretching of microvessels (similar to breathing movements or mechanical ventilation) may increase susceptibility of HUVECs to transfused RBCs (9)

• RBC velocity, vessel diameter, FCD, • Fresh PRBCs more effective at relieving after stored PRBC transfusion (47) microcirculatory hypoxia compared to • FCD, RBC velocity, vessel diameter, stored PRBCs in rat cremaster flap O2 distribution after PRBC transfusion (50) model (47) • RBC adhesion in capillaries (48) • IgG-mediated HTRs induced acute • RBC velocity, vessel diameter (49) vaso-occlusive crisis in the mice cremaster model. CXCR2 blockage prevented HTR-induced vasoocclusive crisis (49)

BC: buffy coat; E. coli: Escherichia coli; FCD: functional capillary density; HTR: hemolytic transfusion reaction; HUVECs: human umbilical vin endothelial cells; LPS: lipopolysaccharide; PC: platelet concentrate; PDMS: polydimethylsiloxane; PMPs: platelet microparticles; PRBC: packed red blood cells; PRP: platelet-rich plasma; RBC: red blood cell; RMP: red cell microparticle; ROS: reactive oxygen species; TNF-α: tumour necrosis factor--alpha.

alignment.6 Lastly, shear stress modifies endothelial interactions with blood cells. For example, monocytes perfused over endothelium activated by tumor necrosis factor-α expressed more tissue factor and CD11b compared to monocytes co-incubated with activated endothelium under static conditions.9 It follows that intravascular transfusion events are affected by blood product-recipient blood interactions and blood mixture-endothelium interactions under flow conditions. Simulation of the post-transfusion intravascular mileau in vitro requires recipient whole blood, blood product, endothelium and perfusate flow. To date, the majority of models used to test transfusion effects have involved static models in which blood products are coincubated with specific cells such as neutrophils,1,17-21 macrophages,3 platelets22 or endothelial cells23 to examine interactions between two cell types. While blood co-incubation experiments enable the investigation of multiple cellular interactions,2,24 they do not recapitulate the frequency and type of cellular interactions that occur under endogenous flow conditions. In vivo models are the “gold standard” for capturing the sum of intravascular interac430

tions that occur after transfusion of stored, packed RBCs. However, such models are limited by increased variability (requiring larger sample sizes), reduced capacity to isolate parameters and longer set-up times. Endothelialized in vitro flow models circumvent these limitations by using cell culture methods, packed RBCs and fresh whole blood, which are easier to acquire with fewer associated ethical considerations, shorter model development time and lower cost.

What can we do with current manufacturing technology? The potential methods to investigate blood mixtureendothelial interactions and current research outputs from these models are summarized in Table 1. Flow models can be divided into planar, microvascular and macrovascular depending on the arrangement of endothelial and blood cells. For planar models, endothelial cells were cultured on coverslips and subjected to various shear stresses generated by laminar flow.25-27 These models were used to study blood-endothelial adhesion under macrovascular flow conditions with shear stresses of 0.3-10 dyn/cm2.25-27 Notably, these models do not recahaematologica | 2019; 104(3)


Perspective Article

pitulate macrovascular flow stream configurations (rectangular flow stream versus circular flow stream) and do not test physiological hematocrits (<10% whole blood/packed RBC suspensions versus undiluted blood).25-

27

The ex vivo artery model recapitulated physiological flow stream configurations by connecting human umbilical arteries in a perfusable circuit.28 However, the use of harvested arteries meant that the model was susceptible

Figure 1. Five key decision points in vascular model design. The five key factors are channel geometry, hydrogel scaffold, endothelial cells, circuit design and perfusate mixture. Branched channels less than 200 mm in diameter are produced using photolithography and soft lithography while straight channels greater than 120 Âľm in diameter are produced using rod/needle casting. The hydrogel scaffold can be produced in one or two pieces and may be seeded with perivascular cells. Endothelial cells are subsequently seeded onto the channels and cultured under flow conditions to form a confluent monolayer. The vascular model is then connected to a circuit with a pump with or without a gas trap. The last decision involves choosing the type of perfusate to be circulated through the circuit. CP: cryoprecipitate; EDTA: ethylenediamine-tetraacetic acid; FFP: fresh-frozen plasma; HAECs: human aortic endothelial cells; HDMECs: human dermal microvascular endothelial cells; HLMECs: human lung microvascular endothelial cells; HUVECs: human umbilical vein endothelial cells; PDMS: polydimethylsiloxane; Plt: platelet concentrates; PRBC: packed red blood cells; PSC: pluripotent stem cells.

haematologica | 2019; 104(3)

431


Perspective Article

to inter-donor variation and ethical concerns. Synthetic arterial models, produced by seeding cells on a biodegradable tubular scaffold, can overcome these issues.29 To date, this model has not been perfused with blood to investigate blood-endothelial interactions. Microvascular models involve the formation of <200 mm channels in polydimethylsiloxane, subendothelium or hydrogels.9,30,31 These channels are available in different branching geometries and can be lined by endothelial cells.31,32 Acellular models can accommodate smaller vessel diameters (5-70 mm) and higher hematocrits (up to 0.4) without blockage.30 In contrast, endothelialized microvascular models are limited by vessel diameters of 50-200 mm with 10-20% (v/v) packed RBC perfusates.9,31 Vessels less than 50 mm in diameter are difficult to produce in hydrogels, while vessels greater than 200 Âľm present issues during endothelial seeding and culture which preclude the formation of a confluent monolayer. Manufacturing limitations associated with vessel diameter can be abrogated by using in vivo microvascular models such as muscle or dorsal skin flaps. However, these models lack the channel diameter and geometrical consistency resulting from in vitro-manufacturing techniques. The presence of endothelial cells in microvascular models enhances thrombosis, thereby hindering the hematocrits that can be tested. Physiological hematocrits are important for viscosity, shear stress and cell-to-cell interactions. For example, an increased hematocrit leads to more RBCendothelial interactions which enhance shear stress at low flow rates. Furthermore, an increased hematocrit has been associated with altered RBC arrangement in the flow stream and margination of blood components.33

Designing vascular models Perfusable in vitro macrovascular and microvascular models are composed of an endothelialized channel in a hydrogel scaffold. The unit is subsequently connected to a circuit and perfused with the desired â&#x20AC;&#x153;testâ&#x20AC;? fluid. There are five key decision points in vascular model design (Figure 1): (i) channel geometry, (ii) scaffold moulding, (iii) endothelial cell seeding, (iv) circuit construction, and (v) perfusate selection. The choice of channel size and shape determines the scaffold manufacturing method. Straight channels with diameters down to 120 mm can be injection-molded using rods, needles or wires stabilized on a mount. This method is cost-effective and technically easy to construct but has limited fidelity for endogenous vascular geometry. Branched channel geometries with diameters of 50200 mm can be formed using a mixture of photolithography and soft lithography methods. Photolithography is used to create a hard negative-profile wafer. Complex channel geometries can be etched onto a photomask which is used to develop the desired pattern on a photoresist-coated crystalline silicon wafer.34 Notably, photolithography requires microfabrication facilities (e.g. clean room, photo pattern generator, spin coater) and specialized consumables (e.g. crystalline silicon wafer, photoresist, developer). The negative-profile wafer becomes the mold for the soft positive-profile stamp used to shape the hydrogel scaffold by soft lithography.31 The soft positiveprofile stamp can be formed from polydimethylsiloxane 432

in any standard laboratory using a vacuum degasser and oven. Polydimethylsiloxane can be printed directly in three dimensions: however, this method has inferior resolution compared to the photo/soft-lithography combination (~760 mm versus 50 mm, resolution limited by soft lithography).35 The hydrogel scaffold can be constructed from collagen, alginates, agarose, poly(ethylene glycol) dimethacrylate or methacrylated gelatin.5 These materials are chosen for their transparency, absence of toxicity, fidelity for micropatterning and transport properties. Perivascular and/or tissue cells can be incorporated into the hydrogel to enhance biological approximation. The hydrogel is subsequently injection-molded onto a stamp and allowed to solidify at specified temperatures. Straight channel scaffolds can be formed as one piece. Scaffolds involving the use of a stamp are formed as two pieces (stamped pieced, flat slab), which are subsequently combined to form closed channels. The key obstacles at this stage are bubble formation during injection molding and channel damage during scaffold assembly. The next decision involves the type of endothelial cells to use to seed the channels. Commercially available multi-donor human umbilical vein endothelial cells are the most commonly used cell type due to their availability, robustness and proliferation consistency. However, multi-donor cells can express multiple antigenic profiles, complicating donor-recipient cross-matching in transfusion simulations. Furthermore, human umbilical vein endothelial cells may behave differently from other primary endothelial cells. Vessel-specific primary endothelial cells such as human lung microvascular endothelial cells and human aortic endothelial cells may improve physiological relevance but these cells are often more difficult to culture and susceptible to variability. Endothelial derivatives from human pluripotent stem cells can also be used.36 The desired endothelial cells can be seeded on the channel surface under static or low flow conditions and grown to confluence under laminar flow. Notably, high seeding flow rates will prevent endothelial cells from attaching to the hydrogel scaffold. We recommend seeding endothelial cells under static conditions and allowing up to 3 hours for the cells to attach, prior to incubating the cells overnight under laminar flow conditions. Endothelial cell growth under laminar flow is important to cultivate the barrier function and monolayer present in endogenous blood vessels. Variables in circuit construction include circulation volume, tubing material and diameter, gas trap and pump. The circulation volume can be adjusted to accommodate the desired outcome measures (e.g. flow cytometry, enzyme-linked immunosorbent assay, mass spectrometry, biochemical analysis). Tubing length and diameter can be varied to reach the desired circulation volume. Notably, the more tubing is used to maximize circulation volume, the greater the ratio between non-endothelialized and endothelized surface area in the circuit. Medical grade tubing made out of Tygon, fluoropolymers or polyvinylchloride are used as they are inert, biocompatible and gas permeable. In our set-up, the gas trap is connected just before the inlet port of the vascular model to minimize bubbles passing through the endothelialized haematologica | 2019; 104(3)


Perspective Article

portion of the circuit. The pump chosen needs to support the desired flow rate (macrovascular: ≤70 mL/min, microvascular: ≤50 µL/min). Syringe pumps deliver consistent continuous flow rates. Peristaltic pumps can deliver continuous and pulsatile flow patterns. Due to the rotor mechanism of peristaltic pumps, the continuous flow setting may deliver a “dampened” pulsatile flow. Pumps which deliver inconsistent flow rates may lead to bubble formation in the circuit. The flow rate can be adjusted to simulate the desired environment (e.g. arteries, capillaries, veins) and shear stress. When calculating the shear rate, it is important to note that blood is a nonNewtonian fluid meaning that its viscosity and the shear stress that it exerts on the endothelium are dependent on the amount of pressure exerted on it. Lastly, the entire circuit can be placed in an incubator or the pump computer can be left outside the incubator (if there are ports available for the passage of cables). Whole blood, blood product, whole blood-blood product mixtures and diluted blood can be perfused through the system to simulate blood flow. Undiluted blood best recapitulates the intravascular milieu by maintaining blood concentration and viscosity. However, undiluted anticoagulated whole blood still carries increased risks of thrombosis (particularly in microvascular channels) and is opaque – making it optically difficult to assess blood cell adhesion to endothelium. In these situations, blood can be diluted to maintain circulation and enable in situ microscopic analysis. Mixing whole blood with packed RBCs at a ratio of 9:1 simulates a one unit transfusion of packed RBCs. The whole blood should be diluted such that the hematocrit is below the transfusion threshold of 7 g/dL to prevent post-transfusion hyperviscosity. While this dilution simulates the hemodilution observed after crystalloid resuscitation, it is unlikely to approximate the blood viscosity observed in patients with chronic anemia. Notably, all blood and blood products circulated through in vitro models are anticoagulated as this treatment is necessary for ex vivo blood storage prior to experimentation. Sodium citrate, ethylenediamine-tetraacetic acid (EDTA) and heparin anticoagulation are non-toxic and suitable for use. Sodium citrate and EDTA work by chelating calcium ions

References 1

Jy W, Mao WW, Horstman L, Tao J, Ahn YS. Platelet microparticles bind, activate and aggregate neutrophils in vitro. Blood Cells Mol Dis. 1995;21(3):217-231. 2. Straat M, Boing AN, Tuip-De Boer A, Nieuwland R, Juffermans NP. Extracellular vesicles from red blood cell products induce a strong pro-inflammatory host response, dependent on both numbers and storage duration. Transfus Med Hemother. 2016;43(4):302-305. 3. Sadallah S, Eken C, Schifferli JA. Erythrocyte-derived ectosomes have immunosuppressive properties. J Leukoc Biol. 2008;84(5):1316-1325. 4. Karsten E, Herbert B. Red blood cells: the immune system's hidden regulator. Annual scientific meeting of Haematology Society

haematologica | 2019; 104(3)

5. 6.

7.

8.

9.

and their anticoagulant effect can be weakened by adding calcium. Sodium citrate is required for coagulation analyses. Furthermore, high levels of sodium citrate can alter the pH of blood, EDTA can activate endothelium and both chelators can interfere with microparticle formation. Heparin provides enduring anticoagulation, but it interferes with coagulation analyses and immune cell function. It follows that the requirement for anticoagulated blood represents a shortcoming of in vitro vascular models. This pitfall can be dealt with to some extent by using in vivo models to corroborate in vitro results. Various outcome measures can be recorded from these in vitro vascular models. Real-time video microscopy can be used to visualize blood cell adhesion to the endothelium. The perfusant can be analyzed for soluble factors (e.g. cytokines, microparticles, biological response mediators) and cell surface markers. After blood circulation, the endothelial cells can be imaged in situ or removed from the scaffold with trypsin for further analysis. In our laboratory, a vascular model with a 3 mm, full-length channel seeded with human umbilical vein endothelial cells is primarily used to simulate transfusion reactions by circulating mixtures of recipient whole blood and donor blood products (Online Supplementary Material). Whole blood circulation leads to increased formation of annexin V-positive microparticles and erythrocyte microparticles compared to statically held whole blood (2.13 x 1010 versus 6.08 x 109 microparticles/mL, P<0.0001 (Online Supplementary Material).

Conclusion In vitro vascular models combine blood, endothelial and flow components into a single system. In this way, endothelialized flow models simulate the blood productrecipient blood and blood-endothelial interactions that occur under flow conditions after blood product transfusion. Five key factors of the experimental set-up (channel geometry, scaffold molding, endothelial cell seeding, circuit construction, and perfusate) can be manipulated depending on the desired organ system, transfusion scenario and outcome measures.

of Australia and New Zealand, the Australian & New Zealand Society of Blood Transfusion and the Thrombosis and Haemostasis Society of Australia and New Zealand. Sydney, Australia, 2017. Rayner SG, Zheng Y. Engineered microvessels for the study of human disease. J Biomech Eng. 2016;138(11). Li YS, Haga JH, Chien S. Molecular basis of the effects of shear stress on vascular endothelial cells. J Biomech. 2005;38(10):1949-1971. Granger D, Senchenkova E. Leukocyte– Endothelial Cell Adhesion. Inflammation and the Microcirculation. San Rafael (CA): Morgan & Claypool Life Sciences, 2010. Hack CE, Zeerleder S. The endothelium in sepsis: source of and a target for inflammation. Crit Care Med. 2001;29(7 Suppl):S2127. Seo J, Conegliano D, Farrell M, et al. A

10. 11.

12.

13.

14.

microengineered model of RBC transfusioninduced pulmonary vascular injury. Sci Rep. 2017;7(1):3413. Hathcock JJ. Flow effects on coagulation and thrombosis. Arterioscler Thromb Vasc Biol. 2006;26(8):1729-1737. Sakariassen KS, Orning L, Turitto VT. The impact of blood shear rate on arterial thrombus formation. Future Sci OA. 2015;1(4):FSO30. Eckstein EC, Bilsker DL, Waters CM, Kippenhan JS, Tilles AW. Transport of platelets in flowing blood. Ann N Y Acad Sci. 1987;516:442-452. Xu C, Wootton DM. Platelet near-wall excess in porcine whole blood in arterysized tubes under steady and pulsatile flow conditions. Biorheology. 2004;41(2):113125. Dimmeler S, Haendeler J, Nehls M, Zeiher AM. Suppression of apoptosis by nitric

433


Perspective Article

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

434

oxide via inhibition of interleukin-1betaconverting enzyme (ICE)-like and cysteine protease protein (CPP)-32-like proteases. J Exp Med. 1997;185(4):601-607. Miao H, Hu YL, Shiu YT, et al. Effects of flow patterns on the localization and expression of VE-cadherin at vascular endothelial cell junctions: in vivo and in vitro investigations. J Vasc Res. 2005;42(1):77-89. Young EW, Watson MW, Srigunapalan S, Wheeler AR, Simmons CA. Technique for real-time measurements of endothelial permeability in a microfluidic membrane chip using laser-induced fluorescence detection. Anal Chem. 2010;82(3):808-816. Biffl WL, Moore EE, Offner PJ, Ciesla DJ, Gonzalez RJ, Silliman CC. Plasma from aged stored red blood cells delays neutrophil apoptosis and primes for cytotoxicity: abrogation by poststorage washing but not prestorage leukoreduction. J Trauma. 2001;50(3):426-432. Cardo LJ, Wilder D, Salata J. Neutrophil priming, caused by cell membranes and microvesicles in packed red blood cell units, is abrogated by leukocyte depletion at collection. Transfus Apher Sci. 2008;38(2):117-125. Jank H, Salzer U. Vesicles generated during storage of red blood cells enhance the generation of radical oxygen species in activated neutrophils. ScientificWorldJournal. 2011;11:173-185. Silliman CC, Moore EE, Kelher MR, Khan SY, Gellar L, Elzi DJ. Identification of lipids that accumulate during the routine storage of prestorage leukoreduced red blood cells and cause acute lung injury. Transfusion. 2011;51(12):2549-2554. Belizaire RM, Prakash PS, Richter JR, et al. Microparticles from stored red blood cells activate neutrophils and cause lung injury after hemorrhage and resuscitation. J Am Coll Surg. 2012;214(4):648-655; discussion 656-657. Morrison A, McMillan L, Hornsey VS, Prowse CV. Stored red-blood-cells inhibit platelet function under physiologic flow. Vox Sang. 2010;99(4):362-368. Bonomini M, Sirolli V, Gizzi F, Di Stante S, Grilli A, Felaco M. Enhanced adherence of human uremic erythrocytes to vascular endothelium: role of phosphatidylserine exposure. Kidney Int. 2002;62(4):1358-1363. Karam O, Tucci M, Toledano BJ, et al. Length of storage and in vitro immunomodulation induced by prestorage leukoreduced red blood cells. Transfusion. 2009;49(11):2326-2334. Macey MG, Wolf SI, Wheeler-Jones CP, Lawson C. Expression of blood coagulation factors on monocytes after exposure to TNF-treated endothelium in a novel whole

26.

27.

28.

29.

30.

31. 32. 33.

34.

35.

36.

37.

38.

blood model of arterial flow. J Immunol Methods. 2009;350(1-2):133-141. Anniss AM, Sparrow RL. Storage duration and white blood cell content of red blood cell (RBC) products increases adhesion of stored RBCs to endothelium under flow conditions. Transfusion. 2006;46(9):15611567. Luk CS, Gray-Statchuk LA, Cepinkas G, Chin-Yee IH. WBC reduction reduces storage-associated RBC adhesion to human vascular endothelial cells under conditions of continuous flow in vitro. Transfusion. 2003;43(2):151-156. Holtom E, Usherwood JR, Macey MG, Lawson C. Microparticle formation after coculture of human whole blood and umbilical artery in a novel in vitro model of flow. Cytometry A. 2012;81(5):390-399. Robert J, Weber B, Frese L, et al. A threedimensional engineered artery model for in vitro atherosclerosis research. PLoS One. 2013;8(11):e79821. Burns JM, Yang X, Forouzan O, Sosa JM, Shevkoplyas SS. Artificial microvascular network: a new tool for measuring rheologic properties of stored red blood cells. Transfusion. 2012;52(5):1010-1023. Morgan JP, Delnero PF, Zheng Y, et al. Formation of microvascular networks in vitro. Nat Protoc. 2013;8(9):1820-1836. Chrobak KM, Potter DR, Tien J. Formation of perfused, functional microvascular tubes in vitro. Microvasc Re. 2006;71(3):185-196. Muller K, Fedosov DA, Gompper G. Margination of micro- and nano-particles in blood flow and its effect on drug delivery. Sci Rep. 2014;4:4871. Martinez-Duarte R, Madou M. SU-8 Photolithography and its impact on microfluidics. 2010. In: Microfluidics and Nanofluidics Handbook: Fabrication, Implementation and Applications.pp239240. Ozbolat V, Dey M, Ayan B, Povilianskas A, Demirel MC, Ozbolat IT. 3D Printing of PDMS improves its mechanical and cell adhesion properties. ACS Biomaterials Science & Engineering. 2018;4(2):682-693. Palpant NJ, Pabon L, Roberts M, et al. Inhibition of beta-catenin signaling respecifies anterior-like endothelium into beating human cardiomyocytes. Development. 2015;142(18):3198-3209. Albanyan AM, Harrison P, Murphy MF. Markers of platelet activation and apoptosis during storage of apheresis- and buffy coatderived platelet concentrates for 7 days. Transfusion. 2009;49(1):108-117. Almizraq R, Tchir JD, Holovati JL, Acker JP. Storage of red blood cells affects membrane composition, microvesiculation, and in vitro

quality. Transfusion. 2013;53(10):2258-2267. 39. Aucar JA, Isaak E, Anthony D. The effect of red blood cell age on coagulation. Am J Surg. 2009;198(6):900-904. 40. Bayer SB, Hampton MB, Winterbourn CC. Accumulation of oxidized peroxiredoxin 2 in red blood cells and its prevention. Transfusion. 2015;55(8):1909-1918. 41. Karger R, Lukow C, Kretschmer V. Deformability of red blood cells and correlation with ATP content during storage as leukocyte-depleted whole blood. Transfus Med Hemother. 2012;39(4):277-282. 42. Antonelou MH, Tzounakas VL, Velentzas AD, Stamoulis KE, Kriebardis AG, Papassideri IS. Effects of pre-storage leukoreduction on stored red blood cells signaling: a time-course evaluation from shape to proteome. J Proteomics. 2012;76 (Spec No.):220-238. 43. Fischer D, Bussow J, Meybohm P, et al. Microparticles from stored red blood cells enhance procoagulant and proinflammatory activity. Transfusion. 2017;57(11):27012711. 44. Boomgaard MN, Gouwerok CW, Homburg CH, de Groot G, IJsseldijk MJ, de Korte D. The platelet adhesion capacity to subendothelial matrix and collagen in a flow model during storage of platelet concentrates for 7 days. Thromb Haemost. 1994;72(4):611-616. 45. Anniss AM, Sparrow RL. Variable adhesion of different red blood cell products to activated vascular endothelium under flow conditions. Am J Hematol. 2007;82(6):439-445. 46. Maier S, Holz-Holzl C, Pajk W, et al. Microcirculatory parameters after isotonic and hypertonic colloidal fluid resuscitation in acute hemorrhagic shock. J Trauma. 2009;66(2):337-345. 47. Gonzalez AM, Yazici I, Kusza K, Siemionow M. Effects of fresh versus banked blood transfusions on microcirculatory hemodynamics and tissue oxygenation in the rat cremaster model. Surgery. 2007;141(5):630-639. 48. Chin-Yee IH, Gray-Statchuk L, Milkovich S, Ellis CG. Transfusion of stored red blood cells adhere in the rat microvasculature. Transfusion. 2009;49(11):2304-2310. 49. Jang JE, Hod EA, Spitalnik SL, Frenette PS. CXCL1 and its receptor, CXCR2, mediate murine sickle cell vaso-occlusion during hemolytic transfusion reactions. J Clin Invest. 2011;121(4):1397-1401. 50. Cabrales P, Intaglietta M, Tsai AG. Transfusion restores blood viscosity and reinstates microvascular conditions from hemorrhagic shock independent of oxygen carrying capacity. Resuscitation. 2007;75 (1):124-134.

haematologica | 2019; 104(3)


REVIEW ARTICLE

Reactivation of hepatitis B virus infection in patients with hematologic disorders

Ferrata Storti Foundation

Bo Wang,1 Ghulam Mufti2 and Kosh Agarwal1

1 Institute of Liver Studies and 2Department of Hematology, King's College Hospital, London, UK

ABSTRACT

Haematologica 2019 Volume 104(3):435-443

H

epatitis B reactivation is the reappearance or rise of hepatitis B virus (HBV) DNA in patients with past or chronic HBV infection, usually occurring in the context of immunosuppression. HBV reactivation has been most commonly reported in patients with hematologic disorders, with potentially serious and life-threatening consequences. In this review, we discuss the basis and presentation of HBV reactivation, and risk factors in terms of the host, the virus and the immunosuppression regimen, including newer agents used to manage hematologic malignancies. We overview the management of HBV reactivation, highlighting an up-dated recommendation on the use of newer nucleoside and nucleotide analogs, such as tenofovir and entecavir, for antiviral prophylaxis.

Introduction Hepatitis B reactivation is the reappearance or rise of hepatitis B virus (HBV) DNA in the serum of patients with past or chronic HBV infection. Reactivation can occur in a variety of clinical settings, usually in the context of an immunosuppressed state or immunosuppressive therapy. HBV reactivation has been most commonly reported in patients receiving chemotherapy for hematologic malignancies and following hematopoietic stem cell transplants.1 An estimated 2 billion people worldwide have serological evidence of either past or present HBV infection, with around 240 million people chronically infected.2 The prevalence varies globally, ranging between 2% in Europe to over 10% in East Asia; in the UK it is estimated to be between 0.5-1.7%, with areas of greater ethnic diversity such as London having a higher prevalence of approximately 2.4%.2,3 Therefore, there is a clear potential for HBV reactivation to cause significant morbidity, and even mortality, if not appropriately diagnosed and managed. Management of HBV in general is undergoing a paradigm shift. Recently up-dated clinical practice guidelines from the European Association for the Study of the Liver (EASL) have redefined the natural history of chronic HBV, driven by a better understanding of the interactions between the virus and the host immune system.4 From a therapeutic point of view, existing agents effectively suppress virus replication and lower serum HBV DNA concentrations, but the goal now is to develop novel agents that can offer functional cure of HBV.5,6 This is defined as the loss of hepatitis B surface antigen (HBsAg), the hallmark of chronic infection. Complete sterilizing cure is not considered possible due to the persistence of HBV DNA within hepatocytes. However, if functional cure becomes a realistic treatment end point, the number of patients with resolved HBV infection but who remain at risk of reactivation may increase significantly. Previous guidelines have been heterogeneous in their recommendations for the assessment of HBV reactivation, especially with regards to patient selection for testing and choice of antiviral prophylaxis. In this review, we aim to provide a practical overview of HBV reactivation at a time when the management of HBV is changing and the therapeutic options are expanding for patients with hematologic disorders, who are at the highest risk of this potentially life-threatening complication.

Correspondence: BO WANG bo.wang@nhs.net Received: November 22, 2018. Accepted: January 18, 2019. Pre-published: February 7, 2019.

doi:10.3324/haematol.2018.210252 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/104/3/435 Š2019 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

Hepatitis B virus reactivation and clinical presentation Chronic HBV infection is defined by the presence of HBsAg in serum with varihaematologica | 2019; 104(3)

435


B. Wang et al.

able HBV DNA levels depending on the balance between HBV replication and immune control.7 Up-dated nomenclature regarding the phases of HBV infection reflect this and broadly classify patients into hepatitis B e antigen (HBeAg) positive or negative, and whether or not there is evidence of a chronic hepatitis (Table 1).4 Those with resolved HBV infection are HBsAg negative and have circulating anti-core antibody (anti-HBc), and often anti-surface antibody (anti-HBs). Although such patients are considered to have past HBV infection, HBV DNA persists within the liver in the form of highly stable covalently closed circular DNA (cccDNA) and integrated DNA.8 Active replication is controlled by both innate and adaptive immune responses, including HBV-specific T-cell responses and neutralizing antibodies produced by activated B cells. However, these responses are not sufficient to eradicate all latent forms of HBV DNA and a reservoir of persistent HBV exists. With immunosuppression due to any cause, immune-mediated control of HBV replication is lost and reactivation can occur.9 Hepatitis B virus reactivation includes both exacerbation of chronic hepatitis B infection in an HBsAg-positive patient (with ≥2 log10 rise in HBV DNA level) and true reactivation of resolved hepatitis B infection, which can either be reverse HBsAg seroconversion (reappearance of HBsAg) or detection of HBV DNA with negative HBsAg. These virological events are often followed by a reactivation-related hepatitis (increase in ALT or AST ≥3 x baseline). In severe cases, or where reactivation is not recognized and there is a delay in treatment, hepatitis may progress to jaundice and potentially fulminant hepatic failure. More commonly, however, HBV DNA falls again either due to immune control or antiviral therapy, and the patient recovers.10 Studies of HBV reactivation during chemotherapy for lymphoma have demonstrated that viral reactivation itself can occur at any time during or after immunosuppression, but the hepatitis and clinical manifestations related to reactivation typically occur after treatment has ended when immune reconstitution takes place.11 In B-cell depletive therapies, such as rituximab, risk of reactivation is protracted, with cases reported up to two years after the last dose.12 Furthermore, HBV reactivation after hematopoietic stem cell transplantation (HSCT) may occur several years after transplantation because of the potential long delay in immune reconstitution.13 Therefore, reactivation may be ongoing over an extended period of time before therapy can be implemented, and requires long-term follow up and surveillance.

Risk factors Factors pertaining to the host, the virus, the immunosuppressive regimen, and the underlying disease itself can all impact on the risk of HBV reactivation. Male sex and older age (≥50 years) have been associated with increased risk. One study of more than 600 HBsAg-positive patients receiving chemotherapy for a range of cancers showed an almost 3-fold increased incidence in men, although the reason for this was not clear.14 Older patients are more likely to have HBsAg seroclearance but persistent levels of total HBV DNA and cccDNA in the liver, hence increasing the risk of reactivation.15 Viral factors associated with reactivation have been shown to include HBsAg positivity, HBeAg positivity, and elevated HBV DNA levels prior to commencing immunosuppressive therapy, all of which reflect a state of poor HBV-specific immune control prior to immunosuppression.16,17 Conversely, possessing antiHBs antibodies has been suggested to be protective against reactivation, although it has not been determined whether the specific titer has any effect.18 More recently, co-infection of HBV with other viruses such as HIV and hepatitis C virus (HCV) has been highlighted as a risk factor for reactivation even without the influence of immunosuppression. Treatment of HBV/HCV co-infected individuals with direct acting antivirals against HCV can result in HBV reactivation, although the clinical significance of this may be minimal.19,20 The mechanism of this observation is thought to be either a direct inhibitory effect of HCV replication on HBV or that immune responses against HCV also suppress HBV replication.21 The risk of HBV reactivation may be determined in part by the underlying disease, although studies comparing similar treatments in different hematologic diseases are lacking. Lymphoma has been the most common underlying condition in reports of reactivation; whether this is a reflection of the disease itself or the treatments given is not clear. An association between chronic HBV infection and non-Hodgkin lymphoma (NHL) has long been postulated and several studies have accumulated evidence to support this; a meta-analysis of more than 3000 patients with NHL and over 1 million controls showed an overall odds ratio of 2.56 for detecting HBsAg positivity in patients with NHL.22 A higher prevalence of anti-HBc positivity alone in NHL patients has also been demonstrated. Possible mechanisms to explain the association include direct HBV infection of lymphocytes, and chronic antigenic stimulation and associated B-cell proliferation.23,24

Table 1. Up-dated nomenclature for natural history phases of chronic hepatitis B virus (HBV) infection, adapted from the 2017 EASL Clinical Practice Guidelines.

HBeAg positive HBeAg HBV DNA ALT Liver disease Old terminology

Chronic infection

Chronic hepatitis

Chronic infection

Positive >107 IU/mL Normal None/minimal Immune tolerant

Positive 104 – 107 IU/mL Elevated Moderate/severe Immune reactive

Negative <2000 IU/mL* Normal None Inactive carrier

HBeAg negative Chronic hepatitis Negative >2000 IU/mL Elevated** Moderate/severe HBeAg negative chronic hepatitis

EASL: European Association for the Study of the Liver; HBeAg: hepatitis B e antigen; ALT: alanine transaminase. *Can be 2000-20000 IU/mL in some patients without signs of chronic hepatitis. **Either persistently or intermittently.

436

haematologica | 2019; 104(3)


HBV reactivation

Finally, the immunosuppression regimen itself is of great importance in the risk of reactivation, and numerous studies have attempted to stratify this risk.25 High risk is considered to be greater than 10%, moderate between 110%, and low risk less than 1%; the degree of risk has implications on management (Table 2). Low-risk regimens are limited to traditional immunosuppressive agents such as azathioprine and oral low-dose methotrexate without combination steroids, short-term low-dose steroids (≤20 mg/day prednisolone or equivalent for ≤7 days), and intraarticular steroids. The specific treatments relevant to hematologic disorders will be discussed in more detail below.

Specific immunosuppressive treatments Systemic cancer chemotherapy The earliest studies of HBV reactivation were in the context of systemic chemotherapy for breast cancer and lymphoma. The greatest rates of reactivation were found in patients with lymphoma, likely due to the potency of immunosuppression offered by the chemotherapy regimens, as well as the immunosuppressive effect of the underlying disease. In one prospective study of patients treated for lymphoma with a variety of chemotherapy regimens, mostly based on CHOP (cyclophosphamide, doxorubicin, vincristine, prednisolone), the rate of hepatitis attributed to HBV reactivation was 48% in HBsAg-positive patients (13 out of 27) and 4% in patients positive for anti-HBc and/or anti-HBs (2 out of 51 patients).11 A similar study of patients receiving chemotherapy for other solid tumors found HBV reactivation in 15 out of 78 HBsAgpositive patients.14 In a large meta-analysis, the risk of HBV reactivation was found to be highest with anthracycline-derived chemotherapy such as doxorubicin and epirubicin.26

Corticosteroids The negative effect of corticosteroids on HBV infection has long been documented, with early studies from the

1980s that aimed to investigate a therapeutic role for prednisolone instead showing a hastened biochemical deterioration and increased complications, including death.27 The mechanism is potentially 2-fold: firstly, the HBV genome contains a glucocorticoid-responsive transcription regulatory element which is up-regulated by corticosteroids resulting in increased viral replication, and secondly, a directly suppressive effect on cytotoxic T cells which are involved in HBV control.28 Reactivation has been reported with steroid monotherapy, with one metaanalysis concluding the risk to be at least 10% in HBsAgpositive patients receiving continuous systemic treatment for four weeks or more.25 It also concluded that doses of ≥20 mg daily of prednisolone or equivalent represented a high risk. In patients with hemato-oncological disorders, corticosteroids are often used in conjunction with other chemotherapy agents, where they have been shown to have an additive deleterious effect. In one randomized study, 50 HBsAg positive patients received the same chemotherapy regimens for NHL either with or without corticosteroids; the incidence of reactivation in the corticosteroid group was significantly higher (18 out of 25 vs. 9 out of 25, respectively), with also a higher incidence of clinically significant hepatitis.29

Anti-CD20-directed monoclonal antibodies This class of drugs is well-reported for causing severe HBV reactivation, with several published cases of fatal fulminant hepatic failure.30-34 These reports and a formal evaluation of the post-marketing data from the US Food and Drug Administration (FDA) adverse events reporting system resulted in a warning on the packaging of all monoclonal antibodies against CD20 regarding the risk of HBV reactivation. Rituximab, ofatumumab and obinutuzumab are currently licensed and predominantly used to treat Bcell malignancies. The risk of reactivation is highest for HBsAg-positive patients, and it has even been suggested that almost all will develop reactivation at some point.25 Patients with resolved HBV infection are also likely to be at high risk. In one large analysis of 326 anti-HBc positive patients receiving rituximab or obinutuzumab as part of

Table 2. Risk groups in terms of immunosuppressive regimen and the recommended management to prevent hepatitis B virus (HBV) reactivation.

Risk group

Treatment regimen

High risk (>10%)

B-cell-depleting anti-CD20-directed monoclonal antibodies (e.g. rituximab) HSCT +/- diagnosis of GvHD

Moderate risk (1-10%)

Low risk (<1%)

Systemic cancer chemotherapy - anthracycline derivatives (e.g. doxorubicin) Tyrosine kinase inhibitors (e.g. imatinib, ibrutinib) Corticosteroids ≥20mg prednisolone, ≥4 weeks Traditional immunosuppressive monotherapy (e.g. azathioprine, methotrexate)

Recommended management HBsAg positive HBsAg negative, anti-HBc positive Treatment/prophylaxis* with TDF/entecavir

Prophylaxis with lamivudine, or TDF/entecavir if expected duration >12 months

Treatment/prophylaxis* with TDF/entecavir

Prophylaxis with lamivudine, or TDF/entecavir if expected duration >12 months

Treatment/prophylaxis* with TDF/entecavir if chronic hepatitis

Monitor HBsAg, ALT and HBV DNA every 3 months

Corticosteroids ≤4 weeks *HBsAg-positive patients with evidence of chronic HBV hepatitis (see Table 1) will require active treatment of HBV infection. HSCT: hematopoietic stem cell transplantation; GvHD: graft-versus-host disease; TDF: tenofovir disoproxil fumarate; HbsAg: hepatitis B surface antigen; anti-HBc: anti-hepatitis B core antibody; ALT: alanine transaminase.

haematologica | 2019; 104(3)

437


B. Wang et al.

chemotherapy for NHL, 27 patients (8.2%) in total had HBV reactivation; of these, 17 had received obinutuzumab and 10 rituximab.35 The mechanism of action of these drugs in causing depletion of circulating B cells and partial depletion in the lymphatic system and bone marrow explains the strong association with HBV reactivation: a fall in anti-HBs antibody titers has been demonstrated in patients undergoing rituximab therapy, with reactivation seen most in those with complete loss of anti-HBs.36 A large meta-analysis of over 800 patients with resolved HBV infection receiving rituximab also demonstrated a protective effect of possessing anti-HBs, with 14% of those who were only anti-HBc positive developing reactivation compared with 5% of those also positive for antiHBs.18 The risk of reactivation associated with this class of drugs also potentially persists for longer than with other therapies. Reactivation events have been reported up to two years after the last dose of rituximab; one prospective study of 63 anti-HBc positive patients with lymphoma reported a cumulative rate of HBV reactivation of 41.5% over two years, with a median time to reactivation of 23 weeks, but a range of up to 100 weeks.12 Again, in this study, possession of anti-HBs antibodies was protective against reactivation, with the 2-year cumulative rate of reactivation being significantly higher in those negative for anti-HBs. The wide range and potential delay in presentation of reactivation mirrors the scenario of a hepatic flare after rituximab treatment of HCV-related lymphomas, and possibly relates to variation in the strength of immune control of HBV in different individuals.37 Monoclonal antibodies directed against other immune cell targets have also been associated with reactivation. Alemtuzumab, a monoclonal antibody against CD52 used for refractory chronic lymphocytic leukemia (CLL) and in HSCT conditioning regimens, has been reported to cause reverse HBsAg seroconversion and significant reactivation-related hepatitis.38,39 Furthermore, as newer agents are developed, emerging cases of HBV reactivation are also being reported, including fatalities: mogamulizumab, a treatment for T-cell lymphoma, and brentuximab vedotin, used in the treatment of refractory or relapsed Hodgkin lymphoma, are two recent examples.40-42 Daratumumab, a monoclonal antibody against CD38 which is overexpressed in B-cell hematologic malignancies, also has the potential for reactivation given its mechanism of action, but so far no reports have emerged.

Hematopoietic stem cell transplantation Hepatitis B virus reactivation in the context of allogeneic HSCT is well-recognized and represents a high risk, ranging from 40% over two years in patients with resolved infection to 70-86% over five years in HBsAg-positive recipients.43,44 Risk factors are similar to those for reactivation in general, including older age (â&#x2030;Ľ50 years) and detectable HBV DNA prior to transplant.45,46 Development of graft-versus-host disease (GvHD) also significantly increases the risk of HBV reactivation, with one study of 85 anti-HBc positive recipients of allogeneic HSCT showing a cumulative rate at two years of 79.5% compared with 21%.44 This increased risk is likely associated with the fact that patients with GvHD receive more immunosuppression therapy, and experience a delay in reconstitution of the immune system for up to 12-18 months.47 HBV reactivation after autologous HSCT is also recognized, although data are more limited. In one study of 32 HBsAg438

positive patients with NHL undergoing high-dose chemotherapy and autologous HSCT, the incidence of hepatitis due to HBV reactivation was 50%.48 In anti-HBc positive patients, the risk is predictably lower; one study found reactivation in 7 out of 107 (6.5%) patients.49 In considering HBV reactivation post HSCT, the HBV status of the donor has been shown to have significant impact. It is known that, in allogeneic HSCT, donor vaccination can result in transfer of immunity against a range of infectious antigens, including hepatitis B.50,51 More recently, it has been demonstrated that recipients of HSCT who have chronic or resolved HBV infection can also benefit from donors who have been vaccinated against HBV with a strong anti-HBs response. In one series, 3 HBsAg-positive patients received allogeneic HSCT from vaccinated donors with a high anti-HBs titer. All 3 recipients became HBsAg-negative post transplant and developed a strong humoral HBV-specific response with high titers of antiHBs antibody as well as detectable T-cell immunity.43 In 2 of the recipients, the underlying hematologic malignancy subsequently relapsed, and in each case, anti-HBs levels declined and the patients again became HBsAg positive with detectable HBV DNA. This demonstrates a unique situation where the risk of HBV reactivation may be modified not only by management of the recipient, but also by careful donor selection.

Other novel agents With the rapid expansion of treatment options for conditions such as multiple myeloma and CLL, there are novel therapies that are worthy of attention as potential causes of HBV reactivation, although evidence is limited to individual reports. Tyrosine kinase inhibitors such as imatinib and nilotinib are thought to be associated with a moderate risk of HBV reactivation.52-55 As tyrosine kinase receptor-mediated signaling pathways are involved in immune activation and proliferation of lymphocytes, it is not unexpected that therapeutic blockade of these pathways may suppress immune control of HBV and result in reactivation. The newer agents ibrutinib (a Bruton tyrosine kinase inhibitor) and idelalisib (a PI3K tyrosine kinase inhibitor) are both B-cell receptor signaling modulators used for the treatment of CLL and certain NHLs. Both have been associated with cases of HBV reactivation; a recent recommendation has been issued by manufacturers of ibrutinib acknowledging the risk of HBV reactivation and advising serological testing for HBV prior to starting treatment.56 Bortezomib, a proteasome inhibitor that has revolutionized the medical management of multiple myeloma, and ruxolitinib, an inhibitor of Janus-activated kinases (JAK) used for treatment of myelofibrosis have both been associated with reported cases of HBV reactivation.57-59 Bortezomib therapy is often given prior to autologous HSCT, and along with the immune dysfunction associated with multiple myeloma itself, it is difficult to isolate the specific risk attributable to bortezomib alone. Nonetheless, reactivation has been reported both in HBsAg-positive and anti-HBc positive patients.57 Immune checkpoint inhibitors as a group of drugs are increasingly used in the treatment of various non-hematologic cancers such as melanoma, renal cell carcinoma and hepatocellular carcinoma.60 Their role in treating hematologic malignancies such as Hodgkin lymphoma is also evolving. Their mechanism of action in overcoming T-cell haematologica | 2019; 104(3)


HBV reactivation

Figure 1. Key points summarizing recommendations. HSCT: hematopoietic stem cell transplantation; HBV: hepatitis B virus; anti-HBc: anti-hepatitis B core; anti-HBs: anti-hepatitis B surface; HBsAg: hepatitis B surface antigen.

dysfunction results in an association with immune-related side effects; hepatotoxicity is not uncommon and usually relates to an autoimmune-type hepatitis.61 Hepatitis related to exacerbation of HBV infection has also been reported in patients who were subsequently found to be HBsAg positive.62,63 The risk of reactivation in those who are only anti-HBc positive is thought to be low.9 Interestingly, checkpoint inhibitors are also being investigated in the treatment of chronic hepatitis B by potentially overcoming the T-cell exhaustion that is observed.64 Venetoclax, a small molecule inhibitor of BCL-2 which is over-expressed in malignant B cells, is used in refractory cases of CLL. No specific cases of HBV reactivation have been reported, but as venetoclax decreases total white cell counts and can cause lymphopenia in addition to neutropenia, it may be capable of inducing reactivation. Similarly, azacitidine and decitabine are hypomethylating agents which have been increasingly used to treat acute myeloid leukemia (AML), especially in the elderly. Their potential to cause myelosuppression raises the potential risk of reactivation, but as yet no reports have emerged. With all these novel agents, reports or series of their use and safety in patients with chronic or resolved hepatitis B are warranted.

Management of hepatitis B virus reactivation Screening Ideal management and prevention of HBV reactivation includes both stringent identification of at risk patients prior to initiation of any immunosuppressive therapy, and appropriate consideration of prophylactic antiviral treathaematologica | 2019; 104(3)

ment. As the majority of people with chronic or past HBV are not aware of their infection, it has been strongly recommended by several international societies and guidelines that all patients should be screened for HBV prior to commencing any immunosuppressive therapy.4,9,25,65 Given the significant risk of reactivation associated with resolved HBV infection and certain regimens, this should include testing for both HBsAg and anti-HBc antibody. Testing for anti-HBs antibodies may also be beneficial, as those without anti-HBs are considered to be at even higher risk. However, at the moment, no recommendations have been made concerning stratifying management according to anti-HBs presence or titer. If during screening a new HBsAg-positive patient is identified, he or she should be referred to an appropriate specialist hepatitis service to undergo full assessment, regardless of the plans for immunosuppression. Assessment will focus on defining the phase of HBV infection (Table 1), differentiating between chronic HBV infection and chronic hepatitis, and staging the liver disease with a combination of imaging, liver biopsy, and/or noninvasive methods such as transient elastography (FibroScan). Patients positive for anti-HBc antibody may also require assessment, as they may still present with advanced liver disease, even in the absence of active HBV viremia. In rare cases, anti-HBc positivity may also represent true occult HBV infection, where the patient is HBsAg-negative but has positive serum HBV DNA. This can be due to virus mutations in surface antigen rendering it undetectable with standard HBsAg assays or, more commonly, strongly suppressed but active viral replication.66 Such patients are managed the same as those who are HBsAg-positive. 439


B. Wang et al.

Treatment Treatment of HBV reactivation can either be prophylactic or pre-emptive; the former approach offers antiviral treatment to all patients considered at moderate or high risk prior to commencing immunosuppression, whereas the latter involves regular monitoring of ALT, HBsAg, and HBV DNA during treatment, with antiviral therapy started when HBV DNA and/or ALT levels rise.9 Studies have compared the two strategies and found prophylactic treatment to be more effective in preventing reactivation.67-69 One study comparing prophylactic versus pre-emptive antiviral therapy in HBsAg-positive patients undergoing chemotherapy for NHL found significantly lower rates of reactivation and hepatitis in the prophylactic group (11.5% vs. 56%; P=0.001).69 Similarly, a study of 80 antiHBc positive patients treated with rituximab for lymphoma also demonstrated that prophylactic antiviral therapy resulted in lower rates of reactivation (4.3% vs. 23.9% at 18 months after chemotherapy; P=0.019).68 In considering who requires prophylactic antiviral treatment, one needs to assess the risk of reactivation, as discussed earlier, both in terms of the individual patient and the immunosuppressive therapy being considered (Table 2).

HBsAg-positive patients Following referral to a specialist service, these patients should receive nucleoside / nucleotide analog antiviral therapy prior to commencement of immunosuppression.4 This will either be active treatment in those with chronic HBV hepatitis who would require antiviral therapy in any event, or prophylaxis in those with chronic infection without hepatitis. HBV DNA and ALT should be monitored every three months throughout immunosuppression. For patients with chronic HBV infection without hepatitis, if the underlying infection remains stable following completion of immunosuppression then it may be appropriate to stop prophylactic antiviral therapy (see below: Duration of therapy). HBsAg-positive patients may also require surveillance for hepatocellular carcinoma, which would be undertaken as part of a specialist review.

Anti-HBc positive patients The risk of reactivation in this group of patients varies and management depends mostly on the immunosuppression regimen proposed. Antiviral prophylaxis is strongly recommended in all patients receiving high-risk (>10%) regimens such as rituximab or HSCT. Those receiving low-risk (<1%) regimens, such as short course low-dose corticosteroids, do not require prophylaxis and can be monitored with regular ALT, HBsAg, and HBV DNA testing. There is some debate regarding those in the intermediate category of moderate risk (1-10%), with some recommending prophylaxis25 and others recommending monitoring and a pre-emptive approach.4 Patients in this group may require an individualized evaluation and consideration of other factors, such as co-morbidities and the likely duration of immunosuppression required. For HBsAg-positive patients requiring active or prophylactic antiviral treatment, third-generation nucleoside / nucleotide analogs such as tenofovir disoproxil fumarate (TDF) and entecavir are generally recommended.4 Until recently, the recommendation for anti-HBc positive patients has been to use lamivudine for prophylaxis. Large meta-analyses have demonstrated the efficacy of lamivudine in significantly reducing the rate of reactivation and 440

related mortality in the setting of systemic chemotherapy.70,71 However, lamivudine is associated with drug-resistant HBV mutants due to its relatively inferior antiviral potency, and rates of lamivudine resistance have been shown to be as high as 56% after two years of treatment.72 Given that some of the immunosuppressive treatments discussed above may continue for more than one year or be given in repeated courses, prolonged antiviral prophylaxis and resistance may be a significant issue. Therefore, there has been a shift in recommendation towards use of TDF or entecavir for prophylaxis in anti-HBc positive patients if immunosuppression is likely to be prolonged or with very high-risk regimens. The efficacy of entecavir in prophylaxis against HBV reactivation has been demonstrated in large comparator studies with lamivudine, whereas the data for TDF are more limited.73,74 One study of HBsAg-positive patients receiving rituximab-based therapy for diffuse large B-cell lymphoma showed significantly lower rates of HBV reactivation in those treated with entecavir (6.6% vs. 30%), as well as lower rates of chemotherapy disruption. Therefore, current recommendations suggest lamivudine should only be considered for anti-HBc positive patients requiring short duration (< 12 months) prophylaxis in the setting of moderate- or lowrisk immunosuppression.

Safety Extensive data regarding long-term TDF and entecavir use in terms of efficacy and safety have been reported in the setting of management of chronic HBV infection. The main concerns surrounding long-term TDF use are renal toxicity with reduction in glomerular filtration rate, and bone toxicity with a decline in bone mineral density.75 These are potentially of additional concern in patients with hematologic disorders, and long-term safety data specifically in such patients are lacking. As for any patient on TDF, close monitoring of renal function with measurement of estimated glomerular filtration rate and serum phosphate is recommended. Recently, tenofovir alafenamide (TAF), a prodrug that results in lower circulating plasma levels of tenofovir, has been licensed for use in chronic HBV infection in patients with TDF-related toxicity and/or renal co-morbidities. Results at 96 weeks of a randomized, double blind study have demonstrated noninferiority compared with TDF in terms of antiviral potency with an improved safety profile.76-78 In patients with existing renal/bone disease, or in whom the potential toxicity associated with TDF is unacceptable, entecavir or TAF are good alternatives.4

Duration of therapy For patients with chronic HBV-related hepatitis, antiviral therapy with nucleoside / nucleotide analogs is long-term. In patients for whom antiviral therapy was started for prophylaxis alone, the evidence to support recommendations on the total duration required is not robust and such recommendations are based on cases of when reactivation has occurred.25 Antiviral prophylaxis for a minimum of six months after completion of chemotherapy or immunosuppression is recommended; in the case of rituximab and Bcell depleting therapies, as reactivation has been reported later, the recommendation is for prophylaxis to continue for a minimum of 12 months. The situation following HSCT is more complex and depends on occurrence of complications such as GvHD, and the viral status of the recipient and haematologica | 2019; 104(3)


HBV reactivation

donor. Decisions to stop prophylaxis will likely need to be made on an individual basis, aided by regular monitoring of viral markers, including anti-HBs antibody titers. For all patients, monitoring including ALT, HBsAg and HBV DNA should continue for at least six months after prophylaxis has been stopped, as cases of reactivation after stopping antiviral treatment have been reported.79

Hepatitis B virus vaccination Given the importance of anti-HBs in potentially reducing the risk of HBV reactivation, the role of HBV vaccination has been investigated, specifically in the setting of HSCT.80,81 In one study of 46 patients with past HBV infection undergoing HSCT, 21 patients received a standard 3dose regimen of HBV vaccine post transplant. None of these patients developed HBV reverse seroconversion compared with 12 out of 25 patients in the non-vaccine group (P=0.0003), even after a median follow up of 67 months. Further studies are required, but vaccination, in its simplicity, presents an attractive strategy to manage reactivation.

Conclusion Hepatitis B virus reactivation is not uncommon in patients with hematologic disorders and malignancies,

References 1. Keam B, Lee JH, Im SA, et al. Why, when, and how to prevent hepatitis B virus reactivation in cancer patients undergoing chemotherapy. J Natl Comp Canc Netw. 2011;9(5):465-477. 2. Schweitzer A, Horn J, Mikolajczyk RT, et al. Estimations of worldwide prevalence of chronic hepatitis B virus infection: a systematic review of data published between 1965 and 2013. Lancet. 2015;386(10003):15461555. 3. Health Protection Agency. Health Protection Report. 2012. 4. European Association for the Study of the Liver. EASL 2017 Clinical Practice Guidelines on the management of hepatitis B virus infection. J Hepatol. 2017;67(2):370398. 5. Liang TJ, Block TM, McMahon BJ, et al. Present and future therapies of hepatitis B: from discovery to cure. Hepatology. 2015;62(6):1893-1908. 6. Bertoletti A, Kennedy PT. The immune tolerant phase of chronic HBV infection: new perspectives on an old concept. Cell Mol Immunol. 2015;12(3):258-263. 7. Chu CJ, Hussain M, Lok AS. Quantitative serum HBV DNA levels during different stages of chronic hepatitis B infection. Hepatology. 2002;36(6):1408-1415. 8. Fong TL, Di Bisceglie AM, Gerber MA, et al. Persistence of hepatitis B virus DNA in the liver after loss of HBsAg in chronic hepatitis B. Hepatology. 1993;18(6):1313-1318. 9. Loomba R, Liang TJ. Hepatitis B reactivation associated with immune suppressive and biological modifier therapies: current concepts, management strategies and future directions. Gastroenterology. 2017;152(6): 1297-1309.

haematologica | 2019; 104(3)

and ongoing reports of significant morbidity, or even fatalities, make this an important topic to understand and manage appropriately. Given the efficacy of antiviral prophylaxis, the key to preventing reactivation is identifying patients at risk. All patients need to be screened for HBV, including HBsAg and anti-HBc antibody, before any immunosuppressive therapy is initiated. Furthermore, it is important to assess and understand the risk posed by individual treatment regimens in order to determine the need for antiviral prophylaxis and the duration of treatment. For those patients identified as requiring antiviral treatment or prophylaxis, a shift towards newer nucleoside / nucleotide analogs tenofovir and entecavir is recommended, especially in high-risk and prolonged regimens. Areas worthy of future research include identifying more sensitive ways of stratifying risk, especially for the large number of patients with resolved HBV infection; these may be serological viral markers that better reflect the burden of latent forms of HBV DNA in the liver and their transcriptional activity, or, indeed, immune markers that indicate the strength or type of immune control and whether they are likely to be affected by immunosuppression. Better risk stratification may allow us to be more selective about who requires prophylaxis, but until then, a low threshold approach is required to prevent the significant morbidity and mortality associated with reactivation.

10. Hoofnagle JH. Reactivation of hepatitis B. Hepatology. 2009; 49(5):156-165. 11. Lok AS, Liang RH, Chiu EK, et al. Reactivation of hepatitis B virus replication in patients receiving cytotoxic therapy. Report of a prospective study. Gastroenterology. 1991;100(1):182-188. 12. Seto WK, Chan TS, Hwang YY, et al. Hepatitis B reactivation in patients with previous hepatitis B virus exposure undergoing rituximab-containing chemotherapy for lymphoma: a prospective study. J Clin Oncol. 2014;32(33):3736-3743. 13. Hammond SP, Borchelt AM, Ukomadu C, et al. Hepatitis B virus reactivation following allogeneic hematopoietic stem cell transplantation. Biol Blood Marrow Transplant. 2009;15(9):1049-1059. 14. Yeo W, Chan PK, Zhong S, et al. Frequency of hepatitis B virus reactivation in cancer patients undergoing cytotoxic chemotherapy: a prospective study of 626 patients with identification of risk factors. J Med Virol. 2000;62(3):299-307. 15. Yuen MF, Wong DK, Fung J, et al. HBsAg seroclearance in chronic hepatitis B in Asian patients: replicative level and risk of hepatocellular carcinoma. Gastroenterology. 2008;135(4):1192-1199. 16. Lau GK, Leung YH, Fong DY, et al. High hepatitis B virus (HBV) DNA viral load as the most important risk factor for HBV reactivation in patients positive for HBV surface antigen undergoing autologous hematopoietic cell transplantation. Blood. 2002;99(7): 2324-2330. 17. Perrillo RP. Acute flares in chronic hepatitis B: the natural and unnatural history of an immunologically mediated liver disease. Gastroenterology. 2001;120(4):1009-1022. 18. Paul S, Dickstein A, Saxena A, et al. Role of surface antibody in hepatitis B reactivation

19.

20.

21.

22.

23.

24.

25.

26.

in patients with resolved infection and hematologic malignancy: a meta-analysis. Hepatology. 2017;66(2):379-388. De Monte A, Courjon J, Anty R, et al. Direct-acting antiviral treatment in adults infected with hepatitis C virus: reactivation of hepatitis B virus coinfection as a further challenge. J Clin Virol. 2016;78:27-30 Wang C, Ji D, Chen J, et al. Hepatitis due to reactivation of hepatitis B virus in endemic areas among patients with hepatitis C treated with direct-acting antiviral agents. Clin Gastroenterol Hepatol. 2017;15(1):132-136. Calvaruso V, Ferraro D, Licata A, et al. HBV reactivation in patients with HCV/HBV cirrhosis on treatment with direct acting antivirals. J Viral Hepat. 2018;25(1):72-79. Nath A, Agarwal R, Malhotra P, et al. Prevalence of hepatitis B virus infection in non-Hodgkin lymphoma: a systematic review and meta-analysis. Intern Med J. 2010;40(9):633-641. Marcucci F, Mele A. Hepatitis viruses and non-Hodgkin lymphoma: epidemiology, mechanisms of tumorigenesis, and therapeutic opportunities. Blood. 2011;117(6): 1792-1798. Marcucci F, Spada E, Mele A, et al. The association of hepatitis B virus infection with Bcell non-Hodgkin lymphoma â&#x20AC;&#x201C; a review. Am J Blood Res. 2012;2(1):18-28. Perrillo RP, Gish R, Falck-Ytter YT. American Gastroenterological Association Institute technical review on prevention and treatment of hepatitis B virus reactivation during immunosuppressive drug therapy. Gastroenterology. 2015;148(1):221-244. Paul S, Saxena A, Terrin N, et al. Hepatitis B virus reactivation and prophylaxis during solid tumour chemotherapy: A systematic review and meta-analysis. Ann Intern Med. 2016;164(1):30-40.

441


B. Wang et al. 27. Lam KC, Lai CL, Trepo C, et al. Deleterious effect of prednisolone in HBsAg-positive chronic active hepatitis. N Engl J Med. 1981;304(7):380-386. 28. Tur-Kaspa R, Shaul Y, Moore DD, et al. The glucocorticoid receptor recognizes a specific nucleotide sequence in hepatitis B virus DNA causing increased activity of the HBV enhancer. Virology. 1988;167(2):630-633. 29. Cheng AL, Hsiung CA, Su IJ, et al. Steroidfree chemotherapy decreases risk of hepatitis B virus (HBV) reactivation in HBV-carriers with lymphoma. Hepatology. 2003;37(6): 1320-1328. 30. Westhoff TH, Jochimsen F, Schmittel A, et al. Fatal hepatitis B virus reactivation by an escape mutant following rituximab therapy. Blood. 2003;102(5):1930. 31. Sarrecchia C, Cappelli A, Aiello P. HBV reactivation with fatal fulminating hepatitis during rituximab treatment in a subject negative for HBsAg and positive for HBsAb and HBcAb. J Infect Chemother. 2005;11(4):189191. 32. Law JK, Ho JK, Hoskins PJ, et al. Fatal reactivation of hepatitis B post-chemotherapy for lymphoma in a hepatitis B surface antigennegative, hepatitis B core antibody-positive patient: potential implications for future prophylaxis recommendations. Leuk Lymphoma. 2005;46(7):1085-1089. 33. Yamagata M, Murohisa T, Tsuchida K, et al. Fulminant B hepatitis in a surface antigen and hepatitis B DNA-negative patient with diffuse large B-cell lymphoma after CHOP chemotherapy plus rituximab. Leuk Lymphoma. 2007;48(2):431-433. 34. Dillon R, Hirschfield GM, Allison ME, et al. Fatal reactivation of hepatitis B after chemotherapy for lymphoma. BMJ. 2008;337:a423. 35. Kusumoto S, Arcaini L, Hong X, et al. Risk fo HBV reactivation in patients with B-cell lymphomas receiving obinutuzumab or rituximab immunochemotherapy. Blood. 2019;133(2):137-146. 36. Pei SN, Ma MC, Wang MC, et al. Analysis of hepatitis B surface antibody titers in B cell lymphoma patients after rituximab therapy. Ann Hematol. 2012;91(7):1007-1012. 37. Visco C, Finotto S. Hepatitis C virus and diffuse large B-cell lymphoma: Pathogenesis behavior and treatment. World J Gastroenterol. 2014;20(32):11054-11061. 38. Iannitto E, Minardi V, Calvaruso G, et al. Hepatitis B virus reactivation and alemtuzumab therapy. Eur J Haematol. 2005;74(3):254-258. 39. Moses SE, Lim ZY, Sudhanva M, et al. Lamivudine prophylaxis and treatment of hepatitis B virus-exposed recipients receiving reduced intensity conditioning hematopoietic stem cell transplants with alemtuzumab. J Med Virol. 2006;78(12):1560-1563. 40. Ifuku H, Kusumoto S, Tanaka Y, et al. Fatal reactivation of hepatitis B virus infection a patient with adult T-cell leukemia-lymphoma receiving the anti-CC chemokine receptor 4 antibody mogamulizumab. Hepatol Res. 2015;45(13):1363-1367. 41. Totani H, Kusumoto S, Ishida T, et al. Reactivation of hepatitis B virus (HBV) infection in adult T-cell leukemia-lymphoma patients with resolved HBV infection following systemic chemotherapy. Int J Hematol. 2015;101(4):398-404. 42. Yang H, Cao Z, Wang Z, et al. Hepatitis B virus reactivation induced by brentuximab vedotin in the treatment of Hodgkin lymphoma: a case report and literature review. Zhonghua Xue Ye Xue Za Zhi.

442

2014;35(10):949-950. 43. Seto WK, Chan TS, Hwang YY, et al. Hepatitis B reactivation in occult viral carriers undergoing hematopoietic stem cell transplantation: A prospective study. Hepatology. 2017;65(5):1451-1461. 44. Lindemann M, Koldehoff M, Fiedler M, et al. Control of hepatitis B virus infection in hematopoietic stem cell recipients after receiving grafts from vaccinated donors. Bone Marrow Transplant. 2016;51(3):428431. 45. Hwang JP, Lok AS. Management of patients with hepatitis B who require immunosuppressive therapy. Nat Rev Gastroenterol Hepatol. 2014;11(4):209-219. 46. Jun CH, Kim BS, Oak CY, et al. HBV reactivation risk factors in patients with chronic HBV infection with low replicative state and resolved HBV infection undergoing hematopoietic stem cell transplantation in Korea. Hepatol Int. 2017;11(1):87-95. 47. Socie G, Ritz J. Current issues in chronic graft-versus-host disease. Blood. 2014;124 (3):374-384. 48. Huang H, Cai Q, Lin T, et al. Lamivudine for the prevention of hepatitis B virus reactivation after high-dose chemotherapy and autologous hematopoietic stem cell transplantation for patients with advanced or relapsed non-Hodgkinâ&#x20AC;&#x2122;s lymphoma single institution experience. Expert Opin Pharmacother. 2009;10(15):2399-2406. 49. Varma A, Biritxinaga L, Saliba RM, et al. Impact of hepatitis B core antibody seropositivity on the outcome of autologous hematopoietic stem cell transplantation for multiple myeloma. Biol Blood Marrow Transplant. 2017;23(4):581-587. 50. Harris AE, Styczynski J, Bodge M, et al. Pretransplant vaccinations in allogeneic stem cell transplantation donors and recipients: an often-missed opportunity for immunoprotection? Bone Marrow Transplant. 2015;50 (7):899-903. 51. Lindemann M, Barsegian V, Runde V, et al. Transfer of humoral and cellular hepatitis B immunity by allogeneic hematopoietic cell transplantation. Transplantation. 2003;75(6): 833-838. 52. Ikeda K, Shiga Y, Takahashi A, et al. Fatal hepatitis B reactivation in a chronic myeloid leukemia patient during imatinib mesylate treatment. Leuk Lymphoma. 2006;47(1): 155-157. 53. Lai GM, Yan SL, Chang CS, et al. Hepatitis B reactivation in chronic myeloid leukaemia patients receiving tyrosine kinase inhibitor. World J Gastroenterol. 2013;19(8):13181321. 54. Lakhani S, Davidson L, Priebat DA, et al. Reactivation of chronic hepatitis B infection related to imtinib mesylate therapy. Hepatol Int. 2008;2(4):498-499. 55. Kang BW, Lee SJ, Moon JH, et al. Chronic myeloid leukemia patient manifesting fatal hepatitis B virus reactivation during treatment with imatinib rescued by liver transplantation: case report and literature review. Int J Haematol. 2009;90(3):383-387. 56. De Jesus Ngoma P, Kabamba B, Dahlqvist G, et al. Occult HBV reactivation induced by ibrutinib treatment: a case report. Acta Gastroenterol Belgica. 2015;78(4):424-426. 57. Li J, Huang B, Li Y, et al. Hepatitis B virus reactivation in patients with multiple myeloma receiving bortezomib-containing regiments followed by autologous stem cell transplant. Leuk Lymphoma. 2015;56(6): 1710-1717. 58. Shen CH, Hwang CE, Chen YY, et al. Hepatitis B virus reactivation associated

59.

60.

61.

62.

63.

64. 65.

66. 67.

68.

69.

70.

71.

72.

73.

74.

75. 76.

with ruxolitinib. Ann Haematol. 2014;93 (6):1075-1076. Caocci G, Murgia F, Podda L, et al. Reactivation of hepatitis B virus infection following ruxolitinib treatment in a patient with myelofibrosis. Leukemia. 2014;28(1): 225-227. Michot JM, Bigenwald C, Champiat S, et al. Immune-related adverse events with immune checkpoint blockade: a comprehensive review. Eur J Cancer. 2016;54:139148. Sanjeevaiah A, Kerr T, Beg MS. Approach and management of checkpoint inhibitorrelated immune hepatitis. J Gastrointest Oncol. 2018;9(1):220-224. Koksal AS, Toka B, Eminler AT, et al. HBVrelated acute hepatitis due to immune checkpoint inhibitors in a patient with malignant melanoma. Ann Oncol. 2017;28 (12):3103-3104. Pandey A, Ezemenari S, Liaukovich M, et al. A rare case of Pembrolizumab-induced reactivation of hepatitis B. Case Rep Oncol Med. 2018;2018:5985131. Bertoletti A, Le Bert N. Immunotherapy for chronic hepatitis B virus infection. Gut Liver. 2018;12(5):497-507. Chaganti S, Illidge T, Barrington S, et al. Guidelines for the management of diffuse large B-cell lymphoma. Br J Haematol. 2016;174(1):43-56. Squadrito G, Spinella R, Raimondo G. The clinical significance of occult HBV infection. Ann Gastroenterol. 2014;27(1):15-19. Huang YH, Hsiao LT, Hong YC, et al. Randomised controlled trial of entecavir prophylaxis for rituximab-associated hepatitis B virus reactivation in patients with lymphoma and resolved hepatitis B. J Clin Oncol. 2013;31(22):2765-2772. Lau GK, Yiu HH, Fong DY, et al. Early is superior to deferred preemptive lamivudine therapy for hepatitis B patients undergoing chemotherapy. Gastroenterology. 2003;125 (6):1742-1749. Hsu C, Hsiung CA, Su IJ, et al. A revisit of prophylactic lamivudine for chemotherapyassociated hepatitis B reactivation in nonHodgkinâ&#x20AC;&#x2122;s lymphoma: a randomised trial. Hepatology. 2008;47(3):844-853. Loomba R, Rowley A, Wesley R, et al. Systematic review: the effect of preventive lamivudine on hepatitis B reactivation during chemotherapy. Ann Intern Med. 2008;148(7):519-528. Kohrt HE, Ouyang DL, Keefe EB. Systematic review: lamivudine prophylaxis for chemotherapy-induced reactivation of chronic hepatitis B virus infection. Aliment Pharmacol Ther. 2006;24(7):1003-1016. Lok AS, Hussain M, Cursano C, et al. Evolution of hepatitis B virus polymerase gene mutations in hepatitis B e antigen-negative patients receiving lamivudine therapy. Hepatology. 2000;32(5):1145-1153. Huang H, Li X, Zhu J, et al. Entecavir vs lamivudine for prevention of hepatitis B virus reactivation among patients with untreated diffuse large B-cell lymphoma receiving R-CHOP chemotherapy: a randomized control trial. JAMA. 2014;312(23): 2521-2530. Hilgendorf I, Loebermann M, Borchert K, et al. Tenofovir for treatment of hepatitis B virus reactivation in patients with chronic GVHD. Bone Marrow Transplant. 2011;46 (9):1274-1275. Woodward CLN, Hall AM, Williams IG, et al. Tenofovir-associated renal and bone toxicity. HIV Med. 2009;10(8):482-487. Agarwal K, Brunetto M, Seto WK, et al. 96

haematologica | 2019; 104(3)


HBV reactivation weeks treatment of tenofovir alafenamide vs. tenofovir disoproxil fumarate for hepatitis B virus infection. J Hepatol. 2018;68(4): 672-681. 77. Buti M, Gane R, Seto WK, et al. Tenofovir alafenamide versus tenofovir disoproxil fumarate for the treatment of patients with HBeAg-negative chronic hepatitis B virus infection: a randomised, double-blind, phase 3, non-inferiority trial. Lancet Gastroenterol Hepatol. 2016;1(3):196-206. 78. Chan HL, Fung S, Seto WK, et al. Tenofovir alafenamide versus tenofovir disoproxil

haematologica | 2019; 104(3)

fumarate for the treatment for HBeAg-positive chronic hepatitis B virus infection: a randomised, double-blind, phase 3, non-inferiority trial. Lancet Gastroenterol Hepatol. 2016;1(3):185-195. 79. Myers RP, Swain MG, Urbanski SJ, et al. Reactivation of hepatitis B e antigen-negative chronic hepatitis B in a bone marrow transplant recipient following lamivudine withdrawal. Can J Gastroenterol. 2001;15 (9):599-603. 80. Takahata M, Hashino S, Onozawa M, et al. Hepatitis B virus (HBV) reverse seroconver-

sion (RS) can be prevented even in nonresponders to hepatitis B vaccine after allogeneic stem cell transplantation: long-term analysis of intervention in RS with vaccine for patients with previous HBV infection. Transpl Infect Dis. 2014; 16(5):797-801. 81. Onozawa M, Hashino S, Darmanin S, et al. HB vaccination in the prevention of viral reactivation in allogeneic hematopoietic stem cell transplantation recipient with previous HBV infection. Biol Blood Marrow Transplant. 2008;14(11):1226-1230.

443


REVIEW ARTICLE Ferrata Storti Foundation

IgG4-related disease: what a hematologist needs to know Luke Y.C. Chen,1 Andre Mattman,2 Michael A. Seidman2,3 and Mollie N. Carruthers3

Division of Hematology, Department of Medicine, University of British Columbia; Department of Pathology and Laboratory Medicine, St. Paulâ&#x20AC;&#x2122;s Hospital and 3Division of Rheumatology, Department of Medicine, University of British Columbia, Vancouver, BC, Canada

1 2

Haematologica 2019 Volume 104(3):444-455

ABSTRACT

I

Correspondence: LUKE Y.C. CHEN lchen2@bccancer.bc.ca Received: October 23, 2018. Accepted: January 7, 2019. Pre-published: January 31, 2019.

doi:10.3324/haematol.2018.205526

gG4-related disease is a fibro-inflammatory condition that can affect nearly any organ system. Common presentations include major salivary and lacrimal gland enlargement, orbital disease, autoimmune pancreatitis, retroperitoneal fibrosis and tubulointerstitial nephritis. This review focuses on the hematologic manifestations of IgG4-related disease, including lymphadenopathy, eosinophilia, and polyclonal hypergammaglobulinemia. The disease can easily be missed by unsuspecting hematologists, as patients may present with clinical problems that mimic disorders such as multicentric Castleman disease, lymphoma, plasma cell neoplasms and hypereosinophilic syndromes. When IgG4-related disease is suspected, serum protein electrophoresis and IgG subclasses are helpful as initial tests but a firm histological diagnosis is essential both to confirm the diagnosis and to rule out mimickers. The central histopathological features are a dense, polyclonal, lymphoplasmacytic infiltrate enriched with IgG4-positive plasma cells (with an IgG4/IgG ratio >40%), storiform fibrosis, and obliterative phlebitis. Importantly for hematologists, the latter two features are seen in all tissues except bone marrow and lymph nodes, making these two sites suboptimal for histological confirmation. Many patients follow an indolent course and respond well to treatment, but a significant proportion may have highly morbid or fatal complications such as peri-aortitis, severe retroperitoneal fibrosis or pachymeningitis. Corticosteroids are effective but cause new or worsening diabetes in about 40% of patients. Initial response rates to rituximab are high but durable remissions are rare. More intensive lymphoma chemotherapy regimens may be required in rare cases of severe, refractory disease, and targeted therapy against plasmablasts, IgE and other disease biomarkers warrant further exploration.

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/104/3/444 Š2019 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

444

Example case An 80-year old Korean man was referred for evaluation of chronic lymphadenopathy, eosinophilia and polyclonal hypergammaglobulinemia. He had had abdominal pain since the 1970s and was initially thought to have Crohn disease, subsequently complicated by idiopathic common bile duct narrowing. In the 1990s, he was found to have a kidney mass on computed tomography which was suspected to be lymphoma, but after the mass was resected, the histology was in keeping with multifocal fibrosclerosis. At the time of referral, his physical examination revealed a low, firm, slightly enlarged, left submandibular gland, no lacrimal gland swelling, and multiple 2 cm or smaller soft inguinal lymph nodes bilaterally. His white blood cell count was 8.3x109/L, eosinophil count 2.0x109/L (normal values <0.7x109/L), creatinine concentration 140 Âľmol/L, total protein 87 g/L (normal values <82 g/L), with a polyclonal increase in gamma globulins on serum protein electrophoresis of 20.5 g/L (normal values <14 g/L), and total IgG of 28.9 g/L (normal values <18.5 g/dL). haematologica | 2019; 104(3)


IgG4-related disease for hematologists

Introduction Immunoglobulin G4-related disease (IgG4-RD) is a chronic immune-mediated disease that may present with tumefactive lesions, fibrosis, and a polyclonal IgG4-positive (IgG4+) plasma cell-enriched infiltrate in nearly any anatomic site. In many centers, systemic therapy is guided by rheumatologists, yet nearly every medical, surgical and pathology specialty must be aware of this entity and its protean manifestations. Involvement of blood-forming and lymphoid organs, manifesting as lymphadenopathy, eosinophilia, and polyclonal hypergammaglobulinemia, is common and IgG4-RD often mimics other hematologic conditions such as multicentric Castleman disease, lymphoma, plasma cell neoplasms, and hypereosinophilic syndromes (HES). This review provides an overview of IgG4-RD with a focus on aspects most relevant to clinical hematology practice. In the early 2000s, while searching for non-invasive biomarkers to distinguish sclerosing (autoimmune) pancreatitis from pancreatic cancer, Japanese investigators noticed a fast-moving band in the beta-gamma region of the serum protein electrophoresis of patients with sclerosing pancreatitis. This band represented markedly elevated serum IgG4.1 Subsequently, abundant polyclonal IgG4+ plasma cells were found within a lymphoplasmacytic infiltrate in tissue samples from patients with autoimmune pancreatitis and in surrounding tissues including the liver and gallbladder.2 Once this entity was recognized as a distinct disease with characteristic histological features, many historically “idiopathic” and eponymous disorders such as multifocal fibrosclerosis (mediastinal and retroperitoneal fibrosis), Kuttner tumor (chronic sclerosing sialadenitis) and Reidel thyroiditis (woody infiltration of the thyroid) were found to be part of the IgG4-RD spectrum.3 Numerous names were proposed in the early days of its discovery, including “IgG4-related sclerosing disease”, “IgG4-related systemic disease”, “IgG4-related multiorgan lymphoproliferative syndrome” and “systemic IgG4-related plasmacytic syndrome”. An international group of investigators, primarily from Japan, the USA and Europe met in Boston in 2011 and agreed upon uniform nomenclature and diagnostic criteria.3 The accepted name “IgG4-related disease” reflects the universality of the IgG4+ plasma cell infiltration in involved organs as well as the frequency of elevation in serum IgG4 rather than a pathogenic role for IgG4 per se.4 The wide variety of disease manifestations stems not only from multi-organ involvement, but also from the fact that different organs may be involved in a metachronous fashion. Common presentations are major salivary (parotid and submandibular) and lacrimal gland enlargement (Mickulicz disease), lymphadenopathy, orbital pseudotumor, pancreatitis, sclerosing cholangitis, retroperitoneal fibrosis and tubulointerstitial nephritis.5

Epidemiology and pathophysiology Underrecognition has hindered accurate estimates of the epidemiological burden of this disease, but a starting point is the prevalence of autoimmune pancreatitis in Japan, which increased from 2.2/100 000 in 2007 to 4.6/100 000 in 2011. The increase is almost certainly due to increased recognition, and given that pancreatic haematologica | 2019; 104(3)

involvement is present in about 20-25% of IgG4-RD cases, the true prevalence of the disease is likely much higher. There is a 2:1 male preponderance and the median age of affected patients at diagnosis is in the sixth to seventh decade of life. Aside from one case report of identical twins with IgG4-RD,6 evidence of genetic susceptibility is sparse. Pediatric cases are rare, but a recent review identified 25 cases in children, of whom 11 had orbital disease and three had autoimmune pancreatitis.7 At first glance, the presence of IgG4 in serum and IgG4+ plasma cells in tissues, increased IgG4+ plasmablasts in serum, and responsiveness to rituximab suggest that B-cell activation drives the disease.8 However, the IgG4 antibody itself is not thought to be pathogenic since it does not bind complement, does not traditionally form immune complexes, and patients with other conditions with markedly elevated serum IgG4, such as IgG4 myeloma, do not develop features of IgG4-RD.9,10 Recent studies have demonstrated that an unconventional population of CD4+SLAMF7+ cytotoxic T lymphocytes is central to the pathogenesis of the disease.11 Histopathologically, polyclonal B cells are found in clusters near these CD4+ T cells, the latter of which are among the most abundant cells within affected tissues. Oligoclonal expansion of these CD4+ cytotoxic T lymphocytes in peripheral blood may explain the high rates of T-cell clonality positivity determined by polymerase chain reaction analysis.12 These CD4+ T cells produce profibrotic cytokines such as interleukin-1, transforming growth factor-beta and interferongamma, as well as cytolytic molecules such as granzyme A and B and perforin.13 These cytotoxic T lymphocytes are likely sustained by continuous antigen presentation by B cells and plasmablasts. The responsible autoantigen(s) have not been definitively identified, but annexin A11 and galactin-3 have both recently been implicated in IgG4RD.14,15 The autoantibody response to galactin-3 is primarily of the IgG4 and IgE isotypes, which correlates with the typical immunoglobulin responses seen in IgG4-RD.

Clinical presentation IgG4-RD can affect nearly any organ except synovial tissue. This “fibro-inflammatory” disease presents with tumefactive (puffy) inflammatory infiltrates and fibrosis with a predilection for glandular tissue. Figure 1 illustrates the manifestations of IgG4-RD by organ system. The most common organ manifestations from two large cohorts, one Japanese and one American, are shown in Table 1.16,17 In patients presenting with the better-known features of IgG4-RD, such as autoimmune pancreatitis, orbital disease and major salivary gland involvement, the disease tends to be recognized and histologically confirmed earlier, but patients referred to hematologists may present with less obvious features of IgG4-RD, and a high index of suspicion is needed to arrive at an accurate diagnosis. Common reasons for referral to a hematologist include lymphadenopathy, eosinophilia, and polyclonal hypergammaglobulinemia.

Lymphadenopathy IgG4-related lymphadenopathy is one of the three most common manifestations of IgG4-RD, affecting 30-60% of patients with IgG4-RD in most large cohorts (Table 1).5,16,17 IgG4-lymphadenompathy may be generalized or local445


L.Y.C. Chen et al.

Figure 1. Manifestations of IgG4-related disease by organ system. The most common primary disease features are indicated in bold.

ized (the latter typically contiguous with involved organs such as pancreas and lungs). Parallel enlargement of salivary, lacrimal and parotid glands is common. Five morphological subtypes, all of which display increased IgG4+ plasma cells, have been described (see Figure 2 for examples):18-20 (i) Multi-centric Castleman disease-like: preserved nodal architecture with patent sinusoids and hyperplastic follicles; abundant mature plasma cells in interfollicular areas with some eosinophils, similar to lymphadenopathy in multicentric Castleman disease or autoimmune disease. (ii) Reactive follicular hyperplasia: increased IgG4+ plasma cells in germinal centers and often in the interfollicular zones with some eosinophils present. (iii) Interfollicular expansion pattern: expanded interfollicular zones with small lymphocytes, plasmablasts and mature plasma cells and eosinophils which may resemble lymphoma (e.g. angioimmunoblastic lymphoma). Examples are given in Figure 2A-C. (iv) Progressive transformation of germinal center-like: 446

Table 1. Clinical characteristics of patients from two large published cohorts.

Japan (n=334)17

Boston (125)16

Mean age at diagnosis Male sex Ethnicity Elevated serum IgG4 Mean number of organs involved (range) Affected organs Salivary glands

63.8 years 61.4% 100% Japanese >95% 3.2 (1-11)

55.2 years 60.8% 76% White 51% 2.3 (1-7)

72.3%

Lacrimal glands/orbit Lymph nodes Pancreas Retroperitoneal/ aorta

57.1% 56.5% 25.5 24.9

Kidney Lung

23.7% 23.4%

28% (submandibular) + 16.8% (parotid) 22.4% 27.2% 19.2% 18.4% (retroperitoneal) + 11.2% (aorta) 12% 17.6%

haematologica | 2019; 104(3)


IgG4-related disease for hematologists

IgG4-related lymphadenopathy has aptly been called both “an underdiagnosed and overdiagnosed entity”:19 underdiagnosed because if it is not included in the differential diagnosis, IgG4 and IgG stains may not be done and the disease may be missed, and overdiagnosed because increased IgG4+ plasma cells may be seen in conditions ranging from Rosai-Dorfman-Destombes disease to inflammatory vasculitis.21,22 Although not the optimal tissue for making the histological diagnosis of IgG4-RD, in a patient with typical clinical features such as autoimmune pancreatitis or retroperitoneal fibrosis, a lymph node biopsy may be sufficient for diagnosis if biopsy of other affected organs is not feasible. Given the low specificity of increased IgG4+ plasma cells in lymph node and the variable histological patterns, the greatest utility of lymph node biopsy is perhaps excluding other diagnoses, such as lymphoma and HHV8-associated Castleman disease. The role of lymph node biopsy is further discussed in the section on “Diagnosis and staging”.

trointestinal tract, and lymph nodes.12 Idiopathic HES and hypereosinophilia of unknown significance are diagnoses of exclusion, and account for a substantial proportion (3050%) of diagnoses of patients evaluated for eosinophilia.2426 Evaluating these patients for IgG4-RD is an important and underappreciated aspect of their care. In fact, we previously published a case report with a diagnostic label of idiopathic HES, reviewed by several world experts in eosinophilia who concurred with the diagnosis, which was subsequently found to be IgG4-RD.27,28 Findings of a myeloid clonal disorder such as increased blast cells, abnormal karyotype, mutations in PDGFR-alpha/beta, FGFR-1 and PCM1-JAK2 are not seen in IgG4-RD. However, differentiating lymphocytic-variant HES from IgG4-RD can be more challenging. The aberrant T-cell phenotypes found in lymphocytic-variant HES, including increased CD4+CD3–, CD3+/CD4–/CD8– and CD4+/CD7– T cells, with or without T-cell clonality as determined by polymerase chain reaction analysis, have all been reported in IgG4-RD.12,29 Increased IgG4 deposits have been found in tissue samples from adult and pediatric patients with eosinophilic esophagitis.30-33 In contrast to HES and chronic eosinophilic leukemia, the eosinophilia secondary to IgG4-RD is typically mild to moderate, rarely exceeding 5x109/L and typically quite evanescent, being ablated by steroids or rituximab therapy.

Eosinophilia

Polyclonal hypergammaglobulinemia

Approximately 40% of patients with IgG4-RD have peripheral blood eosinophilia, often accompanied by asthma and atopy.23 Thus, IgG4-RD is an important and underappreciated cause of reactive or secondary eosinophilia.12 HES and IgG4-RD commonly involve the skin, lungs, gas-

As for eosinophilia, IgG4-RD represents an important new diagnostic consideration in patients with hypergammaglobulinemia. An elevated serum IgG4 level, often accompanied by an increase in IgG1, causes polyclonal hypergammaglobulinemia. Rarely, this elevation can be

scattered larger or transformed follicles containing plasma cells. An example is given in Figure 2D,E. (v) Inflammatory pseudotumor-like: lymph node partially effaced by a fibro-inflammatory infiltrate and storiform fibrosis; this subtype is considered most specific for IgG4RD in lymph nodes. An example is given in Figure 2F.

A

B

C

D

E

F

Figure 2. Lymph nodes in IgG4-related disease. (A,B) An example of the interfollicular pattern of IgG4-related lymphadenopathy, with mature plasma cells, many expressing IgG4, distributed between benign follicles. (A) Hematoxylin and eosin stain. (B) IgG4 immunohistochemistry. (C) A needle core lymph node biopsy from a different case with the interfollicular pattern (hematoxylin and eosin stain). (D,E) A case of IgG4-lymphadenoapthy with a progressive transformation of the follicular center pattern, with the plasma cells within the follicle proper. (D) Hematoxylin and eosin stain. (E) IgG4 immunohistochemistry. (F) An example of a mass-like lesion (inflammatory pseudotumor) with dense fibrosis and associated follicular hyperplasia in a case of IgG4-lymphadenoapthy (hematoxylin and eosin).

haematologica | 2019; 104(3)

447


L.Y.C. Chen et al.

sufficient to result in polyclonal hyperviscosity syndrome.27,28,34 It is not known what causes the exuberant production of IgG4 immunoglobulins, currently considered an epiphenomenon rather than a contributor to the pathogenesis of the disease.8 Not surprisingly, the serum free light chains tend to be abnormally elevated.35 IgE is often markedly increased, particularly in patients with eosinophilia and atopy, whereas IgA and IgM levels are normal or modestly elevated. Serum IgG4 typically runs in the fast gamma or beta-gamma region on the serum protein electrophoresis, and thus the typical electrophoretic profile of a patient with IgG4-RD demonstrates polyclonal hypergammaglobulinemia with beta-gamma bridging. This sometimes prominent pattern is dependent on IgG4 concentration and is highlighted in Figure 3. The hypergammaglobulinemia can be mistaken for a polyclonal increase in IgA, monoclonal gammopathy of undetermined significance or “biclonal” IgG kappa and lambda gammopathy, as laboratory physicians may not be familiar with the dense bands of IgG-lambda and kappa which in fact represent polyclonal IgG4.36,37 Some patients have even been treated for myeloma before subsequently being found to have IgG4-RD as the cause of their protein abnormalities, plasmacytosis and renal disease.9,38 Laboratory physicians must, therefore, consider the differential diagnosis of beta-gamma bridging and order or suggest additional investigations to clarify clonality and heavy chain composition where necessary.39,40 Prior to the recognition of IgG4-RD, a large case series of 130 patients with polyclonal hypergammaglobulinemia >30 g/L on serum protein electrophoresis showed that the most common single diagnoses were liver disease (79/130, 66%), connective tissue disease (28/130, 22%), chronic infection (8/130, 6%) and hematologic disorders (7/130, 5%).41 In a recent, single-center study of 70 patients with polyclonal increases in IgG ≥20 g/L it was found that 14 (20%) had IgG4-RD as the cause of their hypergammaglobulinemia, indicating that a substantial proportion of patients with hypergammaglobulinemia have IgG4-RD as the underlying cause.42 The discovery of IgG4-RD has also led to increased recognition of other IgG subclass elevations with specific diseases, such as hepatitis C and mono-

clonal gammopathy of undetermined significance with IgG1, hypothyroidism and irritable bowel syndrome with IgG2, rheumatoid arthritis with increased IgG3 and IgG1, and celiac disease with IgG4.43 IgG4 myeloma has been described but is rare; one large case series found that 6/158 bone marrow biopsies in myeloma patients expressed IgG4, in keeping with the relatively small fraction of overall circulating gamma globulins normally made up by the IgG4 subtype.10 One case report of IgG4 subtype POEMS has been reported.44 The bone marrow morphology may mimic myeloma with florid plasmacytosis,9,45 although in our experience, bone marrow examination is very insensitive for the diagnosis of IgG4-RD, with many cases showing no increase in plasma cells or lymphocytes despite florid hypergammaglobulinemia.

Important mimickers of IgG4-related disease The diagnostic challenge of IgG4-RD for hematologists is heightened by overlap of clinical and laboratory features with those of a number of other hematologic diseases including lymphoma, plasma cell neoplasms, and histiocyte disorders (Table 2). In addition to the diseases discussed in this section and presented in Table 2, there are numerous non-hematologic mimickers reviewed in detail elsewhere.46 Careful review of histological specimens and correlation with clinical, laboratory and radiological findings are crucial for solidifying the correct diagnosis. Multicentric Castleman disease and IgG4-RD show considerable overlap given the high frequency of lymphadenopathy, IgG4+ plasma cell-enriched tissue infiltrates and elevated serum IgG4 levels seen in both diseases.18 However, IgG4-RD typically affects older patients, rarely exhibits the “hyper-interleukin-6” systemic inflammatory features of multicentric Castleman disease such as fever and elevated C-reactive protein, and the histological features are distinct. The histiocytic disorders RosaiDorfman-Destombes disease and Erdheim-Chester disease both cause inflammatory mass lesions that can mimic IgG4-RD. Histopathological evaluation of RosaiDorfman-Destombes disease can show enrichment of IgG4+ plasma cells,22,47 but typically in the context of

Figure 3. Serum protein electrophoresis showing electrophoretic patterns for four patients with mild to gross elevations in IgG4 concentration in between two for patients with low IgG4 concentration. The physicochemical properties of the IgG4 heavy chain result in a relative anodal position (shift toward albumin) of the gamma globulins when the IgG4 becomes the predominant gamma globulin. Apart from IgG4, IgA immunoglobulins are frequently observed in the boundary between the beta and gamma regions. Monoclonal bands may also migrate in this region as shown in the gel (the monoclonal gammopathy in this case is an IgG1 monoclonal band that has physicochemical properties that are atypical for IgG1 immunoglobulins, which are normally found in a more cathodal position). NC: normal control, MG: monoclonal gammopathy.

448

haematologica | 2019; 104(3)


IgG4-related disease for hematologists

CD68+ S100+ histiocytes, often associated with emperipolesis. The most recent classification of the histiocyte disorders recommends evaluating suspected cases of Rosai-Dorfman-Destombes disease for increased IgG4+

plasma cells,48 but in the absence of other evidence for a common pathophysiological link, Rosai-DorfmanDestombes disease is not considered part of the spectrum of IgG4-RD or vice versa.49 One third of patients with

Table 2. Diseases that mimic hematologic manifestations of IgG4-RD (lymphadenopathy, eosinophilia and polyclonal hypergammaglobulinemia).

Mimicker of IgG4-related disease

Areas of overlap

Distinguishing features of the mimicker not typically seen in IgG4-related disease

Multicentric Castleman disease (MCD)

• Lymphadenopathy (particularly the MCD-like variant of IgG4-LAD) • High serum IgG4 • IgG4+ plasma cells in tissue* • Polyclonal hypergammaglobulinemia • Lymphadenopathy • Polyclonal hypergammaglobulinemia • Polyclonal plasmacytosis, including IgG4+ plasma cells in bone marrow and skin* • Lymphadenopathy • Extranodal RDD can involve the nasal cavity, and retro-orbital tissue • Meningeal involvement of RDD may mimic IgG4-pachymeningitis • Salivary gland involvement • CNS involvement including pachymeningitis • Increased IgG4+ plasma cells in tissue* • Retroperitoneal fibrosis • Central nervous system/hypophysitis • Pulmonary involvement

• MCD is a “hyper-IL-6” syndrome associated with B symptoms and highly elevated CRP and IL-6 not seen in IgG4-RD

Cutaneous and systemic plasmacytosis (CSP)

Rosai-Dorfman-Destombes disease (RDD, a.k.a “Sinus Histiocytosis with Massive Lymphadenopathy; “R” group histiocytosis)

Erdheim Chester disease (ECD; “L” group histiocytosis)

Malignant lymphoma

Sarcoidosis

Hypereosinophilic syndrome (HES)/chronic eosinophilic leukemia (CEL) [particularly lymphocyte-variant HES]

Plasma cell myeloma/monoclonal gammopathy of undetermined significance

Vasculitis, particularly eosinophilic granulomatosis with polyangiitis and granulomatosis with polyangiitis

• Lymphadenopathy • Extranodal mass lesions • Blood and tissue eosinophilia • Lymphadenopathy • Pulmonary nodules • Pachymeningitis and/or hypophysitis • Polyclonal hypergammaglobulinemia • Multi-organ involvement • T-cell clonality by PCR • Atopy/asthma/elevated IgE • Lymphadenopathy • Eosinophilia • Elevated serum IgG4 • Aberrant T cell phenotype in peripheral blood (CD4+/3-, CD4+/7–, CD3+/4–/8–) • Plasma cell infiltrate • Hypergammaglobulinemia • Renal failure • Proteinuria • Eosinophilia in blood and tissue • Polyclonal hypergammaglobulinemia with elevated serum IgG4 • Peri-aortitis and rarely, Kawasaki-like coronary arteritis are seen in IgG4-RD • Multi-organ involvement, including respiratory, gastrointestinal and renal structures

• IgG4-RD rarely involves skin whereas cutaneous lesions (round/oval, red/brown poorly circumscribed macules, papules and plaques) are an obligatory feature of CSP; serum IgG4 in CSP is normal or mildly elevated • Massive cervical lymphadenopathy is atypical for IgG4-RD • RDD may present with B symptoms • Large histiocytic cells with hypochromatic nuclei and pale cytoplasm; emperipolesis; positive for S100, CD68, CD14 and CD163

• >95% of ECD patients have bone involvement, which is rare in IgG4-RD (apart from rare IgG4+ angiocentric eosinophilic fibrosis of the head and neck) • Yellow peri-orbital xanthelasmas common in ECD (skin involvement in IgG4-RD is rare, and when present tends to be erythematous or flesh-colored papules) • Foamy multi-nucleated histiocytes, few Touton cells, fibrosis; CD68+, CD163+, CD1a– • 50% of ECD patients are BRAF V600E-positive • B symptoms, bone involvement, brain parenchymal involvement and hypercalcemia are rare in IgG4-RD • Non-caseating granulomas • Hypercalcemia and elevated ACE levels

• Increased blasts or myeloid clone in CEL • PDGFR-alpha/PDGFR-beta/FGFR1/PCM1-JAK2 positivity are not seen in IgG4-RD • More marked and persistent eosinophilia in HES

• Plasma cell clonality • Monoclonal paraprotein ± suppression of polyclonal immunoglobulins • Lytic boney disease/hypercalcemia • Light chains in urine rather than albuminuria • Highly elevated CRP in vasculitis • Extravascular granulomas, small-medium vessel vasculitis • Mononeuritis multiplex not a feature of IgG4-RD

*Although increased IgG4+ plasma cells have been described in these diseases, the absolute counts and IgG4/IgG ratio are typically well below the thresholds for the diagnosis of IgG4RD and the other key features of IgG4-RD (storiform fibrosis and obliterative phlebitis) are not seen.3 IgG4-LAD: IgG4-lymphadenopathy; IL-6: interleukin-6; CRP: C-reactive protein; IgG4RD: IgG4-related disease; CNS: central nervous system; ACE: angiotensin-converting enzyme; PCR: polymerase chain reaction.

haematologica | 2019; 104(3)

449


L.Y.C. Chen et al.

Erdheim-Chester disease have retroperitoneal fibrosis, which may raise the suspicion of IgG4-RD; however, more than 95% of patients with Erdheim-Chester disease have skeletal involvement, which, aside from rare cases of IgG4-related angiocentric eosinophilic fibrosis (midline destructive lesions of the head and neck), is generally not seen in IgG4-RD.50 The histology of Erdheim-Chester disease shows a CD68+S100–CD1a– histiocyte infiltration often with “foamy histiocytes”.51 Extra-pulmonary sarcoidosis may share clinical features similar to those of IgG4-RD, including polyclonal hypergammaglobulinemia, lymphadenopathy, pulmonary nodules, sclerosing mesenteritis and pachymeningitis. The association between IgG4-RD and malignant lymphoma has been studied extensively. In Asian patients, mucosal-associated lymphoid tissue (MALT) lymphoma, particularly of ophthalmological tissues has been described, whereas in western populations a variety of histologies (diffuse large B-cell, follicular, lymphoplasmacytic, and MALT) have been reported.52,53 There are also case reports of IgG4-RD concomitant with autoimmune lymphoproliferative syndrome.54,55

Case continued The patient was suspected to have IgG4-RD, so serum IgG subclasses were analyzed. His serum IgG4 level was markedly elevated at 11.6 g/L (normal values <1.35). The tissue blocks from his previous nephrectomy were retrieved and pathology review showed a lymphoplasmacytic infiltrate, moderate tissue eosinophilia and interstitial fibrosis and atrophy. Staining for IgG4 and IgG revealed abundant IgG4+ plasma cells with >40 IgG4+ plasma cells per high power field and an IgG4/IgG ratio >40%. Computed tomography scans of the neck, chest, abdomen and pelvis revealed multiple pulmonary nodules, carinal lymphadenopathy and a soft tissue density encasing the main coronary arteries, previous right nephrectomy and pancreatic atrophy. His IgG4-RD Responder Index activity score was 12.

Diagnosis and staging A careful history and thorough physical examination, attending to clues such as a history of waxing and waning glandular swelling, sicca symptoms and unexplained pancreatitis or jaundice must be accompanied by serum protein studies. Histological confirmation of the disease is essential, and once a diagnosis has been established, investigations to assess for symptomatic and subclinical organ involvement, such as early retroperitoneal fibrosis and albuminuria, are important for management planning. The IgG4-RD Responder Index is a standardized, validated tool for the evaluation of disease activity at initial evaluation and subsequent follow up.56,57 Suggested laboratory and imaging tests are summarized in Table 3.

Serum protein studies Approximately 70% of patients with IgG4-RD have an elevated serum IgG4 level. Serum IgG subclasses should be investigated in conjunction with serum protein electrophoresis to exclude a monoclonal paraprotein (Figure 3). Serum IgG4 level has a diagnostic sensitivity ranging from 83-97% and specificity from 60-85% with a general cut-off of “above the upper limit of normal”.58-60 Typically, 1.35 g/L is used as the biomarker cut-off for IgG4-RD (which corresponds to the upper limit of normal for one 450

common commercial method but not another) without specifying the methodology. While mildly elevated serum IgG4 can be seen in many conditions, markedly elevated serum IgG4 (>5 g/L) is approximately 90% specific for IgG4-RD. Aside from methodological differences, serum IgG4 levels in IgG4-RD can vary greatly depending on ethnicity and degree of organ involvement. In the Boston cohort of patients (76% White), only 53 of 103 patients had elevated serum IgG4 levels.16 In contrast, in a cohort of 334 Japanese patients, more than 95% had elevated serum IgG4.17 In our multi-ethnic cohort, we found that Asians have a higher serum IgG4 than non-Asians (median 11.2 g/L versus 2.9 g/L, P=0.0094), and elevated serum IgG4 had a sensitivity of 96% in Asians compared to 67% in non-Asians.61 Patients with multi-organ involvement or of Asian ethnicity typically have elevated serum IgG4, sometimes markedly so, such as the patient in this illustrative case. The serum IgG4/IgG ratio is typically >0.2 in patients with IgG4-RD, although the ratio does not increase the diagnostic specificity of serum IgG4 alone. Flow cytometric detection of plasmablasts may offer a more sensitive modality for diagnosing IgG4-RD, with a reported sensitivity of 95% and specificity of 82% using a cut-off of 900/mL.62 However, the flow cytometry method used to detect plasmablasts is not widely available. Most centers use immunonephelometry to measure IgG subclasses, which can cause some challenges with interpretation. The two most common immunonephelometric methods (Siemens and Binding Site) correlate well with regard to IgG4, but the absolute IgG4 values differ by approximately 50% at the upper limit of normal.63 IgG4 levels may also be markedly under-reported in cases of extreme IgG4 elevations due to the hook effect. The hook effect, or prozone phenomenon, occurs when an excessive amount of analyte prevents binding of the capture antibody in a sandwich assay, yielding a falsely low or normal result. Erroneously low measurements of serum IgG4 reported in the literature reflect this error.64 Furthermore, IgG4 itself interferes with the nephelometric measurement of IgG1 and IgG2, in particular, which can obscure the immunoglobulin profile that would otherwise highlight the disproportionate elevation of serum IgG4.65 Because of the traditional errors in immunonephelometry, some have mistakenly reported increased serum IgG2 levels as a marker of IgG4-RD.66-68 Our group has recently demonstrated that mass spectrometry is an alternative that eliminates these analytical errors and is more costeffective than immunonephelometry.65

Histopathology A firm diagnosis of IgG4-related disease requires histopathological confirmation, except in the case of autoimmune pancreatitis, in which radiological features (diffuse “sausage-like” enlargement of the pancreas with featureless borders and delayed enhancement with or without a capsule-like rim or “halo”) may be sufficiently specific to exclude requirement for tissue biopsy.3,69 As in sarcoidosis, in which non-caseating granulomas may be seen in any of the organs affected by the disease, IgG4-RD demonstrates common histology in most of the multitude of organs that may be affected. The three major histological features of IgG4-RD in tissue are: (i) a dense, polyclonal lymphoplasmacytic infiltrate enriched with IgG4+ plasma cells; (ii) fibrosis; and (iii) obliterative phlebitis. With regards to the lymphoplasmahaematologica | 2019; 104(3)


IgG4-related disease for hematologists Table 3. Diagnostic and staging tests in IgG4-RD.

Test(s)

Typical findings

Complete blood count/differential/blood film

Eosinophilia (~40% of patients, typically mild); rouleaux formation due to polyclonal hypergammaglobulinemia Normal or moderately elevated (CRP typically <20 mg/L) in the absence of peri-aortitis or active infection Mildly elevated IgG4 <1.5-5 g/L is nonspecific, and 30% of IgG4-RD patients have normal serum IgG4 levels. Serum IgG4 >5 g/L is helpful both for diagnosis and as a disease marker. Other IgG subclasses may be moderately elevated (IgG4/IgG ratio is typically >0.2)* IgA and IgM may be normal or mildly increased; IgE may be markedly increased. Immunoglobulin suppression is atypical. SPEP and UPEP are important to rule out monoclonal proteins May be weakly positive

C-reactive protein (CRP); interleukin-6 and other markers of systemic inflammation Serum IgG subclasses

Blood and urine tests

Immunoglobulins (IgG, IgA, IgM, IgE); serum and urine protein electrophoresis (SPEP, UPEP)

Autoantibodies (e.g. antinuclear antibody, rheumatoid factor) Complements (C3/C4)

Often low, especially with tubulointerstitial nephritis

Urinalysis, random albumin/creatinine ratio

Albuminuria is common; nephrotic range proteinuria can be seen with membranoproliferative glomerulonephritis

Other markers of end organ damage: lipase, glucose and hemoglobin A1c, liver enzymes, thyroid-stimulating hormone, creatinine, urinalysis, urine albumin/creatinine ratio

Subclinical pancreatitis with elevated lipase, glucose intolerance, hepatitis and albuminuria are common

Computed tomography (CT) of the neck, chest, abdomen, pelvis

Diffuse “sausage-like” or segmental enlargement of pancreas, often with a “halo”; wedge-shaped hypodensities in kidneys; ductal organs, such as bile duct and bronchus, show diffuse “pipe-stem” wall thickening; thickened aortic wall; hepatic mass lesions; retroperitoneal fibrosis or peri-aortic cuffing Patients with orbital disease typically have lacrimal gland swelling

Imaging

If lacrimal enlargement ⇒CT to head (rule out orbital involvement) Archived specimens

New biopsy

Pathology

Lymph node and bone marrow

As long as tissue blocks are still available, the pathologist should be able to examine the histology and then order immunostaining for IgG4 and IgG if typical features are present Excisional is preferable to core biopsy when possible to look for the central pathological features: • Storiform fibrosis • Obliterative phlebitis • Polyclonal lymphoplasmacytic infiltrate with increased IgG4+ plasma cells and IgG4+/IgG+ plasma cell ratio > 40%** Consider minor salivary gland (lip) biopsy if affected organs are high risk for biopsy These tissues are unusual in that fibrosis and obliterative phlebitis are typically not seen, and thus biopsy of other tissues may be required for a definitive diagnosis. Involved lymph nodes typically have >100 IgG4+ plasma cells/hpf with an IgG4/IgG ratio >40%. Bone marrow involvement with eosinophilia and increased IgG4+ plasma cells is rare and may be absent even in patients with marked serum hypergammaglobulinemia

*IgG2 may be spuriously elevated when nephelometric measurement is used.65 **The number of IgG4+ cells/hpf required varies, depending on the tissue, from >10/hpf in meningeal tissue to >200/hpf in skin. In all tissues, the ratio of IgG4+ to IgG+ plasma cells should be ≥40%.3 Hpf: high-power field.

haematologica | 2019; 104(3)

451


L.Y.C. Chen et al.

cytic infiltrate, the number of IgG4+ plasma cells per highpower field (hpf) considered diagnostic varies according to tissue site, from >10/hpf in meninges to >100/hpf in skin. Regardless of the site, the ratio of IgG4+/IgG+ plasma cells is >40% in IgG4-RD. Fibrosis is a histological requirement for the diagnosis of IgG4-RD and should be arranged at least focally in a storiform pattern. Storiform fibrosis is a swirling, “cartwheel” pattern of fibrosis which may have a patchy distribution and can, therefore, be missed with small biopsies. In the obliterative phlebitis of IgG4-RD, venous channels are obliterated by an inflammatory lymphoplasmacytic infiltrate. Expert pathologists recommend looking for arteries/arterioles where the accompanying venous vessel is not readily apparent and may in fact have been replaced by an inflammatory infiltrate; elastin stains may be helpful in identifying completely obliterated vessels. Other histopathological features include phlebitis without obliteration of the lumen and increased number of eosinophils. As in the illustrative case, archival specimens may be used to confirm a diagnosis, and many patients will have previous biopsies available due to their chronic disease course. As long as a tissue block is still available, IgG4 and IgG stains can be done. When a new biopsy is required, excisional specimens are preferred over core needle biopsies to allow for proper assessment. There are currently no established criteria for the cytological diagnosis of IgG4-RD from a fine needle aspirate. Increased numbers of IgG4+ plasma cells are seen in all tissues affected by IgG4-RD, but this is not, in itself, a specific finding. Many chronic inflammatory conditions such as vasculitis, inflammatory bowel disease and lymphoma may exhibit increased numbers of IgG4+ plasma cells but do not share the other histological features of storiform fibrosis, obliterative phlebitis and absence of granulomatous inflammation.19,21 Unfortunately, for the purposes of hematologists, bone marrow involvement is uncommon in IgG4-RD and obliterative phlebitis and storiform fibro-

A

C 452

sis (not to be confused with myelofibrosis) are not typically seen in bone marrow and lymph nodes.70 Furthermore, even when involved, lymph nodes and bone marrow may not show robust elevation in IgG4-expressing plasma cells as compared to the total IgG population, or the findings may be only focal. Patterns of IgG4-RD have not been well established in the bone marrow, but the presence of mature plasma cells would be supportive of marrow involvement, once plasma cell neoplasms are excluded. Some examples of marrow involvement in IgG4-RD are shown in Figure 4. In general, only organs with clinical or radiological evidence of involvement are likely to show diagnostic features on biopsy, and thus biopsy should be directed at affected organ(s). Patients with proteinuria or renal lesions on imaging may require kidney biopsy, and renal involvement may demonstrate two distinct histological patterns: hypocomplementic tubulointerstitial nephritis in 80% of cases, and membranoproliferative glomerulonephritis in 20% of cases.70 In patients in whom affected organs are not amenable to biopsy, minor salivary gland (lip) biopsy should be considered. Even without clinical evidence of major salivary gland swelling or sicca symptoms, minor salivary gland biopsy can be a minimally invasive way to reach a histological diagnosis in some patients. One study of 66 patients with suspected IgG4-RD reported a sensitivity of 55% and specificity of 100% for labial salivary gland biopsy.71 From a pragmatic perspective, patients who have classic clinical, laboratory and radiological manifestations of IgG4-RD but are too frail to undergo an attempted biopsy attempt, or in whom small biopsies yield insufficient diagnostic material,72 can be given a working diagnosis of “suspected IgG4-RD” and treated as such provided reasonable efforts have been made to exclude mimickers of IgG4-RD. Splenic involvement in IgG4-RD remains an enigma. Overt splenomegaly and splenic lesions are rare in confirmed cases of IgG4-RD. A rare entity known as scleros-

B

D

Figure 4. Bone marrow specimens involved by IgG4-related disease. Both cases show mature plasma cells distributed throughout the marrow. Ancillary studies established that these plasma cells were polyclonal, excluding a plasma cell neoplasm. (A,C) Hematoxylin and eosin stains. (B,D) IgG4 immunohistochemistry.

haematologica | 2019; 104(3)


IgG4-related disease for hematologists

ing angiomatoid transformation of the spleen (SANT) is known to be enriched with IgG4+ plasma cells, but whether this condition is part of the spectrum of IgG4-RD remains unclear, as many patients with sclerosing angiomatoid transformation of the spleen do not exhibit other features of IgG4-RD.73 The inability to biopsy splenic tissue short of full splenectomy hinders the histological characterization of IgG4-RD in this organ.

Imaging and staging Once histopathological confirmation of the diagnosis has been obtained, the disease can be staged by computed tomography of the chest, abdomen and pelvis.74 Although the orbit is a commonly involved organ in IgG4-RD, the absence of dacryoadenitis is suggestive that there is no underlying orbital pseudotumor and therefore dedicated head or orbital computed tomography is not always necessary.75 Positron emission tomography/computed tomography showed greater sensitivity for the detection of disease in arteries, salivary glands and lymph nodes in a study of 21 patients but further evaluation is needed to clarify which patients benefit from the combined tomographic method over conventional imaging.76 It is important to check the urinary albumin/creatinine ratio and serum C3/C4 levels to assess for renal involvement. IgG4-RD often behaves indolently, but accurate diagnosis and staging are crucial because some patients have asymptomatic but potentially organ- or life-threatening disease such as retroperitoneal fibrosis, peri-aortitis or coronary arteritis, the last of which may develop suddenly. Moreover, fibrotic disease is typically irreversible, so early treatment is important.

Case continued Given the patient’s multi-organ involvement, particularly of the coronary arteries, he was treated with two doses of rituximab 1 g administered intravenously 2 weeks apart. His IgG4 level improved from 11.6 g/L to 5.84 g/L after 6 months. A repeat computed tomography scan of the chest and abdomen showed improvement of his coronary artery vasculitis, pulmonary nodules and lymphadenopathy and his post-treatment IgG4-RD Responder Index activity score was 3.

sion for some patients.80-82 In the absence of prospective clinical trials, an international consensus guideline on IgG4-RD treatment had only a 46% expert agreement on whether disease-modifying anti-rheumatic drugs should be started from the outset of treatment or not.83 Rituximab has been shown to be highly effective for IgG4-RD, with a response rate of 97% (29 out of 30 patients) in one prospective trial.74 The majority of patients did not require concomitant corticosteroids from the time of enrollment. However, rituximab is difficult to access due to cost, particularly outside of the USA, and remissions are often short-lived. In a French database study of 33 patients who received rituximab, a clinical response was seen in 29/31 (93.5%) patients, but 13/31 (41.9%) responders relapsed during a mean follow up of 24.8 months (with the relapses occurring at a mean of 19 months after rituximab therapy).84 The severe infection rate was estimated to be 12.1/100 patient years and three patients had hypogammaglobulinemia <5 g/L. In our practice, we often keep patients on a low dose of maintenance prednisone to maintain remission after rituximab treatment. A number of emerging therapeutic options show promise. An open label, phase 2 clinical trial of a humanized anti-CD19 antibody (Xmab®5871) showed that this treatment decreased the IgG-RD Responder Index score by ≥2 points in 12/12 patients who completed the study.85 Successful use of lymphoma chemoimmunotherapy regimens such as fludarabine and rituximab and bendamustine and rituximab have been reported for two steroidand rituximab-refractory cases.28 Whether these patients needed T-cell directed therapy or simply more potent immunosuppression than can be achieved by steroids or rituximab alone requires further exploration. Elotuzumab is an enticing agent because of the expression of SLAMF7 both on circulating plasmablasts and on the CD4+ cytotoxic T lymphocytes thought to drive the disease process. Anti-IgE therapy with omalizumab is also a potential target for those with severe atopic disease or asthma and elevated IgE levels.

Conclusions Treatment Steroids are the first-line therapy for most patients, with an overall response rate of 93% and complete response rate of 66% in a phase 2 trial of 44 patients with IgG4-RD from Japan.77 The regimen used in this trial was prednisone 0.6 mg/kg/day initially with a gradual decrease of 5 mg every 2 weeks.77 Patients with higher eosinophil count, lacrimal gland involvement, five or more organs involved, and higher IgG4-RD Responder Index scores were at higher risk of glucocorticoid failure in a Chinese cohort of 215 patients.78 Some suggest a maintenance dosage of prednisone 5 mg daily for autoimmune pancreatitis specifically, which, as one would expect, decreases relapse rates compared to placebo.79 However, the toxicity associated with steroids is well known, with 40% of patients experiencing new or worsening diabetes. Blood glucose levels should be checked regularly in all patients receiving steroid therapy and many of our patients are also followed by a diabetes specialist. Disease-modifying anti-rheumatic drugs are not very effective for induction of remission but may have a role in maintenance of remishaematologica | 2019; 104(3)

IgG4-RD is an important condition for hematologists to recognize, both because of its common hematologic manifestations of lymphadenopathy, eosinophilia and polyclonal hypergammaglobulinemia, as well as its overlap with other hematologic inflammatory and neoplastic diseases. When IgG4-RD is suspected, measurement of serum IgG subclasses is a simple, non-invasive screening test; levels >5 g/L (normal <1.35 g/L) are highly suspicious for IgG4-RD. Regardless of serum IgG4 level, the definitive diagnosis requires histology, preferably in an affected organ other than lymph node or bone marrow, to confirm IgG4-RD and to exclude its many mimics. Early recognition and treatment with steroids, rituximab, or other immunosuppressive therapies, is essential to prevent complications such as fibrosis, peri-aortitis, and renal failure.

Key points • IgG4-RD is an important cause of eosinophilia, lymphadenopathy and polyclonal hypergammaglobulinemia. 453


L.Y.C. Chen et al.

Hematologists should include IgG4-RD in the differential diagnosis of these abnormalities. • Other common manifestations of IgG4-RD include autoimmune pancreatitis, obstructive jaundice, orbital pseudotumor, lacrimal and salivary gland swelling, retroperitoneal fibrosis, and tubulointerstitial nephritis. • Serum protein electrophoresis and IgG subclass evaluation should be performed in patients with suspected IgG4-RD. Serum IgG4 levels are elevated in approximately 70% of cases. Mildly increased serum IgG4 (1.5-5 g/L) is a non-specific finding, but a markedly elevated level (>5 g/L) is 90% specific for IgG4-RD. • A firm diagnosis requires histological confirmation based on the International Consensus Criteria which

References 1. Hamano H, Kawa S, Horiuchi A, et al. High serum IgG4 concentrations in patients with sclerosing pancreatitis. N Engl J Med. 2001;344(10):732-738. 2. Kamisawa T, Funata N, Hayashi Y, et al. A new clinicopathological entity of IgG4-related autoimmune disease. J Gastroenterol. 2003;38(10):982-984. 3. Deshpande V, Zen Y, Chan JK, et al. Consensus statement on the pathology of IgG4-related disease. Mod Pathol. 2012;25(9):1181-1192. 4. Stone JH, Khosroshahi A, Deshpande V, et al. Recommendations for the nomenclature of IgG4-related disease and its individual organ system manifestations. Arthritis Rheum. 2012;64(10):3061-3067. 5. Mahajan VS, Mattoo H, Deshpande V, Pillai SS, Stone JH. IgG4-related disease. Annu Rev Pathol. 2014;9:315-347. 6. Grados A, Vaysse T, Ebbo M, Carbonnel F, Schleinitz N. IgG4-related disease in monozygotic twins: a case report. Ann Intern Med. 2017;166(2):153-155. 7. Karim F, Loeffen J, Bramer W, et al. IgG4related disease: a systematic review of this unrecognized disease in pediatrics. Pediatr Rheumatol Online J. 2016;14(1):18. 8. Stone JH, Zen Y, Deshpande V. IgG4-related disease. N Engl J Med. 2012;366(6):539-551. 9. Tang SH, Lin MH, Du JS, Liu YC, Hsiao HH, Liu TC. IgG4-related disease with bone marrow involvement mimicking multiple myeloma. Br J Haematol. 2017;177(5):673. 10. Geyer JT, Niesvizky R, Jayabalan DS, et al. IgG4 plasma cell myeloma: new insights into the pathogenesis of IgG4-related disease. Mod Pathol. 2014;27(3):375-381. 11. Mattoo H, Mahajan VS, Maehara T, et al. Clonal expansion of CD4(+) cytotoxic T lymphocytes in patients with IgG4-related disease. J Allergy Clin Immunol. 2016;138(3):825-838. 12. Carruthers MN, Park S, Slack GW, et al. IgG4-related disease and lymphocyte-variant hypereosinophilic syndrome: a comparative case series. Eur J Haematol. 2016;98(4):378-387. 13. Mattoo H, Stone JH, Pillai S. Clonally expanded cytotoxic CD4(+) T cells and the pathogenesis of IgG4-related disease. Autoimmunity. 2017;50(1):19-24. 14. Hubers LM, Vos H, Schuurman AR, et al. Annexin A11 is targeted by IgG4 and IgG1 autoantibodies in IgG4-related disease. Gut. 2018;67(4):728-735. 15. Perugino CA, AlSalem SB, Mattoo H, et al. Identification of galectin-3 as an autoantigen

454

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

include a dense lymphoplasmacytic infiltrate, storiform fibrosis and obliterative phlebitis. There must be an increased number of IgG4+ plasma cells with an IgG4:IgG plasma cell ratio >40%. • Early recognition and diagnosis are essential because patients typically respond well to steroids or rituximab in the early stages of the disease, but fibrotic disease and late complications such as chronic pancreatitis are often irreversible. Acknowledgments The authors thank the clinicians and pathologists of the IgG4 West working group in Vancouver for their invaluable collaboration, and Dr. Erica Peterson for providing critical comments on the draft of this manuscript.

in patients with IgG4-related disease. J Allergy Clin Immunol. 2018 May 29. [Epub ahead of print] Wallace ZS, Deshpande V, Mattoo H, et al. IgG4-related disease: clinical and laboratory features in one hundred twenty-five patients. Arthritis Rheumatol. 2015;67(9): 2466-2475. Yamada K, Yamamoto M, Saeki T, et al. New clues to the nature of immunoglobulin G4-related disease: a retrospective Japanese multicenter study of baseline clinical features of 334 cases. Arthritis Res Ther. 2017;19(1):262. Sato Y, Kojima M, Takata K, et al. Systemic IgG4-related lymphadenopathy: a clinical and pathologic comparison to multicentric Castleman's disease. Mod Pathol. 2009;22(4):589-599. Cheuk W, Chan JK. Lymphadenopathy of IgG4-related disease: an underdiagnosed and overdiagnosed entity. Semin Diagn Pathol. 2012;29(4):226-234. Wick MR, O'Malley DP. Lymphadenopathy associated with IgG4-related disease: diagnosis & differential diagnosis. Semin Diagn Pathol. 2017;35(1):61-66. Chang SY, Keogh KA, Lewis JE, et al. IgG4positive plasma cells in granulomatosis with polyangiitis (Wegener's): a clinicopathologic and immunohistochemical study on 43 granulomatosis with polyangiitis and 20 control cases. Hum Pathol. 2013;44(11): 2432-2437. Menon MP, Evbuomwan MO, Rosai J, Jaffe ES, Pittaluga S. A subset of Rosai-Dorfman disease cases show increased IgG4-positive plasma cells: another red herring or a true association with IgG4-related disease? Histopathology. 2014;64(3):455-459. Della Torre E, Mattoo H, Mahajan VS, Carruthers M, Pillai S, Stone JH. Prevalence of atopy, eosinophilia, and IgE elevation in IgG4-related disease. Allergy. 2014;69(2): 269-272. Gotlib J. World Health Organizationdefined eosinophilic disorders: 2017 update on diagnosis, risk stratification, and management. Am J Hematol. 2017;92(11):12431259. Chen YY, Khoury P, Ware JM, et al. Marked and persistent eosinophilia in the absence of clinical manifestations. J Allergy Clin Immunol. 2014;133(4):1195-1202. Hu Z, Boddu PC, Loghavi S, et al. A multimodality work-up of patients with hypereosinophilia. Am J Hematol. 2018;93(11): 1337-1346. Chen LY, Lai EJ, Collins DR, Ostrow DN, Sreenivasan GM. A young woman with

28.

29. 30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

episodic angioedema, papilledema, and eosinophilia. Am J Hematol. 2010;85(2):124127. Chen LY, Wong PC, Noda S, Collins DR, Sreenivasan GM, Coupland R. Polyclonal hyperviscosity syndrome in IgG4-related disease and associated conditions. Clin Case Rep. 2015;3(4):217-226. Khoury P. Lymphocytic-variant hypereosinophilic syndromes. The Hematologist. 2017;14(6):6. Mohammad N, Avinashi V, Chan E, Vallance BA, Portales-Casamar E, Bush JW. Pediatric eosinophilic esophagitis is associated with increased lamina propria immunoglobulin G4-positive plasma cells. J Pediatr Gastroenterol Nutr. 2018;67(2):204209. Chen C, Chen K, Huang X, Wang K, Qian S. Concurrent eosinophilia and IgG4-related disease in a child: a case report and review of the literature. Exp Ther Med. 2018;15(3): 2739-2748. Clayton F, Fang JC, Gleich GJ, et al. Eosinophilic esophagitis in adults is associated with IgG4 and not mediated by IgE. Gastroenterology. 2014;147(3):602-609. Zukerberg L, Mahadevan K, Selig M, Deshpande V. Esophageal intrasquamous IgG4 deposits: an adjunctive marker to distinguish eosinophilic esophagitis from reflux esophagitis. Histopathology. 2016;68(7): 968-976. Wong PC, Fung AT, Gerrie AS, et al. IgG4related disease with hypergammaglobulinemic hyperviscosity and retinopathy. Eur J Haematol. 2013;90(3):250-256. Grados A, Ebbo M, Boucraut J, et al. Serum immunoglobulin free light chain assessment in IgG4-related disease. Int J Rheumatol. 2013;2013:426759. Jawad MD, Go RS, Witzig TE, Mikhael JR, Ravindran A, Murrray DL. Pseudo-monoclonal gammopathy: a report of four cases. Haematologica. 2017;102(11):e466-e469. Jacobs JF, van der Molen RG, Keren DF. Relatively restricted migration of polyclonal IgG4 may mimic a monoclonal gammopathy in IgG4-related disease. Am J Clin Pathol. 2014;142(1):76-81. Costa MS, Silva A, Costa L, Rodrigues TB, Lemos S, Garrido J. Immunoglobulin G4related disease mimicking multiple myeloma. Port J Nephrol Hypert. 2018;32(3):217222. Finn WG, Gulbranson R, Fisher S, et al. Detection of polyclonal increases in immunoglobulin G4 subclass by distinct patterns on capillary serum protein electrophoresis: diagnostic pitfalls and clinical

haematologica | 2019; 104(3)


IgG4-related disease for hematologists

40.

41.

42.

43.

44. 45. 46. 47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

observations in a study of 303 cases. Am J Clin Pathol. 2016;146(3):303-311. McCudden CR, Jacobs JFM, Keren D, Caillon H, Dejoie T, Andersen K. Recognition and management of common, rare, and novel serum protein electrophoresis and immunofixation interferences. Clin Biochem. 2017;51:72-79. Dispenzieri A, Gertz MA, Therneau TM, Kyle RA. Retrospective cohort study of 148 patients with polyclonal gammopathy. Mayo Clin Proc. 2001;76(5):476-487. Zhao EJ, Carruthers MN, Li CH, Mattman A, Chen LYC. Prevalence of IgG4-Related Disease in Patients with Hypergammaglobulinemia. XXXVII World Congress of the International Society of Hematology (ISH 2018); Sept 14, 2018, 2018; Vancouver, British Columbia, Canada. Engelhart S, Glynn RJ, Schur PH. Disease associations with isolated elevations of each of the four IgG subclasses. Semin Arthritis Rheum. 2017;47(2):276-280. Zheng M, Zhou P, Zheng K, et al. A special subtype of POEMS syndrome: IgG4 subtype. Am J Transl Res. 2016;8(2):588-596. Zhu J, Wu B. Immunoglobulin G4-positive plasma cell myeloma. Blood. 2015;126(19): 2254. Kamisawa T, Zen Y, Pillai S, Stone JH. IgG4related disease. Lancet. 2014;385(9976): 1460-1471. Zhang X, Hyjek E, Vardiman J. A subset of Rosai-Dorfman disease exhibits features of IgG4-related disease. Am J Clin Pathol. 2013;139(5):622-632. 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. Abla O, Jacobsen E, Picarsic J, et al. Consensus recommendations for the diagnosis and clinical management of RosaiDorfman-Destombes disease. Blood. 2018;131(26):2877-2890. Deshpande V, Khosroshahi A, Nielsen GP, Hamilos DL, Stone JH. Eosinophilic angiocentric fibrosis is a form of IgG4-related systemic disease. Am J Surg Pathol. 2011;35(5):701-706. Estrada-Veras JI, O'Brien KJ, Boyd LC, et al. The clinical spectrum of Erdheim-Chester disease: an observational cohort study. Blood Adv. 2017;1(6):357-366. Sato Y, Ohshima K, Ichimura K, et al. Ocular adnexal IgG4-related disease has uniform clinicopathology. Pathol Int. 2008;58(8):465470. Bledsoe JR, Wallace ZS, Stone JH, Deshpande V, Ferry JA. Lymphomas in IgG4-related disease: clinicopathologic features in a Western population. Virchows Arch. 2017;472(5):839-852. van de Ven A, Seidl M, Drendel V, et al. IgG4-related disease in autoimmune lymphoproliferative syndrome. Clin Immunol. 2017;180:97-99. Langan RC, Gill F, Raiji MT, et al. Autoimmune pancreatitis in the autoimmune lymphoproliferative syndrome (ALPS): a sheep in wolves' clothing? Pancreas. 2013;42(2):363-366. Carruthers MN, Stone JH, Deshpande V, Khosroshahi A. Development of an IgG4-

haematologica | 2019; 104(3)

57.

58.

59.

60.

61. 62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

RD Responder Index. Int J Rheumatol. 2012;2012:259408. Wallace ZS, Khosroshahi A, Carruthers MD, et al. An international, multi-specialty validation study of the IgG4-related disease responder index. Arthritis Care Res (Hoboken). 2018;70(11):1671-1678. Carruthers MN, Khosroshahi A, Augustin T, Deshpande V, Stone JH. The diagnostic utility of serum IgG4 concentrations in IgG4related disease. Ann Rheum Dis. 2015;74(1):14-18. Culver EL, Sadler R, Simpson D, et al. Elevated serum IgG4 levels in diagnosis, treatment response, organ involvement, and relapse in a prospective IgG4-related disease UK cohort. Am J Gastroenterol. 2016;111 (5):733-743. Masaki Y, Kurose N, Yamamoto M, et al. Cutoff values of serum IgG4 and histopathological IgG4+ plasma cells for diagnosis of patients with IgG4-related disease. Int J Rheumatol. 2012;2012:580814. Qi R, Chen LYC, Park S, et al. Utility of serum IgG4 levels in a multiethnic population. Am J Med Sci. 2018;355(1):61-66. Wallace ZS, Mattoo H, Carruthers M, et al. Plasmablasts as a biomarker for IgG4-related disease, independent of serum IgG4 concentrations. Ann Rheum Dis. 2015;74(1):190195. Parker AR, Carr-Smith HD. Calibration differences and the prozone phenomenon in IgG4-related disease: comment on the article by Khosroshahi et al. Arthritis Rheumatol. 2014;66(7):1964-1965. Khosroshahi A, Cheryk LA, Carruthers MN, Edwards JA, Bloch DB, Stone JH. Brief Report: spuriously low serum IgG4 concentrations caused by the prozone phenomenon in patients with IgG4-related disease. Arthritis Rheumatol. 2014;66(1):213-217. van der Gugten G, DeMarco ML, Chen LYC, et al. Resolution of spurious immunonephelometric IgG subclass measurement discrepancies by LC-MS/MS. Clin Chem. 2018;64(4):735-742. Chan ASY, Mudhar H, Shen SY, et al. Serum IgG2 and tissue IgG2 plasma cell elevation in orbital IgG4-related disease (IgG4-RD): potential use in IgG4-RD assessment. Br J Ophthalmol. 2017;101(11):1576-1582. Dunkley L, Mudhar HS. IgG4-related disease presenting with raised serum IgG2-real timeline of IgG4-RD? Rheumatology. 2017;57(1):197-199. Mattman A, Chen LYC, van der Gugten G, et al. Comment on: IgG4-related disease presenting with raised serum IgG2-real timeline of IgG4-RD? Rheumatology. 2018;57(6):1126-1127. Chari ST, Takahashi N, Levy MJ, et al. A diagnostic strategy to distinguish autoimmune pancreatitis from pancreatic cancer. Clin Gastroenterol Hepatol. 2009;7(10): 1097-1103. Mann S, Seidman MA, Barbour SJ, Levin A, Carruthers M, Chen LY. Recognizing IgG4related tubulointerstitial nephritis. Can J Kidney Health Dis. 2016;3:34. Moriyama M, Ohta M, Furukawa S, et al. The diagnostic utility of labial salivary gland biopsy in IgG4-related disease. Mod Rheumatol. 2016;26(5):725-729.

72. Arora K, Rivera M, Ting DT, Deshpande V. The histologic diagnosis of IgG4-related disease on small biopsies: challenges and pitfalls. Histopathology. 2018 Nov 8. [Epub ahed of print] 73. Kuo TT, Chen TC, Lee LY. Sclerosing angiomatoid nodular transformation of the spleen (SANT): clinicopathological study of 10 cases with or without abdominal disseminated calcifying fibrous tumors, and the presence of a significant number of IgG4+ plasma cells. Pathol Int. 2009;59(12):844-850. 74. Carruthers MN, Topazian MD, Khosroshahi A, et al. Rituximab for IgG4-related disease: a prospective, open-label trial. Ann Rheum Dis. 2015;74(6):1171-1177. 75. Wallace ZS, Khosroshahi A, Jakobiec FA, et al. IgG4-related systemic disease as a cause of "idiopathic" orbital inflammation, including orbital myositis, and trigeminal nerve involvement. Surv Ophthalmol. 2012;57(1): 26-33. 76. Ebbo M, Grados A, Guedj E, et al. Usefulness of 2-[18F]-fluoro-2-deoxy-D-glucose-positron emission tomography/computed tomography for staging and evaluation of treatment response in IgG4-related disease: a retrospective multicenter study. Arthritis Care Res (Hoboken). 2014;66(1):86-96. 77. Masaki Y, Matsui S, Saeki T, et al. A multicenter phase II prospective clinical trial of glucocorticoid for patients with untreated IgG4-related disease. Mod Rheumatol. 2017;27(5):849-854. 78. Wang L, Zhang P, Wang M, et al. Failure of remission induction by glucocorticoids alone or in combination with immunosuppressive agents in IgG4-related disease: a prospective study of 215 patients. Arthritis Res Ther. 2018;20(1):65. 79. Masamune A, Nishimori I, Kikuta K, et al. Randomised controlled trial of long-term maintenance corticosteroid therapy in patients with autoimmune pancreatitis. Gut. 2017;66(3):487-494. 80. Hart PA, Topazian MD, Witzig TE, et al. Treatment of relapsing autoimmune pancreatitis with immunomodulators and rituximab: the Mayo Clinic experience. Gut. 2013;62(11):1607-1615. 81. Yunyun F, Yu C, Panpan Z, et al. Efficacy of cyclophosphamide treatment for immunoglobulin G4-related disease with addition of glucocorticoids. Sci Rep. 2017;7(1):6195. 82. Yunyun F, Yu P, Panpan Z, et al. Efficacy and safety of low dose mycophenolate mofetil treatment for immunoglobulin G4-related disease. Rheumatology (Oxford, England). 2018 Aug 14. [Epub ahead of print]. 83. Khosroshahi A, Wallace ZS, Crowe JL, et al. International consensus guidance statement on the management and treatment of IgG4related disease. Arthritis Rheumatol. 2015;67(7):1688-1699. 84. Ebbo M, Grados A, Samson M, et al. Longterm efficacy and safety of rituximab in IgG4-related disease: data from a French nationwide study of thirty-three patients. PLoS One. 2017;12(9):e0183844. 85. Stone JH, Wallace ZS, Perugino CA, et al. A Trial of XmAb (R) 5871, a reversible inhibitor of CD19+ cells, in IgG4-related disease. Arthritis Rheumatol. 2016;68(Suppl 10).

455


ARTICLE Ferrata Storti Foundation

Haematologica 2018 Volume 104(3):456-467

Hematopoiesis

Chronic sympathetic driven hypertension promotes atherosclerosis by enhancing hematopoiesis

Annas Al-Sharea,1* Man K. S. Lee,1 Alexandra Whillas,1 Danielle L. Michell,1,2 Waled A. Shihata,1,3 Alyce J. Nicholls,4 Olivia D. Cooney,1 Michael J. Kraakman,1,5 Camilla Bertuzzo Veiga,1 Ann-Maree Jefferis,3 Kristy Jackson,6 Prabhakara R. Nagareddy,7 Gavin Lambert,8,9 Connie H. Y. Wong,4 Karen L. Andrews,3 Geoff A. Head,6 Jaye Chin-Dusting3 and Andrew J. Murphy1,10*

Haematopoiesis and Leukocyte Biology Laboratory, Division of Immunometabolism, Baker Heart and Diabetes Institute, Melbourne, VIC, Australia; 2Department of Medicine, Vanderbilt University School of Medicine, Nashville, TN, USA; 3Department of Pharmacology, Monash University, Clayton, VIC, Australia; 4Monash University, Melbourne, VIC, Australia; 5Naomi Berrie Diabetes Center and Department of Medicine, Columbia University, New York, NY, USA; 6Neuropharmacology Laboratory, Division of Hypertension and Cardiac Disease, Baker Heart and Diabetes Institute, Melbourne, VIC, Australia; 7Department of Nutrition Sciences, University of Alabama at Birmingham, AL, USA; 8Human Neurotransmitters Laboratory, Division of Hypertension and Cardiac Disease, Baker Heart and Diabetes Institute, Melbourne, VIC, Australia; 9Iverson Health Innovation Research Institute, Swinburne University of Technology, Hawthorn, VIC, Australia; 10Department of Immunology, Monash University, Melbourne, VIC, Australia 1

*Corresponding Authors

ABSTRACT

H

Correspondence: ANDREW J. MURPHY andrew.murphy@baker.edu.au ANNAS AL-SHAREA annas.al-sharea@baker.edu.au Received: March 7, 2018. Accepted: October 22, 2018. Pre-published: October 25, 2018. doi:10.3324/haematol.2018.192898 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/104/3/456 Š2019 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

456

ypertension is a major, independent risk factor for atherosclerotic cardiovascular disease. However, this pathology can arise through multiple pathways, which could influence vascular disease through distinct mechanisms. An overactive sympathetic nervous system is a dominant pathway that can precipitate in elevated blood pressure. We aimed to determine how the sympathetic nervous system directly promotes atherosclerosis in the setting of hypertension. We used a mouse model of sympathetic nervous system-driven hypertension on the atherosclerotic-prone apolipoprotein E-deficient background. When mice were placed on a western type diet for 16 weeks, we showed the evolution of unstable atherosclerotic lesions. Fortuitously, the changes in lesion composition were independent of endothelial dysfunction, allowing for the discovery of alternative mechanisms. With the use of flow cytometry and bone marrow imaging, we found that sympathetic activation caused deterioration of the hematopoietic stem and progenitor cell niche in the bone marrow, promoting the liberation of these cells into the circulation and extramedullary hematopoiesis in the spleen. Specifically, sympathetic activation reduced the abundance of key hematopoietic stem and progenitor cell niche cells, sinusoidal endothelial cells and osteoblasts. Additionally, sympathetic bone marrow activity prompted neutrophils to secrete proteases to cleave the hematopoietic stem and progenitor cell surface receptor CXCR4. All these effects could be reversed using the b-blocker propranolol during the feeding period. These findings suggest that elevated blood pressure driven by the sympathetic nervous system can influence mechanisms that modulate the hematopoietic system to promote atherosclerosis and contribute to cardiovascular events. Introduction Hypertension is a major, independent risk factor for atherosclerotic cardiovascular disease (CVD).1 As the pathophysiology of hypertension is both complex and multifactorial, the direct mechanism(s) that ultimately contribute to CVD remain haematologica | 2019; 104(3)


SNS driven hypertension enhances hematopoiesis

unclear.2 The most frequently targeted pathway in reducing blood pressure is the renin-angiotensin system (RAS). The contribution of the RAS to hypertension and atherosclerosis is not exclusive, as angiotensin II (AngII) can also accelerate atherogenesis independent of hypertension.3,4 Another major determinant of hypertension is an overactive sympathetic nervous system (SNS).5,6 There are indications that autonomic input into the bone marrow (BM) may be altered in the setting of hypertension.7-10 However, the mechanisms promoting atherogenesis with SNS activation associated hypertension are not completely elucidated. While there is an overlap in some atherosclerosis promoting mechanisms between the RAS and SNS, a distinct subset of events is also likely to be evoked by the SNS, which requires further investigation. Atherosclerosis is a disease driven by the infiltration of immune cells, in particular monocytes, into the plaque.11-13 It is also well established that the abundance of circulating monocytes predicts cardiovascular (CV) events and is directly linked to atherogenesis.14,15 Interestingly, the SNS plays a direct role in regulating the hematopoietic system from which immune cells, including monocytes, arise.16-19 In the context of CVD, the mobilization of hematopoietic stem and progenitor cells (HSPCs) from the BM to extramedullary tissues such as the spleen results in the generation of atherogenic monocytes that abundantly enter into the atherosclerotic plaque.20 Mobilization of HSPCs can be mediated by sympathetic signaling within the BM, particularly in response to granulocyte-colony stimulating factor (G-CSF). The SNS synergizes with GCSF to promote the breakdown of the HSPC BM microenvironment, which decreases the abundance of key HSPC retention factors and results in the liberation of HSPCs into the circulation.16 This pathway has also been shown to be activated following a myocardial infarction (MI).21 Sympathetic activation, along with raised G-CSF levels that are observed in apolipoprotein E knockout (Apoe-/-) mice, caused HSPC mobilization from the BM and homing to the spleen where monocytes were subsequently produced that infiltrated atherosclerotic lesions. Interestingly, this promoted an unstable plaque phenotype, prone to rupturing and thus provides a plausible explanation for primary heart attack survivors being highly prone to a secondary, often fatal, CV event.21,22 Importantly, the involvement of the SNS in driving aberrant hematopoiesis is not restricted to complications following a MI, as similarities in other models of stress and ischemic stroke are evident, suggesting this to be a more general mechanism. The augmented hematopoietic response in these pathologies caused by overactivation of the SNS were inhibited by administration of b-blockers or genetic deletion of b-adrenergic receptors.21,23-25 There appears to be an important role of the SNS in regulating hematopoiesis in acute stressors (i.e., MI, stroke, variable stress). However, it remains unknown if chronic sympathetic activation invokes this same atherogenic process. Thus, it is plausible that chronic sympathetic activation present in some cases of hypertension could play an important role in regulating atherogenesis by altering hematopoiesis. To address this question, we employed the Schlager hypertensive mice which were crossed onto an Apoe-/- background to produce hypertensive atherosclerosis-prone mice. The Schlager mouse was chosen as it represents a model of hypertension that is almost entirely driven by the SNS, with minimal contribution by the haematologica | 2019; 104(3)

RAS.26 We sought to characterize the contribution of SNS activation associated hypertension to the development of atherosclerosis, with the aim of understanding whether this form of hypertension was also associated with alterations to the hematopoietic system. Moreover, we aimed to investigate whether targeting the SNS could inhibit atherogenesis and, in turn, reveal an additional mechanism of hypertension associated atherosclerosis.

Methods Detailed methods are available in Online supplementary Methods.

Animal Models Apoe-/- mice were purchased from Jackson Laboratories and bred at the AMREP Animal centre. To generate hypertensive Apoe-/mice, BPH/2J mice were crossed with Apoe-/- mice to produce BPH/2J x Apoe-/- (BPH/Apoe-/-) mice. At 6 weeks of age, male Apoe-/and BPH/ Apoe-/- mice were placed on a western type diet (WTD - SF00-219, Specialty Feeds, Australia; 21% fat, 0.15% cholesterol) for 16 weeks. In the first cohort of mice, age-matched mice Apoe-/and BPH/Apoe-/- were placed on a WTD for 16 weeks for end-point analysis. In a second cohort of mice, obtained from a new set of breeders, three groups of aged-matched mice were employed: 1) Apoe-/-, 2) BPH/Apoe-/- and 3) BPH/ Apoe-/- + propranolol (0.5g/L; administered via drinking water for the duration of the WTD feeding). For the propranolol group, mice consumed on average 2.5ml of water amounting to an average daily dose of 3540mg/kg/daily of propranolol. To determine the effect of specific 2-adrenoreceptor blockade on HSPC mobilization and blood pressure we used BPH mice on an Apoe+/+ background. The mice were injected daily with ICI118551 (5mg/kg; Abcam, AUS) for 2 weeks. All animal experiments were approved by the AMREP Animal Ethics Committee and conducted in accordance with the Australian code of practice for the care and use of animals for scientific purposes as stipulated by the National Health and Medical Research Council of Australia. All mice were housed in a normal light and dark cycle and had ad libitum access to food and water. Mice were randomly assigned to treatment and end-point analysis was blinded.

Statistics Data are presented as mean Âą SEM (unless stated otherwise) and were analysed using the two-tailed Student t-test or One-way ANOVA where appropriate. Analysis of baseline and final blood pressure between strains was performed using a two-way ANOVA with the factors strain (Pstrain) and time (Ptime) followed by a Sidak post-hoc test to account for multiple comparisons. A P<0.05 was considered significant. All tests were performed using the Prism software (GraphPad Software, Inc., La Jolla, CA, USA).

Results Hypertension associated with chronic sympathetic activation promotes an unstable atherosclerotic phenotype To determine the contribution of chronic sympathetic activation in hypertension to atherosclerosis we crossed Schlager hypertensive mice with Apoe-/- mice (BPH/Apoe-/-) and compared these to normotensive Apoe-/- mice. Mice were fed a high fat, high cholesterol western type diet (WTD) for 16 weeks. We preferenced this model over con457


A. Al-Sharea et al.

tinual infusions of noradrenaline to allow for circadian fluctuations in blood pressure and heart rate, and to prevent the ongoing immune complications of surgery associated with the use of mini-pumps. Firstly, examining traditional cardiovascular risk factors revealed no change in body weight or cholesterol between the two groups and while blood pressure increased over the feeding period in

A

B

both strains, the BPH/Apoe-/- mice maintained significantly higher blood pressure as measured by tail cuff and radio telemetry (Figure 1, A-C and Online Supplementary Figure S1). The mice were also equally active (Online Supplementary Figure S1). To explore the effect of chronic sympathetic activation associated with hypertension on atherosclerosis, we assessed the atherosclerotic burden in

C

D

E

F

G

H

Figure 1. Hypertensive Apoe-/- mice have characteristics of unstable atherosclerotic lesions. Apoe-/- and BPH/Apoe-/- mice were fed a WTD for 16 weeks A) Weekly body mass progression. B) Total plasma cholesterol was measured at the end of feeding. C) Systolic blood pressure of mice was measured before and at the end of WTD feeding. D) H&E staining for plaque size in the proximal aorta and E) ORO+ lesions in the aortic arch were quantified. Proximal aortas were also stained for F) lipid content (ORO), G) macrophages (CD68) and H) collagen (picosirius red). Lesions were imaged at X4 (insets) and zoomed in to view single lesions, scale bar = 100 mm. Data are presented as mean Âą SEM where *P<0.05 and **P<0.01 (Studentâ&#x20AC;&#x2122;s t-test) ****P<0.0001 (Two-way ANOVA). A,B,C) n= 5-10, D,E,F) n= 8, G) n=9, H) n=7-8.

458

haematologica | 2019; 104(3)


SNS driven hypertension enhances hematopoiesis

the proximal aorta and aortic arch. We observed increases in plaque size between the groups (Figure 1D, E), suggesting that sympathetically driven hypertension may promote accelerated plaque growth. We further explored the lesion characteristics and noted a significant increase in the abundance of lipid within the lesions from the BPH/Apoe-/- mice (Figure 1F). A significant increase in plaque macrophages were accompanied by a decrease in plaque collagen in the BPH/Apoe-/- mice (Figure 1G, H), suggesting that chronic sympathetic activation was promoting remodeling of lesions in an adverse, unstable man-

A

D

ner. This plaque phenotype in the BPH/Apoe-/- mice resonates with the findings of Dutta et al. in the context of acute SNS stimulation during a MI.21

Hypertensive Apoe–/– mice do not develop endothelial dysfunction Endothelial dysfunction is a generally accepted consequence of hypertension.27 To determine if the enhanced atherogenesis in BPH/Apoe-/- mice was the result of endothelial dysfunction we assessed the vascular responses in aortas from the BPH/Apoe-/- mice in comparison with

C

B

E

Figure 2. Hypertensive Apoe-/do not have endothelial dysand function. Apoe-/BPH/Apoe-/- mice were fed a WTD for 16 weeks after which aortas were harvested for ex vivo assessment of endothelial function. Aortas were assessed for A) diameter, B) KPSS and C) L-NAME contraction. Further myograph analyses were preformed to determine aortic relaxation in response to D) ACh and E) SNP with or without L-NAME administration. Aortic T-cell infiltration was assessed by flow cytometry for F) Abundance of CD4+ T-helper cells. Data are presented as mean ± SEM where **P<0.01 (Student’s ttest). A,B) n=6-10,C) n= 6-8, D) n= 4-6, E,F) n=6.

F

haematologica | 2019; 104(3)

459


A. Al-Sharea et al. B

A

D

E

C

F

G

H

I

Figure 3. BPH/Apoe-/- mice have extramedullary hematopoiesis. Apoe-/- and BPH/ Apoe-/- mice were fed a WTD for 16 weeks after which flow cytometry was used to assess A) circulating monocytes(M)/neutrophils(N). B) Levels of BM HSPCs and C) proportion of HSPCs proliferating. D) Levels of BM GMPs and E) proportion of GMPs proliferating. F) Blood HSPC and MPC populations. G) Spleen HSPCs and GMPs were assessed along with H) proliferating HSPCs and GMPs. I) splenic monocyte(M)/neutrophil(N) populations. Data are presented as mean Âą SEM where *P<0.05 and **P<0.01 (Studentâ&#x20AC;&#x2122;s t-test). A-I) n= 5-11.

460

haematologica | 2019; 104(3)


SNS driven hypertension enhances hematopoiesis

control Apoe-/- mice. Firstly, no change in vessel diameter or constrictor responses to a high potassium solution was evident. Nor were there differences in basal nitric oxide (NO) levels when the constriction to L-NAME (L-NGNitroarginine methyl ester) was examined (Figure 2, A-C). These data suggest that alterations in vascular reactivity are not biased by differences in constrictor responses. Surprisingly, endothelium-dependent NO-mediated relaxation in response to acetylcholine (ACh) was worse in the Apoe-/- mice when compared to BPH/Apoe-/- mice (Figure 2D). These differences between the strains were endothelium independent since there were no differences in the

A

B

C

E

D

G

F

H

constriction and relaxation response to the NO donor sodium nitroprusside (SNP) in the presence or absence of L-NAME (Figure 2E). To further confirm no decline in vascular function in these mice, we examined the abundance of T cells, which have been linked to the pathogenesis of hypertension.28 We observed no differences in aortic T cells between the Apoe-/- and BPH/Apoe-/- mice (Figure 2F). Moreover, there was no difference in the activation state of these CD4+ T cells, as assessed by CD62L expression (MFI; data not shown). These data suggest that the enhanced atherogenesis in the BPH/Apoe-/- mice occurs independently of changes to the endothelium.

I

haematologica | 2019; 104(3)

J

Figure 4. Sympathetic nervous system signalling induces a breakdown of the HSPC niche in the BM. Apoe-/- and BPH/Apoe-/- mice were fed a WTD for 16 weeks with BPH/Apoe-/- mice either treated with or without Propranolol (0.5g/L) in drinking water. A) Plasma NA was quantified by HPLC. B) BM sections were immunostained for sympathetic activity as indicated by tyrosine hydroxylase, imaged at X20; scale bar = 25 mm. C) Systolic blood pressure of mice following 16 weeks treatment. D) Osteoblastic lineage cells were quantified by flow cytometry. E) mRNA levels of Runx2 in the BM. F) H&E stained representation of BM vascular morphology, imaged at X10 scale bar = 100 mm. G) BMECs were measured by flow cytometry. H) MMP9 content in the BM extracellular fluid was determined via Zymography. I) CXCR4 expression levels on HSPCs was assessed by flow cytometry from and J) neutrophil supernatant cultured HSPCs. Data are presented as mean Âą SEM where *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001 (Studentâ&#x20AC;&#x2122;s ttest or One-Way ANOVA). A) n= 8, B) n=5, C) n=7, D) n= 7-9, E) n= 5-9, G) n= 7-9, H) 5-9, I) n= 12-15, J) n= 7.

461


A. Al-Sharea et al.

BPH/Apoeâ&#x20AC;&#x201C;/â&#x20AC;&#x201C; mice have enhanced myelopoiesis An overactive SNS has recently been shown to promote the mobilization of BM HSPCs to the spleen, resulting in the generation of splenic monocytes that can infiltrate into atherosclerotic lesions. Therefore, we next assessed the hematopoietic system in these mice.21 We discovered prominent monocytosis and neutrophilia in the BPH/Apoe-/- mice in the blood (Figure 3A). Next, to determine if the increased monocyte and neutrophil numbers were due to activated myelopoiesis, we examined the abundance and proliferation of HSPCs and myeloid progenitor cells in the BM. While the levels and proliferation of HSPCs within the BM were similar (Figure 3B, C), we did observe more granulocyte-macrophage progenitors (GMPs) in the BPH/Apoe-/- mice, which were proliferating at a higher rate (Figure 3D,E). Consistent with SNS activation in promoting the mobilisation of HSPCs from the BM, we detected elevated levels of circulating HSPCs and myeloid progenitors (MPCs) in the BPH/Apoe-/- mice (Figure 3F). Given the higher circulating HSPCs, we were expecting to see more HSPCs in the spleen. However, no such change in the abundance of HSPCs was detected (Figure 3G). Of note, a higher proportion were in the G2M phase of the cell cycle (Figure 3H), suggesting that the chronic activation of the SNS was influencing the HSPCs to proliferate more in the spleens of BPH/Apoe-/- mice. Indeed, monocytes and neutrophils were elevated in the spleens of the BPH/Apoe-/- mice, confirming extramedullary myelopoiesis was occurring in this chronic sympathetic driven model (Figure 3I).

Sympathetic activation contributes to the breakdown of the HSPC bone marrow microenvironment The Schlager mice are an established model of sympathetic activation-mediated hypertension.26 However, we wanted to confirm that there was evidence of increased sympathetic activation in the BM, which could account for the enhanced mobilization of HSPCs observed in Figure 3D. Firstly, to confirm an overall increase in sympathetic tone, we quantified plasma noradrenaline levels, which we found to be higher in the BPH/Apoe-/- mice (Figure 4A). More central to our proposed mechanism for enhanced HSPC mobilization in the BPH/Apoe-/- mice, we found enhanced expression of tyrosine hydroxylase (TH), the rate-limiting enzyme found in nerve terminals responsible for noradrenaline (NA) production, around the blood vessels in the BM of the BPH/Apoe-/- mice (Figure 4B). Together, these data reveal a more global increase in sympathetic tone in SNS-driven hypertension. Next, we sought to determine if overactive sympathetic signaling in the BM led to changes within the BM microenvironment and whether these changes could be reversed with the use of a b-blocker. Given the importance of sympathetic overdrive in mediating hypertension, as expected, propranolol normalized blood pressure in the BPH/Apoe-/- mice (Figure 4C). Having demonstrated that propranolol could reverse the systemic responsiveness of b-receptors to sympathetic activation in BPH/Apoe-/- mice, we examined key HSPC niche cells in the BM to determine if sympathetic overdrive influenced myelopoiesis via effects on the BM niche. Interestingly, we found a significant reduction in the abundance of CD51+ osteoblasts in the BM on the BPH/Apoe-/mice, which were restored when these mice were treated with propranolol (Figure 4D). Consistent with this find462

ing, analysis of BM mRNA for Runx2, the transcription factor that drives osteoblast production, showed a reduction in Runx2 expression in the BPH/Apoe-/- mice relative to Apoe-/- mice. Similar to our flow cytometry data, treatment with propranolol prevented the suppression of Runx2 expression (Figure 4E). When we assessed the gross morphological changes in the BM, it appeared that the vascular structures were altered with the BPH/Apoe-/- mice showing smaller sinusoidal structures relative to the Apoe/mice, with propranolol reverting the sinusoids back to that seen in the Apoe-/- mice (Figure 4F). Furthermore, in examining the endothelial cell population, we noted a trend towards a decrease in the abundance of these cells, which, again, could be restored with the administration of propranolol (Figure 4G). As these niche cells are an important source of the HSPC retention factor CXCL12, we measured its mRNA expression and found that propranolol greatly increased Cxcl12 expression, thereby potentially aiding in promoting HSPC retention and reduced quiescence in the BM (Online Supplementary Figure S2, A). The changes in these two key niche cells may provide a mechanism for increased HSPC release from the BM in BPH/Apoe-/- mice. Other cells within the BM express b-adrenoreceptors, which we profiled using gene array data from Novershtern et al. and analysed using online software (BloodSpot) to generate a hierarchical differentiation tree.29,30 Firstly, HSPCs and myeloid progenitors did not display any enrichment for the adrenoceptors. However, neutrophils were identified as one of the cells enriched in transcripts for the b2-adrenoreceptor, but not b1- or b3-adrenoreceptors (Online Supplementary Figure S2, B-D). We pharmacologically confirmed the requirement for b2-adrenoreceptor stimulation in HSPC mobilization using the BPH mice on an Apoe+/+ background by administering the b2 specific antagonist ICI-118551 (Online Supplementary Figure S2, E). Furthermore, neutrophils have previously been shown to be responsive to NA in vitro.31,32 Mechanistically, activated neutrophils can release MMP9 which can cleave CXCR4 on HSPCs, providing another avenue to HSPC liberation from the BM.22,33 We measured levels of MMP9 in the BM extracellular fluid (BMEF) via zymography and found that both active and latent MMP9 levels increased in the BPH/Apoe-/- mice, a phenotype reversed with propranolol treatment (Figure 4H and Online Supplementary Figure S2, F). In support of this we found reduced surface CXCR4 expression on the HSPCs from the BPH/Apoe-/- mice, which was restored in mice treated with propranolol (Figure 4I). These data were further supported by BM mRNA analysis indicating that propranolol treatment increases Cxcr4 expression (Online Supplementary Figure S2, G). To explore this mechanism further, we cultured HSPCs in the supernatants of neutrophils treated with NA and examined CXCR4 cell surface abundance. We found significantly less CXCR4 on HSPCs cultured in supernatants from NA activated neutrophils (isolated from wild-type mice), compared to vehicle treated neutrophils, which was prevented when MMP9 was inhibited (Figure 4J). When we included the b2-adrenoreceptor specific inhibitor ICI-118551 into the BM neutrophil stimulation media with NA, the harvested supernatant caused less efficient cleavage of CXCR4 (Online Supplementary Figure S2, H) thereby confirming the role for neutrophil b2-adrenoreceptors. These data support the hypothesis that sympathetic activation is present in the BM of the BPH/Apoe-/- mice and responsible for the haematologica | 2019; 104(3)


SNS driven hypertension enhances hematopoiesis

mobilization of HSPCs by acting on key niche cells along with stimulating neutrophils to secrete proteases that cleave the retention receptor CXCR4 on HSPCs.

Suppressing chronic sympathetic signaling dampens myelopoiesis in hypertensive Apoe–/– mice Having observed a restoration in the HSPC BM microenvironment when BPH/Apoe-/- mice were treated with propranolol, we explored if this was also reflected by normalization of myelopoiesis in these mice. Following

A

B

D

F

haematologica | 2019; 104(3)

C

E

G

H

I

treatment with propranolol, we observed a reduction in circulating monocytes and neutrophils (Figure 5A). We determined if this reduction was echoed by changes in the BM stem and progenitor populations. Following administration of propranolol, the abundance of BM HSPCs was not affected; however, these cells were proliferating at lower rates and giving rise to fewer GMPs (Figure 5, B-D). Consistent with an improvement in the HSPC microenvironment, there were fewer mobilized HSPCs and MPCs in the blood of the BPH/Apoe-/- mice treated with propra-

Figure 5. Propranolol prevents extramedullary hematopoiesis in BPH/Apoe-/- mice. BPH/Apoe/mice were fed a WTD for 16 weeks and treated with vehicle or Propranolol (0.5 g/L) in drinking water. Flow cytometry was used to assess A) circulating monocytes(M)/neutrophils(N), B) BM HSPCs and C) proportion of which are proliferating along with D) of BM GMPs. E) Blood HSPC and MPC populations. F) Spleen HSPC and G) GMP populations along with H) proliferating splenic HSPCs and GMPs were assessed by flow cytometry. I) Splenic monocyte(M)/neutrophil(N) populations. Data are presented as mean ± SEM where *P<0.05, **P<0.01 and ***P<0.001 (Student’s t-test). A-I) n= 6-10.

463


A. Al-Sharea et al.

nolol (Figure 5E). This was paralleled by a decrease in extramedullary hematopoiesis in the spleen as evidenced by fewer proliferating HSPCs, GMPs and less monocytes and neutrophils (Figure 5, F-I). Taken together, these data suggest that lowering responsiveness to chronic sympathetic signaling in the BPH/Apoe-/- mice results in an overall dampening of myelopoiesis.

Blocking sympathetic signalling decreases atherosclerosis in BPH/Apoe–/– mice To examine if the reduction in sympathetic tone and dampening of myelopoiesis was associated with reduced

atherosclerotic plaque progression, we assessed the size and complexity of lesions in the proximal aorta. Firstly, we noted a reduction in lesion size in the proximal aorta and aortic arch of BPH/Apoe-/- mice treated with propranolol (Figure 6A, B). Exploring the lesion characteristics, we noted that propranolol treated mice had reduced plaque lipid accumulation along with a reduction in plaque macrophages (Figure 6C, D). We also observed a trend for increased collagen (Figure 6E). These changes were seen in the absence of any changes in plasma cholesterol levels (Figure 6F). Given that the hypertension in the BPH/Apoe/- mice did not promote endothelial dysfunction, it sug-

A

B

C

D

E

F

Figure 6. Propranolol inhibits plaque progression in BPH/Apoe-/- mice. BPH/Apoe-/- mice were fed a WTD for 16 weeks and treated with vehicle or Propranolol (0.5g/L) in drinking water. At the end point, atherosclerosis in the proximal aorta was assessed for A) H&E staining for plaque size in the proximal aorta and B) lipid content (ORO+) lesions in the aortic arch were quantified. Proximal aortas were also stained for C) lipid content (ORO), D) macrophages (CD68) and E) collagen (picosirius red). Lesions were imaged at X4 (insets) and zoomed in to view single lesion, scale bar = 100 mm. F) Total plasma cholesterol levels. Data are presented as mean ± SEM where *P<0.05 and **P<0.01 (Student’s t-test). A,B) n= 9, C-E) n=7, F) n=9.

464

haematologica | 2019; 104(3)


SNS driven hypertension enhances hematopoiesis

gests that the improvements in plaque size and complexity are due to dampened myelopoiesis and, subsequently, reduced monocyte infiltration.

Discussion Chronic hypertension is arguably one of the most common risk factors associated with atherosclerotic CVD.1 However, delving into the responsible mechanism(s), it remains unclear if an increase in blood pressure alone, or in conjunction with a change in concurrent signaling events such activation of the RAS or sympathetic activation, directly contributes to atherosclerotic CVD. Using a genetic model of hypertension driven by sympathetic activation, we show that this form of hypertension (compared to atherosclerotic prone mice without hypertension) alters the characteristics of the atherosclerotic lesion to a more unstable phenotype, hallmarked by increased macrophage accumulation. We also found that chronic sympathetic activation caused changes in hematopoiesis. In particular, increased sympathetic activity was found in the BM, altering the HSPC microenvironment and causing the liberation of certain stem cells to the spleen where monocytes were generated. This was accompanied by an increase in blood monocytes, likely explaining the increased macrophage burden observed in the atherosclerotic lesions. These atherogenic pathways could all be inhibited pharmacologically by blocking sympathetic signalling through b-adrenoreceptors using propranolol. These findings suggest that chronic sympathetic activation, present in many forms of hypertension, likely contributes to the increased CVD risk by modulating hematopoiesis, independent of endothelial dysfunction. With respect to understanding the contribution of hypertension to vascular disease, the majority of research has focused on the effects on the endothelium. Perhaps the most common belief is that hypertension causes endothelial dysfunction and activation, which in turn recruits immune cells and forms the main mechanism propagating the atheroma. Interestingly, we found no evidence of endothelial dysfunction in our hypertensive mice, at least in this model of a dominant sympathetic driven form of hypertension, the endothelial dysfunction was not contributing significantly to atherogenisis.34 Supporting our theory that underlying sympathetic nervous signaling, that may be independent of pressure itself, can drive atherogenesis, moderate increases in AngII are sufficient to promote accelerated atherogenesis, without elevations in blood pressure.3,4 Additionally, AngII has also been shown to invoke a T-helper cell (TH1) immune response to promote atherogenesis independent of its hemodynamic effects. Thus, signaling events that can cause hypertension are likely important in driving CVD through their immune modulatory responses.7-10,35,36 Further, with the discovery of accelerated vascular disease driven by acute events triggering sympathetic activation leading to enhanced monocyte production, it is plausible that this pathway is triggered in chronic SNS-driven hypertension and would contribute to accelerated atherosclerosis.21,23,24 We hypothesized that the overactive SNS seen in subgroups of patients with hypertension would contribute to atherogenesis by stimulating hematopoiesis. Importantly, elevated WBCs are associated with the incidence of hypertension and predicts CV outcomes in this haematologica | 2019; 104(3)

patient group.37-39 However, the cause of increased WBCs in hypertensive patients has not been resolved. Consistent with recent studies which have observed monocytosis following acute scenarios of sympathetic activation, we too observed monocytosis in the hypertensive BPH/Apoe-/- mice.21,24 The initial predominant change driven by the overactive sympathetic signaling in our study, relevant to increased myelopoiesis, appears to occur within the BM. We noted a decreased abundance of two key niche cells, endothelial cells and osteoblasts, which harbour anchoring points in the marrow for HSPCs, preventing their release into circulation.40-43 The contribution of the SNS in regulating this process was first described by a seminal study from the Frenette laboratory, detailing the requirement of a functional SNS in the BM, which is required for G-CSF mediated HSPC mobilization.16 Almost a decade later, the Nahrendorf group discovered the importance of this pathway in respect to CVD, revealing that sympathetic activation following an acute myocardial infarction promotes HSPC liberation to the spleen where the production of an additional atherogenic pool of monocytes occurs.21 The absence of an expanded HSPC population in the spleen is likely due to the chronic nature of our study and suggests that these cells likely rapidly matured into myeloid committed progenitors. The recent studies then suggest that the monocytes generated, migrated into the atherosclerotic lesion, and enhanced macrophage burden, potentiating the risk of a secondary CV event. Our data reveal that this process is occurring chronically and identifies an important mechanism that likely contributes to atherogenesis and the increased risk of a CV event in hypertension. We also identified another pathway by which sympathetic signaling can induce the liberation of BM HSPCS by causing a decrease in the HSPC-expressed retention receptor CXCR4. Given that HSPCs do not appear to express b adrenoreceptors, it suggested a cell extrinsic mechanism resulting in less HSPC cell surface CXCR4. Interestingly, neutrophils express b2 adrenoceptors and can be activated after sensing NE (Online Supplementary Figure S2, B-C). Modelling this in vitro revealed that NE-activated neutrophils produce MMP9, which cleaves CXCR4 on HSPCs. Thus, BM sympathetic activation likely liberates HSPCs via multiple mechanisms, some of which are independent of the previously described SNS/G-CSF axis. As mentioned above, there are several studies that have identified a role for b adrenoreceptors in influencing HSPC release via modulating the BM niche. There is strong evidence for the role of b-3 adrenergic receptor in regulating nestin+ stromal cell production of key factors such as CXCL12, angiopoietin and stem cell factor, thereby influencing HSPC retention and proliferation. b-3 antagonism following ischemic events has shown reduced HSPC mobilisation and proliferation leading to dampened extramedullary hematopoiesis.19,21,23,24 However, there is also strong evidence pointing to a role for the b-2 adrenergic receptor in regulating BM niche components and HSPC mobilisation. Although this has been suggested to occur through other niche components such as osteoblasts and other stromal cells and not specifically nestin+ cells.16,18 These findings regarding the role of b adrenoceptors in modulating HSPC mobilisation suggest that in our study there is a likely contribution of both b-2 and b-3 receptors to changes in the BM. However, considering the role of b-2 in the setting of hypertension and elevated 465


A. Al-Sharea et al. blood pressure we focused on the specific role of b-2. Interestingly, a recent study by Mendez-Ferrerâ&#x20AC;&#x2122;s group has also highlighted the chronic role of nestin+ cells present in the BM and other tissues in regulating myeloid cell movement in the setting of atherosclerosis.44 Given that nestin+ cells in the BM express b-3 receptors and the data we have presented above regarding chronic sympathetic driven hypertension and its contribution to atherosclerosis, it is likely that this pathway may also play a role and warrants further investigation. The obvious limitation of this study is that our findings were generated in mice. However, this also allowed us to isolate a prominent form of hypertension to reveal a novel atherogenic mechanism, which appears to be independent of endothelial dysfunction, and thus our findings also permit the current dogma to be challenged. While we revealed the effectiveness of propranolol in this model, there is a need to further investigate the effects of directly reducing enhanced hematopoiesis without targeting systemic blood

References 1. (2015) WHO. Non Communicable diseases in Health in 2015: From MDG to SDG. Contract No: ISBN 978954156511. 2015; 2. Bondjers G, Glukhova M, Hansson GK, Postnov YV, Reidy MA, Schwartz SM. Hypertension and atherosclerosis. Cause and effect, or two effects with one unknown cause? Circulation. 1991;84(6 Suppl):VI2-16. 3. Daugherty A, Manning MW, Cassis LA. Angiotensin II promotes atherosclerotic lesions and aneurysms in apolipoprotein Edeficient mice. J Clin Invest. 2000;105(11):1605-1612. 4. Mazzolai L, Duchosal MA, Korber M, et al. Endogenous angiotensin II induces atherosclerotic plaque vulnerability and elicits a Th1 response in ApoE-/mice. Hypertension. 2004;44(3):277-282. 5. Noll G, Wenzel RR, Binggeli C, Corti C, Luscher TF. Role of sympathetic nervous system in hypertension and effects of cardiovascular drugs. Eur Heart J. 1998;19 Suppl F:F32-38. 6. Esler M, Jennings G, Korner P, et al. Assessment of human sympathetic nervous system activity from measurements of norepinephrine turnover. Hypertension. 1988; 11(1):3-20. 7. Zubcevic J, Jun JY, Kim S, et al. Altered inflammatory response is associated with an impaired autonomic input to the bone marrow in the spontaneously hypertensive rat. Hypertension. 2014;63(3):542-550. 8. Santisteban MM, Zubcevic J, Baekey DM, Raizada MK. Dysfunctional brain-bone marrow communication: a paradigm shift in the pathophysiology of hypertension. Curr Hypertens Rep. 2013;15(4):377-389. 9. Santisteban MM, Kim S, Pepine CJ, Raizada MK. Brain-gut-bone marrow axis: implications for hypertension and related therapeutics. Circ Res. 2016;118(8):1327-1336. 10. Santisteban MM, Ahmari N, Carvajal JM, et al. Involvement of bone marrow cells and neuroinflammation in hypertension. Circ Res. 2015;117(2):178-191.

466

pressure. It is likely that with the development of antiinflammatory drugs targeted at the hematopoietic system, it would be possible to dampen the effects on the hematopoietic system without affecting blood pressure which would hypothetically provide the same conclusions as in the present study. Finally, we only studied one form of hypertension, driven by sympathetic signaling. It would be of specific importance to extend a modified version of this hypothesis to hypertension driven by the RAS. Funding AJM is Career Development Fellow of the NHMRC (APP1085752) and a Future Leader Fellowship from the National Heart Foundation (100440) and a recipient of a CSL Centenary Award. This study was also supported by NHMRC project grants (APP1106154 and APP1142938) to AJM. and J.C-D. M.J.K is a Russell Berrie Foundation Scholar in Diabetes Research from the Naomi Berrie Diabetes Centre. PRN was supported by grants from the NIH (R01HL1379 & R00HL1225).

11. Ross R. Atherosclerosis is an inflammatory disease. Am Heart J. 1999;138(5 Pt 2):S419420. 12. Woollard KJ, Geissmann F. Monocytes in atherosclerosis: subsets and functions. Nat Rev Cardiol. 2010;7(2):77-86. 13. Murphy AJ, Tall AR. Disordered haematopoiesis and athero-thrombosis. Eur Heart J. 2016;37(14):1113-1121. 14. Qiao JH, Tripathi J, Mishra NK, et al. Role of macrophage colony-stimulating factor in atherosclerosis: studies of osteopetrotic mice. Am J Pathol. 1997;150(5):1687-1699. 15. van der Valk FM, Kuijk C, Verweij SL, et al. Increased haematopoietic activity in patients with atherosclerosis. Eur Heart J. 2017;38(6):425-432. 16. Katayama Y, Battista M, Kao WM, et al. Signals from the sympathetic nervous system regulate hematopoietic stem cell egress from bone marrow. Cell. 2006;124(2):407421. 17. Stiekema LCA, Schnitzler JG, Nahrendorf M, Stroes ESG. The maturation of a 'neural-hematopoietic' inflammatory axis in cardiovascular disease. Curr Opin Lipidol. 2017;28(6):507-512. 18. Mendez-Ferrer S, Battista M, Frenette PS. Cooperation of beta(2)- and beta(3)-adrenergic receptors in hematopoietic progenitor cell mobilization. Ann N Y Acad Sci. 2010; 1192:139-144. 19. Mendez-Ferrer S, Lucas D, Battista M, Frenette PS. Haematopoietic stem cell release is regulated by circadian oscillations. Nature. 2008;452(7186):442-447. 20. Robbins CS, Chudnovskiy A, Rauch PJ, et al. Extramedullary hematopoiesis generates Ly-6C(high) monocytes that infiltrate atherosclerotic lesions. Circulation. 2012;125(2):364-374. 21. Dutta P, Courties G, Wei Y, et al. Myocardial infarction accelerates atherosclerosis. Nature. 2012;487(7407):325-329. 22. Westerterp M, Gourion-Arsiquaud S, Murphy AJ, et al. Regulation of hematopoietic stem and progenitor cell mobilization by cholesterol efflux pathways. Cell Stem Cell. 2012;11(2):195-206. 23. Heidt T, Sager HB, Courties G, et al.

24.

25.

26.

27. 28.

29.

30.

31.

32.

33.

Chronic variable stress activates hematopoietic stem cells. Nat Med. 2014; 20(7):754-758. Courties G, Herisson F, Sager HB, et al. Ischemic stroke activates hematopoietic bone marrow stem cells. Circ Res. 2015; 116(3):407-417. Esler M, Jennings G, Lambert G. Measurement of overall and cardiac norepinephrine release into plasma during cognitive challenge. Psychoneuroendocrinology. 1989;14(6):477-481. Davern PJ, Nguyen-Huu TP, La Greca L, Abdelkader A, Head GA. Role of the sympathetic nervous system in Schlager genetically hypertensive mice. Hypertension. 2009;54(4):852-859. Brandes RP. Endothelial dysfunction and hypertension. Hypertension. 2014;64(5): 924-928. Guzik TJ, Hoch NE, Brown KA, et al. Role of the T cell in the genesis of angiotensin II induced hypertension and vascular dysfunction. J Exp Med. 2007;204(10):24492460. Novershtern N, Subramanian A, Lawton LN, et al. Densely interconnected transcriptional circuits control cell states in human hematopoiesis. Cell. 2011;144(2):296-309. Bagger FO, Sasivarevic D, Sohi SH, et al. BloodSpot: a database of gene expression profiles and transcriptional programs for healthy and malignant haematopoiesis. Nucleic Acids Res. 2016;44(D1):D917924. Kim MH, Gorouhi F, Ramirez S, et al. Catecholamine stress alters neutrophil trafficking and impairs wound healing by beta2-adrenergic receptor-mediated upregulation of IL-6. J Invest Dermatol. 2014; 134(3):809-817. Nicholls AJ, Wen SW, Hall P, and MJH, Wong CHY. Activation of the sympathetic nervous system modulates 1 neutrophil function. J Leukoc Biol. 2018;103(2):295309. Levesque JP, Hendy J, Takamatsu Y, Simmons PJ, Bendall LJ. Disruption of the CXCR4/CXCL12 chemotactic interaction during hematopoietic stem cell mobiliza-

haematologica | 2019; 104(3)


SNS driven hypertension enhances hematopoiesis

34. 35.

36.

37.

tion induced by GCSF or cyclophosphamide. J Clin Invest. 2003;111(2):187196. Susic D. Hypertension, aging, and atherosclerosis. The endothelial interface. Med Clin North Am. 1997;81(5):1231-1240. Zubcevic J, Santisteban MM, Pitts T, et al. Functional neural-bone marrow pathways: implications in hypertension and cardiovascular disease. Hypertension. 2014; 63(6):e129-139. Wei Z, Spizzo I, Diep H, Drummond GR, Widdop RE, Vinh A. Differential phenotypes of tissue-infiltrating T cells during angiotensin II-induced hypertension in mice. PloS One. 2014;9(12):e114895. Schillaci G, Pirro M, Pucci G, et al.

haematologica | 2019; 104(3)

38. 39.

40.

41.

Prognostic value of elevated white blood cell count in hypertension. Am J Hypertens. 2007;20(4):364-369. Karthikeyan VJ, Lip GY. White blood cell count and hypertension. J Hum Hypertens. 2006;20(5):310-312. Nakanishi N, Sato M, Shirai K, Suzuki K, Tatara K. White blood cell count as a risk factor for hypertension; a study of Japanese male office workers. J Hypertens. 2002; 20(5):851-857. Calvi LM, Adams GB, Weibrecht KW, et al. Osteoblastic cells regulate the haematopoietic stem cell niche. Nature. 2003; 425(6960):841-846. Semerad CL, Christopher MJ, Liu F, et al. G-CSF potently inhibits osteoblast activity

and CXCL12 mRNA expression in the bone marrow. Blood. 2005;106(9):30203027. 42. Hooper AT, Butler JM, Nolan DJ, et al. Engraftment and reconstitution of hematopoiesis is dependent on VEGFR2mediated regeneration of sinusoidal endothelial cells. Cell Stem Cell. 2009; 4(3):263-274. 43. Tamplin OJ, Durand EM, Carr LA, et al. Hematopoietic stem cell arrival triggers dynamic remodeling of the perivascular niche. Cell. 2015;160(1-2):241-252. 44. Del Toro R, Chevre R, Rodriguez C, et al. Nestin(+) cells direct inflammatory cell migration in atherosclerosis. Nat Commun. 2016;7:12706.

467


ARTICLE Ferrata Storti Foundation

Red Cell Biology & its Disorders

Association of anemia with health-related quality of life and survival: a large population-based cohort study

Hanneke J.C.M. Wouters,1,2 Melanie M. van der Klauw,2 Theo de Witte,3 Reinhard Stauder,4 Dorine W. Swinkels,5,6 Bruce H.R. Wolffenbuttel2 and Gerwin Huls1

Department of Hematology, University Medical Center Groningen, the Netherlands; Department of Endocrinology, University Medical Center Groningen, the Netherlands; 3 Department of Hematology, Radboud University Medical Center, Nijmegen, the Netherlands; 4Department of Internal Medicine V (Hematology and Oncology), Medical University, Innsbruck, Austria; 5Department of Laboratory Medicine, Radboud University Medical Center, Nijmegen, the Netherlands and 6Hepcidinanalysis.com, Nijmegen, the Netherlands 1

Haematologica 2019 Volume 104(3):468-476

2

HJCMW and MMvdK contributed equally to this work. BHRW and GH contributed equally to this work.

ABSTRACT

Correspondence: HANNEKE J.C.M. WOUTERS h.j.c.m.wouters@umcg.nl Received: April 11, 2018. Accepted: October 2, 2018. Pre-published: October 11, 2018. doi:10.3324/haematol.2018.195552 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/104/3/468 Š2019 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

468

A

nemia is highly prevalent, especially in older individuals. In selected populations, anemia has been reported to be associated with impaired survival and health-related quality of life. However, data on this impact in the general population are rare. Furthermore, discussions on the optimal definition of anemia have not been conclusive. We investigated these issues using survival data, scores from a health-related quality of life questionnaire (RAND-36), and hemoglobin concentration from 138670 subjects, aged 18-93 years, participating in the Lifelines cohort. Anemia was defined according to World Health Organization criteria and was further subclassified in participants over 60 years old. Anemia was present in 5510 (4.0%) of all 138670 subjects and 516 (2.8%) in the 18667 individuals older than 60 years. Anemia had no impact on overall survival and limited impact on health-related quality of life in individuals less than 60 years old. In contrast, in individuals over 60 years old anemia significantly impaired overall survival and healthrelated quality of life. The lower health-related quality of life was mainly observed in subscales representing physical functioning. Although consensus on the subclassification of anemia is lacking, our data suggest that particularly anemia of chronic inflammation was associated with worse overall survival and decreased health-related quality of life. Multivariate models confirmed that anemia was an independent risk factor for decreased health-related quality of life in older individuals. Finally, women with a hemoglobin concentration between 12.0-13.0 g/dL (considered anemia in men, but not in women) experienced a significantly lower health-related quality of life. This large, prospective, populationbased study indicates that anemia is associated with worse overall survival and health-related quality of life in older individuals, but not in younger individuals. The findings of this study challenge the definition of anemia in women over 60 years old, and suggest that the optimal definition of anemia, in the perspective of health-related quality of life, in women over 60 years old should be altered to a hemoglobin concentration below 13.0 g/dL (8.0 mmol/L), which is comparable to that in men. haematologica | 2019; 104(3)


Anemia, survival and health-related quality of life

Introduction Anemia, according to the World Health Organization (WHO) criteria, is defined as a hemoglobin concentration <13.0 g/dL (8.0 mmol/L) in adult men and <12.0 g/dL (7.5 mmol/L) in adult, non-pregnant women. Anemia affects nearly 25% of the world population.1,2 In developed countries, anemia is one of the most frequent conditions in older individuals, with a presumed prevalence of over 10% individuals who are 65 years or older in selected populations.3,4 More than half of anemic older individuals can be diagnosed with nutritional deficiencies or have anemia of chronic inflammation (ACI). The etiology of anemia remains unknown in about one-third of the older population.4 In older individuals the presence of anemia, even mild anemia, is associated with an increased risk of falls,5 decreased physical performance,6 longer and more frequent stays in hospital,7-9 and increased mortality.7,9,10 Suggestions to raise the hemoglobin thresholds for the definition of anemia in older individuals have been made in several studies.11,12 The impact of anemia on health-related quality of life (HRQoL) has been studied in different populations of patients, including patients with chronic kidney disease,13 chronic obstructive pulmonary disease,14 cancer15 and heart failure.16 In these studies, anemia was reported to be associated with a reduced HRQoL. However, little research has been done on this potential association in community-dwelling populations. Better understanding of the impact of anemia on HRQoL in a communitydwelling general population is therefore essential and could provide critical entry points for interventions that affect anemia, especially in older individuals. In this study, we investigated the association between anemia and survival, and between anemia and HRQoL. In addition, we used this association to evaluate the definition of anemia, considering the impact of age and gender on this association in the large community-dwelling population from the Lifelines cohort.17

Methods Subjects In this study we used data from 138670 subjects participating in the Lifelines cohort study. Data from quality of life questionnaires and on hemoglobin concentration were available for these subjects. Lifelines is a multidisciplinary, prospective, population-based cohort study examining, in a unique three-generation design, the health and health-related behaviors of persons living in the north of the Netherlands. It has been shown that the Lifelines cohort is representative of the population of the northern part of the Netherlands.18 The local ethics committee approved the research protocol and informed consent was signed by every participant.

Health-related quality of life HRQoL was measured using the RAND 36-Item Health Survey.19 Since HRQoL scores are not normally distributed, for each subscale we defined a sex-specific and age-specific cut-off point at the 25th percentile of the non-anemic Lifelines population. Participants with a score lower than 25th percentile were considered to have an abnormally low score for that specific subscale.

Survival All participants were followed from the moment of their inclusion in the Lifelines cohort until death or up to May 2018 haematologica | 2019; 104(3)

(maximum follow-up 138 months, median 79 months). Information on participantsâ&#x20AC;&#x2122; deaths was obtained from the municipal personal records database.

Definition of anemia and classification into subtypes of anemia In accordance with the WHO criteria, anemia was defined as a hemoglobin concentration <13.0 g/dL (8.0 mmol/L) in adult men and <12.0 g/dL (7.5 mmol/L) in adult, non-pregnant women.20 In anemic subjects older than 60 years of age additional biochemical tests were performed, using stored plasma, to determine the type of anemia: anemia due to nutritional deficiency, ACI or unexplained anemia. There is no worldwide accepted classification into subtypes of anemia. Reference values were taken from the University Medical Center Groningen or from published literature, as indicated. Iron deficiency was considered present if the participant had two or three of the following criteria: serum ferritin concentration <30 mg/L, transferrin saturation rate <16% or hepcidin concentration <0.5 nmol/L.21-23 Transferrin saturation was calculated by dividing serum iron by total iron-binding capacity [transferrin (g/L)x25]. Folate deficiency was defined as a serum folate level <9.8 nmol/L. Vitamin B12 deficiency was defined as a serum methylmalonic acid concentration >340 nmol/L, if the estimated glomerular filtration was >30 mL/min (because methylmalonic acid levels may be elevated in people with severely impaired renal function24). If there was no evidence of nutrient deficiency, subjects with anemia were evaluated for other causes. Subjects were classified as having anemia related to chronic renal disease if the estimated glomerular filtration rate was <30 mL/min. The Chronic Kidney Disease-Epidemiology Collaboration (CDKEPI) formula was used to calculate the estimated glomerular filtration rate. ACI was defined as present if the participant had (i) a C-reactive protein concentration >5.0 mg/L or an absolute number of leukocytes >10x109/L and (ii) two or more of the following criteria: transferrin saturation rate <16%, serum ferritin concentration >100 mg/L, serum iron <10 mmol/L and hepcidin >14.7 nmol/L in men or hepcidin>15.6 nmol/L in women.25,26 If subjects with anemia could not be classified into any of these categories, they were considered, by exclusion, to have unexplained anemia. Details of the Lifelines cohort, clinical examination, biochemical measurements, RAND 36-Item Health Survey and data description and statistical analysis are provided in the Online Supplementary Data.

Results Relevant baseline characteristics are shown in Table 1 for anemic and non-anemic participants from the population-based Lifelines cohort. In total 551 men and 4959 women met the WHO criteria for anemia. The overall prevalence of anemia was 4.0% (Figure 1A). The prevalence of anemia in women was highest (about 8%) in the age cohort 40-49 years and showed a second peak of about 4% at an older age (older than 70 years). In contrast, the prevalence of anemia in men gradually increased with age, with a peak of about 5% in older age (older than 70 years). In the population of individuals 60 years of age and older the overall prevalence of anemia was 2.8%, with a prevalence of 2.7% in men and 3.0% in women. The distribution of the hemoglobin concentration according to age is shown in Online Supplementary Figure S2A,C. In men, unlike in women, the mean hemo469


H.J.C.M. Wouters et al.

globin concentration decreased with age (Online Supplementary Figure S2B,D). Over 90% of the anemic individuals had mild anemia [>10 g/dL, (> 6.2 mmol/L)]. In individuals younger than 60 years the presence of anemia did not have any impact on overall survival (Figure

1B). In contrast, the overall survival of anemic individuals older than 60 years was significantly lower than that of non-anemic individuals older than 60 years (Figures 1C). The mean of the total HRQoL score was significantly lower in anemic individuals in all age groups than in non-

Table 1. Baseline characteristics of the study cohort.

Anemic individuals (N=5510) Men/women 551 (10.0)/4959 (90.0) Age (years) 43.4 ± 11.9 Body mass index (kg/m2) 24.2 (22.1 - 27.2) Creatinine (µmol/L) 69 ± 18 Estimated glomerular filtration rate (mL/min) 98 ± 17 HbA1c (%) 5.6 ± 0.5 Glucose (mmol/L) 4.9 ± 0.8 Leukocytes (109/L) 5.6 ± 1.7 Platelets (109/L) 278 ± 70 Number of medications 2 (1 – 3) Multiple drug use 712 (12.9) Non-smokers/former 3048 (55.3)/1754 (31.8)/708 (12.8) smokers/current smokers

Non-anemic individuals (N=133160)

P-value

57283 (43.0)/75877 (57.0) 44.3 ± 12.7 25.4 (23.1 - 28.2) 74 ± 13 97± 15 5.5 ± 0.4 5.0 ± 0.8 6.1 ± 1.8 249 ± 56 2 (1 – 3) 12360 (9.3) 62228 (46.7)/42411 (31.8)/28521 (21.4)

<0.001 <0.001 <0.001 <0.001 0.003 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001

Data are given as mean ± standard deviation, median (interquartile range) when not normally distributed, or n (%). HbA1c: glycated hemoglobin.

Figure 1. Prevalence of anemia and overall survival in individuals younger and older than 60 years. (A) The prevalence of anemia as a function of sex and age. (B) Overall survival of non-anemic and anemic individuals younger than 60 years. (C) Overall survival of non-anemic and anemic individuals older than 60 years.

470

haematologica | 2019; 104(3)


Anemia, survival and health-related quality of life

anemic individuals (Online Supplementary Figure S3A). At a younger age the mean difference was relatively small, but after the age of 60 years this difference in mean HRQoL increased substantially. Because HRQoL scores are not normally distributed, we also studied the percentage of individuals with a HRQoL score below a cutoff value. This analysis revealed that the percentage of individuals with a total score below the 25th percentile was significantly higher among the group of individuals over 60 years old (Online Supplementary Figure S3B). Older individuals had a significantly lower mean HRQoL in the subscales physical functioning, social functioning, physical role functioning, emotional role functioning, vitality, bodily pain and general health (Online

Supplementary Figure S4). Moreover, a significantly larger proportion of anemic subjects older than 70 years had a score below the cut-off value, indicating a poor score, in six of the eight subscales: physical functioning (40% versus 25%), social functioning (42% versus 33%), physical role functioning (35% versus 23%), vitality (36% versus 27%), bodily pain (45% versus 32%) and general health (42% versus 32%). In subjects between 60-70 years old there was a difference in the subscales physical functioning (36% versus 27%) and physical role functioning (29% versus 21%) (Figure 2). Between 18-60 years no significant differences were observed when comparing percentages of individuals with a HRQoL score below the cut-off value. Finally, we used multivariate logistic

Figure 2. Percentages of individuals, divided according to age group, with scores below the sex- and age-specific cut-off values for the different subscales of the RAND-36 health survey. An asterisk indicates a significantly (P-value <0.01) larger percentage below the sex- and age-specific cut-off value in anemic individuals compared with non-anemic individuals. PF: physical functioning; SF: social functioning; RF: physical role functioning; RE: emotional role functioning; MH: mental health; VT: vitality; BP: bodily pain; GH: general health.

haematologica | 2019; 104(3)

471


H.J.C.M. Wouters et al.

regression analysis to investigate whether anemia was an independent risk factor for HRQoL. This analysis showed that anemia, adjusted for body mass index, being a current smoker, multiple drug use, educational level and living situation independently increased the risk of having a lower total HRQoL (OR 1.28; 95% CI: 1.00 â&#x20AC;&#x201C; 1.64; P=0.046). Since anemia had a particular impact on overall survival and HRQoL in subjects older than 60 years within the Lifelines cohort, the type of anemia in these individuals was further characterized (see the Methods section). In total 174 (35.0%) subjects were classified as having nutrient deficiency anemia, 81 (16.3%) as having ACI, and 242 (48.7%) as having unexplained anemia (Online Supplementary Table S1). Compared with individuals from the general population, a significantly larger proportion of individuals with ACI had a HRQoL score lower than the cut-off value in seven of the eight subscales (all except mental health) (Table 2). Nutrient deficiency anemia was associated with a significantly lower score in only one subscale (physical functioning). Individuals with unexplained anemia had a significantly lower score in the subscale bodily pain. These results were more explicit in men than in women (Online Supplementary Tables S2 and S3). In order to determine which type of anemia was an independent risk factor for HRQoL in individuals older than 60 years, we performed logistic regression analysis in individuals older than 60 years (Table 3 and Online Supplementary Table S4). After adjusting for body mass index, being a current smoking, multiple drug use, educational level and living situation, ACI was associated with a higher risk of having a score below the cut-off in all seven subscales which were affected by anemia (i.e. all subscales except mental health) (Table 3). Overall survival was significantly lower in individuals with ACI than in those with nutrient deficiency anemia or unexplained anemia (Figure 3).

Although the difference between the median values of the hemoglobin concentration decreased during aging, (Online Supplementary Figure S5), the definition of anemia is based on hemoglobin concentrations obtained in younger adults. Consequently, there are persistent discussions regarding whether the definition of anemia in older individuals is appropriate or should be corrected using a higher hemoglobin threshold. Because of the clear correlation between anemia in individuals older than 60 years and survival and HRQoL, we used these associations to reassess the definition of anemia in individuals older than 60 years. We assumed that anemia should be considered in the case that a certain hemoglobin level was associated with a significant (negative) impact on HRQoL. Logistic regression was used to model the relationship between hemoglobin concentration and total HRQoL in both men and women. These analyses revealed that men older than 60 years with a hemoglobin concentration between 13.0-13.7 g/dL (8.08.5 mmol/L) (just above the current WHO defined threshold) did not have a lower HRQoL compared with men with a hemoglobin concentration >13.7 g/dL (8.5 mmol/L), suggesting that the current threshold which defines anemia in men is appropriate. However, when specifically comparing women older than 60 years with a hemoglobin concentration between 12.0-13.0 g/dL (7.5-8.0 mmol/L) (considered anemic for men but not anemic for women according to the current WHO definition), this group had a lower HRQoL than that of women with a hemoglobin concentration >13.0 g/dL (8.0 mmol/L). These data strongly suggest that the optimal definition of anemia for women older than 60 years should be a hemoglobin concentration <13.0 g/dL (8.0 mmol/L) (Figure 4). So, in the case that HRQoL is used to define anemia, there would be no difference in the definition between older men and women. Due to the limited number of deaths among groups of individuals divid-

Table 2. Percentage of individuals older than 60 years with anemia with HRQoL below the sex- and age-specific cut-off value for the different HRQoL subscales as a function of their type of anemia.

General population older than 60 years (N=18152)

Nutrient deficiency anemia (N=174)

Anemia of chronic inflammation (N=81)

Unexplained anemia (N=242)

Physical functioning

4869 (26.8) 5482 (30.2)

Physical role functioning

3820 (21.0)

49 (28.2)

Emotional role functioning

1936 (10.7)

20 (11.5)

Mental health Vitality

5216 (28.7) 4820 (26.5)

42 (24.1) 49 (28.2)

Bodily pain

5618 (30.9)

60 (34.5)

General health

5104 (28.1)

60 (34.5)

40 (49.4)* P<0.001 41 (50.6)* P<0.001 41 (50.6)* P<0.001 17 (21.0)* P=0.003 26 (32.1) 39 (47.0)* P<0.001 40 (49.4)* P<0.001 40 (49.4)* P<0.001

78 (32.2)

Social functioning

67 (38.5)* P <0.001 53 (30.5)

85 (26.0) 63 (26.0) 30 (12.4) 72 (29.8) 75 (31.0) 95 (39.3)* P=0.006 79 (32.6)

Data are given as n (%) below the sex- and age-specific cut-off value for the different HRQoL subscales.* P-value <0.01 between individuals from the non-anemic group of the general population and individuals with the given type of anemia.

472

haematologica | 2019; 104(3)


Anemia, survival and health-related quality of life

ed according to the various intervals of hemoglobin concentration, survival could not be used to model the optimal definition of anemia.

Discussion In this study we observed that anemia is associated with worse overall survival and is an independent risk factor for HRQoL in individuals older than 60 years, in contrast to younger individuals. Although consensus on the subclassification of anemia is lacking, our data suggest that particularly ACI was an independent risk factor for HRQoL in subscales representing physical functioning. Furthermore, we observed that women (but not men) older than 60 years with a hemoglobin concentration in the lower normal WHO range experienced decreased HRQoL. This suggests that the definition of anemia in women older than 60 years should be altered to a hemoglobin concentration <13.0 g/dL (8.0 mmol/L), which is comparable to the definition of anemia in men. The prevalence of anemia according to the WHO criteria was 4.0% in the whole cohort and 2.8% among subjects older than 60 years, which is consistent with the findings of population-based studies from Sweden (3.8% in patients 44 - 73 years old)27 and Germany (3.2% in

Table 3. Risk of having a lower score than the (age- and sex- specific) 25th percentile cut-off due to anemia per HRQoL subscale among individuals older than 60 years.

HRQoL subscale

Type of anemia

Physical functioning

Nutritional deficiency anemia Anemia of chronic inflammation

Social functioning

Unexplained anemia Nutritional deficiency anemia Anemia of chronic inflammation

Physical role functioning

Unexplained anemia Nutritional deficiency anemia Anemia of chronic inflammation

Unexplained anemia Emotional role Nutritional deficiency anemia functioning Anemia of chronic inflammation

Vitality

Unexplained anemia Nutritional deficiency anemia Anemia of chronic inflammation

Bodily pain

Unexplained anemia Nutritional deficiency anemia Anemia of chronic inflammation

Unexplained anemia General health Nutritional deficiency anemia Anemia of chronic inflammation Unexplained anemia

Odds ratio (95% CI) 1.16 (0.75 – 1.79) 3.30 (1.76 – 6.21) **P<0.001 1.04 (0.74 – 1.48) 0.89 (0.58 - 1.35) 2.39 (1.32 – 4.31) **P =0.004 1.04 (0.75 - 1.44) 1.15 (0.74 – 1.78) 3.75 (2.05- 6.83) **P<0.001 1.03 (0.72 - 1.48) 0.89 (0.49 – 1.59) 2.53 (1.31 – 4.90) **P=0.006 1.04 (0.66 – 1.65) 0.92 (0.60 - 1.42) 2.89 (1.59 – 5.24) **P<0.001 1.13 (0.81 - 1.57) 0.96 (0.63 - 1.45) 1.84 (1.02 – 3.33) *P=0.04 1.16 (0.84 – 1.60) 0.87 (0.57 - 1.34) 2.44 (1.33 – 4.45) **P =0.004 0.99 (0.71 – 1.38)

Adjusted for body mass index, smoking status, multiple drug use, educational level and living situation. Data are shown as odds ratio and 95% confidence interval (95% CI). **P-value <0.01; *Pvalue <0.05.

haematologica | 2019; 104(3)

subjects 45-75 years old28 and 6.9% in subjects over 65 years old29). However, the prevalence of anemia in older individuals in our community-dwelling population was remarkably lower than that found in previous studies.3,4,30 A possible explanation for this discrepancy could be a participation bias in the Lifelines cohort, although earlier observations showed that the Lifelines study population is representative of the general population, with a low risk of selection.18 The relatively low prevalence of anemia in older individuals could also be related to the high quality and accessibility of the general health system nowadays, especially for older individuals. Furthermore, the prevalence of anemia and the types of anemia might differ according to race and geographical area.2,4 Anemia did not have an impact on survival in individuals younger than 60 years. In contrast, among individuals older than 60 years, anemic subjects had a significantly decreased survival. The decreased survival in older anemic individuals is in accordance with that in multiple previous studies.7,9,10 In our study there was increased mortality in older individuals with ACI: this finding is partly in line with the findings of a study by Willems et al. who reported increased mortality in subjects with nutrient deficiency anemia and ACI, but not in those with unexplained anemia.31 Since an aging population results in an increased prevalence of anemia, an understanding of the association between anemia and HRQoL is important. Remarkably, only a few, relatively small studies of communitydwelling populations have described this association. In a study of 717 subjects, Lucca et al. showed, after adjustment for a large number of demographic and clinical confounders, that anemia in elderly Italians (65 - 84 years old) was significantly associated with lower disease-specific QoL, measured with the Functional Assessment of Cancer Therapy-Anemia.32 Thein et al. studied an American cohort of 328 individuals and showed that anemia in subjects older than 65 years was independently associated with impairments in multiple subscales of HRQoL (measured with the Short Form-36), especially in measures of functional limitation.33 Bang et al. also found that anemia was associated with physical impairment in 695 Korean individuals older than 65 years.34 Our data, in an unprecedented large cohort, confirms the relationship between anemia and HRQoL in older individuals (but not in younger individuals). There is no worldwide accepted classification into subtypes of anemia. To be able to study the subtypes, we based our criteria on literature or reference values from the local laboratory. In addition to those criteria, we also analyzed the data with different classification criteria (see Online Supplementary Data for definitions). Although misclassification within the various subtypes of anemia cannot be excluded, our data suggest that the impact of anemia on HRQoL is mainly present in individuals with ACI (Online Supplementary Tables S5 and S6). The pathophysiology of ACI is multifactorial, including increased levels of inflammatory cytokines that lead to altered iron metabolism, with a key role for the iron regulatory peptide hormone hepcidin.35,36 Indeed, elevated levels of inflammatory markers, which are found in older anemic individuals, have been shown to be associated with physical disability.37,38 Our observation that anemia mainly affects HRQoL of subjects with ACI is supported by studies showing that erythropoiesis-stimulating agents 473


H.J.C.M. Wouters et al. Figure 3. Kaplan-Meier curves of survival of individuals with different types of anemia.

have an impact on the HRQoL of patients with cancer, HIV/AIDS and chronic kidney disease.39 In older individuals a positive effect of erythropoiesis-stimulating agents on QoL has been shown in patients with chronic kidney disease,40 and chemotherapy-induced anemia.41 In one randomized, blinded, placebo-controlled trial in predominantly African-American elderly women with ACI or unexplained anemia an increase of the hemoglobin concentration by 2 g/dL led to improvements in QoL as measured by the Functional Assessment of Chronic Illness Therapy instrument.42 The most pronounced association between HRQoL and anemia was observed in older individuals and in subscales related to physical health. This suggests that younger individuals can compensate better for the physical consequences of anemia, limiting the impact of anemia on their HRQoL. For older individuals these compensatory mechanisms might be limited compared to those in younger persons. The ability to engage in physical activities is known to be important for maintaining QoL in older persons.43,44 In this study we defined anemia according to the WHO criteria, which define anemia as a hemoglobin concentration <13.0 g/dL (8.0 mmol/L) in adult men and <12.0 g/dL (7.5 mmol/L) in adult, non-pregnant women. This is the most commonly used definition of anemia, but the use of this definition in older/elderly populations has been criticized.11,12 The WHO criteria were derived from a small study population of healthy, young participants nearly 50 years ago.20 Based on data from two large American databases, Beutler et al. proposed new, slightly higher cut-offs for white women older than 50 years and white men older than 60 years.12 Although the different distributions of median hemoglobin concentration between men and women (in general and in older individuals) suggest that it is ‘logic’ to use different definitions of anemia in men and women, 474

our data raise doubts about this logic. Indeed, HRQoL data from our study suggest that the definition of anemia in men – a hemoglobin concentration <13.0 g/dL (8.0 mmol/L) – should also be applied for women older than 60 years. Apparently, the hemoglobin values which start to hamper (physical) functioning do not differ between older men and women. We were not able to study the optimal definition of anemia in perspective of mortality because of the limited number of deaths in our cohort. Martinsson et al. observed an increased mortality defining anemia as <14.0 g/dL (8.7 mmol/L) in men and <13.0 g/dL (8.0 mmol/L) in women.27 A recent, large study assessing the effect of hemoglobin concentrations on cardiovascular and all-cause mortality in nearly 300,000 participants showed that men with a hemoglobin concentration in the lower normal range had a higher risk of mortality whereas this was not observed in women.45 Our study has some strengths and limitations. It is the first study to investigate the possible effect of anemia on HRQoL in individuals of all ages in a very large cohort from the general population. Additionally, we used information on HRQoL from a large number of participants with a wide range of age, socio-economic status and comorbidities. All subjects were uniformly characterized and were not aware of hemoglobin status when filling out the questionnaires, preventing the outcome of the questionnaire from being affected by the knowledge of a diagnosis of anemia. Several potential limitations should be acknowledged. As mentioned earlier, the subtypes of anemia might have been misclassified due to a lack of clear classification criteria. Since this is a cross-sectional study, the analyses do not provide information about causality, for example for the fact that despite lower median hemoglobin concentrations in older women than in older men, the impact of anemia on HRQoL occurred at the same hemoglobin concentration. Additionally, given the observational haematologica | 2019; 104(3)


Anemia, survival and health-related quality of life

Figure 4. Risk of having a lower total health-related quality of life score than the (sex- and age-specific) 25th percentile cut-off according to hemoglobin concentrations. The red rectangles indicate a hemoglobin concentration below the currently used World Health Organization definition of anemia. 95% CI: 95% confidence interval.

nature of the study, it was not possible to exclude the role of unknown or unmeasured confounding variables. A third limitation is that HRQoL was measured with the RAND-36 questionnaire, which is a generic health status questionnaire and not specifically developed to measure QoL in anemic individuals. Finally, it is possible that the results were influenced by volunteer bias. This study indicates that anemia is associated with decreased survival and HRQoL, especially that concerning physical health, in individuals older than 60 years. After adjusting for confounding variables the impact of anemia was particularly present in individuals with ACI. These results suggest that older individuals might benefit from treatment to increase their hemoglobin values. Furthermore, this study challenges the sex-dependent definition of anemia in individuals older than 60 years, and suggests that in both men and women older than 60 years, the definition currently used in men [hemoglobin concentration <13.0 g/dL (8.0 mmol/L)] could be applied in perspective of HRQoL. Acknowledgments The authors would like to thank all participants of the Lifelines cohort study and everybody involved in the set-up and implementation of the study. We would also like to thank Anneke Geurts-Moespot and Siem Klaver of haematologica | 2019; 104(3)

Hepcidinanalysis.com, who contributed to the measurement of hepcidin. Funding Lifelines has been funded by a number of public sources, notably the Dutch Government, The Netherlands Organization of Scientific Research NOW (grant 175.010.2007.006), the Northern Netherlands Collaboration of Provinces (SNN), the European fund for regional development, Dutch Ministry of Economic Affairs, Pieken in de Delta, Provinces of Groningen and Drenthe, the Target project, BBMRI-NL, the University of Groningen, and the University Medical Center Groningen, the Netherlands. This work was supported by the National Consortium for Healthy Aging, and funds from the European Union’s Seventh Framework program (FP7/2007–2013) through the BioSHaRE-EU (Biobank Standardisation and Harmonisation for Research Excellence in the European Union) project, grant agreement 261,433. Lifelines (BRIF4568) is engaged in a Bioresource Research Impact Factor (BRIF) policy pilot study, details of which can be found at: http://bioshare.eu/content/bioresource-impact-factor. This work is part of the MDS-RIGHT activities, which has received funding from the European Union’s Horizon 2020 research and innovation program under grant agreement n, 634789 - 'Providing the right care to the right patient with myelodysplastic syndrome at the right time'. 475


H.J.C.M. Wouters et al.

References 1. McLean E, Cogswell M, Egli I, Wojdyla D, de Benoist B. Worldwide prevalence of anaemia, WHO Vitamin and Mineral Nutrition Information System, 1993-2005. Public Health Nutr. 2009;12(4):444-454. 2. Stevens GA, Finucane MM, De-Regil LM, et al. Global, regional, and national trends in haemoglobin concentration and prevalence of total and severe anaemia in children and pregnant and non-pregnant women for 1995-2011: a systematic analysis of population-representative data. Lancet Glob Health. 2013;1(1):e16-25. 3. Gaskell H, Derry S, Moore A, McQuay HJ. Prevalence of anaemia in older persons: systematic review. BMC Geriatr. 2008;8:1. 4. Guralnik JM, Eisenstaedt RS, Ferrucci L, Klein HG, Woodman RC. Prevalence of anemia in persons 65 years and older in the United States: evidence for a high rate of unexplained anemia. Blood. 2004;104(8): 2263-2268. 5. Penninx BW, Pluijm SM, Lips P, et al. Latelife anemia is associated with increased risk of recurrent falls. J Am Geriatr Soc. 2005;53(12):2106-2111. 6. Penninx BW, Pahor M, Cesari M, et al. Anemia is associated with disability and decreased physical performance and muscle strength in the elderly. J Am Geriatr Soc. 2004;52(5):719-724. 7. Culleton BF, Manns BJ, Zhang J, Tonelli M, Klarenbach S, Hemmelgarn BR. Impact of anemia on hospitalization and mortality in older adults. Blood. 2006;107(10):38413846. 8. Dharmarajan TS, Pankratov A, Morris E, et al. Anemia: its impact on hospitalizations and length of hospital stay in nursing home and community older adults. J Am Med Dir Assoc. 2008;9(5):354-359. 9. Penninx BW, Pahor M, Woodman RC, Guralnik JM. Anemia in old age is associated with increased mortality and hospitalization. J Gerontol A Biol Sci Med Sci. 2006;61(5):474-479. 10. Halawi R, Moukhadder H, Taher A. Anemia in the elderly: a consequence of aging? Expert Rev Hematol. 2017;10(4):327335. 11. Chaves PH, Ashar B, Guralnik JM, Fried LP. Looking at the relationship between hemoglobin concentration and prevalent mobility difficulty in older women. Should the criteria currently used to define anemia in older people be reevaluated? J Am Geriatr Soc. 2002;50(7):1257-1264. 12. Beutler E, Waalen J. The definition of anemia: what is the lower limit of normal of the blood hemoglobin concentration? Blood. 2006;107(5):1747-1750. 13. Farag YM, Keithi-Reddy SR, Mittal BV, et al. Anemia, inflammation and health-related quality of life in chronic kidney disease patients. Clin Nephrol. 2011;75(6):524-533. 14. Ferrari M, Manea L, Anton K, et al. Anemia and hemoglobin serum levels are associated with exercise capacity and quality of life in chronic obstructive pulmonary disease. BMC Pulm Med. 2015;15:58. 15. Wasada I, Eguchi H, Kurita M, et al. Anemia affects the quality of life of Japanese cancer

476

16.

17.

18.

19.

20.

21.

22. 23.

24.

25.

26. 27.

28.

29.

30.

31.

patients. Tokai J Exp Clin Med. 2013;38 (1):7-11. Kraai IH, Luttik ML, Johansson P, et al. Health-related quality of life and anemia in hospitalized patients with heart failure. Int J Cardiol. 2012;161(3):151-155. Stolk RP, Rosmalen JG, Postma DS, et al. Universal risk factors for multifactorial diseases: LifeLines: a three-generation population-based study. Eur J Epidemiol. 2008;23 (1):67-74. Klijs B, Scholtens S, Mandemakers JJ, Snieder H, Stolk RP, Smidt N. Representativeness of the LifeLines cohort study. PLoS One. 2015;10(9):e0137203. VanderZee K I KI. Psychometric qualities of the RAND 36-Item Health Survey 1.0: a multidimensional measure of general health status. Int J Behav Med. 1996;3(2):104-122. Blanc B, Finch CA, Hallberg L, et al. Nutritional anaemias. Report of a WHO scientific group. World Health Organ Tech Rep Ser. 1968;405:5-37. Mast AE, Blinder MA, Gronowski AM, Chumley C, Scott MG. Clinical utility of the soluble transferrin receptor and comparison with serum ferritin in several populations. Clin Chem. 1998;44(1):45-51. Lopez A, Cacoub P, Macdougall IC, PeyrinBiroulet L. Iron deficiency anaemia. Lancet. 2016;387(10021):907-916. Kroot JJ, Laarakkers CM, Geurts-Moespot AJ, et al. Immunochemical and mass-spectrometry-based serum hepcidin assays for iron metabolism disorders. Clin Chem. 2010;56(10):1570-1579. Ganji V, Kafai MR. Population prevalence, attributable risk, and attributable risk percentage for high methylmalonic acid concentrations in the post-folic acid fortification period in the US. Nutr Metab (Lond). 2012;9(1):2. Guralnik JM, Eisenstaedt RS, Ferrucci L, Klein HG, Woodman RC. Prevalence of anemia in persons 65 years and older in the United States: evidence for a high rate of unexplained anemia. Blood. 2004;104(8): 2263-2268. Weiss G, Goodnough LT. Anemia of chronic disease. N Engl J Med. 2005;352(10):10111023. Martinsson A, Andersson C, Andell P, Koul S, Engstrom G, Smith JG. Anemia in the general population: prevalence, clinical correlates and prognostic impact. Eur J Epidemiol. 2014;29(7):489-498. Eisele L, Durig J, Broecker-Preuss M, et al. Prevalence and incidence of anemia in the German Heinz Nixdorf Recall Study. Ann Hematol. 2013;92(6):731-737. Endres HG, Wedding U, Pittrow D, Thiem U, Trampisch HJ, Diehm C. Prevalence of anemia in elderly patients in primary care: impact on 5-year mortality risk and differences between men and women. Curr Med Res Opin. 2009;25(5):1143-1158. Stauder R, Thein SL. Anemia in the elderly: clinical implications and new therapeutic concepts. Haematologica. 2014;99(7):11271130. Willems JM, den Elzen WP, Vlasveld LT, et al. No increased mortality risk in older persons with unexplained anaemia. Age Ageing. 2012;41(4):501-506.

32. Lucca U, Tettamanti M, Mosconi P, et al. Association of mild anemia with cognitive, functional, mood and quality of life outcomes in the elderly: the "Health and Anemia" study. PLoS One. 2008;3(4):e1920. 33. Thein M, Ershler WB, Artz AS, et al. Diminished quality of life and physical function in community-dwelling elderly with anemia. Medicine (Baltimore). 2009;88(2): 107-114. 34. Bang SM, Lee JO, Kim YJ, et al. Anemia and activities of daily living in the Korean urban elderly population: results from the Korean Longitudinal Study on Health and Aging (KLoSHA). Ann Hematol. 2013;92(1):59-65. 35. Langer AL, Ginzburg YZ. Role of hepcidinferroportin axis in the pathophysiology, diagnosis, and treatment of anemia of chronic inflammation. Hemodial Int. 2017;21 (Suppl 1):S37-S46. 36. Weiss G, Goodnough LT. Anemia of chronic disease. N Engl J Med. 2005;352(10):10111023. 37. Ferrucci L, Penninx BW, Volpato S, et al. Change in muscle strength explains accelerated decline of physical function in older women with high interleukin-6 serum levels. J Am Geriatr Soc. 2002;50(12):19471954. 38. Olivares M, Hertrampf E, Capurro MT, Wegner D. Prevalence of anemia in elderly subjects living at home: role of micronutrient deficiency and inflammation. Eur J Clin Nutr. 2000;54(11):834-839. 39. Kimel M, Leidy NK, Mannix S, Dixon J. Does epoetin alfa improve health-related quality of life in chronically ill patients with anemia? Summary of trials of cancer, HIV/AIDS, and chronic kidney disease. Value Health. 2008;11(1):57-75. 40. Roger SD, Jassal SV, Woodward MC, Soroka S, McMahon LP. A randomised single-blind study to improve health-related quality of life by treating anaemia of chronic kidney disease with Aranesp(R) (darbepoetin alfa) in older people: STIMULATE. Int Urol Nephrol. 2014;46(2):469-475. 41. Boccia R, Lillie T, Tomita D, Balducci L. The effectiveness of darbepoetin alfa administered every 3 weeks on hematologic outcomes and quality of life in older patients with chemotherapy-induced anemia. Oncologist. 2007;12(5):584-593. 42. Agnihotri P, Telfer M, Butt Z, et al. Chronic anemia and fatigue in elderly patients: results of a randomized, double-blind, placebo-controlled, crossover exploratory study with epoetin alfa. J Am Geriatr Soc. 2007;55(10):1557-1565. 43. Vagetti GC, Barbosa Filho VC, Moreira NB, Oliveira V, Mazzardo O, Campos W. Association between physical activity and quality of life in the elderly: a systematic review, 2000-2012. Rev Bras Psiquiatr. 2014;36(1):76-88. 44. Molzahn A, Skevington SM, Kalfoss M, Makaroff KS. The importance of facets of quality of life to older adults: an international investigation. Qual Life Res. 2010;19(2):293-298. 45. Lee G, Choi S, Kim K, et al. Association of hemoglobin concentration and its change with cardiovascular and all-cause mortality. J Am Heart Assoc. 2018;7(3).

haematologica | 2019; 104(3)


ARTICLE

Red Cell Biology & its Disorders

Sotatercept, a novel transforming growth factor b ligand trap, improves anemia in b-thalassemia: a phase II, open-label, dose-finding study

Maria Domenica Cappellini,1 John Porter,2 Raffaella Origa,3 Gian Luca Forni,4 Ersi Voskaridou,5 Frédéric Galactéros,6 Ali T. Taher,7 Jean-Benoît Arlet,8,9,10 JeanAntoine Ribeil,11 Maciej Garbowski,2 Giovanna Graziadei,1 Chantal Brouzes,11 Michaela Semeraro,11 Abderrahmane Laadem,12 Dimana Miteva,13 Jun Zou,12 Victoria Sung,14 Tatiana Zinger,13 Kenneth M. Attie15 and Olivier Hermine9,10,16

Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico, Department of Clinical Sciences and Community Health, University of Milan, Italy; 2Department of Haematology, University College London, UK; 3Day Hospital Talassemia, Ospedale Pediatrico Microcitemico “A.Cao”, A.O. “G. Brotzu”, Cagliari, Italy; 4Centro della Microcitemia, Ospedali Galliera, Genova, Italy; 5Thalassemia Center, Laikon General Hospital, Athens, Greece; 6 UMGGR, Hôpital Henri Mondor; Assistance Publique-Hôpitaux de Paris (APHP); UPEC, Créteil, France; 7Department of Internal Medicine, American University of Beirut Medical Center, Lebanon; 8Department of Internal Medicine, APHP, Hôpital Européen GeorgesPompidou, Paris, France; 9INSERM UMR1163, CNRS ERL 8254, Institut Imagine, Université Paris Descartes-Sorbonne Paris Cité, France; 10Laboratory of Excellence GR-Ex, Paris, France; 11Laboratory of Onco-hematology, Hôpital Necker-Enfants Malades, Paris, France; 12 Celgene Corporation, Summit, NJ, USA; 13Celgene Corporation, Boudry, Switzerland; 14 Celgene Corporation, San Francisco, CA, USA; 15Acceleron Pharma, Cambridge, MA, USA and 16Department of Hematology, APHP, Hôpital Necker, Paris, France 1

b

Haematologica 2019 Volume 104(3):477-484

ABSTRACT

-thalassemia, a hereditary blood disorder caused by defective synthesis of hemoglobin b globin chains, leads to ineffective erythropoiesis and chronic anemia that may require blood transfusions. Sotatercept (ACE-011) acts as a ligand trap to inhibit negative regulators of late-stage erythropoiesis in the transforming growth factor b superfamily, correcting ineffective erythropoiesis. In this phase II, open-label, dose-finding study, 16 patients with transfusion-dependent b-thalassemia and 30 patients with non-transfusion-dependent b-thalassemia were enrolled at seven centers in four countries between November 2012 and November 2014. Patients were treated with sotatercept at doses of 0.1, 0.3, 0.5, 0.75, or 1.0 mg/kg to determine a safe and effective dose. Doses were administered by subcutaneous injection every 3 weeks. Patients were treated for ≤22 months. Response was assessed as a ≥20% reduction in transfusion burden sustained for 24 weeks in transfusion-dependent b-thalassemia patients, and an increase in hemoglobin level of ≥1.0 g/dL sustained for 12 weeks in non-transfusiondependent b-thalassemia patients. Sotatercept was well tolerated. After a median treatment duration of 14.4 months (range 0.6-35.9), no severe lifethreatening adverse events were observed. Thirteen percent of patients reported serious but manageable adverse events. The active dose of sotatercept was ≥0.3 mg/kg for patients with non-transfusion-dependent b-thalassemia and ≥0.5 mg/kg for those with transfusion-dependent b-thalassemia. Of 30 non-transfusion-dependent b-thalassemia patients treated with ≥0.1 mg/kg sotatercept, 18 (60%) achieved a mean hemoglobin increase ≥1.0 g/dL, and 11 (37%) an increase ≥1.5 g/dL, sustained for ≥12 weeks. Four (100%) transfusion-dependent b-thalassemia patients treated with 1.0 mg/kg sotatercept achieved a transfusion-burden reduction of ≥20%. Sotatercept was effective and well tolerated in patients with b-thalassemia. Most patients with non-transfusion-dependent b-thalassemia treated with higher doses achieved sustained increases in hemoglobin level. Transfusion-dependent b-thalassemia patients treated with higher doses of sotatercept achieved notable reductions in transfusion requirements. This trial was registered at ClinicalTrials.gov with the number NCT01571635. haematologica | 2019; 104(3)

Ferrata Storti Foundation

Correspondence: MARIA DOMENICA CAPPELLINI maria.cappellini@unimi.it OLIVIER HERMINE olivier.hermine@nck.aphp.fr Received: May 30, 2018. Accepted: October 12, 2018. Pre-published: October 18, 2018. doi:10.3324/haematol.2018.198887 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/104/3/477 ©2019 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

477


M.D. Cappellini et al.

Introduction b-thalassemia is a hereditary blood disorder caused by defective synthesis of the b globin chains of hemoglobin1 characterized by ineffective erythropoiesis.2-4 Mutations in the b-globin genes lead to reduced or absent b-globin chain synthesis, increasing the ratio of α-globin to non-αglobin chains. Due to the relative excess of α-globin chains, α-globin precipitates within erythroblasts as hemichromes, leading to oxidative stress, maturation arrest, membrane damage and apoptosis of late-stage erythroid precursors, and reduced red blood cell (RBC) life span.4-6 Although erythropoiesis-stimulating agents have been used in patients with b-thalassemia, ineffective erythropoiesis is not corrected.7 Use of erythropoiesis-stimulating agents is, therefore, not recommended for the treatment of b-thalassemia.7,8 b-thalassemia phenotypes vary in severity, ranging from asymptomatic thalassemia minor to non-transfusiondependent thalassemia (NTDT) (including thalassemia intermedia and hemoglobin E - b-thalassemia) to transfusion-dependent thalassemia (TDT) (thalassemia major). Treatment of TDT involves regular and lifelong blood transfusions leading to iron overload; long-term management of iron overload requires regular iron chelation therapy.9,10 However, iron chelation therapy is associated with significant toxicities and requires a high level of treatment adherence and monitoring that can be difficult to manage and may have a negative impact on patients’ quality of life.11,12 Bone marrow transplantation offers potentially curative treatment,13 but is not possible in all patients,14 and is associated with significant morbidity and mortality.15 Gene therapy has shown early promise but is still under investigation.16 The treatment of NTDT is based on managing the long-term complications of ineffective erythropoiesis, including chronic anemia and iron overload,17 using iron chelation therapy and occasional RBC transfusion.18,19 Sotatercept is a ligand trap that inhibits transforming growth factor beta (TGF-b) superfamily members including growth differentiation factor 11 (GDF-11) and activin B.20,21 GDF-11 is overexpressed in immature erythroblasts in b-thalassemia.21 Aberrant GDF-11 production may induce expansion of erythroid progenitors and increase oxidative stress, leading to maturation arrest of late erythroid precursors and ineffective erythropoiesis.21 Preclinical work has shown that administration of an activin receptor IIA (ActRIIA) ligand trap decreases GDF11 concentration, reduces reactive oxidative stress levels, and promotes terminal maturation in immature erythroblasts.21 Sotatercept is a novel recombinant fusion protein consisting of the extracellular domain of the human ActRIIA (ACVR2A) linked to the human immunoglobulin G1 Fc domain.20 When administered subcutaneously, sotatercept increased hemoglobin levels and RBC count in healthy, postmenopausal women.22 In a phase II trial of patients with lower-risk myelodysplastic syndromes and anemia, sotatercept reduced transfusion burden in 47% of patients with a high transfusion burden, and increased hemoglobin levels in 58% of patients with a low transfusion burden.23 In a b-thalassemia mouse model, RAP-011 (a murine analog of sotatercept) improved hematologic parameters including RBC count, total hemoglobin, hematocrit, and mean corpuscular volume.21 The aim of this phase II study was to determine a safe, 478

tolerable, and effective dose of sotatercept to increase hemoglobin levels and reduce blood transfusion burden in adults with TDT and NTDT b-thalassemia.

Methods In this phase II, open-label, dose-finding study, patients were enrolled at seven centers in France, Greece, Italy, and the United Kingdom (listed in the Online Supplementary Data) between November 2012 and November 2014. The study was approved by individual institutional review boards at participating centers and, where appropriate, national health authorities, and was conducted in compliance with the Declaration of Helsinki. All patients provided written informed consent. ClinicalTrials.gov registration number: NCT01571635.

Inclusion criteria

Eligible patients were aged ≥18 years with a diagnosis of TDT or NTDT, and an Eastern Cooperative Oncology Group Performance Status of 0 to 1. Transfusion dependence was defined as receiving ≥2 RBC units every 30 days for ≥168 days prior to study enrollment, with no transfusion-free period of >45 consecutive days during this period. The mean hemoglobin level before transfusion was ≤10.5 g/dL in the 168 days prior to enrollment, with the last pretransfusion hemoglobin level preceding enrollment being ≤10.5 g/dL. Non-transfusion dependence was defined as ≤1 episode of transfusion during the 168 days prior to enrollment; an episode of transfusion was defined as ≤4 RBC units received during the 168 days prior to enrollment. Exclusion criteria are listed in the Online Supplementary Data.

Study design To determine a safe and tolerable dose of sotatercept, a doseescalation study was carried out. Patients were initially enrolled in two cohorts, receiving doses of 0.1 mg/kg and 0.3 mg/kg, administered by subcutaneous injection every 3 weeks. Four dose-escalation cohorts, with doses of 0.5, 0.75, 1.0, and 1.5 mg/kg, were subsequently opened to enrollment. Details of the study design are included in the Online Supplementary Data.

Study endpoints Primary efficacy endpoints were a reduction in transfusion burden of ≥20% from pretreatment levels, sustained for 24 weeks in TDT patients, and an increase in hemoglobin level of ≥1.0 g/dL sustained for 12 weeks from mean pretreatment hemoglobin levels in NTDT patients. Hematologic parameters, including hemoglobin levels and RBC counts, were measured on days 1, 8, and 15 (±3 days) of each 3-week sotatercept dose period. Secondary endpoints included reduction in RBC transfusion burden in TDT patients, hemoglobin level increase from baseline in NTDT patients, and safety. Exploratory endpoints included iron metabolism markers (including serum ferritin and hepcidin), and clinical symptoms associated with ineffective erythropoiesis and anemia (including extramedullary hematopoiesis, leg ulcers, and bilirubin levels).

Statistical analysis Efficacy analyses were carried out on the intent-to-treat population, which included all patients enrolled for treatment. Efficacy data are presented by assigned dose group. Safety analyses were conducted on the safety population, defined as those patients who received one or more dose of sotatercept. Safety data are presented prior to intrapatient dose escalation for sotatercept dose groups and presented post-intrapatient dose escalation for patients overhaematologica | 2019; 104(3)


Sotatercept TGF-b ligand trap in β-thalassemia

all. The data cutoff date for this analysis was November 27, 2015. Detailed statistical methods are included in the Online Supplementary Data.

Results Patients As of November 27, 2015, 46 patients had been enrolled between November 2012 and November 2014 – 16 with TDT and 30 with NTDT (Online Supplementary Figure S1). The patients’ baseline demographic and disease characteristics are shown in Table 1; the data are presented by assigned dose level of sotatercept prior to dose escalation. All patients received two or more doses of sotatercept. The median duration of treatment was 19.6 months (range, 0.6-35.9) for NTDT patients and 13.8 months (range, 1.4-27.7) for TDT patients.

Erythroid response

Non-transfusion-dependent b-thalassemia patients

Of the 30 patients with NTDT treated with sotatercept

doses of 0.1 to 1.0 mg/kg, 18 (60%) achieved a mean hemoglobin increase of ≥1.0 g/dL, and 11 (37%) had a mean hemoglobin increase of ≥1.5 g/dL sustained for ≥12 weeks (Figure 1A). The highest rate of response was seen with 0.75 mg/kg sotatercept: 86% (6 of 7) of the patients treated with this dose achieved an increase in hemoglobin of ≥1.5 g/dL. Responders receiving sotatercept 0.5 mg/kg experienced the greatest maximum increase in hemoglobin levels within 12 weeks versus baseline (3.2±0.2 g/dL) (Figure 1B). No patients receiving sotatercept 0.1 mg/kg achieved a response. The mean change in hemoglobin levels from baseline in NTDT patients is shown in Figure 2. No significant differences in reticulocyte count or fetal hemoglobin (HbF) levels were reported for NTDT patients during the study (data not shown). The starting active dose of sotatercept in patients with NTDT, based on the observed responses, was ≥0.3 mg/kg. Of four patients with NTDT who were receiving concomitant hydroxyurea therapy at enrollment, three continued to receive hydroxyurea during the study (dose range 54-84 mg/kg) without interruption or modification;

Table 1. Baseline characteristics of patients with non-transfusion-dependent and transfusion-dependent b-thalassemia treated with sotatercept by assigned dose level.

Characteristic

NTDT patients Age, years Median Range Female, n (%) Weight, kg Median Range Hemoglobin, g/dL Median Range Mean±SD Mean corpuscular volume, fL Mean Median Range TDT patients Age, years Median Range Female, n (%) Weight, kg Median Range b-thalassemia major, n (%) b-thalassemia intermedia*, n (%) Transfusion burden, RBC units/24 weeks† Hemoglobin, g/dL Median Range

Sotatercept dose 0.5 mg/kg 0.75 mg/kg (n=8) (n=12)

0.1 mg/kg (n=8)

0.3 mg/kg (n=9)

1.0 mg/kg (n=9)

Overall (N=46)

n=6

n=6

n=6

n=7

n=5

n=30

39.5 32.0-53.0 3 (50)

47.0 37.0-55.0 3 (50)

39.0 34.0-54.0 3 (50)

43.0 29.0-65.0 6 (86)

41.0 19.0-52.0 1 (20)

42.0 19.0-65.0 16 (53)

54.0 48.0-85.0

58.0 53.6-65.0

64.2 41.0-75.0

56.0 47.6-75.0

62.0 56.0-67.0

58.9 41.0-85.0

8.8 5.9-10.7 8.6±1.68

8.5 6.0-9.5 8.3±1.26

8.5 6.4-9.3 8.2±1.13

8.7 7.1-9.6 8.5±0.89

7.6 6.6-9.4 7.7±1.04

8.4 5.9-10.7 8.3±1.18

76.2 75.8 67.7-83.8 n=2

75.6 75.8 61.6-90.5 n=3

75.5 68.4 62.6-103.4 n=2

63.2 61.1 54.6-80.4 n=5

78.6 80.3 70.6-89.3 n=4

73.3 72.1 54.6-103.4 n=16

41.5 34.0-49.0 1 (50)

34.0 23.0-39.0 1 (33)

35.0 34.0-36.0 0 (0)

45.0 33.0-51.0 2 (40)

39.5 27.0-46.0 2 (50)

35.5 23.0-51.0 6 (38)

49.4 46.9-51.8 2 (100) 0 15, 33

65.7 52.0-71.0 1 (33) 2 (67) 14, 18, 33

70.7 60.5-80.8 2 (100) 0 30, 30

58.0 48.5-83.0 3 (60) 2 (40) 8, 18, 18, 30, 35

63.0 53.5-85.9 4 (100) 0 18, 18, 18, 24

60.1 46.9-85.9 12 (75) 4 (25) -

9.5 9.3-9.7

8.6 8.5-10.6

9.8 9.4-10.1

8.9 8.0-10.0

9.6 8.6-10.9

9.3 8.0-10.9

NTDT: non-transfusion-dependent b-thalassemia; TDT: transfusion-dependent b-thalassemia; SD: standard deviation; RBC: red blood cell. *Patients with b-thalassemia intermedia gene mutations who met the transfusion burden requirement were classified as having TDT. †Values presented for transfusion burden for individual TDT patients.

haematologica | 2019; 104(3)

479


M.D. Cappellini et al. two achieved a mean hemoglobin increase of ≥1 g/dL sustained for ≥12 weeks.

Transfusion-dependent b-thalassemia patients

levels were lower than pretransfusion levels. The remaining three (19%) patients (2 receiving 0.75 mg/kg sotatercept, 1 receiving 0.3 mg/kg sotatercept) experienced an increase in transfusion frequency, requiring between 0.71.0 additional transfusions per 24 weeks, representing an increase of 6.7-17.2% from baseline. Based on the observed responses, the active starting dose of sotatercept in patients with TDT was ≥0.5 mg/kg.

Of 16 patients with TDT treated with sotatercept, ten (63%) achieved a transfusion burden reduction of ≥20% sustained for ≥24 weeks; seven patients (44%) achieved a reduction of ≥33%, and two patients (13%), a reduction of ≥50% (Figure 3). Five of the ten patients achieved transfusion burden reduction while maintaining stable or improved hemoglobin levels at 12 weeks. The mean change in hemoglobin level from baseline to the end of treatment was 0.7 g/dL in all patients with TDT. Nine patients (56%) achieved a reduction in transfusion frequency over 24 weeks, receiving fewer transfusions per 24 weeks during treatment, while maintaining stable or improved hemoglobin levels versus pretransfusion levels. Four patients (25%) also achieved a reduction in transfusion frequency over 24 weeks; however, their hemoglobin

For NTDT patients, indirect bilirubin and total bilirubin decreased by 5-25% for patients receiving 0.3, 0.75, and 1.0 mg/kg (Online Supplementary Figure S2A). Among TDT patients with available bilirubin data at baseline, indirect bilirubin and total bilirubin increased by between 15% and 80% during treatment versus baseline at all dose levels (Online Supplementary Figure S2B). However, the changes were not statistically significant, and there was wide variability in the level of change.

A

B

Bilirubin

Figure 1. Response to sotatercept treatment in patients with non-transfusion-dependent β-thalassemia. (A) Percentage of sotatercept-treated non-transfusiondependent b-thalassemia patients achieving mean hemoglobin level increases from baseline of ≥1.0 g/dL and ≥1.5 g/dL sustained for ≥12 weeks by assigned dose group. (B) Average maximum increase in hemoglobin levels within 12 weeks versus baseline levels in responders versus non-responders by dose group. Responders were those patients achieving a ≥1.0 g/dL increase in hemoglobin levels sustained for ≥12 weeks. Error bars show the standard deviation of the mean. Hb: hemoglobin.

Figure 2. Mean change in hemoglobin levels from baseline up to day 400 in patients with non-transfusion-dependent b-thalassemia treated with different doses of sotatercept. Data are presented by assigned dose level, prior to intrapatient dose escalation. Hb: hemoglobin.

480

haematologica | 2019; 104(3)


Sotatercept TGF-b ligand trap in β-thalassemia

Clinical response

Red blood cell morphology

Extramedullary masses

Changes to RBC morphology were recorded for all patients during the study. Representative images of the changes observed are provided in Online Supplementary Figure S5. Reductions in hypochromia, anisocytosis, and poikilocytosis, and the proportion of target cells were reported, and were associated with increased hemoglobin levels.

The effect of sotatercept on extramedullary masses, as measured by magnetic resonance imaging, was reported for four patients with NTDT and one with TDT. The TDT patient experienced a reduction in extramedullary mass volume at 12 months after treatment with sotatercept 0.75 mg/kg (data not shown); however, due to an unequal distribution of the mass, it was not possible to determine the exact volume of the mass. One NTDT patient experienced reduction in extramedullary mass volume at 12 months after treatment with sotatercept 0.3 mg/kg, with the reduction ranging from 0.4% to 59.3% across three separate extramedullary masses (Online Supplementary Figure S3A), alongside a stable increase in hemoglobin level. Another NTDT patient treated with sotatercept 1.0 mg/kg experienced volume increases of 67% and 55% in two extramedullary masses, at 7 months after treatment (Online Supplementary Figure S3B). This correlated with an increase in hemoglobin levels (6 g/dL at baseline versus 8 g/dL at 7 months after treatment). Two NTDT patients experienced changes in volumes of extramedullary masses of between −6.3% and +22.5% at 11 months after treatment with sotatercept 1.0 mg/kg (data not shown).

Leg ulcers Seven NTDT patients had a history of leg ulcers at baseline; the effect of sotatercept on chronic leg ulcers was reported for one NTDT patient. The chronic leg ulcers improved, and the dischromic area was reduced following treatment with sotatercept 0.5 mg/kg for 6 months (Online Supplementary Figure S4).

Serum ferritin Among NTDT patients who responded to sotatercept treatment, mean levels of serum ferritin decreased regardless of iron chelation therapy status (Online Supplementary Figure S6A,B). In contrast to NTDT patients, all TDT patients received iron chelation therapy. Levels of serum ferritin decreased over time in TDT patients (Online Supplementary Figure S6C).

Safety Twenty-five of 46 patients (54%) experienced one or more treatment-related adverse events; the most common treatment-related adverse events in all patients were bone pain [26% (n=12)], arthralgia [15% (n=7)], back pain [11% (n=5)], asthenia/fatigue [11% (n=5)], and hypertension [11% (n=5)] (Table 2). Treatment-related adverse events led to discontinuation of sotatercept in eight patients (17%) (Table 2); of these eight patients, seven were not dependent on transfusions and one was transfusion-dependent. All 46 patients experienced one or more treatment-emergent adverse events (Online Supplementary Table S1); six (13%) experienced one or more serious adverse events (Online Supplementary Table S2). Nine patients (20%) experienced one or more grade 3-4 treatment-emergent adverse

Table 2. Incidence of treatment-related adverse events of any grade occurring in sotatercept-treated patients in any dose cohort.

Adverse events, n (%) 0.1 mg/kg (n=8)

Patients with ≥1 treatment-emergent AE* 8 (100) Patients with ≥1 treatment-related AE* 2 (25) Bone pain 2 (25) Arthralgia 0 Back pain 0 Asthenia/fatigue 0 Hypertension 0 Pain in extremity 0 Patients with treatment-emergent 2 (25) AE leading to discontinuation† Treatment-emergent AE leading to discontinuation‡ Hypertension 0 Superficial thrombophlebitis 1 (13) Injection site erythema 0 Pyrexia 0 Extramedullary hematopoiesis 0 Ventricular extrasystoles 0 Hypersensitivity 0 Bone pain 1 (13)

Sotatercept dose (data presented prior to intrapatient dose escalation) 0.3 mg/kg 0.5 mg/kg 0.75 mg/kg (n=9) (n=8) (n=12)

Overall (data presented post-intrapatient dose 1.0 mg/kg escalation) (n=9) (N=46)

9 (100) 2 (22) 2 (22) 0 0 0 0 0 0

8 (100) 4 (50) 2 (25) 1 (13) 1 (13) 1 (13) 1 (13) 1 (13) 1 (13)

12 (100) 8 (67) 1 (8) 3 (25) 2 (17) 3 (25) 2 (17) 2 (17) 3 (25)

9 (100) 6 (67) 4 (44) 1 (11) 2 (22) 1 (11) 1 (11) 0 2 (22)

46 (100) 25 (54) 12 (26) 7 (15) 5 (11) 5 (11) 5 (11) 3 (7) 8 (17)

0 0 0 0 0 0 0 0

0 0 0 0 0 1 (13) 0 0

2 (17) 0 1 (8) 0 0 0 1 (8) 0

0 0 0 1 (11) 1 (11) 0 0 0

2 (4) 1 (2) 1 (2) 1 (2) 1 (2) 1 (2) 1 (2) 1 (2)

Adverse events (AE) in each dose cohort are presented prior to intrapatient dose escalation. Total AE are presented after-intrapatient dose escalation. *AE occurring in ≥5% of sotatercept-treated patients. †AE occurring in any sotatercept-treated patient. ‡One patient had both injection site erythema and hypersensitivity as AE leading to treatment discontinuation.

haematologica | 2019; 104(3)

481


M.D. Cappellini et al.

events; hypertension and anemia were the most frequent [4% (n=2)] (Online Supplementary Table S3). The incidences of anemia were not related to treatment with sotatercept and anemia resolved within one treatment cycle. The most common treatment-emergent adverse events of any grade among the 30 NTDT patients were headache [53% (n=16)], arthralgia [47% (n=14)], cough [40% (n=12)], and asthenia/fatigue [37% (n=11)] (Online Supplementary Table S4). The most common treatment-emergent adverse events among the 16 TDT patients were bone pain [63% (n=10)], back pain (56% (n=9)], asthenia/fatigue [56% (n=9)]), headache [44% (n=7)], and arthralgia [44% (n=7]) (Online Supplementary Table S4). Most treatment-emergent adverse events were mild and did not lead to treatment discontinuation (Table 2). Bone pain was reported equally among the different sotatercept dose cohorts; the mean time to first onset of bone pain was similar between dose cohorts and the duration of bone pain was short (mean 12.0 days). Asthenia was less frequent at higher dose levels; 13% (3 incidences/23 doses received) versus 5% (4 incidences/81 doses received) for TDT patients and 36% (47 incidences/132 doses received) versus 5% (3 incidences/67 doses received) for NTDT patients at 0.1 mg/kg and 1.0 mg/kg, respectively) (data not shown). Observed differences between baseline and on-treatment laboratory values for liver and kidney function, including alanine aminotransferase, aspartate aminotransferase, and serum creatinine levels, were within the ranges expected for this population and were not related to treatment (data not shown). Dose delays for safety reasons were reported in eight patients, all with NTDT.

Discussion b-thalassemia is characterized by chronic anemia due to ineffective erythropoiesis and peripheral hemolysis. The current treatment for TDT is demanding, requiring regular blood transfusions and lifelong iron chelation therapy. No standardized treatment for NTDT is available, and no

treatment is currently available to improve ineffective erythropoiesis. Previous investigational treatments for NTDT have included HbF modulators such as butyrates, azacitidine, or hydroxyurea, with or without erythropoietin; however, results with these and other modulators of HbF have been inconsistent.24,25 Sotatercept, a recombinant fusion protein that acts as a ligand trap for TGF-b superfamily ligands, is hypothesized to improve late-stage erythropoiesis by reducing proliferation of early erythroid progenitors and precursors while increasing differentiation and maturation of late-stage RBC precursors.20,21 This drug may therefore present a novel approach to restoring effective erythropoiesis in bthalassemia. In this phase II, open-label, dose-finding study, sotatercept exhibited a good safety profile, and was tolerated by most patients. The most frequent adverse events were bone, articular or back pain, and asthenia/fatigue. Treatment discontinuation due to adverse events was rare, and the incidence of grade 3-4 adverse events was low. Changes in laboratory values for liver and kidney function were not thought to be treatment-related and were in line with fluctuations seen in b-thalassemia patients without treatment. The active dose of sotatercept was ≥0.3 mg/kg for NTDT patients and ≥0.5 mg/kg for TDT patients. Importantly, concomitant administration of hydroxyurea did not appear to interfere with the response to sotatercept, or with treatment compliance. The majority (75%) of NTDT patients treated with higher doses (0.75-1.0 mg/kg) of sotatercept achieved sustained increases in hemoglobin of ≥1.0 g/dL. Similarly, 66% of TDT patients treated with higher doses of sotatercept (0.75-1.0 mg/kg) achieved reductions of ≥33% in RBC transfusion requirements. The increase in hemoglobin and reduction in RBC transfusion correlated with increased serum exposure to sotatercept (data not shown), although responses were not proportional to sotatercept dose because of interpatient variability in serum drug exposure. The small number of patients in each dose group makes comparisons between sotatercept levels difficult. Studies are ongoing to identify any differences in baseline characteristics between responders and nonresponders.

Figure 3. Percentages of transfusion-dependent b-thalassemia patients treated with different doses of sotatercept who achieved a reduction in transfusion burden of ≥20%, ≥33%, and ≥50% sustained for ≥24 weeks.

482

haematologica | 2019; 104(3)


Sotatercept TGF-b ligand trap in β-thalassemia Ineffective erythropoiesis in b-thalassemia is associated with increased iron absorption, and patients with TDT often require regular RBC transfusions, further increasing the risk of iron overload. Reducing ineffective erythropoiesis and transfusion burden will decrease the rate of iron loading and associated complications such as heart, liver, and endocrine disorders. In this study, decreases in serum ferritin levels were observed in sotatercept-treated patients with NTDT and TDT, with the decreases occurring in a dose-dependent manner, regardless of concurrent treatment with iron chelation therapy. Although these results suggest that sotatercept may reduce iron absorption by reducing ineffective erythropoiesis, other mechanisms such as removal or redistribution of iron from overloaded organs may also contribute to reduced iron levels. Further studies will be required to elucidate the mechanism of action; however, a reduction in iron overload with long-term use may translate into improved outcomes for patients. Sotatercept treatment was associated with a reduction in the volume of extramedullary masses, as measured by magnetic resonance imaging, in two patients with recorded data – one with NTDT and one with TDT. However, this change in extramedullary mass volume was not observed for all patients, and, due to technical difficulties, precise estimation of extramedullary mass volume was not always possible. Reductions in the volumes of these masses did not appear to correlate with response. Further study is needed to determine the effect of sotatercept on extramedullary masses. In rodent models, sotatercept acts on ineffective erythropoiesis to reduce production of α-globin aggregates and apoptosis of late erythroid precursors, thereby increasing the efficiency of RBC production and hemoglobin levels.21,26 Although hemoglobin levels increased in NTDT patients during the study, no significant differences in HbF or reticulocyte count were reported. In a small subset of patients, normalization of RBC morphology was reported and associated with increasing hemoglobin levels over time. These data suggest that sotatercept may increase the lifespan of RBC in part by improving reticulocyte quality. These changes are consistent with the mode of action of sotatercept as a ligand trap for TGF-b superfamily members, including GDF-11. Binding of sotatercept to GDF-11 inhibits SMAD2/3 signaling, reducing TGF-b superfamily ligand signaling and thereby promoting terminal differentiation of erythroblasts.20 There are some limitations to this study, notably the small number of patients enrolled, which limited the abil-

References 1. Galanello R, Origa R. Beta-thalassemia. Orphanet J Rare Dis. 2010;5:11. 2. Malamos B, Belcher EH, Gyftaki E, Binopoulos D. Simultaneous studies with Fe59 and Cr51 in congenital haemolytic anaemias. Nucl Med (Stuttg). 1961;2:1-20. 3. Weatherall DJ, Clegg JB. The Thalassemia Syndromes. 2nd ed. Oxford, England: Blackwell Scientific Publications; 1972. 4. Arlet JB, Dussiot M, Moura IC, Hermine O,

haematologica | 2019; 104(3)

ity to draw comparisons between different sotatercept dose groups. The inclusion of both NTDT and TDT bthalassemia patients also resulted in patients being grouped into smaller subgroups that further limited the scope of the study, especially as intrapatient dose escalation was allowed. The short duration of follow-up may be another limitation, and longer follow-up would provide information on the long-term clinical efficacy and safety of sotatercept. This study demonstrated the safety and efficacy of sotatercept in patients with b-thalassemia. An improvement in hemoglobin levels and reduction in transfusion burden with sotatercept treatment have also been demonstrated in a phase II study of anemia patients with lower-risk myelodysplastic syndromes.23 This suggests that sotatercept may work to decrease ineffective erythropoiesis in multiple disease states via a single underlying mode of action. Preliminary data with sotatercept led to the initiation of a similar phase II study in b-thalassemia of the related recombinant fusion protein, luspatercept. Luspatercept has more selective activity on GDF11, and is also safe and effective in the treatment of b-thalassemia (Piga A, et al. manuscript submitted) and myelodysplastic syndromes.27 Luspatercept comprises the modified extracellular domain of human ActRIIB linked to the human IgG Fc domain,26,28 and has a similar mode of action to sotatercept but does not bind to other members of the TGF-b superfamily, such as activin A.20,26 A double-blind, randomized, placebo-controlled phase III trial of luspatercept in patients who require regular RBC transfusions due to b-thalassemia has recently completed recruitment (BELIEVE; NCT02604433). Luspatercept is also being studied in a phase III trial of patients with very low-, low- and intermediate-risk myelodysplastic syndromes (MEDALIST; NCT02631070). While the decision was made not to advance trials of sotatercept in b-thalassemia due to binding of sotatercept to activin A, sotatercept represents the first drug developed in its class, functioning as a TGF-b superfamily inhibitor to correct ineffective erythropoiesis. TGF-b superfamily inhibition may provide an alternative or complementary treatment option for patients with b-thalassemia. Acknowledgments The authors would like to thank all the patients and their families who participated in this study. The authors received editorial and writing support from Victoria Edwards, PhD, from Excerpta Medica, funded by Celgene Corporation. Celgene Corporation and Acceleron Pharma provided funding for this study.

Courtois G. Novel players in b-thalassemia dyserythropoiesis and new therapeutic strategies. Curr Opin Hematol. 2016;23(3): 181-188. 5. Rund D, Rachmilewitz E. Beta-thalassemia. N Engl J Med. 2005;353(11):1135-1146. 6. Melchiori L, Gardenghi S, Rivella S. Betathalassemia: HiJAKing ineffective erythropoiesis and iron overload. Adv Hematol, 2010;2010:938640. 7. Fibach E, Rachmilewitz EA. Does erythropoietin have a role in the treatment of b-

hemoglobinopathies? Hematol Oncol Clin North Am. 2014;28(2):249-263. 8. Gardenghi S, Grady RW, Rivella S. Anemia, ineffective erythropoiesis, and hepcidin: interacting factors in abnormal iron metabolism leading to iron overload in b-thalassemia. Hematol Oncol Clin North Am. 2010;24(6):1089-1107. 9. Cappellini MD, Cohen A, Porter J, Taher A, Viprakasit V. Guidelines for the Management of TD Thalassemia (TDT). 3rd ed. Nicosia, Cyprus: Thalassemia International

483


M.D. Cappellini et al. Federation; 2014. 10. Rund D. Thalassemia 2016: modern medicine battles an ancient disease. Am J Hematol. 2016;91(1):15-21. 11. Haghpanah S, Nasirabadi S, Ghaffarpasand F, et al. Quality of life among Iranian patients with beta-thalassemia major using the SF-36 questionnaire. Sao Paulo Med J. 2013;131(3):166-172. 12. Dahlui M, Hishamshah MI, Rahman AJ, Aljunid SM. Quality of life in transfusiondependent thalassemia patients on desferrioxamine treatment. Singapore Med J. 2009;50(8):794-799. 13. Baronciani D, Angelucci E, Potschger U, et al. Hematopoietic stem cell transplantation in thalassemia: a report from the European Society for Blood and Bone Marrow Transplantation Hemoglobinopathy Registry, 2000â&#x20AC;&#x201C;2010. Bone Marrow Transplant. 2016;51(4):536-541. 14. Qari MH, Wali Y, Albagshi MH, et al. Regional consensus opinion for the management of beta thalassemia major in the Arabian Gulf area. Orphanet J Rare Dis. 2013;8:143. 15. Angelucci E, Matthes-Martin S, Baronciani D, et al. Hematopoietic stem cell transplantation in thalassemia major and sickle cell disease: indications and management recommendations from an international expert panel. Haematologica. 2014;99(5): 811-820.

484

16. Negre O, Eggimann AV, Beuzard Y, et al. Gene therapy of the b-hemoglobinopathies by lentiviral transfer of the b(A(T87Q))-globin gene. Hum Gene Ther. 2016;27(2):148165. 17. Saliba AN, Taher AT. Morbidities in nontransfusion-dependent thalassemia. Ann N Y Acad Sci. 2016;1368(1):82-94. 18. Taher A, Vichinsky E, Musallam K, Cappellini MD, Viprakasit V. Blood transfusion. In: Weatherall D, ed. Guidelines for the Management of Non Transfusion Dependent Thalassaemia (NTDT). Nicosia, Cyprus: Thalassaemia International Federation; 2013. 19. Rachmilewitz EA, Giardina PJ. How I treat thalassemia. Blood. 2011;118(13):34793488. 20. Carrancio S, Markovics J, Wong P, et al. An activin receptor IIA ligand trap promotes erythropoiesis resulting in a rapid induction of red blood cells and haemoglobin. Br J Haematol. 2014;165(6):870-882. 21. Dussiot M, Maciel TT, Fricot A, et al. An activin receptor IIA ligand trap corrects ineffective erythropoiesis in b-thalassemia. Nat Med. 2014;20(4):398-407. 22. Ruckle J, Jacobs M, Kramer W, et al. Singledose, randomized, double-blind, placebocontrolled study of ACE-011 (ActRIIAIgG1) in postmenopausal women. J Bone Miner Res. 2009;24(4):744-752.

23. Komrokji R, Garcia-Manero G, Ades L, et al. Sotatercept with long-term extension for the treatment of anemia in patients with lower-risk myelodysplastic syndromes: a phase 2, dose-ranging trial. Lancet Haematol. 2018;5(2):e63-e72. 24. Porter J, Garbowski M. Novel erythropoiesis stimulating agents in thalassemia. European Hematology Association Education Programme. 2015;9:311-320. 25. Taher AT, Musallam KM, Karimi M, et al. Overview on practices in thalassemia intermedia management aiming for lowering complication rates across a region of endemicity: the OPTIMAL CARE study. Blood. 2010;115(10):1886-1892. 26. Suragani RN, Cadena SM, Cawley SM, et al. Transforming growth factor-b superfamily ligand trap ACE-536 corrects anemia by promoting late-stage erythropoiesis. Nat Med. 2014;20(4):408-414. 27. Platzbecker U, Germing U, GĂśtze KS, et al. Luspatercept for the treatment of anaemia in patients with lower-risk myelodysplastic syndromes (PACE-MDS): a multicentre, open-label, phase 2 dose-finding study with long-term extension study. Lancet Oncol. 2017;18(10):1338-1347. 28. Attie KM, Allison MJ, McClure T, et al. A phase 1 study of ACE-536, a regulator of erythroid differentiation, in healthy volunteers. Am J Hematol. 2014;89(7):766-770.

haematologica | 2019; 104(3)


ARTICLE

Myelodysplastic Syndromes

The high NRF2 expression confers chemotherapy resistance partly through up-regulated DUSP1 in myelodysplastic syndromes

Peipei Lin,1,2,3,4* Yanling Ren,1,2,3* Xiaomei Yan,4 Yingwan Luo,1,2,3 Hua Zhang,1,2,3 Meenu Kesarwani,4 Jiachen Bu,4 Di Zhan,4 Yile Zhou,1,2,4 Yuting Tang,4 Shuanghong Zhu,1,2,3 Weilai Xu,1,2,3 Xinping Zhou,1,2,3 Chen Mei,1,2,3 Liya Ma,1,2,3 Li Ye,1,2,3 Chao Hu,1,2 Mohammad Azam,4 Wei Ding,5 Jie Jin,1,2 Gang Huang4# and Hongyan Tong1,2,3#

Department of Hematology, the First Affiliated Hospital, College of Medicine, Zhejiang University, Hangzhou, China; 2Institute of Hematology, Zhejiang University, Hangzhou, China; 3Myelodysplastic Syndromes Diagnosis and Therapy Center, the First Affiliated Hospital, College of Medicine, Zhejiang University, Hangzhou, China; 4Divisions of Pathology and Experimental Hematology and Cancer Biology, Cincinnati Children's Hospital Medical Center, OH, USA and 5Department of Pathology, the First Affiliated Hospital, College of Medicine, Zhejiang University, Hangzhou, China 1

Ferrata Storti Foundation

Haematologica 2019 Volume 104(3):485-496

*PL and YR contributed equally to this work. #HYT and GH contributed equally to this study as joint senior authors.

ABSTRACT

A

lthough cytarabine has been widely considered as one of the chemotherapy drugs for high-risk myelodysplastic syndromes (MDS), the overall response rate is only approximately 20-30%. Nuclear factor erythroid 2-related factor 2 (NRF2, also called NFE2L2) has been shown to play a pivotal role in preventing cancer cells from being affected by chemotherapy. However, it is not yet known whether NRF2 can be used as a prognostic biomarker in MDS, or whether elevated NRF2 levels are associated with cytarabine resistance. Here, we found that NRF2 expression levels in bone marrow from high-risk patients exceeded that of low-risk MDS patients. Importantly, high NRF2 levels are correlated with inferior outcomes in MDS patients (n=137). Downregulation of NRF2 by the inhibitor Luteolin, or lentiviral shRNA knockdown, enhanced the chemotherapeutic efficacy of cytarabine, while MDS cells treated by NRF2 agonist Sulforaphane showed increased resistance to cytarabine. More importantly, pharmacological inhibition of NRF2 could sensitize primary high-risk MDS cells to cytarabine treatment. Mechanistically, downregulation of dual specificity protein phosphatase 1, an NRF2 direct target gene, could abrogate cytarabine resistance in NRF2 elevated MDS cells. Silencing NRF2 or dual specificity protein phosphatase 1 also significantly sensitized cytarabine treatment and inhibited tumors in MDS cells transplanted mouse models in vivo. Our study suggests that targeting NRF2 in combination with conventional chemotherapy could pave the way for future therapy for highrisk MDS.

Introduction Myelodysplastic syndromes (MDS) are a heterogeneous disease of clonal hematopoietic stem cell neoplasms characterized by ineffective hematopoiesis, cytopenia, dysplasia of the myeloid cells, and an inherent risk of progression to acute myeloid leukemia (AML).1 There are several prognostic scores, and the International Prognostic Scoring System (IPSS) and its revised vision (IPSS-R) are commonly used to stratify patients into two risk groups, defining lower and higher risk patients.2,3 Higher risk MDS patients have an increased risk of developing AML and are associated with poor clinical outcomes. Treatment strategies were made haematologica | 2019; 104(3)

Correspondence: HONGYAN TONG tonghongyan@zju.edu.cn GANG HUANG Gang.Huang@cchmc.org Received: May 15, 2018. Accepted: September 26, 2018. Pre-published: September 27, 2018. doi:10.3324/haematol.2018.197749 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/104/3/485 Š2019 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

485


P. Lin et al.

according to the risk categories of MDS. Cytarabine (AraC) is a pyrimidine nucleoside analog that interferes with the synthesis of DNA when the cycle holds in the S phase. Over the last decades, Ara-C-based therapies have been widely used to manage MDS patients, especially those at higher risk.4 However, the overall response rate of single Ara-C treatment was only approximately 20-30%.5,6 Clinically, MDS patients who remained unresponsive to the routine treatment of Ara-C (100mg/m2) were defined as Ara-C-resistant MDS patients. Nuclear factor erythroid 2-related factor 2 (NRF2, also called NFE2L2) is a transcription factor that protects cells from oxidative damage.7,8 Under oxidative stress, NRF2 is released from its cytosolic inhibitor Kelch-like ECH-associated protein 1 (KEAP1) and translocates to the nucleus.9 It has recently been shown that NRF2 underlies drug resistance in acute myeloid leukemia (AML), chronic myeloid leukemia (CML), and chronic lymphocytic leukemia (CLL).10-12 NRF2 binding to antioxidant responsive element (ARE) allows induction of a number of cytoprotective and detoxification genes, such as NAD(P)H: quinone oxidoreductase 1 (NQO1), heme oxygenase-1 (HO-1), and glutamate-cysteine ligase (GCL).7,13,14 Few studies have shown the mechanisms of NRF2 in drug resistance. NRF2 target genes, such as HO-1, have been reported to facilitate resistance of tumor cells to chemotherapy in AML cells.15,16 Here, we aimed to correlate NRF2 expression and its clinical outcome in a large cohort of MDS patients (n=137). We also performed in vitro and in vivo experiments to validate our findings regarding the function of NRF2 in chemo-resistance in MDS. We found that NRF2 expressions were elevated in higher risk MDS and correlated with inferior clinical outcomes. High levels of NRF2 reduced MDS cell sensitivity to Ara-C treatment partly through its direct target gene DUSP1.

Methods Immunohistochemistry Immunohistochemistry (IHC) was performed on 4 µm thick bone marrow (BM) sections. BM samples were stained for NRF2 expression (dilution 1:200; Abcam, UK) or DUSP1 expression (dilution 1:100; Abcam, UK). Samples were incubated using primary antibody for 30 minutes (min) at 37°C. Secondary antibody (dilution 1:50; Dako, Denmark) was applied for 15 min. Binding was visualized by the horseradish peroxidase (HRP)/3,3’diaminobenzidine (DAB) kit (ZSGB-BIO, China). Staining results were semi-quantified using an arbitrary score as follows: No staining, 0; Pale yellow, 1; Tan, 2; Brown, 3; Nuclear staining in 0-25% of cells, 0; Nuclear staining in 25-50% of cells, 1; Nuclear staining in 50-75% of cells, 3; Nuclear staining in 75-100% of cells, 4. The stain color score multiplied by the nuclear staining proportion score is the final IHC score.

In vitro cytotoxicity assay Myelodysplastic syndrome cell lines (5×105/mL) and primary MDS cells (1×106/mL) were seeded in 96-well flat bottom plates and treated with increasing concentrations of Ara-C. Cell proliferation was determined using the MTS proliferation assay. 20 ml of MTS (Promega, USA) was added to 100 ml of cell suspension, and cells were further incubated in 5% CO2 for 3-4 hours at 37 °C. The plates were then analyzed on an enzyme immunoassay plate reader at 490 nm. The half inhibitory concentration (IC50) values 486

of Ara-C were calculated by Prism Graphpad software. All experiments were performed in triplicate.

Mice models NOD/SCID-IL2Rγnull-SGM3 (NSGS) mice were bred and maintained in Cincinnati Children’s Hospital Medical Center (CCHMC).17 Mice were randomized into six groups. NRF2 shRNA SKM-1 (transfected with shRNA targeting NRF2), DUSP1 shRNA SKM-1 (transfected with shRNA targeting DUSP1), and scramble shRNA SKM-1 were resuspended in 300 ml phosphate buffer saline (PBS) and then injected intravenously into the non-irradiated mice (1 million cells per mouse). Ten days after cell inoculation (Day 0), the mice received 50 milligram/ kilogram (mg/kg) of AraC or PBS once a day for five consecutive days (Day 10-15). Ara-C and PBS were injected intraperitoneally. All experiments were performed in accordance with protocols approved by the Institutional Review Board of CCHMC.

Statistical analysis Data were analyzed using SPSS 16.0 and GraphPad Prism 6. Statistical analyses were performed using Student t-test or one-/two-way ANOVA with multiple comparisons correction. P<0.05 was considered statistically significant.

Results NRF2 is elevated in higher risk MDS and correlates with inferior overall survival To explore NRF2 expression and its clinical outcome in MDS, we performed immunohistochemistry (IHC) on 137 MDS patients and 17 controls (Figure 1A). IHC staining results showed that NRF2 was over-expressed in the BM biopsies from MDS patients (P<0.01) (Figure 1B). The NRF2 levels of intermediate-, high-, and very high-risk IPSS-R patients exceeded that of low-risk IPSS-R patients (P=0.004) (Online Supplementary Table S1). To validate the expression of NRF2 with its downstream target signature in MDS, we analyzed published gene expression profiles of CD34+ BM cells purified from samples obtained from 183 MDS patients and 17 healthy controls.18 We performed gene set enrichment analysis (GSEA) to explore the downstream targets’ signature of NRF2 in this cohort.19 Although the expression of NRF2 is not significantly activated in MDS patients compared to the healthy controls (P=0.225) (Online Supplementary Figure S1A), NRF2 expression is significantly enriched in higher risk MDS patients (MDS-EB-1/2) when compared to those with lower risk MDS (MDS-SLD/MDS-RS) (P=0.020) (Figure 1C). Leading edge genes are shown in Online Supplementary Appendix Lists 1 and 2. It is worth noting that MDS patients with higher NRF2 expression levels (IHC scores, 4-6) displayed worse overall survival (OS) than patients with lower NRF2 levels (IHC scores, 0-3) (median, 391 vs. 554 days; P=0.011) (Figure 1D). We further performed CD34 and NRF2 double staining by immunofluorescence with MDS patient BM aspiration samples (Online Supplementary Figure S1B). CD34 and NRF2 double staining in particular cells demonstrated that NRF2 was also expressed at protein levels in CD34+ cells (Online Supplementary Figure S1C). High NRF2 expression levels were closely associated with higher risk according to the 2016 WHO subtype (P=0.022), IPSS cytogenetics (P=0.001), and IPSS categories (P=0.001). There were no significant differences in other clinical features between haematologica | 2019; 104(3)


NRF2 and Ara-C resistance in MDS

MDS patients with higher and lower NRF2 levels (Online Supplementary Table S2).

Pharmacological modulations of NRF2 regulate chemotherapeutic efficacy of Ara-C in MDS cells The effect of NRF2 on Ara-C resistance was first evaluated in primary MDS cells. After 72 h exposure to increasing doses of Ara-C, proliferation of primary MDS cells with 2 mM NRF2 inhibitor Luteolin was significantly reduced compared to vehicle control-treated MDS cells (# 1 MDS-EB-2 IC50: 5.7 mM vs. 4.8 mM, P=0.04; # 2 MDSEB-2 IC50: 3.2 mM vs. 1.9 mM, P=0.006; # 3 MDS-EB-2 IC50: 7.5 mM vs. 6.9 mM; P=0.03) (Figure 2A). To further identify the function of NRF2 in MDS, we examined the pharmacological effects of NRF2 inhibitor and activator in MDS-patient-derived SKM-1 and murine MDS model cells MLLPTD/WT/RUNX1-S291fs cells.20 The highest dose of Ara-C resulted in approximately 90% inhibition of SKM-1 at 72 h (IC50, 1.72 mM) and MLLPTD/WT/RUNX1-S291fs cells at 48 h (IC50, 0.17 mM) (Online Supplementary Figure S2A and B). In agreement with previous reports on human lung carcinoma and col-

orectal cancer cell lines,21,22 we found that the NRF2 inhibitor Luteolin (3, 4, 5, 7-tetrahydroxy flavone) suppressed the protein expression of NRF2 in SKM-1 (Figure 2B). Sulforaphane (SFN) has been shown to be a potent NRF2 activator.23 SFN treatments in SKM-1 cells increased the protein expression of NRF2 (Figure 2C). NRF2 mRNA levels in MDS cells treated with the NRF2 inhibitor or agonist were measured. There was little change at mRNA levels, but obvious changes of NRF2 were seen at protein levels (Online Supplementary Figure S2C-F). Lower doses of Luteolin treatment had little effect on cell proliferation (Online Supplementary Figure S3A), but significantly enhanced the cytotoxicity of Ara-C (0-4 mM) to SKM-1. The IC50 values of Ara-C in SKM-1 cells were 1.41 mM and 0.93 mM with 5 mM and 10 mM Luteolin treatment, respectively (P<0.001) (Figure 2D). Similar results were also found in MLLPTD/WT/RUNX1-S291fs cells (Online Supplementary Figure S3B-D). The IC50 was reduced from 0.17 mM to 0.11 mM by 1 mM Luteolin in MLLPTD/WT/RUNX1-S291fs (P=0.007) (Online Supplementary Figure S3D). 1mM SFN treatments had little effect on the proliferation of SKM-1 cells (Online Supplementary Figure

A

B

C

D Figure 1. Expression and clinical outcomes of NRF2 in myelodysplastic syndrome (MDS) patients. (A) NRF2 immunohistochemistry (IHC) staining of bone marrow biopsy samples (magnification Ă&#x2014;400). (B) MDS patients had higher NRF2 IHC scores compared to controls. (C) Gene set enrichment plot showed that NRF2 target genes were enriched in higher-risk MDS patients. (D) MDS patients with higher NRF2 levels displayed worse overall survival (OS). *P<0.05; **P<0.01; ***Pâ&#x2030;¤0.001. Int: intermediate; MDS-SLD: myelodysplastic syndrome single-lineage dysplasia; MDS-RS: myelodysplastic syndrome with ring sinderoblasts.

haematologica | 2019; 104(3)

487


P. Lin et al.

Re-sensitizing MDS cells to Ara-C treatment in vitro by knockdown of NRF2

S3E). Interestingly, SFN treatment can decrease the chemotherapeutic effect of Ara-C (0-40 mM) in SKM-1. The IC50 was raised from 1.72 mM to 5.73 mM in SKM-1 by 2.5 mM SFN treatment (P=0.008) (Figure 2E). A similar effect of SFN could also be found in MLLPTD/WT/RUNX1S291fs (Online Supplementary Figure S3F-H). 1 mM SFN treatment increased the Ara-C IC50 from 0.17 mM to 0.26 mM compared to vehicle treatment (P=0.001) (Online Supplementary Figure S3H).

Transduction of NRF2 shRNA plasmid in human and mouse MDS cell lines repressed NRF2 mRNA levels by approximately 40-60%, compared with scramble shRNA plasmid transduction (Online Supplementary Figure S4A and B). Immunoblot analysis revealed that NRF2 shRNA robustly reduced the expression of NRF2 protein (Online Supplementary Figure S4C and D). Knockdown of NRF2

A

B

C

D

E

F

G

Figure 2. NRF2 inhibitor and activator regulate the sensitivity of myelodysplastic syndrome (MDS) cells to cytarabine (Ara-C) treatment. (A) Ara-C IC50 was significantly decreased in primary MDS cells treated with the NRF2 inhibitor Luteolin. (B) Luteolin decreased NRF2 protein levels in SKM-1. (C) The NRF2 agonist Sulforaphane increased NRF2 protein levels in SKM-1. (D) Ara-C IC50 was significantly decreased by Luteolin in SKM-1. (E) Ara-C IC50 was significantly increased by Sulforaphane in SKM-1. (F) NRF2 silencing significantly decreased IC50 of Ara-C in SKM-1. (G) NRF2 shRNA enhanced apoptosis induced by Ara-C in SKM-1 cell lines. *P<0.05; **P<0.01; ***Pâ&#x2030;¤0.001. h: hours.

488

haematologica | 2019; 104(3)


NRF2 and Ara-C resistance in MDS

resulted in a significant reduction of Ara-C IC50 in SKM1 (72 h Ara-C IC50, 2.20 mM vs. 0.87 mM; P=0.001) (Figure 2F) and MLLPTD/WT/RUNX1-S291fs (48 h Ara-C IC50, 0.37 mM vs. 0.25 mM; P=0.049) (Online Supplementary Figure S4E). Knockdown of NRF2 enhanced apoptosis induced by Ara-C in MDS cell lines (Figure 2G and Online Supplementary Figure S4F). We also found that NRF2 silenced MDS cell lines after Ara-C treatment tended to be arrested in the S phase (Online Supplementary Figure S4G-J).

A

DUSP1 is an NRF2 direct target gene in MDS To further investigate the mechanisms involved in NRF2-mediated Ara-C resistance, we also analyzed published gene expression profiles of Ara-C-sensitive and AraC-resistant AML patient samples. Our analysis indicated that a group of NRF2 target genes might be responsible for Ara-C resistance in AML (P=0.032) (Figure 3A). Leading edge genes are shown in Online Supplementary Appendix List 3. After overlapping the up-regulated NRF2 target genes in high-risk MDS patients (total 132 genes) and Ara-

B

C

D

E

F

Figure 3. DUSP1 is an NRF2 target gene in myelodysplastic syndrome (MDS). (A) Gene set enrichment plot showed that NRF2 target genes were enriched in cytarabine (Ara-C)--resistant acute myeloid leukemia (AML) patients. (B) Overlap of up-regulated NRF2 target genes in higher-risk MDS patients and Ara-C-resistant AML patients. (C) The gene list of 37 overlapped genes. (D) ChIP sequence analysis of published data24 indicated the NRF2 binding site in the region of DUSP1 gene. (E) NRF2 binding sites in the regions of NQO1 and DUSP1 genes. TSS: transcription start site; TTS: transcription termination site. (F) NRF2 ChIP q-PCR analysis of SKM1 cells.

haematologica | 2019; 104(3)

489


P. Lin et al.

C resistant AML patients (total 153 genes), we found a list of common up-regulated NRF2 target genes (n=37) (Figure 3B). Interestingly, dual-specificity protein phosphatase 1 (DUSP1) was one of the genes up-regulated in both highrisk MDS patients and Ara-C-resistant AML patients (Figure 3C). We then performed NRF2 ChIP-seq analysis based on a published dataset from human lymphoblastoid cell lines.24 In cells treated with NRF2 agonist, ChIP analysis vaildated the NRF2 binding site in the region of DUSP1 gene loci (Figure 3D). The NRF2 binding regions proximal to NQO1 and DUSP1 genes contained a conserved NRF2

binding TGAnnnnGG motif, as previously reported (Figure 3E).25 ChIP q-PCR analysis revealed that the NRF2 binding signals in the NQO1 and DUSP1 genes were significantly higher than the negative control loci. Lower NRF2 signals were detected in SKM-1 with 5 mM NRF2 inhibitor treatment (48 h, P<0.01) (Figure 3F). Consistent with the mRNA expression of NRF2, mRNA expression of DUSP1 could also be inhibited by 2 mM NRF2 inhibitor Luteolin treatment in primary MDS cells (Figure 4A). Our q-PCR results confirmed that DUSP1 was an NRF2 direct target gene in SKM-1 and

A

B

C

D

E

490

Figure 4. NRF2 and DUSP1 expressions were elevated in higher-risk myelodysplastic syndrome (MDS) or cytarabine (Ara-C)--resistant MDS patients. (A) NRF2 and DUSP1 mRNA levels were both repressed by Luteolin in primary MDS cells. (B) DUSP1 immunohistochemistry (IHC) staining of bone marrow (BM) biopsy samples (magnification Ă&#x2014;400). (C) NRF2 and DUSP1 IHC scores in controls and MDS. (D) Immunoblotting analysis was conducted for NRF2 and DUSP1 protein levels in healthy controls, MDS cell lines, and primary MDS cells. (E) Elevations of NRF2 and DUSP1 were confirmed in the BM samples of Ara-C-resistant MDS by immunoblotting analysis. *P<0.05; **P<0.01.

haematologica | 2019; 104(3)


NRF2 and Ara-C resistance in MDS

MLLPTD/WT /RUNX1-S291fs cells (Online Supplementary Figure S5A-D).We also performed immunohistochemistry assay to detect DUSP1 expression in 11 controls and 26 MDS patients (Figure 4B). NRF2 and DUSP1 IHC scores were both significantly increased in the higher-risk MDS group (high-risk and very high-risk by IPSS-R) compared to the control group (P<0.05) (Figure 4C). To better detect the small differences at the protein levels, we further compared the levels of NRF2 and DUSP1 with the BM mononuclear cells from Ara-C sensitive and Ara-C-resistant MDS patients by immunoblotting (Figure 4D). Samples from MDS patients who were responsive to AraC treatment were selected as Ara-C-sensitive MDS samples. MDS patients in whom Ara-C treatment was seen to be ineffective were chosen as Ara-C-resistant cases. Elevated levels of NRF2 and DUSP1 were seen in the BM samples of Ara-C-resistant MDS patients by immunoblot-

A

C

ting analysis; this was statistically significant based on intensities (Figure 4E).

NRF2 mediates Ara-C resistance partly through its direct target gene DUSP1 A Dusp1 and Dusp6 inhibitor,26,27 was used to test the potential of targeting DUSP1 for Ara-C therapy. The combined inhibitory effect was predicted using the Bliss independent model.28,29 Our data indicate that BCI and Ara-C have statistically significant synergistic effects on NRF2 activated SKM-1 cells (%survival, predicted 70.1% vs. experimental 51.0%; P=0.005) and MLLPTD/WT/RUNX1S291fs cells (%survival, predicted 57.0% vs. experimental 41.0%; P=0.048) (Figure 5A and Online Supplementary Figure S5E). To further demonstrate that the therapeutic synergistic effect is due to the inhibition of DUSP1, we conducted DUSP1 shRNA on SKM-1 cells. DUSP1

B

D

E

Figure 5. NRF2 confers cytarabine (Ara-C) resistance partly through the activation of DUSP1 in myelodysplastic syndrome (MDS). (A) Ara-C and (E)-2-benzylidene3-(cyclohexylamino)-2,3-dihydro-1H-inden-1-one (BCI) have synergistic effects in NRF2 agonist-treated SKM-1 cells. (B) DUSP1 shRNA-1 sensitized SKM-1 cells to AraC treatment. (C) DUSP1 shRNA-1 re-sensitized NRF2 agonist treated SKM-1 cells to Ara-C treatment. (D) DUSP1 shRNA-1 enhanced apoptosis induced by Ara-C in SKM-1 cell lines. (E) DUSP1 shRNA-1 induced S phrase arrest in SKM-1. *P<0.05; **P<0.01; ***Pâ&#x2030;¤0.001.

haematologica | 2019; 104(3)

491


P. Lin et al.

shRNA-1 and Dusp1 shRNA-1 were chosen because the data indicated that both of them mediated the best knockdown efficiency in human and mouse MDS cell lines, respectively (Online Supplementary Figure S6A-D). Downregulation of DUSP1 by lentivirus shRNA could sensitize SKM-1 cells to Ara-C treatment (72 h Ara-C IC50, scramble shRNA 1.91 mM vs. DUSP1 shRNA-1 1.23 mM; P=0.002) (Figure 5B). Treatment of the NRF2 agonist SFN significantly mitigated Ara-C toxicity in scramble shRNA SKM-1 cells (72 h Ara-C IC50, control treatment 2.05 mM vs. SFN treatment 2.50 mM; P=0.044). Significantly, DUSP1 shRNA on NRF2 elevated SKM-1 cell lines abrogate Ara-C resistance (72 h Ara-C with SFN treatment IC50, scramble shRNA 2.50 mM vs. DUSP1 shRNA-1 1.35 mM; P=0.020) (Figure 5C). In MLLPTD/WT /RUNX1-S291fs cells, Dusp1 downregulation could sensi-

tize MDS mouse cells to Ara-C treatment (48 h Ara-C IC50, scramble shRNA 0.20 µM vs. Dusp1 shRNA-1 0.15 mM; P=0.001) (Online Supplementary Figure S6E). Consistent with the data in SKM-1 cells, Dusp1 shRNA could also re-sensitize Ara-C-resistant MDS mouse cells [48 h Ara-C IC50, scramble shRNA with control treatment 0.20 mM vs. scramble shRNA with SFN treatment 0.22 mM (P=0.025) scramble shRNA with SFN treatment 0.22 mM vs. Dusp1 shRNA-1 with SFN treatment 0.16 mM (P=0.001)] (Online Supplementary Figure S6F). Knockdown of DUSP1 led to an increase in the apoptotic rate of SKM1 treated with Ara-C compared to scramble shRNA SKM1 cells (Figure 5D). DUSP1 silenced SKM-1 cell lines with Ara-C treatment also tended to be arrested in the S phase (Figure 5E). These results indicated that DUSP1 was a downstream gene of NRF2 and partially responsible for

A

B

C

Figure 6. Knockdown of NRF2 significantly sensitizes myelodysplastic syndrome (MDS) cells to cytarabine (Ara-C) in vivo. (A) Flow cytometry showed appearance of human CD45+ cells in liver tumors of scramble, NRF2 and DUSP1 shRNA SKM-1 transplanted NSGS mice. (B) Liver tumors in scramble, NRF2 and DUSP1 shRNA SKM-1 transplanted NSGS mice with phosphate buffer saline (PBS) or Ara-C treatment. (C) Liver tumor volumes were significantly smaller in NRF2 or DUSP1 shRNA SKM-1 transplanted MDS mice treated with Ara-C compared with PBS. *P<0.05; **P<0.01; ***P≤0.001. g: grams.

492

haematologica | 2019; 104(3)


NRF2 and Ara-C resistance in MDS

NRF2-mediated Ara-C resistance. To determine whether silencing of NRF2 or DUSP1 compromised the reactive oxygen specis (ROS) levels in MDS cells, we analyzed ROS production by flow cytometry. No significant difference in ROS levels was observed between NRF2 or DUSP1 knockdown MDS cells and control cells (Online Supplementary Figure S7A-D). To further investigate other possible pathways involved in Ara-C resistance, we adjusted the fold change value to 1.1 and the rawp value to 0.05 so that we obtained two larger cohorts of up-regulated genes in high-risk MDS (total 5477 genes) or Ara-Cresistant AML (total 3074 genes). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis was performed on 331 overlapped common genes (Online Supplementary Figure S7E), which revealed potential involvement of genes in a number of pathways (Online Supplementary Figure S7F).

Re-sensitizing MDS cells to Ara-C treatment in vivo by knockdown of NRF2 or DUSP1 To determine the effect of NRF2 and DUSP1 on chemoresistance in vivo, we established xenograft mouse models through intravenously injecting NRF2 shRNA, DUSP1 shRNA, or scramble shRNA SKM-1 cells into NOD/SCID-IL2RÎłnull-SGM3 (NSGS) mice. The ratios of human CD45+ cells to mouse CD45+ cells were more than 80% in tumors (Figure 6A) but less than 5% in bone marrow (Online Supplementary Figure S8A) or peripheral blood (Online Supplementary Figure S8B). However, no significant differences were observed in the survival of transplanted NSGS mice (Online Supplementary Figure S8C-E). In scramble shRNA MDS mice, the tumor weight showed a trend to decrease, but did not reach a significant change in the Ara-C treatment group (1.91 g vs. 1.53 g with PBS vs. AraC treatment group, respectively; P=0.062) (Figure 6B and C). Treatment of NRF2 or DUSP1 silencing MDS mice

with Ara-C resulted in significantly smaller tumors in the liver (NRF2 silencing MDS mice, 2.30 g vs. 1.77 g with PBS vs. Ara-C treatment group, respectively, P=0.010; DUSP1 silencing MDS mice, 1.84 g vs. 1.41 g, respectively, P=0.001).

Discussion The present study was aimed to investigate the role of NRF2 in MDS and its molecular mechanism involved in chemoresistance, particularly in Ara-C-based therapy. Using IHC and unbiased analysis, we identified the predictive role of NRF2 in clinical outcomes amongst MDS patients. Based on our data and published evidence, we proposed a model of NRF2 in high-risk MDS with Ara-C treatment (Figure 7). Activation of NRF2-DUSP1 signaling and other pathways might lead to Ara-C resistance in high-risk MDS. Inhibition of NRF2 could re-sensitize MDS cells to Ara-C treatment. Garcia-Manero et al. had previously measured NRF2 mRNA levels in peripheral blood mononuclear cells of AML or MDS cases (n=31) and reported that higher mRNA levels of NRF2 were associated with longer survival.30 Here, we explored the prognostic impact of NRF2 in a larger cohort of MDS patients (n=137). Our IHC data indicated that higher risk MDS patients had higher NRF2 expression levels in BM samples compared to lower risk patients by IPSS-R (P=0.004). GSEA results of CD34+ BM cell gene expressions from published MDS patient cohort data (n=183) further confirmed NRF2 was elevated in higher-risk MDS patients (MDS-EB-1/2) compared to lower-risk patients (MDS-SLD/RS) (Figure 1C). We speculate that activated NF-ÎşB signaling may drive the overexpression of NRF2 in high-risk MDS.10,31 It has also been reported that mitochondrial dynamics could regulate neu-

Figure 7. A proposed model of NRF2 in higher-risk myelodysplastic syndrome (MDS) with cytarabine (Ara-C) treatment. (A) Ara-C treatment inhibits the cell viability of the MDS cells with low NRF2 levels. (B) NRF2 confers Ara-C resistance partly through its downstream target gene DUSP1 in MDS cells. (C) The inhibition of NRF2 re-sensitizes MDS cells to Ara-C treatment.

haematologica | 2019; 104(3)

493


P. Lin et al.

ral stem cell fate by modifying ROS signaling to activate NRF2-dependent pathways.32,33 Future studies may help to elucidate the possible mechanisms involved in high NRF2 expression levels in higher-risk MDS patients. Importantly, our survival analysis indicated that MDS patients with higher NRF2 levels in BM cells correlated with worse OS than patients with lower NRF2 levels (P=0.011) (Figure 1D). The discrepancy between our results and the previously published data may be related to the methods, cells studied (mononuclear cells of PB vs. BM), or sample size (31 vs. 137). This needs to be further investigated in larger independent cohorts. Ara-C is widely used as a treatment approach in higherrisk MDS patients; however, single Ara-C treatment has limited therapeutic effect. Drug resistance is the major cause of treatment failure. High NRF2 proteins in human primary AML cells have been shown to be driven by NFκB and knockdown of NRF2 reduced colony formation of AML cells in response to treatment of Ara-C and daunorubine.10 However, there are no research reports on the role of NRF2 in mediating drug resistance in MDS. The SKM1 cell line was established from a Japanese male patient in 1985 who was initially diagnosed as higher-risk MDS (MDS-EB-2).34 It has been reported that SKM-1 cells had a higher IC50 of Ara-C than other myeloid leukemia cell lines, indicating that SKM-1 cells were more Ara-C resistant.35,36 The incidences of MLL-PTD and RUNX1 mutations showed an increase in higher-risk MDS compared to lower-risk MDS.37 Thus, SKM-1 cells and MDS mouse model cell line RUNX1 mutant-transduced MllPTD/WT BM cells (MllPTD/WT/RUNX1-S291fs) were used in our study. BM mononuclear cells from Ara-C-sensitive and Ara-C-resistant MDS patients were also studied. Our results indicated that NRF2-mediated drug resistance in MDS was similar to other conditions.10,38 Luteolin is a potential NRF2 inhibitor that can promote the degradation of NRF2 mRNA. Our results revealed that NRF2 downregulation in primary MDS cells, by inhibitor Luteolin, decreased IC50 of Ara-C. As primary MDS cells were mostly composed of non-transformed cells, we also validated our results in human and mouse MDS cell lines. Downregulation of NRF2 by Luteolin could also enhance the chemotherapeutic efficacy of Ara-C in MDS cell lines. SFN is a well-known NRF2 agonist. SFN is an isothiocyanate that forms a KEAP1– Sulforaphane thionoacyl adduct to stabilize NRF2.39 We found that upregulation of NRF2, mediated by agonist SFN, induced resistance of MDS cells to Ara-C. Previous reports in AML indicated the protective role of NRF2 against apoptosis.15 NRF2 regulates homologs miR-125B1 and miR-29B1 to repress the apoptosis induced by the front-line AML chemotherapy agent daunorubicin.40 To better define how suppression of NRF2 sensitized MDS cells to Ara-C, we used lentivirus-mediated shRNA for knockdown of NRF2. Knockdown of NRF2 markedly enhanced apoptosis and triggered S-phrase arrest in MDS cell lines treated with Ara-C. Taken together, NRF2 levels regulate MDS cells’ sensitivity to Ara-C therapy. There is increasing evidence to suggest that NRF2 target genes, such as HO-1, NQO1, and multidrug resistance-associated protein (MRP), are involved in cytoprotection and detoxification, thus providing drug resistance in anti-cancer therapy.12,41,42 To determine the mechanisms of NRF2-mediated Ara-C resistance in MDS, we performed GSEA analysis on published data and then 494

discovered a list of the genes (n=37) up-regulated in both high-risk MDS patients and Ara-C-resistant AML patients. One of the genes on the list is DUSP1 (also known as MKP1), which regulates mitogen-activated protein kinase (MAPKinase) by dephosphorylation of threonine and tyrosine residues.43 DUSP1 may play an important role in the cellular response to environmental oxidative stress and agents that damage DNA.44 However, little is known about the relationship between DUSP1 and NRF2 or the effect of DUSP1 on chemo-resistance in MDS. Our ChIP q-PCR and q-PCR data indicated DUSP1 was an NRF2 direct target gene. Our IHC and immunoblotting data showed that DUSP1 expressions were elevated in higher-risk or Ara-C-resistant MDS. Given the small number of MDS patients studied, future validation with larger cohorts is needed. Interestingly, downregulation of DUSP1 by inhibitor or lentivirus shRNA could abrogate Ara-C resistance in NRF2-elevated MDS cells. There is growing evidence to demonstrate that NRF2 activation by antioxidant interventions increased cancer cell migration and induced tumor metastasis by decreasing ROS levels.45,46 It has been suggested that NRF2 improved sensitivity of AML cells to chemotherapy by compromising the ability of the AML cell to scavenge the ROS.38 Our results suggested that ROS signaling pathways may play limited roles in NRF2-mediated Ara-C resistance in MDS cells. We identified a larger cohort of the genes (n=331) up-regulated in both high-risk MDS patients and Ara-C-resistant AML patients. We found significant enrichment of the up-regulated genes in 12 KEGG pathways pertaining to cell signaling (e.g. MAPK and JAK-STAT signaling), immune responses (e.g. chemokine signaling and lysosome signaling), and cell death (e.g. apoptosis and FoxO signaling). It has been reported that alterations of SETD2 (encoding the histone 3 lysine 36 trimethyltransferase) and EZH2 (catalyzing the trimethylation of lysine 27 of histone H3) also led to resistance to DNA damagingchemotherapy such as Ara-C in leukemia via different mechanisms.47,48 Our current study indicated that NRF2 conferred Ara-C resistance partly through DUSP1 in MDS. Other signaling pathways identified in this study warrant further investigation in the future. Our data showed that silencing NRF2 or DUSP1 significantly sensitized tumors to Ara-C by measuring tumor size in the livers of our SKM-1-transplanted mouse models. SKM-1 is a cell line carrying a large number of mutations, affecting ASXL1, BCORL1, EZH2, SF1, STAG2, TET2, TP53, and WT1, which are close to the characteristics of high-risk MDS.49 Thus, SKM-1 cell lines were used in this study. To better establish stable NRF2 or DUSP1 knockdown MDS mouse models, SKM-1 cells were transfected and then injected into NSGS mice. SKM-1 cells mainly generated tumors in the livers but not in BM or peripheral blood of NSGS mice. Although not the ideal model, our in vivo results indicated that silencing NRF2 or DUSP1 increased the sensitivity of SKM-1 cells to Ara-C treatment. Future MDS patientderived xenograft models are needed to validate our findings.50 In conclusion, our clinical and experimental results revealed that NRF2 expression levels are elevated in high-risk MDS patients and serve as a statistically significant prognostic variable for OS in MDS patients. Pharmacological inhibition of NRF2 re-sensitizes MDS haematologica | 2019; 104(3)


NRF2 and Ara-C resistance in MDS

cells to Ara-C treatment while activation of NRF2 by agonist resulted in the reduced sensitivity to Ara-C. NRF2 mediates Ara-C resistance partly through its direct target gene DUSP1. Taken together, our findings suggest that silencing NRF2 re-sensitizes high-risk MDS cells to Ara-C treatment. Targeting NRF2 in combination with conventional chemotherapy could overcome drug resistance in high-risk MDS patients.

References 1. Cazzola M, Malcovati L. Myelodysplastic syndromes--coping with ineffective hematopoiesis. N Engl J Med. 2005; 352(6):536-538. 2. Greenberg P, Cox C, LeBeau MM, et al. International scoring system for evaluating prognosis in myelodysplastic syndromes. Blood. 1997;89(6):2079-2088. 3. Greenberg PL, Tuechler H, Schanz J, et al. Revised international prognostic scoring system for myelodysplastic syndromes. Blood. 2012;120(12):2454-2465. 4. Killick SB, Carter C, Culligan D, et al. Guidelines for the diagnosis and management of adult myelodysplastic syndromes. Br J Haematol. 2014;164(4):503-525. 5. Burnett AK, Milligan D, Prentice AG, et al. A comparison of low-dose cytarabine and hydroxyurea with or without all-trans retinoic acid for acute myeloid leukemia and high-risk myelodysplastic syndrome in patients not considered fit for intensive treatment. Cancer. 2007;109(6):1114-1124. 6. Miller KB, Kim K, Morrison FS, et al. The evaluation of low-dose cytarabine in the treatment of myelodysplastic syndromes: a phase-III intergroup study. Ann Hematol. 1992;65(4):162-168. 7. Hayes JD, Dinkova-Kostova AT. The Nrf2 regulatory network provides an interface between redox and intermediary metabolism. Trends Biochem Sci. 2014;39(4):199218. 8. Ma Q. Role of nrf2 in oxidative stress and toxicity. Annu Rev Pharmacol Toxicol. 2013;53:401-426. 9. Jaramillo MC, Zhang DD. The emerging role of the Nrf2-Keap1 signaling pathway in cancer. Genes Dev. 2013;27(20):21792191. 10. Rushworth SA, Zaitseva L, Murray MY, Shah NM, Bowles KM, MacEwan DJ. The high Nrf2 expression in human acute myeloid leukemia is driven by NF-kappaB and underlies its chemo-resistance. Blood. 2012;120(26):5188-5198. 11. Zhang BP, Zhao J, Li SS, et al. Mangiferin activates Nrf2-antioxidant response element signaling without reducing the sensitivity to etoposide of human myeloid leukemia cells in vitro. Acta Pharmacol Sin. 2014;35(2):257-266. 12. Wu RP, Hayashi T, Cottam HB, et al. Nrf2 responses and the therapeutic selectivity of electrophilic compounds in chronic lymphocytic leukemia. Proc Natl Acad Sci U S A. 2010;107(16):7479-7484. 13. Niture SK, Khatri R, Jaiswal AK. Regulation of Nrf2-an update. Free Radic Biol Med. 2014;66:36-44. 14. Rushworth SA, Macewan DJ. The role of nrf2 and cytoprotection in regulating chemotherapy resistance of human

haematologica | 2019; 104(3)

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

Funding This work was supported by grants from the National Key Technology R&D Program (2014BAI09B13), National Natural Science Foundation of China Grants (81270582, 81470290, 81700121, 81800121), Major Program Fund of the Science Technology Department of Zhejiang Province (2013c03043-2), the Taub Foundation (to GH), and National Institutes of Health (NIH) (R01DK105014 to GH).

leukemia cells. Cancers (Basel). 2011; 3(2):1605-1621. Rushworth SA, Bowles KM, MacEwan DJ. High basal nuclear levels of Nrf2 in acute myeloid leukemia reduces sensitivity to proteasome inhibitors. Cancer Res. 2011; 71(5):1999-2009. Cho JM, Manandhar S, Lee HR, Park HM, Kwak MK. Role of the Nrf2-antioxidant system in cytotoxicity mediated by anticancer cisplatin: implication to cancer cell resistance. Cancer Lett. 2008;260(1-2):96108. Wunderlich M, Stockman C, Devarajan M, et al. A xenograft model of macrophage activation syndrome amenable to antiCD33 and anti-IL-6R treatment. JCI Insight. 2016;1(15):e88181. Pellagatti A, Cazzola M, Giagounidis A, et al. Deregulated gene expression pathways in myelodysplastic syndrome hematopoietic stem cells. Leukemia. 2010;24(4):756764. Subramanian A, Tamayo P, Mootha VK, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005;102(43):15545-15550. Hayashi Y, Zhang Y, Yokota A, Yan XM, Liu JQ, et al. Pathobiologic pseudohypoxia as a putative mechanism underlying myelodysplastic syndromes. Cancer Discov. 2018 August 23. [Epub ahead of print] Tang X, Wang H, Fan L, et al. Luteolin inhibits Nrf2 leading to negative regulation of the Nrf2/ARE pathway and sensitization of human lung carcinoma A549 cells to therapeutic drugs. Free Radic Biol Med. 2011;50(11):1599-1609. Chian S, Li YY, Wang XJ, Tang XW. Luteolin sensitizes two oxaliplatin-resistant colorectal cancer cell lines to chemotherapeutic drugs via inhibition of the Nrf2 pathway. Asian Pac J Cancer Prev. 2014;15(6):2911-2916. Jazwa A, Rojo AI, Innamorato NG, Hesse M, Fernandez-Ruiz J, Cuadrado A. Pharmacological targeting of the transcription factor Nrf2 at the basal ganglia provides disease modifying therapy for experimental parkinsonism. Antioxid Redox Signal. 2011;14(12):2347-2360. Chorley BN, Campbell MR, Wang X, et al. Identification of novel NRF2-regulated genes by ChIP-Seq: influence on retinoid X receptor alpha. Nucleic Acids Res. 2012;40(15):7416-7429. Kobayashi EH, Suzuki T, Funayama R, et al. Nrf2 suppresses macrophage inflammatory response by blocking proinflammatory cytokine transcription. Nat Commun. 2016;7:11624. Molina G, Vogt A, Bakan A, et al. Zebrafish chemical screening reveals an inhibitor of

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

Dusp6 that expands cardiac cell lineages. Nat Chem Biol. 2009;5(9):680-687. Kesarwani M, Kincaid Z, Gomaa A, et al. Targeting c-FOS and DUSP1 abrogates intrinsic resistance to tyrosine-kinase inhibitor therapy in BCR-ABL-induced leukemia. Nat Med. 2017;23(4):472-482. Greco WR, Bravo G, Parsons JC. The search for synergy: a critical review from a response surface perspective. Pharmacol Rev. 1995;47(2):331-385. Zhao W, Sachsenmeier K, Zhang L, Sult E, Hollingsworth RE, Yang H. A New Bliss Independence Model to Analyze Drug Combination Data. J Biomol Screen. 2014;19(5):817-821. Garcia-Manero G, Tambaro FP, Bekele NB, et al. Phase II trial of vorinostat with idarubicin and cytarabine for patients with newly diagnosed acute myelogenous leukemia or myelodysplastic syndrome. J Clin Oncol. 2012;30(18):2204-2210. Braun T, Carvalho G, Coquelle A, et al. NFkappaB constitutes a potential therapeutic target in high-risk myelodysplastic syndrome. Blood. 2006;107(3):1156-1165. Khacho M, Clark A, Svoboda DS, et al. Mitochondrial Dynamics Impacts Stem Cell Identity and Fate Decisions by Regulating a Nuclear Transcriptional Program. Cell Stem Cell. 2016;19(2):232247. Li L, Li M, Sun C, et al. Altered hematopoietic cell gene expression precedes development of therapy-related myelodysplasia/ acute myeloid leukemia and identifies patients at risk. Cancer Cell. 2011; 20(5):591-605. Nakagawa T, Matozaki S, Murayama T, et al. Establishment of a leukaemic cell line from a patient with acquisition of chromosomal abnormalities during disease progression in myelodysplastic syndrome. Br J Haematol. 1993;85(3):469-476. Liu Z, Chen J, Wang H, et al. Chidamide shows synergistic cytotoxicity with cytarabine via inducing G0/G1 arrest and apoptosis in myelodysplastic syndromes. Am J Transl Res. 2017;9(12):5631-5642. Murase M, Iwamura H, Komatsu K, et al. Lack of cross-resistance to FF-10501, an inhibitor of inosine-5'-monophosphate dehydrogenase, in azacitidine-resistant cell lines selected from SKM-1 and MOLM-13 leukemia cell lines. Pharmacol Res Perspect. 2016;4(1):e00206. Dicker F, Haferlach C, Sundermann J, et al. Mutation analysis for RUNX1, MLL-PTD, FLT3-ITD, NPM1 and NRAS in 269 patients with MDS or secondary AML. Leukemia. 2010;24(8):1528-1532. Karathedath S, Rajamani BM, Musheer Aalam SM, et al. Role of NF-E2 related factor 2 (Nrf2) on chemotherapy resistance in acute myeloid leukemia (AML) and the

495


P. Lin et al.

39.

40.

41.

42.

496

effect of pharmacological inhibition of Nrf2. PLoS One. 2017;12(5):e0177227. Hong F, Freeman ML, Liebler DC. Identification of sensor cysteines in human Keap1 modified by the cancer chemopreventive agent sulforaphane. Chem Res Toxicol. 2005;18(12):1917-1926. Shah NM, Zaitseva L, Bowles KM, MacEwan DJ, Rushworth SA. NRF2-driven miR-125B1 and miR-29B1 transcriptional regulation controls a novel anti-apoptotic miRNA regulatory network for AML survival. Cell Death Differ. 2015;22(4):654664. Cai H, Scott E, Kholghi A, et al. Cancer chemoprevention: Evidence of a nonlinear dose response for the protective effects of resveratrol in humans and mice. Sci Transl Med. 2015;7(298):298ra117. Cole SP. Targeting multidrug resistance protein 1 (MRP1, ABCC1): past, present, and

43.

44.

45.

46.

future. Annu Rev Pharmacol Toxicol. 2014;54:95-117. Checker R, Gambhir L, Sharma D, Kumar M, Sandur SK. Plumbagin induces apoptosis in lymphoma cells via oxidative stress mediated glutathionylation and inhibition of mitogen-activated protein kinase phosphatases (MKP1/2). Cancer Lett. 2015;357(1):265-278. Keyse SM, Emslie EA. Oxidative stress and heat shock induce a human gene encoding a protein-tyrosine phosphatase. Nature. 1992;359(6396):644-647. Wang H, Liu X, Long M, et al. NRF2 activation by antioxidant antidiabetic agents accelerates tumor metastasis. Sci Transl Med. 2016;8(334):334ra351. Tschop MH, Stumvoll M, Ristow M. Opposing Effects of Antidiabetic Interventions on Malignant Growth and Metastasis. Cell Metab. 2016;23(6):959-960.

47. Mar BG, Chu SH, Kahn JD, et al. SETD2 alterations impair DNA damage recognition and lead to resistance to chemotherapy in leukemia. Blood. 2017;130(24):2631-2641. 48. Gollner S, Oellerich T, Agrawal-Singh S, et al. Loss of the histone methyltransferase EZH2 induces resistance to multiple drugs in acute myeloid leukemia. Nat Med. 2017;23(1):69-78. 49. Palau A, Mallo M, Palomo L, et al. Immunophenotypic, cytogenetic, and mutational characterization of cell lines derived from myelodysplastic syndrome patients after progression to acute myeloid leukemia. Genes Chromosomes Cancer. 2017;56(3):243-252. 50. Medyouf H, Mossner M, Jann JC, et al. Myelodysplastic cells in patients reprogram mesenchymal stromal cells to establish a transplantable stem cell niche disease unit. Cell Stem Cell. 2014;14(6):824-837.

haematologica | 2019; 104(3)


ARTICLE

Myelodysplastic Syndromes

Dyserythropoiesis evaluated by the RED score and hepcidin:ferritin ratio predicts response to erythropoietin in lower-risk myelodysplastic syndromes Sophie Park,1,2 Olivier Kosmider,3 Frédéric Maloisel,4 Bernard Drenou,5 Nicolas Chapuis,6 Thibaud Lefebvre,7 Zoubida Karim,6 Hervé Puy,6 Anne Sophie Alary,3 Sarah Ducamp,7 Frédérique Verdier,7 Cécile Bouilloux,1,2 Alice Rousseau,7 Marie-Christine Jacob,8 Agathe Debliquis,9 Agnes Charpentier,10 Emmanuel Gyan,11 Bruno Anglaret,12 Cecile Leyronnas,13 Selim Corm,14 Borhane Slama,15 Stephane Cheze,16 Kamel Laribi,17 Shanti Amé,18 Christian Rose,19 Florence Lachenal,20 Andrea Toma,21 Gian Matteo Pica,22 Martin Carre,1,2 Frédéric Garban,1,2 Clara Mariette,1,2 Jean-Yves Cahn,1,2 Mathieu Meunier,1,2 Olivier Herault,23 Pierre Fenaux,24 Orianne Wagner-Ballon,25 Valerie Bardet,26 Francois Dreyfus27 and Michaela Fontenay3

Department of Hematology, CHU Grenoble-Alpes, Grenoble; 2Institute for Advanced Biosciences, INSERM U1209, CNRS UMR 5309, Grenoble; 3Assistance PubliqueHôpitaux de Paris (AP-HP), Service d’Hématologie Biologique, Hôpitaux Universitaires Paris Centre, Institut Cochin, Université Paris Descartes; 4SOL Hematology, Clinique Saint Anne, Strasbourg; 5Department of Hematology, Hôpital Emile Muller, CH de Mulhouse; 6INSERM UMR1149, CNRS 8252 - Centre de Recherche sur l'Inflammation (CRI) Equipe “Hème, Fer et Pathologies Inflammatoires”, Labex GREX, Centre Français des Porphyries - Hôpital Louis Mourier HUPNVS, Paris; 7Institut Cochin, INSERM U1016, CNRS UMR 8104, Paris Descartes University; 8Institut de Biologie et Pathologie, Immunology, CHU Grenoble-Alpes, Grenoble; 9Hematology Laboratory, Mulhouse Hospital; 10Department of Hematology, Hôpital St Philibert, Lille; 11Department of Hematology, CHU de Tours; 12CH de Valence; 13Institut Daniel Hollard, Grenoble; 14 Medipole de Savoie, Challes les Eaux; 15Department of Hematology, CH d'Avignon; 16 Department of Hematology, CHU Caen; 17Department of Hematology , CH Le Mans; 18 Department of Hematology, Hôpital Civil, CHU Strasbourg; 19Department of Hematology, Hôpital Saint Vincent de Paul, Lille; 20Department of Hematology CH Pierre Oudot, Bourgoin-Jallieu; 21Department of Hematology, Hôpital Universitaire Henri Mondor, AP-HP, Université Paris 12, Créteil; 22Department of Hematology, CH Métropole Savoie, Chambery; 23Laboratoire d’Hématologie, CHU Tours; 24Department of Hematology, Saint Louis Hospital, AP-HP, Université Paris Diderot; 25Département d’Hématologie et Immunologie Biologiques, Hôpital Universitaire Henri Mondor, Creteil; 26 Service d’Hématologie Immunologie Transfusion, Hôpitaux Universitaires Paris Ile de France-Ouest, AP-HP and 27Department of Hematology, Cochin Hospital, Paris V, France

Ferrata Storti Foundation

Haematologica 2019 Volume 104(3):497-504

1

ABSTRACT

E

rythropoiesis-stimulating agents are generally the first line of treatment of anemia in patients with lower-risk myelodysplastic syndrome. We prospectively investigated the predictive value of somatic mutations, and biomarkers of ineffective erythropoiesis including the flow cytometry RED score, serum growth-differentiation factor-15, and hepcidin levels. Inclusion criteria were no prior treatment with erythropoiesis-stimulating agents, low- or intermediate-1-risk myelodysplastic syndrome according to the International Prognostic Scoring System, and a hemoglobin level <10 g/dL. Patients could be red blood cell transfusion-dependent or not and were given epoetin zeta 40 000 IU/week. Serum erythropoietin level, iron parameters, hepcidin, flow cytometry Ogata and RED scores, and growthdifferentiation factor-15 levels were determined at baseline, and molecular analysis by next-generation sequencing was also conducted. Erythroid response (defined according to the International Working Group 2006 criteria) was assessed at week 12. Seventy patients, with a median age of 78 years, were included in the study. haematologica | 2019; 104(3)

Correspondence: SOPHIE PARK spark@chu-grenoble.fr Received: Juy 27, 2018. Accepted: October 2, 2018. Pre-published: October 4, 2018. doi:10.3324/haematol.2018.203158 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/104/3/497 ©2019 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

497


S. Park et al.

There were 22 patients with refractory cytopenia with multilineage dysplasia, 19 with refractory cytopenia with unilineage dysplasia, 14 with refractory anemia with ring sideroblasts, four with refractory anemia with excess blasts-1, six with chronic myelomonocytic leukemia, two with del5qand three with unclassifiable myelodysplastic syndrome. According to the revised International Prognostic Scoring System, 13 had very low risk, 47 had low risk, nine intermediate risk and one had high-risk disease. Twenty patients were transfusion dependent. Forty-eight percent had an erythroid response and the median duration of the response was 26 months. At baseline, non-responders had significantly higher RED scores and lower hepcidin:ferritin ratios. In multivariate analysis, only a RED score >4 (P=0.05) and a hepcidin:ferritin ratio <9 (P=0.02) were statistically significantly associated with worse erythroid response. The median response duration was shorter in patients with growth-differentiation factor-15 >2000 pg/mL and a hepcidin:ferritin ratio <9 (P=0.0008 and P=0.01, respectively). In multivariate analysis, both variables were associated with shorter response duration. Erythroid response to epoetin zeta was similar to that obtained with other erythropoiesis-stimulating agents and was correlated with higher baseline hepcidin:ferritin ratio and lower RED score. ClinicalTrials.gov registration: NCT 03598582.

Introduction Myelodysplastic syndromes (MDS) are acquired, neoplastic disorders of hematopoietic stem cells characterized by ineffective and dysplastic myeloid cell differentiation, in particular dyserythropoiesis leading to anemia in more than 80% of cases.1 The main prognostic factors in MDS, which define the risk of MDS progression to acute myeloid leukemia and patientsâ&#x20AC;&#x2122; survival, include the number and importance of cytopenias as well as the percentage of bone marrow blasts and the cytogenetic abnormalities. These factors are combined in a recently revised International Prognostic Scoring System (IPSS-R).2 For lower-risk MDS, essentially characterized by anemia related to ineffective erythropoiesis, first-line treatments generally include (in the absence of the 5q chromosomal deletion) erythropoiesis-stimulating agents (ESA) (mainly epoetins and darbepoetin) with or without granulocyte - colony - stimulating factors to improve anemia. Subsequent treatments include lenalidomide, hypomethylating agents, luspatercept and other investigational drugs (APG-101, imetelstat, etc.). The response rate to ESA according to International Working Group 2006 criteria [i.e. erythroid hematologic improvement (HI-E), including red blood cell (RBC) transfusion independence or an increase of hemoglobin level >1.5 g/dL] is approximately 50% for patients with favorable prognostic factors, with a median response duration of 20 to 24 months. Our results, confirmed by other groups, suggest that this treatment does not increase the risk of transformation into acute myeloid leukemia and improves overall survival.3,4 Known prognostic factors for a favorable response to ESA in MDS patients include early stage disease, serum erythropoietin (sEPO) levels below 500 U/L and no or low requirement for RBC transfusion,5 while trilineage dysplasia may be associated with a lower response rate to ESA and/or shorter duration of response.6 We also found that a score based on IPSS-R, serum ferritin level and sEPO concentration can efficiently predict the response,7 while patients with mutations in at least three genes have a lower response rate.8 The mechanisms of primary resistance or loss of response are unknown. Relapse may be explained in only 25% to 30% of patients by a detectable progression of MDS with, as shown by worsening of 498

cytopenias other than anemia, increasing numbers of bone marrow blasts or frank progression to acute myeloid leukemia. Biomarkers of ineffective erythropoiesis may, therefore, help to predict the response to ESA. Over the past 20 years, new approaches in multiparameter flow cytometry have contributed to the diagnosis of MDS by quantifying myeloblasts, hematogones and dysgranulopoiesis using the Ogata score.9-13 However, few studies have specifically addressed the late stages of erythroid differentiation.10,14 The flow cytometry RED score10 was developed as a whole bone marrow flow cytometry protocol using the nuclear dye CyTRAK orange to gate nucleated cells without lysing red blood cells. The RED score is based on the evaluation of dyserythropoiesis with CD71 and CD36 coefficient of variation values and hemoglobin levels according to gender. It ranges from 0 to 7, with a RED score â&#x2030;Ľ3 predicting MDS with a sensitivity of 77.5% and a specificity of 90%. By combining the RED and the Ogata scores, the sensitivity can reach 87.9% and the specificity 88.9%.10 Interestingly, ineffective erythropoiesis has been correlated with high levels of serum growth differentiation factor-15 (GDF-15), a member of the transforming growth factor-b superfamily that is produced by erythroid precursors at the end of the differentiation process.15,16 Upon chronic RBC transfusions, patients with lowerrisk MDS may develop iron overload. A greater insight into iron homeostasis in MDS patients can be obtained by monitoring specific markers of iron metabolism including hepcidin, in correlation with ferritin and levels of transferrin saturation. In this trial, we aimed to find biomarkers predicting response to epoetin zeta in patients with lower-risk MDS and to use these to evaluate the efficacy of a biosimilar drug to epoetin alfa in such patients.

Methods Patients and study design Eligibility criteria for inclusion in the study were: (i) a diagnosis of MDS according to the World Health Organization (WHO) 2008 criteria [refractory anemia (RA), refractory anemia with ring sideroblasts (RARS), refractory cytopenia with multilineage dysplasia (RCMD), refractory cytopenia with unilineage dysplasia haematologica | 2019; 104(3)


RED score and hepcidin:ferritin in lower-risk MDS

(RCUD), refractory anemia with excess blasts-1 (RAEB-1), del 5q syndrome], or chronic myelomonocytic leukemia-1 (CMML-1) with a white blood cell count <13x109/L, (ii) low/intermediate (int)-1 risk according to the IPSS determined locally, (iii) hemoglobin <10 g/dL or RBC transfusion dependence, (iv) Eastern Cooperative Oncology Group performance status <2 and (v) sEPO level <500 IU/L. Exclusion criteria were: non-controlled hypertension, or cardiovascular disease (uncontrolled angina pectoris, heart failure), renal insufficiency, sEPO level >500 IU/L, systemic infection or chronic inflammatory disease, serum folate <2 ng/mL or vitamin B12 <200 pg/mL, and other non-MDS-related causes of anemia (e.g., hemolysis, hemorrhage, iron deficiency). All patients gave their written informed consent to biological investigations according to the recommendations of the local ethics committee (Comité de Protection des Personnes Paris V, CPP n. RCB 212A01395-38, EUDRACT 2012-002990-7338) and the study was conducted in accordance with the Helsinki Declaration and registered in ClinicalTrials.gov as NCT03598582.

Treatment Patients received subcutaneous epoetin zeta 40 000 IU/week for 12 weeks. Response was evaluated after 12 weeks of treatment according to International Working Group 2006 criteria. Nonresponders were excluded from the study while responsive patients continued on epoetin zeta for another 52 weeks. Patients still responding at week 52 could continue treatment, based on the physician’s decision. If hemoglobin levels exceeded 12 g/dL at any time before week 12, the dose of epoetin zeta was reduced to 20 000 IU/week. After week 12, the intervals between injections were increased by 1 week if hemoglobin levels exceeded 13 g/dL. The purpose of this dose adjustment was to reach epoetin zeta doses allowing hemoglobin levels to be maintained between 11 and 12 g/dL. During the dose adjustment period, weekly blood counts were performed. No prescription of iron was allowed in this trial, in order not to perturb iron metabolism markers. Each patient had a minimal follow-up of 52 weeks.

Biological endpoints The primary endpoint of this study was to find new biomarkers capable of predicting the response to epoetin zeta. Bone marrow aspirates were collected from all 70 patients at inclusion and then after 12 weeks. No samples were excluded based on clinical parameters. Fresh bone marrow aspirates were sent to Cochin Hospital, Paris, for centralized flow cytometry analysis of dyserythropoiesis, using the RED score, and gene sequencing. Ogata scores were also assessed locally in the hospitals of Mulhouse, Creteil, Tours, Grenoble and Cochin. Patients were re-evaluated at week 12 by flow cytometry using both the RED and Ogata scores, assessed centrally in Cochin Hospital.10 Blood plasma was also collected for quantitative analyses of hepcidin and GDF-15. Hepcidin levels from plasma samples collected in EDTA were measured by liquid chromatography coupled to tandem mass spectrometry in Louis Mourier Hospital using the method described by Lefebvre et al.17 The results are expressed as hepcidin:ferritin (x100) ratios which represent a measure of the adequacy of hepcidin level relative to iron body stores because hepcidin levels are known to be modulated by transfusion and inflammation. All patients, except one, had C-reactive protein values <5 mg/L. GDF-15 was measured by enzyme-linked immunosorbent assay at Cochin hospital, using kits obtained from R&D Systems (Minneapolis, MN, USA).18

extraction using the DNA/RNA Kit (Qiagen, Hilden, Germany). All 70 samples were screened for mutations in a panel of 26 genes (ASXL1, CBL, DNMT3A, ETV6, EZH2, FLT3, IDH1, IDH2, JAK2, KIT, KRAS, NRAS, MPL, NPM1, PHF6, PTPN11, RIT1, RUNX1, SETBP1, SF3B1, SRSF2, TET2, TP53, U2AF1, WT1 and ZRSR2) by a next-generation sequencing assay using the Ion AmpliSeq™ library kit 2 384 n. 4480442 (Life Technologies, Chicago, IL, USA). All the samples were also screened for ASXL1 (including c.1934dupG; p.G646WfsX12) and SRSF2 mutations by Sanger sequencing. JAK2, NPM1 and FLT3-ITD mutations were analyzed by real-time polymerase chain reaction and fluorescent polymerase chain reaction to confirm the next-generation sequencing data. Bioinformatic analysis was performed as previously described.19

Sample size justification and statistical analysis The sample size was computed based on an expected treatment response rate of 50% to 60%, i.e. about 30 responders and 30 nonresponders. Allowing for 10%-15% of the biological data being non-evaluable, 70 patients had to be included. For continuous variables, values were expressed as medians and interquartile ranges (IQR) or means and compared using the Wilcoxon test. Categorical variables are reported as counts or percentages with 95% confidence intervals (95% CI) and were compared using a Fisher exact or chi-squared test. Evaluation of the erythroid response (according to International Working Group 2006 criteria) after 12 weeks of treatment, i.e. the efficacy endpoint, was considered as a binary variable (response versus nonresponse). The search for factors predictive of response was performed by logistic regression and results were presented as odds ratio (OR) with 95% CI. P values <0.05 were considered statistically significant. Receiving operating characteristic (ROC) curves were used to determine the best thresholds of the prognostic factors. All analyses were performed using JMP® software (version 10, SAS Institute Inc, Cary, NC, USA).

Results Patients’ baseline characteristics Seventy patients (31 males and 39 females) were recruited in 16 French centers between January 2013 and March 2017. Their median age was 78 years (range, 57-93 years). At inclusion, the WHO classification of the patients was the following: 22 RCMD, 19 RCUD, 14 RARS, four RAEB1, six CMML, two del 5q-, and three MDS-unclassifiable. The IPSS classification was low in 43 (61.5%) and int-1 in 27 (38.5%) patients, whereas the IPSS-R classification was very low in 13 (18.5%), low in 47 (67%), intermediate in nine (13%) and high in one (1.4%) patient. Twenty (28%) patients were dependent on RBC transfusions (median >2 RBC transfusions/8 weeks) receiving a cumulated number of RBC concentrates ranging from 2 to 7 (median 3) (Table 1). Before treatment, next-generation sequencing was performed in 68 patients. The most frequent mutations involved SF3B1 (n=28), TET2 (n=20), ASXL1 (n=15), SRSF2 (n=9), DNMT3A (n=9), U2AF1 (n=7), IDH1/2 (n=6), and EZH2 (n=6) genes. One, two, three, four or five mutations were detected in 12, 12, 26, 10, seven and one patients, respectively, and 26% of patients had more than two mutations (Online Supplementary Figure S1).

Genomic studies and bioinformatic analysis

Efficacy and safety of epoetin zeta

Mononuclear cells from bone marrow aspirates were purified on a Ficoll gradient. Cell pellets were further processed for DNA

Overall the HI-E was 47.6% (33 patients). Among the 20 transfusion-dependent patients, eight obtained HI-E

haematologica | 2019; 104(3)

499


S. Park et al.

(40%) (decrease of transfusion of at least 4 RBC concentrates/8 weeks or transfusion independence). The median duration of response was 26.1 months. In three, MDS progressed to acute myeloid leukemia. Twelve patients died (including 3 with acute myeloid leukemia and 1 with progression of MDS). The median overall survival was 41 months with a median follow-up of 15.5 months. No grade >2 adverse events were reported. No cases of hypertension or deep venous thrombosis were reported. No patients had a hemoglobin greater than 13g/dL.

Flow cytometry results and correlation between iron parameters The median values of the iron parameters were: hemoglobin 9.5 g/dL (IQR 8.9-10), transferrin saturation 39% (IQR 28-56), serum ferritin 411 ng/mL (IQR 258-831) and sEPO 49 U/L (IQR 25-122). The median Ogata score was 1 (IQR 1-2) at baseline and 2 (IQR 1-2) at week 12. The median RED score was 7 (IQR 5-7) at baseline (with 90% of patients having a score ≥3), and 5 (IQR 4.75-7) at week 12 (Table 1). The median GDF-15 concentration was 1971 pg/mL (IQR 1236-2860). GDF-15 level was lower in patients with IPSS low-risk MDS than in those with int-1 MDS and correlated with transferrin saturation (P<0.003). The median hepcidin concentration was 27 ng/mL (IQR 14-42; normal 1.3-26.4) and the median hepcidin:ferritin ratio was 5.5 (IQR 1.9-11.85). The median GDF-15 levels were elevated particularly in patients with RARS and CMML (3163 pg/mL and 3520 pg/mL, respectively, compared to those in patients with other types of MDS (RA, RCMD, RAEB-1, Wilcoxon test P=0.02) (Online Supplementary Figure S2). The hepcidin:ferritin ratio was not correlated with WHO 2008 classification despite a tendency to a lower ratio in RARS. The hepcidin:ferritin ratio was inversely correlated with GDF-15 levels (R2=0.245, P=0.04), Online Supplementary Figure S3). Patients with ferritin >400 ng/mL had a significantly higher mean GDF-15 level (3011 pg/mL) than patients with ferritin <400 ng/mL (mean GDF-15 level 2018 pg/mL, P=0.005). Hepcidin:ferritin ratios were correlated with the cumulative number of RBC transfusions before the trial onset: the mean number of transfusions was 2.3 in patients with a hepcidin:ferritin ratio >9 versus 5 in patients with a hepcidin:ferritin ≤9 (P=0.03). On the basis of ROC curves, we chose the threshold of 9 for hepcidin:ferritin ratio for the remaining statistical analyses (Online Supplementary Figure S4A).

Biomarkers of response to erythropoiesis-stimulating agents Baseline age, sex, blood counts (absolute neutrophil count, platelets, hemoglobin), WHO 2008 classification, number and type of mutations, IPSS-R classification, serum ferritin, sEPO levels, Ogata score and previous RBC transfusions were not significantly associated with response (data not shown). Therefore, mutation analysis failed to identify biological markers of response to ESA. Non-responders had a significantly higher sEPO level (117.3 IU/L versus 65.5 IU/L in responders, P=0.001) (Table 2A). Non-responders had a significantly higher baseline RED score (mean 6.2 in non-responders versus 4.9 in responders, P=0.01) (Figure 1A) and lower hepcidin:ferritin ratio (mean 4.8 versus 9, P=0.04) (Figure 1B). GDF-15 levels tended to be higher in non-responders (2180 pg/mL 500

Table 1. Clinical and biological characteristics of the patients.

Variables

N. or median (range)

Sex Female/male 39/31 Age years ≤65 6 65-70 7 >70-75 14 >75-80 12 >80 31 WHO classification RCMD 22 RARS 14 RCUD 19 RAEB-1 4 Del5 q 2 MDS-U 3 CMML 6 IPSS classification Low 43 Intermediate 27 IPSS-R classification Very low 13 Low 47 Intermediate 9 High 1 Transfusion dependency Yes 20 No 50 Number of RBC transfusions/8 weeks before ESA treatment 2 or 3 15 3-6 4 5-7 1 Hb at inclusion (g/dL) 9.5 (6.7-10.7) sEPO level (IU/L) 49 (6.4-490) Serum ferritin (ng/mL) 418 (32-2566) Transferrin saturation (%) 40 (16.76-95) Hepcidin (ng/mL) 27 (7.95-163) Hepcidin/ferritin x 100 5.5 (0.76-25.2) Patients with hepcidin: ferritin ≤9 Yes 44 No 26 GDF-15 (pg/mL) 1971 (879-11788) Ogata score at T0 1 (0-4) Ogata score at week 12 2 (0-4) RED score at T0 7 (2.3-7) Patients with RED score at T0 >4 Yes 40 No 30 RED score at week 12 5 (2.2-7) Patients with RED score at week 12 >4 32 Yes 32 No 38 RED score + Ogata score 7 (1.3-11) (median, range)

% 56%/44% 9 10 20 17 44 31.5% 20% 27 6% 3% 4% 8.5% 61% 39% 19% 67% 13% 1 28% 72%

21 6 1

62% 38%

57% 43% 78% 46% 54%

WHO: World Health Organization; RCMD: refractory cytopenia with multilineage dysplasia; RARS: refractory anemia with ring sideroblasts; RCUD: refractory cytopenia with unilineage dysplasia; RAEB-1: refractory anemia with ring sideroblasts; MDS-U: myelodysplastic syndrome-unclassifiable; CMML: chronic myelomonocytic leukemia; IPSS: International Prognosis Scoring System; IPSS-R: Revised IPSS; RBC: red blood cell; ESA: erythropoietin-stimulating agent; Hb: hemoglobin; sEPO: serum erythropoietin; GDF-15: growth-differentiation factor-15: T0: baseline.

haematologica | 2019; 104(3)


RED score and hepcidin:ferritin in lower-risk MDS

versus 1822 pg/mL in responders; P=0.47) (Figure 1C). Using ROC curves, thresholds were subsequently defined as 4 for the RED score, 9 for hepcidin:ferritin ratio and 2000 pg/mL for GDF-15. ROC curves, sensitivity and specificity for these three parameters are presented in Online Supplementary Figure S4. In multivariate analysis, only RED score ≤4 and hepcidin:ferritin ratio >9 were significantly associated with better HI-E (OR 4.02, 95% CI: 1.0-22.7, P=0.05, and OR 4.44, 95% CI: 1.23-18.2, P=0.02, respectively) (Table 2B). The HI-E rate was 39% in patients with a RED score >4 versus 75% in patients with a RED score ≤4. Patients with a low hepcidin:ferritin ratio (≤9) had relatively higher sEPO levels than those with higher hepcidin:ferritin ratios (mean sEPO 108 versus 65 IU/L, respectively, P=0.003).

cidin:ferritin ratio >9 (P=0.01) (Figure 2B). The RED score was not significantly associated with response duration (P=0.4). In multivariate analysis, taking into account GDF-15 level, hepcidin:ferritin ratio and IPSS classificaiton, only GDF-15 level >2000 pg/mL and hepcidin:ferritin ratio ≤9 predicted shorter response (Table 3).

A

Biomarkers of duration of response to erythropoiesis-stimulating agents The median duration of response to epoetin zeta was 26.1 months, being 8 months in patients with GDF-15 >2000 pg/mL versus 35 months for those with GDF-15 ≤2000 pg/mL (P=0.0008) (Figure 2A). The median duration of response in patients with a hepcidin:ferritin ratio ≤9 was 10 months versus 30 months in patients with a hepTable 2A. Baseline biological markers associated with response to erythropoiesis-stimulating agents.

Variables

Responders n=33

Hb level ng/mL, mean sEPO level IU/L, mean Ferritin (ng/mL), mean Transferrin saturation (%), mean Hepcidin (ng/mL), mean Hepcidin/ferritin (x100), mean Patients with hepcidin:ferritin ≤9; n (%) GDF-15 (pg/mL) mean Patients with GDF-15 >2000 pg/mL; n (%) Ogata score, mean RED score at T0, mean Patients with RED score >4; n (%) RED score+ Ogata score, mean Patients with >2 mutations; n (%)

9.4 65.5 613 36.2 39.3 9 15 (45%)

B

Non-responders Univariate n=37 analysis (P) 9.4 117 618 50.7 30.2 6 27 (73%)

0.37 0.001 0.62 0.04 0.45 0.04 0.006

1822

2180

0.47

11 (33%) 1.26 4.9 15 (45%)

20 (54%) 1.57 6.2 23 (62%)

0.046 0.4 0.05 0.029

6.5 7 (21%)

7.6 25 (67%)

0.15 0.37

C

Table 2B. Multivariate analysis of predictors of erythroid hematologic improvement (HI-E) to erythropoiesis-stimulating agents.

Variables Hepcidin/ferritin (x100) GDF-15 (pg/mL) RED score at T0

>9 ≤9 (ref) ≤2000 >2000 (ref) ≤4 >4 (ref)

OR (95%CI)

P

4.44 [1.23-18.2] 1.00 2.86 [0.80-11.6] 1.00 4.02 [1.0-22.7] 1

0.02 0.11 0.05

Hb: hemoglobin; sEPO: serum erythropoietin; GDF-15: growth-differentiation factor-15: T0: baseline.

haematologica | 2019; 104(3)

Figure 1. Biological markers of dyserythropoiesis and correlation with response to erythropoiesis-stimulating agents. Response was defined according to the International Working Group 2006 criteria (IWG 2006) for hematologic improvement-erythroid. (A) Mean RED score before treatment for patients who did have a response (yes) or did not have a response (no). Non-responders had higher RED scores (P=0.01). (B) Mean hepcidin:ferritin ratio in patients who did or did not have a response; the hepcidin:ferritin ratio was lower in nonresponders (P=0.04). (C) Mean GDF-15 level in patients who did or did not respond to erythropoietin-stimulating agent treatment (P=0.4).

501


S. Park et al. A

Table 3. Multivariate analysis of predictors of duration of response to erythropoiesis-stimulating agents.

Variables

Threshold values ≤9 >9 (ref) >2000 ≤2000 (ref) Intermediate Low (ref)

Hepcidin:ferritin (x100) GDF-15 (pg/mL) IPSS

HR on duration of response ESA (95% CI)

P

2.46 [1.23-5.30] 1.00 3.47 [1.74-7.21] 1.00 1.73 [0.90-3.33] 1.00

0.01 0.0004 0.09

GDF-15: growth-differentiation factor-15; IPSS: International Prognosis Scoring System.

B

Figure 2. Duration of response to erythropoiesis-stimulating agents. Kaplan Meier curves showing duration of response (in months) according to (A) GDF-15 level (P=0.0008) and (B) hepcidin:ferritin ratio (P=0.01). Hepcidin:ferritin >9: red curve; hepcidin:ferritin ≤9: blue curve.

Discussion The findings of this study suggest that epoetin zeta is as effective as other ESA3,4 in this group of patients with lower-risk MDS with relatively favorable characteristics (only 28% of them were transfusion-dependent and their median sEPO level was relatively low), with a HI-E response rate of 47.6% and a median duration of response of 26.1 months. Safety findings were consistent with the known safety profiles of epoetin alfa and darbepoetin.20,21 The two major results of our study are that a low hepcidin:ferritin ratio and a high RED score were predictive of lower HI-E, while GDF-15 >2000 pg/mL and hepcidin:ferritin <9 could be predictive factors of shorter response duration. To the best of our knowledge, this is the first time that iron homeostasis parameters and dyserythropoiesis have been observed to have a predictive value with regards to response to ESA. Hepcidin, a 25-amino acid peptide, is produced mainly by hepatocytes and secreted into the plasma. This peptide lowers the amount of iron in the serum by inhibiting iron export by ferroportin, a membrane-bound cellular iron exporter present on macrophages and at the basolateral site of enterocytes, which release iron into the circulation. Hepcidin is suppressed by increased erythropoietic iron 502

demand and is upregulated in the presence of increased iron levels or elevated body iron stores.22,23 In this study, a tendency to lower hepcidin:ferritin ratios was found in RARS patients. Theoretically, hepcidin suppression should lead to an increase of iron transport across the cell membrane and, therefore, greater delivery of iron to transferrin for the needs of increased erythropoiesis, but as MDS patients have ineffective erythropoiesis, hepcidin suppression actually results in iron overload and a potential toxicity to erythropoiesis with the Fenton reaction leading to an excess of reactive oxygen species particularly in RARS. This suggests that erythropoiesis and iron homeostasis are tightly related and that dyserythropoiesis, as measured by the RED score, influences iron equilibrium by suppressing hepcidin and increasing iron absorption. The erythroid mediator GDF-15, produced in response to erythropoietin by erythroblasts, has been described as a potential suppressor of hepcidin expression.15 However, the correlation between GDF-15 and hepcidin levels was poor in studies in phlebotomized mice and in MDS patients.24,25 Recently, the role of GDF-15 in the regulation of hepcidin production in physiological and pathological settings has been challenged by the discovery of erythroferrone, which has been demonstrated to be the real suppressor of hepcidin.26 GDF-15 levels are elevated in thalassemia and congenital dyserythropoietic anemia, two disorders with a high degree of ineffective erythropoiesis15,16,27 and correlate with the severity of anemia. GDF-15 level is, therefore, more the reflection of erythroid precursor activity and a useful indicator of ineffective erythropoiesis. Taken altogether, hepcidin suppression by increased erythropoietic activity appears to underlie the development of abnormal iron overload in anemia, independently of RBC transfusions and is strongly correlated with worse erythroid response to ESA. Hepcidin, routinely available in some countries, can be measured by mass spectrometry and international laboratories are working to uniform hepcidin analysis.28,29 This assay could also be used for MDS studies. To conclude, our results could lead to the use of hepcidin measurements to predict response to ESA. Patients with a low hepcidin:ferritin ratio, high RED score and high GDF-15 level may require other treatments than ESA alone. These treatments should aim at improving the use of iron in MDS. We recently showed potential positive effects of low doses of deferasirox in MDS patients. It is postulated that this drug quenches levels of reactive oxygen species, leading to activation of nuclear factor-κB in erythroblasts, haematologica | 2019; 104(3)


RED score and hepcidin:ferritin in lower-risk MDS

which in turns induces proliferation of these cells and possible improvement of anemia.30 An ongoing clinical trial is investing the use of low-dose deferasirox in anemic patients with lower-risk MDS who have failed to respond to treatment with ESA (ClinicalTrials.gov identifier: NCT03387475). Secondly, increasing hepcidin level could prevent further accumulation of iron and reduce ironmediated tissue injury by redistributing iron from parenchymal tissues to macrophages where iron is less toxic. A clinical trial with an agonist of hepcidin (LJPC410) is ongoing in patients with b-thalassemia with myocardial iron overload (ClinicalTrials.gov identifier: NCT03381833). In patients with dysregulated iron homeostasis, hepcidin agonists or low-dose deferasirox could be considered in combination with ESA. Concerning the RED score, the data are consistent with those of other flow cytometry scores described by the European LeukemiaNet predicting a lower response to ESA or hypomethylating agents in MDS patients.31-33 Flow cytometry techniques have been recommended by international expert panels for use in the diagnosis of MDS, particularly in order to measure dyserythropoiesis objectively.14,34 A French national survey on the practical use of flow cytometry in the diagnosis of MDS showed that about 12 of the 50 French centers that routinely perform

References 1. Fontenay-Roupie M, Bouscary D, Guesnu M, et al. Ineffective erythropoiesis in myelodysplastic syndromes: correlation with Fas expression but not with lack of erythropoietin receptor signal transduction. Br J Haematol. 1999;106(2):464-473. 2. Greenberg PL, Tuechler H, Schanz J, et al. Revised International Prognostic Scoring System for myelodysplastic syndromes. Blood. 2012;120(12):2454-2465. 3. Park S, Grabar S, Kelaidi C, et al. Predictive factors of response and survival in myelodysplastic syndrome treated with erythropoietin and G-CSF: the GFM experience. Blood. 2008;111(2):574-582. 4. Jadersten M, Montgomery SM, Dybedal I, Porwit-MacDonald A, Hellstrom-Lindberg E. Long-term outcome of treatment of anemia in MDS with erythropoietin and GCSF. Blood. 2005;106(3):803-811. 5. Hellstrom-Lindberg E, Gulbrandsen N, Lindberg G, et al. A validated decision model for treating the anaemia of myelodysplastic syndromes with erythropoietin + granulocyte colony-stimulating factor: significant effects on quality of life. Br J Haematol. 2003;120(6):1037-1046. 6. Howe RB, Porwit-MacDonald A, Wanat R, Tehranchi R, Hellstrom-Lindberg E. The WHO classification of MDS does make a difference. Blood. 2004;103(9):3265-3270. 7. Santini V, Schemenau J, Levis A, et al. Can the revised IPSS predict response to erythropoietic-stimulating agents in patients with classical IPSS low or intermediate-1 MDS? Blood. 2013;122(13):2286-2288. 8. Kosmider O, Passet M, Santini V, et al. Are somatic mutations predictive of response to erythropoiesis stimulating agents in lower risk myelodysplastic syndromes? Haematologica. 2016;101(7):e280-283.

haematologica | 2019; 104(3)

flow cytometry for the diagnosis of myeloid neoplasms use flow cytometry for the diagnosis of MDS and only four centers use the RED score (Orianne Wagner-Ballon, personal communication). However, if other larger studies confirm the usefulness of the RED score in the prediction of response to ESA, these techniques will be used more often in a practical way. To conclude, in our series, the flow cytometry RED score and hepcidin:ferritin ratio enabled identification of patients who would not respond to ESA, and GDF-15 levels and hepcidin:ferritin ratio predicted shorter duration of response. These results warrant confirmation in larger series. In patients with MDS, other treatments aimed at restoring redox and iron homeostasis could efficiently improve the erythroid response. Acknowledgments This study received the support of a research grant from Hospira/Pfizer and the institutional support of Assistance Publique-Hôpitaux de Paris through the Unité de Recherche Clinique Paris Descartes Necker Cochin. The authors thank all investigators of the Groupe Francophone des Myélodysplasies (GFM) and Audrey Gauthier for technical assistance. We wish to pay tribute to Christian Rose who died recently. He was an active member of the GFM and a leader in the field of iron chelation.

9. Ogata K, Nakamura K, Yokose N, et al. Clinical significance of phenotypic features of blasts in patients with myelodysplastic syndrome. Blood. 2002;100(12):3887-3896. 10. Mathis S, Chapuis N, Debord C, et al. Flow cytometric detection of dyserythropoiesis: a sensitive and powerful diagnostic tool for myelodysplastic syndromes. Leukemia. 2013;27(10):1981-1987. 11. Westers TM, Ireland R, Kern W, et al. Standardization of flow cytometry in myelodysplastic syndromes: a report from an international consortium and the European LeukemiaNet Working Group. Leukemia. 2012;26(7):1730-1741. 12. Della Porta MG, Picone C, Pascutto C, et al. Multicenter validation of a reproducible flow cytometric score for the diagnosis of low-grade myelodysplastic syndromes: results of a European LeukemiaNET study. Haematologica. 2012;97(8):1209-1217. 13. Ogata K, Kishikawa Y, Satoh C, Tamura H, Dan K, Hayashi A. Diagnostic application of flow cytometric characteristics of CD34+ cells in low-grade myelodysplastic syndromes. Blood. 2006;108(3):1037-1044. 14. Westers TM, Cremers EM, et al. Immunophenotypic analysis of erythroid dysplasia in myelodysplastic syndromes. A report from the IMDSFlow working group. Haematologica. 2017;102(2):308-319. 15. Tanno T, Bhanu NV, Oneal PA, et al. High levels of GDF15 in thalassemia suppress expression of the iron regulatory protein hepcidin. Nat Med. 2007;13(9):1096-1101. 16. Tamary H, Shalev H, Perez-Avraham G, et al. Elevated growth differentiation factor 15 expression in patients with congenital dyserythropoietic anemia type I. Blood. 2008;112(13):5241-5244. 17. Lefebvre T, Dessendier N, Houamel D, et al. LC-MS/MS method for hepcidin-25 measurement in human and mouse serum:

18.

19.

20.

21.

22.

23.

24.

25.

clinical and research implications in iron disorders. Clin Chem Lab Med. 2015;53(10): 1557-1567. Forejtnikova H, Vieillevoye M, Zermati Y, et al. Transferrin receptor 2 is a component of the erythropoietin receptor complex and is required for efficient erythropoiesis. Blood. 2010;116(24):5357-5367. Chesnais V, Arcangeli ML, Delette C, et al. Architectural and functional heterogeneity of hematopoietic stem/progenitor cells in non-del(5q) myelodysplastic syndromes. Blood. 2017;129(4):484-496. Park S, Fenaux P, Greenberg P, et al. Efficacy and safety of darbepoetin alpha in patients with myelodysplastic syndromes: a systematic review and meta-analysis. Br J Haematol. 2016;174(5):730-747. Platzbecker U, Symeonidis A, Oliva EN, et al. A phase 3 randomized placebo-controlled trial of darbepoetin alfa in patients with anemia and lower-risk myelodysplastic syndromes. Leukemia. 2017;31(9):19441950. Delaby C, Pilard N, Goncalves AS, Beaumont C, Canonne-Hergaux F. Presence of the iron exporter ferroportin at the plasma membrane of macrophages is enhanced by iron loading and down-regulated by hepcidin. Blood. 2005;106(12):3979-3984. Nemeth E, Tuttle MS, Powelson J, et al. Hepcidin regulates cellular iron efflux by binding to ferroportin and inducing its internalization. Science. 2004;306(5704):2090-2093. Casanovas G, Vujic Spasic M, Casu C, Rivella S, et al. The murine growth differentiation factor 15 is not essential for systemic iron homeostasis in phlebotomized mice. Haematologica. 2013;98(3):444-447. Santini V, Girelli D, Sanna A, et al. Hepcidin levels and their determinants in different types of myelodysplastic syndromes. PLoS

503


S. Park et al. One. 2011;6(8):e23109. 26. Kautz L, Jung G, Du X, et al. Erythroferrone contributes to hepcidin suppression and iron overload in a mouse model of betathalassemia. Blood. 2015;126(17):20312037. 27. Ramirez JM, Schaad O, Durual S, et al. Growth differentiation factor 15 production is necessary for normal erythroid differentiation and is increased in refractory anaemia with ring-sideroblasts. Br J Haematol. 2009;144(2):251-262. 28. Lefebvre T, Lasocki S, Feneant-Thibault M, et al. Added value of hepcidin quantification for the diagnosis and follow-up of anemia-related diseases. Ann Biol Clin (Paris). 2017;75(1):9-18.

504

29. van der Vorm LN, Hendriks JC, Laarakkers CM, et al. Toward worldwide hepcidin assay harmonization: identification of a commutable secondary reference material. Clin Chem. 2016;62(7):993-1001. 30. Meunier M, Ancelet S, Lefebvre C, et al. Reactive oxygen species levels control NFkappaB activation by low dose deferasirox in erythroid progenitors of low risk myelodysplastic syndromes. Oncotarget. 2017;8(62):105510-105524. 31. Westers TM, Alhan C, Chamuleau ME, et al. Aberrant immunophenotype of blasts in myelodysplastic syndromes is a clinically relevant biomarker in predicting response to growth factor treatment. Blood. 2010;115(9):1779-1784.

32. Alhan C, Westers TM, Cremers EM, et al. High flow cytometric scores identify adverse prognostic subgroups within the revised International Prognostic Scoring System for myelodysplastic syndromes. Br J Haematol. 2014;167(1):100-109. 33. Alhan C, Westers TM, van der Helm LH, et al. Absence of aberrant myeloid progenitors by flow cytometry is associated with favorable response to azacitidine in higher risk myelodysplastic syndromes. Cytometry B Clin Cytom. 2014;86(3):207-215. 34. Valent P, Orazi A, Steensma DP, et al. Proposed minimal diagnostic criteria for myelodysplastic syndromes (MDS) and potential pre-MDS conditions. Oncotarget. 2017;8(43):73483-73500.

haematologica | 2019; 104(3)


ARTICLE

Acute Myeloid Leukemia

Transglutaminase 2 programs differentiating acute promyelocytic leukemia cells in all-trans retinoic acid treatment to inflammatory stage through NF-κB activation

Ferrata Storti Foundation

Károly Jambrovics,1 Iván P. Uray,2 Zsolt Keresztessy,1,3 Jeffrey W. Keillor,4 László Fésüs1,5 and Zoltán Balajthy1

Department of Biochemistry and Molecular Biology, Faculty of Medicine, University of Debrecen, Hungary; 2Department of Clinical Oncology, Faculty of Medicine, University of Debrecen, Hungary; 3Genome Medicine and Bioinformatics Core Facility, Research Center for Molecular Medicine, University of Debrecen, Hungary; 4Department of Chemistry and Biomolecular Sciences, University of Ottawa, ON Canada and 5MTA DE Apoptosis, Genomics and Stem Cell Research Group of the Hungarian Academy of Sciences, University of Debrecen, Hungary 1

Haematologica 2019 Volume 104(3):505-515

ABSTRACT

D

ifferentiation syndrome (DS) is a life-threatening complication arising during retinoid treatment of acute promyelocytic leukemia (APL). Administration of all-trans retinoic acid leads to significant changes in gene expression, among the most induced of which is transglutaminase 2, which is not normally expressed in neutrophil granulocytes. To evaluate the pathophysiological function of transglutaminase 2 in the context of immunological function and disease outcomes, such as excessive superoxide anion, cytokine, and chemokine production in differentiated NB4 cells, we used an NB4 transglutaminase knock-out cell line and a transglutaminase inhibitor, NC9, which inhibits both transamidase- and guanosine triphosphate-binding activities, to clarify the contribution of transglutaminase to the development of potentially lethal DS during all-trans retinoic acid treatment of APL. We found that such treatment not only enhanced cell-surface expression of CD11b and CD11c but also induced high-affinity states; atypical transglutaminase 2 expression in NB4 cells activated the nuclear factor kappa (κ)-light-chain-enhancer of the activated B-cell pathway, driving pathogenic processes with an inflammatory cascade through the expression of numerous cytokines, including tumor necrosis factor alpha (TNF-α), interleukin 1 beta (IL-1b), and monocyte chemoattractant protein 1. NC9 decreased the amount of transglutaminase 2, p65/RelA, and p50 in differentiated NB4 cells and their nuclei, leading to attenuated inflammatory cytokine synthesis. NC9 significantly inhibits transglutaminase 2 nuclear translocation but accelerates its proteasomal breakdown. This study demonstrates that transglutaminase 2 expression induced by alltrans retinoic acid treatment reprograms inflammatory signaling networks governed by nuclear factor κ-light-chain-enhancer of activated Bcell activation, resulting in overexpression of TNF-α and IL-1b in differentiating APL cells, suggesting that atypically expressed transglutaminase 2 is a promising target for leukemia treatment.

Introduction Acute promyelocytic leukemia (APL), an acute myeloid leukemia (AML) subtype, is identified by clonal proliferation of promyelocytic precursor cells with reduced ability to differentiate into mature neutrophil granulocytes.1-6 Expression of PML/RARα in APL suppresses differentiation along the neutrophil lineage.7-9 In clinical settings, the target is primarily the PML/RARα chimeric protein and its degradation, initiated by all-trans retinoic acid (ATRA) or arsenic trioxide.10-12 ATRA-induced differentiation therapy leads to differentiation syndrome (DS), which can be fatal in 2.5-30% of cases. DS is characterized by large numbers of haematologica | 2019; 104(3)

Correspondence: ZOLTAN BALAJTHY balajthy@med.unideb.hu Received: March 6, 2018. Accepted: September 19, 2018. Pre-published: September 28, 2018. doi:10.3324/haematol.2018.192823 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/104/3/505 ©2019 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

505


K. Jambrovics et al.

inflammatory differentiating leukemic cells in the bloodstream, releasing chemokines and cytokines in a so-called cytokine storm, which shifts endothelial cell function from normal toward inflammatory processes. DS is also characterized by manifestation of unexplained fever, respiratory distress, pleural and pericardial effusions, pulmonary edema, episodic hypotension, and vascular capillary leakage, which may lead to acute renal failure.13,14 Although glucocorticoid treatment leads to recovery in most patients within 12 hours (h) and resolution of symptoms within 24 h, the condition is fatal in 1-5% of patients. Dexamethasone treatment will not inhibit the induction of chemokines in differentiating APL cells.15,16 ATRA-induced differentiation can be modeled to a certain extent using NB4 APL cells.17-19 The differentiation process involves modulation of thousands of genes to produce functional neutrophil granulocytes. The most highly up-regulated gene in ATRA-activated maturation of NB4 cells is tissue transglutaminase (TG2). TG2 expression silencing in NB4 cells has revealed functional TG2 participation in modulation of gene expression, reactive oxygen species (ROS) generation, cytokine expression, adhesion, and migration, and phagocytic capacity of differentiated neutrophil granulocytes.20,21 TG2 is a Ca2+-dependent protein cross-linking enzyme that also adds amines to proteins and is capable of deamidating γ-carboxamide groups of particular protein-bound glutamines.22,23 In addition, TG2 has several enzymatic activities that do not require Ca2+; it can hydrolyze guanosine triphosphate (GTP) and adenosine triphosphate (ATP), can mediate signal transduction via G-protein-coupled receptors, and has protein kinase and protein disulfide isomerase activities. Recent evidence shows that TG2 in the GTP-bound/closed (signaling) conformation drives cancer cell survival.24,25 To provide firm evidence for the critical involvement of TG2 in ATRA-induced differentiation of promyelocytic leukemia cells to inflammatory neutrophils, we generated TG2-deleted NB4 cells and applied a cell-penetrable, irreversible TG2 inhibitor to observe how TG2 influences the development of inflammatory states. Our results demonstrate that ATRA-induced atypical TG2 expression enhances NF-κB gene expression, nuclear translocation, and transcriptional activation of NF-κB target genes, leading to unregulated production of inflammatory cytokines and chemokines.

Methods Cell lines, treatments and measurements The cell culture conditions of the NB4 APL cell line have been described previously.18 The NB4 TG2-KO cell line was generated from the wild-type cell line by TALEN which is described in detail in the Online Supplementary Appendix. NB4 cell lines were treated with 1 µM ATRA (Sigma-Aldrich) or 1 mM ATRA + 30 mM NC9 (30 mM stock solution) and collected at the indicated time points. Phorbolmyristate acetate (PMA) in 10 nM or tumor necrosis factor alpha (TNF-α) at a concentration of 2 ng/mL were used. Western blot analysis (preparation of cell lysate and subcellular fractions), cell preparation, and staining methods for fluorescenceactivated cell sorting (FACS) analysis and the nitro-blue-tetrazolium test (NBT) have been described previously.20,21 To evaluate NF-κB pathway activity, an NF-κB promoter-driven 506

luciferase construct was used, stably integrated into the genomic DNA of NB4 cell lines (CLS-013-L8-QIAGEN). The assay was performed according to the manufacturer’s protocol. The transfected cells were selected by administration of puromycin (SigmaAldrich) at a final concentration of 10 mg/mL. Measurement of luciferase activity was performed using a Bright-Glo™ Luciferase Assay System (Promega). The data were validated by GraphPad Prism 7.0 using a parallel normalizing method based on cell numbers and protein concentration. The preparation of the RNA samples has been published previously.20,21 Real-time Q-PCR reaction utilized the following probes (ABI, Applied Biosystems): TGM2, MCP-1, TNF-α, IL-1b, GP91PHOX, NCF2, GAPDH, and CYCLO-D. The analysis was performed using an ABI Prism 7900 (ABI, Applied Biosystems). Relative expression levels were normalized to the level of cyclophilin-D and glyceraldehyde 3-phosphate dehydrogenase (GAPDH). Secreted cytokine concentration was measured using an ELISA Kit (BioLegend/RayBiotech) according to the manufacturer’s instructions. NB4-WT cells were treated as described above for 11 days. At day 11, cells were treated with 5 mM MG132 for 3 h. Samples were harvested after 30 minutes, 1 h, 2 h, and 3 h, and handled as described previously.20,21 A more detailed description of the methods used are available in the Online Supplementary Appendix.

Results TG2 accelerates phagocytotic and antimicrobial functions of differentiating NB4 cell lines The NBT test is a simple method for examining phagocytic and oxygen-dependent antimicrobial ability of neutrophil granulocytes. We previously reported that NB4 TG2-KD (TG2 knockdown) cells reduced NBT but to a lesser extent than wild-type NB4 (NB4-WT) cells after three days of ATRA treatment.21 To determine the contribution and correlation of TG2 expression levels to differentiated neutrophil granulocyte status, NB4 human acute promyelocytic leukemia cells (NB4-WT) and sublines NB4 TG2-C (virus control), NB4 TG2-KD, NB4 TG2-ha (heterozygous allele), NB4 TG2-KO (knockout) (see Online Supplementary Figures S1-S3) were treated with ATRA. First, the level of TG2 mRNA transcription and protein expression was determined in cell lines; it was found that while the TG2-KO cell line did not express TG2, the NB4WT and NB4 TG2-C cell lines expressed TG2 abundantly and comparably. However, the NB4 TG2-KD and NB4 TG2-ha cell lines expressed TG2 at intermediate levels between the total and TG2-deficient level (Figure 1A and B). TG2-deficient conditions (TG2-KO) led to a rightward shift in the NBT-positivity response curve compared to NB4-WT cells, indicating that the moderate or total expression of TG2 accelerated the differentiation process in NB4 TG2-KD or TG2-ha and NB4 TG2-C or NB4-WT cells, respectively, resulting in an early increase in NBTpositivity and a saturation curve over the 11-day time scale (Figure 1C and Online Supplementary Figure S4).

ATRA induces expression of leukocyte b2 integrin receptors MAC-1 and p150,95 and their high-affinity state on the cell surface of NB4 cell lines

Analysis of the cell-surface expression of L-selectin, CD11b, and CD11c, which are indicators of the degree of haematologica | 2019; 104(3)


Atypical TG2 expression activates the NF-κB pathway

A

B

C

Figure 1. Tissue transglutaminase (TG2) accelerates antimicrobial ability of differentiating NB4 cell lines. (A) Relative mRNA expression of TG2 in NB4-wild-type (WT) treated with 1 μM all-trans retinoic acid (ATRA), virus control TG2-C, shRNA-silenced TG2-KD, hetero-allelic TG2-ha, and TALEN-TG2 knocked-out (KO) cells measured at the indicated days by real-time Q-PCR and normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) mRNA expression (n=3). (B) Representative Western blot showing TG2 protein expression levels upon ATRA treatment over 11 days (n=3). (C) NB4 cell lines undergoing differentiation characterized by ability to reduce nitro-blue-tetrazolium (NBT). The assays were performed at the indicated time points in triplicates. Percentage of the NBT-positivity expressed as mean %±Standard Deviation for 3 parallel experiments. Statistical analysis was performed via two-way analysis of variance (ANOVA; Bonferroni post-hoc test; *P<0.05, **P<0.01, ***P<0.005 ****P<0.001).

differentiation on ATRA-treated NB4 cell lines, showed that while the expression of L-selectin, CD11b, and CD11c increased significantly from day 0 to day 3, subsequently remaining almost unaltered regarding cell surface positivity (Online Supplementary Figure S5), the mean fluorescence intensity (MFI) of the cells increased during the 11-day treatment. ATRA-induced differentiation is associated with increasing TG2 expression parallel to the progression of the differentiation of NB4 cell lines. Neither knockdown of TG2 expression nor use of the NB4 TG2KO cell line changed the cell-surface expression of Lselectin, CD11b, and CD11c in ATRA-induced differentiation (Figure 2A-C). In addition, PMA did not stimulate cell-surface expression of CD11b over basal expression (Figure 2B). Notably, some non-activated adhesion receptors were observed on the cell surface, while others were restricted to cytoplasmic granules, like CD11b.26 To gain insight into the activation states of ATRA-treated NB4 cell lines, we used CBRM1/5 mAb, which is specific for highaffinity CD11b and does not recognize CD11b on resting myeloid cells, and examined the cell-surface expression and affinity status of CD11b. We found that ATRA not haematologica | 2019; 104(3)

only enhanced the cell-surface expression of CD11b but also induced its high-affinity state (Figure 2B). Assuming that the measured surface expression levels of CD11b might be the “basal expression levels” on NB4 cell lines, we determined the PMA-stimulated CD11b cell-surface expression levels by CBRM1/5, revealing that PMA cannot further increase cell-surface expression of CD11b integrin (Figure 2B). Antibody anti-CD11c Clone 3.9 binds to CD11c in an activation-dependent manner and is specific to the I domain of CD11c.27 Anti-CD11c antibody revealed gradually increasing expression of activated CD11c integrin receptors on ATRA-treated NB4 cell lines, which could not be further enhanced by PMA (Figure 2C).

TG2 drives the respiratory burst of NB4 cell lines Some phagocytes and neutrophil granulocytes can generate large amounts of reactive oxidants in response to particulate or soluble inflammatory stimuli. We had previously found that neutrophils of TG2-KO mice showed lower expression of both GP91PHOX mRNA and protein with less ROS production compared to wild-type mice.20 The TG2-dependence of PMA-induced ROS capacity was 507


K. Jambrovics et al. A

B

C

Figure 2. All-trans retinoic acid (ATRA) induces both CD11b and CD11c b2 integrin expression and their high-affinity state on the cell surface of NB4 cell lines. FACS analysis of (A) cell-surface expression of L-selectin, differentiation marker (B) CD11b and (C) CD11c in 1 mM ATRA, 1 mM ATRA + 30 mM NC9, or after 20 minutes stimulation of 10 nM PMA-treated NB4-WT, TG2-C, NB4 TG2-KD, TG2-ha, and TG2-KO cells, respectively, at the indicated days. Measurements were conducted in triplicates, and values were validated via Flowing software. Graphs show the representation of the mean±Standard Deviation fluorescent intensity (MFI) values, in parallel. MFI values were calculated based on each treatment’s respective isotype control (n=9). Statistical analysis was performed via two-way analysis of variance (ANOVA; Bonferroni post-hoc test; *P<0.05, **P<0.01, ***P<0.005 ****P<0.001).

further investigated by measuring the expression levels of GP91PHOX and NCF-2/P67PHOX mRNA, the two major components of the neutrophil NADPH-oxidase system. The TG2-dependent expression pattern of these genes (Figure 3A and B) was also consistent with the findings that ATRA-differentiated total TG2 expressing NB4-WT cells produced more than 20-fold higher levels of ROS (70996.33±9034.96: RLU/100 µg protein) than the ATRAdifferentiated TG2 deficient NB4 TG2-KO cells (2767.33±205.90 RLU/ 100 mg protein) (Figure 3C).

TG2 induces typical proinflammatory cytokines and chemokine expression through NF-κB and transcriptional activation Proinflammatory cytokine TNFα, IL-1b, and chemokine monocyte chemoattractant protein-1 (MCP-1/CCL2) have many functions in the progression of inflammation and activation of other leukocytes. In the case of NB4 cell lines, the expression patterns of TNFα, IL-1b, and MCP-1 in both mRNA and protein levels showed very similar trends in the context of TG2 expression. While high TG2 expression was accompanied by elevated mRNA and protein expression of cytokines and MCP-1 chemokine (NB4-WT; TNFα 3120.74±321.39 pg/mL, IL-1b 6721.27±510.05 pg/mL, and MCP-1 2053.95±55.53 pg/mL) at low (NB4 TG2-KD and TG2-ha) or deficient expression of TG2-KO, both remained low in each case (TG2-KO; TNFα 658.28±36.33 pg/mL, IL1b 818.00±10.02 pg/mL, and MCP-1, 576.16±133.86 pg/mL) (Figure 4A1-C2, black bars and lines). 508

Because the expression levels of GP91PHOX (Figure 3B), MCP-1/CCL2, MDC/CCL22,21 TNF-α, and IL1b had changed proportionally with the amount of TG2 in differentiating NB4 cell lines, and as all these proteins were found to be NF-κB-dependent genes, we hypothesized that TG2 mediated the transcriptional activity of NF-κB proportional to TG2 abundance. To confirm that TG2 level supports NF-κB-mediated transcriptional activity, we carried out a luciferase reporter assay using an NF-κB promoter-driven luciferase construct that was stably integrated into the genomic DNA of the NB4 cell lines. We found that the various expression levels of TG2 in the cell lines during the 11-day time frame were associated with proportionate NF-κB promoter-driven luciferase reporter TG2-C activity (WT: 1.41x105±3.14x104, 5 4 1.29x10 ±1.46x10 , TG2-KD 2.40x104±7.35x103, TG2-ha 1.39x104±3.63x103, TG2-KO 5.22x102±1.44x102/100 mg protein) and could not be further enhanced by exogenous TNF-α (Figure 4D).

TG2 expression level drives inflammatory cytokine expression quantitatively in resting ATRA-differentiated NB4 cells We determined the secreted proteins appearing in the supernatant of NB4 cells at normal, reduced, or abrogated TG2 expression. Inflammatory biomarkers were quantified from the supernatants of ATRA-differentiated NB4WT, TG2-KD, TG2-KO, and NB4-WT NC9-treated cells using the ELISA-based RayBiotech 200 Human Biomarker haematologica | 2019; 104(3)


Atypical TG2 expression activates the NF-κB pathway

A

B

C

Figure 3. Tissue transglutaminase (TG2) expression drives both expression of NCF2 and GP91PHOX respiratory burst oxidase genes and the generation of reactive oxygen species (ROS). (A and B) Relative mRNA expressions of NCF2 and GP91PHOX upon ATRA and ATRA + NC9 treatment over 11 days were determined at the indicated time points by real-time Q-PCR and normalized to cyclophilin-D mRNA levels in NB4-WT, TG2-C, TG2-KD, TG2-ha, and TG2-KO cells. Results are the mean±Standard Deviation of 3 independent experiments. (C) Production of ROS was determined for each cell line by using a luminescence-based method in triplicates (n=5) and reported as relative light units (RLU). Values are normalized to 100 mg protein of cell lysate content, respectively. Statistical significance was determined via two-way analysis of variance (ANOVA; Bonferroni post-hoc test; *P<0.05, **P<0.01, ***P<0.005 ****P<0.001). GAPDH: glyceraldehyde 3-phosphate dehydrogenase; ATRA: all-trans retinoic acid.

Testing Service. Among the 200 proteins available, 50 were detectable in the supernatants. Among the 50 detectable proteins, 44 were expressed in a TG2-dependent manner, from which 18 were identified as NF-κB transcription-factor target genes in the Boston NF-κB target gene database.28 Of these 18 proteins, the expression level of the following 10 cytokines was observed to change in parallel with changes in the expression of TG2 : T N F α , I-309 (CCL-1), IP-10 (CXCL10), MIP-3α (CCL20), IL-10, ICAM-1, MCSF, IL-1ra, MDC (CCL22), and PAI-1 in NB4WT, TG2-KD, and TG2-KO cells (Figure 5A). The levels of 8 NF-κB-controlled chemokines [MCP-1 (CCL2), MIP-1a (CCL3), MIP-1b (CCL4), cytokines IL-1b, IL-8, and IL-9, CCL-28, and OPN (SPP1)] were found to vary in parallel with TG2 expression (Figure 5B). Furthermore, expression of MCP-3 (CCL7), MCSF R, TNF RII, GDF-15, angiogenin, VEGF R1, PECAM-1, lymphotactin (XCL1), and CXCL16, which do not depend on NF-κB, was also TG2 expressiondependent (Figure 5C). Similarly, Cathepsin S, TNF RI, Resistin, and IL-2 Rb, but not Eotaxin-3 (CCL26) or Eotaxin-2 (CCL24), were expressed in a TG2-dependent manner (Figure 5D). The supernatant concentration of the remaining 16 biomarkers is shown in Online Supplementary Figure S6.

TG2 is a new potential chemotherapeutic target in APL to prevent development of DS NC9 is a novel penetrating, irreversible transamidase site-specific inhibitor of TG2, which can transform TG2 from its closed/folded (signaling) conformation to its open haematologica | 2019; 104(3)

(non-signaling) form, modulating both its conformation and activity.29,30 We used NC9 to test its effect on NBT-positivity, cell-surface adhesion receptors, ROS, cytokine, and chemokine production in differentiating NB4 cell lines. We did not observe any significant difference in NBT-positivity (Figure 1C, right) between NC9-treated and -untreated NB4 TG2-KO cells. However, we did observe effective inhibition of ROS production, with almost 10 times lower magnitude at day 11 in all cell lines (WT 8884.00±759.29 RLU/100 mg protein and TG2-KO 133.66±17.67 RLU/100 mg protein) (Figure 3C). We also found that at low TG2 expression levels, the mRNA and protein expression levels of TNFα and IL-1b and the expression of MCP-1 remained low, similar to TG2-deficient states in the NB4 cells (Figure 4A-C, red bars and lines). In the ATRA-treated NB4-WT and NB4 TG2-C cells, where TG2 expression was maximal, the total expression levels of mRNA of TNFα, IL-1b , and MCP-1 were reduced by at least 50% or more in the presence of NC9. The lower mRNA levels were directly reflected in the amounts of secreted TNFα, IL-1b , and MCP-1 (WT; TNFα 1133.01±214.01pg/mL, IL-1b 3467.16±310.59 pg/mL, and MCP-1 1220.14±53.09pg/mL) (Figure 4A-C, red bars and lines). In parallel, the stably integrated NF-κB promoter-driven luciferase reporter activity was at least 15 times less inducible in the NB4-WT cell lines in the presence of the TG2 inhibitor (WT: 1.12x104±7.48x103, TG2-C 1.58x104±9.73x103, TG2-KD 6.15x102±3.40x102, TG2-ha 2.54x103±2.05x103, and TG2-KO 4.15x102±1.58x102/100 mg protein) (Figure 4D). The reduced endogenous NF-κB activities also became apparent in the amounts of secreted 509


K. Jambrovics et al.

A1

A2

B1

B2

C1

C2

D

Figure 4. Tissue transglutaminase (TG2) up-regulates both expression of tumor necrosis factor alpha (TNF-α), interleukin 1 beta (IL-1b), and monocyte chemoattractant protein 1 (MCP-1) through nuclear factor kappa (κ)-light-chain-enhancer of the activated B-cell (NF-κB) pathway transcriptional activation and their secretion. Relative mRNA expression of (A1) TNF-α, (B1) IL-1b and (C1) MCP-1 in 1 mM ATRA or 1 mM ATRA plus 30 mM NC9-treated NB4-WT, TG2-C, TG2-KD, TG2-ha, and TG2-KO cells measured at the indicated days by real-time Q-PCR and normalized to cyclophilin-D mRNA expression. In the cell culture, supernatant-secreted (A2) TNFα chemokine, (B2) IL-1β and (C2) MCP-1 inflammatory cytokines were quantified by ELISA and normalized to 100 mg protein content of cell lysate. Figures show secreted protein levels from 3 independent experiments measured in triplicate and graphs show the representation of the mean±Standard Deviation. (D) All-trans retinoic acid (ATRA) induced NF-κB transcriptional activity on NF-κB responsive luciferase reporter gene. Treated and harvested cells were analyzed for luciferase reporter activity as relative light units (RLU). Luciferase activity measurements were performed in triplicates (n=9). Statistical significance was determined via the two-way analysis of variance (ANOVA; Bonferroni post-hoc test; *P<0.05, **P<0.01, ***P<0.005 ****P<0.001).

510

haematologica | 2019; 104(3)


Atypical TG2 expression activates the NF-κB pathway

A

B

C

D

Figure 5. Atypical expression levels of tissue transglutaminase (TG2) drives inflammatory cytokine expression. In the cell culture, supernatant-secreted chemokines and cytokines were quantified by ELISA-based fluorescent detection cytokine array (RayBiotech Human Cytokine Array). Figures represent the mean±Standard Deviation of secreted protein levels from 3 independent experiments. (A) TG2-quantity-dependent-regulated nuclear factor kappa (κ)-light-chainenhancer of the activated B-cell (NF-κB) transcription factor target genes, and TG2 inhibitor, NC9-sensitive secretory proteins. (B) TG2-modulated NF-κB transcription factor target genes, and TG2 inhibitor, NC9-insensitive secretory proteins. (C) TG2-dependent and TG2 inhibitor, NC9-sensitive secretory proteins. (D) TG2-dependent and TG2 inhibitor, NC9-insensitive secretory proteins. Statistical significance was determined via the Student t-test; asterisks show the significant differences compared to the wild type (Student t-test; *P<0.05, **P<0.01, ***P<0.005 ****P<0.001).

haematologica | 2019; 104(3)

511


K. Jambrovics et al. A

B

C2 C1

D1 D2

E1

E2

Figure 6. Tissue transglutaminase (TG2) both induces and guides nuclear translocation of nuclear factor kappa (κ)-light-chain-enhancer of the activated B-cell (NFκB). (A) Relative mRNA expression of TG2 in 1 mM all-trans retinoic acid (ATRA) + 30 mM NC9 treated NB4-WT, TG2-C, TG2-KD, TG2-ha, and TG2-KO cells measured at the indicated days by real-time Q-PCR and normalized to GAPDH mRNA expression (n=3). (B) Representative Western blot showing TG2 protein expression levels upon ATRA + NC9 treatment through 11 days (n=3). (C1-2) Representative Western blots showing total cell lysate samples TG2, p65/RelA, phospho-p65/RelA, and p50 protein expression levels in NB4 cell lines and its densitometric analysis, where graphs show the representation of the mean±Standard Deviation values of integrated density (n=6). (D1-2) Western blots of total cell lysate, cytosolic and nuclear fractions of TG2, and p65/RelA, phospho-p65/RelA, and p50 protein expression and its densitometry, respectively (n=10). (E1-2) Western blot analysis of NB4-WT cells upon ATRA, ATRA + NC9, and 5 mM MG132 treatment (n=6). Statistical significance was determined via the two-way analysis of variance (ANOVA; Bonferroni post-hoc test; *P<0.05, **P<0.01, ***P<0.005 ****P<0.001).

512

haematologica | 2019; 104(3)


Atypical TG2 expression activates the NF-κB pathway

inflammatory proteins, which were expressed in a TG2quantity-dependent manner (Figure 5A). Secretory proteins not targeting NF-κB genes but expressed in a TG2dependent manner also included NC9 sensitive secretory proteins (Figure 5C).

and considerably increases cytosolic levels (Figure 6D2, upper left, red bar). In addition, both phospho-p65/RelA and p50 showed similar distribution patterns upon NC9 treatment (Figure 6D2, lower left and right, black and red bars).

TG2 contributes to expression and nuclear translocation of NF-κB, which is significantly reduced by the TG2 inhibitor NC9

Discussion

NF-κB promoter-driven luciferase reporter activity and the extent of the expression of several NF-κB-controlled inflammatory biomarkers suggested that NF-κB transcriptional activity might depend on the magnitude of the expression of TG2 (Figures 4D and 5A). Our starting observation was that NB4 cells treated with ATRA plus NC9 expressed lower amounts of TG2 protein than cells treated only with ATRA (Figures 1B and 6B). First, we examined the impact of NC9 on the transcription level of TG2 mRNA and found that it did not change. The amount of TG2 protein was affected by NC9, while TG2 mRNA levels remained unaffected suggesting that the proteolytic degradation of TG2 might be increased by NC9. To test this hypothesis, 11-day differentiated NB4 cells were treated for 3 h with MG132 proteasome inhibitor. Western blot analysis of TG2 expression confirmed the enhanced proteolytic degradation of TG2 during the NC9 treatment. In the presence of NC9, the amount of TG2 decreased to approximately one-tenth that measured in cells treated with ATRA alone (Figure 6E1 and E2). Next, we investigated the effect of the lower TG2 protein levels on the expression of p65/RelA, p50, and phospho(Ser536)-p65/RelA components of NF-κB and found that the absence, or reduced gene expression, or suppression of TG2 by NC9 significantly restricted the expression of p65/RelA, phospho-p65/RelA and p50 proteins in NB4WT, TG2-C, TG2-KD, TG2-ha, and TG2-KO cells (Figure 6C1 and C2, upper and lower, left and right panels). However, we have long been aware that in the ATRA-differentiated NB4-WT cells, significant amounts of TG2 protein translocate to the nucleus during differentiation.20 To demonstrate the role of nuclear TG2, after 11 days of differentiation, the cytosolic and nuclear fractions of NB4WT were separated and analyzed for the expression of TG2, p65/RelA, phospho-p65/RelA, and p50 (Figure 6D1 and D2). In the differentiated cells, NC9 treatment was associated with significantly increased cytosolic TG2 levels and lower levels in the nucleus compared to cells treated only with ATRA (Figure 6D2, upper left). Concurrently, the amount of total and nuclear p65/RelA also decreased with NC9 treatment (Figure 6D2, upper right). Distribution patterns of phospho(Ser536)-p65/RelA, which is the transcriptionally active form of p65/RelA, were very similar to the TG2 protein detected in the total, cytosolic, and nuclear fractions (Figure 6D2, upper and lower right, black and red bars). Both total and nuclear phospho-p65/RelA protein contents were significantly reduced in the presence of NC9 inhibitor (Figure 6D2, lower left, black and red bars). The p50 subunit of NF-B showed patterns of cellular distribution very similar to p65/RelA (Figure 6D2, lower right and upper right, black bars), with a distinctive difference in the amount of nuclear p50 protein remaining low, around the detection limit, in NC9-treated NB4-WT cells (Figure 6D2, lower right, black and red bars). The cellular distribution of TG2 indicated that NC9 inhibits nuclear translocation of TG2

Treatment of APL principally comprises ATRA-based therapy, which leads to terminal differentiation of APL cells to neutrophil granulocytes. In adverse events, this treatment is frequently associated with severe hyperinflammatory reactions, leading to organ infiltration of differentiating APL cells. The pathogenic processes of the hyper-inflammatory cascade in DS are not completely understood. At least two different mechanisms appear to play important roles in DS development: differentiation of APL cells with cytokine release and adhesion, and migration of differentiated APL cells to different organs. ATRA-induced differentiation of APL cells is associated with elevated expression of inflammatory cytokines and adhesion molecules called integrins. Activation of neutrophils is elicited by mediators, including chemokines, selectins, and integrin-mediated outsidein signaling. Locally produced chemokines mediate a sequence of events leading to extravasation of leukocytes at the inflammatory site. Neutrophils express different chemokine receptors, including CXCR2, whose most potent ligand is IL8 (CXCL8).31 Shibakura et al. first demonstrated that ATRA could induce synthesis and secretion of IL8 in NB4 cells. Thereafter, it was shown that not only do ATRA-treated APL cells express IL8 mRNA in ex vivo cell cultures, but also IL8, as well as chemokines such as MCP-1 (CCL2), MIP-1a (CCL3), and MIP-1b (CCL4), are present in the serum of APL patients who developed DS during ATRA treatment.32 It has also become obvious that NB4 cells secrete IL8 constitutively, which is further enhanced during ATRA-induced differentiation, and cannot be inhibited by dexamethasone.33 In the presence of IL8, neutrophil granulocytes increase the amounts of CD11b integrin receptor along with its binding activity and the amounts of CD11c on their cell surfaces.34,35 Simultaneously, secreted TNF-α also stimulates neutrophils, triggering activation of CD11c.36 Since 1998, CD11b has been used as a surface marker of granulocytic cell differentiation of NB4 cells in published research, in spite of the fact that CD11b is mainly stored intracellularly in both specific and gelatinase granules and secretory vesicles under normal physiological conditions.37-39 Our observations confirm that during ATRA-induced differentiation of NB4 cell lines, CD11b and CD11c receptors were translocated in large quantities to the cell surface, and show that the amount of surface CD11b cannot be further increased by chemoattractant or phorbol-esters (Figure 3A-C and Online Supplementary Figure S5). Nevertheless, Nupponen et al. found that while “neutrophil CD11b expression and circulating interleukin-8 represent a diagnostic marker for early-onset neonatal sepsis,” CD11c is also a potential diagnostic biomarker for sepsis and systemic inflammation.40 An antibody detecting activation-associated CD11c revealed gradually increasing cell-surface receptor expression in an activated state, with a high affinity for ligands on ATRA-treated NB4 cell lines (Figure 2C). In this manner, inflammatory cytokines could

haematologica | 2019; 104(3)

513


K. Jambrovics et al.

result in rolling, stable adhesion, microvascular sequestration, and infiltration of differentiated APL cells, with increasing potential for the APL-mediated organ damage observed in DS.38,41 TNFα is among the most effective physiological inducers of NF-κB activity. In NB4-WT cells, TNFα did not enhance further NF-κB activity already highly elevated by ATRA, but it increased NF-κB transcriptional activity from 5.2x102±1.4x102 to 3.2x104±7.2x103 RLU in NB4 TG2-KO cells (Figure 4D). This may indicate that NF-κB maintains a low basal transcriptional activity in the absence of TG2 in NB4 TG2-KO cells, and adding exogenous TNFα could trigger more NF-κB transcription activity from the IκB:p65:p50 complex. In NB4-WT cells, ATRA-induced atypical expression of TG2 can exacerbate low-level inflammation (IL8 expression, see above) as its expression is associated with both upregulation and activation of NFκB, resulting in increased endogenous TNFα synthesis and secretion, which may become a sort of self-accelerating process (Figure 4A2-D and Figure 6C1 and 2). Our findings strongly indicate that ATRA-induced TG2 expression is associated with translocation of NF-κB into the nucleus, upregulation of numerous inflammatory genes, and secretion of their products. These include TG2-quantity-dependent-regulated NF-κB transcription factor target genes: TNFα, I-309 (CCL-1), IP-10 (CXCL10), MIP-3α (CCL20), IL10, ICAM-1, MCSF, IL1ra, MDC (CCL22), and PAI-1, whose amounts were significantly reduced in the absence of TG2 in NB4 TG2-KO cells, with the exception of IL-1ra (Figure 5A). Similarly, among TG2-modulated NF-κB transcription factor target genes and TG2 inhibitor NC9-insensitive secretory proteins [MCP-1 (CCL2), MIP-1a (CCL3), MIP-1b (CCL4), cytokines IL-1b, IL-8, IL-9, CCL-28, and OPN (SPP1)], only MIP-1a (CCL3) did not show a significant decrease in TG2 deficiency (Figure 5B). In addition, of the remaining 15 TG2 expression-dependent secretory proteins, only one was not significantly changed in the absence of TG2 expression. Together, these results suggest that expression of TG2 in ATRA-induced differentiating APL cells is a crucial factor in the developing inflammatory process. The suppression of TG2 parallels the considerably low levels of p50 and phospho (Ser536) p65/RelA (Figure 6C2), suggesting that ATRA-induced expression of TG2 reprograms APL cells to be inflammatory neutrophils and induces NF-κB nuclear translocation. Destruction of IκB is stimulated by several signals such as lipopolysaccharide, ROS, TNFα, and Il-b. One possibility for TG2-dependent NF-κB activation is that TG2 interacts with IκB to initiate its non-proteasomal degradation, causing both activation and nuclear translocation of NF-κB-s.42 Alternatively, TG2-mediated cross-linking of IκBα and activation of NFκB has been previously described, yet we could not detect any forms of IκBα polymer in differentiated NB4 cells (K Jambrovics, 2018, unpublished observation). It was demonstrated that TG2 can form a complex in cytosol as well as in nuclei, with p65 binding to the promoter of the HIF-1α transcription factor, and in this sense, TG2 could become a transcriptional co-regulator in the nucleus.42 In the presence of NC9, an irreversible transamidase site-specific inhibitor of TG2, the conformational equilib-

514

rium of TG2 shifts from the closed GTP-binding form to the open conformation characterized by disorganized GTP-binding sites.25 According to our studies, NC9induced conformational changes in TG2 significantly affect the NF-κB signaling pathway. In the presence of NC9, total TG2 is reduced, as is the nuclear translocation of the decreased amount of TG2, significantly increasing the amount of cytosolic TG2. In turn, the reduced amount of nuclear TG2 correlates with reduced total levels of nuclear p50, p65/RelA, and phospho-p65/RelA proteins in NB4-WT cells, while the level of the transcriptionally active form of p65 and phospho-p65 is increased significantly in the cytosol. This raises the possibility that TG2 in its GTP-bound closed conformation can readily translocate to the nucleus and assist the nuclear translocation of p65/RelA. When TG2 is no longer capable of adopting its GTP-bound closed conformation, due to modification by NC9, the accumulation of both TG2 and p65/RelA in the cytosol results in low NF-κB transcription activity, and, consequently, in the significantly reduced production of inflammatory cytokines and chemokines. Finally, the balance of TG2 synthesis and degradation is changed by degradation triggered by NC9. Among the secreted cytokines and chemokines, TNFα and IL-1b are the most powerful inflammatory agents. At the highest concentrations of TNFα (0.8-1.2 ng/L in DS patients), they cause capillary leakage and reduced cardiac, lung, and renal function.38,43,44 Our data indicate that TG2 is a critical component of this process, and NC9induced inhibition of TG2 may prevent development of this dangerous side effect of retinoid therapy. Overall, our study revealed a novel, active role of TG2 in expression and activation of the components of NF-κB and thus in the development of atypical response to conventional ATRA treatment of APL. Targeted suppression of the TG2-dependent process may alleviate the common and potentially fatal toxicity of retinoid treatment in APL, representing an important potential therapeutic strategy. Funding This work was supported by Hungarian grants from the National Research Found OTKA NK105046, and in part by both OTKA K 129139 and TÁMOP-4.2.2.D-15/1/KONV2015-0016 project implemented through the New Széchenyi Plan, co-financed by the European Social Fund. The work was in part also supported by the GINOP-2.3.2-15-2016-00020 | 2.3.2-15-2016-00020 MolMedEx TUMORDNS grant and the EFOP-3.6.1-16-2016-00022|3.6.1-16-2016-00022 “Debrecen Venture Catapult Program”. The work is supported by the GINOP-2.3.2-15-2016-00006|2.3.2-15-2016-00006 project co-financed by the European Union and the European Regional Development Fund. KJ received a fellowship from the DOTE Apoptosis Research Foundation. The research was partly financed by the Higher Education Institutional Excellence Programme of the Ministry of Human Capacities in Hungary, within the framework of the Biotechnology thematic programme of the University of Debrecen. Acknowledgments The authors would like to thank Dr. István Szatmári and Pál Botó for assistance with cell sorting, Orsolya Molnár for clone selection and Western blot analyses of the KO cell lines.

haematologica | 2019; 104(3)


Atypical TG2 expression activates the NF-κB pathway

References 1. Bennett JM, Catovsky D, Daniel MT, et al. Proposals for the Classification of the Acute Leukaemias French-American-British (FAB) Co-operative Group. Br J Haematol. 1976;33(4):451-458. 2. Bennett J, Catovsky D, Daniel M, et al. A variant form of hypergranular promyelocytic leukemia (M3). Ann Intern Med. 1980;92:261. 3. Bennet J. Proposed revised criteria for the classification of acute leukemia: A report from the French-American-British Cooperative Group. Ann Intern Med. 1985; 3:620-625. 4. Arber DA. Realistic pathologic classification of acute myeloid leukemias. Am J Clin Pathol. 2001;115(4):552-560. 5. Kühnl A, Grimwade D. Molecular markers in acute myeloid leukaemia. Int J Hematol. 2012;96(2):153-163. 6. Wiernik PH, Gallagher RE, Tallman MS. Acute promyelocytic leukemia. Neoplastic Diseases of the Blood: Springer, 2013:403453. 7. Grignani F, Ferrucci PF, Testa U, et al. The acute promyelocytic leukemia-specific PML-RAR fusion protein inhibits differentiation and promotes survival of myeloid precursor cells. Cell. 1993;74(3):423-431. 8. Kogan SC, Hong SH, Shultz DB, et al. Leukemia initiated by PMLRAR : the PML domain plays a critical role while retinoic acid–mediated transactivation is dispensable. Blood. 2000;95(5):1541-1550. 9. Ablain J. Revisiting the differentiation paradigm in acute promyelocytic leukemia. Blood. 2011;117(22):5795-5802. 10. Breitman T, Collins SJ, Keene B. Terminal differentiation of human promyelocytic leukemic cells in primary culture in response to retinoic acid. Blood. 1981; 57(6):1000-1004. 11. Raelson JV, Nervi C, Rosenauer A, et al. The PML/RAR alpha oncoprotein is a direct molecular target of retinoic acid in acute promyelocytic leukemia cells. Blood. 1996;88(8):2826-2832. 12. Fenaux P, Wang Z, Degos L. Treatment of acute promyelocytic leukemia by retinoids. Acute Promyelocytic Leukemia: Springer, 2007:101-128. 13. Montesinos P, Sanz MA. The differentiation syndrome in patients with acute promyelocytic leukemia: experience of the pethema group and review of the literature. Mediterr J Hematol Infect Dis. 2011; 3(1):e2011059. 14. Patatanian E, Thompson D. Retinoic acid syndrome: a review. J Clin Pharm Ther. 2008;33(4):331-338. 15. Luesink M, Pennings JL, Wissink WM, et al. Chemokine induction by all-trans retinoic acid and arsenic trioxide in acute promyelocytic leukemia: triggering the differentiation syndrome. Blood. 2009;114(27):55125521. 16. Montesinos P, Bergua JM, Vellenga E, et al.

haematologica | 2019; 104(3)

17.

18.

19.

20.

21.

22.

23.

24.

25.

26. 27.

28.

29.

Differentiation syndrome in patients with acute promyelocytic leukemia treated with all-trans retinoic acid and anthracycline chemotherapy: characteristics, outcome, and prognostic factors. Blood. 2009; 113(4):775-783. Yoshida H, Kitamura K, Tanaka K, et al. Accelerated degradation of PML-retinoic acid receptor (PML-RARA) oncoprotein by all-trans-retinoic acid in acute promyelocytic leukemia: possible role of the proteasome pathway. Cancer Res. 1996;56(13):2945-2948. Lanotte M, Martin-Thouvenin V, Najman S, et al. NB4, a maturation inducible cell line with t (15; 17) marker isolated from a human acute promyelocytic leukemia (M3). Blood. 1991;77(5):1080-1086. Tallman MS, Andersen JW, Schiffer CA, et al. Clinical description of 44 patients with acute promyelocytic leukemia who developed the retinoic acid syndrome. Blood. 2000;95(1):90-95. Balajthy Z, Csomós K, Vámosi G, et al. Tissue-transglutaminase contributes to neutrophil granulocyte differentiation and functions. Blood. 2006;108(6):2045-2054. Csomós K, Német I, Fésüs L, et al. Tissue transglutaminase contributes to the alltrans-retinoic acid–induced differentiation syndrome phenotype in the NB4 model of acute promyelocytic leukemia. Blood. 2010;116(19):3933-3943. Fesus L, Piacentini M. Transglutaminase 2: an enigmatic enzyme with diverse functions. Trends Biochem Sci. 2002; 27(10):534-539. Lorand L, Graham RM. Transglutaminases: crosslinking enzymes with pleiotropic functions. Nat Rev Mol Cell Biol. 2003; 4(2):140-156. Eckert RL, Fisher ML, Grun D, et al. Transglutaminase is a tumor cell and cancer stem cell survival factor. Mol Carcinog. 2015;54(10):947-958. Kerr C, Szmacinski H, Fisher ML, et al. Transamidase site-targeted agents alter the conformation of the transglutaminase cancer stem cell survival protein to reduce GTP binding activity and cancer stem cell survival. Oncogene. 2017;36(21):2981-2990. Borregaard N, Cowland JB. Granules of the human neutrophilic polymorphonuclear leukocyte. Blood. 1997;89(10):3503-3521. Sadhu C, Hendrickson L, Dick KO et al. Novel Tools for Functional Analysis of CD11c: Activation-Specific, ActivationIndependent, and Activating Antibodies. J Immunoassay Immunochem. 2007; 29(1):42-57. Boston University Biology. 2018. Target Genes of NF-kB. [ONLINE] Available at: h t t p s : / / w w w. b u . e d u / n f - k b / g e n e resources/target-genes/. [Last accessed 15 May 2018] Caron NS, Munsie LN, Keillor JW, et al. Using FLIM-FRET to measure conformational changes of transglutaminase type 2 in live cells. PloS One. 2012;7(8):e44159.

30. Keillor JW, Chica RA, Chabot N, et al. The bioorganic chemistry of transglutaminase: from mechanism to inhibition and engineering. Can J Chem. 2008;86(4):271-276. 31. Harada A, Sekido N, Akahoshi T, et al. Essential involvement of interleukin-8 (IL8) in acute inflammation. J Leukoc Biol. 1994;56(5):559-564. 32. Shibakura M, Niiya K, Niiya M, et al. Induction of CXC and CC chemokines by all-trans retinoic acid in acute promyelocytic leukemia cells. Leuk Res. 2005;29(7):755759. 33. Tsai W-H, Hsu H-C, Lin C-C et, al. Role of interleukin-8 and growth-regulated oncogene- in the chemotactic migration of alltrans retinoic acid-treated promyelocytic leukemic cells toward alveolar epithelial cells. Crit Care Med. 2007;35(3):879-885. 34. Detmers PA, Lo SK, Olsen-Egbert E, et al. Neutrophil-activating protein 1/interleukin 8 stimulates the binding activity of the leukocyte adhesion receptor CD11b/CD18 on human neutrophils. J Exp Med.1990;171(4):1155-1162. 35. Takami M1, Terry V, Petruzzelli L. Signaling pathways involved in IL-8-dependent activation of adhesion through Mac-1. J Immunol. 2002;168(9):4559-4566. 36. Loike JD, Sodeik B, Cao L, et al. CD11c/CD18 on neutrophils recognizes a domain at the N terminus of the A alpha chain of fibrinogen. Proc Natl Acad Sci. 1991;88(3):1044-1048. 37. Borregaard N, Cowland JB. Granules of the human neutrophilic polymorphonuclear leukocyte. Blood. 1997;89(10):3503-3521. 38. Frankel SR, Eardley A, Lauwers G, et al. The “retinoic acid syndrome” in acute promyelocytic leukemia. Ann Intern Med. 1992;117(4):292-296. 39. Bainton DF, Miller LJ, Kishimoto T, et al. Leukocyte adhesion receptors are stored in peroxidase-negative granules of human neutrophils. J Exp Med. 1987;166(6):16411653. 40. Nupponen I, Andersson S, Järvenpää A-L, et al. Neutrophil CD11b expression and circulating interleukin-8 as diagnostic markers for early-onset neonatal sepsis. Pediatrics. 2001;108(1):E12. 41. Tang L, Chai W, Ye F, et al. HMGB1 promotes differentiation syndrome by inducing hyperinflammation via MEK/ERK signaling in acute promyelocytic leukemia cells. Oncotarget. 2017;8(16):27314-27327. 42. Kumar S, Mehta K. Tissue transglutaminase constitutively activates HIF-1 promoter and nuclear factor- B via a non-canonical pathway. PloS One. 2012;7(11):e49321. 43. Flombaum CD, Isaacs M, Reich L, et al. Acute renal failure associated with the retinoic acid syndrome in acute promyelocytic leukemia. Am J Kidney Dis. 1996; 27(1):134-137. 44. Mitrovic M, Suvajdzic N, Elezovic I, et al. Thrombotic events in acute promyelocytic leukemia. Thromb Res. 2015;135(4):588593.

515


ARTICLE Ferrata Storti Foundation

Acute Myeloid Leukemia

Clinical implications of subclonal TP53 mutations in acute myeloid leukemia

Katharina T. Prochazka,1* Gudrun Pregartner,2* Frank G. Rücker,3 Ellen Heitzer,4 Gabriel Pabst,1 Albert Wölfler,1 Armin Zebisch,1 Andrea Berghold,2 Konstanze Döhner3** and Heinz Sill1**

Haematologica 2019 Volume 104(3):516-523

1 Division of Hematology, Medical University of Graz, Austria; 2Institute for Medical Informatics, Statistics and Documentation, Medical University of Graz, Austria; 3 Department of Internal Medicine III, University Hospital of Ulm, Germany and 4Institute of Human Genetics, Medical University of Graz, Austria

*KTP and GP contributed equally to this work.

**KD and HS contributed equally to this work.

ABSTRACT

T

Correspondence: HEINZ SILL heinz.sill@medunigraz.at Received: August 21, 2018. Accepted: October 9, 2018. Pre-published: October 11, 2018. doi:10.3324/haematol.2018.205013 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/104/3/516 ©2019 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

516

he role of subclonal TP53 mutations, defined by a variant allele frequency of <20%, has not been addressed in acute myeloid leukemia yet. We, therefore, analyzed their prognostic value in a cohort of 1,537 patients with newly diagnosed disease, prospectively treated within three trials of the “German-Austrian Acute Myeloid Leukemia Study Group”. Mutational analysis was performed by targeted deep sequencing and patients with TP53 mutations were categorized by their variant allele frequency into groups with frequencies >40%, 20%40% and <20%. A total of 108 TP53 mutations were found in 98 patients (6.4%). Among these, 61 patients had variant allele frequencies >40%, 19 had variant allele frequencies between 20%-40% and 18 had frequencies <20%. Compared to specimens with clonal TP53 mutations, those with subclonal ones showed significantly fewer complex karyotypes and chromosomal losses. In either TP53-mutated group, patients experienced significantly fewer complete responses (P<0.001) and had worse overall and event-free survival rates (P<0.0001). In Cox regression analyses adjusting for age, white blood cell count, cytogenetic risk and type of acute myeloid leukemia, the adverse prognostic effect of TP53 mutations remained significant for all TP53-mutated subgroups. These data suggest that subclonal TP53 mutations are a novel prognostic parameter in acute myeloid leukemia and emphasize the usefulness of next-generation sequencing technologies for risk stratification in this disorder. The study was registered at ClinicalTrials.gov with number NCT00146120. Introduction Acute myeloid leukemia (AML) is an aggressive malignancy with an annual, ageadjusted incidence of 3.5 cases per 100,000 adults, rising to 15-20 cases per 100,000 above the age of 60 years.1 It, therefore, contributes substantially to morbidity and mortality of the elderly. The pathogenesis of AML represents a multistep process involving mutagenesis, epigenetic dysregulation and formation of copy number aberrations. During the process of leukemogenesis, initiating mutations affect hematopoietic stem and progenitor cells, giving rise to preleukemic/leukemic stem cells and, ultimately, frank leukemia. Virtually every AML genome is characterized by a number of subclones of variable size. These subclones may have different pathobiological properties as well as responses to antileukemic treatments.2-4 Aberrations of the TP53 tumor suppressor gene have been described in a variable number of cases of AML. They are present in less than 10% of cases of de novo AML, whereas their rates in therapy-related AML and erythroid leukemias exceed 20% and 90%, respectively.5-7 These aberrations include gene mutations, most of which are located within the DNA binding domain of the gene, and/or deletions of different sizes affecting the TP53 locus on chromosome 17p13. Although the majority of haematologica | 2019; 104(3)


TP53 subclones in AML

TP53 aberrations are somatically acquired, constituting an early leukemogenic event, germline mutations are increasingly being recognized, predominantly in patients with therapy-related AML.8,9 Aberrations of TP53 are associated with an exceedingly adverse prognosis as demonstrated by several independent reports.5,10,11 Recently, it was shown that TP53 mutations and deletions encompassing the TP53 locus have a different prognostic impact in AML, with only mutations but not deletions significantly influencing survival of these patients.12 As a consequence, testing for TP53 mutations has been introduced into the 2017 recommendations of the European LeukemiaNet.13 However, in the studies performed so far, TP53 mutations were assessed as a dichotomous variable only. With the advent of next-generation sequencing technologies, mutational subclones can now be detected with high sensitivity. Here, we aimed to investigate the clinical characteristics associated with subclonal TP53 mutations and their prognostic impact in a large cohort of AML patients prospectively treated within studies of the “GermanAustrian AML Study Group” (AMLSG).

Kaplan-Meier method and compared by the log-rank test. In addition, TP53-mutated patients were further categorized according to their VAF (>40%, 20%-40%, <20%). This categorization into three different groups according to the TP53 VAF was based on different biological features regarding concomitant chromosomal aberrations as outlined in the “Results” section. It also allowed a comparison with previous reports on the impact of TP53 VAF in patients with myelodysplastic syndromes.18,19 Univariable and multivariable Cox regression analyses were performed to identify relevant prognostic factors. We assessed age, white blood cell count, cytogenetic risk group and type of AML (de novo, secondary or therapy-related) in addition to the TP53 status. Hazard ratios (HR) are presented along with their 95% CI. To compare patients’ characteristics among the groups defined by their TP53 status and VAF, we performed a Fisher exact test for categorical parameters and the Kruskal-Wallis or Mann-Whitney-U tests, respectively, for continuous parameters. If the overall test between the three TP53 VAF groups showed statistically significant differences (α=0.05), post-hoc tests were performed. Due to the multiple groups tested, we employed a Bonferroni correction and considered a P-value of <0.017 as statistically significant. All statistical analyses were conducted using R version 3.4.4 (https://www.r-project.org).

Methods Results The study was approved by the ethics committees of the University of Ulm, Germany, and the Medical University of Graz, Austria and conducted in accordance with REMARK guidelines (“REporting recommendations for tumor MARKer prognostic studies”).14

Study participants Data from a total of 1537 intensively treated AML patients enrolled in three prospective, multicenter, clinical trials of the AMLSG were analyzed.15-17 Details of these studies as well as a list of AMLSG investigators and centers are provided in the Online Supplementary Appendix.

Sequence analysis Genetic profiling of a total of 1,537 diagnostic AML specimens using a targeted sequencing approach with 111 genes associated with myeloid neoplasms was reported previously for this combined cohort.2 Sequencing data were deposited in the European Genome-Phenome Archive (www.ebi.ac.uk/ega, accession number EGAS00001000275) and retrieved for the present study. Methodological details are provided in the Online Supplementary Methods together with information on ultradeep sequencing used for the analysis of selected, sequential patients’ samples.

Statistical analysis The study was designed to assess differences in overall survival between AML patients exhibiting a TP53 wild-type status and those with subclonal TP53 mutations. Based on data from Papaemmanuil et al.2 an absolute reduction of 35% in 3-year overall survival compared to that of patients with a TP53 wild-type status should be detectable using a sample size of 574 patients assuming an overall survival rate of 55% for TP53 wild-type patients and a frequency of 5% for TP53 muted subclones with a variant allele frequency (VAF) as low as 5% (power=90%; α=0.05). The main outcome parameters assessed were overall survival and event-free survival, as defined by the European LeukemiaNet.13 We determined median survival times and estimated 3-year survival rates along with their 95% confidence intervals (95% CI). Survival rates of patients with a TP53 wild-type status and patients with TP53 mutations were plotted using the haematologica | 2019; 104(3)

Characteristics of TP53-mutated subclones in acute myeloid leukemia A total of 1,537 patients, enrolled in the AMLSG trials HD98A, HD98B and 07-04, were evaluated in this study. Their clinical characteristics are shown in Table 1. Of those patients, 1,408 (91.6%) had de novo AML, 61 (4.0%) had secondary AML following myelodysplastic syndromes and 68 (4.4%) had therapy-related AML. The median follow-up of all patients was 25 months (range, 0 - 219.6). When analyzing diagnostic AML specimens, 108 pathogenic TP53 mutations were found in 98 (6.4%) patients. Seven patients exhibited two TP53 mutations each and one patient had four. A detailed presentation of the mutations detected in this cohort is given in Online Supplementary Table S1. When categorizing the 98 patients with TP53 mutations according to their maximum VAF, we found 61 (62.2%) with a VAF >40% representing the major AML clone, 19 (19.4%) with a VAF between 20% and 40% and 18 (18.4%) with a VAF <20%, this last group representing subclones. The vast majority of TP53 mutations in all groups were missense mutations located in the DNA binding domain of the gene. Neither type of the mutation nor its location showed a statistically significant difference among the three TP53-mutated subgroups (P=0.279 for mutation type and P=0.687 for mutation location) (Figure 1). As compared to AML patients with TP53 wild-type, those with clonal and subclonal TP53 mutations showed significantly lower white blood cell counts as well as peripheral blood and bone marrow blasts (Table 1). It is well documented that clonal TP53 mutations are highly associated with complex karyotypes as well as marker chromosomes arising from chromothripsis.20,21 As shown in Table 2, a significant association with complex karyotypes was observed for TP53 mutations with a VAF of >40% and between 20% and 40%, but not for subclonal ones with a VAF of <20%. There was also a statistically significant difference between TP53 wild-type and TP53 517


K.T. Prochazka et al. Table 1. Patientsâ&#x20AC;&#x2122; characteristics.

Gender male female Age (median, years) [range] Type of AML de novo secondary therapy-related WBC (median, 109/L) [range] PB blasts (median, %) [range] BM blasts (median, %) [range] Cytogenetic risk low intermediate high ELN 2010 risk favorable intermediate-1 intermediate-2 high Complete remission Consolidation therapy chemotherapy autologous HSCT allogeneic HSCT Relapse

TP53 wild-type (N=1439)

TP53 mutated (N=98)

758 (52.7%) 681 (47.3%) 49.4 [18.3-84.5]

60 (61.2%) 38 (38.8%) 57.4 [20.0-78.8]

1326 (92.1%) 57 (4.0%) 56 (3.9%) 15.6 [0.2-532.7] 38.0 [0.0-100.0] 76.0 [0.0-100.0]

82 (83.7%) 4 (4.1%) 12 (12.2%) 4.8 [0.5-145.4] 19.0 [0.0-96.0] 60.0 [14.0-100.0]

202 (15.1%) 930 (69.9%) 199 (15.0%)

2 (2.4%) 11 (13.1%) 71 (84.5%)

421 (33.2%) 402 (31.7%) 258 (20.4%) 186 (14.7%) 1221 (84.9%) 998 (69.4%) 578 (57.9%) 87 (8.7%) 333 (33.4%) 584 (47.8%)

2 (2.4%) 7 (8.3%) 4 (4.8%) 71 (84.5%) 47 (48.0%) 33 (33.7%) 14 (42.4%) 1 (3.0%) 18 (54.5%) 32 (68.1%)

P-value

VAF<20% (N=18)

TP53 mutated VAF 20%-40% (N=19)

VAF>40% (N=61)

9 (50.0%) 9 (50.0%) 58.1 [20.0-73.3]

14 (73.7%) 5 (26.3%) 56.9 [28.4-76.4]

37 (60.7%) 24 (39.3%) 56.9 [28.2-78.8]

18 (100%) 0 (0.0%) 0 (0.0%) 3.0 [1.0-133.2] 5.0 [0.0-96.0] 46.0 [18.0-100.0]

14 (73.7%) 1 (5.3%) 4 (21.1%) 3.3 [0.5-43.3] 11.0 [0.0-83.0] 60.0 [14.0-90.0]

50 (82.0%) 3 (4.9%) 8 (13.1%) 7.7 [1.0-145.4] 31.0 [0.0-92.0] 69.0 [20.0-95.0]

0 (0.0%) 6 (42.9%) 8 (57.1%)

1 (6.2%) 0 (0.0%) 15 (93.8%)

1 (1.9%) 5 (9.3%) 48 (88.9%)

1 (7.1%) 4 (28.6%) 1 (7.1%) 8 (57.1%) 10 (55.6%) 6 (33.3%) 3 (50.0%) 0 (0.0%) 3 (50.0%) 5 (50.0%)

1 (6.2%) 0 (0.0%) 0 (0.0%) 15 (93.8%) 11 (57.9%) 8 (42.1%) 4 (50.0%) 1 (12.5%) 3 (37.5%) 9 (81.8%)

0 (0.0%) 3 (5.6%) 3 (5.6%) 48 (88.9%) 26 (42.6%) 19 (31.1%) 7 (36.8%) 0 (0.0%) 12 (63.2%) 18 (69.2%)

0.116

<0.001 0.002

<0.001 0.001 0.002 <0.001

<0.001

<0.001 <0.001

0.007

VAF: variant allele frequency; WBC: white blood cells; PB: peripheral blood; BM: bone marrow; ELN: European LeukemiaNet; HSCT: hematopoietic stem cell transplantation. Numbers presented are N (%) if not otherwise specified.

Figure 1. Distribution of 108 TP53 mutations found in diagnostic specimens of 98/1537 patients with acute myeloid leukemia. Top panel: TP53 mutations with a variant allele frequency (VAF) of >40%; middle panel: mutations with a VAF of 20%-40%; lower panel: mutations with a VAF <20%. Missense mutations are marked in black, nonsense in red, insertions/deletions in blue and essential splice site mutations in purple. Despite different VAF, the vast majority of TP53 mutations are missense mutations located within the DNA binding domain of the gene.

518

haematologica | 2019; 104(3)


TP53 subclones in AML Table 2. Chromosomal aberrations and cooperating gene mutations found in 98 acute myeloid leukemia patients with TP53 mutations according to their variant allele frequency.

Complex karyotype Marker chromosome -5/5q-7/7q-17/17pN. of losses, median [range] N. of gains, median [range] N. of cooperating mutations, median [range]

VAF<20%

VAF 20%-40%

VAF>40%

Overall P-value*

<20% vs. 20%-40%

Post-hoc tests <20% vs. >40%

20%-40% vs. >40%

7 (50.0%) 4 (28.6%) 7 (38.9%) 4 (22.2%) 1 (5.6%) 1.0 [0.0-8.0] 1.0 [0.0-6.0]

14 (87.5%) 11 (73.3%) 8 (42.1%) 6 (31.6%) 4 (21.1%) 4.0 [0.0-13.0] 4.5 [0.0-9.0]

46 (85.2%) 34 (64.2%) 31 (50.8%) 15 (24.6%) 16 (26.2%) 5.0 [0.0-13.0] 3.0 [0.0-9.0]

0.019 0.030 0.625 0.799 0.201 0.028 0.050

0.046 0.027

0.009 0.031

1.000 0.758

0.056

0.008

0.806

1.0 [0.0-4.0]

0.0 [0.0-3.0]

1.0 [0.0-5.0]

0.255

* Post-hoc tests were only performed if the overall test was significant at the 0.05 significance level. Due to multiple group comparisons in the post-hoc tests, a P-value of <0.017 was considered statistically significant therein.(a) With respect to marker chromosomes, data refer to a total of 82 analyzed, TP53 mutated cases (VAF <20%, n=14; 20%-40%, n=15; >40%, n=53).(b) The number (N.) of losses, gains and cooperating mutations refers to the total number of specific aberrations observed per sample.

mutated cases with respect to marker chromosomes (2.9% versus 59.8%; P<0.001), however, no significant difference was observed among the three TP53-mutated subgroups. Furthermore, in the group with subclonal TP53 mutations, a significantly lower number of chromosomal losses was detected as compared to the number in the group with clonal TP53 mutations. The frequency of the chromosomal aberrations -5/5q-, -7/7q- and -17/17p- did not differ significantly among the three TP53-mutated groups, nor did the total number of chromosomal gains. The number of cooperating driver mutations was low with zero to five mutations detected per sample and did not differ significantly between the TP53-mutated groups. Mutations in DNMT3A (14/98), NRAS (8/98), FLT3 (7/98) and PTPN11 (7/98) were the most frequent cooperating mutations (Figure 2).

Impact of TP53-mutated subclones on survival of patients with acute myeloid leukemia Of 1,537 patients undergoing induction therapy, 1,268 (82.5%) achieved complete remission. The complete remission rate was significantly inferior in the TP53mutated cohort (48.0% versus 84.9% for TP53 wild-type patients, P<0.001) but equally distributed among the VAFbased groups (TP53 VAF >40%, 42.6%; 20%-40%, 57.9%; <20%, 55.6%; P=0.424) (Online Supplementary Figure S1). A total of 1,031 patients (67.1%) underwent consolidation therapy and of those, 439 (42.6%) received hematopoietic stem cell transplantation (351 allogeneic and 88 autologous; 42.1% of consolidated TP53 wild-type patients and 57.6% of TP53-mutated patients) (Table 1). The relapse rate was significantly higher in the TP53mutated cohort (68.1% versus 47.8% for TP53 wild-type patients, P=0.007) but did not differ significantly among the TP53-mutated subgroups (TP53 VAF >40%, 69.2%; 20%-40%, 81.8%; <20%, 50.0%; P=0.336) (Table 1). Within the TP53-mutated cohort, patients who underwent allogeneic hematopoietic stem cell transplantation had a significantly better event-free survival than those given chemotherapy or autologous hematopoietic stem cell transplantation [HR, 0.25 (95% CI: 0.11-0.58); P=0.001]. No significant difference was observed for overall survival [HR, 0.47 (95% CI: 0.22-1.01); P=0.054]. haematologica | 2019; 104(3)

The estimated median overall survival for the entire cohort was 28.1 months (95% CI: 24.3-33.5) but differed substantially between TP53 wild-type and TP53-mutated patients [33.6 months (95% CI: 28.4-45.0) versus 6.5 months (95% CI: 5.0-8.2)]. The median overall survival was short in all TP53-mutated subgroups (TP53 VAF >40%, 5.8 months; 20%-40%, 6.9 months; <20%, 6.9 months). The estimated 3-year overall survival rate for the entire cohort was 46.5% (95% CI: 44.0-49.1) with notable differences between TP53 wild-type and TP53-mutated patients [49.1% (95% CI: 46.5-51.8) versus 8.3% (95% CI: 4.3-16.2)]. Again, 3-year overall survival rates were low in each of the TP53-mutated groups (TP53 VAF >40%, 11.5%; 20%-40%, 5.3%; <20%, 0%) (Figure 3A,B, Online Supplementary Table S2). The estimated median event-free survival for the entire cohort was 15.0 months (95% CI: 13.6-16.5) with that for TP53 wild-type patients being 16.5 months (95% CI: 15.018.2) and that for TP53-mutated patients being 5.7 months (95% CI: 4.3-7.4). The median event-free survival was short in all TP53-mutated subgroups (TP53 VAF >40%, 5.2 months; 20%-40%, 6.9 months; <20%, 6.5 months). The estimated 3-year event-free survival rate for the entire cohort was 36.3% (95% CI: 33.9-38.8) with a pronounced difference between TP53 wild-type and TP53-mutated patients [38.3% (35.9-40.9) versus 6.3% (2.9-13.6)]. As for overall survival, 3-year event-free survival rates were low in all TP53-mutated subgroups (TP53 VAF >40%, 9.8%; 20%-40%, 0%; <20%, 0%).(Figure 4A,B, Online Supplementary Table S3). Table 3 and Online Supplementary Table S4 present the results of the Cox regression analyses assessing the impact of TP53 mutational status on overall and event-free survival. The models contain the different VAF groups, age, white blood cell count, cytogenetic risk group and type of AML as predictors. Whereas all of the factors assessed showed a significant impact on both outcome parameters in the initial univariable analysis, type of AML did not remain significant for event-free survival in the multivariable model. Importantly, each of the TP53-mutated groups showed significantly worse overall and event-free survival compared to the TP53 wild-type group. For overall survival, the HR (95% CI), for the comparison with TP53 519


K.T. Prochazka et al.

Figure 2. Concurrent gene mutations in 98 TP53-mutated cases of acute myeloid leukemia divided by functional classes. Of those specimens with two or more TP53 mutations, only the one with the highest variant allele frequency (VAF) is listed. Each column represents an individual patient with the total number of cooperating events stated at the bottom. Each colored box represents a mutation of the gene listed at the left. The red vertical lines indicate the 40% and 20% cut-offs with respect to TP53 VAF. #: number.

A

B

Figure 3. Kaplan-Meier analysis of overall survival in 1,537 patients with acute myeloid leukemia stratified by TP53 mutational status. (A) Overall survival: TP53 wild-type patients versus TP53-mutated patients. (B) Overall survival: TP53 wild-type patients versus patients in the three groups with the defined variant allele frequencies of mutated TP53.

520

haematologica | 2019; 104(3)


TP53 subclones in AML

A

B

Figure 4. Kaplan-Meier analysis of event-free survival in 1537 patients with acute myeloid leukemia stratified by TP53 mutational status. (A) Event-free survival: TP53 wild-type patients versus TP53-mutated patients. (B) TP53 wild-type patients versus patients in the three groups with the defined variant allele frequencies of mutated TP53.

Table 3. Univariable and multivariable Cox regression analysis for overall survival.

HR TP53 wild-type VAF <20% VAF 20%-40% VAF >40% Age White blood cell count (log) Cytogenetic risk high intermediate low Type of AML de novo secondary therapy-related

Univariable 95% CI

P-value

HR

Multivariable 95% CI

<0.001 1 4.95 3.44 3.73 1.04 1.08

3.05 - 8.04 2.16 - 5.50 2.84 - 4.88 1.03 - 1.04 1.04 - 1.13

1 0.45 0.20

0.38 - 0.52 0.15 - 0.26

<0.001 <0.001 <0.001

<0.001 1 3.71 1.78 2.03 1.03 1.17

2.11 - 6.51 1.04 - 3.02 1.47 - 2.79 1.03 - 1.04 1.11 - 1.22

1 0.49 0.27

0.41 - 0.58 0.20 - 0.36

<0.001 1 1.74 1.43

<0.001 <0.001 <0.001

0.019 1 1.53 1.21

1.30 - 2.33 1.07 - 1.92

P-value

1.13 - 2.07 0.88 - 1.66

HR:hazard ratio; CI: confidence interval; VAF: variant allele frequency.

wild-type patients, was 2.03 (1.47-2.79) for patients with a TP53 VAF >40%, 1.78 (1.04-3.02) for those with a VAF of 20%-40% and 3.71 (2.11-6.51) for those with a VAF <20%. For event-free survival, the estimates were 1.89 (1.38-2.59) for patients with a TP53 VAF >40%, 2.27 (1.353.80) for those with a VAF of 20%-40% and 3.57 (2.046.26) for those with a VAF <20%.

Case study on longitudinal TP53 mutational assessment In one subject with secondary AML treated at the Medical University of Graz, Austria, outside a clinical trial, serial assessment of bone marrow specimens by error-corrected ultradeep sequencing was performed.22,23 This 68year old male presented with myelodysplastic syndrome and an International Prognostic Scoring System risk score of “intermediate 2”. Within 2 months after diagnosis, the patient’s disease transformed into secondary AML and he was then treated with daunorubicin and cytarabine (3+7) as well as Ida-FLAG as a salvage regimen. Nevertheless, the haematologica | 2019; 104(3)

patient died of resistant disease 4 months after diagnosis. Analysis of the diagnostic sample revealed the clonal TP53 p.273H mutation with a VAF of 42.3% and a subclonal TP53 Q104X mutation with a VAF of 0.4%. Both TP53mutated clones expanded slightly in the course of transformation to secondary AML which was also characterized by a novel KRAS Q61H mutation. However, during two courses of intensive chemotherapy, both TP53 mutations substantially expanded to a clone size of 64.2% and 18.6%, respectively. (Figure 5 and Online Supplementary Table S5).

Discussion In this study, we analyzed a large cohort of intensively treated AML patients focusing on biological and clinical characteristics associated with subclonal TP53 mutations. We found that these aberrations represent a substantial proportion of TP53-mutated AML. Similarly to clonal TP53 mutations, subclonal ones are mainly missense 521


K.T. Prochazka et al.

mutations and are associated with an adverse outcome as evidenced by significantly inferior complete remission, overall survival and event-free survival rates. Results similar to those presented here have been obtained for patients with chronic lymphocytic leukemia in whom those with TP53-mutated subclones with a median VAF as low as 2.1% showed comparable clinical phenotypes and survival as poor as those with clonal mutations.24,25 Interestingly, divergent data have been published for patients with myelodysplastic syndromes.18,19 Analyzing two independent cohorts of 219 and 150 patients, also including a few cases with secondary AML, Sallman et al. stratified TP53 VAF into the same categories as defined here (>40%, 20%-40%, <20%). Importantly, there was a significant difference in overall survival between patients exhibiting TP53 mutant VAF of >40% and <20%. Whereas myelodysplastic syndrome patients with a TP53 VAF >40% had a median overall survival of 124 days, overall survival was not reached in patients with a VAF <20%. In contrast, Goel et al. investigated two cohorts of 81 and 1,117 patients and did not find any significant difference with respect to TP53 VAF and their negative prognostic impact on overall survival. It was argued that one of the reasons for these discrepancies might have been the small number of TP53-mutated patients investigated. In addition, early therapeutic intervention leading to clonal expansion of a drug-resistant clone may also be relevant to these differences in survival.26 In our sufficiently powered study of 1,537 patients predominantly with de novo AML, 98 had TP53-mutated clones of various sizes. We provide evidence here that small TP53-mutated subclones share biological characteristics with clonal ones, such as mutation type, location of the mutation within TP53 domains and cooperating mutations. We also demonstrate that even TP53-mutated subclones, defined by a VAF <20%, have a statistically significant negative prognostic impact with respect to complete remission rate, overall survival and event-free survival. These findings may have implications for TP53 screening methods and future risk stratification in AML as classical Sanger sequencing with a detection limit of mutant clones set at 20% VAF should be replaced by high sensitivity next-generation sequencing approaches. Furthermore, when confirmed by others, subclonal TP53 mutations should be incorporated into future risk classification for AML. The mechanisms by which mutant p53 mediates resistance to cytotoxic treatments are poorly understood and possibly involve novel gain-of-function properties. TP53 mutations affect preleukemic stem cells in AML and it could be shown in vitro and in vivo that genotoxic stress leads to expansion of murine hematopoietic stem and progenitor cells exhibiting either p53 haploinsufficiency or expressing a heterozygous TP53 mutation.27-29 These data are in line with the AML case presented here, in which substantial expansion of both a clonal and a subclonal TP53 mutation was observed following high-dose chemotherapy but not during the transition from myelodysplastic syndrome to frank leukemia. Similar clinical data have recently been published for patients with lymphomas showing clonal hematopoiesis including TP53 mutations at diagnosis or before undergoing autologous hematopoietic stem cell transplantation. These individuals were at increased risk of clonal expansion and, ultimately, development of therapyrelated myeloid neoplasms following intensive, lymphomaspecific chemo- and radiotherapies.30-33 522

Figure 5. Longitudinal mutational analyses of a patient with secondary acute myeloid leukemia showing a clonal (R273H) and a subclonal (Q104X) TP53 mutation. VAF: variant allele frequency; MDS, myelodysplastic syndrome: sAML: secondary acute myeloid leukemia; PD, progressive diesease. Bottom line: cooperating mutations.

In our study we were able to demonstrate TP53-mutated clones with a VAF as low as 4.66% (Online Supplementary Table S1). However, recently developed next-generation sequencing techniques such as error-corrected ultradeep sequencing allow an unambiguous identificantion of mutant alleles with a frequency as low as 0.1%.34 In chronic lymphocytic leukemia, it has been demonstrated that even very small TP53-mutated clones constitute an adverse prognostic parameter.24,25 Given the mechanism of selection of such clones following genotoxic therapies as outlined above, it is very likely that a similar effect may also be operational in AML. Another issue is related to the assessment of VAF using clinical specimens. Specification of VAF refers either to mononuclear cell fractions or with a correction to represent exclusively neoplastic cells. However, it must be taken into account that mutations affecting preleukemic stem cells in AML are being propagated into mature blood cells.35 By analyzing highly purified T-lymphocytes of diagnostic AML specimens with TP53 mutations, we were able to demonstrate the specific aberration being present in up to 20% of these mature blood cells.27 Thus, analyzing mononuclear cell fractions may be the more comprehensive approach to assess VAF in AML and, possibly, other myeloid disorders. Finally, the diagnostic specimens in this study were derived from both bone marrow and peripheral blood (Online Supplementary Methods) raising the issue of comparability of these sources. Recently, Duncavage et al. demonstrated that the mutational landscape is conserved in peripheral blood of AML and myelodysplastic syndrome patients at diagnosis as well as during treatment.36 Similar data were obtained for TP53 mutations in patients with myelodysplastic syndromes.37 In this study we investigated the prognostic impact of subclonal TP53 mutations in intensively treated AML patients. Results may be somewhat different when assessing AML patients with TP53 mutations treated with hypomethylating agents such as decitabine. Although a substantial impact on overall survival could not be shown with that approach, similar response rates as those of AML patients with TP53 wild-type have been reported recently.38 It might, therefore, be necessary to prospectivehaematologica | 2019; 104(3)


TP53 subclones in AML

ly address the impact of subclonal TP53 mutations in such an AML cohort. In conclusion, we have shown that subclonal TP53 mutations represent an adverse prognostic marker in AML, independently of the clone size. Our data further corroborate the use of next-generation sequencing technologies for risk stratification of patients with AML. Future work will assess the clonal architecture of these specimens and investigate whether TP53-mutated subclones exhibit preleukemic/leukemic stem cell properties which may explain their persistence following cytotoxic therapies.

References 1. Dores GM, Devesa SS, Curtis RE, Linet MS, Morton LM. Acute leukemia incidence and patient survival among children and adults in the United States, 2001-2007. Blood. 2012;119(1):34-43. 2. Papaemmanuil E, Gerstung M, Bullinger L, et al. Genomic classification and prognosis in acute myeloid leukemia. N Engl J Med. 2016;374(23):2209-2221. 3. Thomas D, Majeti R. Biology and relevance of human acute myeloid leukemia stem cells. Blood. 2017;129(12):1577-1585. 4. Hirsch P, Zhang Y, Tang R, et al. Genetic hierarchy and temporal variegation in the clonal history of acute myeloid leukaemia. Nat Commun. 2016;7:12475. 5. Rucker FG, Schlenk RF, Bullinger L, et al. TP53 alterations in acute myeloid leukemia with complex karyotype correlate with specific copy number alterations, monosomal karyotype, and dismal outcome. Blood. 2012;119(9):2114-2121. 6. Andersen MK, Christiansen DH, PedersenBjergaard J. Centromeric breakage and highly rearranged chromosome derivatives associated with mutations of TP53 are common in therapy-related MDS and AML after therapy with alkylating agents: an M-FISH study. Genes Chromosomes Cancer. 2005;42(4):358-371. 7. Montalban-Bravo G, Benton CB, Wang SA, et al. More than 1 TP53 abnormality is a dominant characteristic of pure erythroid leukemia. Blood. 2017;129(18):2584-2587. 8. Schulz E, Valentin A, Ulz P, et al. Germline mutations in the DNA damage response genes BRCA1, BRCA2, BARD1 and TP53 in patients with therapy related myeloid neoplasms. J Med Genet. 2012;49(7):422-428. 9. Zebisch A, Lal R, Muller M, et al. Acute myeloid leukemia with TP53 germ line mutations. Blood. 2016;128(18):2270-2272. 10. Middeke JM, Herold S, Rucker-Braun E, et al. TP53 mutation in patients with high-risk acute myeloid leukaemia treated with allogeneic haematopoietic stem cell transplantation. Br J Haematol. 2016;172(6):914-922. 11. Grossmann V, Schnittger S, Kohlmann A, et al. A novel hierarchical prognostic model of AML solely based on molecular mutations. Blood. 2012;120(15):2963-2972. 12. Stengel A, Kern W, Haferlach T, Meggendorfer M, Fasan A, Haferlach C. The impact of TP53 mutations and TP53 deletions on survival varies between AML, ALL, MDS and CLL: an analysis of 3307 cases. Leukemia. 2017;31(3):705-711. 13. Dohner H, Estey E, Grimwade D, et al. Diagnosis and management of AML in adults: 2017 ELN recommendations from an

haematologica | 2019; 104(3)

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

Acknowledgments We thank Ms. Daniela Weber, Ulm, Germany, for excellent assistance and Dr. Karl Kashofer, Graz, Austria, for providing DNA for sequencing analysis. Funding This work was supported by the Austrian Science Fund FWF (P31430-B26 and P28949-B28), Leukämiehilfe Steiermark, DFG-Forschergruppe FOR 2674 ‚Aging-related epigenetic remodeling in acute myeloid leukemia’, subproject A02, and Österreichische Nationalbank (16917).

international expert panel. Blood. 2017;129 (4):424-447. McShane LM, Altman DG, Sauerbrei W, et al. REporting recommendations for tumor MARKer prognostic studies (REMARK). Breast Cancer Res Treat. 2006;100(2):229235. Schlenk RF, Dohner K, Krauter J, et al. Mutations and treatment outcome in cytogenetically normal acute myeloid leukemia. N Engl J Med. 2008;358(18):1909-1918. Schlenk RF, Frohling S, Hartmann F, et al. Phase III study of all-trans retinoic acid in previously untreated patients 61 years or older with acute myeloid leukemia. Leukemia. 2004;18(11):1798-1803. Schlenk RF, Dohner 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. Sallman DA, Komrokji R, Vaupel C, et al. Impact of TP53 mutation variant allele frequency on phenotype and outcomes in myelodysplastic syndromes. Leukemia. 2016;30(3):666-673. Goel S, Hall J, Pradhan K, et al. High prevalence and allele burden-independent prognostic importance of p53 mutations in an inner-city MDS/AML cohort. Leukemia. 2016;30(8):1793-1795. Rausch T, Jones DT, Zapatka M, et al. Genome sequencing of pediatric medulloblastoma links catastrophic DNA rearrangements with TP53 mutations. Cell. 2012;148(1-2):59-71. Bochtler T, Granzow M, Stolzel F, et al. Marker chromosomes can arise from chromothripsis and predict adverse prognosis in acute myeloid leukemia. Blood. 2017;129(10):1333-1342. Wolfler A, Erkeland SJ, Bodner C, et al. A functional single-nucleotide polymorphism of the G-CSF receptor gene predisposes individuals to high-risk myelodysplastic syndrome. Blood. 2005;105(9):3731-3736. Kinde I, Wu J, Papadopoulos N, Kinzler KW, Vogelstein B. Detection and quantification of rare mutations with massively parallel sequencing. Proc Natl Acad Sci U S A. 2011;108(23):9530-9535. Malcikova J, Stano-Kozubik K, Tichy B, et al. Detailed analysis of therapy-driven clonal evolution of TP53 mutations in chronic lymphocytic leukemia. Leukemia. 2015;29(4): 877-885. Rossi D, Khiabanian H, Spina V, et al. Clinical impact of small TP53 mutated subclones in chronic lymphocytic leukemia.

Blood. 2014;123(14):2139-2147. 26. Sallman DA, Komrokji R, List A, Padron E. Reply to Goel et al. 'TP53 mutation alleleburden and disease outcome in MDS/AML'. Leukemia. 2017;31(3):767-768. 27. Lal R, Lind K, Heitzer E, et al. Somatic TP53 mutations characterize preleukemic stem cells in acute myeloid leukemia. Blood. 2017;129(18):2587-2591. 28. Wong TN, Ramsingh G, Young AL, et al. Role of TP53 mutations in the origin and evolution of therapy-related acute myeloid leukaemia. Nature. 2015;518(7540):552555. 29. Chen S, Gao R, Yao C, et al. Genotoxic stresses promote clonal expansion of hematopoietic stem cells expressing mutant p53. Leukemia. 2018;32(3):850-854. 30. Schulz E, Kashofer K, Heitzer E, et al. Preexisting TP53 mutation in therapy-related acute myeloid leukemia. Ann Hematol. 2015;94(3):527-529. 31. Gillis NK, Ball M, Zhang Q, et al. Clonal haemopoiesis and therapy-related myeloid malignancies in elderly patients: a proof-ofconcept, case-control study. Lancet Oncol. 2017;18(1):112-121. 32. Takahashi K, Wang F, Kantarjian H, et al. Preleukaemic clonal haemopoiesis and risk of therapy-related myeloid neoplasms: a case-control study. Lancet Oncol. 2017;18(1):100-111. 33. Gibson CJ, Lindsley RC, Tchekmedyian V, et al. Clonal hematopoiesis associated with adverse outcomes after autologous stem-cell transplantation for lymphoma. J Clin Oncol. 2017;:JCO2016716712. 34. Stahlberg A, Krzyzanowski PM, Egyud M, Filges S, Stein L, Godfrey TE. Simple multiplexed PCR-based barcoding of DNA for ultrasensitive mutation detection by nextgeneration sequencing. Nat Protoc. 2017;12(4):664-682. 35. Reinisch A, Chan SM, Thomas D, Majeti R. Biology and clinical relevance of acute myeloid leukemia stem cells. Semin Hematol. 2015;52(3):150-164. 36. Duncavage EJ, Uy GL, Petti AA, et al. Mutational landscape and response are conserved in peripheral blood of AML and MDS patients during decitabine therapy. Blood. 2017;129(10):1397-1401. 37. Belickova M, Vesela J, Jonasova A, et al. TP53 mutation variant allele frequency is a potential predictor for clinical outcome of patients with lower-risk myelodysplastic syndromes. Oncotarget. 2016;7(24):3626636279. 38. Welch JS, Petti AA, Miller CA, et al. TP53 and decitabine in acute myeloid leukemia and myelodysplastic syndromes. N Engl J Med. 2016;375(21):2023-2036.

523


ARTICLE Ferrata Storti Foundation

Haematologica 2019 Volume 104(3):524-532

Acute Myeloid Leukemia

Haploidentical versus unrelated allogeneic stem cell transplantation for relapsed/ refractory acute myeloid leukemia: a report on 1578 patients from the Acute Leukemia Working Party of the EBMT

Eolia Brissot,1 Myriam Labopin,1,2 Gerhard Ehninger,3 Matthias Stelljes,4 Arne Brecht,5 Arnold Ganser,6 Johanna Tischer,7 Nicolaus Kröger,8 Boris Afanasyev,9 Jürgen Finke,10 Ahmet Elmaagacli,11 Herman Einsele,12 Mohamad Mohty1,2* and Arnon Nagler2,13*

Service d’Hématologie Clinique et de Thérapie Cellulaire, Hôpital Saint Antoine, APHP, Paris, France; 2Acute Leukemia Working Party office, Hôpital Saint Antoine, APHP, Paris, France; 3Universitaetsklinikum Dresden, Medizinische Klinik und Poliklinik I, Germany; 4 University of Münster, Department of Medicine A / Hematology and Oncology, Germany; 5 Deutsche Klinik für Diagnostik, KMT Zentrum, Wiesbaden, Germany; 6Hannover Medical School, Department of Haematology, Hemostasis, Oncology, and Stem Cell Transplantation, Germany; 7Klinikum Grosshadern, Med. Klinik III, Munich, Germany; 8 University Hospital Eppendorf, Bone Marrow Transplantation Centre, Hamburg, Germany; 9 First State Pavlov Medical University of St. Petersburg, Raisa Gorbacheva Memorial Research Institute for Pediatric Oncology, Hematology, and Transplantation, Russia; 10 University of Freiburg, Faculty of Medicine and Department of Medicine -Hematology, Oncology and Stem Cell Transplantation, Germany; 11Asklepios Klinik St. Georg, Department of Hematology, Hamburg, Germany; 12Universitaetsklinikum Würzburg, Med. Klinik und Poliklinik II, Germany and 13Hematology Division, Chaim Sheba Medical Center and Tel Aviv University, Tel-Hashomer, Ramat-Gan, Israel 1

* MM and AN contributed equally to this study and share senior authorship

ABSTRACT

Correspondence: EOLIA BRISSOT eolia.brissot@aphp.fr Received: February 4, 2018. Accepted: October 19, 2018. Pre-published: October 25, 2018. doi:10.3324/haematol.2017.187450 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/104/3/524 ©2019 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

524

P

rimary refractory or relapsed acute myeloid leukemia is associated with a dismal prognosis. Allogeneic stem cell transplantation is the only therapeutic option that offers prolonged survival and cure in this setting. In the absence of a matched sibling donor, transplantation from unrelated 10/10 HLA allele-matched or 9/10 HLA allele-mismatched donors and haploidentical donors are potential alternatives. The current study aimed to compare the outcomes of acute myeloid leukemia patients with active disease who received allogeneic stem cell transplantation from a haploidentical donor with post-transplant cyclophosphamide (n=199) versus an unrelated 10/10-matched donor (n=1111) and versus an unrelated 9/10-mismatched donor (n=383) between 2007 and 2014 and who were reported to the European Society for Blood and Marrow Transplantation registry. Propensity score weighted analysis was conducted in order to control for disease risk imbalances between the groups. The leukemia-free survival rates at 2 years of recipients of grafts from a haploidentical donor, an unrelated 10/10-matched donor and an unrelated 9/10-mismatched donor were 22.8%, 28% and 22.2%, respectively (P=NS). In multivariate analysis, there were no significant differences in leukemia-free survival, overall survival, relapse incidence, non-relapse mortality, or graft-versus-host-disease-free relapse-free survival between the three groups. Two predictive factors were associated with a higher relapse incidence: transplantation during first or second relapse compared to primary refractory acute myeloid leukemia and poor cytogenetics. Allogeneic stem cell transplantation may rescue about 25% of acute myeloid leukemia patients with active disease. Importantly, the outcomes of transplants from haploidentical donors were comparable to those from 10/10-matched and 9/10-mismatched unrelated donors. Therefore, a haploidentical donor is a valid option for acute myeloid leukemia patients with active disease.

haematologica | 2019; 104(3)


Haploidentical versus unrelated donor HSCT for active AML

Introduction

Methods

After initiation of intensive chemotherapy for acute myelogenous leukemia (AML), failure to respond is a major unfavorable prognostic factor.1,2 Obtaining a morphological complete remission (CR) after induction has been defined as a prognostic factor and even, until recently, considered as a prerequisite for allogeneic stem cell transplantation (HSCT). However, up to 30% of adults with newly diagnosed AML fail to achieve CR after two courses of intensive chemotherapy. 1 Moreover, once a first or second CR has been obtained, approximately half of younger patients and 80% of older patients relapse.1,3,4 In both clinical situations, refractory and/or relapsed AML, active disease remains a major therapeutic challenge. Consequently, the accumulating evidence that HSCT can deliver long-term disease-free survival in a proportion of patients with AML with active disease represents an importance advance in the treatment in this very high-risk patient population.59 Defining the impact of donor selection is still a major issue. It has been demonstrated that HSCT from matched sibling donors is a valid option, leading to a disease-free survival rate in the range of 20-30% for this very high-risk patient population.7,10-13 More recently, HSCT from unrelated donors (UD) was used for patients with primary refractory AML, with an overall survival (OS) rate of about 22%.13-15 Since 2010, the use of haploidentical HSCT has surged by about 300% among European Society of Blood and Marrow Transplantation (EBMT) centers.16,17 Indeed, over recent years, haploidentical donors have been increasingly adopted as a valid source of donor cells for HSCT in AML in the absence of HLA-compatible matched sibling donors or matched UD. Based on several non-randomized comparative studies evaluating HSCT from haploidentical donors (Haplo HSCT),18-21 the combined data suggest similar outcomes for Haplo and UD HSCT.22,23 However, only small series are available for patients with resistant and/or relapsed AML undergoing HSCT from alternative donors. Craddock et al. reported that the OS rate of 36 patients who received an UD transplant with a reduced-intensity conditioning regimen (RIC) was 36% at 5 years, which was similar to that of 18 patients given myeloablative conditioning (MAC).14 For patients with active leukemia, HSCT from alternative or mismatched donors may, theoretically, be of advantage, as HLA disparities may augment donor/recipient alloreactivity. However, relatively few data are available for the very high-risk population of patients with refractory or relapsing AML transplanted from alternative donors while in active disease. In view of the fact that the development of Haplo HSCT is significantly influenced by the use of post-transplantation cyclophosphamide (PTCy), and because advances in supportive care influence outcomes, a safety and efficacy update comparison between Haplo and UD HSCT in a large cohort of patients with active disease is highly warranted to further support decision-making. With this aim, the present study, based on the EBMT - Acute Leukemia Working Party (ALWP) database, was conducted in order to compare outcomes of AML patients with active disease after Haplo HSCT versus 10/10 or 9/10 HLA-matched UD HSCT.

Study design and data retrieval

haematologica | 2019; 104(3)

This is a retrospective, multicenter, registry-based analysis. Data for this study were provided and approved by the ALWP of the EBMT group registry. The EBMT registry is a voluntary working group of more than 600 transplant centers, mostly located in Europe, which are required to report all consecutive stem-cell transplantations and follow-up data once a year. Data are entered, managed, and maintained in a central database with internet access; each EBMT center is represented in this database. There are no restrictions on centers for reporting data, except for those required by law on patients’ consent, data confidentiality and accuracy. Quality control measures include several independent systems: confirmation of the validity of the entered data by the reporting team, selective comparison of the survey data with MED-A data sets in the EBMT registry database, cross-checking with national registries, and regular in-house and external data audits. All patients provided informed consent to the use of their data in retrospective studies. The Review Board of the ALWP as well as the ethics committee of the EBMT approved this study. Eligibility criteria for this analysis included adult patients (aged >18 years) with active AML including primary refractory AML, AML in first relapse and second relapse who had received a first HSCT from a 10/10 HLA allele-matched UD (UD 10/10), or a 9/10 HLA allele-mismatched UD (UD 9/10) or a haploidentical donor (≥2 antigen mismatches or more out of 8) with post-transplant cyclophosphamide (Haplo PTCy) as graft-versus-host disease (GvHD) prophylaxis. Active AML was defined by the failure to achieve CR (bone marrow blasts >5%) despite induction chemotherapy. Cytogenetic groups were defined according to Grimwade et al.24 The source of stem cells could be either bone marrow or granulocyte colony-stimulating factor-mobilized peripheral blood stem cells. All UD were HLA-matched (10/10) or mismatched at one loci (9/10) (-A, -B, -C, -DRB1, -DQB1). We excluded patients who had undergone umbilical cord blood HSCT, so that the analysis was restricted to a more homogeneous study population. MAC was defined, according to the EBMT, as a regimen containing total body irradiation with a dose >6 Gy, a total dose of oral busulfan >8 mg/kg or a total dose of intravenous busulfan >6.4 mg/kg.25 The FLAMSA sequential conditioning regimen consisted of a combination of a short course of intensive chemotherapy using fludarabine 30 mg/m2/day, intermediate-dose cytosine arabinoside 2 g/m2/day, and amsacrine 100 mg/m2/day from day –12 to –9, followed, after a 3-day rest, by RIC using 4 Gy total body irradiation on day –5, cyclophosphamide 40 to 60 mg/kg/day on days –4 and –3, and antithymocyte globulin from days –4 to –2; the 4 Gy total body irradiation could have been replaced by a total dose of intravenous busulfan of 6.4 mg/kg (or an equivalent oral dose).26-28

Endpoints OS was calculated from the date of transplantation until death or last observation alive. Leukemia-free survival (LFS) was calculated from the date of transplantation until relapse or last diseasefree follow-up. Relapse and death from any cause were considered events. Non-relapse mortality (NRM) was defined as death without prior relapse. Neutrophil recovery was defined as achieving absolute neutrophil counts greater than or equal to 0.5×109/L for three consecutive days. The diagnosis and grading of acute29 and chronic GvHD30 were performed by transplant centers using the standard criteria. Cytogenetic abnormalities were classified according to Medical Research Centre criteria, and graft-versushost-free, relapse-free survival (GRFS) as previously published.31

525


E. Brissot et al.

Statistical analysis Patient-, disease-, and transplant-related variables were compared between the three groups (Haplo PTCy, UD 10/10, UD 9/10) using the chi-square test for categorical variables and the Mann-Whitney test for continuous variables. The median followup was estimated using the reverse Kaplan-Meier method. Variables considered were patient’s age at transplantation, donor/recipient gender, interval from diagnosis to transplantation, cytogenetic group, type of conditioning (RIC/MAC/FLAMSA), source of stem cells (peripheral blood stem cells versus bone marrow), patient/donor cytomegalovirus serology, Karnofsky Performance Status (KPS) at the time of transplantation, in vivo Tcell depletion, and year of transplantation. Factors that significantly differed between the three groups with P-values of <0.05, and all known as potential prognostic factors were included in the final models. Cumulative incidence functions were used to estimate relapse incidence (RI) and NRM in a competing risk setting, because death and relapse compete with each other. To study chronic GvHD, we considered relapse and death to be competing events. Probabilities of LFS and OS were calculated using KaplanMeier estimates. Univariate analyses were performed using the Gray test for cumulative incidence functions and the log-rank test for LFS and OS. Associations of patient and graft characteristics with outcomes were evaluated in multivariate analysis, using a Cox proportional hazards model. All tests were two-sided. We used propensity score weighing to control for pre-treatment imbalances in observed variables. The following factors were included in the propensity score model: patient’s age, time from diagnosis to transplantation, year of transplant, status at transplant, cytogenetic group, donor/patient cytomegalovirus serology, conditioning (RIC versus MAC), and gender matching (female donor to male recipient versus other). Propensity scores were estimated using generalized boosted models. As the study question was whether Haplo PTCy could replace UD 10/10 or UD 9/10, we weighted the groups receiving either UD 10/10 or UD 9/10 HSCT to match the characteristics of patients receiving Haplo HSCT, by estimating the average treatment effect among the treated group (Haplo HSCT being the treated group). We then used pairwise average treatments to fit the weighted Kaplan-Meier and Cox models separately for Haplo PTCy versus UD 10/10 HSCT and Haplo PTCy versus UD 9/10 HSCT. The type I error rate was fixed at 0.05 for determination of factors associated with time to events. Analyses were performed using the R statistical software version 3.2.3 (R Development Core Team, Vienna, Austria). Propensity score analysis was performed using the mnps function of the Twang package and weighted analyses using the survey package.

MAC regimens in the Haplo PTCy group in comparison to both the UD 10/10 and the UD 9/10 groups. There were more cytomegalovirus-positive recipient-donor pairs in the Haplo-PTCy group. Peripheral blood stem cells were, as expected, the main source of stem cells in the UD 10/10 and UD 9/10 groups, while peripheral blood stem cells represented 52.8% of the stem cell source in the HaploPTCy group.

Engraftment and graft-versus-host disease The cumulative incidence of engraftment at day 30 was 85.5% [95% confidence interval (95% CI): 79-90.2], 92.3% (95% CI: 90.5-93.7) and 92.2% (95% CI: 88.9-94.6) in the Haplo PTCy, UD 10/10 and 9/10 groups, respectively (P<10-3 for both comparisons). Lower incidences of all acute GvHD grades were observed after Haplo PTCy than after UD 9/10. The cumulative incidences of grade II–IV acute GvHD were 28.2% and 36.3%, respectively (P=0.03) and those of severe grade III–IV acute GvHD were 8.9% and 16.1%, respectively (P=0.02) (Table 3). No difference was observed in the incidence of grade II-IV acute GvHD between the Haplo PTCy and UD 10/10 groups (P=NS). At 2 years, the cumulative incidence of chronic GvHD was lower in the Haplo PTCy group than in the UD 9/10 group (19.3% and 27.4%, respectively, P=0.04), while no difference was found in the incidence of chronic GvHD between the Haplo PTCy and UD 10/10 groups (P=NS) (Table 3). The cumulative incidence of extensive chronic GvHD was similar in the three groups of patients (Haplo PTCy - 11%, UD 10/10 -11.6% and UD 9/10 -11.6%). The percentages of patients who achieved CR within 100 days were 79.7%, 77% and 78.3% in the Haplo PTCy, UD 10/10 and UD 9/10 groups, respectively (P=NS). Performing a landmark analysis at day 100 for comparing outcomes between patients who achieved CR before day 100 to those who did not indicated that CR seems to be a surrogate marker for subsequent outcome. The probability of being alive and free of disease 1 year after HSCT was only 13.7% for patients who did not achieve CR before day 100 (data not shown). We also analyzed chronic GvHD as a time-dependent variable demonstrating that the association of chronic GvHD with a lower RI [hazard ratio (HR)=0.77, 95% CI: 0.60-0.99, P=0.04) was counterbalanced by a higher NRM (HR=1.98, 95% CI: 1.38-2.84, P<10-3), and thus did not translate into better LFS (P=NS).

Leukemia-free survival, overall survival, relapse incidence and non-relapse mortality Results Patients, disease and transplant characteristics Data were obtained from 218 reporting centers (Online Supplementary Data). The patients’ and disease characteristics are summarized in Table 1. Of the total 1693 HSCT, 1111 were UD 10/10, 383 were UD 9/10 and 199 were Haplo PTCy. The three cohorts of patients differed for several variables (Table 1). The median follow-up was longer for the UD 10/10 and the UD 9/10 groups than for the Haplo-PTCy group. The follow-up completeness index at 2 years, which is the ratio of total observed person-time and the potential person-time of follow-up at 2 years32 was 73% for Haplo PTCy, 76% for UD 10/10 and 80% for UD 9/10. Significantly more patients received 526

In univariate analysis, the LFS rate at 2 years was 22.8% in the Haplo PTCy group versus 28% in the UD 10/10 group and 22.2% in the UD 9/10 group (P=NS) (Figure 1A, Online Supplementary Table S1). Multivariate analysis showed lower LFS rates in patients with poor cytogenetics (HR=1.35, 95% CI: 1.11-1.62, P=0.002), those transplanted in second relapse in comparison to those transplanted in primary refractory AML (HR=1.31, 95% CI, 1.03-1.65, P=0.03), and in patients transplanted from cytomegalovirus seropositive donors (HR=1.22, 95% CI: 1.05-1.41, P=0.01). In contrast, better LFS was associated with KPS ≥90 at transplantation (HR=0.66, 95% CI: 0.580.76, P<10-3), with a shorter time from diagnosis to transplantation (HR=0.99, 95% CI: 0.98-0.99, P=0.02), and with RIC in comparison to MAC (HR=0.84, 95% CI: 0.71haematologica | 2019; 104(3)


Haploidentical versus unrelated donor HSCT for active AML

0.99, P=0.03). Of note, no effect was observed for donor type (Table 4). The OS rate at 2 years did not differ between the three groups of patients (29.3% in the Haplo PTCy group versus 34.7% in the UD 10/10 group and 27.6% in the UD 9/10 group, P=NS) (Figure 1B). These results were confirmed by multivariate analysis. In the latter, three predictive fac-

tors were associated with lower OS: disease status (second relapse versus primary refractory AML), poor cytogenetics and the patient being positive for cytomegalovirus, whereas KPS ≥90 at transplant, RIC versus MAC and shorter time from diagnosis to transplantation were associated with a better OS (Table 4). We did not find any differences in terms of RI between

Table 1. Baseline characteristics of patients.

Number

383

Haplo versus UD 10/10 P value

Haplo versus UD 9/10 P value

18.1 (0.6- 113) 52.4 (18.1-77.3) 2011 (2007-2015) 6.8 (2-474.8)

22.9 (1.8 - 104.9) 51.7 (18-76) 2011(2007-2015) 8.1 (2.1-121.4)

0.05 NS <10-3 NS

0.02 NS <10-3 NS

501 (45.1) 491 (44.2) 119 (10.7)

141 (36.8) 190 (49.6) 52 (13.6)

NS

NS

40 (10.2) 235 (59.8) 118 (30) 718

15 (9.3) 98 (60.9) 48 (29.8) 222

NS

NS

412 (49) 619 (61)

143 (39.7) 217 (60.3)

0.01

0.03

587 (52.9) 523 (47.1)

209 (54.6) 174 (45.4)

NS

NS

769 (72.8) 287 (27.2)

247 (66) 127 (34 )

<10-3

NS

925 (87.7) 130 (12.3)

313 (83.7) 61 (16.3)

<10-3

NS

308 (29.1) 81 (7.6) 305 (28.8) 366 (34.5)

89 (24.2) 38 (10.3) 128 (34.9) 112 (30.5)

<10-3

<10-3

465 (41.9) 380 (34.3) 263 (23.7)

144 (37.7) 128 (33.5) 110 (28.8)

<10-3

<10-3

72 (6.5) 1039 (93.5)

30 (7.8) 353 (92.2)

<10-3

<10-3

325 (30) 484 (44.6) 54 (5) 20 (2) 25 (2.3)

113 (29.7) 161 (42.4) 27 (7.1) 11 (2.9) 7 (2)

<10-3

<10-3

265 (24.2) 830 (75.8)

54 (14.2) 327 (85.8)

<10-3

<10-3

Haplo PTCy

UD 10/10

UD 9/10

199

1111

Follow-up in months, median (range) 16 (2.1 - 63.5) Age at transplant in years, median (range) 51.9 (18.2-77.8) Year of transplant in years, median (range) 2014 (2009-2015) Time from diagnosis in transplant in years 7.6 (2-122.1) in months (range) Status of AML, n (%) Primary refractory 82 (41.2) 1st relapse 87 (43.7) 2nd relapse 30 (15.1) Cytogenetics, n (%) Favorable 7 (5.6) Intermediate 83 (65.9) Adverse 36 (28.6) Unknown/failed 73 KPS at transplant, n (%) <90% 91 (49.7) ≥90% 92 (50.3) Patients’ gender, n (%) Male 108 (54.3) Female 91 (45.7) Donors’ gender, n (%) Male 119 (59.8) Female 80 (40.2) Female D to male R, n (%) No 155 (77.9) Yes 44 (22.1) CMV status, n (%) D-/R30 (15.7) D+/R9 (4.71) D-/R+ 39 (20.4) D+/R+ 113 (59.2) Conditioning regimen, n (%) Myeloablative 106 (53.5) Reduced intensity 81 (40.9) Sequential strategy 11 (5.6) Source of stem cells, n (%) Bone marrow 94 (47.2) Peripheral blood 105 (52.7) GvHD prophylaxis, n (%) CSA+MTX 0 CSA+MMF 0 Tacrolimus+MMF 0 CSA+MMF+MTX 0 PTCy 199 (100) In vivo T-cell depletion, n (%) No 188 (94.5) Yes 11 (5.5)

AML: acute myeloid leukemia; BM: bone marrow; CsA: cyclosporine; D: donor; GvHD: graft-versus-host disease; Haplo: haplo-identical; KPS: Karnofsky Performance Status; MMF: mycophenolate mofetil; MTX: methotrexate; NS: not significant; PTCy: post-transplant cyclophosphamide; R: recipient; UD: unrelated donor.

haematologica | 2019; 104(3)

527


E. Brissot et al.

the three groups (Table 2, Figure 1C). This result was also confirmed in multivariate analysis showing that poor cytogenetics and disease status (first and second relapse versus primary refractory AML) were the only risk factors associated with increased RI, whereas KPS, age at transplantation and time from diagnosis to transplantation were protective

factors (HR=0.76, 95% CI: 0.64-0.90, P=0.001; HR=0.92, 95% CI: 0.87-0.98, P=0.007; and HR=0.98, 95% CI: 0.970.99, P<10-3, respectively) (Table 4). No differences in NRM were noted between the three groups of patients in univariate analysis (Table 2, Figure 1D). Multivariate analysis demonstrated that patientsâ&#x20AC;&#x2122; age

Table 2. Transplantation outcomes.

Leukemia-free survival Overall survival Relapse incidence Non-relapse mortality GRFS

Haplo PTCy

UD 10/10

UD 9/10

P value

P value Haplo versus UD 10/10

P value Haplo versus UD 9/10

22.8% (16.3-29.2) 29.3% (22.1-36.6) 52% (44.3-59.1) 25.3% (19.2-31.8) 16.3% (10.6-21.9)

28% (25-30.9) 34.7% (31.5-37.8) 46.3% (43.1-49.4) 25.7% (23.1-28.5) 16.4% (14-18.8)

22.2% (17.6-26.7) 27.6% (22.7-32.5) 51.1% (45.7-56.3) 26.7% (22.2-31.4) 16% (12.1-19.9)

NS NS NS NS NS

NS NS NS NS NS

NS NS NS NS NS

Data are presented as percentage with 95% confidence intervals in brackets. GRFS: graft-versus-host disease-free, relapse-free survival; haplo: haploidentical; PTCy: post-transplant cyclophosphamide; NS: not significant; UD: unrelated donors.

A

B

C

D

Figure 1. Leukemia-free survival, overall survival, relapse incidence and non-relapse mortality in patients with acute myeloid leukemia allografted during active disease. (A) The 2-year probability of leukemia-free survival (LFS) was 22.8% (95% CI: 16.3-29.2) in the group transplanted from a haploindentical donor with posttransplant cyclophosphamide (Haplo) versus 28% (95% CI: 25-30.9) in the 10/10 HLA-matched unrelated donor group (UD 10/10), and 22.2% (95% CI: 17.6-26.7) in the 9/10 HLA-mismatched unrelated donor group (UD 9/10) (P=NS). (B) The 2-year probability of overall survival (OS) was 29.3% (95% CI: 22.1-36.6) in the Haplo group versus 34.7% (95% CI: 31.5-37.8) in the UD 10/10 and 27.6% (95% CI: 22.7-32.5) in the UD 9/10 groups (P=NS). (C) The 2-year cumulative incidence of relapse (RI) was 52% (95% CI: 44.3-59.1) in the Haplo group versus 46.3% (95% CI: 43.1-49.4) in the UD 10/10 and 51.1% (95% CI: 45.7-56.3) in the UD 9/10 groups (P=NS). (D) The 2-year cumulative incidence of non-relapse mortality (NRM) was 25.3% (95% CI: 19.2-31.8) in the Haplo group versus 25.7% (95% CI: 23.128.5) in the UD 10/10 and 26.7% (95% CI: 22.2-31.4) in the UD 9/10 groups (P=NS).

528

haematologica | 2019; 104(3)


Haploidentical versus unrelated donor HSCT for active AML

(per 10 years) and cytomegalovirus positivity were associated with higher NRM (HR=1.18, 95% CI: 1.08-1.28, P<10-3; and HR=1.37, 95% CI: 1.07-1.77, P=0.01, respectively), while RIC compared to MAC and KPS â&#x2030;Ľ90 were associated with lower NRM (HR=0.64, 95% CI: 0.49-0.84, P=0.001; and HR=0.52, 95% CI: 0.41-0.64, P<10-3; respectively) (Table 4). Notably, no effect was observed for the type of donor. In addition, no significant differences were found in GRFS according to donor type in the multivariate analysis. Three factors were associated with a better GRFS: longer time from diagnosis to transplantation, RIC versus MAC, and a KPS â&#x2030;Ľ90. Patients with poor cytogenetics had a lower GRFS (Table 4). As shown in Online Supplementary Table S2, most events happened within the first year after HSCT. Finally, in order to reduce the effects of confounding factors, we performed a weighted analysis on propensity scores (weighted average treatment). The results of the weighted Kaplan-Meier and Cox analyses confirmed the study results as described in Table 5. In the weighted analysis on propensity scores, the Haplo PTCy group had a significantly lower incidence of grade III-IV acute GvHD compared to that of patients in the UD 10/10 group (8.9% versus 14.5%, respectively, P=0.04), as confirmed by Cox analysis (P=0.049).

Causes of death Leukemia was the most common cause of death (accounting for 50% of the deaths in the Haplo PTCy group, 54% in the UD 10/10 group, and 54.5% in the UD 9/10 group). GvHD was the second most common, being the cause of death in 11.5% of the patients in the Haplo PTCy group, 12.1% in the UD 10/10 group, and 15.2% in the UD 9/10 group. Infection was the cause of death in 27%, 20.8%, and 20.6% of the patients in the Haplo PTCy, UD10/10 and UD 9/10 groups, respectively.

Discussion In the present study, we compared the transplantation outcomes after Haplo HSCT with PTCy versus transplantation from matched (10/10) or mismatched (9/10) UD in AML patients with active disease. The LFS rate was about 25% and the OS rate about 30% after HSCT for this highrisk population with advanced disease with no difference between Haplo PTCy and UD 10/10 or 9/10. Rates of acute GvHD grades II-IV and of chronic GVHD were similar between Haplo PTCy and UD 10/10, and the same

held true for the 2-year NRM. It is important to note that higher rates of grades II-IV acute GvHD and chronic GvHD were observed in the UD 9/10 group without there being an impact on RI. Although we could hypothesize a stronger graft-versus-leukemia effect after Haplo PTCy than after UD HSCT, we did not observe differences in terms of RI or LFS. This finding is in accordance with those of our previous study, in which we compared relapse rates between patients with primary refractory AML undergoing allogeneic transplantation from unrelated versus sibling donors and found no difference.13 One hypothesis is that, besides the very strong tolerance induction mediated by the PTCy in the Haplo setting in the case of active leukemia, the very aggressive biology of the disease and its refractoriness to several lines of chemotherapy lead to fast disease progression and relapse early after transplantaion. Thus, the graft-versus-leukemia effect, even across broad HLA disparities, is too weak or too slow to control the leukemia. Of note, about 37% of patients received a RIC regimen in our study. As expected, the NRM rate was significantly lower in the RIC group than in the MAC group, with no difference in RI between the two groups. In all, LFS, OS, and GRFS were significantly better after RIC than after MAC. We could hypothesize that this is because even a high intensity conditioning regimen does not have a strong impact on this chemo-refractory leukemia. In our current study, 384 of the patients received a sequential approach with aplasia-inducing chemotherapy followed by the conditioning regimen. However, the question of which treatment should be used in a given patient cannot yet be answered. In a recent meta-analysis of leukemic patients with induction failure, Wattad et al. concluded that HSCT without prior salvage chemotherapy and HSCT in CR after salvage therapy resulted in comparable survival outcomes, and both strategies were significantly superior to HSCT performed after failure of salvage therapy.33 In the present study, no differences in outcome were found between patients who received MAC or sequential regimens. One hypothesis to explain this is that the refractoriness of the malignant leukemic clone to chemotherapy makes the conditioning regimen unable to induce remission, or even a transient response allowing sufficient time for the alloreactive cells to mediate the graft-versusleukemia effect.34 Importantly, an interval from diagnosis to transplant longer than the median was a negative prognostic factor for LFS, OS, RI and GRFS in multivariate analysis. These data, which are consistent with those of a study by Craddock et al.14 and our previous results,13 highlight the

Table 3. Univariate analysis for acute graft-versus-host disease and chronic graft-versus-host disease.

Acute GvHD II-IV Acute GvHD III-IV Chronic GvHD Extensive chronic GvHD PMN day 30

Haplo PTCy

UD 10/10

UD 9/10

P value

P value Haplo PTCy versus UD 10/10

P value Haplo PTCy versus UD 9/10

28.2% (21.8-34.9) 8.9% (5.3-13.7) 19.3% (13.6-25.7) 11% (6.7-16.4) 85.5% (79-90.2)

30.6% (27.8-33.4) 14% (11.9-16.2) 25.6% (22.7-28.6) 11.6% (9.6-13.9) 92.3% (90.5-93.7)

36.3% (31.3-41.2) 16.1% (12.5-20.1) 27.4% (22.6-32.4) 11.6% (8.3-15.4) 92.2% (88.9-94.6)

0.04 NS NS NS <10-3

NS NS NS NS <10-3

0.03 0.02 0.04 NS <10-3

Data are presented as percentages with 95% confidence intervals in brackets. ext: extensive; GvHD: graft-versus-host disease; haplo: haploidentical; PTCy: post-transplant; PMN: polymorphonuclear neutrophil; PTCy: post-transplant cyclophosphamide; UD: unrelated donors.

haematologica | 2019; 104(3)

529


E. Brissot et al. Table 4. Multivariate analysis for leukemia-free survival, overall survival, relapse incidence, non-relapse mortality and graft-versus-host disease-free, relapse-free survival.

HR

CI

P value

Leukemia-free survival UD 10/10 versus Haplo UD 9/10 versus Haplo Disease status: PRF-AML 1st relapse versus PRF-AML 2nd relapse versus PRF-AML Time from diagnosis to transplant Age at transplant (per 10 years) Year of transplant Conditioning regimen (ref=MAC) RIC versus MAC Sequential strategy versus MAC Poor cytogenetics Karnofsky Performance Score ≥90 Stem cell source: PBS versus BM Previous autologous transplant Female to male ratio Patient CMV positive Donor CMV positive Center (frailty)

0.97 1.05 1 1.10 1.31 0.99 1.00 1.00 1 0.84 0.97 1.35 0.66 0.97 0.95 0.89 1.22 0.99

0.75 - 1.24 0.80 - 1.37

NS NS

0.95 - 1.28 1.03 - 1.65 0.98 - 0.99 0.95 - 1.05 0.97 - 1.03

NS 0.03 0.02 NS NS

0.71 - 0.99 0.8 - 1.17 1.11- 1.62 0.58 - 0.76 0.78 - 1.21 0.68 - 1.33 0.74 - 1.06 1.05 - 1.41 0.87 - 1.14

0.03 NS 0.002 <10-3 NS NS NS 0.01 NS <10-3

Overall survival UD 10/10 versus Haplo UD 9/10 versus Haplo Disease status: PRF-AML 1st relapse versus PRF-AML 2nd relapse versus PRF-AML Time from diagnosis to transplant Age at transplant (per 10 years) Year of transplant Conditioning regimen (ref=MAC) RIC versus MAC Sequential strategy versus MAC Poor cytogenetics Karnofsky Performance Score ≥90 Stem cell source: PBS versus BM Previous autologous transplant Female to male ratio Patient CMV positive Donor CMV positive Center (frailty)

0.98 1.05 1 1.14 1.35 0.99 1.04 1.00 1 0.74 0.91 1.28 0.62 0.97 0.94 0.90 1.27 0.96

0.75 - 1.27 0.79 - 1.39

NS NS

0.97 - 1.33 1.05 - 1.72 0.99 - 1.00 0.99 - 1.10 0.97 - 1.03

NS 0.018 0.02 NS NS

0.62 - 0.88 0.75 - 1.12 1.06 - 1.56 0.54 - 0.71 0.77 - 1.22 0.66 - 1.35 0.75 - 1.09 1.09 - 1.49 0.83 - 1.11

-3

<10 NS 0.01 <10-3 NS NS NS 0.002 NS <10-3

0.67 - 1.25 0.74 - 1.47

NS NS

1.09 - 1.60 1.21 - 2.21 0.97 - 0.99 0.87 - 0.98 0.94 - 1.01

0.005 0.001 <10-3 0.007 NS

0.79 - 1.19

NS

Relapse incidence UD 10/10 versus Haplo UD 9/10 versus Haplo Disease status: PRF-AML 1st relapse versus PRF-AML 2nd relapse versus PRF-AML Time from diagnosis to transplant Age at transplant (per 10 years) Year of transplant Conditioning regimen (ref=MAC) RIC versus MAC

0.92 1.03 1 1.32 1.64 0.98 0.92 0.98 1 0.97

continued in the next coloum

530

continued from the previous coloum

HR

CI

P value

1.07 1.48 0.76 0.90 0.99 0.89 1.15 0.97

0.85 - 1.36 1.18 - 1.85 0.64 - 0.90 0.69 - 1.18 0.65 - 1.54 0.71 - 1.12 0.96 - 1.38 0.82 - 1.15

NS <10-3 0.001 NS NS NS NS NS <10-3

Non-relapse mortality UD 10/10 versus Haplo UD 9/10 versus Haplo Disease status: PRF-AML 1st relapse versus PRF-AML 2nd relapse versus PRF-AML Time from diagnosis to transplant Age at transplant (per 10 years) Year of transplant Conditioning regimen (ref=MAC) RIC versus MAC Sequential strategy versus MAC Poor cytogenetics Karnofsky Performance Score ≥90 Stem cell source: PBS versus BM Previous autologous transplant Female to male Patient CMV positive Donor CMV positive Center (frailty)

1.01 1.03 1 0.88 1.03 1.00 1.18 1.04 1 0.64 0.80 1.10 0.52 1.19 0.99 0.87 1.37 1.03

0.67 - 1.52 0.66 - 1.61

NS NS

0.70 - 1.12 0.71 - 1.50 0.99 - 1.01 1.08 - 1.28 0.99 - 1.09

NS NS NS <10-3 NS

0.49 - 0.84 0.59 - 1.09 0.79 - 1.55 0.41 - 0.64 0.81 - 1.75 0.57 - 1.72 0.63 - 1.18 1.07 - 1.77 0.83 - 1.29

0.001 NS NS <10-3 NS NS NS 0.01 NS NS

GRFS UD 10/10 versus Haplo UD 9/10 versus Haplo Disease status: PRF-AML 1st relapse versus PRF-AML 2nd relapse versus PRF-AML Time from diagnosis to transplant Age at transplant (per 10 years) Year of transplant Conditioning regimen (ref=MAC) RIC versus MAC Sequential strategy versus MAC Poor cytogenetics Karnofsky Performance Score ≥90 Stem cell source: PB versus BM Previous autologous transplant Female to male ratio Patient CMV positive Donor CMV positive Center (frailty)

1.09 1.11 1 1.05 1.22 0.99 0.98 1.02 1 0.86 0.93 1.20 0.68 1.15 0.95 0.85 1.14 0.98

0.87 - 1.39 0.86 - 1.43

NS NS

0.91 - 1.21 0.98 - 1.52 0.98 - 0.99 0.93 - 1.02 0.99- 1.05

NS NS 0.02 NS NS

0.74 - 0.99 0.78 - 1.11 1.00 - 1.44 0.60 - 0.77 0.94 - 1.42 0.69 - 1.31 0.71 - 1.01 0.99 - 1.31 0.86 - 1.12

0.047 NS 0.046 <10-3 NS NS NS NS NS 0.01

Sequential strategy versus MAC Poor cytogenetics Karnofsky Performance Score ≥90 Stem cell source: PBS versus BM Previous autologous transplant Female to male ratio Patient CMV positive Donor CMV positive Center (frailty)

AML: acute myeloid leukemia; BM: bone marrow; CI: confidence interval; CMV: cytomegalovirus; haplo: haploidentical; HR: hazard ratio; MAC: myeloablative conditioning; NS: not significant; PB: peripheral blood PRF-AML; PRF: primary refractory acute myeloid leukemia; Rel: relapse; RIC: reduced intensity conditioning; UD: unrelated donor.


Haploidentical versus unrelated donor HSCT for active AML

Table 5. Weighted Cox model for leukemia-free survival, overall survival, relapse incidence, non-relapse mortality, graft-versus-host disease-free, relapse-free survival and graft-versus-host disease.

Leukemia-free survival Overall survival Relapse incidence Non-relapse mortality GRFS Acute GvHD II-IV Acute GvHD III-IV Chronic GvHD Extensive chronic GvHD

Haplo PTCy

UD 10/10

UD 9/10

P value Haplo PTCy versus UD 10/10

P value Haplo PTCy versus UD 9/10

22.8% (17.1-30.2) 29.3% (22.9-37.5) 52% (43.9-58.8) 25.3% (18.7-31.3) 16.2% (11.5-22.9) 28.2% (21.3-34.5) 8.9% (4.6-13) 19.3% (13-25.1) 11.3% (6.2-16.1)

25.6% (19.4-33.7) 32.2% (25.4-40.8) 48.2% (43.6-52.5) 26.2% (22.3-29.9) 17.1% (11.9-24.4) 31.3% (27-35.3) 14.5% (11.4-17.4) 22.9% (19.1-26.6) 11.2% (8.5-13.8)

21.6% (15-30.9) 25.4% (18.3-35.4) 53.2% (45.6-59.7) 25.2% (19.3-30.6) 16.1% (10.5-24.7) 40.4% (32.8-47.2) 20.1% (13.5-26.2) 25.5% (19.1-31.4) 11.1% (6.6-15.3)

NS NS NS NS NS NS 0.04 NS NS

NS NS NS NS NS 0.01 0.005 0.04 NS

Data are presented as percentages with 95% confidence intervals in brackets; GvHD: graft-versus-host disease; GRFS: graft-versus-host disease-free, relapse-free survival; Haplo: haploidentical; not significant; PTCy: post-transplant cyclophosphamide; UD: unrelated donors.

urgent need to search for a donor for AML patients with active disease who lack a matched sibling donor. Haploidentical donors are available for the majority of patients, providing access to further stem cell donations or donor lymphocyte infusions as needed.35 Furthermore, it is a factor on which the physicians can have an influence, unlike many other factors. In accordance, Wattad et al. showed that patients transplanted during refractory disease after salvage therapy had a significantly poorer outcome compared to that of patients who proceeded directly to transplantation, and those transplanted in CR after salvage therapy.12 Thus, we could recommend not to try additional lines of chemotherapy to achieve CR in patients with active disease, but to take advantage of the first available donor in order to proceed to transplantation. The cytogenetic characterization of the leukemia represents a major prognostic factor for LFS, RI, OS, and GRFS.36,37 We previously reported that poor-risk cytogenetics was an adverse pre-HSCT variable in patients with primary refractory AML who underwent HSCT with a graft from a matched sibling or matched UD.13 Accordingly, in the present study, primary refractory AML with poor cytogenetic characteristics was associated, at 2 years, with a significant decrease in LFS and OS, and increase in RI. These data pave the road for investigating additional approaches relying on sequential conditioning regimens and/or post-transplant treatments.38-42 Being retrospective and registry-based, this study has some limitations: several of the patientsâ&#x20AC;&#x2122; characteristics differed between the groups. We addressed this limitation, at least in part, by using the propensity score technique. In addition, there was a relative inherent selection process for HSCT in our study and a relative lack of infor-

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.

haematologica | 2019; 104(3)

mation on the reasons for an EBMT center allocating patients to HSCT from a haploidentical donor versus UD, so that distinguishing the choice of the donor from the role of a potential center effect is difficult. Finally, the counts of circulating and bone marrow blasts at the time of HSCT were missing for a substantial number of patients. However, the aim of this analysis was to compare the two types of donors using EBMT registry data. The design of the study and inclusion criteria were intended to answer this clinical question and were not, therefore, adapted for developing a prognostic score. There are ongoing trials aiming to compare outcomes after Haplo PTCy versus UD in hematologic malignancies, but they do not focus specifically on the setting of active AML disease (NCT02623309). Therefore, in the absence of any prospect of such comparative studies, our data suggest that haploidentical donors are equally effective as 10/10 matched and 9/10 mismatched UD for allogeneic transplantation in patients with active AML. In conclusion, our results indicate that, when an HLAidentical sibling donor is not available for an AML patient with active disease who is, otherwise, a candidate for HSCT, a haploidentical donor may be used with the expectation of similar rates of NRM, LFS, OS, and GRFS at 2 years, compared with those achieved with 10/10 matched and 9/10 mismatched UD. Acknowledgments We thank all the European Society for Blood and Marrow Transplantation (EBMT) participating centers and national registries for providing patientsâ&#x20AC;&#x2122; data for the study, and data managers for their valuable contribution (Online Supplementary Data). Further information is available at the EBMT web site.

2. Thol F, Schlenk RF, Heuser M, Ganser A. How I treat refractory and early relapsed acute myeloid leukemia. Blood. 2015;126 (3):319-327. 3. Schlenk RF, Frech P, Weber D, et al. Impact of pretreatment characteristics and salvage strategy on outcome in patients with relapsed acute myeloid leukemia. Leukemia. 2017;31(5):1217-1220.

4. Dombret H, Gardin C. An update of current treatments for adult acute myeloid leukemia. Blood. 2016;127(1):53-61. 5. Revesz D, Chelghoum Y, Le QH, Elhamri M, Michallet M, Thomas X. Salvage by timed sequential chemotherapy in primary resistant acute myeloid leukemia: analysis of prognostic factors. Ann Hematol. 2003;82 (11):684-690.

531


E. Brissot et al. 6. Duval M, Klein JP, He W, et al. Hematopoietic stem-cell transplantation for acute leukemia in relapse or primary induction failure. J Clin Oncol. 2010;28(23):37303738. 7. Nagler A, Savani BN, Labopin M, et al. Outcomes after use of two standard ablative regimens in patients with refractory acute myeloid leukaemia: a retrospective, multicentre, registry analysis. Lancet Haematol. 2015;2(9):e384-392. 8. Gyurkocza B, Lazarus HM, Giralt S. Allogeneic hematopoietic cell transplantation in patients with AML not achieving remission: potentially curative therapy. Bone Marrow Transplant. 2017;52(8):1083-1090. 9. Dohner H, Weisdorf DJ, Bloomfield CD. acute myeloid leukemia. N Engl J Med. 2015;373(12):1136-1152. 10. Fung HC, Stein A, Slovak M, et al. A longterm follow-up report on allogeneic stem cell transplantation for patients with primary refractory acute myelogenous leukemia: impact of cytogenetic characteristics on transplantation outcome. Biol Blood Marrow Transplant. 2003;9(12):766-771. 11. Michallet M, Thomas X, Vernant JP, et al. Long-term outcome after allogeneic hematopoietic stem cell transplantation for advanced stage acute myeloblastic leukemia: a retrospective study of 379 patients reported to the Societe Francaise de Greffe de Moelle (SFGM). Bone Marrow Transplant. 2000;26(11):1157-1163. 12. Wattad M, Weber D, Dohner K, et al. Impact of salvage regimens on response and overall survival in acute myeloid leukemia with induction failure. Leukemia. 2017;31(6):1306-1313. 13. Brissot E, Labopin M, Stelljes M, et al. Comparison of matched sibling donors versus unrelated donors in allogeneic stem cell transplantation for primary refractory acute myeloid leukemia: a study on behalf of the Acute Leukemia Working Party of the EBMT. J Hematol Oncol. 2017;10(1):130. 14. Craddock C, Labopin M, Pillai S, et al. Factors predicting outcome after unrelated donor stem cell transplantation in primary refractory acute myeloid leukaemia. Leukemia. 2011;25(5):808-813. 15. Jabbour E, Daver N, Champlin R, et al. Allogeneic stem cell transplantation as initial salvage for patients with acute myeloid leukemia refractory to high-dose cytarabinebased induction chemotherapy. Am J Hematol. 2014;89(4):395-398. 16. Lee CJ, Savani BN, Mohty M, et al. Haploidentical hematopoietic cell transplantation for adult acute myeloid leukemia: a position statement from the Acute Leukemia Working Party of the European Society for Blood and Marrow Transplantation. Haematologica. 2017;102 (11):1810-1822. 17. Passweg JR, Baldomero H, Bader P, et al. Use of haploidentical stem cell transplantation continues to increase: the 2015 European Society for Blood and Marrow Transplant activity survey report. Bone Marrow Transplant. 2017;52(6):811-817. 18. Bashey A, Zhang X, Sizemore CA, et al. Tcell-replete HLA-haploidentical hematopoietic transplantation for hematologic malignancies using post-transplantation

532

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

cyclophosphamide results in outcomes equivalent to those of contemporaneous HLA-matched related and unrelated donor transplantation. J Clin Oncol. 2013;31(10): 1310-1316. Ciurea SO, Zhang MJ, Bacigalupo AA, et al. Haploidentical transplant with posttransplant cyclophosphamide vs matched unrelated donor transplant for acute myeloid leukemia. Blood. 2015;126(8):1033-1040. Bashey A, Zhang X, Jackson K, et al. Comparison of outcomes of hematopoietic cell transplants from T-replete haploidentical donors using post-transplantation cyclophosphamide with 10 of 10 HLA-A, -B, -C, -DRB1, and -DQB1 allele-matched unrelated donors and HLA-identical sibling donors: a multivariable analysis including disease risk index. Biol Blood Marrow Transplant. 2016;22(1):125-133. Rashidi A, DiPersio JF, Westervelt P, et al. Comparison of outcomes after peripheral blood haploidentical versus matched unrelated donor allogeneic hematopoietic cell transplantation in patients with acute myeloid leukemia: a retrospective singlecenter review. Biol Blood Marrow Transplant. 2016;22(9):1696-1701. Piemontese S, Ciceri F, Labopin M, et al. A comparison between allogeneic stem cell transplantation from unmanipulated haploidentical and unrelated donors in acute leukemia. J Hematol Oncol. 2017;10(1):24. Versluis J, In 't Hout FE, Devillier R, et al. Comparative value of post-remission treatment in cytogenetically normal AML subclassified by NPM1 and FLT3-ITD allelic ratio. Leukemia. 2017;31(1):26-33. Grimwade D, Hills RK, Moorman AV, et al. Refinement of cytogenetic classification in acute myeloid leukemia: determination of prognostic significance of rare recurring chromosomal abnormalities among 5876 younger adult patients treated in the United Kingdom Medical Research Council trials. Blood. 2010;116(3):354-365. Bacigalupo A, Ballen K, Rizzo D, et al. Defining the intensity of conditioning regimens: working definitions. Biol Blood Marrow Transplant. 2009;15(12):1628-1633. Schmid C, Schleuning M, Ledderose G, Tischer J, Kolb HJ. Sequential regimen of chemotherapy, reduced-intensity conditioning for allogeneic stem-cell transplantation, and prophylactic donor lymphocyte transfusion in high-risk acute myeloid leukemia and myelodysplastic syndrome. J Clin Oncol. 2005;23(24):5675-5687. Malard F, Labopin M, Stuhler G, et al. Sequential intensified conditioning Regimen allogeneic hematopoietic stem cell transplantation in adult patients with intermediate- or high-risk acute myeloid leukemia in complete remission: a study from the Acute Leukemia Working Party of the European Group for Blood and Marrow Transplantation. Biol Blood Marrow Transplant. 2017;23(2):278-284. Ringden O, Labopin M, Schmid C, et al. Sequential chemotherapy followed by reduced-intensity conditioning and allogeneic haematopoietic stem cell transplantation in adult patients with relapse or refractory acute myeloid leukaemia: a survey from the Acute Leukaemia Working Party of

EBMT. Br J Haematol. 2017;176(3):431-439. 29. Glucksberg H, Storb R, Fefer A, et al. Clinical manifestations of graft-versus-host disease in human recipients of marrow from HL-Amatched sibling donors. Transplantation. 1974;18(4):295-304. 30. Terwey TH, Le Duc TM, Hemmati PG, et al. NIH-defined graft-versus-host disease and evidence for a potent graft-versus-leukemia effect in patients with acute lymphoblastic leukemia. Ann Oncol. 2013;24(5):1363-1370. 31. Ruggeri A, Labopin M, Ciceri F, Mohty M, Nagler A. Definition of GvHD-free, relapsefree survival for registry-based studies: an ALWP-EBMT analysis on patients with AML in remission. Bone Marrow Transplant. 2016;51(4):610-611. 32. Clark TG, Altman DG, De Stavola BL. Quantification of the completeness of follow-up. Lancet. 2002;359(9314):1309-1310. 33. Wattad M, Weber D, Dohner K, et al. Impact of salvage regimens on response and overall survival in acute myeloid leukemia with induction failure. Leukemia. 2017;31 (6):1306-1313. 34. Raanani P, Dazzi F, Sohal J, et al. The rate and kinetics of molecular response to donor leucocyte transfusions in chronic myeloid leukaemia patients treated for relapse after allogeneic bone marrow transplantation. Br J Haematol. 1997;99(4):945-950. 35. Kanakry CG, Fuchs EJ, Luznik L. Modern approaches to HLA-haploidentical blood or marrow transplantation. Nat Rev Clin Oncol. 2016;13(2):132. 36. Byrd JC, Mrozek K, Dodge RK, et al. Pretreatment cytogenetic abnormalities are predictive of induction success, cumulative incidence of relapse, and overall survival in adult patients with de novo acute myeloid leukemia: results from Cancer and Leukemia Group B (CALGB 8461). Blood. 2002;100(13):4325-4336. 37. Grimwade D, Hills RK, Moorman AV, et al. Refinement of cytogenetic classification in acute myeloid leukemia: determination of prognostic significance of rare recurring chromosomal abnormalities among 5876 younger adult patients treated in the United Kingdom Medical Research Council trials. Blood. 2010;116(3):354-365. 38. Tsirigotis P, Byrne M, Schmid C, et al. Relapse of AML after hematopoietic stem cell transplantation: methods of monitoring and preventive strategies. A review from the ALWP of the EBMT. Bone Marrow Transplant. 2016;51(11):1431-1438. 39. Stein EM, DiNardo CD, Pollyea DA, et al. Enasidenib in mutant IDH2 relapsed or refractory acute myeloid leukemia. Blood. 2017;130(6):722-731. 40. Kayser S, Levis MJ, Schlenk RF. Midostaurin treatment in FLT3-mutated acute myeloid leukemia and systemic mastocytosis. Expert Rev Clin Pharmacol. 2017;10(11):1177-1189. 41. Antar A, Otrock ZK, El-Cheikh J, et al. Inhibition of FLT3 in AML: a focus on sorafenib. Bone Marrow Transplant. 2017;52 (3):344-351. 42. Konopleva M, Pollyea DA, Potluri J, et al. Efficacy and biological correlates of response in a phase II study of venetoclax monotherapy in patients with acute myelogenous leukemia. Cancer Discov. 2016;6(10):11061117.

haematologica | 2019; 104(3)


ARTICLE

Acute Lymphoblastic Leukemia

ActivinA: a new leukemia-promoting factor conferring migratory advantage to B-cell precursor-acute lymphoblastic leukemic cells

Federica Portale,1 Giulia Cricrì,1 Silvia Bresolin,2 Monica Lupi,3 Stefania Gaspari,4.5 Daniela Silvestri,6,7 Barbara Russo,1 Noemi Marino,1 Paolo Ubezio,3 Fabio Pagni,8 Patrizia Vergani,9 Geertruy Te Kronnie,2 Maria Grazia Valsecchi,6 Franco Locatelli,4 Carmelo Rizzari,7 Andrea Biondi,1,7 Erica Dander1* and Giovanna D’Amico1* *ED and GD’A contributed equally to this work

Centro Ricerca Tettamanti, Department of Pediatrics, University of Milano-Bicocca, Fondazione MBBM, Monza; 2Department of Women’s and Children’s Health, University of Padova; 3Department of Oncology, Istituto di Ricerche Farmacologiche Mario Negri IRCCS, Milano; 4Department of Paediatric Haematology-Oncology and Cell and Gene Therapy, IRCCS Ospedale Pediatrico Bambino Gesù and Sapienza University of Rome; 5 Medical Statistics Unit, Department of Clinical Medicine and Prevention, University of Milano-Bicocca, Monza; 6Medical Statistics Unit, Department of Clinical Medicine and Prevention, University of Milano-Bicocca; 7School of Medicine and Surgery, University of Milano-Bicocca, Monza; 8School of Medicine and Surgery, University of Milano-Bicocca and 9Department of Obstetrics and Gynecology, University of Milano-Bicocca, Monza, Italy 1

Ferrata Storti Foundation

Haematologica 2019 Volume 104(3):533-545

ABSTRACT

B

-cell precursor-acute lymphoblastic leukemia modulates the bone marrow (BM) niche to become leukemia-supporting and chemoprotective by reprogramming the stromal microenvironment. New therapies targeting the interplay between leukemia and stroma can help improve disease outcome. We identified ActivinA, a TGF-b family member with a well-described role in promoting several solid malignancies, as a factor favoring leukemia that could represent a new potential target for therapy. ActivinA resulted over-expressed in the leukemic BM and its production was strongly induced in mesenchymal stromal cells after culture with leukemic cells. Moreover, MSCs isolated from BM of leukemic patients showed an intrinsic ability to secrete higher amounts of ActivinA compared to their normal counterparts. The pro-inflammatory leukemic BM microenvironment synergized with leukemic cells to induce stromal-derived ActivinA. Gene expression analysis of ActivinAtreated leukemic cells showed that this protein was able to significantly influence motility-associated pathways. Interestingly, ActivinA promoted random motility and CXCL12-driven migration of leukemic cells, even at suboptimal chemokine concentrations, characterizing the leukemic niche. Conversely, ActivinA severely impaired CXCL12induced migration of healthy CD34+ cells. This opposite effect can be explained by the ability of ActivinA to increase intracellular calcium only in leukemic cells, boosting cytoskeleton dynamics through a higher rate of actin polymerization. Moreover, by stimulating the invasiveness of the leukemic cells, ActivinA was found to be a leukemia-promoting factor. Importantly, the ability of ActivinA to enhance BM engraftment and the metastatic potential of leukemic cells was confirmed in a xenograft mouse model of the disease. Overall, ActivinA was seen to be a key factor in conferring a migratory advantage to leukemic cells over healthy hematopoiesis within the leukemic niche. Introduction Acute lymphoblastic leukemia (ALL) is the most frequent childhood malignancy worldwide. B-cell precursor (BCP)-ALL represents about 80% of ALL cases and mainly affects children, with an incidence of 3-4 cases per 100,000 each year.1 Even haematologica | 2019; 104(3)

Correspondence: GIOVANNA D’AMICO giovanna.damico@hsgerardo.org Received: January 24, 2018. Accepted: September 21, 2018. Pre-published: September 27, 2018. doi:10.3324/haematol.2018.188664 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/104/3/533 ©2019 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

533


F. Portale et al.

though the cure rate exceeds 80% in children, BCP-ALL is the leading cause of cancer-related death in children and young adults.2 In spite of the notable improvements in disease management, the emergence of chemoresistance decreases the probability that therapy will be successful, and leads to relapse in more than 20% of treated patients.3 BCP-ALL cells critically depend on interactions with the bone marrow (BM) microenvironment, which provides essential regulatory cues for proliferation, survival and drug resistance, and such interactions contribute to treatment failure and disease relapse.4 In particular, mesenchymal stromal cells (MSCs) have been recognized as an essential supportive element of the leukemic hematopoietic microenvironment because of their ability to define exclusive BM niches that sustain leukemic cells to the detriment of normal hematopoiesis and resist chemotherapy.5 In this complex network, it has been shown that chemokines could contribute to BCP-ALL development by driving the migration of leukemic cells toward protective BM niches, as well as by providing anti-apoptotic signals.6 ActivinA is a pleiotropic cytokine that belongs to the TGF-b superfamily. It has a broad tissue distribution, being involved in multiple physiological and pathological processes, including inflammation, metabolism, immune response, and endocrine function. Recent studies have demonstrated that ActivinA is an important regulator of carcinogenesis. Indeed, it can directly modulate cancer cell proliferation and migration. It can also enhance tumor progression by regulating the tumor microenvironment.7 ActivinA sends signals through its transmembrane serine/threonine kinase receptors. It binds to type II Activin receptors (ACVR2A or ACVR2B), causing recruitment, phosphorylation and activation of type I Activin receptors (ALK2 or ALK4). ActivinA signaling is inhibited by Inhibins, through competitive binding for Activin receptors, and by Follistatin (FST) and Follistatin like-3 (FSTL3), which act as trap molecules.8 The Activin receptor II ligand trap ACE-011 is currently under investigation in a Phase II clinical trial on multiple myeloma.9 The aim of the current study was to explore the role of ActivinA in the leukemic BM niche, with a particular focus on its supportive role for BCP-ALL cells to the detriment of healthy hematopoiesis.

Methods Patients’ and healthy donors’ samples Bone marrow plasma samples were collected from 125 BCPALL patients at diagnosis and from 56 healthy donors (HDs). Primary BCP-ALL cells were isolated at diagnosis from 22 BM aspirates and used for in vitro assays. Details of the study cohort are shown in the Online Supplementary Appendix. The study was approved by the Institutional Review Board (AIEOP-BFM ALL 2009 protocol; EudraCT-2007-004270-43).

Culture of B-cell precursor-acute lymphoblastic leukemia cell lines The leukemic cell lines 697, NALM-6, RS4;11, SUP-B15 and REH were cultured as described in the Online Supplementary Appendix.

Isolation of bone marrow-mesenchymal stromal cells Bone marrow-mesenchymal stromal cells (BM-MSCs) from 15 534

BCP-ALL patients (ALL-MSCs) and 15 age-matched HDs (HDMSCs) (Online Supplementary Table S1) were cultured as described in the Online Supplementary Appendix.

Cord blood-CD34+ and bone marrow-CD34+ cell isolation CD34+ cells were isolated from cord blood (CB) units and HD BM aspirates, as described in the Online Supplementary Appendix.

Co-culture of primary leukemic cells with bone marrow-mesenchymal stromal cells Bone marrow-mesenchymal stromal cells were co-cultured with primary leukemic cells at an MSC:leukemia ratio of 1:30, either in the presence or in the absence of 0.4 mm Transwell inserts (Costar Transwell® Permable Supports, Corning Inc., MA, USA) in DMEM 2% FBS. After 72 hours (h), supernatants were collected and cryopreserved at -80°C for further analyses.

ELISA assay for quantification of ActivinA, CXCL12 and pro-inflammatory cytokines The levels of ActivinA, pro-inflammatory cytokines (IL-1b, IL-6 and TNF-α) and CXCL12 were assessed in BM plasma samples and culture supernatants using commercially available ELISA kits (R&D Systems, USA), according to the manufacturer’s instructions.

Quantitative RT-PCR

qRT-PCR was performed using LightCycler® 480 (Roche, Basel, Switzerland), as reported in the Online Supplementary Appendix.

Gene expression profile analysis Gene expression profile analysis of 697 cells treated or not with ActivinA (50 ng/mL) for 6 h and 24 h was evaluated by GeneChip Human Genome U133 Plus 2.0 arrays (Affymetrix Inc., Santa Clara, CA, USA). Details of the procedure are described in the Online Supplementary Appendix.

Time-lapse microscopy Leukemic cells were seeded in Gelatin B-coated wells of an 8well chamber slide (Ibidi, Martinsried, Germany). Cell tracks were recorded as described in the Online Supplementary Appendix.

Chemotaxis assays Both leukemic and healthy CD34+ cells, pretreated or not with ActivinA (24 h stimulation), were tested for chemotaxis in Transwell-based assays (Costar Transwell® Permable Supports, Corning, MA, USA). Selected experiments were performed using the ALK4 blocker SB431542 (SigmaAldrich, St Louis, MO, USA). Details are described in the Online Supplementary Appendix.

Invasion assays The invasive capacity of leukemic cells was evaluated using Matrigel-coated Transwells, as described in the Online Supplementary Appendix.

Activin receptor analyses 697 leukemic cell line and primary blasts were analyzed for ALK4, ACVR2A and ACVR2B expression by flow cytometry and qRT-PCR and for ALK2 expression by western blot, as described in the Online Supplementary Appendix.

CXCR4 and CXCR7 staining Expression of CXCR4 and CXCR7 was analyzed by flow cytometry in 697 cells, pretreated or not with ActivinA (Online Supplementary Appendix). haematologica | 2019; 104(3)


ActivinA enhances BCP-ALL cell migration

A

B

Filamentous (F)-actin polymerization assay F-actin polymerization was analyzed in 697 cells pretreated or not with ActivinA using AlexaFluor 647-labeled phalloidin (Invitrogen, Carlsbad, CA, USA) before and after CXCL12 stimulation, as reported in the Online Supplementary Appendix.

Calcium mobilization Intracellular calcium mobilization was measured by flow cytometry using the Fluo-4NW Assay (Invitrogen), as reported in the Online Supplementary Appendix.

B-cell acute lymphoblastic leukemia xenograft model Female 7-9-week-old NOD-SCID-Îłchain-/- (NSG) mice (Charles River, Calco, Italy) were intravenously (i.v.) transplanted with 0.5x106 697 cells or 106 NALM-6 cells, either pretreated or not with ActivinA for 24 h. Details are described in the Online Supplementary Appendix. The study was approved by the Italian Ministry of Health (approval n. 64/2014).

Statistical analyses Differences between subgroups were compared with the Mann-Whitney test or Wilcoxon matched-pairs signed rank test in the case of matched values.

Results Stroma-derived ActivinA increased in response to leukemia It has been demonstrated that ActivinA could exert a pro-tumoral role in several types of cancer both through direct effects on tumoral cells and indirect effects on the haematologica | 2019; 104(3)

Figure 1. B-cell precursor-acute lymphoblastic leukemia (BCP-ALL) cells reprogrammed the bone marrow (BM) stroma to produce high levels of ActivinA. (A) BM plasma levels of ActivinA were assessed by ELISA in healthy donors (HDs) (n=44) and BCP-ALL patients at the onset of the disease (n=108). Each box plot shows the median and the mean (+) and extends from the lowest to the highest value. *P<0.05: Mann-Whitney test. (B) Primary leukemic blasts were either directly co-cultured or separated by a 0.4 mm Transwell insert with HD-mesenchymal stromal cells (MSCs). After 72 hours of culture, supernatants were collected and ActivinA concentration was analyzed by ELISA (n=17 independent co-cultures). Each box plot shows the median and extends from the lowest to the highest value. **P<0.01; ***P<0.001; ****P<0.0001: Wilcoxon matched-pairs signed rank test.

tumor microenvironment.7 Therefore, we first measured ActivinA levels in BM plasma samples from 44 HDs and 108 BCP-ALL patients at disease onset. ELISA assay revealed that ActivinA was significantly increased in the BM plasma of BCP-ALL patients compared to that of HDs (Figure 1A). The median concentration (mc) of ActivinA was 400 pg/mL (range: 62.5-7241 pg/mL) in BCP-ALL patients and 273.4 pg/mL (range: 62.5-2338 pg/mL) in HDs (P<0.05). To determine whether ActivinA plasma levels at BCPALL diagnosis were related to disease outcome/severity, we analyzed 98 patients with available follow up out of the 108 tested, considering several clinical and biological parameters. In detail, 3- and 4-year event-free survival (EFS) and sensitive quantitative PCR-based minimal residual disease (MRD) at days +33 and +78 and final risk were taken into account.10 ActivinA levels did not impact on EFS, on PCR-MRD risk, or on patientsâ&#x20AC;&#x2122; final risk stratification (data not shown). It had previously been shown that BM-MSCs exhibit a basal level of ActivinA secretion.11 In view of the pivotal role of BM-MSCs in sustaining the leukemic niche,1 we explored the regulation of MSC-derived ActivinA in the context of BCP-ALL. We first confirmed that MSCs isolated from the BM of HDs (HD-MSCs) were able to constitutively produce the molecule (mc: 103.2, range: 62.5526.7 pg/mL). Then, to test whether BCP-ALL cells could modulate MSC-derived ActivinA, we set up co-culture experiments of HD-MSCs with primary leukemic blasts and quantified ActivinA in supernatants after 72 h of co-culture. Interestingly, we found that primary leukemic cells significantly induced ActivinA in MSCs both through 535


F. Portale et al. A

B

Figure 2. B-cell precursor-acute lymphoblastic leukemia (BCP-ALL) cells expressed type I and type II ActivinA receptors. (A) The expression of the type I ActivinA receptors ALK2 and ALK4 and the type II ActivinA receptors ACVR2A and ACVR2B was quantified in five leukemic cell lines (697, NALM-6, RS4;11, SUP-B15, REH) and (B) in bone marrow (BM) primary blasts from BCP-ALL patients by western blot (ALK2, n=12) and flow cytometry (ALK4, ACVR2A, ACVR2B, n=18). Data are presented as the ratio of ALK2 to b-actin obtained by densitometric analysis and mean fluorescence intensity (MFI) for ALK4, ACVR2A and ACVR2B.

soluble factors (mc: 268.1, range: 62.5-842.8 pg/mL) and a cell-to-cell contact-mediated mechanism (mc: 777.9, range: 96.2-1456 pg/mL), with a 2.6 and a 7-fold increase, respectively, compared to the basal condition (P<0.001 and P<0.0001). Notably, primary BCP-ALL cells secreted either very low or even undetectable levels of ActivinA (Figure 1B).

Leukemic cells expressed ActivinA receptors To determine whether BCP-ALL cells could be targets of ActivinA, the expression of Activin receptors was assessed on five leukemic cell lines (697, NALM-6, RS4;11, SUPB15, REH) (Figure 2A) and eighteen primary BCP-ALL blasts by flow cytometry and western blot analyses (Figure 2B). Western blot images are shown in Online Supplementary Figure S1. Both type I and type II Activin receptors were found to be expressed by all primary blasts and cell lines tested, with a markedly wide range of expression. The expression level of ActivinA receptors in primary BCP-ALL cells was highly patient-specific and was shown to be independent of the commonly investigated leukemia-related genetic alterations. Taken together, these data suggest that BCP-ALL cells could possibly respond to ActivinA. Moreover, we showed that ActivinA was able to significantly increase the expression of its type I receptors, thus suggesting a positive loop underlying the responsiveness of leukemic cells to ActivinA (Online Supplementary Figure S2). Subsequent analyses were performed on the 697 and 536

NALM-6 cell lines as their different Activin receptor expression makes them representative of the high interpatient variability observed.

Gene expression analysis revealed ActivinA involvement in regulating cell motility For a more in depth analysis of the molecular changes induced by ActivinA in BCP-ALL cells, we performed gene expression profile (GEP) analysis of 697 cells upon stimulation with ActivinA for 6 h and 24 h. We found that 122 genes were differentially expressed in ActivinA-treated cells versus untreated cells after 6 h of stimulation (FDR<0.05) and that 151 genes were differentially expressed after 24 h of stimulation (FDR<0.05). Gene Ontology (GO) analysis of differentially expressed genes identified enriched GO categories (Online Supplementary Figure S3A) critically linked to cancerogenesis such as “regulation of cell activation”, “positive regulation of antigen receptor-mediated signaling pathway”, “pathways in cancer”, etc. Interestingly, we also observed that ActivinA was able to influence migration-associated pathways, such as “calcium ion homeostasis and transport into cytosol", “PI3K/AKT activation”, “Ras signaling pathway”, “focal adhesion", suggesting its possible effect on leukemic cell motility. These data are in agreement with the recently recognized role of ActivinA in the regulation of cell migration and invasion in the context of several solid malignancies.12-15 On the basis of this evidence, we first used qRTPCR assays to validate the ActivinA-mediated changes in haematologica | 2019; 104(3)


ActivinA enhances BCP-ALL cell migration

A

C

B

D

E

Figure 3. ActivinA enhanced cell motility, migration and invasion of leukemic cells. (A) Leukemic cells were treated or not with ActivinA and then tracked for 24 hours (h) by time-lapse microscopy. Dead cells were excluded using PI staining. Data represent the meanÂąStandard Error of Mean (SEM) of the migrated distance over time of three independent experiments. The distance migrated after 24 h by all tracked cells was compared between ActivinA-stimulated (gray) and not stimulated (white) cells. (Left) 697 cells treated with 50 ng/mL ActivinA. (Right) B-cell precursor-acute lymphoblastic leukemia (BCP-ALL) primary blasts treated with 100 ng/mL ActivinA. (B) Chemotaxis assay was performed using 697 cells stimulated with ActivinA for 24 h (50 ng/mL) and allowed to migrate toward CXCL12-containing medium (100 ng/mL) for 4 h (5 mM pore Transwell). Each box plot shows the median and the mean (+) of the percentage of migrated cells and extends from the lowest to the highest value. The graphs represent the results of six independent experiments. The percentage of migrated cells was determined as described in the Online Supplementary Methods. (C) Primary BCP-ALL cells from 13 patients were exposed to ActivinA (100 ng/mL) for 24 h and employed for chemotaxis assays toward CXCL12-containing medium (100 ng/mL). The average percentage of cells migrated after 1 h of culture is shown. (D) Chemotaxis assay was performed using BCPALL primary cells pretreated for 1 h with SB431542 or vehicle (DMSO) before the stimulation with or without ActivinA for 24 h (100 ng/mL). Cells were allowed to migrate toward CXCL12-containing medium (100 ng/mL) for 1 h (5 mM pores Transwell). The percentage of migrated cells was determined as described in the Online Supplementary Methods. The average percentage of inhibitionÂąSEM is represented (one out of three representative experiments). (E) Primary BCP-ALL cells pretreated or not with ActivinA for 24 h (100 ng/mL) were allowed to migrate through Transwell inserts (8 mM pores) coated with a Matrigel-barrier (1 mg/mL) for 24 h in the presence of CXCL12 (100 ng/mL) in the lower chamber. Each box plot shows the median and the mean (+) of the percentage of invaded cells and extends from the lowest to the highest value. The graphs represent the results of 3 different patients. *P<0.05; ***P<0.001; ****P<0.0001; Wilcoxon matched-pairs signed rank test (A-C); *P<0.05: comparison with vehicle (0 mM), Mann-Whitney test; #P<0.05: comparison with 10 mM SB431542, Mann-Whitney test (D); *P<0.05: MannWhitney test (E).

the expression of several genes linked to Ca2+ homeostasis (ATP2B2, ATP2B4), Ras pathway activation (VAV3), and cell motility and movement regulation (ARHGAP25, CORO1A, DOCK4, LCK, PTPRC). Data obtained in qRTPCR on 697 cells were highly concordant with microarray data (Online Supplementary Figure S3B). Raw data are shown in Online Supplementary Figure S4. In addition, qRT-PCR analyses of the above-mentioned genes were performed on 7 primary BCP-ALL patients' samples stimulated or not with ActivinA (Online Supplementary Figure S5). Among them, 3 presented the t(1;19) (gray dots) typical of the 697 cell line. Interestingly, four of the tested genes (ARHGAP25, CORO1A, DOCK4 and PTPRC) were significantly modulated in at least one stimulation time point (Online Supplementary Figure S5B) by ActivinA in more than 70% of patients, similarly to our observations in 697 cells. It is worth noting that the t(1;19) patients showed a modulation of the ATP2B4, VAV3 and LCK genes (Online Supplementary Figure S5B) more similar to 697 cells than the translocation negative ones.

ActivinA increased random motility, chemotaxis and invasion of B-cell precursor-acute lymphoblastic leukemia cells To test whether ActivinA was able to modulate BCPALL movement, we performed time-lapse microscopy haematologica | 2019; 104(3)

(TLM) analyses and migration assays. TLM studies showed that ActivinA was able to increase random motility both in the 697 cell line (P<0.0001) (Figure 3A, left) and primary leukemic cells (P<0.0001) (Figure 3A, right). It has been demonstrated that chemokines, and in particular the CXCR4/CXCL12 axis, play a key role in the homing and retention of ALL cells within the BM niche. Therefore, we tested whether ActivinA was able to modulate CXCL12-induced migration of leukemic cells using a Transwell-based migration assay. Notably, we found that ActivinA-pretreated 697 cells showed a significant increase in CXCL12-driven migration (P<0.05) (Figure 3B). Importantly, this finding was confirmed in primary leukemic cells obtained from the BM of 13 BCP-ALL patients collected at diagnosis (Figure 3C). The average percentage of primary unstimulated BCP-ALL blasts migrated in response to CXCL12 (mean: 15.6%, range: 1.9-43.5%) was significantly increased following ActivinA stimulation (mean: 22.9%, range: 4.1-61.1%) (P<0.001). To ensure the specificity of the observed effect on CXCL12 mediated migration, we blocked ActivinA/Activin receptor axis by using SB431542, a wellcharacterized specific inhibitor of transforming growth factor-b superfamily type I Activin receptor-like kinase (ALK) receptors ALK4, ALK5, and ALK7.16,17 SB431542 inhibited the migration of ActivinA-pretreated 697, 537


F. Portale et al. A

B

NALM-6 cells (Online Supplementary Figure S6) and primary leukemic blasts (Figure 3D) in response to CXCL12 in a dose-dependent manner. In 3 different patients, we demonstrated that 10 mM SB431542 inhibited ActivinA stimulatory effect on CXCL12-driven migration of 78.8% (range: 74.5-84.0%; P<0.05). Interestingly, it has been reported that ActivinA expression is associated with an invasive phenotype in several types of cancer, including ovarian cancer, esophageal adenocarcinoma, breast cancer, and oral squamous cell carcinomas.12-15 Therefore, we tested whether ActivinA was able to modulate leukemic cell invasive capacity using Matrigel-coated Transwells. We found that ActivinA increased the ability of primary BCP-ALL cells to migrate through a complex matrix in the presence of CXCL12 (Figure 3E), with a 2-fold increase compared to the untreated condition (P<0.05).

ActivinA enhanced leukemic cells responsiveness to low levels of CXCL12 CXCL12 reduction is one of the microenvironmental alterations occurring in the leukemic BM, as observed in both mice models and leukemic patients,18,19 which is associated with impairment of normal hematopoiesis.5 Here, in a large cohort of 70 patients, we confirmed a significant reduction of approximately 6 times the CXCL12 of BM plasma level in BCP-ALL patients (mc: 77.7, range: 15.7488.9 pg/mL) compared to HDs (mc: 476.8, range: 99.11763 pg/mL, n=46) (P<0.0001) (Figure 4A). To test the potential ability of ActivinA to increase the responsiveness of leukemic cells to suboptimal concentrations of CXCL12, we performed dose-response chemotaxis assays. We demonstrated that ActivinA enhanced 697 cell 538

Figure 4. ActivinA enhanced leukemic cell responsiveness to CXCL12. (A) CXCL12 bone marrow plasma levels were assessed by ELISA in 46 healthy donors (HDs) and 70 B-cell precursor-acute lymphoblastic leukemia (BCPALL) patients at the onset of the disease. Each box plot shows the median and the mean (+) and extends from the lowest to the highest value. ****P<0.0001: Mann-Whitney test. (B) 697 cells were pretreated with ActivinA (50 ng/mL) for 24 hours (h) and then incubated for 4 h in Transwell chambers toward decreasing concentration of CXCL12 (100-10-1 ng/mL). Each box plot shows the median and the mean (+) of the percentage of migrated cells and extends from the lowest to the highest value. The graphs represent the results of six independent experiments. *P<0.01: ActivinA-treated versus untreated 697 cells, Wilcoxon matched-pairs signed rank test; #P<0.05; ## P<0.01: comparison with 100 ng/mL CXCL12-induced migration, Mann-Whitney test.

line migration toward CXCL12 used at a concentration 10- or 100-fold lower than that classically used in in vitro migration assays (100 ng/mL). Indeed, ActivinA pretreatment induced a 10-fold increase in the CXCL12-driven chemotaxis toward 10 ng/mL CXCL12 (P<0.05) and a 7fold increase toward 1 ng/mL CXCL12 (P<0.05), compared to untreated cells, that showed a responsiveness to these low chemokine concentrations comparable to that of empty medium (Figure 4B).

Intracellular calcium levels and actin polymerization were increased by ActivinA in leukemic cells To further investigate the enhanced leukemic cell responsiveness to CXCL12, we first evaluated whether ActivinA treatment could affect chemokine receptor expression. Flow cytometry analysis of CXCL12 chemokine receptors on 697 cells showed that the levels of CXCR4 and CXCR7, evaluated both as extracellular receptors and intracellular pool, were not affected by ActivinA (Figure 5A). We, therefore, performed flow cytometry analysis to determine the effect of ActivinA on the intracellular calcium level of BCP-ALL cells. Our data indicated that both in 697 cells (Figure 5B) and in primary BCP-ALL cells (Figure 5C) the basal intracellular calcium content was increased in ActivinA-pretreated cells as compared to untreated cells (mean range 1; P<0.05). Interestingly, on addition of CXCL12, ActivinA-treated cells showed a further significant increase in the concentration of free cytosolic Ca2+ compared to the untreated cells (Figure 5B and 5C, peak and mean range 2; P<0.05). Moreover, we evaluated the effect of ActivinA on actin cytoskeleton dynamics. Since the conversion of globular haematologica | 2019; 104(3)


ActivinA enhances BCP-ALL cell migration

A

B

C

D

Figure 5. ActivinA increased CXCL12-induced calcium mobilization and actin polymerization. (A) The extracellular and intracellular levels of CXCR4 and CXCR7 were analyzed by flow cytometry in 697 cells treated (black line) or not (gray line) with ActivinA for 24 hours (h). Representative data are shown from three independent experiments. 697 cells (B) and B-cell precursor-acute lymphoblastic leukemia (BCP-ALL) primary blasts (C) were cultured for 24 h either in the presence or in the absence of ActivinA (50 ng/mL or 100 ng/mL, respectively). Cells were loaded with Fluo-4 NW and free cytosolic Ca2+ changes were measured by FACS. Background was recorded for 30 seconds (s) and signal upon CXCL12 addition was registered for an additional 90 s. The black line represents data obtained from ActivinA-treated cells; the gray line represents untreated cells. The arrow indicates CXCL12 addition. The box plots represent the mean fluorescence intensity (MFI) before (mean range 1) and after (mean range 2) the addition of CXCL12 and the maximum peak reached upon CXCL12 addition (peak range 2). Each box plot shows the median and the mean (+), and extends from the lowest to the highest value. The results are representative of one out of six independent experiments. (D) 697 cells were starved in low serum medium for 24 h and then stimulated or not with ActivinA (50 ng/mL) for an additional 24 h. Cells were then stained with AF647-phalloidin and MFI quantified by flow cytometry. Percentage of MFI change was defined as follows: (MFI after CXCL12 addition/ MFI before CXCL12 addition) x 100. Mean values (ÂąStandard Error of Mean) of one out three independent experiments are represented in the graph. *P<0.05: Wilcoxon matched-pairs signed rank test (B and C). *P<0.05: Mann-Whitney test (D).

into filamentous actin (F-actin) is a prerequisite for sitedirected migration, we analyzed whether the increased chemotactic response upon ActivinA treatment was associated with enhanced chemokine-induced actin polymerization. Pretreatment of 697 cells with ActivinA for 24 h resulted in a more prominent conversion of globular into F-actin starting from 15 seconds (s) after addition of CXCL12 (P<0.05) (Figure 5C). Notably, ActivinA-pretreated cells maintained a higher amount of F-actin compared to untreated cells, even 180 s after CXCL12 stimulation (P<0.001). These data strongly support our GEP results highlighting the role of ActivinA as a modulator of several genes involved in cytoskeleton remodeling and regulation of calcium dynamics. Results on migration, invasion, chemokine receptors, and calcium flux were confirmed in the NALM-6 BCPALL cell line (Online Supplementary Figure S7). haematologica | 2019; 104(3)

ActivinA impaired CXCL12-driven migration of healthy CD34+ cells We further evaluated whether ActivinA promoted a selective advantage to BCP-ALL cells compared to healthy CD34+ cells. CB- and BM-derived CD34+ cells expressed both type I and type II Activin receptors, thus suggesting that they could both respond to this molecule (Figure 6A). The effect of ActivinA on CXC12-driven chemotaxis of CD34+ cells was evaluated by Transwell-based migration assays. Surprisingly, we observed the opposite effect to that observed with leukemic cells. ActivinA pretreatment resulted in an average reduction of approximately 55% of CXCL12-driven chemotaxis, compared to untreated CB-CD34+ cells (Figure 6B). Of note, the regulation of cell viability did not account for the reduced chemotaxis (data not shown). These data were confirmed in BM-CD34+ cells derived from three HDs, with an average reduction of 539


F. Portale et al.

approximately 25% in CXCL12-driven migration (Figure 6B). This effect on CD34+ cell migration was not due to an ActivinA-mediated regulation of the CXCL12 chemokine receptors, CXCR4 and CXCR7, as demonstrated by flow cytometry analysis of both membrane-bound receptors and intracellular pool (Figure 6C). In addition, ActivinA pretreatment significantly decresed free cytosolic Ca2+ of CD34+ cells after the addition of CXCL12 in 2 out of 3

independent experiments (Figure 6D). Overall, these data suggest that leukemic cells could displace healthy hematopoietic stem cells from their niches through an ActivinA-mediated mechanism.

Acute lymphoblastic leukemia-mesenchymal stromal cells secrete high amounts of ActivinA Finally, we focused our attention on the capacity of the

A

B

C

D

Figure 6. ActivinA impaired CXCL12-driven migration of healthy CD34+ cells. (A) Expression of ActivinA receptors ALK2, ALK4, ACVR2A and ACVR2B was quantified in both cord blood (CB)-CD34+ and bone marrow (BM)-CD34+ cells by qRT-PCR. Data are presented as mRNA fold change of ActivinA receptors normalized to GAPDH mRNA (endogenous control). DAUDI cell line was employed as calibrator because of its low expression of ActivinA receptors (www.proteinatlas.org). (B) CB-CD34+ cells (top) and BM-CD34+ (bottom) were pretreated or not with ActivinA (100 ng/mL) for 24 hours (h) and allowed to migrate through 5 mM pores in Transwell chambers toward CXCL12 (100 ng/mL) for 1 h. The graphs represent one representative experiment (CB-CD34+ n=5, BM-CD34+ n=3). Each box plot shows the median and the mean (+) of the percentage of migrated cells, and extends from the lowest to the highest value. *P<0.05: Mann-Whitney test. (C) The extracellular and intracellular levels of CXCR4 and CXCR7 were analyzed in cells either treated (black line) or not (gray line) with ActivinA for 24 h by flow cytometry. Data from one representative experiment out of three are shown. (D) CB-CD34+ cells were cultured for 24 h in the presence or absence of ActivinA (100 ng/mL). Cells were loaded with Fluo-4 NW and free cytosolic Ca2+ changes were measured by FACS. Background was recorded for 30 seconds (s) and signal upon CXCL12 addition was registered for an additional 90 s. The black line represents results obtained from ActivinA-treated cells, while the gray line represents untreated cells. The results are representative of three independent experiments. *P<0.05: Mann-Whitney test (Ppeak R2: comparison between MFI peak range 2 in treated vs. untreated cells; Pmean R2: comparison between MFI mean range 2 in treated vs. untreated cells).

540

haematologica | 2019; 104(3)


ActivinA enhances BCP-ALL cell migration

BCP-ALL BM microenvironment to influence MSCderived ActivinA. For this purpose, we isolated BM-MSCs from 15 HDs and 15 BCP-ALL patients at the onset of the disease. ALL-MSCs resulted comparable in terms of immunophenotype and adipogenic/osteogenic differentiation capacity to HD-MSCs (Online Supplementary Figure S8). After 24 h of culture, ELISA assay showed a significantly higher production of ActivinA (P<0.05) by ALL-MSCs (mc: 222.2, range: 62.5-4855 pg/mL) compared to their normal counterparts (mc: 220.7, range: 62.5-518 pg/mL) (Figure 7A). Therefore, we hypothesized that BM-MSCs primed by the leukemic microenvironment could account for the high amount of ActivinA in the BM of BCP-ALL patients.

The role of inflammation in the editing of the microenvironment has been defined in several types of cancer, including hematologic malignancies. Recent evidence highlighted that the BM of ALL patients is a highly proinflammatory environment.20 These data were confirmed in our cohort of patients. Indeed, higher levels of the proinflammatory cytokines IL-1b (P<0.0001), IL-6 (P<0.01), and TNF-α (P<0.01) were detected in the BM plasma of BCP-ALL patients compared to HDs (Online Supplementary Figure S9). We then investigated whether the pro-inflammatory cytokines IL-1b, IL-6 and TNF-α could regulate ActivinA levels in the BM of BCP-ALL patients by stimulating both HD-MSCs and ALL-MSCs with a cocktail of the abovementioned pro-inflammatory cytokines for 24 h. ELISA

A

B

Figure 7. Inflammation contributed to ActivinA production by bone marrow-mesenchymal stem cells (BM-MSCs). (A) ActivinA secretion by BM-MSCs from B-cell precursor-acute lymphoblastic leukemia (BCP-ALL) patients (ALL-MSCs; n=15) and from healthy donors (HDs) (HD-MSCs; n=15) was assessed by ELISA after 24 hours (h) of culture ± IL-1b (50 ng/mL), IL-6 (40 ng/mL) and TNF-α (100 ng/mL). Each box plot shows the median and the mean (+), and extends from the lowest to the highest value. *P<0.05; **P<0.01: Wilcoxon matched-pairs signed rank test. (B) Primary BCP-ALL cells were co-cultured with HDMSCs directly (bottom) or separated by a Transwell insert (top) in presence of IL-1b, IL-6 and TNF-α for 72 h. ActivinA expression was assessed by ELISA on culture supernatants (n=17 independent co-cultures). The expected additive effect was calculated as the sum of the single effects produced by the two stimulating factors, leukemic cells (second column) and inflammation (third column). Synergism was defined as a “greater-than-the-expected-additive effect”. Each box plot shows the median and the mean (+), and extends from the lowest to §§§ P<0.001; the highest value. §§§§ P<0.0001: stimulated versus unstimulated MSC; ***P<0.001; ****P<0.0001: measured effect versus expected additive effect, indirect contact and direct contact, respectively; Wilcoxon matched-pairs signed rank test.

haematologica | 2019; 104(3)

541


F. Portale et al.

A

B

C

Figure 8. NALM-6 cells stimulated with ActivinA showed an increased ability to engraft in vivo in the bone marrow (BM), meninges and brain of NSG mice. (A) NSG mice received intravenous (i.v.) transfer of 1x106 NALM-6 cells, previously cultured for 24 hours (h) in the presence or absence of ActivinA (50 ng/mL). (B) Weight loss was periodically monitored over two weeks after transplantation. The graph shows mean percentagesÂąStandard Error of mean (SEM) of body weight change. Data from 12 mice/group in four independent experiments are shown. *P<0.05; **P<0.01: Mann-Whitney test. (C) Mice were sacrificed on days 11 and 14 after transplantation and the percentage of infiltrating hCD19+ hCD10+ leukemic cells was determined by flow cytometry in the BM, peripheral blood, spleen, liver, meninges and brain. Each box plot shows the median and the mean (+), and extends from the lowest to the highest value (n=12 mice/group, four independent experiments). *P<0.05: Mann-Whitney test.

assays revealed a significant induction of ActivinA release in BM-MSCs compared to their respective basal condition. Indeed, upon stimulation, ActivinA production by HDMSCs reached a 28-fold increase compared to the basal condition (mc: 5713, range: 1446-14221 pg/mL vs. basal condition) (P<0.0001) (Figure 7A). Interestingly, the molecule was released to a higher extent by ALL-MSCs in a pro-inflammatory condition compared to their normal counterparts (mc: 10085, range: 2904-19776 pg/mL vs. inflamed HD-MSCs; P<0.01). Notably, by mimicking an inflamed BM niche through the simultaneous stimulation of HD-MSCs with leukemic blasts and pro-inflammatory cytokines (Figure 7B), we showed a strong increase in the secretion of ActivinA both in the direct (Figure 7B, bottom, mc: 27860, range: 1315092391 pg/mL, n=17) and the indirect (Figure 7B, top, mc: 25409, range: 9050-65714 pg/mL) co-culture condition. Of note, the combination of both leukemic blasts and proinflammatory cytokines (Figure 7B, fourth column) produced a synergistic induction of ActivinA, since the extent of the release was higher compared to the sum of separately used stimuli21 (Figure 7B, expected additive effect: fifth column=second+third columns; top: P<0.001; bottom: P<0.0001). 542

ActivinA increased the in vivo engraftment of B-cell precursor acute lymphoblastic leukemia cells to bone marrow and extramedullary sites in a xenograft mouse model With the aim of testing the efficacy of ActivinA to induce leukemia dissemination in vivo, we performed a set of experiments in which 697 or NALM-6 cells in vitro pretreated with ActivinA for 24 hours were injected (i.v.) into NSG mice. Interestingly, on day +7, NSG mice injected with 697 cells (Online Supplementary Figure S10) pretreated with ActivinA showed a higher leukemic engraftment in the liver (median percentage of leukemic cells: 48.4%, range: 19.2-54.1%, n=9) compared to untreated cells (median percentage of leukemic cells: 27.0%, range: 10.9-43.2%, n=9), suggesting a migratory advantage in ActivinA-treated cells. As expected from the literature, our data showed a high tropism of 697 cells for the liver.22,23 On the contrary, the percentages of leukemia engraftment in BM and in other leukemic target organs were modest and were comparable between the two experimental groups. To test a more physiological environment for leukemia and overcome the low engraftment of 697 in BM, we transplanted mice with NALM-6 cells that are known to haematologica | 2019; 104(3)


ActivinA enhances BCP-ALL cell migration

have high engraftment levels into the BM.24 NALM-6 cells, either pretreated or not with ActivinA, were injected (i.v.) at day 0 in NSG mice (1×106/mouse) to evaluate their engraftment in different leukemia-targeted organs (Figure 8A). Mice were subsequently monitored for weight loss over two weeks after transplantation, and leukemia burden was evaluated 11 and 14 days after injection by flow cytometry, determining the percentage of human CD19 and CD10 positive cells in several organs. First, we observed a significant difference in terms of change in body weight between the two experimental groups starting from day 4 after injection (P<0.05) with an even greater difference on days 11 and 14 after injection (P<0.01), suggesting a different disease progression (Figure 8B). Moreover, we found that, on injection, both untreated and pretreated cells disseminated through the peripheral blood to different organs, such as BM, spleen, liver, meninges, and brain. Notably, we found that ActivinApretreated cells were able to engraft more rapidly in the BM of recipient mice (median leukemic percentage at day +11: 17.7%, range: 6.1-42.3%) compared to their untreated counterparts (median leukemic percentage at day +11: 10.3%, range: 0-21.4%, n=12). We also observed an increased leukemic percentage in the central nervous system (meninges and brain) of NSG mice following in vitro exposure to ActivinA (P<0.05) (Figure 8C), indicating that ActivinA stimulation was able to enhance the metastatic potential of leukemic cells in vivo. Therefore, in these two mouse models, we were able to demonstrate that ActivinA stimulation makes leukemic cells more aggressive.

Discussion There is ample evidence correlating aberrant TGF-b family growth factor activity to carcinogenesis. Despite its prominent role in solid cancer progression,7 the involvement of ActivinA, a member of the TGF-b family, in the pathogenesis of hematologic malignancies has never been explored. To the best of our knowledge, here we show for the first time that ActivinA is highly expressed in the BM of BCP-ALL patients at diagnosis compared to HDs. Interestingly, we demonstrated that BM-MSCs are an important source of ActivinA, the production of which is strongly up-regulated following direct contact with leukemic cells or through leukemia-released soluble factors. This finding is in accordance with the recently revised “seed and soil” theory, showing that leukemic cells are able to alter the BM stroma, creating a fertile ground which fuels tumor cell survival and progression.1,25,26 Of note, we observed that MSCs isolated from the BM of BCP-ALL patients were able to produce higher levels of ActivinA compared to HD-MSCs, even after several in vitro passages. This means that they 'remember' the profound alterations that had occurred within the leukemic BM niche. Nowadays, there is general agreement that inflammation could play a pivotal role in the transformation, survival and proliferation of leukemic cells. In particular, several studies have demonstrated that BM cells in ALL are able to create a pro-inflammatory microenvironment that impairs frequency and function of normal HSCs within the BM.18,20 In line with this evidence, we demonstrate that haematologica | 2019; 104(3)

the BM of BCP-ALL patients is characterized by increased levels of the pro-inflammatory cytokines IL-1b, IL-6 and TNF-α, which can synergize with BCP-ALL cells in stimulating ActivinA production and release by BM-MSCs. Accordingly, it has been demonstrated that ActivinA expression was increased in several inflammatory diseases, such as septicemia, inflammatory bowel disease, and rheumatoid arthritis.27 The abundance of ActivinA within the leukemic BM niche, and its identification as a new MSC-secreted leukemia-driven molecule, prompted us to investigate its possible effects on BCP-ALL cells. In accordance with recent literature reporting that ActivinA increases the migration and invasive properties of several solid tumors,12-14,28-33 our GEP analysis of ActivinA-treated leukemic cells showed a crucial effect on different biological processes linked to cell motility. In detail, we used time-lapse microscopy to demonstrate that this molecule was able to increase the spontaneous cell motility of both immortalized and primary BCP-ALL cells. In addition to increased random motility, ActivinAtreated leukemic cells were more responsive to the CXCL12 chemokine, which plays an essential role in maintaining the quiescent BM HSC pool, thus regulating physiological hematopoiesis.34 The increase in BCP-ALL migration towards CXCL12 was selectively inhibited by ActivinA/Activin receptor blocking through SB431542,16 ensuring the specific contribution of ActivinA in mediating this advantage. Importantly, the effect of ActivinA on cell chemotaxis was highly cell-specific. While increasing leukemic cell migration in response to even suboptimal concentrations of CXCL12, ActivinA markedly impaired the ability of healthy CD34+ cells to migrate toward a CXCL12 gradient. This opposite effect could be particularly relevant in the context of the altered BCP-ALL BM niche, where we observed a significantly decreased CXCL12 concentration, in agreement with recent literature.19 Concerning the molecular mechanisms underlying ActivinA action, a protein-mediated regulation of the CXCL12 receptors CXCR4 and CXCR7 was ruled out. Moreover, in contrast to what Sozzani et al. described in dendritic cells,16 ActivinA did not stimulate either leukemic cell chemotaxis itself (Online Supplementary Figure S10) or their secretion of CXCL12 (data not shown). On the contrary, our GEP analysis demonstrated that this molecule mainly induces an overall positive regulation of pathways associated with cell motility, such as RAS, PI3K/AKT, and calcium homeostasis. It has been demonstrated that Ca2+ co-ordinates structural components of the cell migration machinery and signaling molecules crucial for proper cell motility. Through the activation of actin-interacting molecules such as protein kinase C35 and calmodulin-dependent kinases,36 and the regulation of Rho GTPases, Ca2+ signaling finely tunes actin cytoskeleton dynamics.37 Interestingly, we demonstrated that ActivinA is able to increase the motility of BCP-ALL cells through an increase in the pool of free cytosolic calcium, resulting in an increased rate of F-actin polymerization. Strikingly, the increase in intracellular Ca2+ was not observed in ActivinA-stimulated CD34+ cells, thus explaining the possible molecular mechanism mediating the differential activity of this molecule on the migration of leukemic versus healthy hematopoietic cells. Recent literature28 describing the ability of ActivinA to 543


F. Portale et al.

stimulate the migration of ovarian cancer cells suggests that Ca2+ increase could be achieved also in leukemic cells through the activation of non-canonical phospo-AKT, phospho-ERK and Rac1 signaling. In line with this hypothesis, our GEP data demonstrate an ActivinAdependent increase in DOCK438 and CORO1A39 in the 697 cell line and primary BCP-ALL cells. This increase positively regulates the CDC42/RAC1 pathway, leading to the generation of phosphatidylinositol-3-phosphate, responsible for calcium release from intracellular stores.40 In agreement with this, ActivinA-mediated downmodulation of the Rac-GTPase activating protein ARHGAP25, that physiologically counterbalances the Rac-activating effect of nucleotide exchange factors,41 could play a prominent role in shaping calcium levels and modulating actin cytoskeleton42 also in leukemic cells. Moreover, ActivinA could also regulate the extent and duration of calcium responses through PTPRC downregulation. Indeed, in immature B cells, it has been demonstrated that the lack of the PTPRC product, CD45, induces enhanced levels of intracellular calcium that can last longer than in CD45 expressing cells, upon BCR engagement.43 Further studies will be crucial to better understand the possible molecular mechanisms mediating ActivinA differential activity in order to identify potential selective targets to counteract its action. The ActivinA-mediated migratory advantage observed on BCP-ALL cells was further confirmed in a xenograft mouse model in which we demonstrated that leukemic cells prestimulated with ActivinA were able to engraft in the BM of NSG mice more rapidly than in their untreated counterparts. Overall, our in vivo data corroborate in vitro findings suggesting the effect of the molecule in favoring leukemia. Future studies testing the efficacy of ActivinA ligand traps on BCP-ALL patient-derived xenografts will be crucial to establish the impact of ActivinA on leukemia propagation. Recent studies importantly linked ActivinA with the enhancement of cell invasion in several solid cancers (colorectal cancer, prostate cancer, breast cancer, glioblastoma, non-small cell lung cancer).44-48 In the context of BCP-ALL, relapse represents the most common cause of treatment failure, mainly occurring in the BM either in an isolated form or in combination with other extramedullary sites.49 Besides its key role in regulating homing processes in the BM niche, CXCL12 is thought to be involved in the widespread infiltration of other organs because of its constitutive expression in

References 1. Chiarini F, Lonetti A, Evangelisti C, et al. Advances in understanding the acute lymphoblastic leukemia bone marrow microenvironment: From biology to therapeutic targeting. Biochim Biophys Acta. 2016;1863(3):449-463. 2. Pui CH, Robison LL, Look AT. Acute lymphoblastic leukaemia. Lancet. 2008; 371(9617):1030-1043. 3. Hunger SP, Mullighan CG. Acute lymphoblastic leukemia in children. N Engl J Med. 2015;373(16):1541-1552. 4. Ayala F, Dewar R, Kieran M, Kalluri R. Contribution of bone microenvironment to

544

5.

6. 7.

8.

extramedullary tissues such as liver, spleen, thymus, lung, kidney, and brain.50 Interestingly, our in vitro observation demonstrated that ActivinA significantly increased the ability of leukemic cells to pass through an extracellular matrix in response to CXCL12. In addition, in vivo injected ActivinA-stimulated leukemic cells were more able than untreated cells to reach extramedullary disease target organs such as the meninges and the brain, suggesting a possible role for ActivinA in the promotion of leukemic cell invasiveness. Notably, ActivinA was able to up-regulate the expression of its type I receptors in leukemic cells, thus creating a self-reinforcing signaling cascade. Overall, our data suggest the establishment of a positive feedback loop between BCP-ALL cells and MSCs, which, through the key action of MSC-secreted ActivinA, generates a microenvironment favoring leukemia at the expense of normal hematopoiesis. Indeed, it is conceivable that the abundance of ActivinA, along with the decrease in CXCL12 within the BM niche, could lead to a reduction in the healthy HSC pool in favor of leukemic cells. On the other hand, the leukemic cells could access the BM sanctuaries where they can achieve signals necessary for cell survival and therapy resistance. Indeed, our in vitro findings were confirmed by in vivo studies and provide the biological rationale for designing therapeutical approaches targeting ActivinA in patients with BCP-ALL. Therefore, the direct targeting of ActivinA or its key downstream mediators could represent a valuable therapeutic option to be combined with conventional chemotherapeutic agents for decreasing the frequency of relapse in BCP-ALL. Acknowledgments The authors would like to thank the nursing and medical staff of the Fondazione MBBM/San Gerardo Hospital and the Bambino GesĂš Hospital. We thank Fondazione Matilde Tettamanti, Comitato Maria Letizia Verga, Comitato Stefano Verri, Fondazione MBBM, the Department of Medicine and Surgery of the University of Milano-Bicocca, Whirlpool and GEICO TAIKI-SHA for their generous support. Funding This work was partially supported by Associazione Italiana Ricerca sul Cancro (project number IG 2014 Id.15494, to GDâ&#x20AC;&#x2122;A). This article was also funded by AIRC Special Program Molecular Clinical Oncology - 5 per mille 2018 (project number 21147 to AB).

leukemogenesis and leukemia progression. Leukemia. 2009;23(12):2233-2241. Colmone A, Amorim M, Pontier AL, Wang S, Jablonski E, Sipkins DA. Leukemic cells create bone marrow niches that disrupt the behavior of normal hematopoietic progenitor cells. Science. 2008;322(5909):18611865. Hanahan D, Weinberg Robert A. Hallmarks of cancer: the next generation. Cell. 144(5):646-674. Loomans HA, Andl CD. Intertwining of ActivinA and TGFb signaling: dual roles in cancer progression and cancer cell invasion. Cancers. 2015;7(1):70-91. Pangas SA, Woodruff TK. Activin signal

transduction pathways. Trends Endocrinol Metab. 2000;11(8):309-314. 9. Abdulkadyrov KM, Salogub GN, Khuazheva NK, et al. Sotatercept in patients with osteolytic lesions of multiple myeloma. Br J Haematol. 2014;165(6):814823. 10. Conter V, Bartram CR, Valsecchi MG, et al. Molecular response to treatment redefines all prognostic factors in children and adolescents with B-cell precursor acute lymphoblastic leukemia: results in 3184 patients of the AIEOP-BFM ALL 2000 study. Blood. 2010;115(16):3206-3214. 11. Vallet S, Mukherjee S, Vaghela N, et al. ActivinA promotes multiple myeloma-

haematologica | 2019; 104(3)


ActivinA enhances BCP-ALL cell migration

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

induced osteolysis and is a promising target for myeloma bone disease. Proc Natl Acad Sci U S A. 2010;107(11):5124-5129. Bashir M, Damineni S, Mukherjee G, Kondaiah P. Activin-A signaling promotes epithelial–mesenchymal transition, invasion, and metastatic growth of breast cancer. NPJ Breast Cancer. 2015;1:15007. Taylor C, Loomans HA, Le Bras GF, et al. ActivinA signaling regulates cell invasion and proliferation in esophageal adenocarcinoma. Oncotarget. 2015;6(33):3422834244. Bufalino A, Cervigne NK, de Oliveira CE, et al. Low miR-143/miR-145 cluster levels induce ActivinA overexpression in oral squamous cell carcinomas, Which contributes to poor prognosis. PLoS One. 2015;10(8):e0136599. Steller MD, Shaw TJ, Vanderhyden BC, Ethier JF. Inhibin resistance is associated with aggressive tumorigenicity of ovarian cancer cells. Mol Cancer Res. 2005;3(1):5061. Salogni L, Musso T, Bosisio D, et al. ActivinA induces dendritic cell migration through the polarized release of CXC chemokine ligands 12 and 14. Blood. 2009; 113(23):5848-5856. Watabe T, Nishihara A, Mishima K, et al. TGF-beta receptor kinase inhibitor enhances growth and integrity of embryonic stem cell-derived endothelial cells. J Cell Biol. 2003;163(6):1303-1311. Balandrán JC, Purizaca J, Enciso J, et al. Proinflammatory-related loss of CXCL12 niche promotes acute lymphoblastic leukemic progression at the expense of normal lymphopoiesis. Front Immunol. 2017; 7:666. van den Berk LC, van der Veer A, Willemse ME, et al. Disturbed CXCR4/CXCL12 axis in paediatric precursor B-cell acute lymphoblastic leukaemia. Br J Haematol. 2014; 166(2):240-249. Vilchis-Ordoñez A, Contreras-Quiroz A, Vadillo E, et al. Bone marrow cells in acute lymphoblastic leukemia create a proinflammatory microenvironment influencing normal hematopoietic differentiation fates. Biomed Res Int. 2015; 2015:386165. Roell KR, Reif DM, Motsinger-Reif AA. An introduction to terminology and methodology of chemical synergy-perspectives from across disciplines. Front Pharmacol. 2017; 8:158. Yuan N, Song L, Zhang S, et al. Bafilomycin A1 targets both autophagy and apoptosis pathways in pediatric B-cell acute lymphoblastic leukemia. Haematologica. 2015;100 (3):345-356. Pillozzi S, Masselli M, De Lorenzo E, et al. Chemotherapy resistance in acute lymphoblastic leukemia requires hERG1 chan-

haematologica | 2019; 104(3)

24.

25.

26.

27. 28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

nels and is overcome by hERG1 blockers. Blood. 2011;117(3):902-914. Barrett DM, Seif AE, Carpenito C, et al. Noninvasive bioluminescent imaging of primary patient acute lymphoblastic leukemia: a strategy for preclinical modeling. Blood. 2011;118(15):e112-117. Raaijmakers MH, Mukherjee S, Guo S, et al. Bone progenitor dysfunction induces myelodysplasia and secondary leukaemia. Nature. 2010;464(7290):852-857. Sison EAR, Brown P. The bone marrow microenvironment and leukemia: biology and therapeutic targeting. Expert Rev Hematol. 2011;4(3):271-283. Sozzani S, Musso T. The yin and yang of ActivinA. Blood. 2011;117(19):5013-5015. Dean M, Davis DA, Burdette JE. ActivinA stimulates migration of the fallopian tube epithelium, an origin of high-grade serous ovarian cancer, through non-canonical signaling. Cancer Lett. 2017;391:114-124. Zhu J, Liu F, Wu Q, Liu X. ActivinA regulates proliferation, invasion and migration in osteosarcoma cells. Mol Med Rep. 2015;11(6):4501-4507. Kang HY, Huang HY, Hsieh CY, et al. ActivinA enhances prostate cancer cell migration through activation of androgen receptor and is overexpressed in metastatic prostate cancer. J Bone Miner Res. 2009;24(7):1180-1193. Yoshinaga K, Yamashita K, Mimori K, Tanaka F, Inoue H, Mori M. ActivinA causes cancer cell aggressiveness in esophageal squamous cell carcinoma cells. Ann Surg Oncol. 2008;15(1):96-103. Chang WM, Lin YF, Su CY, et al. Dysregulation of RUNX2/Activin-A axis upon miR-376c downregulation promotes lymph node metastasis in head and neck squamous cell carcinoma. Cancer Res. 2016;76(24):7140-7150. Tamminen JA, Yin M, Ronty M, et al. Overexpression of activin-A and -B in malignant mesothelioma - attenuated Smad3 signaling responses and ERK activation promote cell migration and invasive growth. Exp Cell Res. 2015;332(1):102-115. Sugiyama T, Kohara H, Noda M, Nagasawa T. Maintenance of the hematopoietic stem cell pool by CXCL12-CXCR4 chemokine signaling in bone marrow stromal cell niches. Immunity. 2006;25(6):977-988. Lipp P, Reither G. Protein Kinase C: The “Masters” of Calcium and Lipid. Cold Spring Harb Perspect Biol. 2011; 3(7):a004556-a004556. Swulius MT, Waxham MN. Ca(2+)/Calmodulin-dependent Protein Kinases. Cell Mol Life Sci. 2008; 65(17): 2637-2657. Parsons JT, Horwitz AR, Schwartz MA. Cell adhesion: integrating cytoskeletal

38.

39.

40. 41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

dynamics and cellular tension. Nat Rev Mol Cell Biol. 2010;11(9):633-643. Kobayashi M, Harada K, Negishi M, Katoh H. Dock4 forms a complex with SH3YL1 and regulates cancer cell migration. Cell Signal. 2014;26(5):1082-1088. Pick R, Begandt D, Stocker TJ, et al. Coronin 1A, a novel player in integrin biology, controls neutrophil trafficking in innate immunity. Blood. 2017;130(7):847858. Downward J. Targeting RAS signalling pathways in cancer therapy. Nat Rev Cancer. 2003;3(1):11-22. Csepanyi-Komi R, Wisniewski E. Rac GTPase Activating Protein ARHGAP25 Regulates Leukocyte Transendothelial Migration in Mice. J Immunol. 2016;197 (7):2807-2815. Csepanyi-Komi R, Sirokmany G, Geiszt M, Ligeti E. ARHGAP25, a novel Rac GTPaseactivating protein, regulates phagocytosis in human neutrophilic granulocytes. Blood. 2012;119(2):573-582. Ogimoto M, Katagiri T, Mashima K, Hasegawa K, Mizuno K, Yakura H. Negative regulation of apoptotic death in immature B cells by CD45. Int Immunol. 1994;6(4):647-654. Li M, Zhang S, Wu N, Wu L, Wang C, Lin Y. Overexpression of miR-499-5p inhibits non-small cell lung cancer proliferation and metastasis by targeting VAV3. Sci Rep. 2016;6:23100. Salhia B, Tran NL, Chan A, et al. The guanine nucleotide exchange factors trio, Ect2, and Vav3 mediate the invasive behavior of glioblastoma. Am J Pathol. 2008; 173(6):1828-1838. Jiang K, Lu Q, Li Q, Ji Y, Chen W, Xue X. Astragaloside IV inhibits breast cancer cell invasion by suppressing Vav3 mediated Rac1/MAPK signaling. Int Immunopharmacol. 2017;42:195-202. Lin KT, Gong J, Li CF, et al. Vav3-rac1 signaling regulates prostate cancer metastasis with elevated Vav3 expression correlating with prostate cancer progression and posttreatment recurrence. Cancer Res. 2012; 72(12):3000-3009. Uen YH, Fang CL, Hseu YC, et al. VAV3 oncogene expression in colorectal cancer: clinical aspects and functional characterization. Sci Rep. 2015;5(1):9360. Locatelli F, Schrappe M, Bernardo ME, Rutella S. How I treat relapsed childhood acute lymphoblastic leukemia. Blood. 2012;120(14):2807-2816. Crazzolara R, Kreczy A, Mann G, et al. High expression of the chemokine receptor CXCR4 predicts extramedullary organ infiltration in childhood acute lymphoblastic leukaemia. Br J Haematol. 2001;115(3):545553.

545


ARTICLE Ferrata Storti Foundation

Acute Lymphoblastic Leukemia

Targeting the endoplasmic reticulummitochondria interface sensitizes leukemia cells to cytostatics Fabian Koczian,1 Olga Nagło,1 Jan Vomacka,2 Binje Vick,3 Phil Servatius,4 Themistoklis Zisis,1 Britta Hettich,1 Uli Kazmaier,4 Stephan A. Sieber,2 Irmela Jeremias,3 Stefan Zahler1 and Simone Braig1

Department of Pharmaceutical Biology, Ludwig Maximilian University of Munich; Department of Chemistry, Technical University of Munich, Garching; 3Research Unit Gene Vectors, Helmholtz Zentrum München, German Center for Environmental Health, Munich and 4Institute of Organic Chemistry, Saarland University, Saarbrücken, Germany

1

Haematologica 2019 Volume 104(3):546-555

2

ABSTRACT

C

Correspondence: SIMONE BRAIG simone.braig@cup.uni-muenchen.de Received: May 9, 2018. Accepted: October 4, 2018. Pre-published: October 11, 2018.

ombination chemotherapy has proven to be a favorable strategy to treat acute leukemia. However, the introduction of novel compounds remains challenging and is hindered by a lack of understanding of their mechanistic interactions with established drugs. In the present study, we demonstrate a highly increased response of various acute leukemia cell lines, drug-resistant cells and patient-derived xenograft cells by combining the recently introduced protein disulfide isomerase inhibitor PS89 with cytostatics. In leukemic cells, a proteomics-based target fishing approach revealed that PS89 affects a whole network of endoplasmic reticulum homeostasis proteins. We elucidate that the strong induction of apoptosis in combination with cytostatics is orchestrated by the PS89 target B-cell receptor-associated protein 31, which transduces apoptosis signals at the endoplasmic reticulum -mitochondria interface. Activation of caspase-8 and cleavage of B-cell receptor-associated protein 31 stimulate a pro-apoptotic crosstalk including release of calcium from the endoplasmic reticulum and an increase in the levels of reactive oxygen species resulting in amplification of mitochondrial apoptosis. The findings of this study promote PS89 as a novel chemosensitizing agent for the treatment of acute leukemia and uncovers that targeting the endoplasmic reticulum - mitochondrial network of cell death is a promising approach in combination therapy.

doi:10.3324/haematol.2018.197368

Introduction

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/104/3/546

Despite the significant success in the management of childhood acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML) with survival rates of >80% and >60%, respectively,1 the outcome of patients with relapsed or chemoresistant leukemia is still dismal.2,3 Especially in older patients, the balance of tolerable dosing versus effective cytotoxicity remains a major challenge. This issue is further exacerbated by the development of leukemic cell chemoresistance, which has been demonstrated for several cytostatics including tubulin binders and topoisomerase inhibitors.4,5 In addition, the emergence of relapse-specific mutations of cancer cells is often associated with resistance to thiopurines and glucocorticoids.6,7 Thus, novel pharmaceutical options are urgently needed for the improvement of current treatment regimens. There is general consenius that combination therapies benefit from the crosstalk of antileukemic agents, however the mechanisms of interaction have only been explored for a few.8 Therefore, drug discovery is not only encouraged to identify novel compounds and targets, but also to enhance the understanding of their interdependence with established cytostatics. The concept of network pharmacology has raised great interest in recent years, especially regarding complex disease systems such as cancer.8,9 Following this principle, multi-target strategies, rather than the ‘one drug, one target’ paradigm, are proposed to be superior in rewiring cancer-

©2019 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

546

haematologica | 2019; 104(3)


PS89 - a novel option for combination therapy in acute leukemia specific networks and for overcoming the system robustness of cancer cell phenotypes.10,11 Translating this concept to combinatorial drug treatment, a highly interesting issue is not only how networks are locally perturbed by individual compounds, but moreover how interventions at multiple cellular loci cooperate. Considering potential proapoptotic target networks, the crucial role of the endoplasmic reticulum (ER)-mitochondria ‘social network of cell death’ was recently stressed in several studies highlighting the dynamic interaction of these two cellular elements.12,13 In this context, the B-cell receptor-associated protein 31 (BAP31) was described as a substrate of caspase-8 and emerges as a communicator of apoptosis signals from the ER to mitochondria.14,15 Consistently, a role of the caspase 8-BAP31 axis has been demonstrated in ER stress-triggered apoptosis of B-cell lymphocytic leukemia cells.16 ER stress results from an imbalance between ER protein load and folding capacity. Protein disulfide isomerases (PDI) constitute a crucial family of enzymes for maintaining oxidative protein folding and ER homeostasis.17 Hence, these proteins have been recognized as exciting novel targets in cancer research.18 Furthermore, overexpression of PDI has been discovered in leukemia and linked to chemoresistance.19-21 Recently, we introduced the first reversible small-molecule PDI inhibitor PS89 which binds in close proximity to the catalytic centers of PDI.22 Moreover and contrary to other PDI inhibitors that are severely cytotoxic,23,24 PS89 is not toxic up to micromolar concentrations, although it has been shown to greatly enhance etoposide-induced apoptosis. This exceptional feature of effective chemosensitization at subtoxic doses motivated not only further combination therapy studies with PS89, but also a deeper analysis of its interactive signaling. In the present work, PS89 is set on stage as a novel therapeutic option for the treatment of acute leukemia. The favorable attributes of PS89 and its broad applicability are highlighted in ALL and AML cell lines, drug-resistant cells as well as patientderived xenograft (PDX) cells. The critical networks integrated in the synergistic pro-apoptotic signaling of PS89 in combination with cytostatics were identified, thus emphasizing the crucial function of ER-mitochondria communication for successful combination therapies.

absence of compounds. Peripheral blood mononuclear cells were freshly isolated from EDTA-anticoagulated blood of healthy donors by gradient centrifugation using Ficoll-Paque PLUS (GE Healthcare, Chicago, IL, USA) according to the manufacturer's instructions. The peripheral blood mononuclear cells were maintained in RPMI 1640 with 2 mM glutamine supplemented with 20% (v/v) fetal calf serum and 1 mM pyruvate. CD34-positive cells were identified by staining with fluorescein isothiocyanate-conjugated anti-human CD34 antibody (BD Biosciences, Heidelberg, Germany) according to the manufacturer’s instructions and analyzed by flow cytometry as described by Fukuda et al.29

Ethical statements Written informed consent was obtained from all patients or legal guardians when the patients were minors. The study was performed in accordance with the ethical standards of the responsible committee on human experimentation (written approval from the ethics committee of the Ludwig Maximilian University Hospital, Munich, number 068-08) and with the Helsinki Declaration of 1975, as revised in 2000. All animal experiments were performed in accordance with the current ethical standards of the official committee on animal experimentation (written approval from Regierung von Oberbayern, number 55.2-1-542532-95-10).

Activity-based protein profiling

Jurkat cells were incubated with unmodified PS89 (100 mM) or dimethyl sulfoxide (DMSO) as a control for 45 min at 37°C and in a second step with the PS89 photo probe (20 mM) or DMSO as a control for 45 min at 37°C. Cells were lysed in 1 mL phosphatebuffered saline with 1% (v/v) NP40 and 1% (w/v) sodium deoxycholate and sonication for 15 s on ice. Sample preparation and mass spectrometry analysis of target proteins by gel-free activitybased protein profiling and dimethyl labeling was performed as previously described.30 Cutoff criteria for target identification were: (i) enrichment by photo probe log2 Probe/DMSO >1.6, -log10 P-value >2 and (ii) PS89 competition log2 Probe/PS89 >0. The data shown are the results of three biological replicates. Details of the methods used are available in the Online Supplementary Appendix.

Results Methods

PS89 sensitizes acute leukemia cell lines and patientderived xenograft cells to cytostatics

Cell cultures

The concept of chemosensitization with the recently introduced PDI inhibitor PS89 was initially evaluated in a dose-response apoptosis assay. While PS89 was applied at a fixed subtoxic dose, the concentration of etoposide could be reduced by at least half to achieve equal cytotoxicity applying the combination treatment. This is in line with a more than 2-fold shift in the half maximal effective concentration (EC50) (Figure 1A). Bliss values indicating synergistic effects of PS89 in combination with etoposide are shown in Online Supplementary Table S1A. Whereas etoposide-treated cells showed a pronounced G2 arrest, PS89 had no effect on the cell cycle (Online Supplementary Figure S1A). PS89 treatment in combination with subtoxic concentrations of etoposide synergistically inhibited Jurkat cell proliferation and colony formation (Online Supplementary Figure S1B,C). The ability of PS89 to induce synergistic apoptosis with diverse cytostatics could be translated to acute leukemia cells of different lineages.

Jurkat cells (wild-type, CASP8-deficient, Bcl-2- and Bcl-xL-overexpressing) were kindly provided by P. H. Krammer (Heidelberg, Germany). CCRF-CEM and vincristine-resistant CEM cells25,26 were obtained from M. Kavallaris (Sydney, Australia), HEK 293 and HeLa cells from the Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ; Braunschweig, Germany) and HL-60 from the American Type Culture Collection (ATCC; Manassas, VA, USA). All cell lines were maintained in ATCC-recommended culture conditions.

Patient-derived xenograft cells, peripheral blood mononuclear cells and CD34-positive cells The model of ALL and AML patients’ leukemia cells growing in mice has been described previously.27,28 Ethical statements and approvals are outlined in the Online Supplementary Information. In the present study, PDX cells were freshly isolated from the bone marrow or spleen of NSG mice and cultured in the presence or haematologica | 2019; 104(3)

547


F. Koczian et al. A

C

E

B

D

F

Figure 1. Chemosensitization of acute leukemia cells with PS89. (A,B) Apoptosis of Jurkat, HL-60, CCRF-CEM and vincristine resistant (VCR-R) CEM cells treated with cytostatics (ETO: etoposide, VCR: vincristine, DNR: daunorubicin) in the presence or absence of 25 mM PS89. The percentage of apoptotic cells was determined by FACS analysis after 48 h. (C) Jurkat and VCR-R CEM were cultured for 48 h in drug-supplemented medium and apoptosis was analyzed by immunoblotting. (D,E) Freshly isolated peripheral blood mononuclear cells (PBMC) and acute lymphoblastic leukemia (ALL) or acute myeloid leukemia (AML) patient-derived xenograft cells were treated with PS89 and cytostatics for 48 h or 72 h, respectively. Apoptotic cells were determined by FACS analysis and specific apoptosis was calculated in relation to untreated controls. Synergism was calculated using the Bliss independence model. (F) PBMC were treated for 48 h with the indicated drugs. CD34-positive cells were identified by flow cytometry using a fluorescein isothiocyanate-conjugated CD34 antibody. Cell death was analyzed by propidium iodide staining.

548

haematologica | 2019; 104(3)


PS89 - a novel option for combination therapy in acute leukemia This was demonstrated by combinations of PS89 with daunorubicin in HL-60 cells and vincristine in CCRF-CEM cells (Figure 1B, synergism calculations according to the Bliss independence model are shown in Online Supplementary Table S1B) as well as with 6-mercaptopurine and dexamethasone in Jurkat and HL-60 cells, respectively (Online Supplementary Figure S1D-F). Furthermore, vincristine-resistant CEM cells showed marked apoptosis in response to treatment with 100 nM vincristine in combi-

A

nation with PS89 (38.3% apoptotic cells), while being resistant to 10-fold higher vincristine concentrations without co-stimulation (1000 nM vincristine; 6.8% apoptotic cells) (Figure 1B). Moreover, clonogenic growth of vincristine-resistant CEM and HL-60 cells was significantly abrogated upon treatment with PS89 in combination with vincristine or 6-mercaptopurine, respectively (Online Supplementary Figure S1G,H). Following treatment of Jurkat, vincristine-resistant CEM, CCRF-CEM and HL-60

B

C

D

E

Figure 2. Analysis of PS89 target proteins. (A) Protein disulfide isomerase (PDI)-genetically modified Jurkat cells (siRNA silencing, 48 h) or HEK cells (PDI overexpression, 24 h) were treated with etoposide and/or PS89 for 48 h and 72 h, respectively. Apoptosis was determined by FACS analysis and inhibition of proliferation by CellTiter-Blue staining. PDI expression was analyzed by immunoblotting, with actin used as a loading control. (B) Principle of activity-based protein profiling with the PS89 photo probe covalently linked to its cellular targets (adapted from Vomacka et al.30). (C) Volcano plot of target proteins enriched by the PS89 photo probe versus dimethyl sulfoxide (DMSO) (left). Targets with >3-fold enrichment (log2 Probe/DMSO >1.6) and -log10 P-value >2 are highlighted. Ranking of targets according to the degree of competition by unmodified PS89 (right). (D,E) Target network analysis of the 42 most enriched PS89-binding proteins with STRING v10. Proteins involved in the most prominent gene ontology classes are highlighted (Cellular Component - green circles; Biological Process - green dots).

haematologica | 2019; 104(3)

549


F. Koczian et al.

cells with PS89 in combination with etoposide, vincristine or 6-mercaptopurine, respectively, activation of caspase-3 and PARP cleavage indicated a clear induction of apoptotic cell death (Figure 1C, Online Supplementary Figure S2). In addition, apoptosis was prevented by the pan-caspase inhibitor QVD-OPh (Online Supplementary Figure S3). Since pharmacokinetic studies demonstrated that PS89 has a very short half-life in the blood (data not shown), in vivo

experiments at not feasible at the moment. However, the broad applicability of PS89 as a chemosensitizing agent was confirmed in ALL and AML PDX cells of diverse backgrounds (Online Supplementary Table S2). PDX samples treated with PS89 and vincristine (Figure 1D) or PS89 and daunorubicin (Figure 1E) showed clearly increased apoptosis rates compared to those treated with the single cytostatics (distinct P-values are presented in Online

A

B

C

D

E

F

Figure 3. Signal transduction via the BAP31-caspase-8 axis. (A) Immunofluorescence staining of BAP31 (green) and localization of the PS89 photo probe linked to a rhodamine reporter (red) in HeLa cells. Representative original images (upper row) and areas of co-localization analyzed with Leica LAS X software (bottom row) are shown. (B) Diffusion values of PS89 and PS89+BAP31 showing an interaction between the probe and protein. Control samples of 50 nM freely diffusing PS89 in buffer solution and 50 nM PS89 probe plus 50 nM BAP31 protein in buffer solution were analyzed by single-point fluorescence correlation spectroscopy measurements. The diffusion coefficient of the probe alone was measured after one-species fitting of the autocorrelation curves (n>15). Diffusion coefficients (D1, D2) of S1 and S2 were measured after two-species fitting of the autocorrelation curves (n>10). The diffusion time (D2) of the mixtures was confined to the diffusion value obtained in the control experiment with PS89 probe alone. Bars represent the mean + standard error of mean. (C) Cleavage of caspase-8 (CASP8) and BAP31 was determined by immunoblotting in Jurkat cells treated with PS89 and etoposide (ETO) for 24 or 48 h. (D) Co-immunoprecipitation of BAP31 and CASP8 from Jurkat cell lysates after 24 h stimulation with PS89 and ETO. Blots were probed for BAP31 and pro- and intermediate p43/41 CASP8. (E) Apoptosis of Jurkat cells treated with the PS89 and ETO combination in the presence of the specific CASP8 inhibitor Z-IETD-FMK after 48 h. (F) Apoptosis of BAP31-silenced HeLa cells treated for 48 h with PS89 and ETO (6 h after transfection). The percentage of apoptotic cells was determined by FACS analysis and normalized to controls. The effect of the PS89 + ETO combination versus treatment with ETO alone was analyzed in siCtrl and siBAP31 cells (one-way ANOVA, Tukey test, P<0.05).

550

haematologica | 2019; 104(3)


PS89 - a novel option for combination therapy in acute leukemia Supplementary Table S3A,B). The Bliss independence model confirmed the synergistic effect of PS89 in combination with vincristine or daunorubicin in ALL and AML PDX cells. Nevertheless, in accordance with previous results, PS89 was non-toxic. Notably, compared to ALL and AML patients’ samples, both healthy peripheral blood mononuclear cells and CD34-positive hematopoietic stem cells showed weak responses to combination treatments (Figure 1D-F, Online Supplementary Table S3).

Proteomics identified a PS89 target network affecting endoplasmic reticulum homeostasis To elucidate the mechanisms underlying the impressive pro-apoptotic effect of PS89 combination treatments, the role of the prominent PS89 target, PDI, was studied by genetic knockdown and overexpression experiments. Since PDI silencing did not mimic and PDI overexpression did not rescue the sensitizing effect of PS89 on apoptosis induction or inhibition on proliferation (Figure 2A and

A Figure 4. Pro-apoptotic crosstalk at the endoplasmic reticulum-mitochondria interface. (A) Intracellular calcium levels of Jurkat cells treated with PS89 and etoposide (ETO) for 24 and 48 h. Fluorescence of Cal-520 stained cells was determined by FACS analysis and mean values normalized towards a dimethyl sulfoxide (DMSO) control. The dotted gray line represents unstained controls. (B) Mitochondrial depolarization of Jurkat cells treated with PS89 and ETO for 24 and 48 h. The percentage of JC-1-stained cells with dissipated versus intact membrane potential (ΔΨm) was determined by FACS analysis (populations as shown by FACS dot plots). (C) Cytochrome c release from mitochondria into the cytosol. Fractionation of Jurkat cell lysates after 48 h treatment with PS89 and ETO was confirmed by voltage-dependent anion channel (VDAC) immunoblotting. Stain-free gels served as the loading control. (D) Intracellular levels of reactive oxygen species (ROS) in Jurkat cells treated with PS89 and ETO for 24 h and 48 h. Fluorescence of carboxy-H2DCFDA-stained cells was determined by FACS analysis and mean values normalized to a dimethyl sulfoxide (DMSO) control. The dotted gray line represents unstained controls. (E) Apoptosis of Jurkat vector control (Jurkat/neo), Bcl-2 overexpressing (Jurkat/Bcl-2) and Bcl-xL overexpressing (Jurkat/Bcl-xL) cells treated with ETO and PS89 for 48 h. (F) Apoptosis of Jurkat cells treated with ABT 199 (0.5 - 50 mM) in the presence or absence of 25 mM PS89. The percentage of apoptotic cells was determined by FACS analysis after 48 h and synergism was calculated using the Bliss independence model.

B

C

D

E F

haematologica | 2019; 104(3)

551


F. Koczian et al.

Online Supplementary Figure S4), we assumed that PS89 affects additional cellular structures. In order to identify the proposed multi-target characteristics of PS89, we performed activity-based protein profiling in Jurkat cells, as depicted in Figure 2B. The PS89 photo probe modified by an alkyne handle (structure shown in Figure 2B) was covalently linked to cellular targets in the presence or absence of PS89. Co-incubation with the unmodified compound acting as a competitor was performed to exclude the identification of targets that were only enriched by the probe, but not by PS89. Including both datasets as well as their reproducibility expressed as P-values (cutoff criteria described in the Methods section), a total of 42 target proteins were identified (Figure 2C and Online Supplementary Table S4). Performing protein-protein interaction analysis using the STRING database,31 23 out of the 42 PS89 target proteins were involved in a protein interaction network (Figure 2D). This was further analyzed by gene ontology functional classification for common cellular components and biological processes.32 A highly significant number of PS89 target proteins was assigned to be located in the ER (false discovery rate 5.9x10-12) and described to be involved in cellular homeostasis, in particular cell redox homeostasis (false discovery rate 9.1x10-7 and 6.7x10-9) (Figure 2D,E). In this way, BAP31, which is described to be involved in ER stress-mediated apoptosis signaling pathways,15 was identified as one of the most prominent target proteins (Figure 2C and Online Supplementary Table S4).

Apoptosis induced by PS89 combination treatments is mediated via the BAP31-caspase-8 axis To validate BAP31 as a direct target of PS89, co-staining was performed with a BAP31-specific antibody and a photo probe linked to a rhodamine reporter dye by click chemistry. Besides the supposed ER specificity, overlap-

ping fluorescence revealed distinct co-localized ER network structures of PS89 photo probe-rhodamine and BAP31-Alexa 488 (Figure 3A). No background staining of either the rhodamine-azide or the Alexa 488 secondary antibody was detected (Online Supplementary Figure S5). Direct binding of PS89 photo probe to BAP31 was further evaluated by single-point fluorescence correlation spectroscopy, a technique in which random motion of fluorescent molecules into and out of a stationary laser focus results in fluctuations in fluorescence intensity, which can be monitored by confocal laser scanning microscopy. Hence, diffusion and concentration values of the PS89 photo probe with or without different amounts of recombinant BAP31 protein can be calculated by fitting of the autocorrelation curves (Figure 3B, Online Supplementary Figure S6). As shown in Figure 3B and Online Supplementary Figure S6, whereas the PS89 photo probe alone has a distinct diffusion value of ~274 mm2/s, two diffusing components were detected in the presence of recombinant BAP31 protein (D2 and D1). The fast diffusing species D2 characterizes remaining unbound photo probe (showing a diffusion value of ~264 mm2/s), whereas the slowly diffusing part D1 reflects the PS89 bound to BAP31. This decrease of PS89 diffusion after addition of the BAP31 protein indicates a strong, direct interaction between the two and the resulting diffusion value of ~70 mm2/s is in agreement with literature values for proteins of that size.33 Of note, no significant change in concentration was observed after addition of the protein in the solution (Online Supplementary Figure S6D). As the BAP31 protein complex has been shown to serve as a platform for caspase-8 activation upon apoptotic stimuli,14 the influence of PS89 on caspase-8 activation after etoposide treatment was examined. Whereas stimulation with etoposide and PS89 alone had only modest

Figure 5. Communication at the endoplasmic reticulum-mitochondria interface in cells treated with PS89 in combination with cytostatics. For details see text. PDI: protein disulfide isomerase; BAP31: B-cell receptorassociated protein 31; ER: endoplasmic reticulum; ROS: reactive oxygen species; Î&#x201D;Ψm: mitochondrial membrane potential; PARP: poly (ADP-ribosome) polymerase.

552

haematologica | 2019; 104(3)


PS89 - a novel option for combination therapy in acute leukemia effects on caspase-8 cleavage, the combination of both compounds resulted in strong activation of caspase-8 in Jurkat, CCRF-CEM as well as ALL PDX samples (Figure 3C, Online Supplementary Figure S7A,B). Moreover, cleavage of BAP31 into the pro-apoptotic p20BAP31 fragment or decreased expression of BAP31 proform was only present in Jurkat, CCRF-CEM and ALL PDX cells treated with the PS89 combination and, as an early trigger, already detectable after 24 h (Figure 3B, Online Supplementary Figure S7). Treatment-independent interaction of procaspase-8 with BAP31 could be demonstrated by coimmunoprecipitation (Figure 3D). Interestingly, binding of cleaved p43/p41 caspase-8 to BAP31 was only detected in the presence of both stimulants, PS89 and etoposide (Figure 3D, for normalization to BAP31 see Online Supplementary Figure S8). In order to investigate whether the induction of apoptosis by the combination of PS89 and etoposide is critically dependent on caspase-8 activity, cells were stimulated in the presence of the specific and irreversible caspase-8 inhibitor Z-IETD-FMK. As shown in Figure 3E, inhibiting the activity of caspase-8 resulted in diminished apoptosis upon treatment with etoposide and PS89. In accordance, the ratio of apoptotic cells among the cells treated with the PS89 and etoposide combination versus etoposide alone was reduced in caspase-8-deficient Jurkat cells compared to wildtype cells (1.7-fold in CASP8-/- versus 2.8 fold in wildtype Jurkat cells (Online Supplementary Figure S9). Next, we examined the functional effects of impairing the expression of the direct PS89 target BAP31 by short interfering (si) RNA. Whereas PS89 significantly enhanced etoposide-triggered apoptosis in control cells, no synergistic effect on apoptosis induction upon treatment with PS89 and etoposide could be detected in siBAP31-transfected cells (Figure 3F).

Endoplasmic reticulum and mitochondrial stress triggers are amplified by PS89 To investigate the consequences of PS89 and etoposide at the ER-mitochondria interface, calcium release from the ER into the cytosol was evaluated by FACS analysis. A shifted Cal 520 fluorescence intensity, indicating higher levels of cytosolic calcium, was observed in Jurkat, CCRFCEM and HL-60 and ALL PDX cells treated with PS89 in combination than in those treated with etoposide alone (Figure 4A, Online Supplementary Figure S10). Interestingly, amplification of calcium release was observed from 24 h to 48 h in cells exposed to PS89 combination treatment, but not in cells treated with etoposide, daunorubicin or vincristine alone. Loss of mitochondrial membrane integrity, analyzed by JC-1 staining, was increased by coincubation of etoposide with PS89 (Figure 4B) resulting in release of cytochrome c into the cytosol (Figure 4C, Online Supplementary Figure S11) and elevated levels of reactive oxygen species in Jurkat, CCRF-CEM and HL-60 cells (Figure 4D, Online Supplementary Figure S12). Reactive oxygen species signaling from mitochondria to the ER provokes further disturbance of ER redox homeostasis and finally closes the feedback loop. The eminent role of functional mitochondrial apoptosis signaling was reinforced by studying stable leukemia cells overexpressing the anti-apoptotic proteins Bcl-2 and BclxL. Interestingly, these clones showed a significantly lower sensitivity towards PS89 in combination with etoposide than the empty vector cell line Jurkat/neo (Figure 4E). Moreover, specific targeting of mitochondria with the haematologica | 2019; 104(3)

Bcl-2 inhibitor ABT-199 resulted in synergistic apoptosis in combination with PS89 (Figure 4F, Bliss values are shown in Online Supplementary Table S5), thus substantiating the importance of cytostatic-induced mitochondrial damage provoking the chemosensitizing effect of PS89.

The endoplasmic reticulum-mitochondria interface mediates mutual amplification of PS89 and cytostatic-triggered apoptosis In summary, PS89 strongly increases mitochondrial apoptosis through a crosstalk and mutual amplification of pro-apoptotic stress signals triggered by cytostatics (Figure 5). The polypharmacological profile of PS89 affecting a network of ER homeostasis proteins is represented by its main targets PDI and BAP31. Upon apoptotic stimuli of cytostatics and exclusively in the presence of PS89, BAP31 is cleaved by caspase-8 to pro-apoptotic p20BAP31. Calcium release from the ER and increased ER stress promote loss of mitochondrial membrane potential (Î&#x201D;Ψm) and apoptosis. In turn, elevated production of reactive oxygen species feeds back to the ER and provokes further ER stress and calcium release. The mutual amplification of ER-mitochondrial stress triggers finally leads to synergistic activation of caspases and apoptosis.

Discussion In the present study, we demonstrated that activating the apoptotic machinery at the ER-mitochondria interface is a highly promising approach for combination drug treatment. Co-stimulation of cytostatics with subtoxic doses of the novel PDI inhibitor PS89 resulted in a highly synergistic apoptotic response in a broad range of ALL and AML cell lines and human xenograft cells derived from both newly diagnosed and relapsed patients. In order to exploit this successful strategy, our work examined the intriguing question of how a drug at non-toxic concentrations could become highly effective in combination with cytostatics. The small molecule PS89 was previously identified as a potent chemosensitizing agent, which inhibits PDI.22 Notably, although PDI plays a key role in the maintenance of oxidative protein folding in the ER, no induction of ER stress or unfolded protein response was observed when applying PS89 alone, but only when the molecule was administered in combination with etoposide. This indicated that activation of the ER stress response results from the disability of the ER to resolve a stressful condition which is provoked in cooperation with cytostatics.22 As it is known that many cytostatics induce cell death via activation of the mitochondrial apoptosis pathway34 and, moreover, an increasing number of studies indicate a pivotal role of ER-mitochondria communication in cell fate decisions,35-37 we investigated a potential PS89-triggered crosstalk between ER stress and mitochondrial damage. Activity-based protein profiling conducted in Jurkat ALL cells identified that, besides PDI, PS89 targets a network of proteins located at the ER, including BAP31. Interestingly, it has been described that etoposide stimulates caspase-8-mediated cleavage of BAP31 to pro-apoptotic p20BAP31 at the ER-mitochondria interface, which results in calcium release from the ER and induction of mitochondrial apoptosis.14 Moreover, under conditions of ER stress, BAP31 interacts with cell death-inducing p53553


F. Koczian et al.

target protein 1 (CDIP1) leading to cleavage of BAP31, recruitment of Bcl-2 and mitochondrial apoptosis via Bax oligomerization.15 We, therefore, presumed that by inhibiting PDI and further ER stress-related proteins in addition to directly targeting BAP31, PS89 tunes pro-apoptotic feedback from the ER to mitochondria, which results in amplification of cell death signaling in combination therapy. To confirm the central role of the caspase-8-BAP31 axis in PS89 combination treatment, the processing and activation of the respective proteins was investigated. Indeed, BAP31 cleavage in co-stimulated cells could be observed already at early time points. Interestingly, we were able to show for the first time the intermediate p43/41 cleavage product of caspase-8 associated with BAP31, which still holds a death effector domain. This supports the suggestion that further processing into the finally active p18 fragment does in fact happen at the BAP31 complex.14 As we detected the p43/41 caspase-8 association as well as BAP31 cleavage only in cells exposed to PS89 combination treatment and silencing of BAP31 rescued the chemosensitizing effect of PS89, we conclude that BAP31 binding is crucial for PS89 to mediate efficient ER-mitochondria communication. Subsequently, an amplification of calcium release was shown in cells exposed to PS89 combination treatment, but not in those treated with etoposide alone. Ultimately, mitochondria-directed calcium flux promotes mitochondrial outer membrane permeabilization, accumulation of reactive oxygen species and release of cytochrome c,13,46 which were demonstrated to be increased in PS89 costimulated cells as well. Hence, PS89 is able to augment apoptotic triggers of cytostatics by interfering with the ER-mitochondria feedback loop. It is noteworthy that cells stably overexpressing the anti-apoptotic mitochondrial proteins Bcl-2 and Bcl-xL are less sensitive to the combination treatment, presumably because of an impaired mitochondrial apoptosis machinery. With reference to the concept of the ER-mitochondria ‘social network of cell death’,13 it is conceivable that the ER-mitochondrial feedback loop procures the crucial pro-apoptotic amplification effect by compromising numerous mitochondria, even if the original stimulus targeted only a few. In order to further comprehend how stress triggers are communicated from mitochondria to the ER and back, closer examination of the BAP31 complex is required. As shown in previous studies, Fis1 bridges mitochondria and ER-located BAP31 which seems to be further under the control of ER stress-inducible CDIP1 as well as the anti-

References 1. Ward E, DeSantis C, Robbins A, et al. Childhood and adolescent cancer statistics, 2014. CA Cancer J Clin. 2014;64(2):83-103. 2. Pui C-H, Robison LL, Look AT. Acute lymphoblastic leukaemia. Lancet. 2008;371 (9617):1030-1043. 3. Döhner H, Weisdorf DJ, Bloomfield CD. Acute myeloid leukemia. New Engl J Med. 2015;373(12):1136-1152. 4. Kavallaris M. Microtubules and resistance to tubulin-binding agents. Nat Rev Cancer. 2010;10(3):194-204. 5. Nitiss JL. Targeting DNA topoisomerase II in

554

apoptotic proteins, Bcl-2 and Bcl-xL.14,15 However, the dynamics regulating the balance of pro- and anti-apoptotic proteins within the complex have not been clarified yet. As PS89 is, to our knowledge, the first BAP31-binding small-molecule compound, which facilitates BAP31 cleavage, it might serve as a valuable tool not only for studying the dynamics of the BAP31 protein complex and manipulating decisive BAP31 interactions that favor the pro-apoptotic output, but also for enabling in-depth characterization of BAP31 as a prospective pharmacologically addressable target protein in different diseases. With reference to hematologic malignancies, this is further encouraged by the finding that overexpression of BAP31 seems to correlate with chemoresistance, as shown in fludarabine-resistant mantle cell lymphoma38 as well as proteasome inhibitor-adapted myeloma cells.39 In terms of prospective anti-cancer therapies, targeting pro-apoptotic ER-mitochondria crosstalk by combinatory pharmaceutical intervention offers versatile options. For example, BH3 mimetics are a valuable novel class of compounds that trigger intrinsic apoptosis and encouraging results were recently shown in an AML phase II trial with ABT-199.40 As shown here, ABT-199 in combination with PS89 strongly increases cell death in Jurkat cells compared to the cell death induced by the agents administered singly. This further underlines the concept of amplification by communication between ER and mitochondria as a promising strategy for developing new drugs able to trigger apoptosis and overcome therapy resistance. Moreover, besides PDI-targeting agents, proteasome or HSP90 inhibitors are potentially highly promising candidates for combination with mitochondria-damaging substances to tune the pro-apoptotic ER stress response.41 In conclusion and in response to the question of how PS89 is able to sensitize acute leukemia cells, the ER-mitochondria interface was identified as the key platform in the pro-apoptotic signaling cascades mediating the cytotoxic effects of PS89 in combination with cytostatics. By directly affecting PDI and BAP31, PS89 mutually amplifies ER and mitochondrial stress triggers, resulting in strong chemosensitizing effects. Hence, this study reveals the potential of targeting the ER-mitochondria apoptosis network as a novel and encouraging strategy in anti-cancer therapy. Acknowledgments We thank Judith Hoffmann for the supply of PS89 and the photo probe, and Kerstin Loske as well as Silvia Schnegg for technical assistance. The project was financially supported by the Dr. Robert Pfleger foundation.

cancer chemotherapy. Nat Rev Cancer. 2009;9(5):338-350. 6. Dieck CL, Tzoneva G, Forouhar F, et al. Structure and mechanisms of NT5C2 mutations driving thiopurine resistance in relapsed lymphoblastic leukemia. Cancer Cell. 2018;34(1):136-147.e6. 7. Nicholson L, Evans CA, Matheson E, et al. Quantitative proteomic analysis reveals maturation as a mechanism underlying glucocorticoid resistance in B lineage ALL and re-sensitization by JNK inhibition. Br J Haematol. 2015;171(4):595-605. 8. Jia J, Zhu F, Ma X, et al. Mechanisms of drug combinations: interaction and network per-

9.

10.

11. 12.

spectives. Nat Rev Drug Discov. 2009;8(2): 111-128. Barabási A-L, Gulbahce N, Loscalzo J. Network medicine: a network-based approach to human disease. Nat Rev Genet. 2011;12(1):56-68. Zimmermann GR, Lehar J, Keith CT. Multitarget therapeutics: when the whole is greater than the sum of the parts. Drug Discov Today. 2007;12(1):34-42. Hopkins AL. Network pharmacology: the next paradigm in drug discovery. Nat Chem Biol. 2008;4(11):682-690. Eletto D, Chevet E, Argon Y, et al. Redox controls UPR to control redox. J Cell Sci.

haematologica | 2019; 104(3)


PS89 - a novel option for combination therapy in acute leukemia 2014;127(17):3649-3658. 13. Grimm S. The ER–mitochondria interface: the social network of cell death. Biochim Biophys Acta. 2012;1823(2):327-334. 14. Iwasawa R, Mahul‐Mellier AL, Datler C, et al. Fis1 and Bap31 bridge the mitochondria– ER interface to establish a platform for apoptosis induction. EMBO J. 2011;30(3):556-568. 15. Namba T, Tian F, Chu K, et al. CDIP1-BAP31 complex transduces apoptotic signals from endoplasmic reticulum to mitochondria under endoplasmic reticulum stress. Cell Rep. 2013;5(2):331-339. 16. Rosati E, Sabatini R, Rampino G, et al. Novel targets for endoplasmic reticulum stressinduced apoptosis in B-CLL. Blood. 2010; 116(15):2713-2723. 17. Galligan JJ, Petersen DR. The human protein disulfide isomerase gene family. Hum Genomics. 2012;6:6. 18. Xu S, Sankar S, Neamati N. Protein disulfide isomerase: a promising target for cancer therapy. Drug Discov Today. 2014;19(3): 222-240. 19. Haefliger S, Klebig C, Schaubitzer K, et al. Protein disulfide isomerase blocks CEBPA translation and is up-regulated during the unfolded protein response in AML. Blood. 2011;117(22):5931-5940. 20. Higa A, Taouji S, Lhomond S, et al. Endoplasmic reticulum stress-activated transcription factor ATF6α requires the disulfide isomerase PDIA5 to modulate chemoresistance. Mol Cell Biol. 2014;34(10):1839-1849. 21. Trivedi R, Müller G, Rathore M, et al. Antileukemic activity of shikonin: role of ERP57 in shikonin induced apoptosis in acute myeloid leukemia. Cell Physiol Biochem. 2016;39(2):604-616. 22. Eirich J, Braig S, Schyschka L, et al. A small molecule inhibits protein disulfide isomerase and triggers the chemosensitization of cancer cells. Angew Chem Int ED Engl. 2014;53(47):12960-12965. 23. Xu S, Butkevich AN, Yamada R, et al.

haematologica | 2019; 104(3)

24.

25.

26.

27.

28.

29.

30.

31.

Discovery of an orally active small-molecule irreversible inhibitor of protein disulfide isomerase for ovarian cancer treatment. Proc Natl Acad Sci USA. 2012;109(40):1634816353. Vatolin S, Phillips JG, Jha BK, et al. Novel protein disulfide isomerase inhibitor with anticancer activity in multiple myeloma. Cancer Res. 2016;76(11):3340-3350. Haber M, Norris MD, Kavallaris M, et al. Atypical multidrug resistance in a therapyinduced drug-resistant human leukemia cell line (LALW-2): resistance to vinca alkaloids independent of P-glycoprotein. Cancer Res. 1989;49(19):5281-5287. Verrills NM, Walsh BJ, Cobon GS, et al. Proteome analysis of vinca alkaloid response and resistance in acute lymphoblastic leukemia reveals novel cytoskeletal alterations. J Biol Chem. 2003;278(46):4508245093. Terziyska N, Alves CC, Groiss V, et al. In vivo imaging enables high resolution preclinical trials on patients’ leukemia cells growing in mice. PLoS One. 2012;7(12): e52798. Vick B, Rothenberg M, Sandhöfer N, et al. An advanced preclinical mouse model for acute myeloid leukemia using patients' cells of various genetic subgroups and in vivo bioluminescence imaging. PLoS One. 2015; 10(3):e0120925. Fukuda J, Kaneko T, Egashira M, et al. Direct measurement of CD34+ blood stem cell absolute counts by flow cytometry. Stem Cells. 1998;16(4):294-300. Vomacka J, Korotkov VS, Bauer B, et al. An aromatic hydroxyamide attenuates multiresistant Staphylococcus aureus toxin expression. Chemistry. 2016;22(5):1622-1630. Szklarczyk D, Franceschini A, Wyder S, et al. STRING v10: protein–protein interaction networks, integrated over the tree of life. Nucleic Acids Res. 2015;43(Database issue): D447-452.

32. Consortium GO. Gene ontology consortium: going forward. Nucleic Acids Res. 2015;43(Database issue):D1049-D1056. 33. Krouglova T, Vercammen J, Engelborghs Y. Correct diffusion coefficients of proteins in fluorescence correlation spectroscopy. Application to tubulin oligomers induced by Mg2+ and Paclitaxel. Biophys J. 2004;87(4): 2635-2646. 34. Chonghaile TN, Sarosiek KA, Vo T-T, et al. Pretreatment mitochondrial priming correlates with clinical response to cytotoxic chemotherapy. Science. 2011;334(6059): 1129-1133. 35. Bhat TA, Chaudhary AK, Kumar S, et al. Endoplasmic reticulum-mediated unfolded protein response and mitochondrial apoptosis in cancer. Biochim Biophys Acta Rev Cancer. 2017;1867(1):58-66. 36. Shore GC, Papa FR, Oakes SA. Signaling cell death from the endoplasmic reticulum stress response. Curr Opin Cell Biol. 2011;23 (2):143-149. 37. Simmen T, Herrera-Cruz MS. Cancer: untethering mitochondria from the ER? Front Oncol. 2017;7:105. 38. Lorkova L, Scigelova M, Arrey TN, et al. Detailed functional and proteomic characterization of fludarabine resistance in mantle cell lymphoma cells. PLoS One. 2015;10(8):e0135314. 39. Soriano G, Besse L, Li N, et al. Proteasome inhibitor-adapted myeloma cells are largely independent from proteasome activity and show complex proteomic changes, in particular in redox and energy metabolism. Leukemia. 2016;30(11):2198-2207. 40. Konopleva M, Pollyea DA, Potluri J, et al. A phase 2 study of ABT-199 (GDC-0199) in patients with acute myelogenous leukemia (AML). Blood. 2014;124(21):118. 41. Wang M, Kaufman RJ. The impact of the endoplasmic reticulum protein-folding environment on cancer development. Nat Rev Cancer. 2014;14(9):581-597.

555


ARTICLE Ferrata Storti Foundation

Haematologica 2019 Volume 104(3):556-563

Acute Lymphoblastic Leukemia

Trypsin-encoding PRSS1-PRSS2 variations influence the risk of asparaginase-associated pancreatitis in children with acute lymphoblastic leukemia: a Ponte di Legno toxicity working group report

Benjamin O. Wolthers,1 Thomas L. Frandsen,1 Chirag J. Patel,2 Rachid Abaji,3 Andishe Attarbaschi,4 Shlomit Barzilai,5 Antonella Colombini,6 Gabriele Escherich,7 Marie Grosjean,8 Maja Krajinovic,3,9 Eric Larsen,10 Der-Cherng Liang,11 Anja Möricke,12 Kirsten K. Rasmussen,1 Sujith Samarasinghe,13 Lewis B. Silverman,14 Inge M. van der Sluis,15 Martin Stanulla,16 Morten Tulstrup,1 Rachita Yadav,8,17 Wenjian Yang,18 Ester Zapotocka,19 Ramneek Gupta8 and Kjeld Schmiegelow1,20 on behalf of the Ponte di Legno toxicity working group

Department of Pediatrics and Adolescent Medicine, University Hospital Rigshospitalet, Copenhagen, Denmark; 2Department of Biomedical Informatics, Harvard Medical School, Boston, MA, USA; 3CHU Sainte-Justine Research Center and Department of Pharmacology, University of Montreal, QC, Canada; 4Department of Pediatric Hematology and Oncology, St Anna Children's Hospital and Department of Pediatric and Adolescent Medicine, Medical University of Vienna, Austria; 5Pediatric Hematology and Oncology, Schneider Children's Medical Center of Israel, Petah-Tikva, Israel and Sackler Faculty of Medicine, Tel Aviv University, Israel; 6Department of Pediatrics, Ospedale San Gerardo, University of Milano-Bicocca, Fondazione MBBM, Monza, Italy; 7University Medical Center Eppendorf, Clinic of Pediatric Hematology and Oncology, Hamburg, Germany; 8Center for Biological Sequence Analysis, Technical University of Denmark, Lyngby, Denmark; 9Department of Pediatrics, University of Montreal, QC, Canada; 10 Maine Children's Cancer Program, Scarborough, ME, USA; 11Division of Pediatric Hematology-Oncology, Mackay Memorial Hospital, Taipei, Taiwan; 12Christian-AlbrechtsUniversity Kiel and University Medical Center Schleswig-Holstein, Department of Pediatrics, Kiel, Germany; 13Great Ormond Street Hospital for Children, London, UK; 14 Department of Pediatric Oncology, Dana-Farber Cancer Institute and Division of Hematology/Oncology, Boston Children's Hospital, Boston, MA, USA; 15Dutch Childhood Oncology Group, The Hague and Princess Máxima Center for Pediatric Oncology, Utrecht, the Netherlands; 16Department of Pediatric Hematology and Oncology, Hannover Medical School, Germany; 17Molecular Neurogenetics Unit, Psychiatric and Neurodevelopmental Genetics Unit, Center for Genomic Medicine, Massachusetts General Hospital, Boston, MA, USA; 18St. Jude Children's Research Hospital, Department of Pharmaceutical Sciences, Memphis, TN, USA; 19University Hospital Motol, Department of Pediatric Hematology/Oncology, Prague, Czech Republic and 20Institute of Clinical Medicine, University of Copenhagen, Denmark. 1

Correspondence: KJELD SCHMIEGELOW Kjeld.Schmiegelow@regionh.dk Received: June 6, 2018. Accepted: November 16, 2018. Pre-published: November 22, 2018. doi:10.3324/haematol.2018.199356 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/104/3/556 ©2019 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

556

ABSTRACT

A

sparaginase-associated pancreatitis is a life-threatening toxicity to childhood acute lymphoblastic leukemia treatment. To elucidate genetic predisposition and asparaginase-associated pancreatitis pathogenesis, ten trial groups contributed remission samples from patients aged 1.0−17.9 years treated for acute lymphoblastic leukemia between 2000 and 2016. Cases (n=244) were defined by the presence of at least two of the following criteria: (i) abdominal pain; (ii) levels of pancreatic enzymes ≥3 x upper normal limit; and (iii) imaging compatible with pancreatitis. Controls (n=1320) completed intended asparaginase therapy, with 78% receiving ≥8 injections of pegylated-asparaginase, without developing asparaginase-associated pancreatitis. rs62228256 on 20q13.2 showed the strongest association with the development of asparaginase-associated pancreatitis (odds ratio=3.75; P=5.2x10-8). Moreover, rs13228878 (OR=0.61; P=7.1x10-6) and rs10273639 (OR=0.62; P=1.1x10-5) on 7q34 showed significant association with the risk of asparaginase-associated pancreatitis. A Dana Farber Cancer Institute ALL Consortium cohort consisting of haematologica | 2019; 104(3)


Pancreatitis in childhood ALL

patients treated on protocols between 1987 and 2004 (controls=285, cases=33), and the Children’s Oncology Group AALL0232 cohort (controls=2653, cases=76) were available as replication cohorts for the 20q13.2 and 7q34 variants, respectively. While rs62228256 was not validated as a risk factor (P=0.77), both rs13228878 (P=0.03) and rs10273639 (P=0.04) were. rs13228878 and rs10273639 are in high linkage disequilibrium (r2=0.94) and associated with elevated expression of the PRSS1 gene, which encodes for trypsinogen, and are known risk variants for alcohol-associated and sporadic pancreatitis in adults. Intra-pancreatic trypsinogen cleavage to proteolytic trypsin induces autodigestion and pancreatitis. In conclusion, this study finds a shared genetic predisposition between asparaginase-associated pancreatitis and non-asparaginase-associated pancreatitis, and targeting the trypsinogen activation pathway may enable identification of effective interventions for asparaginase-associated pancreatitis.

Introduction Intensification of chemotherapy for childhood acute lymphoblastic leukemia (ALL) has generated 5-year survival rates greater than 90%, but has been associated with an increase in therapy-related toxicity.1 Asparaginase is a key drug in the treatment of childhood ALL and there is growing interest in its use as an anti-metastatic agent in breast cancer.2 Asparaginase depletes the body of the nonessential amino acid asparagine through deamidation of asparagine into aspartic acid and ammonia,3 and targets protein synthesis in malignant lymphoblasts by impairing the ability to synthesize asparagine.4,5 Pancreatitis associated with asparaginase therapy (AAP) is a frequent toxicity affecting 4−10% of children treated on contemporary ALL protocols, and is associated with severe complications.6–9 In addition, re-exposure to asparaginase after AAP is associated with a high risk (~50%) of a second episode of AAP, and thus AAP often entails truncation of asparaginase therapy, thereby decreasing the patients’ chance of survival.4,5,9 The mechanism(s) by which asparaginase causes pancreatitis are elusive, thus hampering attempts to identify patients with an altered risk of AAP.10 The Ponte di Legno toxicity working group, therefore, initiated a study with three main purposes: (i) to define diagnostic consensus criteria for AAP;11 (ii) to describe the phenotype of AAP in patients across multiple ALL trial groups;9 and (iii) to explore genotype-phenotype associations, using a genome-wide approach, to identify patients with altered risk of AAP.9,11 Genome-wide association (GWA) studies are agnostic by design, reporting phenotype-genotype associations without prior hypotheses and often including speculative mechanisms. Replication of GWA study results are thus a requisite for credibility Accordingly, this study presents results from the largest AAP GWA study so far, with a strong focus on investigating previously validated variants associated with non-asparaginase-induced pancreatitis and replicating top results in similar childhood ALL cohorts.

Methods Study design and participants Ten international childhood ALL trial groups (Online Supplementary Table S1) contributed to the discovery cohort. Postremission DNA was collected from June 2015 to January 2017, three groups collected DNA on AAP cases only while seven haematologica | 2019; 104(3)

groups did so on both cases and controls (Online Supplementary Figure S1). The database containing phenotype data was approved by the regional ethical review board of The Capital Region of Denmark (H-2-2010-022), the Danish Data Protection Authorities (j.nr.: 2012-58-0004), and by relevant regulatory authorities in all participating countries. Genotype data were stored at the Technical University of Denmark’s server Computerome.12 Children (aged 1.0–17.9 years) with newly diagnosed ALL between January 2000 and January 2016 were eligible, irrespective of ethnicity. Pancreatitis was defined as asparaginase-associated if diagnosed within 50 days of the last injection of native E. coli asparaginase or polyethylene glycolated E. coli asparaginase (PEGasparaginase) and cases fulfilled the Ponte di Legno toxicity working group consensus definition for AAP: i.e., at least two of (i) amylase, pancreatic amylase, or pancreatic lipase ≥3 x upper normal limit; (ii) abdominal pain; and (iii) imaging compatible with imaging compatible with pancreatitis. All controls received the planned amount of asparaginase therapy in their respective protocols, with more than 78% (1024/1320) receiving at least eight injections of PEG-asparaginase without developing AAP. A subset of 62 AAP cases was previously included in a Nordic Society of Pediatric Hematology and Oncology (NOPHO) GWA study.13 These samples were genotyped on identical genotyping arrays as the remaining cohort, and raw genotyping data on these patients were pooled with those of the remaining cohort prior to quality control, and association analyses were done in one cohort.

Genotyping Post-remission DNA was genotyped by Aros Applied Biotechnology A/S (Aarhus, Denmark) on Illumina Omni2.5exome-8 BeadChip arrays using the human genome assembly (GRCh37) for reference. Quality control was performed using the PLINK tool,14 and single nucleotide polymorphisms (SNPs) were annotated in Ensembl Variant Effect Predictor GCRCh37.15 Alleles given are refSNP alleles according to dbSNP (not necessarily the alleles supplied by the Illumina map).16

Quality control Quality control was performed according to previously published criteria17 (Online Supplementary Figures S2-4), excluding individuals with: (i) a discordance in number of X chromosomes between geno- and phenotypes; (ii) missing data on >3% of SNPs; (iii) excess heterozygosity between autosomal SNPs; and (iv) high relatedness between samples. SNPs were excluded based on: (i) missing data on >2% of individuals (call rate); (ii) Hardy-Weinberg equilibrium; (iii) minor allele frequency <0.01; (iv) difference in call rate between cases and controls (Fisher exact test P<1.10-5); and (v) duplicated genomic position. 557


B.O. Wolthers et al.

Replication Three top SNPs were tested for validation in two separate cohorts. The Children’s Oncology Group AALL0232 ALL cohort included previously genotyped data on PRSS1-2 variants (but not on the NFATC2 variant). The NFATC2 variant was genotyped de novo in a cohort of patients from the Dana Farber Cancer Institute (DFCI) ALL Consortium protocols 87−01, 91−01, 95−01 and 00−01 (1987-2004). The AALL0232 ALL cohort included 76 cases diagnosed using National Cancer Institute’s Common Terminology Criteria for Adverse Events (CTCAE criteria) (Online Supplement pages 6−9) and 2577 controls.8 The cohort is described in detail in the appendix.8 The AALL0232 cohort was genotyped on the Affymetrix Genome-Wide Human SNP 6.0 Array, and imputed genotypes were generated using 1000 Genomes as the reference population as reported (ref: PMID 26265699). Timedependent analysis (Cox proportional hazards regression) was performed adjusting for age and ancestry. The cohort of patients treated on the DFCI ALL Consortium protocols 87−01, 91−01, 95−01 and 00−01 received 20−30 weeks of post-induction asparaginase therapy (Online Supplement pages 9-11). Thirty-three cases diagnosed according to the CTCAE criteria and 285 controls were included in this cohort, and genotyped by allele-specific oligonucleotide hybridization as described elsewhere.18,19 The Pearson correlation coefficient was used to investigate associations between genotype and AAP.

performed in the statistical program R version 3.3.3.20 Linear and logistic regression analyses were used to identify associations between genotypes and continuous or categorical clinical variables adjusting for age and ancestry. Genotype was treated as a numerical value (0, 1 or 2 minor alleles) for additive effect. The KaplanMeier method was used to estimate probability of event-free survival according to genotypes, and differences were compared with the two-sided log-rank test. No prior sample size calculations were applied for pre-study power calculations. Two-sided P-values below 0.05 were regarded as statistically significant. SNPs were annotated to genes 10 kb up- or downstream from transcription start- and end-sites, respectively, and all SNPs with a P-value below 5 x 10-5 were manually inspected for associations with genes and pathways previously related to pancreatitis. Investigated SNPs were explored using dbSNP16 and Ensembl,15 linkage disequilibrium (LD) between SNPs by the National Cancer Institute LDassoc tool,21 tissue expression (expressive quantitative trait loci) by GTeX,22 regulatory effect by RegulomeDB,23 and regional association plots were produced by the LocusZoom tool.24 Genes and SNPs previously associated with pancreatitis were investigated by searching PubMed for reports published in English within the last 10 years, using the search terms “pancreatitis” AND “genome” OR “genetic” OR “genotype” in the title. Gene functions were defined by Genecards25 (www.genecards.org).

Statistical analysis

Results

Association analysis was done in PLINK using logistic regression, assuming an additive genetic model, and adjusting for genetic ancestry and age. Genetic ancestry was determined by clustering analysis, and non-CEU ancestry was defined as individuals >16 standard deviations away from the HapMap-defined CEU (Northern and Western European) centroid mean. Using this model, multidimensional scaling plots showed an equal distribution of cases and controls according to ancestry (Online Supplementary Figure S3) and QQ plots showed no sign of population substructure (λ = 1.02, Online Supplementary Figure S4). Statistical analysis of phenotype and genotype associations was

After quality control filtering, 244 cases, 1320 controls and 1401908 SNPs were eligible for association analysis. Two hundred and five of the 244 (84%) cases and 1185/1320 (90%) of controls were of European (CEU) ancestry (Online Supplementary Figure S4). The median age (interquartile range, IQR) of the cases was 8.1 years (IQR, 4.3 to 13.1) while that of the controls was 5.0 years (IQR, 3.0 to 9.0). Fifty-five percent of both cases (133/244) and controls (724/1320) were male. Figure 1 shows the significance of SNPs associated with

Figure 1.Manhattan plot. Manhattan plot showing single nucleotide polymorphisms (SNPs) associated with asparaginase-associated pancreatitis in 244 cases and 1320 controls. The x axis represents genomic location, and the y axis represents the P value for the SNP associations calculated using logistic regression adjusting for age and ancestry. Genes previously associated with pancreatitis are represented in color. SNPs are annotated to genes based on genomic location (10 kb upstream and downstream of the transcription start site and transcription terminator, respectively. The human assembly GRCh37 was used for reference.

558

haematologica | 2019; 104(3)


Pancreatitis in childhood ALL

AAP. The variant rs62228256 [reference allele=C, minor allele=T (C>T)] on 20q13.2 showed the strongest association with AAP [odds ratio (OR), 3.75; 95% confidence interval (95% CI): 2.33 to 6.04; P=5.2 x 10-8). rs62228256 is located 274 kilobase pairs upstream of Nuclear factor of activated T cells (NFATC2) and has been documented to be an expressive quantitative trait locus for this gene in pancreatic tissue (Online Supplementary Figure S5). The 30 SNPs most associated with AAP, with P-values of 5 x 10-5 or lower, included rs13228878 (A>G; OR, 0.61; 95% CI: 0.5 to 0.76; P=7.1 x 10-6) and rs10273639 (C>T; OR, 0.62; 95% CI: 0.5 to 0.77; P=1.1 x 10-5) (Table 1). These SNPs reside on the same haplotype and are in high LD (CEU population LD; r2=0.94) in the PRSS1-PRSS2 locus on chromosome 7 (Figure 2). PRSS1 and PRSS2 encode for the proteases cationic and anionic trypsinogen, respectively. Both minor alleles rs13228878_G and rs10273639_T reduce the risk of AAP. When performing association analysis in the CEU population of 205 AAP cases and 1185 controls, rs62228256 (OR, 3.75; 95% CI: 2.27 to 6.2;

P=2.47 x 10-7), rs13228878 (OR, 0.60; 95% CI: 0.48 to 0.76; P=2.1 x 10-5) and rs10273639 (OR, 0.62; 95% CI: 0.49 to 0.78; P=3.8 x 10-5) (Online Supplementary Figures S6-8 and Online Supplementary Table S2) remained strongly associated with AAP. Further investigation of previously validated SNPs within genes known to regulate trypsin activation26 (Online Supplementary Table S3) showed that AAP was associated with rs17107315 in pancreatic secretory trypsin inhibitor (SPINK1; OR, 2.87; 95% CI: 1.36 to 5.8; P=4 x 10-3), rs10436957 in chymotrypsin C (CTRC; OR, 0.69; 95% CI: 0.53 to 0.89; P=5 x 10-3) and rs4409525 in Claudin2 (CLDN2; OR, 1.41; 95% CI: 1.08 to 1.83; P=0.01) with all minor alleles altering AAP risk in the direction and with effects very similar to those previously reported. In a logistic regression model, testing whether the effect of the most associated PRSS1-PRSS2 variant (rs13228878) was modified by rs17107315 (SPINK1), rs10436957 (CTRC) and rs4409525 (CLDN2), no significant interactions were identified (P=0.48, P=0.95 and P=0.93, respectively).

Table 1. Top 30 single nucleotide polymorphisms.

SNP

Chr

Position

Major>minor allele

MAF cases

MAF controls

OR (95% CI)

rs62228256 rs7270119 rs16996276 rs62228230 rs934350 rs170623 rs75245362 rs368819120 rs4769201 rs7851954 rs61734424 rs9912225 rs7155612 rs2167730 rs80170196 rs62228228 rs5010616 rs12494164 rs16848986 rs34375180 rs7139808 rs12582343 rs13228878 rs6477109 rs74109922 rs1505495 rs1791520 rs10273639 rs4655107 rs55634345

20 20 20 20 13 9 14 12 13 9 19 17 14 8 19 20 12 3 3 12 13 12 7 9 13 4 18 7 1 4

50454447 50436587 50455925 50445082 103589776 101984936 95990645 71747240 22698015 6796167 50747533 4680732 95976755 78103417 50747159 50443845 71748290 164967758 164979570 71779640 22693228 71766297 142473466 6794938 103582300 172973580 22118315 142456928 23094454 19846813

C>T A>G A>C G>A A>G C>G C>T G>A>G C>T T>C A>G T>G T>C C>T G>A C>T A>C T>C G>A T>C A>G A>G C>T A>G A>C T>C C>T G>A G>A

0.07 0.07 0.07 0.07 0.32 0.38 0.11 0.48 0.04 0.28 0.08 0.10 0.12 0.24 0.08 0.06 0.48 0.23 0.22 0.49 0.04 0.48 0.35 0.29 0.12 0.10 0.32 0.35 0.13 0.46

0.02 0.02 0.02 0.02 0.22 0.28 0.05 0.37 0.01 0.40 0.04 0.05 0.06 0.34 0.04 0.02 0.37 0.14 0.14 0.38 0.01 0.38 0.44 0.40 0.06 0.17 0.23 0.44 0.24 0.34

3.75 (2.33-6.04) 3.64 (2.28-5.8) 3.64 (2.27-5.85) 3.54 (2.23-5.63) 1.84 (1.47-2.3) 1.78 (1.42-2.22) 2.46 (1.73-3.51) 1.67 (1.35-2.06) 4.66 (2.45-8.86) 0.59 (0.47-0.74) 2.75 (1.8-4.2) 2.42 (1.67-3.5) 2.24 (1.6-3.15) 0.57 (0.45-0.73) 2.73 (1.78-4.18) 3.15 (1.94-5.13) 1.63 (1.32-2) 1.77 (1.39-2.27) 1.77 (1.38-2.26) 1.61 (1.31-1.98) 4.59 (2.38-8.87) 1.62 (1.31-1.99) 0.61 (0.5-0.76) 0.61 (0.49-0.75) 2.13 (1.52-2.97) 0.48 (0.34-0.66) 1.65 (1.32-2.07) 0.62 (0.5-0.77) 0.53 (0.39-0.7) 1.58 (1.29-1.94)

P-value

Gene (distance from gene)

5.18x10-8 5.52x10-8 8.64x10-8 9.19x10-8 1.16x10-7 3.93x10-7 ALG2 (+0.69 kb) & SEC61B (0) -7 6.33x10 SCARNA13 (-9.05 kb) & SNHG10 (-8.6 kb) 2.04x10-6 2.67x10-6 3.02x10-6 KDM4C (0) 3.04x10-6 MYH14 (0) -6 3.10x10 TM4SF5 (0) & VMO1 (-7.8 kb) 3.15x10-6 3.69 x10-6 3.77x10-6 MYH14 (0) 3.80x10-6 4.76x10-6 4.83x10-6 5.19x10-6 5.53x10-6 5.59x10-6 5.90x10-6 7.06x10-6 PRSS2 (-6.44 kb) & PRSS3P2 (-5.29 kb) 7.64x10-6 KDM4C (0) 1.03x10-5 1.06x10-5 GALNTL6 (0) -5 1.12x10 1.13x10-5 PRSS1 (-0.4 kb) 1.16x10-5 EPHB2 (0) 1.19x10-5

Top 30 SNPs most associated with AAP in 244 cases and 1320 controls. The model used here includes covariates for age and genetic ancestry. SNPs were annotated to genes if â&#x2030;¤10 kb upstream (-) or downstream (+) of the transcription start site or transcription terminator, respectively. SNP: single nucleotide polymorphism; Chr,: chromosome; MAF: minor allele frequency; OR: odds ratio; CI: confidence interval.

haematologica | 2019; 104(3)

559


B.O. Wolthers et al.

Validation of results In validating our top SNP in the NFATC2 locus (rs62228256) and top SNPs in the PRSS1-PRSS2 locus (rs13228878 and rs10273639) we used two cohorts of children with ALL: one treated according to the DFCI ALL Consortium 87−01, 91−01, 95−01 and 00−01 protocols,27,28 and the other according to Children’s Oncology Group AALL0232 protocol8 (Online Supplement pages 6−12). Whereas the association between AAP and rs6228256 genotype was not replicated in the DFCI ALL Consortium cohort (P=0.77, Online Supplement page 10), both rs13228878 (hazard ratio, 0.68; 95% CI: 0.48 to 0.96; P=0.03) and rs10273639 (hazard ratio, 0.69; 95% CI: 0.49 to 0.98; P=0.04) were significantly associated with the risk of developing pancreatitis in the AALL0232 cohort (Online Supplement page 8). Two previous studies have investigated SNPs associated with pancreatitis in children with ALL. Using different diagnostic criteria for pancreatitis and including controls with less than 5 weeks of asparaginase therapy, Liu et al. associated the rare variant (general population minor allele frequency=0.009%) rs199695765 in carboxypeptidase A2 encoding CPA2 (hazard ratio, 587; 95% CI: 66.8 to 5166; P=9 x 10-9) with AAP.8 We did not directly genotype this SNP, and none of the genotyped SNPs in the CPA2 region was in LD with rs199695765. Of 32 SNPs 10 kb up- and down-stream of CPA2, rs66839817 (T>C, OR, 1.28; 95% CI: 1.03 to 1.57; P=0.02) showed the strongest association. In a NOPHO GWA study we previously found the ULK2 variant rs281366 (P=5.8 x 10-7) and RGS6 variant rs17179470 (P=1.3 x 10-6) to be most associated with AAP.13 Excluding cases and controls from the NOPHO study, we failed to validate these results in 184 cases and 712 controls. Both rs281366 and rs17179470 were directly genotyped in this cohort with non-significant P-values (P=0.84 and P=0.32, respectively).

Genotype-phenotype associations Table 2 shows associations between PRSS1-PRSS2 genotype and amount of PEG-asparaginase given prior to AAP, days from PEG-asparaginase injection to AAP, complications of pancreatitis, and risk of a second episode of

AAP after re-exposure to asparaginase. The risk allele was not associated with number of PEG-asparaginase injections prior to AAP or time from injection of PEG-asparaginase to diagnosis of AAP. Furthermore, the risk of developing acute complications was not found to be associated with PRSS1-PRSS2 genotype (Table 2). In the Nordic subset of cases (n=92) and controls (n=1024) we found no association between PRSS1-PRSS2 genotype and 5-year event-free survival (Online Supplementary Figure S9; P=0.4). Out of 46 children who were re-exposed to asparaginase, 17 (37%) developed a second episode of AAP; although not in a statistically significant manner, the PRSS1-PRSS2 minor allele indicated the same protective effect as in the risk of initial AAP (OR, 0.49; 95% CI: 0.15 to 1.41; P=0.20).

Discussion This study is the first to find and validate variants in the PRSS1-PRSS2 locus associated with the risk of AAP in children with ALL. In doing so, we found that activation of trypsin within pancreatic acinar cells is a key initiating event in the pathogenesis of pancreatitis, regardless of the exposure i.e. alcohol, hyperlipidemia, or asparaginase. The role of trypsin activation in the pathogenesis of pancreatitis had long been suspected, but an underlying genetic susceptibility was not documented until 1996 when Whitcomb et al. documented mutations in the PRSS1 gene causing hereditary pancreatitis,29 and later associated a common genetic variant in the PRSS1-PRSS2 locus (rs10273639) with the risk of alcohol-related and sporadic pancreatitis.30 This association was recently validated in larger European and Asian cohorts, and the haplotype has been studied in detail.31–33 rs10273639 is located 408 base pairs upstream of the translation initiation codon of cationic trypsinogen. A recent functional study documented that the proximal rs4726576 (C>A) variant (204 kb upstream of the translation initiation codon) drives the association.34 The rs4726576 and rs10273639 variants are in high LD (r2 >99%) in European and Asian populations but have a r2 = 0.8 in the African meta-population, and

Figure 2. Regional association plot of the PRSS1-2 locus on chromosome 7. Regional asssociation plot showing single nucleotide polymorphisms (SNPs) associated with asparaginase-associated pancreatitis in 244 cases and 1320 controls. The x axis represents genomic location, and the y axis represents the P values for the SNP associations calculated using logistic regression adjusting for age and ancestry. rs13228878 (P=7.1 x 10-6) is represented in purple and rs10273639 (P=1.1 x 10-5) in red. The color of the dots reflects linkage disequilibrium (LD) of the genotyped SNPs. LD is based on 1000 genomes European samples, November 2014. The human assembly GRCh37 was used for reference.

560

haematologica | 2019; 104(3)


Pancreatitis in childhood ALL

pinpointing the association driving the signal is thus of importance in the latter population. Sequencing the PRSS1-PRSS2 risk allele has not revealed disease-associated coding variants accounting for the association with pancreatitis in GWA studies; however, the risk allele (rs4726576; rs10273639) is an expressive quantitative trait

locus for PRSS1 shown to elevate gene expression.30,34 The activation of trypsinogen is mediated by cleavage of the N-terminus extension of cationic trypsinogen (a calcium binding site) to active trypsin. Higher levels of calcium have been shown to lead to trypsin activation, and with higher expression of trypsin as seen in patients with

Figure 3. Schematic drawing illustrating the effects of asparaginase on pancreatic acinar cells. Schematic drawing illustrating the likely effect of asparaginase (red triangle) on the protease-activated receptor 2 (PAR2) receptor, and how this leads to increased calcium (Ca++) efflux from the endoplasmic reticulum. This in turn leads to opening of Ca++ release activated (CRAC) channels, further increasing intracellular calcium levels, reducing ATP levels and allowing activation of inactive trypsinogen to active trypsin. The drawing is heavily influenced by diagrams by Peng et al. (Phil Trans Royal Soc, 2015) and Whitcomb et al. (Nature Gen, 2012).

Table 2. Baseline characteristics, pancreatic enzyme levels and complications of pancreatitis in children according to the PRSS1-2 (rs13228878) genotype.

Heterozygote for rs13228878_A risk allele n=109

Homozygote for rs13228878_G non- risk allele n=30

P

7.75 (4–12.06) 53% (55/104) 397 (263–673; n= 43)

9.21 (4.46–13.43) 55% (60/109) 382 (206–671; n=51)

6.22 (4.9–12.1) 57% (17/30) 222 (151–617; n=14)

0.27

1255 (758–2140; n=39) 10.5 (6–16; n=70)

1096 (415–1754; n=48) 10.5 (6–14; n=82)

867 (193–2862; n=11) 11 (8–13.25; n=16)

0.24

4 (2–7; n=74)

3 (2–6; n=85)

3.5 (1.25–6.75; n=18)

0.34

5% (5/93) 21% (17/80) 25% (22/89) 3% (3/93)

7% (7/101) 23% (20/87) 31% (31/100) 1% (1/101)

0% 0/26 18% (4/22) 8% (2/24) 0% (0/27)

0.38

Homozygote for rs13228878_A risk allele n=104 Age (years), median (median; IQR) Sex, % males n. males/total)) Total amylase at diagnosis of AAP (U/L) median (IQR; n. available data) Lipase at diagnosis of AAP (U/L) median (IQR; n. available data) Days from last PEG- asparaginase exposure to AAP, median (IQR; n. available data) Number of PEG-asparaginase administrations prior to AAP, median (IQR; n. available data) Assisted mechanical ventilation, % (yes (n.)/available data (n.)) Acute insulin therapy, % (yes (n.)/available data (n.)) Pseudocysts, % (yes (n.)/available data (n.)) Death, % (yes (n.)/available data (n.))

0.92 0.33

0.96

0.88 0.07 0.38

Baseline characteristics, pancreatic enzyme levels and complications of pancreatitis according to rs13228878 genotype. Differences among groups were analyzed with the Kruskal-Wallis rank sum test (continuous variables) and chi-square test (categorical variables).

haematologica | 2019; 104(3)

561


B.O. Wolthers et al.

the (rs4726576; rs10273639) risk allele, these patients are at higher risk of trypsin activation and pancreatitis.26 In a study investigating the effect of asparaginase on pancreatic acinar cells, asparaginase evoked intracellular calcium release from the endoplasmic reticulum mediated by the protease-activated receptor 2 (Figure 3). This elevation in calcium levels in turn activated calcium release activated calcium (CRAC) channels further increasing intracellular calcium levels, leading to decreased ATP levels, trypsin activation and necrosis.35 The pancreatitis-causing mechanism seems to be independent of the anti-neoplastic effect of asparaginase, and drugs inhibiting CRAC channels could thus be used to prevent AAP during asparaginase therapy, not least in patients who are re-exposed to asparaginase after having had AAP, since these patients have a ~50% risk or developing a second episode of AAP.9,13 The NFATC2 associated variant rs62228256 was most significantly associated with AAP. However, with no association found in the replication study and no association with pancreatitis found in adult studies on non-asparaginase associated pancreatitis,26 the association between NFATC2 and AAP seems to have low credibility. We were not able to associate the PRSS1-PRSS2 genotype to risk of AAP-related complications, indicating that this allele only alters the risk of AAP, while the complications are a result of other factors. The validation of our PRSS1-PRSS2 top SNPs in the AALL0232 cohort strengthens the credibility of this result. The association was of similar effect size, but of borderline statistical significance, which may reflect three key issues: (i) diagnostic criteria differ between the cohorts, and cases are not completely comparable; (ii) pancreatitis is strongly associated with asparaginase exposure, and it was a prerequisite that included controls received a significant amount of asparaginase to reduce the risk of

References 1. Schmiegelow K, Müller K, Mogensen SS, et al. Non-infectious chemotherapy-associated acute toxicities during childhood acute lymphoblastic leukemia therapy. F1000Res. 2017;6444. 2. Knott SRV, Wagenblast E, Khan S, et al. Asparagine bioavailability governs metastasis in a model of breast cancer. Nature. 2018;554(7692):378-381. 3. Müller HJ, Boos J. Use of L-asparaginase in childhood ALL. Crit Rev Oncol Hematol. 1998;28(2):97–113. 4. Pession A, Valsecchi MG, Masera G, et al. Long-term results of a randomized trial on extended use of high dose L-asparaginase for standard risk childhood acute lymphoblastic leukemia. J Clin Oncol. 2005;23(28):7161–7167. 5. Silverman LB. Improved outcome for children with acute lymphoblastic leukemia: results of Dana-Farber Consortium Protocol 91-01. Blood. 2001;97(5):1211–1218. 6. Haskell CM, Canellos GP, Leventhal BG, et al. L-Asparaginase. N Engl J Med. 1969;281 (19):1028–1034. 7. Raja RA, Schmiegelow K, Frandsen TL. Asparaginase-associated pancreatitis in children. Br J Haematol. 2012;159(August):18– 27.

562

false negative controls; and (iii) the validation cohort was relatively small.36,37 Our results need to be judged in the light of their limitations. The association analyses were strongly influenced by individuals of CEU ethnicity, and we cannot determine the effects in non-CEU populations. Moreover, our results for the PRSS1-PRSS2 locus did not reach genome-wide significance with a P-value <5 x 10-8. This highlights a challenge when doing GWA studies in cohorts of limited size, as will often be the case in childhood ALL, and requires strategies for validation in independent and similar cohorts. In this present study, we attempted to limit the problem of limited sample size by improving the quality of phenotyping, collecting individual clinical data on AAP cases and including controls with documented completion of extensive asparaginase therapy. In conclusion, we found that children who develop AAP possess identical genetic risk variants as adults with nonasparaginase-associated pancreatitis. This may allow future preventive measures for reduction of AAP. Acknowledgments We thank all the researchers who scrutinized patients’ files and completed phenotype questionnaires, colleagues at Harvard Department of Biomedical Informatics for valuable insights, and organizational support from the research staff at Bonkolab, at the University Hospital Rigshospitalet. Furthermore, we thank the Bloodwise Childhood Leukaemia Cell Bank, UK, for providing samples and data for this research. Funding This study was funded by the Kirsten and Freddy Johansen Foundation, the Danish Childhood Cancer Foundation, the Swedish Childhood Cancer Foundation and the Danish Cancer Society (R150-A10181).

8. Liu C, Yang W, Devidas M, et al. Clinical and genetic risk factors for acute pancreatitis in patients with acute lymphoblastic leukemia. J Clin Oncol. 2016;34(18):2133-2140. 9. Wolthers BO, Frandsen TL, Baruchel A, et al. Asparaginase-associated pancreatitis in childhood acute lymphoblastic leukaemia: an observational Ponte di Legno Toxicity Working Group study. Lancet Oncol. 2017;18(9):1238–1248. 10. Pemmaraju N, Rytting ME. Questions on asparaginase-associated pancreatitis. Lancet Oncol. 2017;18(9):1148–1149. 11. Schmiegelow K, Attarbaschi A, Barzilai S, et al. Consensus definitions of 14 severe acute toxic effects for childhood lymphoblastic leukaemia treatment: a Delphi consensus. Lancet Oncol. 2016;17(6):e231–e239. 12. Computerome. http://www.computerome.dtu.dk/. 13. Wolthers BO, Frandsen TL, Abrahamsson J, et al. Asparaginase-associated pancreatitis. A study on pheno-and genotype in the NOPHO ALL2008 protocol. Leukemia. 2017;31(2):325-332. 14. Purcell S, Neale B, Todd-Brown K, et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet. 2007;81(3):559–575. 15. Yates A, Akanni W, Amode MR, et al. Ensembl 2016. Nucleic Acids Res.

2016;44(D1):D710–D716. 16. Sherry ST, Ward MH, Kholodov M, et al. dbSNP: the NCBI database of genetic variation. Nucleic Acids Res. 2001;29(1):308–311. 17. Anderson CA, Pettersson FH, Clarke GM, Cardon LR, Morris P, Zondervan KT. Data quality control in genetic case-control association studies. Nat Protoc. 2011;5(9):1564– 1573. 18. Bourgeois S, Labuda D. Dynamic allele-specific oligonucleotide hybridization on solid support. Anal Biochem. 2004;324(2):309– 311. 19. Labuda D, Krajinovic M, Richer C, et al. Rapid detection of CYP1A1, CYP2D6, and NAT variants by multiplex polymerase chain reaction and allele-specific oligonucleotide assay. anal biochem. 1999;275(1): 84–92. 20. Team RC. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing. 21. Machiela MJ, Chanock SJ. LDassoc: an online tool for interactively exploring genome-wide association study results and prioritizing variants for functional investigation. Bioinformatics. 2018;34(5):887-889. 22. Lonsdale J, Thomas J, Salvatore M, et al. The genotype-tissue expression (GTEx) project. Nat Genet. 2013;45(6):580–585. 23. Boyle AP, Hong EL, Hariharan M, et al.

haematologica | 2019; 104(3)


Pancreatitis in childhood ALL

24.

25.

26.

27.

28.

Annotation of functional variation in personal genomes using RegulomeDB. Genome Res. 2012;22(9):1790–1797. Pruim RJ, Welch RP, Sanna S, et al. LocusZoom: regional visualization of genome-wide association scan results. Bioinformatics. 2010;26(18):2336–2337. Rebhan M, Chalifa-Caspi V, Prilusky J, Lancet D. GeneCards: integrating information about genes, proteins and diseases. Trends Genet. 1997;13(4):163. Zator Z, Whitcomb DC. Insights into the genetic risk factors for the development of pancreatic disease. Therap Adv Gastroenterol. 2017;10(3):323-336. Vrooman LM, Supko JG, Neuberg DS, et al. Erwinia asparaginase after allergy to E. coli asparaginase in children with acute lymphoblastic leukemia. Pediatr Blood Cancer. 2010;54(2):199-205. Silverman LB, Stevenson KE, Brien JEO, et al. Long-term results of Dana-Farber Cancer Institute ALL Consortium protocols for children with newly diagnosed acute lym-

haematologica | 2019; 104(3)

29.

30.

31.

32.

33.

phoblastic leukemia (1985–2000). Leukemia. 2010;24(2):617–632. Whitcomb DC, Gorry MC, Preston RA, et al. Hereditary pancreatitis is caused by a mutation in the cationic trypsinogen gene. Nat Genet. 1996;14(2):141–145. Whitcomb DC, LaRusch J, Krasinskas AM, et al. Common genetic variants in the CLDN2 and PRSS1-PRSS2 loci alter risk for alcohol-related and sporadic pancreatitis. Nat Genet. 2012;44(12):1349–1354. Rosendahl J, Kirsten H, Hegyi E, et al. Genome-wide association study identifies inversion in the CTRB1-CTRB2 locus to modify risk for alcoholic and non-alcoholic chronic pancreatitis. Gut. 2018;67(10):18551863. Masamune A, Nakano E, Hamada S, Kakuta Y, Kume K, Shimosegawa T. Common variants at PRSS1–PRSS2 and CLDN2–MORC4 loci associate with chronic pancreatitis in Japan. Gut. 2015;64(8):1345–1346. Paliwal S, Bhaskar S, Reddy DN, et al. Association analysis of PRSS1-PRSS2 and

34.

35.

36.

37.

CLDN2-MORC4 variants in nonalcoholic chronic pancreatitis using tropical calcific pancreatitis as model. Pancreas. 2016;45(8): 1153–1157. Boulling A, Sato M, Masson E, Génin E, Chen J-M, Férec C. Identification of a functional PRSS1 promoter variant in linkage disequilibrium with the chronic pancreatitisprotecting rs10273639. Gut. 2015;64(11): 1837–1838. Peng S, Gerasimenko JV, Tsugorka T, et al. Calcium and adenosine triphosphate control of cellular pathology: asparaginase-induced pancreatitis elicited via protease-activated receptor 2. Philos Trans R Soc Lond B Biol Sci. 2016;371(1700). Sud A, Kinnersley B, Houlston RS. Genomewide association studies of cancer: current insights and future perspectives. Nat Rev Cancer. 2017;17(11):692–704. Ioannidis JPA, Thomas G, Daly MJ. Validating, augmenting and refining genome-wide association signals. Nat Rev Genet. 2009;10(5):318–329.

563


ARTICLE Ferrata Storti Foundation

Hodgkin Lymphoma

CCR5 antagonism by maraviroc inhibits Hodgkin lymphoma microenvironment interactions and xenograft growth

Naike Casagrande,1 Cinzia Borghese,1 Lydia Visser,2 Maurizio Mongiat,1 Alfonso Colombatti1 and Donatella Aldinucci1

Unit of Molecular Oncology, Centro di Riferimento Oncologico di Aviano (CRO) IRCCS, Italy and 2Department of Pathology and Medical Biology, University Medical Center Groningen (UMcG), the Netherlands

1

Haematologica 2019 Volume 104(3):564-575

NC and CB contributed equally to this study.

ABSTRACT

C

Correspondence: DONATELLA ALDINUCCI daldinucci@cro.it Received: May 3, 2018. Accepted: October 9, 2018. Pre-published: October 11, 2018. doi:10.3324/haematol.2018.196725 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/104/3/564 ©2019 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

564

lassic Hodgkin lymphoma tumor cells express a functional CCR5 receptor, and tumor tissues express high CCL5 levels, suggesting that CCL5-CCR5 signaling is involved in tumor-microenvironment formation and tumor growth. Using the CCR5 antagonist, maraviroc, and a neutralizing anti-CCL5 antibody, we found that CCL5 secreted by classic Hodgkin lymphoma cells recruited mesenchymal stromal cells and monocytes. The “education” of mesenchymal stromal cells by tumor cell-conditioned medium enhanced mesenchymal stromal cells’ proliferation and CCL5 secretion. In turn, educated mesenchymal stromal cell-conditioned medium increased the clonogenic growth of tumor cells and monocyte migration, but these effects were reduced by maraviroc. Monocyte education by tumor cell-conditioned medium induced their growth and reprogrammed them towards immunosuppressive tumor-associated macrophages that expressed IDO and PD-L1 and secreted IL-10, CCL17 and TGF-b. Educated monocyte-conditioned medium slowed the growth of phytohemagglutinin-activated lymphocytes. Maraviroc decreased tumor cell growth and synergized with doxorubicin and brentuximab vedotin. A three-dimensional heterospheroid assay showed that maraviroc counteracted both the formation and viability of heterospheroids generated by co-cultivation of tumor cells with mesenchymal stromal cells and monocytes. In mice bearing tumor cell xenografts, maraviroc reduced tumor growth by more than 50% and inhibited monocyte accumulation, without weight loss. Finally, in classic Hodgkin lymphoma human tumor tissues, CCL5 and CD68 expression correlated positively, and patients with high CCL5 levels had poor prognosis. In conclusion, since the present challenges are to find molecules counteracting the formation of the immunosuppressive tumor microenvironment or new, less toxic drug combinations, the repurposed drug maraviroc may represent a new opportunity for classic Hodgkin lymphoma treatment.

Introduction Inflammatory chemokines are indispensable “gate-keepers” of inflammation and immunity against cancer, but tumor cells can subvert chemokines into acting as tumor-promoting molecules.1 C-C motif chemokine ligand 5 (CCL5) is one such chemokine that can favor tumor development in multiple ways; for example, by acting as a growth factor for tumor cells, stimulating angiogenesis, recruiting stromal and inflammatory cells, and taking part in immune evasion mechanisms.2-6 CCL5 belongs to the C-C chemokine family whose members include CCL3 and CCL4.1,2 Its activity is mediated through binding to CCR1, CCR3, and CCR5, while CD44 serves as an auxiliary receptor.2 CCL5 and other chemokines are expressed at higher levels in classic Hodgkin lymphoma (cHL) tumor tissues than in healthy lymph nodes and in tissues with haematologica | 2019; 104(3)


Maraviroc inhibits cHL crosstalk and xenograft growth

reactive lymphoid hyperplasia.7,8 Both CCL5 and its receptor CCR5 are constitutively expressed by cHL-derived cell lines7 by tumor cells from cHL lymph node tissues and by bystander cells including stromal cells and lymphocytes.7 The CCR5 receptor expressed by cHL cells is fully functional and its ligands function as both paracrine and autocrine growth factors.7 The interactions of cHL tumor cells with a variety of non-tumor reactive cells accumulating in cHL tissues mediate tumor cell growth, formation of an immunosuppressive, protective tumor microenvironment (TME), neoangiogenesis,9 and drug resistance.10,11 Increasing evidence suggests that not only T cells,12 but also mesenchymal stromal cells (MSCs)13 and monocytes,14,15 contribute to the TME in cHL.11,16 MSCs, by modulating NKG2D expression in T cells and its ligand in tumor cells, reduce the immune response against cHL cells.13 A high number of infiltrating macrophages,17,18 predominantly derived from circulating monocytes,19 and a high absolute monocyte count in peripheral blood both correlate with poor cHL prognosis.20,21 These observations likely reflect the ability of cHL cells to reprogram macrophages towards immunosuppressive tumor-associated macrophages (TAMs).20,21 Given current knowledge about cell-cell interactions in cHL, there is interest in drugs that can interfere with this crosstalk.22-25 But since drug discovery is expensive and time-consuming, drug repurposing is an attractive approach for finding new cancer treatments.26 One such repurposed drug is the CCR5 antagonist maraviroc.27 Approved by the US Food and Drug Administration for the treatment of HIV infection, maraviroc causes few side effects in humans, even during long-term therapy.28,29 As an anticancer drug, maraviroc has different effects: it blocks metastasis of basal breast cancer cells;30 it decreases the migration of regulatory T cells; it reduces metastatic breast cancer growth in the lungs;31 and it inhibits the accumulation of fibroblasts in human colorectal cancer.32 Maraviroc reprograms immunosuppressive myeloid cells and reinvigorates antitumor immunity by targeting the autocrine CCL5-CCR5 axis in bone marrow.6 It also polarizes macrophages towards an M1-like functional state.27 Our working hypothesis is that cHL cancer cells, by secreting CCL5, recruit both MSCs and monocytes to the TME, and then reprogram these cell types to make them pro-tumorigenic. Thus, blocking the CCR5 receptor should inhibit not only tumor growth, as we previously observed,7 but also the recruitment of cells to form the protective, immunosuppressive TME. Here, we investigated the role of CCL5-CCR5 signaling in the interactions of monocytes and MSCs with cHL cells, using, in particular, three-dimensional multicellular heterospheroids33 formed by tumor cells, monocytes and MSCs, as well as an in vivo cHL model and tissues from cHL patients.

Methods Maraviroc (Sigma-Aldrich) was dissolved in DMSO at 51.8 mM. Other reagents are detailed in Online Supplementary Methods, together with protocols for cell migration, proliferation, clonogenic growth and senescence assays, immunosuppression, synergy, flow cytometry, ELISAs and other cell-based assays, immunofluorescence, survival of tumor xenografts and statistical analysis. haematologica | 2019; 104(3)

Cell culture and conditioned media Authenticated cHL-derived cell lines L-1236, L-428, KMH2, HDLM-2, and L-540 (DSMZ, Germany) were cultured in RPMI-1640 medium containing 10% fetal calf serum (FCS). To prepare conditioned medium (CM) from cHL cell lines, cells were seeded at 2.0×105/ml in RPMI-1640 plus 10% FCS, and the medium was collected after 72 h. Human bone marrow-derived and adipose tissue-derived mesenchymal stromal cells (MSCs) (BM-MSCs and ATMSC, respectively) were purchased from Lonza (Verviers, Belgium). cHL-MSCs from frozen lymph nodes were generated as described in Online Supplementary Methods. BMMSCs, AT-MSC and cHL-MSCs were maintained in Mesenchymal Stem Cell Growth Medium Bulletkit (Lonza) to avoid differentiation. Monocytes were isolated from peripheral blood mononuclear cells (PBMCs) from healthy donor blood using CD14 Microbeads, Human (Miltenyi Biotec). To generate tumor-educated MSCs (E-MSCs) and monocytes (E-monocytes), MSCs and monocytes were cultured separately for 6 days in complete culture medium (RPMI-1640, 10% FCS) containing 20% CM from cHL cell lines; half the volume of medium was replaced every other day. To prepare CM from these tumor-educated cells, they were washed and cultured in fresh medium for 72 h.

3D culture of heterospheroids Heterospheroids were generated by co-culturing various combinations of cHL cells, HL-MSCs and monocytes (1.0 × 104/mL of each cell type) in RPMI-1640 medium containing 1% FCS, using plates coated with 20 mg/mL polyHEMA (Sigma) to prevent adhesion. After 4 days, heterospheroids were tested for CCL5 secretion into the medium by ELISA. In some experiments, heterospheroids were treated with maraviroc, alone or with doxorubicin, for 6 days. Growth was evaluated using the PrestoBlue Cell Viability Reagent (Invitrogen).

Tumor xenograft experiments Animal experiments were approved by the Italian Ministry of Health (no. 671/2015/PR). We used ten 4week-old female athymic nude/nude mice (Harlan Laboratories) and ten 4-week-old male NOD/SCID gamma chain deficient (NSG) mice (Charles River). L-540 (200×106 cells/animal) and L-428 cells (10×106 cells/animal) were suspended in Matrigel (diluted 1:3 in PBS) and inoculated into the flank of nude mice (L-540) or NSG mice (L-428). When tumors were palpable, animals were divided into two equal groups and treated every day (L-540) or every other day (L-428) with maraviroc (intraperitoneal injection, 10 mg/kg)34 or vehicle (PBS). Body weight and tumor volume were measured daily.

Immunohistochemistry tissue microarray analysis of cHL patients The study protocol was approved by the institutional review board of the University Medical Center Groningen. We recruited 65 patients with cHL (Online Supplementary Table S1). All study subjects provided written informed consent. Immunohistochemistry was performed for CCL5 (C-19 antibody, 1:200 dilution, Santa Cruz Biotechnology) (antigen retrieval in 10 mM Tris-HCl pH 9, 1 mM EDTA). CD68 was detected with KP1 antibody (1:4000 dilution, Dako) (antigen retrieval in 10 mM citrate buffer, pH 6). 565


N. Casagrande et al.

A

B

C

D

E

F

G

H

Figure 1. Maraviroc inhibits crosstalk between cHL cells and both MSCs and monocytes. (A) Percentages of BM-MSCs and AT-MSCs that migrated (in 5 h) through a fibronectin-coated Boyden chamber towards conditioned medium (CM) from L-1236 or KM-H2 cells, in the presence of increasing concentrations of maraviroc (MSCs were treated for 1 h prior to migration). (B) Effect of a neutralizing anti-CCL5 antibody (5 mg/mL) in cHL-conditioned medium on MSC migration. Transmigrated cells were revealed using a computer-interfaced GeniusPlus microplate reader (Tecan). Results are the mean and SD of three replicate wells from three independent experiments. (C) BM-MSCs, AT-MSCs and HL-MSCs (100 cells/well; 24-well plates) were cultured in RPMI-1640 medium containing 10% cHL CM. After 9 days, cells were fixed with methanol and stained with crystal violet. (D) BM-MSCs (500 cells/well; 96-well plates) were cultured in RPMI-1640 medium containing 20% cHL CM, with or without a neutralizing anti-FGF2 (1 mg/ml), anti-TGFb1 (2 mg/mL) or anti-TNFα (0.5 mg/mL) antibody. After 9 days, growth was evaluated using the MTT assay. Results are the mean and SD of three replicate wells from three independent experiments. (E) BM-MSCs were cultured for 6 days with 20% CM from L-1236, L-428, KM-H2, HDLM-2, and L-540 cells. Then, the medium was changed with fresh medium, and 3 days later MSC CM was recovered and quantified for CCL5 by ELISA. All samples were tested in triplicate; conditioned media from three different experiments were evaluated. (F) BM-MSCs were cultured for 6 days with 20% CM (from KMH2 or HDLM-2 cells) in the presence or absence of a neutralizing anti-TNFα antibody (0.5 mg/mL). Then, the medium was changed and after 3 days CCL5 was quantified by ELISA. All samples were run in triplicate; conditioned media from three different experiments were evaluated. (G) Clonogenic growth. L-1236 (103/mL), HDLM2 (5 × 102/mL), L-540 (2.5 × 102/mL) cells were cultured in methylcellulose-containing medium in the absence or presence of 5% E-BM-MSC CM and with increasing concentrations of maraviroc. After 14 days of incubation, plates were observed under phase contrast microscopy and aggregates with 40 cells or more were scored as colonies (8 replicate wells). Each experiment was done in triplicate; conditioned media from three different experiments were evaluated. (H) Percentage of CD14+ monocytes that migrated (in 1 h) through fibronectin-coated Boyden chambers towards medium (RPMI-1640 plus 10% FCS) or E-BM-MSC CM. Prior to migration towards E-BM-MSC CM, monocytes were pretreated with maraviroc (0.1-100 mM) for 1 h. Results are means and SD of transmigrated monocytes for three different experiments. AT-MSCs: Adipose Tissue; BM-MSCs: Bone Marrow; cHL: classical Hodgkin lymphoma; CM: conditioned medium; HL-MSCs: Hodgkin lymphoma; MSC: Mesenchymal stromal cells; E-BM-MSCs: tumor Educated.

566

haematologica | 2019; 104(3)


Maraviroc inhibits cHL crosstalk and xenograft growth

Results

pose tissue, and cHL lymph nodes) increased in the presence of cHL-conditioned medium in a dose-dependent manner (Figure 1C and Online Supplementary Figure S1B). This effect was partially mediated by FGF2, TGFb1 and TNFα secreted by cHL cells, since addition of antibodies against these growth factors significantly, but incompletely, reduced growth (Figure 1D). cHL-conditioned medium almost totally abolished MSC senescence induced by serum starvation (Online Supplementary Figure S1C) and reduced apoptosis induced by doxorubicin treatment (Online Supplementary Figure S1D). To survive and proliferate, cancer cells not only recruit but also shape or “educate” normal cells.11 BM-MSCs released very low amounts of CCL5 under normal culture conditions (Figure 1E). After being cultured with cHL-conditioned medium, thereby

Maraviroc inhibits crosstalk between cHL cells and both MSCs and monocytes Conditioned medium from L-1236 and KM-H2 cHL-derived cell lines stimulated the migration of BM-MSCs (Figure 1A, left) and AT-MSCs (Figure 1A, right). Since MSCs express CCR5 (Online Supplementary Figure S1A)35 and cHL cells secret CCL5,7 we investigated whether this chemokine is directly involved in MSC migration. Addition of the CCR5 antagonist maraviroc (Figure 1A) or a neutralizing anti-CCL5 antibody (Figure 1B) significantly reduced the MSC migration induced by cHL-conditioned medi um. Growth of MSCs from different sources (bone marrow, adi-

A

B

D

C

E

Figure 2. cHL cells induce monocyte migration and proliferation. (A) Percentages of monocytes that migrated (in 1 h) through fibronectin-coated Boyden chambers towards conditioned medium (CM) from L-1236 and L-428 cells, in the presence of increasing concentrations of maraviroc (monocytes were treated for 1 h prior to migration). (B) Effect of a neutralizing anti-CCL5 antibody (5 mg/mL) in cHL-conditioned medium on monocyte migration. Results are means and SD of three replicate wells from three independent experiments. (C) cHL cells (2x105 cells/mL) were cultured for 3 days before CM was collected and tested for M-CSF by ELISA. All samples were run in triplicate; conditioned media from three different experiments were evaluated. (D) Monocytes (2.0x104 cells/well; 96-well flat-bottomed plates) were exposed to increasing concentrations (percentage, v/v) of cHL CM. After 9 days, monocyte growth was evaluated using the MTT assay. Results are mean and SD of three experiments. (E) Clonogenic growth in methylcellulose-containing medium. L-1236 (103/mL), HDLM-2 (5x102/mL), L-540 (2.5x102/mL) cells were cultured in the absence or presence of 5% (v/v) E-monocyte CM, with increasing concentrations of maraviroc. After 14 days of incubation, plates were observed under phase contrast microscopy and aggregates with 40 cells or more were scored as colonies (8 replicate wells). Each experiment was done in triplicate; conditioned media from three different experiments were evaluated. cHL: classical Hodgkin lymphoma; CM: conditioned medium, (E)-monocytes: tumor Educated.

haematologica | 2019; 104(3)

567


N. Casagrande et al.

Table 1. Combination index (CI) values for L-1236 and HDLM-2 cell lines treated with maraviroc (first column) and with doxorubicin, brentuximab vedotin, cisplatin, gemcitabine or vinorelbine for 72 h.

L-1236

HDLM-2

MVC (mM)

Doxo (ng/mL)

CI

BV (mg/mL)

CI

CDDP (µM)

CI

GCB (nM)

CI

VRB (ng/mL)

CI

25 50 100 25 50 100

6.0 12.5 *25.0 3.1 6.2 *12.5

0.397 0.594 0.628 0.512 0.526 0.753

3.8 7.5 *15.0 75 150 *300

0.253 0.357 0.495 0.373 0.478 0.655

0.13 0.25 *0.50 0.31 0.63 *1.25

0.606 0.955 1.152 0.571 0.706 1.169

0.15 0.30 *0.60 0.80 1.60 *3.20

0.662 0.912 1.302 0.699 1.054 1.173

0.13 0.25 *0.50 0.25 0.50 *1.00

0.849 0.946 1.051 0.818 0.887 1.045

Cell viability was determined by trypan blue dye exclusion. Combination indexes were calculated using CalcuSyn software. *Indicates the IC50 for each drug. (IC50) Concentration required for 50% in vitro inhibition of growth. MVC: maraviroc; Doxo: doxorubicin; BV: brentuximab vedotin; CDDP: cisplatin; GCB: gemcitabine; VRB: vinorelbine

becoming tumor educated (E-BM-MSCs), they strongly secreted CCL5 (Figure 1E). This response was partially reduced by treatment with an anti-TNFα antibody (Figure 1F). Education of BMMSCs did not, however, induce the secretion of CCL3 or CCL4 (data not shown). Conditioned medium from E-BM-MSCs increased the clonogenic growth of L-1236, HDLM-2 and L-540 cells, and this effect was reduced by maraviroc in a dose-dependent manner (Figure 1G and Online Supplementary Figure S2A). This effect was due to CCL5-CCR5 interactions, because it was partially inhibited by inclusion of a neutralizing anti-CCL5 antibody (Online Supplementary Figure S2B). Conditioned medium from BM-MSCs educated with L-1236-conditioned medium (and thereby containing CCL5; Figure 1E) increased the migration of CD14+ monocytes, and this effect was reduced in a dose-dependent manner by maraviroc (Figure 1H). Taken together, these results suggest that E-BM-MSCs, by secreting CCL5, stimulate tumor growth as well as monocyte recruitment in the TME.

cHL cells induce monocyte migration and proliferation cHL-conditioned medium increased CCR5 expression in monocytes (Online Supplementary Figure S3A) and enhanced their migration through fibronectin-coated Boyden chambers (Online Supplementary Figure S3B). Representative photomicrographs of transmigrated monocytes (red Fast-Dil colored cells) are shown in Online Supplementary Figure S3C. This enhanced monocyte migration was significantly reduced when maraviroc (Figure 2A) or a neutralizing anti-CCL5 antibody (Figure 2B) was added. cHL cells, especially L-1236 and L-428 cells, secreted macrophage colonystimulating factor (M-CSF) (Figure 2C), a cytokine involved in monocyte proliferation and differentiation. In accordance, conditioned medium from cHL cells increased monocyte growth (Figure 2D and Online Supplementary Figure S3D). E-monocytes secreted CCL3 and low amounts of CCL4 and CCL5 (Online Supplementary Figure S3E). Treatment of L-1236, HDLM-2 and L-540 cells with conditioned medium from E-monocytes increased the number of viable tumor cells (Online Supplementary Figure S3F) and stimulated clonogenic growth (Figure 2E), but this growth was inhibited by maraviroc treatment (Figure 2E and Online Supplementary Figure S3G).

Reprogramming of monocytes by cHL cells TAMs express high levels of CD206, PD-L1 and IDO; they secrete IL-10, CCL17/TARC and TGF-b, and can inhibit growth of activated T cells.16 When monocytes were cultivated in conditioned medium from L-1236 or L-428 cells, they upregulated the secretion of IL-10, CCL17, and TGFb (Figure 3A) and increased the 568

expression of CD206, PD-L1, and IDO (Figure 3B). Conditioned medium from L-540 cells did not induce IL-10 or CCL17 secretion or IDO expression by monocytes, but did enhance CD206, PD-L1, and especially TGFb secretion (Figures 3A-B). Conditioned medium from E-monocytes slowed, in a dose-dependent manner, the growth of phytohemagglutinin-activated lymphocytes (Figure 3C). These results demonstrate that cHL cells recruit, induce proliferation, and then reprogram monocytes towards an immunosuppressive phenotype.

Maraviroc slows tumor cell growth and synergizes with doxorubicin and brentuximab vedotin Considering that cHL cells express a functional CCR5 receptor, we evaluated cHL cell growth in the presence of increasing concentrations of maraviroc. This treatment slightly slowed the growth of cHL cells (Figure 4A) and slightly increased the percentage of cells in the G1 phase of the cell cycle (Figure 4B and Online Supplementary Figure S4A). On the contrary, maraviroc treatment did not induce apoptosis-necrosis as shown by the lack of change in annexin-V and 7-AAD staining (Online Supplementary Figure S4B) and in levels of activated caspase-3/7 (Online Supplementary Figure S4C). Maraviroc (25, 50, 100 mM) synergized with doxorubicin and brentuximab vedotin as indicated by the combination index as being <0.9 in all three conditions tested (Table 1 and Online Supplementary Figure S5A-B). Moreover, maraviroc exerted additive or antagonist effects (combination index ≥0.9) in combination with cisplatin, gemcitabine and vinorelbine (Table 1). Synergistic effects with both doxorubicin and brentuximab vedotin were also obtained using ten-fold lower maraviroc concentrations (2.5, 5, 10 mM), but not with hundred-fold lower concentrations (0.25, 0.5, 1.0 mM) (Online Supplementary Table S2).

Maraviroc inhibits 3D heterospheroid formation To mimic TME interactions, we co-cultured cHL cells with MSCs and monocytes in a non-adherent, 3D setting in polyHEMA-coated wells. The cells were dispersed in the medium at time 0 but at 24 h had started to self-assemble into 3D heterospheroids (Figure 4C and Online Supplementary Figure S6A). Heterospheroids containing all three cell types secreted high levels of CCL5, and those containing only cHL cells (either L-1236 or HDLM-2) and HL-MSCs also had high CCL5 secretion, whereas the combinations of just monocytes with cHL cells or with HLMSCs expressed low levels (Figure 4D). However, when combined with HL-MSCs and cHL cells, monocytes were able to induce a further increase of CCL5 secretion (Figure 4D). Next, we evaluated the effects of maraviroc on the self-assembling ability and viability of heterospheroids. Maraviroc inhibited the selfhaematologica | 2019; 104(3)


Maraviroc inhibits cHL crosstalk and xenograft growth

A

B

C

Figure 3. Monocyte conversion towards immunosuppressive E-monocytes by cHL cells. (A) Purified CD14+ monocytes were cultured for 6 days in the absence or presence of 20% conditioned medium (CM) from L-1236, L-428, and L-540 cells (to convert monocytes into E-monocytes), then washed and cultured for another 72 h in fresh medium. This E-monocyte CM was recovered for IL-10, CCL17, and TGFb ELISAs. Results are means and SD of three independent experiments. (B) Representative flow cytometric histograms showing surface expression of CD206 and PD-L1 and intracellular expression of IDO in monocytes (control) and E-monocytes (treated with cHL CM). (C) E-monocyte CM was tested for immunosuppressive activity. Phytohemagglutinin (PHA)-activated lymphocytes were treated with increasing concentrations (percentage, v/v) of monocyte CM and E-monocyte CM. After 72 h, growth was assayed using the Cell Proliferation ELISA, BrdU. All samples were tested in triplicate; conditioned media from three different experiments were evaluated. cHL: classical Hodgkin lymphoma; CM: conditioned medium, (E)-monocytes: tumor Educated.

assembling of cHL, HL-MSCs and monocytes into heterospheroids (Figure 4E). It also reduced the total number of viable cells in the heterospheroids (Figure 4F). Considering this reduced viability, and the finding that maraviroc synergized with doxorubicin (Table 1), we applied drug combinations to the heterospheroids and found that doxorubicin and maraviroc exerted synergistic activity (combination index <1) against heterospheroids too (Online Supplementary Figure S6B). When cHL cells (CD30+) were recovered from heterospheroids by trypsinization and purification with anti-CD30 beads, fewer viable tumor cells were recovered from maraviroc-treated than from untreated heterospheroids (Online Supplementary Figure S6C). The cells from treated heterospheroids produced fewer colonies (Online Supplementary Figure S6D) and were proportionately more in G1 than in G2M phase than untreated cells (Online Supplementary Figure S6E).

Maraviroc slows the growth of cHL tumor xenografts and reduces infiltrated TAMs To analyze the anticancer activity of maraviroc in vivo, we studied L-540 tumor cell xenografts in female athymic nude mice, treated every day with an intraperitoneal injection of 10 mg/kg maraviroc or vehicle. By day 12, untreated tumors had grown to a mean volume of 880 mm3 (SD = 88 mm3), whereas maraviroctreated tumors were more than 50% smaller (435±75 mm3; P<0.0001, Student’s t-test; Figure 5A and Online Supplementary haematologica | 2019; 104(3)

Figure S7A). Maraviroc treatment was not toxic as the animals were normal on physical inspection and had similar weight to untreated animals (Figure 5B). Maraviroc-treated mice had significantly better “survival” (i.e., tumor volume <800 mm3) than untreated mice (P=0.002, log rank test) (Online Supplementary Figure S7B). Since maraviroc inhibited the migration of monocytes in vitro, we evaluated whether similar activity was also detectable in vivo by examining infiltrating TAMs (CD68+) in L-540 tumor xenografts. Immunofluorescence analysis of CD30 on tissue sections showed no difference between untreated and maraviroctreated mice (Online Supplementary Figure S7C). However, the animals differed substantially in staining for CD68, a marker of TAM infiltration, which was almost completely absent in maraviroctreated mice (Figure 5C-D and Online Supplementary Figure S7D). Similar results were obtained when male NSG mice were injected with L-428 tumor cells (Figure 5E-H). In particular, maraviroc treatment reduced xenograft growth by about 60% (Figure 5E), without weight loss (Figure 5F), and it reduced CD68 staining by 75% (Figure 5G-H and Online Supplementary Figure S7D).

High CCL5 expression positively correlates with CD68 and poor survival To confirm our in vitro and in vivo results, we studied cHL tissues from 65 patients (Online Supplementary Table S1). cHL tissues had a median CCL5 expression level of 12% positive pixels (range, 0%569


N. Casagrande et al.

(21%-100%). These groups had 24, 22, and 16 patients, respectively (data missing for 3 patients). A Kaplan-Meier plot showed that the group of patients with high CCL5 expression had significantly worse survival (P=0.0072) than patients with low or medium expression (Figure 6C). The hazard ratio for progression-free survival of low/medium CCL5 expression to that of high expression was 0.015 (P=0.0016; 95% CI, 0.0011-0.2064).

44%) and a median CD68 expression level of 17% positive pixels (range, 0%-58%) (Figure 6A). No difference in CCL5 expression was found in EBV-positive versus EBV-negative HL samples (data not shown). On the contrary, a significant correlation between CCL5 and CD68 levels was found (Spearman r=0.251, P=0.0487) (Figure 6B). Patients were then divided into three arbitrary categories of CCL5 expression: low (0%-10%), medium (11%-20%) and high

A

B

C

D

E

F

Figure 4. Maraviroc slows tumor cell growth and inhibits heterospheroid formation. (A) cHL cells (2x105 cells/ml) were cultured with increasing concentrations of maraviroc for 72 h, and viable cells were counted by trypan blue dye exclusion. Results are means and SD of three replicate wells from three independent experiments. (B) Percentages of cHL cells in the various cell cycle phases after a 24 h treatment with maraviroc (100 mM). Results are means and SD of at least three experiments. (C) Representative image of 3D heterospheroids generated by plating HDLM-2 cells (stained green with CFSE), MSCs (red with fluorescent DiI), and monocytes (blue with DiD) under non-adherent conditions (poly-HEMA-coated wells). (D) L-1236 or HDLM-2 cells, HL-MSCs and monocytes were cultured in RPMI 1640 medium plus 1% FCS, alone or in combination under non-adherent conditions. After 4 days, conditioned medium was collected for CCL5 ELISAs; three different experiments were evaluated. (E) Heterospheroids generated by co-cultivation of cHL cells (L-1236, HDLM-2, or L-540 cells) with HL-MSCs and monocytes under nonadherent conditions in the presence or absence of maraviroc (100 mM) and photographed after 24 h using an inverted microscope (Eclipse TS/100, Nikon). (F) Heterospheroids (cHL + HL-MSCs + monocytes) were cultured with and without maraviroc. After 48 h, viable cells were evaluated using PrestoBlue Cell Viability Reagent. Relative fluorescence units (RFU). Values are means and SD of three experiments. cHL: classical Hodgkin lymphoma; (HL)-MSCs: Hodgkin lymphoma; MSC: mesenchymal stromal cells; 3D: three dimensional.

570

haematologica | 2019; 104(3)


Maraviroc inhibits cHL crosstalk and xenograft growth

Discussion The TME plays an active role in cHL,11 suggesting the possibility of developing alternative treatment strategies that target not only tumor cells, but also the TME’s protective effects.11,22 Here, we found that maraviroc, a CCR5 antagonist, inhibited cHL cell recruitment of monocytes and MSCs, reduced the cHL cell growth-promoting effects of CCR5 ligands secreted by monocytes and MSCs, synergized with doxorubicin and brentuximab vedotin, and decreased cHL tumor xenograft growth and monocyte infiltration in vivo. In cHL patients, high CCL5 levels corre-

A

B

C

D

E

G

haematologica | 2019; 104(3)

F

H

lated with monocyte infiltration and poor prognosis. Our in vitro results suggest that there is a “domino effect” within the cHL TME: tumor cells, by secreting CCL5, among other molecules, recruit, expand, and educate MSCs and monocytes; these cells, in turn, secrete CCR5 ligands (i.e., CCL3 and CCL4) to recruit other normal cells and stimulate the growth of tumor cells, which reprogram (educate) monocytes to become immunosuppressive TAMs. A schematic view of the possible mechanisms leading to cHL cell proliferation and TME formation by CCR5 ligands, and the counteracting effects of maraviroc, is shown in Figure 7.

Figure 5. Maraviroc reduces cHL xenograft growth and TAM infiltration. (A-D) Xenografts in nude mice inoculated with L540 cells (20x106 cells/animal) and treated every day with an intraperitoneal injection of 10 mg/kg maraviroc (n = 5) or vehicle (n=5). (A) Xenograft tumor growth. (B) Body weights of xenografted mice. (C) Quantification of CD68+ staining in immunofluorescent cryosections using Volocity software provided by PerkinElmer (arbitrary units). Data are means and SD. (D) Immunofluorescent photomicrographs of CD68+ staining in tumor cryosections from maraviroc-treated and untreated xenografted mice. Nuclei were stained with TO-PRO-3 dye. Representative images were acquired using a confocal microscope (Leica DM IRE2). (E-H) Xenografts in NSG mice inoculated with L-428 cells (10x106 cells/animal) and treated every other day with an intraperitoneal injection of 10 mg/kg maraviroc (n = 5) or vehicle (n=5). (E) Xenograft tumor growth. (F) Body weights of xenografted mice. (G) Quantification of CD68+ staining in immunofluorescent cryosections using Volocity software (arbitrary units). Data are means and SD. (H) Immunofluorescent photomicrographs of CD68+ staining in tumor cryosections from maraviroc-treated and untreated xenografted mice. Nuclei were stained with TO-PRO-3 dye. Representative images were acquired using a confocal microscope (Leica DM IRE2). NSG: NOD/SCID gamma chain deficient; TAM: tumor associated macrophages.

571


N. Casagrande et al.

We found that conditioned medium from cHL cells increased the growth of MSCs from different sources, and maraviroc decreased MSC recruitment by cHL cells. Thus, the development of fibrosis in cHL tissues could be explained by recruitment of MSCs by cytokines, including CCL5, secreted by primary tumors and expansion or activation by FGF2, TGFb or TNFÎą secreted by tumor cells. Moreover, our findings that cHL education of MSCs (E-MSCs) increased the secretion of CCL5 and that maraviroc reduced monocyte recruitment by E-MSC-conditioned medium suggest an active role of E-MSCs in TME formation. In this perspective, MSCs of the cHL TME not only may down-regulate anti-tumor immune responses

through NKG2D-NKG2DL interactions,36,37 but may also enhance the number of infiltrated TAMs by secreting CCL5 and, consequently, supporting tumor progression. Macrophages seem to be involved in the pathogenesis of cHL, since high levels of TAMs as well as the absolute monocyte count in the blood correlate with an unfavorable clinical outcome.17,20,21,38 Despite the importance of monocyte levels in TME, their education (i.e., conditioning by tumor cells) seems an essential prerequisite for their pro-tumor activity.39,40 Recently, it was demonstrated that conditioned medium from cHL cell lines induced an immunosuppressive phenotype in macrophages obtained by pre-cultivation of monocytes with M-CSF or GM-

A

B

C

P=0.0072

HR (low-med/hi)=0.015

Figure 6. CCL5 expression positively correlates with CD68 expression in human cHL tissues and with lower progression-free survival in cHL patients. (A) Representative photomicrographs of cHL tissues stained for CCL5 and CD68 (marker of macrophages/monocytes). Top row, a case with low expression; bottom row, a case with high expression. (B) Correlation of expression levels of CCL5 and CD68 in 65 cHL patients. Spearman r=0.251; P=0.0487. (C) Kaplan-Meier survival plot for 5-year progression-free survival in cHL patients, subdivided according to the percentage of CCL5 positivity (low, 0%-10%; medium, 11%-20%; high, 21%-100%). Patients with high CCL5 expression had worse survival (P=0.0072, HR=0.015 vs. low-medium levels). cHL: classical Hodgkin lymphoma; HR: hazard ratio; r: Spearman correlation coefficient.

572

haematologica | 2019; 104(3)


Maraviroc inhibits cHL crosstalk and xenograft growth

CSF.16 Consistent with the finding that cHL cells secrete M-CSF, the education of monocytes by cHL-cell conditioned medium, also without a preconditioning with MCSF or GM-CSF,16,17 was sufficient to shape monocytes to secrete and express immunosuppressive molecules, including IDO and PD-L1, and to inhibit PHA-activated lymphocyte growth. Recently, it was found that, in cHL tissues, TAMs are not randomly distributed; in fact, PD-L1+ TAMs lie in greater proximity to PD-L1+ tumor cells and PD1+ T cells preferentially localize in proximity to PD-L1+ TAMs, suggesting a model in which the inflammatory microenviron-

ment of cHL is highly organized with PD-L1+ TAMs immediately surrounding Hodgkin and Reed-Sternberg cells to engage PD-1+ T cells and augment immune suppression.41 Thus, our results suggest that cHL cells, by inducing PD-L1 expression in monocytes, contribute to the building of the immunosuppressive niche.41 Maraviroc, by reducing monocyte recruitment, may counteract this phenomenon. Since current two-dimensional (2D) in vitro methods often fail to adequately replicate tumor cell interactions with the TME and to properly assess drug activity, here, to evaluate the effects of maraviroc, we developed and

Figure 7. Proposed mechanism for the inhibitory effects of the CCR5 antagonist maraviroc on TME formation and cHL tumor growth. (1) Maraviroc, a CCR5 antagonist, inhibits the recruitment of monocytes and MSCs by cHL cells (Hodgkin and Reed/Sternberg, HRS). (2) The tumor â&#x20AC;&#x153;educationâ&#x20AC;? of MSCs cells (E-MSCs) induces the secretion of CCL5. (3) Maraviroc inhibits the recruitment of monocytes by E-MSCs secreting CCL5. (4) Maraviroc decreases the cHL cell growth-promoting effects of CCR5 ligands secreted by tumor educated-monocytes (E-monocytes) and E-MSCs (CCL5+). (5) Maraviroc inhibits the growth of cHL cells (autocrine growth). (6) cHL: classical Hodgkin lymphoma; HRS: Hodgkin and Reed/Sternberg; MVC: Maraviroc; MSC: Mesenchymal stromal cells; (E)-MSCs: tumor Educated; (E)-monocytes: tumor Educated; TME: tumor microenvironment.

haematologica | 2019; 104(3)

573


N. Casagrande et al.

used a three-dimensional (3D) multicellular heterospheroid model33 formed by tumor cells, monocytes and MSCs. By using 3D heterospheroids, in which tumor cells and different types of normal cells interact and organize their positions, we found that interactions between MSCs, monocytes and cHL cells increased the overall secretion of CCL5. Maraviroc decreased heterospheroid selfassembling, cell viability and cHL clonogenic growth ability, suggesting that it may counteract TME formation and, as a consequence, its protective effects. In a mouse xenograft model, maraviroc decreased the in vivo growth of both L-540 and L-428 cells by more than 50% and reduced monocyte infiltration without apparent toxicity to the animals. These findings confirm that CCR5 signaling contributes to determining the fate of cHL tumor cells. Recently, Jiao et al.42 demonstrated that maraviroc dramatically enhanced cell killing of CCR5+ breast cancer cells by the DNA-damaging chemotherapeutic agent doxorubicin. Here, maraviroc synergized not only with doxorubicin, but also with brentuximab vedotin. These results suggest that CCR5 inhibitors, by enhancing the activity of other drugs, may allow a dose reduction of the two

References 1. Chow MT, Luster AD. Chemokines in cancer. Cancer Immunol.Res. 2014;2(12):11251131. 2. Aldinucci D, Colombatti A. The inflammatory chemokine CCL5 and cancer progression. Mediators Inflamm. 2014; 2014:292376. 3. Zhang Q, Qin J, Zhong L, et al. CCL5mediated Th2 immune polarization promotes metastasis in luminal breast cancer. Cancer Res. 2015;75(20):4312-4321. 4. Wang LH, Lin CY, Liu SC, et al. CCL5 promotes VEGF-C production and induces lymphangiogenesis by suppressing miR507 in human chondrosarcoma cells. Oncotarget. 2016;7(24):36896-36908. 5. Atretkhany KN, Drutskaya MS, Nedospasov SA, Grivennikov SI, Kuprash DV. Chemokines, cytokines and exosomes help tumors to shape inflammatory microenvironment. Pharmacol Ther. 2016;168:98-112. 6. Ban Y, Mai J, Li X, et al. Targeting autocrine CCL5-CCR5 axis reprograms immunosuppressive myeloid cells and reinvigorates antitumor immunity. Cancer Res. 2017; 77(11):2857-2868. 7. Aldinucci D, Lorenzon D, Cattaruzza L, et al. Expression of CCR5 receptors on ReedSternberg cells and Hodgkin lymphoma cell lines: involvement of CCL5/Rantes in tumor cell growth and microenvironmental interactions. Int J Cancer. 2008;122(4):769776. 8. Mathas S, Hartmann S, Kuppers R. Hodgkin lymphoma: pathology and biology. Semin Hematol. 2016;53(3):139-147. 9. Linke F, Harenberg M, Nietert MM, et al. Microenvironmental interactions between endothelial and lymphoma cells: a role for the canonical WNT pathway in Hodgkin lymphoma. Leukemia. 2017;31(2):361-372. 10. Celegato M, Borghese C, Casagrande N, et

574

11.

12.

13.

14.

15.

16.

17.

18.

19.

chemotherapeutic agents already in use in cHL patients. As shown here, CCL5 expression levels positively correlated with CD68 levels in cHL tissues, and patients with high CCL5 levels had lower progression-free survival than patients with low or medium expression of this chemokine; this difference may be due to the higher level of macrophages recruited and then reprogrammed by cHL cells towards immunosuppressive tumor-associated PDL1-positive macrophages as well as by enhanced tumor growth by CCL5. In conclusion, in light of the significant reduction of tumor mass obtained with maraviroc alone, its low toxicity and ability to inhibit monocyte infiltration, our results provide the rationale for its clinical assessment in cHL, as a single agent or in combination therapy. Funding This work was supported by grant IG 15844 from the Italian Association for Cancer Research (to D.A.) and by an Intramural Grant (5X1000CRO-2011). Acknowledgments Valerie Matarese provided scientific editing.

al. Preclinical activity of the repurposed drug Auranofin in classical Hodgkin lymphoma. Blood. 2015;126(11):1394-1397. Aldinucci D, Celegato M, Casagrande N. Microenvironmental interactions in classical Hodgkin lymphoma and their role in promoting tumor growth, immune escape and drug resistance. Cancer Lett. 2016;380(1):243-252. Wein F, Weniger MA, Hoing B, et al. Complex immune evasion strategies in classical Hodgkin lymphoma. Cancer Immunol Res. 2017;5(12):1122-1132. Poggi A, Musso A, Dapino I, Zocchi MR. Mechanisms of tumor escape from immune system: Role of mesenchymal stromal cells. Immunol Lett. 2014;159(12):55-72. Tan KL, Scott DW, Hong F, et al. Tumorassociated macrophages predict inferior outcomes in classic Hodgkin lymphoma: a correlative study from the E2496 Intergroup trial. Blood. 2012;120(16):32803287. Scott DW, Steidl C. The classical Hodgkin lymphoma tumor microenvironment: macrophages and gene expression-based modeling. Hematology Am Soc Hematol Educ Program. 2014;2014(1):144-150. Ruella M, Klichinsky M, Kenderian SS, et al. Overcoming the immunosuppressive tumor microenvironment of Hodgkin lymphoma using chimeric antigen receptor T cells. Cancer Discov. 2017;7(10):1154-1167. Tudor CS, Bruns H, Daniel C, et al. Macrophages and dendritic cells as actors in the immune reaction of classical Hodgkin lymphoma. PLoS One. 2014;9(12):e114345. Koh YW, Park CS, Yoon DH, Suh C, Huh J. CD163 expression was associated with angiogenesis and shortened survival in patients with uniformly treated classical Hodgkin lymphoma. PLoS One. 2014; 9(1):e87066. Crane GM, Samols MA, Morsberger LA, et al. Tumor-infiltrating macrophages in post-

20.

21.

22.

23.

24.

25. 26.

27.

28.

29.

transplant, relapsed classical Hodgkin lymphoma are donor-derived. PLoS One. 2016;11(9):e0163559. Tadmor T, Bari A, Marcheselli L, et al. Absolute monocyte count and lymphocytemonocyte ratio predict outcome in nodular sclerosis Hodgkin lymphoma: evaluation based on data from 1450 patients. Mayo Clin Proc. 2015; 90(6):756-764. Gotti M, Nicola M, Lucioni M, et al. Independent prognostic impact of tumourinfiltrating macrophages in early-stage Hodgkin's lymphoma. Hematol Oncol. 2017;35(3):296-302. Carlo-Stella C, Santoro A. Microenvironment-related biomarkers and novel targets in classical Hodgkin's lymphoma. Biomark Med. 2015;9(8):807-817. Montanari F, Diefenbach CS. Hodgkin lymphoma: targeting the tumor microenvironment as a therapeutic strategy. Clin Adv Hematol Oncol. 2015;13(8):518-524. Younes A, Ansell SM. Novel agents in the treatment of Hodgkin lymphoma: biological basis and clinical results. Semin Hematol. 2016;53(3):186-189. von TB, Moskowitz CH. Treatment of relapsed and refractory Hodgkin lymphoma. Semin Hematol. 2016;53(3):180-185. Weir SJ, DeGennaro LJ, Austin CP. Repurposing approved and abandoned drugs for the treatment and prevention of cancer through public-private partnership. Cancer Res. 2012;72(5):1055-1058. Halama N, Zoernig I, Berthel A, et al. Tumoral immune cell exploitation in colorectal cancer metastases can be targeted effectively by Anti-CCR5 therapy in cancer patients. Cancer Cell. 2016;29(4):587-601. Abel S, van der Ryst E, Rosario MC, et al. Assessment of the pharmacokinetics, safety and tolerability of maraviroc, a novel CCR5 antagonist, in healthy volunteers. Br J Clin Pharmacol. 2008;65 Suppl 1:5-18. Genebat M, Ruiz-Mateos E, Pulido I, et al. Long-term immunovirogical effect and tol-

haematologica | 2019; 104(3)


Maraviroc inhibits cHL crosstalk and xenograft growth

30.

31.

32.

33.

erability of a maraviroc-containing regimen in routine clinical practice. Curr HIV Res. 2010;8(6):482-486. Velasco-Velazquez M, Jiao X, De La Fuente M, et al. CCR5 antagonist blocks metastasis of basal breast cancer cells. Cancer Res. 2012;72(15):3839-3850. Halvorsen EC, Hamilton MJ, Young A, et al. Maraviroc decreases CCL8-mediated migration of CCR5(+) regulatory T cells and reduces metastatic tumor growth in the lungs. Oncoimmunology. 2016; 5(6):e1150398. Tanabe Y, Sasaki S, Mukaida N, Baba T. Blockade of the chemokine receptor, CCR5, reduces the growth of orthotopically injected colon cancer cells via limiting cancer-associated fibroblast accumulation. Oncotarget. 2016;7(30):48335-48345. Weiswald LB, Bellet D, Dangles-Marie V. Spherical cancer models in tumor biology. Neoplasia. 2015;17(1):1-15.

haematologica | 2019; 104(3)

34.

35.

36.

37.

38.

Mencarelli A, Graziosi L, Renga B, et al. CCR5 antagonism by maraviroc reduces the potential for gastric cancer cell dissemination. Transl Oncol. 2013;6(6):784-793. Lejmi E, Perriraz N, Clement S, et al. Inflammatory chemokines MIP-1delta and MIP-3alpha are involved in the migration of multipotent mesenchymal stromal cells induced by hepatoma cells. Stem Cells Dev. 2015;24(10):1223-1235. Poggi A, Zocchi MR. Stress immunity in lymphomas: mesenchymal cells as a target of therapy. Front Biosci (Landmark Ed). 2014;19:281-290. Zocchi MR, Catellani S, Canevali P, et al. High ERp5/ADAM10 expression in lymph node microenvironment and impaired NKG2D ligands recognition in Hodgkin lymphomas. Blood. 2012;119(6):1479-1489. Barros MH, Segges P, Vera-Lozada G, Hassan R, Niedobitek G. Macrophage polarization reflects T cell composition of tumor

39 .

40.

41.

42.

microenvironment in pediatric classical Hodgkin lymphoma and has impact on survival. PLoS One. 2015; 10(5):e0124531. Dorsam B, Bosl T, Reiners KS, et al. Hodgkin lymphoma-derived extracellular vesicles change the secretome of fibroblasts toward a CAF phenotype. Front Immunol. 2018;9:1358. Mantovani A, Marchesi F, Malesci A, Laghi L, Allavena P. Tumour-associated macrophages as treatment targets in oncology. Nat RevClin Oncol. 2017;14(7):399416. Carey CD, Gusenleitner D, Lipschitz M, et al. Topological analysis reveals a PD-L1associated microenvironmental niche for Reed-Sternberg cells in Hodgkin lymphoma. Blood. 2017;130(22):2420-2430. Jiao X, Velasco-Velazquez MA, Wang M, et al. CCR5 governs DNA damage and breast cancer stem cell expansion. Cancer Res. 2018;78(7):1657-1671.

575


ARTICLE Ferrata Storti Foundation

Haematologica 2019 Volume 104(3):576-586

Chronic Lymphocytic Leukemia

Mutations in the RAS-BRAF-MAPK-ERK pathway define a specific subgroup of patients with adverse clinical features and provide new therapeutic options in chronic lymphocytic leukemia

Neus Giménez,1,2* Alejandra Martínez-Trillos,1,3* Arnau Montraveta,1 Mónica Lopez-Guerra,1,4 Laia Rosich,1 Ferran Nadeu,1 Juan G. Valero,1 Marta Aymerich,1,4 Laura Magnano,1,4 Maria Rozman,1,4 Estella Matutes,4 Julio Delgado,1,3 Tycho Baumann,1,3 Eva Gine,1,3 Marcos González,5 Miguel Alcoceba,5 M. José Terol,6 Blanca Navarro,6 Enrique Colado,7 Angel R. Payer,7 Xose S. Puente,8 Carlos López-Otín,8 Armando Lopez-Guillermo,1,3 Elias Campo,1,4 Dolors Colomer1,4** and Neus Villamor1,4** Institut d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), CIBERONC, Barcelona; 2Anaxomics Biotech, Barcelona; 3Hematology Department and 4 Hematopathology Unit, Hospital Clinic, Barcelona; 5Hematology Department, University Hospital- IBSAL, and Institute of Molecular and Cellular Biology of Cancer, University of Salamanca, CIBERONC; 6Hematology Department, Hospital Clínico Universitario, Valencia: 7Hematology Department, Hospital Universitario Central de Asturias, Oviedo, and 8Departamento de Bioquímica y Biología Molecular, Instituto Universitario de Oncología, Universidad de Oviedo, CIBERONC, Spain. 1

*NG and AM-T contributed equally to the study.

**DC and NV share senior authorship of the manuscript.

ABSTRACT

Correspondence: DOLORS COLOMER dcolomer@clinic.cat Received: May 1, 2018. Accepted: September 26, 2018. Pre-published: September 27, 2018. doi:10.3324/haematol.2018.196931 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/104/3/576 ©2019 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

576

M

utations in genes of the RAS-BRAF-MAPK-ERK pathway have not been fully explored in patients with chronic lymphocytic leukemia. We, therefore, analyzed the clinical and biological characteristics of chronic lymphocytic leukemia patients with mutations in this pathway and investigated the in vitro response of primary cells to BRAF and ERK inhibitors. Putative damaging mutations were found in 25 of 452 patients (5.5%). Among these, BRAF was mutated in nine patients (2.0%), genes upstream of BRAF (KITLG, KIT, PTPN11, GNB1, KRAS and NRAS) were mutated in 12 patients (2.6%), and genes downstream of BRAF (MAPK2K1, MAPK2K2, and MAPK1) were mutated in five patients (1.1%). The most frequent mutations were missense, subclonal and mutually exclusive. Patients with these mutations more frequently had increased lactate dehydrogenase levels, high expression of ZAP-70, CD49d, CD38, trisomy 12 and unmutated immunoglobulin heavy-chain variable region genes and had a worse 5-year time to first treatment (hazard ratio 1.8, P=0.025). Gene expression analysis showed upregulation of genes of the MAPK pathway in the group carrying RAS-BRAF-MAPK-ERK pathway mutations. The BRAF inhibitors vemurafenib and dabrafenib were not able to inhibit phosphorylation of ERK, the downstream effector of the pathway, in primary cells. In contrast, ulixertinib, a pan-ERK inhibitor, decreased phospho-ERK levels. In conclusion, although larger series of patients are needed to corroborate these findings, our results suggest that the RAS-BRAF-MAPK-ERK pathway is one of the core cellular processes affected by novel mutations in chronic lymphocytic leukemia, is associated with adverse clinical features and could be pharmacologically inhibited. haematologica | 2019; 104(3)


Altered RAS-BRAF-MAPK-ERK pathway in CLL

Introduction The clinical course of patients with chronic lymphocytic leukemia (CLL) is highly heterogeneous.1,2 The mutational status of the immunoglobulin heavy-chain variable-region genes (IGHV) and deletions/mutations of 11q/ATM/BIRC3 and 17p/TP53 are important determinants of the clinical outcome of patients with CLL.3–6 Whole genome sequencing and whole exome sequencing have identified recurrent acquired mutations in the coding and non-coding regions of several genes. A few of them are mutated with moderate/low frequencies (11-15%), whereas the majority are mutated at much lower frequencies (2-5%).7–10 This mutational landscape highlights the patients’ heterogeneity. Several of the mutations, including some with a low incidence, have been reported to be associated with particular clinical features and disease evolution.9,11–13 BRAF is a member of the serine-threonine kinase RAF family, comprising RAF-1/CRAF, ARAF, and BRAF. In normal cells, BRAF functions as a mitotic signal transporter in the RAS/RAF/mitogen-extracellular signal-regulated kinase 1/2 (MEK1/2)/ extracellular signal-regulated kinase 1/2 (ERK1/2)/mitogen activated protein kinase (MAPK) pathway. This pathway plays a pivotal role in regulating embryogenesis, cell proliferation, differentiation, migration, and survival.14 In the last decade, a high frequency of BRAF point mutations has been identified in melanoma and other human cancers.15,16 BRAF mutations are also a characteristic of hairy cell leukemia (HCL), being detected in 95% to 100% of patients with this type of leukemia.17,18 The most common BRAF mutation leads to the substitution of a valine for glutamic acid at amino acid 600 (V600E) in the kinase domain of the protein. This substitution mimics the phosphorylation of the activation loop, thereby leading to its constitutive activation and phosphorylation of MEK1 and MEK2, which in turn phosphorylate and activate the effector kinases ERK1 and ERK2.19 ERK proteins target numerous substrates, such as protein kinases, transcription factors, and cytoskeletal or nuclear proteins. Moreover, they are able to affect protein functions either by phosphorylating proteins in the cytoplasm or by translocating them into the nucleus where they activate transcription factors that regulate proliferation- and cell survival-associated genes.20 BRAF mutations have been recurrently reported in CLL patients with a frequency of approximately 3%;21–24 most of these mutations cluster within or near the activation loop. Recently, novel CLL drivers (NRAS, KRAS, NRAS and MAP2K1) of the RAS-BRAF-MAPK-ERK pathway have also been described.9-24 However, the impact of BRAF mutations and other mutations in the RAS-BRAF-MAPKERK pathway in CLL is not well established. We analyzed the clinical and biological characteristics and the impact of mutations in genes of the RAS-BRAFMAPK-ERK pathway in CLL patients, the functional implications of these mutations and the in vitro response to different MAPK inhibitors.

Consortium for CLL (ICGC-CLL)7 were analyzed. All patients gave informed consent to inclusion in this study, according to the guidelines of the ICGC-CLL project and the local ethics committees. The study was conducted in accordance with the Declaration of Helsinki.

Primary chronic lymphocytic leukemia cells CLL cells were isolated, cryopreserved and stored in the Hematopathology collection registered at the Biobank (Hospital Clínic-IDIBAPS; R121004-094) (Online Supplementary Methods). Functional studies were done in all patients with mutations in genes of the RAS-BRAF-MAPK-ERK pathway for whom cryopreserved material was available.

Mutational analysis Whole exome sequencing or whole genome sequencing was performed in 452 CLL patients. DNA from purified CLL cells (>95% tumor cells) was obtained before administration of any treatment, as described elsewhere.7 The median interval between diagnosis and sample analysis was 36 months (range, 0-300 months). Mutations in genes of the RAS-BRAF-MAPK-ERK pathway according to the Kyoto Encyclopedia of Genes and Genomes (KEGG) database (KITLG, KIT, SOS2, PTPN11, GNB1, KRAS, NRAS, BRAF, MAP2K1, MAP2K2 and MAPK1) were selected for further analysis. Clonal mutations were considered when the variant allele frequency (VAF) was ≥0.40 and subclonal when the VAF was <0.40. PolyPhen-2, SIFT and CADD algorithms were used for in silico prediction of the pathogenicity of the mutations. Coding mutations were considered pathogenic if they were reported as such by at least two algorithms (probably damaging by PolyPhen2 and/or damaging by SIFT and/or with a phred-like score >20 by CADD).

Gene expression analysis The gene expression profile of 143 purified CLL samples with unmutated IGHV genes (U-IGHV) from the CLL-ICGC project7 was analyzed using the Gene Set Enrichment Analysis (GSEA) package version 2.0. Enrichment of the MAPK gene signature was investigated using the C2 Biocarta and C2 KEGG collection version 6.1 as reported in the Online Supplementary Methods. Gene sets with a P≤0.05, a false discovery rate (FDR) q-value ≤10% and a normalized enrichment score (NES) ≥1.5 were considered to be significantly enriched in the group with mutations in the RASBRAF-MAPK-ERK pathway.

Western blot analysis Whole-cell protein extracts were obtained from CLL cells and peripheral blood mononuclear cells from healthy donors and western blot was performed with antibodies against phosphorylated-T202/Y204 ERK 1/2 and total ERK (Santa Cruz Biotechnology, Santa Cruz, CA, USA) (Online Supplementary Methods).

Analysis of viability Vemurafenib, dabrafenib, and ulixertinib (BVD-523) were purchased from Selleckchem (Houston, TX, USA). Primary CLL cells were incubated for 24 or 48 h with the indicated doses of the drugs and then stained and analyzed as reported in the Online Supplementary Methods.

Methods Patients

B-cell receptor stimulation and quantification of phosphorylated ERK by flow cytometry

Four hundred fifty-two patients (276 males/176 females) diagnosed with CLL according to the World Health Organization criteria25 and included in the International Cancer Genome

B-cell receptors were stimulated by incubating CLL cells with 10 mg/mL of anti-IgM (Southern Biotech, Birmingham, AL, USA) and cells were stained for phospho (T202 and Y204)-ERK1/2-phyco-

haematologica | 2019; 104(3)

577


N. Giménez et al.

erythrin (Becton Dickinson, Franklin Lakes, NJ, USA) (Online Supplementary Methods).

Results

Statistical analysis

Clinical and biological impact of mutations in the RASBRAF-MAPK-ERK pathway

A Fisher test or non-parametric tests were used to correlate clinical and biological variables according to the presence of mutations in the RAS-BRAF-MAPK-ERK pathway. Time to first treatment (TTFT) was calculated from the date of sampling to the first treatment or last follow-up. Overall survival was calculated from the date of sampling to the date of death or last follow-up. All the analyses were conducted using SPSS 20 (www.ibm.com) software and are detailed in the Online Supplementary Methods. For primary cell cultures data are presented as the mean ± standard error of the mean. Comparisons between groups were evaluated with a Wilcoxon paired test using GraphPad Prism 4.0 software. Results were considered statistically significant when the P-value was ≤0.05.

Four hundred fifty-two patients (276 males/176 females) with CLL were analyzed for the clinical and biological impact of mutations in genes of the RAS-BRAFMAPK-ERK pathway (see Online Supplementary Table S1 for the main characteristics of the series). A total of 31 mutations affecting genes of the RASBRAF-MAPK-ERK pathway were observed in 30 of the 452 CLL patients (7%) (Online Supplementary Figure S1 and Table 1). Mutations were missense (25/31; 81%) or noncoding mutations at the 3’ or splice donor regions (6/31; 19%). The mean VAF for the 31 individual mutations was 0.36 ± 0.13. According to the results of the PolyPhen-2, SIFT and CADD algorithms used to predict the patho-

Table 1. Description of the mutations in genes of the RAS-BRAF-MAPK-ERK pathway in patients with chronic lymphocytic leukemia.

Case Patient 1 2 3 4 5 6 7 8 9* 10 11 12 13 9* 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

723 33 1078 850 191 677 1192 1226 155 15 1564 398 598 155 1371 27 100 134 148 279 721 824 1079 1431 44 1365 884 761 1477 1568 442

Gene name KITLG KIT KIT SOS2 PTPN11 PTPN11 PTPN11 PTPN11 PTPN11 GNB1 GNB1 KRAS KRAS KRAS NRAS BRAF BRAF BRAF BRAF BRAF BRAF BRAF BRAF BRAF MAP2K1 MAP2K1 MAP2K2 MAP2K2 MAP2K2 MAP2K2 MAPK1

HGVS.p

Annotation

n.a. 3' UTR p.Val833Leu missense n.a. 3' UTR p.Pro7Ser missense p.Ala72Val missense p.Glu76Lys missense p.Asp61Val missense p.Asp61Val missense p.Ser502Pro missense p.Ile80Thr missense n.a. 3' UTR p.Gly12Val missense p.Gln61His missense p.Gly12Asp missense p.Gln61Arg missense p.Glu501Lys missense p.Lys601Glu missense p.Gly469Ala missense p.Lys601Asn missense p.Asp594Gly missense p.Asn581Ser missense p.Leu597Gln missense p.Val600Glu missense p.Gly534Arg missense p.Phe53Cys missense p.Gly128Asp missense n.a. splice donor p.Gln60Pro missense n.a. 3' UTR p.Tyr134Cys missense n.a. 3' UTR

PolyPhen-2 predictiona n.a. Probably damaging n.a. Benign Probably damaging Probably damaging Probably damaging Probably damaging Possibily damaging Probably damaging n.a. Probably damaging Benign Possibily damaging Benign Probably damaging Possibily damaging Probably damaging Possibily damaging Probably damaging Probably damaging Probably damaging Probably damaging Possibily damaging Probably damaging Probably damaging n.a. Probably damaging n.a. Probably damaging n.a.

SIFT CADD predictionb phred-like scorec

VAF

IGHV

TP53

BIRC3

ATM

n.a. Damaging n.a. Tolerated Damaging Damaging Damaging Damaging Damaging Damaging n.a. Damaging Damaging Damaging Damaging Damaging Damaging Damaging Damaging Damaging Damaging Damaging Damaging Damaging Damaging Damaging n.a. Damaging n.a. Damaging n.a.

0.39 0.24 0.55 0.50 0.58 0.54 0.17 0.50 0.15 0.42 0.31 0.18 0.42 0.30 0.22 0.15 0.20 0.54 0.38 0.49 0.48 0.25 0.33 0.46 0.19 0.29 0.39 0.26 0.43 0.33 0.43

UM M UM M M UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM M UM M UM UM

UM UM UM UM UM M UM UM UM UM UM UM UM UM UM UM UM UM UM M UM UM UM UM UM UM UM UM UM M UM

M UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM M UM UM UM

M UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM UM M UM M UM UM UM

4.25 22.70 5.91 10.21 32.00 33.00 28.20 28.20 31.00 28.10 1.21 29.90 23.50 25.30 23.10 34.00 24.50 27.50 24.30 29.70 19.38 28.80 32.00 34.00 29.10 32.00 23.30 24.70 11.64 27.00 12.71

*CLL case with two mutations in genes of the RAS-BRAF-MAPK-ERK pathway; aAdzhubei I, Jordan DM, Sunyaev SR. Predicting functional effect of human missense mutations using PolyPhen-2. Curr Protoc Hum Genet. 2013; Chapter 7: Unit 7.20. bNg PC, Henikoff S. SIFT: Predicting amino acid changes that affect protein function. Nucleic Acids Res. 2003 Jul 1;31(13):3812-4. cKircher M, Witten DM, Jain P, O'Roak BJ, Cooper GM, Shendure J. A general framework for estimating the relative pathogenicity of human genetic variants. Nat Genet. 2014;46(3):310-5. HGVS.p: Human Genome Variation Society protein sequence; PolyPhen-2: Polymorphism Phenotyping v2; SIFT: Sorting Intolerant From Tolerant; CADD: Combined Annotation-Dependent Depletion; VAF: variant allele frequency; IGVH: immunoglobulin variant heavy chain genes; 3’UTR: 3’ untranslated region; n.a. not applicable; M: mutated, UM: unmutated.

578

haematologica | 2019; 104(3)


Altered RAS-BRAF-MAPK-ERK pathway in CLL

genicity of the mutations, five mutations in the 3’ untranslated region (cases 1, 3, 11, 28 and 30) and one missense mutation (case 4, SOS2 gene) were discarded as not being pathogenic. We were able to demonstrate that the mutation in the 3’ untranslated region of KITLG (case 1) was functional as we detected high levels of phosphorylated ERK, a surrogate marker of RAS-BRAF-MAPK-ERK pathway activation (Figure 3A). Due to the absence of cryopreserved material, we could not analyze the functionality of these mutations in the remaining cases. Therefore, considering only the putative functional mutations, a total of 26 functional mutations affecting genes of the RAS-BRAFMAPK-ERK pathway were observed in 25 of 452 CLL patients (5.5%). In 11 of the 25 patients (44%) these mutations were clonal (VAF ≥0.40) and in the other 14 patients (56%) they were subclonal (VAF <0.40). Mutations were detected in genes upstream of BRAF (KITLG, KIT, PTPN11, GNB1, KRAS and NRAS) in 12/452 patients (2.6%), in BRAF in 9/452 patients (2.0%), and in genes downstream of BRAF (MAP2K1 alias MEK1, MAP2K2 alias MEK2) in 5/452 patients (1.1%). The most frequent single mutated gene was BRAF (n=9/26, 34.6%) followed by PTPN11 (n=5/26, 19.2%), MAP2K2 (n=3/26, 11.5%), KRAS (n=3/26, 11.5%), and MAP2K1, (2/26 cases, 7.7%); mutations of GNB1, NRAS, KIT, and KITLG were each found in one patient. One patient had concomitant mutations of PTPN11 and KRAS. Interestingly, BRAF mutations were localized between exons 11 to 15 and most of them occurred in the activation loop (A-loop) near the V600 position or near the phosphate-binding loop (P-loop) at residues 464-469. Only in one case did the BRAF mutation correspond to V600E, the most common mutation described in a variety of human malignancies including HCL.17

Association of mutations in the RAS-BRAF-MAPK-ERK pathway with clinical and biological features The main clinical and biological characteristics of the 25 patients with functional mutations in the RAS-BRAFMAPK-ERK pathway are listed in Table 2. The age, sex and clinical stage of the patients with mutations in the RAS-BRAF-MAPK-ERK pathway were similar to those of the patients without mutations. However, patients with mutations in RAS-BRAF-MAPKERK pathway genes more frequently had abnormal values of lactate dehydrogenase, high expression of ZAP-70, CD38 and CD49d, trisomy 12 and most of them had UIGHV (21/24, 87%) (P≤0.05 in all comparisons) (Table 2). Patients with mutations in the RAS-BRAF-MAPK-ERK pathway more frequently had three or more driver mutations than patients without mutations in the pathway, but no differences were observed in the genes most frequently mutated in CLL (NOTCH1, SF3B1, BIRC3, TP53 or ATM) (Table 2). Six cases contemporaneously carried mutations in TP53, ATM or BIRC3. As most patients with mutations in the RAS-BRAF-MAPK-ERK pathway had U-IGHV, we conducted a similar analysis including only the subgroup of U-IGHV patients. As seen in Table 3, only lactate dehydrogenase and trisomy 12 maintained statistical significance. Figure 1 shows a brick-plot of concomitant gene mutations/cytogenetic aberrations for cases with RASBRAF-MAPK-ERK pathway mutations. Patients with mutations in the RAS-BRAF-MAPK-ERK pathway required treatment more frequently, considering both the whole group (88% versus 43%; P<0.001) and haematologica | 2019; 104(3)

within the U-IGHV subgroup (95% versus 75%; P<0.048). There were no differences in the type of treatment received or the response achieved according to the presence or absence of mutations in the pathway (Table 2). Five-year TTFT of patients with Binet A or B disease was 82% [95% confidence interval (95% CI): 66-98%] in patients with mutations in the RAS-BRAF-MAPK-ERK pathway versus 50% (95% CI: 42-58%) in the unmutated group; P<0.001]. The comparison between clonal and subclonal mutated cases showed that the 5-year TTFT was

Table 2. Main clinical and biological characteristics of patients according to mutations in the RAS-BRAF-MAPK-ERK pathway.

Parameter

Category

Gender Male (%) Age (years), median (range) A Binet stage B C 0 Rai stage I-II III-IV Lymphocytes count (x109/L), median (range) Absolute CLL cell count (x109/L), median (range) Hemoglobin (g/L), median (range) Platelets (x109/L), median (range) B2 microglobulin UNV* Lacate dehydrogenase UNV* IGHV Unmutated CD49d >30% CD38 >30% ZAP-70 ≥20% Genetics del(13q)(q14.3) Trisomy 12 del(11q)(q22.3) del(17p)(p13.1) Driver mutations ≥3 NOTCH1 Mutated SF3B1 Mutated TP53 Disrupted BIRC3 Disrupted ATM Disrupted Treated 184/427 (43%) Response to treatment* CR PR Failure 5-year TTFT (95% CI)* A&B 5-year OS (95% CI) All 5-year t-DLBCL All

Unmutated (n=427)

Mutated P-value (n=25)

257 (60%) 61 (18-93) 366 (87%) 47 (11%) 8 (2%) 278 (66%) 130 (31%) 12 (3%) 11 (1-203)

19 (76%) 61 (44-84) 21 (88%) 1 (4%) 2 (8%) 13 (54%) 9 (38%) 2 (8%) 11 (1-75)

8 (0.4-192)

6 (0.7-83)

ns ns ns

ns ns ns

141 (45-177)

147 (125-159) ns

204 (49-791)

170 (99-315)

ns

119/373 (32%) 26/407 (6%) 145/421 (34%) 92/290 (32%) 96/403 (24%) 98/394 (25%) 148/308(48%) 48/308 (16%) 26/307 (8%) 11/308 (4%) 159/427(37%) 52/427 (12%) 38/427(9%) 21/397 (5%) 38/427 (9%) 47/427 (11%) 22/25 (88%)

7/18 (39%) 6/19 (32%) 21/24 (87%) 9/13 (69%) 10/23 (43%) 14/21 (67%) 3/13 (23%) 6/13 (46%) 0/13 (0%) 1/13 (8%) 17/25 (68%) 5/25 (20%) 1/25 (4%) 2/23 (9%) 2/25 (8%) 3/25 (12%) <0.001

ns 0.002 <0.001 0.012 0.046 <0.001 ns 0.011 ns ns 0.003 ns ns ns ns ns

102 (55%) 48 (26%) 13 (7%) 50% (42-58) 80% (74-86) 2% (1-3)

12 (57%) 4 (19%) ns 3 (14%) 82% (66-98) <0.001 78% (60-96) ns 11% (0-25) 0.080

*It was not possible to assess the response to treatment in 21/184 (11%) of the unmutated patients and in 2/21 (9%) of the mutated patients. CLL: chronic lymphocytic leukemia; UNV: above normal value; CR: complete response, PR: partial response, TTFT: time to first treatment; OS: overall survival; 95% CI: 95% confidence interval; t-DLBCL: transformation into diffuse large B-cell lymphoma (Richter syndrome); ns: not significant.

579


N. Giménez et al.

92% (95 CI: 76-100%) for patients with subclonal mutations, 70% (95 CI: 42-98%) for patients with clonal mutations, and 51% (95% CI: 42-60%; P≤0.001) for those without mutations. The adverse effect of mutations in genes of the RAS-BRAF-MAPK-ERK pathway was observed independently of the mutated gene (Online Supplementary Figure S2). Overall, patients with mutations in the RAS-BRAF-MAPK-ERK pathway had a worse TTFT than that of patients without mutations (P<0.001) (Figure 2A). However, when other adverse mutations (TP53, ATM or BIRC3)26,27 were taken into account, patients with mutations in both the RAS-BRAF-MAPK-ERK pathway and in TP53, ATM or BIRC3 (n=6, 1%) had the shortest 5-year TTFT (100%) followed by patients with mutations in TP53, ATM or BIRC3 [n=64,15%; 5-year TTFT of 83% (CI 95%: 71-95%)], patients with mutations only in the RASBRAF-MAPK-ERK pathway [n=16, 4%; 5-year TTFT of 75% (CI 95%: 54-96%)], and patients without mutations [n=337, 79%; 5-year TTFT of 44% (CI 95%: 34-54%)]

(P≤0.001) (Figure 2B). In the subgroup of patients with Binet A or B CLL with U-IGHV, those patients with adverse gene mutations concomitantly with mutations in RAS-BRAF-MAPK-ERK pathway genes (n=6, 4%) again had a worse 5-year TTFT (all treated) than patients with only mutations in TP53, ATM or BIRC3 (n=45, 30%; 5year TTFT: 87%, CI 95%: 77-97%), patients with only mutations in RAS-BRAF-MAPK-ERK pathway genes (n=13, 8%; 5-year TTFT: 85%, CI 95%: 65-100%), and patients without mutations in these genes (n=88, 56%; 5year TTFT: 71%, CI: 95%: 60-82%) (P=0.001) (Figure 2C). A multivariate analysis including IGHV status, mutations in RAS-BRAF-MAPK-ERK pathway genes, and mutations in TP53, ATM or BIRC3 in a final model with 418 patients showed an independent impact on TTFT for IGHV status [hazard risk (HR) 3.4 (95% CI: 2.5-4.8), P<0.001], mutations in the RAS-BRAF-MAPK-ERK pathway [HR 1.8 (95% CI: 1.1- 3), P=0.016] and adverse mutations [HR 2.0 (95% CI: 1.5-2.8), P<0.001].

Table 3. Main clinical and biological characteristics of patients according to the presence or absence of mutations in genes of the RAS-BRAFMAPK-ERK pathway in the subgroup with unmutated IGHV chronic lymphocytic leukemia.

Parameter

Category

Unmutated (n=145)

Mutated (n=21)

P-value

Gender Age (years), median (range)

Male (%)

94 (65%) 61 (18-93) 105/142 (74%) 32/142 (22%) 5/142 (4%) 67/141 (47%) 66/141 (47%) 8/141 (6%) 10.7 (1-106) 8 (0.8-114) 140 (45-166) 210 (49-470) 57/128 (45%) 15/137 (11%) 42/89 (47%) 61/136 (45%) 77/1131 (59%) 38/102 (37%) 22/102 (22%) 21/102 (20%) 5/102 (5%) 96/145 (66%) 43/145 (30%) 21/145 (15%) 9/134 (7%) 30/145 (21%) 38/145 (26%) 108/145 (75%) 78% (68-88) 70% (60-80) 9% (5-13)

16 (76%) 61 (44-78)ns 18/20 (90%) 1/20 (5%) 1/20 (5%) 11/20 (55%) 8/20 (40%) 1/20 (5%) 12 (1-26) 7 (0.7-83) 149 (125-159) 163 (99-315) 7/15 (47%) 6/16 (37%) 8/11 (73%) 10/19 (53%) 13/18(72%) 1/10 (10%) 6/10 (60%) 0/10 (9%) 1/10 (10%) 14/21(67%) 5/21 (24%) 1/21 (5%) 2/20 (10%) 2/21 (12%) 3/21 (14%) 20/21 (95%) 90% (76-100) 84% (64-100) 12% (0-26)

ns

Binet stage

Rai stage Lymphocytes count (x109/L), median (range) Absolute CLL cells count (x109/L), median (range) Hemoglobin (g/L), median (range) Platelets (x109/L), median (range) B2 microglobulin Lactate dehydrogenase CD49d CD38 ZAP-70 Genetics

Driver mutations NOTCH1 SF3B1 TP53 BIRC3 ATM Treated 5-year TTFT (95% CI) 5-year OS (95% CI) 5-year t-DLBCL

A B C 0 I-II III-IV

UNV UNV >30% >30% ≥20% del(13q)(q14.3) Trisomy 12 del(11q)(q22.3) del(17p)(p13.1) ≥3 Mutated Mutated Disrupted Disrupted Disrupted A&B U-IGHV All

ns

ns ns ns ns ns ns 0.011 ns ns ns ns 0.015 ns ns ns ns ns ns ns ns 0.048 0.025 0.020 ns

TTFT: time to first treatment; OS: overall survival; 95% CI: 95% confidence interval; t-DLBCL: transformation into diffuse large B-cell lymphoma (Richter syndrome); ns: not significant: UNV: above normal value: 95% CI: 95% confidence interval; U-IGHV: unmutated IGHV genes

580

haematologica | 2019; 104(3)


Altered RAS-BRAF-MAPK-ERK pathway in CLL

The overall survival of patients with mutations in RASBRAF-MAPK-ERK pathway genes was similar to that of patients without mutations in this pathway (Table 2). When mutations in TP53, ATM or BIRC3 were taken into account, the overall survival of patients with mutations in genes of the RAS-BRAF-MAPK-ERK pathway alone was similar to that of patients without adverse mutations (Figure 2D) [5-year overall survival of patients without mutations, 84% (95% CI: 78-92%); with mutations only in the RAS-BRAF-MAPK-ERK pathway, 80% (95% CI: 64-99%); with adverse mutations only, 66% (95% CI: 5379%); and with both abnormalities in RAS-BRAF-MAPKERK pathway genes and adverse mutations, 66% (95% CI: 45-100%), P=0.003]. Multivariate analysis including IGHV status, mutations in genes of the RAS-BRAFMAPK-ERK pathway, and adverse mutations in a final model with 439 patients showed an independent impact on overall survival for IGHV status [HR 3.3 (95% CI: 1.95.9), P<0.001] and adverse mutations [HR 1.7 (95% CI: 1.1-2.8), P=0.02].

Functional and gene expression analysis To assess the functional impact of these genomic alterations on the RAS-BRAF-MAPK-ERK pathway, we analyzed the phosphorylation status of ERK as a surrogate marker of activation of the pathway. Western blotting with an antibody that specifically recognizes the dually phosphorylated and active forms of ERK1 and ERK2 showed higher levels of endogenous ERK phosphorylation (3.3- to 4.4-fold induction) in CLL cases with mutations in KITLG, BRAF, MAP2K2 and MAP2K1 genes compared to U-IGHV CLL cases with no alterations in the MAPK/ERK pathway (Figure 3A). The same results were obtained when analyzing the phosphorylated forms of ERK by flow cytometry, labeling cells with phospho (T202/Y204)-ERK1/2-phycoerythrin. Figure 3B shows that cases with mutations in genes of the RAS-BRAF-MAPK-

ERK pathway (PTPN11, BRAF, and MAP2K1 mutations) had higher basal levels of phosphorylated ERK than cases of U-IGHV CLL (5- to 10-fold). To identify the differential biological characteristics of cells carrying mutations in the RAS-BRAF-MAPK-ERK pathway, we conducted a gene expression profiling study in CD19+ tumor CLL cells from 143 CLL cases, 17 of which carrying functional mutations according to PolyPhen-2, SIFT and CADD phred-like predictions. With the C2 Biocarta analysis, we detected 126 of 149 gene sets upregulated in the group carrying mutations in genes of the RAS-BRAF-MAPK-ERK pathway, including the Biocarta MAPK pathway (NES=1.90; P<0.001; FDR=0.013) (Online Supplementary Table S2 and Figure 3C). Similar results were obtained when carrying out a C2 KEGG analysis. We detected 104 of 178 gene sets upregulated in the group carrying mutations in genes of the RAS-BRAF-MAPK-ERK pathway, including the KEGG MAPK signaling pathway (NES=1.85; P<0.001; FDR=0.013) (Online Supplementary Table S3 and Figure 3D). Genes belonging to the Biocarta and KEGG MAPK pathways are listed in Online Supplementary Tables S4 and S5, respectively.

Response to MAPK pathway inhibitors We next evaluated the effect of BRAF inhibitors (vemurafenib, a specific inhibitor of the BRAF V600E mutation, and dabrafenib, specific for BRAF V600E and V600K variants) in cells from 17 CLL cases, nine containing mutations in genes of the RAS-BRAF-MAPK-ERK pathway (KITLG, PTPN11, KRAS, BRAF, MAPK1, MAP2K1 and MAP2K2) and eight U-IGHV CLL cases with no alterations in this pathway. Vemurafenib, at a dose of 2.5 mM, was not able to inhibit basal ERK phosphorylation or after anti-IgM stimulation in mutated cases, while a slight effect was observed after treatment with 2.5 mM of dabrafenib. Furthermore, upregulation of

Figure 1. Brick-plot showing gene mutations, cytogenetic abnormalities and the type of RAS-BRAF-MAPK-ERK pathway mutations. Clonal mutations are labeled in dark blue, subclonal mutations in light blue, normal genes or chromosomal regions in light gray, and mutated/deleted genes or chromosomal regions in dark gray. Adverse alterations: TP53, ATM or BIRC3.

haematologica | 2019; 104(3)

581


N. GimĂŠnez et al.

phosphorylated ERK, was observed in the U-IGHV CLL cases with no mutations in the RAS-BRAF-MAPK-ERK pathway after incubation with 2.5 mM of dabrafenib (P<0.05) (Figure 4A). We next analyzed the cytotoxic effect of these drugs at different doses (0.5 to 5 mM) and times (24 h and 48 h): vemurafenib did not have any cytotoxic effect, while dabrafenib exerted some degree of cytotoxicity at the higher doses in both mutated RAS-BRAF-MAPK-ERK cases and U-IGHV CLL cases after 24 h of incubation (P<0.05) and at all doses after 48 h of incubation (P<0.05 at 0.5 mM and P<0.01 at 1-5 mM) (Figure 4B).

A

B

C

D

Finally, we compared the effect of the pan-ERK inhibitor ulixertinib (BVD-523) in six patients carrying mutations in the RAS-BRAF-MAPK-ERK pathway (KITLG, PTPN11, BRAF, MAP2K1, MAP2K2 and MAPK1) and six U-IGHV CLL cases without mutations. In contrast to the lack of effect of vemurafenib and dabrafenib at 2.5 mM, ulixertinib was able to inhibit basal ERK phosphorylation (by 60%) in all cases with mutations in the RAS-BRAF-MAPKERK pathway at doses of 2.5 mM, and after stimulation with anti-IgM at much lower doses (100 nM) (Figure 4C). This effect was not observed in RAS-BRAF-MAPK-ERK pathway unmutated, U-IGHV cells.

Figure 2. Outcome of patients according to mutations in genes of the RAS-BRAF-MAPK-ERK pathway. (A) Time to first treatment (TTFT) in Binet stage A and B patients according to mutations in the RAS-BRAF-MAPK-ERK pathway (the green line represents patients with clonal mutations, the orange line represents patients with subclonal mutations and the blue line represents patients with no mutations in the RAS-BRAF-MAPK-ERK pathway). (B) TTFT in Binet stage A and B patients according to the presence or absence of mutations in the RAS-BRAF-MAPK-ERK pathway and/or adverse mutations (TP53, ATM or BIRC3). (C) TTFT in U-IGHV CLL Binet A and B patients according to the presence or absence of mutations in the RAS-BRAF-MAPK-ERK pathway and/or adverse mutations (TP53, ATM or BIRC3). (D) Overall survival of all CLL patients according to the presence or absence of mutations in the RAS-BRAF-MAPK-ERK pathway and/or adverse mutations (TP53, ATM or BIRC3).

582

haematologica | 2019; 104(3)


Altered RAS-BRAF-MAPK-ERK pathway in CLL

Discussion CLL is characterized by a heterogeneous mutational landscape, with the presence of certain mutations being associated with progression of the disease and refractoriness to immuno-chemotherapy, which lead to a poor outcome.6,13,28 Recently, it has been proposed that the MAPKâ&#x20AC;&#x201C; ERK pathway could be one of the cellular processes affected in CLL through mutations in novel CLL drivers such as NRAS, KRAS, BRAF, PTPN11 and MAP2K1.9,24 The RASBRAF-MAPK-ERK pathway plays a central role not only in regulating normal cellular processes involved in proliferation, growth, and differentiation, but also in oncogenesis,29 and it is an important key dysregulated pathway in cancer.30 In our series, we observed mutations in genes belonging to the RAS-BRAF-MAPK-ERK pathway in 5% of CLL patients, a frequency similar to that already described.13 When we evaluated each mutation specifically, BRAF mutations were detected in 2% of our CLL series, as previously reported.9,21 BRAF mutations did not involve the canonical hotspot (V600E) seen in other malignancies,17 which leads to constitutive activation of BRAF, but rather were clustered around the activation segment of the

kinase domain.9,23 Mutations in these positions confer variable but increased signaling and have oncogenic capacity.31 Mutations in exon 15 of BRAF have been associated with refractoriness to fludarabine22 although they do not seem to be selected during progression to refractory CLL.21 Furthermore, the frequency of BRAF V600E mutations is higher in Richter syndrome than in untransformed CLL32, and this mutation could be acquired during the evolution of CLL. Recently, our group reported that the mere detection of a BRAF mutation, even at a very low frequency, had a prognostic impact on TTFT.33 However, given the low frequency of mutations observed in CLL patients, larger series of patients are needed to corroborate these observations. Mutations in genes upstream and downstream of BRAF were observed in 64% (16/25) of cases. MAP2K1 mutations have already been described in HCL-variant and conventional HCL with rearranged IGHV4-34,34 Langerhans cell histiocytosis,35 and pediatric-type follicular lymphoma.36 This mutation, similar to those of BRAF, leads to activation of the downstream target, ERK.36 Moreover, we found mutations in additional genes of this pathway, such as MAP2K2, which encodes MEK2, and PTPN11, which encodes SHP-2. Both these proteins participate in the reg-

A

B

C

D

haematologica | 2019; 104(3)

Figure 3. Activation of the RAS-BRAF-MAPK-ERK pathway. (A) Basal phosphorylated (p)-ERK and ERK levels analyzed by western blot in cases of chronic lymphocytic leukemia (CLL) with mutations in genes of the RAS-BRAF-MAPK-ERK pathway (case 1: KITLG mutation, case 16: BRAF mutation, case 27: MAP2K2 mutation and case 25: MAP2K1 mutation), in unmutated IGHV (U-IGHV) CLL and in peripheral blood mononuclear cells (PBMC). Îątubulin was used as a loading control. p-ERK/ERK levels were quantified relative to the U-IGHV case. (B) Basal p-ERK levels were analyzed by flow cytometry in CLL cases with mutations in genes of the RAS-BRAF-MAPK-ERK pathway (case 6: PTPN11 mutation, case 15: BRAF mutation and case 25: MAP2K1 mutation). Expression levels are relative to those in U-IGHV CLL. (C) Gene set enrichment analysis (GSEA) plots of the BiocartaMAPK and KEGG MAPK signaling pathway gene sets regarding mutational status in genes of the RAS-BRAF-MAPK-ERK pathway in U-IGHV cases. The enrichment plot contains profiles of the running enrichment scores (ES) and positions of gene set members on the rank ordered list in GSEA (126 unmutated and 17 mutated CLL cases). NES, normalized enrichment score. FDR, false discovery rate.

583


N. Giménez et al. A

B

C

Figure 4. Effect of RAS-BRAF-MAPK-ERK inhibitors in cases of RAS-BRAF-MAPK-ERK-mutated and unmutated IGHV chronic lymphocytic leukemia. (A) Cells from 17 cases of chronic lymphocytic leukemia (CLL), nine containing mutations in the RAS-BRAF-MAPK-ERK pathway (KITLG, PTPN11, KRAS, BRAF, MAPK1, MAP2K1, MAP2K2) and eight with unmutated IGHV genes (U-IGHV) with no alterations in genes of the RAS-BRAF-MAPK-ERK pathway were treated with vemurafenib 2.5 mM or dabrafenib 2.5 mM. p-ERK levels were analyzed by flow cytometry after 1.5 h of treatment and expressed relative to untreated cells (Ct) at basal levels (unstimulated) or after stimulation with anti-IgM (stimulated) (*P<0.05). Histograms showing anti-IgM stimulation of ERK (T202/Y204) phosphorylation with and without vemurafenib or dabrafenib (2.5 mM) treatment in representative CLL cases (U-IGHV CLL and case 17 with a BRAF mutation). (B) Cell viability after treatment for 24 and 48 h with vemurafenib or dabrafenib at doses of 0.5-5 mM. Bars represent the mean ± standard error of the mean (SEM) of all samples analyzed (n=9 in the group of CLL cases with mutations in genes of the RAS-BRAF-MAPK-ERK pathway and n=8 in the unmutated CLL group) (*P<0.05; **P<0.01). (C) p-ERK levels after treatment with 0.1 or 2.5 mM ulixertinib (UT) relative to untreated (Ct) samples analyzed by flow cytometry at basal levels (unstimulated) or after stimulation with anti-IgM (stimulated). Bars represent the mean ± SEM of six samples analyzed in each group, six with mutations in genes of the RAS-BRAF-MAPK-ERK pathway (KITLG, PTPN11, BRAF, MAP2K1, MAP2K2, and MAPK1) and six U-IGHV CLL cases. Histograms showing anti-IgM stimulation of ERK (T202/Y204) phosphorylation and its inhibition by 100 nM and 2.5 mM ulixertinib (UT) in representative CLL cases (U-IGHV: CLL and case 15: BRAF mutation). Each patient is represented by a different color depending on the RAS-BRAF-MAPK-ERK mutational status and the mutation position relative to BRAF.

584

haematologica | 2019; 104(3)


Altered RAS-BRAF-MAPK-ERK pathway in CLL

ulation of the RAS-BRAF-MAPK-ERK signaling pathway.37 Mutations in this pathway seem to be mutually exclusive as only in one case were two different mutations observed simultaneously in the pathway. In this way, oncogene mutations that activate common downstream pathways often occur in a mutually exclusive fashion,38 as has been reported for BRAF and MAP2K1 in HCL-variant.34 The upregulation of genes of the MAPK pathway observed in the gene expression profiling analysis as well as the higher levels of phosphorylated ERK, a surrogate marker of MAPK pathway activation,39 in cases with mutations in genes of the RAS-BRAF-MAPK-ERK pathway suggested the activation of this pathway in this subgroup of patients. Importantly, no ERK phosphorylation was observed in unmutated cases. Overall, these results agree with those found in other cancers, in which it has been postulated that the activation of RAS-RAF-MEK-ERK signaling can occur through mutations in several genes in the pathway.40 Our data suggest that mutations in the RAS-BRAFMAPK-ERK pathway are associated with adverse biological features such as U-IGHV, high expression of ZAP-70, CD38 and CD49d, abnormal values of lactate dehydrogenase, and accumulation of three or more driver mutations. Importantly, mutated CLL cases had a 5-year TTFT similar to that of patients with adverse mutations (TP53, ATM or BIRC3), whereas patients carrying both types of mutations simultaneously had the worst 5-year TTFT, as reported by our group and others.7,9,22,33 In our series of patients, the impact of mutations in genes of the RASBRAF-MAPK-ERK pathway on TTFT was independent of that of IGHV status and mutations in TP53, ATM or BIRC3. However, mutations in genes of the RAS-BRAFMAPK-ERK pathway did not affect overall survival. Recently it was reported that BRAF mutations were associated with adverse overall survival, whereas KRAS and NRAS mutations were not.24 Vemurafenib (in 2011) and dabrafenib (in 2013) were the first selective BRAF inhibitors clinically approved for the treatment of melanoma with BRAF mutations.30 MEK inhibitors have also shown efficacy in BRAF-mutant melanoma and in 2014 and 2015 the Food and Drug Administration approved the use of MEK inhibitors in combination with BRAF inhibitors as standard-of-care for BRAF-mutant advanced melanoma.41 With these compounds, clinical response rates of around 50% and increased survival have been reported in BRAF-mutant melanoma42 as well as in cases of HCL refractory to conventional therapy.43,44 However, the majority of responses are transient and resistance is often associated with a plethora of different mechanisms that allow tumor cells to bypass BRAF/MEK inhibition and restore ERK-dependent signaling.45 Our results showed that vemurafenib and dabrafenib were not able to decrease levels of ERK phosphorylation significantly in mutated cases, although a slight effect was observed after dabrafenib treatment which could be an off-target effect. Accordingly, a different spectrum of efficacy against non-V600 BRAF mutants has been described for vemurafenib and dabrafenib.46 In contrast, activation of ERK was detected in unmutated CLL cases, potentially due to ERK activation by the B-cell receptor signaling complex as it has been described that

haematologica | 2019; 104(3)

BRAF inhibitor-related ERK phosphorylation can be partially abrogated by blocking B-cell receptor signaling with SYK inhibitors.47 It has been postulated that cancer cells can dynamically rewire their signaling networks to restore ERK activity and override the actions of inhibitors that act upstream of ERK.48 We, therefore, consider ERK itself as one of the “best” nodes for effective disruption of ERK signaling. Our results demonstrated that ulixertinib (BVD-523), a potent and highly selective inhibitor of ERK1/2, was able to inhibit ERK phosphorylation in vitro in all CLL cases with mutations in genes of the RAS-BRAF-MAPK-ERK pathway. Ulixertinib has shown activity in BRAF- and RASmutant cell lines. Results of phase I studies in solid tumors have documented a safe and well-tolerated effect in patients who harbored BRAF-, NRAS- and MEK-mutant solid tumors, supporting the ongoing development of ulixertinib for patients with MAPK-activating alterations.49 Recently it was reported that CLL cells with trisomy 12 showed increased sensitivity to MEK and ERK inhibitors, pointing to an essential role for MEK/ERK signaling in CLL with trisomy 12.50 In conclusion, we showed that the RAS-BRAF-MAPKERK pathway is one of the cellular processes affected in CLL and identified novel CLL drivers. Patients with mutations in genes of the RAS-BRAF-MAPK-ERK pathway had adverse biological features and most of them required treatment. Furthermore, our results suggest that inhibition of ERK phosphorylation in this subgroup of mutated CLL patients can be achieved using new, specific ERK inhibitors that have recently entered clinical trials. Pharmacological inhibition of the RAS-BRAF-MAPK-ERK pathway may represent a therapeutic approach to improve responses in this subgroup of CLL patients. Acknowledgments This study was supported by the Ministerio de Economía y Competitividad, Grant n. SAF2015-67633-R ,and PI16/00420 which are part of Plan Nacional de I+D+I and are co-financed by the European Regional Development Fund (FEDER-“Una manera de hacer Europa”) and the CERCA program from Generalitat Catalunya. European Union’s Seventh Framework Programme for research, technological development and demonstration under grant agreement n. 306240; Generalitat de Catalunya Suport Grups de Recerca AGAUR 2017-SGR1009, and Departament de Salut (SLT002-16-00350), Instituto de Salud Carlos III (ISCIII) International Cancer Genome Consortium for Chronic Lymphocytic Leukemia (ICGC-CLL Genome Project), and project PM15/00007, which is part of Plan Nacional de I+D+I and are co-financed by FEDER. NG is a recipient of a predoctoral fellowship from Agaur and EC is an Academia Researcher of the "Institució Catalana de Recerca i Estudis Avançats" (ICREA) of the Generalitat de Catalunya. This work was mainly developed at the Centre Esther Koplowitz (CEK), Barcelona, Spain. We are indebted to the Genomics core facility of the Institut d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS) for technical help. We are grateful to N Villahoz and MC Muro for their excellent work in the coordination of the CLL Spanish Consortium and also thank L Jimenez, S Cabezas, and A Giró for their excellent technical assistance. Finally, we are very grateful to all patients with CLL who participated in this study.

585


N. Giménez et al.

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. Fabbri G, Dalla-Favera R. The molecular pathogenesis of chronic lymphocytic leukaemia. Nat Rev Cancer. 2016;16(3):145– 162. 3. Delgado J, Salaverria I, Baumann T, et al. Genomic complexity and IGHV mutational status are key predictors of outcome of chronic lymphocytic leukemia patients with TP53 disruption. Haematologica. 2014;99 (11):e231-234. 4. Zenz T, Eichhorst B, Busch R, et al. TP53 mutation and survival in chronic lymphocytic leukemia. J Clin Oncol. 2010;28(29): 4473–4479. 5. Rossi D, Rasi S, Spina V, et al. Integrated mutational and cytogenetic analysis identifies new prognostic subgroups in chronic lymphocytic leukemia. Blood. 2013;121(8): 1403–1412. 6. Baliakas P, Hadzidimitriou A, Sutton L-A, et al. Recurrent mutations refine prognosis in chronic lymphocytic leukemia. Leukemia. 2015;29(2):329–336. 7. Puente XS, Beà S, Valdés-Mas R, et al. Noncoding recurrent mutations in chronic lymphocytic leukaemia. Nature. 2015;526 (7574):519–524. 8. Quesada V, Conde L, Villamor N, et al. Exome sequencing identifies recurrent mutations of the splicing factor SF3B1 gene in chronic lymphocytic leukemia. Nat Genet. 2012;44(1):47–52. 9. 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. 10. Wang L, Lawrence MS, Wan Y, et al. SF3B1 and other novel cancer genes in chronic lymphocytic leukemia. N Engl J Med. 2011;365(26):2497–2506. 11. Landau DA, Carter SL, Stojanov P, et al. Evolution and impact of subclonal mutations in chronic lymphocytic leukemia. Cell. 2013;152(4):714–726. 12. Jeromin S, Weissmann S, Haferlach C, et al. SF3B1 mutations correlated to cytogenetics and mutations in NOTCH1, FBXW7, MYD88, XPO1 and TP53 in 1160 untreated CLL patient. Leukemia. 2014;28(1):108–117. 13. Nadeu F, Delgado J, Royo C, et al. Clinical impact of clonal and subclonal TP53, SF3B1, BIRC3, NOTCH1, and ATM mutations in chronic lymphocytic leukemia. Blood. 2016;127(17):2122–2130. 14. Lavoie H, Therrien M. Regulation of RAF protein kinases in ERK signalling. Nat Rev Mol Cell Biol. 2015;16(5):281–298. 15. Davies H, Bignell GR, Cox C, et al. Mutations of the BRAF gene in human cancer. Nature. 2002;417(6892):949–954. 16. Forbes SA, Beare D, Boutselakis H, et al. COSMIC: Somatic cancer genetics at highresolution. Nucleic Acids Res. 2017;45(D1): D777–D783. 17. Tiacci E, Trifonov V, Schiavoni G, et al. BRAF mutations in hairy-cell leukemia. N Engl J Med. 2011;364(24):2305–2315. 18. Tiacci E, Pettirossi V, Schiavoni G, Falini B.

586

19.

20.

21. 22.

23.

24.

25.

26.

27.

28.

29. 30.

31.

32.

33.

34.

Genomics of Hairy cell leukemia. J Clin Oncol. 2017;35(9):1002–1010. Pakneshan S, Salajegheh A, Smith RA, Lam AK. Clinicopathological relevance of BRAF mutations in human cancer. Pathology. 2013;45(4):346–356. Buscà R, Pouysségur J, Lenormand P. ERK1 and ERK2 map kinases: specific roles or functional redundancy? Front Cell Dev Biol. 2016;453. Jebaraj BMC, Kienle D, Bühler A, et al. BRAF mutations in chronic lymphocytic leukemia. Leuk Lymphoma. 2013;54(6):1177–1182. Pandzic T, Larsson J, He L, et al. Transposon mutagenesis reveals fludarabine-resistance mechanisms in chronic lymphocytic leukemia. Clin Cancer Res. 2016;22(24): 6217–6227. Damm F, Mylonas E, Cosson A, et al. Acquired initiating mutations in early hematopoietic cells of CLL patients. Cancer Discov. 2014;4(9):1088–1101. Leeksma AC, Taylor J, Wu B, et al. Clonal diversity predicts adverse outcome in chronic lymphocytic leukemia. Leukemia. 2018 Jul 23. [Epub ahead of print] Müller-Hermelink H, Montserrat E, Catovsky D et al. Chronic lymphocytic leukemia/small lymphocytic lymphoma. WHO classification of tumours of haematopoietic and lymphoid tissues. IARC: International Agency for Research on Cancer; Lyon, 2008; 4th edition. Pages 180183. Rossi D, Cerri M, Deambrogi C, et al. The prognostic value of TP53 mutations in chronic lymphocytic leukemia is independent of del17p13: Implications for overall survival and chemorefractoriness. Clin Cancer Res. 2009;15(3):995–1004. Rossi D, Fangazio M, Rasi S, et al. Disruption of BIRC3 associates with fludarabine chemorefractoriness in TP53 wildtype chronic lymphocytic leukemia. Blood. 2012;119(12):2854–2862. Lazarian G, Guièze R, Wu CJ. Clinical Implications of novel genomic discoveries in chronic lymphocytic leukemia. J Clin Oncol. 2017;35(9):984–993. Downward J. Targeting RAS signalling pathways in cancer therapy. Nat Rev Cancer. 2003;3(1):11–22. Imperial R, Toor OM, Hussain A, Subramanian J, Masood A. Comprehensive pancancer genomic analysis reveals (RTK)RAS-RAF-MEK as a key dysregulated pathway in cancer: its clinical implications. Semin Cancer Biol. 2017 Nov 22. [Epub ahead of print]. Wan PTC, Garnett MJ, Roe SM, et al. Mechanism of activation of the RAF-ERK signaling pathway by oncogenic mutations of B-RAF. Cell. 2004;116(6):855–867. Sellar RS, Fend F, Akarca AU, et al. BRAFV600E mutations are found in Richter syndrome and may allow targeted therapy in a subset of patients. Br J Haematol. 2015;170(2):282–285. Nadeu F, Clot G, Delgado J, et al. Clinical impact of the subclonal architecture and mutational complexity in chronic lymphocytic leukemia. Leukemia. 2017;32(3):645653. Waterfall JJ, Arons E, Walker RL, et al. High

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48. 49.

50.

prevalence of MAP2K1 mutations in variant and IGHV4-34-expressing hairy-cell leukemias. Nat Genet. 2014;46(1):8–10. Brown NA, Furtado L V, Betz BL, et al. High prevalence of somatic MAP2K1 mutations in BRAF V600E-negative Langerhans cell histiocytosis. Blood. 2014;124(10):1655– 1658. Schmidt J, Ramis-Zaldivar JE, Nadeu F, et al. Mutations of MAP2K1 are frequent in pediatric-type follicular lymphoma and result in ERK pathway activation. Blood. 2017;130(3):323–327. Yang SH, Sharrocks AD, Whitmarsh AJ. MAP kinase signalling cascades and transcriptional regulation. Gene. 2013;513(1):1– 13. Thomas RK, Baker AC, Debiasi RM, et al. High-throughput oncogene mutation profiling in human cancer. Nat Genet. 2007;39(3):347–351. Warden DW, Ondrejka S, Lin J, Durkin L, Bodo J, Hsi ED. Phospho-ERK(THR202/ Tyr214) is overexpressed in hairy cell leukemia and is a useful diagnostic marker in bone marrow trephine sections. Am J Surg Pathol. 2013;37(2):305–308. Burotto M, Chiou VL, Lee J-M, Kohn EC. The MAPK pathway across different malignancies: a new perspective. Cancer. 2014;120(22):3446–3456. Eroglu Z, Ribas A. Combination therapy with BRAF and MEK inhibitors for melanoma: latest evidence and place in therapy. Ther Adv Med Oncol. 2016;8(1):48–56. Chapman PB, Hauschild A, Robert C, et al. Improved survival with vemurafenib in melanoma with BRAF V600E mutation. N Engl J Med. 2011;364(26):2507–2516. Dietrich S, Glimm H, Andrulis M, et al. BRAF inhibition in refractory hairy-cell leukemia. N Engl J Med. 2012;366(21):2038– 2040. Follows GA, Sims H, Bloxham DM, et al. Rapid response of biallelic BRAF V600E mutated hairy cell leukaemia to low dose vemurafenib. Br J Haematol. 2013;161(1): 150–153. Shi H, Hugo W, Kong X, et al. Acquired resistance and clonal evolution in melanoma during BRAF inhibitor therapy. Cancer Discov. 2014;4(1):80–93. Kordes M, Röring M, Heining C, et al. Cooperation of BRAF(F595L) and mutant HRAS in histiocytic sarcoma provides new insights into oncogenic BRAF signaling. Leukemia. 2016;30(4):937-946. Yaktapour N, Meiss F, Mastroianni J, et al. BRAF inhibitor-associated ERK activation drives development of chronic lymphocytic leukemia. J Clin Invest. 2014;124(11):5074– 5084. Ryan MB, Der CJ, Wang-Gillam A, Cox AD. Targeting RAS-mutant cancers: Is ERK the key? Trends Cancer. 2015;1(3):183–198. Sullivan RJ, Infante JR, Janku F, et al. First-inclass ERK1/2 inhibitor ulixertinib (BVD-523) in patients with MAPK mutant advanced solid tumors: results of a phase I dose-escalation and expansion study. Cancer Discov. 2018;8(2):184-195. Dietrich S, Oleś M, Lu J, et al. Drug-perturbation-based stratification of blood cancer. J Clin Invest. 2018;128(1):427-445.

haematologica | 2019; 104(3)

586


ARTICLE

Hemostasis

Unraveling the effect of silent, intronic and missense mutations on VWF splicing: contribution of next generation sequencing in the study of mRNA Nina Borràs,1,2 Gerard Orriols,1 Javier Batlle,3 Almudena Pérez-Rodríguez,3 Teresa Fidalgo,4 Patricia Martinho,4 María Fernanda López-Fernández,3 Ángela Rodríguez-Trillo,3 Esther Lourés,3 Rafael Parra,1,2 Carme Altisent,2 Ana Rosa Cid,5 Santiago Bonanad,5 Noelia Cabrera,5 Andrés Moret,5 María Eva Mingot-Castellano,6 Nira Navarro,7 Rocío Pérez-Montes,8 Sally Marcellin,9 Ana Moreto,10 Sonia Herrero,11 Inmaculada Soto,12 Núria Fernández-Mosteirín,13 Víctor Jiménez-Yuste,14 Nieves Alonso,15 Aurora de Andrés-Jacob,16 Emilia Fontanes,17 Rosa Campos,18 María José Paloma,19 Nuria Bermejo,20 Ruben Berrueco,21 José Mateo,22 Karmele Arribalzaga,23 Pascual Marco,24 Ángeles Palomo,25 Nerea Castro Quismondo,26 Belén Iñigo,27 María del Mar Nieto,28 Rosa Vidal,29 María Paz Martínez,30 Reyes Aguinaco,31 Jesús María Tenorio,33 María Ferreiro,33 Javier García-Frade,34 Ana María Rodríguez-Huerta,35 Jorge Cuesta,36 Ramón Rodríguez-González,37 Faustino García-Candel,38 Manuela Dobón,39 Carlos Aguilar,40 Francisco Vidal1,2,41 and Irene Corrales1,2

1 Banc de Sang i Teixits, Barcelona, Spain; 2Institut de Recerca Vall d’Hebron Universitat Autònoma de Barcelona (VHIR-UAB), Spain; 3Complexo Hospitalario Universitario A Coruña, INIBIC, Spain; 4Centro Hospitalar e Universitário de Coimbra, Portugal; 5Hospital Universitario y Politécnico La Fe, Valencia, Spain; 6Hospital Regional Universitario de Málaga, Spain; 7Hospital Universitario Dr. Negrín, Las Palmas de Gran Canaria, Spain; 8Hospital Universitario Marqués de Valdecilla, Santander, Spain; 9Salud Castilla y León, Segovia, Spain; 10Hospital Universitario Cruces, Barakaldo, Spain; 11 Hospital Universitario de Guadalajara, Spain; 12Hospital Universitario Central de Asturias, Oviedo, Spain; 13Hospital Universitario Miguel Servet, Zaragoza, Spain; 14 Hospital Universitario La Paz, Madrid, Spain; 15Hospital Infanta Cristina, Badajoz, Spain; 16Complexo Hospitalario Universitario Santiago de Compostela, Spain; 17Hospital Universitario Lucus Augusti, Lugo, Spain; 18Hospital Jerez de la Frontera, Cádiz, Spain; 19 Hospital Virgen del Camino, Pamplona, Spain; 20Hospital San Pedro de Alcántara, Cáceres, Spain; 21Hospital Sant Joan de Deu, Barcelona, Spain; 22Hospital Sta Creu i St Pau, Barcelona, Spain; 23Hospital Universitario Fundación de Alcorcón, Madrid, Spain; 24 Hospital General de Alicante, Spain; 25Hospital Regional Universitario Carlos Haya, Málaga, Spain; 26Hospital Universitario 12 de Octubre, Madrid, Spain; 27Hospital Clínico San Carlos, Madrid, Spain; 28Complejo Hospitalario de Jaén, Spain; 29Fundación Jiménez Díaz, Madrid, Spain; 30Hospital Nuestra Sra. de Sonsoles de Ávila, Spain; 32Hospital Joan XXIII, Tarragona, Spain; 32Hospital Ramón y Cajal, Madrid, Spain; 33Hospital Montecelo, Pontevedra, Spain; 34Hospital Río Hortega, Valladolid, Spain; 35Hospital Gregorio Marañón, Madrid, Spain; 36Hospital Virgen de la Salud, Toledo, Spain; 37Hospital Severo Ochoa, Madrid, Spain; 38Hospital Universitario Virgen Arrixaca, Murcia, Spain; 39Hospital Lozano Blesa, Zaragoza, Spain; 40Hospital Santa Bárbara, Soria, Spain and 41CIBER de Enfermedades Cardiovasculares, Madrid, Spain

ABSTRACT

L

arge studies in von Willebrand disease patients, including Spanish and Portuguese registries, led to the identification of >250 different mutations. It is a challenge to determine the pathogenic effect of potential splice site mutations on VWF mRNA. This study aimed to elucidate the true effects of 18 mutations on VWF mRNA processing, investigate the contribution of next-generation sequencing to in vivo mRNA study in von Willebrand disease, and compare the findings with in silico prediction. RNA extracted from patient platelets and leukocytes was amplified by RT-PCR and sequenced using Sanger and next generation sequencing techniques. Eight mutations affected VWF splicing: c.1533+1G>A, c.5664+2T>C and c.546G>A (p.=) prompted exon skipping; c.3223-7_3236dup and c.7082-2A>G resulted in activation of cryptic sites; c.3379+1G>A and c.7437G>A) demonstrated both molecular pathogenic mechanisms simultaneously; and the p.Cys370Tyr missense mutation generated two aberrant transcripts. Of note, the complete haematologica | 2019; 104(3)

Ferrata Storti Foundation

Haematologica 2019 Volume 104(1):587-598

Correspondence: IRENE CORRALES icorrales@bst.cat/fvidal@bst.cat Received: August 3, 2018. Accepted: October 19, 2018. Pre-published: October 25, 2018. doi:10.3324/haematol.2018.203166 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/104/3/587 ©2019 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

587


N. Borràs et al.

effect of three mutations was provided by next generation sequencing alone because of low expression of the aberrant transcripts. In the remaining 10 mutations, no effect was elucidated in the experiments. However, the differential findings obtained in platelets and leukocytes provided substantial evidence that four of these would have an effect on VWF levels. In this first report using next generation sequencing technology to unravel the effects of VWF mutations on splicing, the technique yielded valuable information. Our data bring to light the importance of studying the effect of synonymous and missense mutations on VWF splicing to improve the current knowledge of the molecular mechanisms behind von Willebrand disease. clinicaltrials.gov identifier:02869074.

Introduction Von Willebrand disease (VWD), the most common congenital bleeding disorder, is caused by a genetic defect in the von Willebrand factor gene (VWF).1 VWF mutational analysis can be valuable for diagnosing and investigating the molecular etiology of VWD, as was seen in the Spanish (PCM-EVW-ES project) and Portuguese cohort of VWD patients.2-4 One interesting challenge in this condition is to elucidate the pathogenic mechanism of VWF mutations. In silico analysis is considered a suitable supporting tool to predict the pathogenicity of the variants identified.5,6 However, functional studies remain essential to unequivocally determine their deleterious effect.7 Functional studies can be performed by analyzing the potential effect of splice site mutations (PSSM) in RNA. In addition to splice site consensus sequence mutations, deep intronic, missense, and synonymous mutations can also disturb splicing. Along this line, 25% of synonymous mutations positioned at exon-intron boundaries result in altered splicing, which, in itself, can cause disease, modify the severity of the disease phenotype, or be linked with disease susceptibility.8 In VWF, heterozygotes for PSSM may be associated with mild forms of VWD type 1 or be phenotypically silent, but when two such mutations are found in different alleles, the phenotype is associated with VWD type 3.9 Almost all related studies of splicing effects have examined peripheral blood platelets, as VWF is exclusively expressed in these cells and endothelial cells.10,11 Platelets are anucleated, but they contain small amounts of translationally active megakaryocytic mRNA.12,13 In contrast, the amount of mRNA obtained from leukocytes is higher and contains mRNA transcripts for genes that are not normally expressed in these cells, known as ‘‘ectopic transcripts’’. Their analysis has been used to investigate mutations in several inherited disorders,14,15 as they facilitate the study of mRNA of those genes expressed in hard to reach tissues.16 PSSM can affect mRNA reorganization and introduce premature termination codons (PTCs) into open reading frames, a common cause of genetic disorders. Most nonsense transcripts are recognized and degraded by nonsense-mediated mRNA decay (NMD),17 a degradation pathway to control synthesis of truncated proteins.18 The efficiency of NMD varies between cell types; hence, the use of RNA from platelets and leukocytes for in vivo study of VWF PSSM offers complementary results, particularly when NMD occurs in the allele carrying the mutation in platelets, as we reported.7 Since its development, next-generation sequencing (NGS) has been increasingly used in molecular genetics to identify mutations causing disease. However, few groups 588

have explored its potential for analyzing splicing variants following RT-PCR.19,20 In this new scenario, the procedure we previously described to analyze the effects of PSSM in VWF7 has been optimized and adapted to an NGS-based technique to investigate its value in this field. Our main objective was to elucidate the true effects of 18 selected mutations (intronic, synonymous, delins, and missense) on mRNA processing and their genotype/phenotype correspondence by analysis of leukocytes and platelets from clinical samples. Finally, the in vivo effects of the mutations were compared with the in silico predictions.

Methods Patients We studied 15 patients diagnosed with different types of VWD, 5 from Complexo Hospitalario Universitario A Coruña, 8 from Hospital Universitari Vall d’Hebron (HUVH), and 2 from Centro Hospitalar e Universitário de Coimbra. Samples from 4 healthy individuals were used as controls. The study was performed according to the guidelines of the Declaration of Helsinki and was approved by the local Research Ethics Committee. All participants provided written informed consent.

Splice site prediction software The predicted impact of potential splice site mutations was analyzed with NetGene221 and the splicing prediction module of Alamut Visual v.2.6.1 software (Interactive Biosoftware, Rouen, France), which integrates data from three methods: Splice Site Prediction by Neural Network (NNSplice), MaxEntScan, and Human Splicing Finder (HSF).

Platelet and leukocyte separation and RNA isolation Leukocyte and platelet RNA from patients and controls was isolated from 10 mL of peripheral blood collected in EDTA tubes, as previously described.7

VWF mRNA amplification After RNA isolation, cDNA was synthesized using the HighCapacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s recommendations. The region including the mutation was amplified by Platinum Taq DNA Polymerase (Thermo Fisher Scientific) in leukocyte and platelet cDNA (Online Supplementary Methods and Table S1). PCR products were separated on 1% agarose gel and visualized by SYBR Safe DNA Gel Stain (Thermo Fisher Scientific).

Sanger sequencing and analysis PCR products were sequenced as previously described.22 However, multiple-band PCR products were previously agarosehaematologica | 2019; 104(3)


Effect of VWF mutations on mRNA splicing

Table 1. Laboratory and molecular data of VWD patients

Patients code

VWD

VWF:Ag

VWF:RCo

FVIII:C

NT change

AA change

UMP01 UMP02

3 carrier 1H

70 36

70.5 49

111 101

UMP03

1

26

21.2

64.8

UMP04 UMP05 UMP06

1 1 2A/2M

50 27 12

52 32 4

47 72 24

UMP07 UMP08

1 1

37 42.4

28 41.7

60 73.4

UMP09‡ UMP10 UMP11

1 3 carrier 1

51 46.4 31

49.8 45 24

71 105 57

c.1533+1G>A c.3379+1G>A * c.-2627C>T * c.5664+2T>C * c.7220T>C * c.7081+6G>T c.7730-56C>T c.4120C>T * c.7730-4C>G * c.8254-5T>G c.546G>A† c.4866C>T† c.3291C>T c.1109G>A c.3223-7_3236dup

UMP12

3

3.7

5

6.7

UMP13

2A

117

6.5

UMP14

1

6

6

UMP15||

3

2

<1

c.449T>C† c.7082-2A>G† 90 c.3426T>C§ c.3485_3486delinsTG§ c.8318G>C† 4,5 c.6699_6702dup† c.7437G>A† 5 c.546G>A§ c.7082-2A>G§ c.8155+3G>C†

p.Leu2407Pro p.Arg1374Cys p.Ser182 p.Asp1622 p.Cys1097 p.Cys370Tyr p.Pro1079_Tyr1080ins LeuGlnValAspProGluPro p.Leu150Pro p.Cys1142 p.Pro1162Leu p.Cys2773Ser p.Cys2235ArgfsTer8 p.Ser2479 p.Ser182 -

Exon 42 28 6 28 25 9 25 5 26 26 52 38 43 6 -

Intron

Domain

13 25 33 41 45 45 51 -

intronic intronic upstream intronic C2 intronic intronic A1 intronic intronic D1 A1-A3 D3 D1 D3

41 41 50

D1 intronic D3 D3 CK D4 C3 D1 intronic intronic

All mutations were identified in heterozygous state. In bold, mutations selected to study their effect on VWF mRNA. The subtype 1H (historical) refers to patients previously diagnosed as type 1 VWD that, at the time of enrollment at the PCM-EVW-ES project, show a slight decrease or even a normal VWF plasma levels. AA:amino acid; NT: nucleotide;VWD: von Willebrand disease. *The allelic phase (cis/trans) of mutations could not be determined since no informative relatives are available. †Mutations in trans. ‡Patient with an additional mutation in the F8 gene. §Mutations in cis. ||Patient previously studied at mRNA level by Corrales et al.7

purified using the MiniElute Gel Extraction kit (Qiagen). The sequences obtained were assembled and aligned against the consensus wild-type (WT) VWF mRNA sequence (GenBank NM_000552) using SeqScape v2.7 software (Thermo Fisher Scientific).

Next-generation sequencing and analysis PCR amplicons obtained per patient were equimolarly mixed in a single tube in a total amount of 250 ng. Subsequently, the libraries were fragmented and the adapter and barcodes were ligated using the NGSgo protocol (GenDX, Utrecht, Netherlands) following the manufacturer’s recommendations. Resulting libraries were combined and sequenced on a MiSeq platform (Illumina, San Diego, CA). After sequencing, barcoded sequences were demultiplexed and analyzed individually. The paired sequence files (fastq format) were used as input for analysis with the CLC Genomic Workbench v.11 software (Qiagen, Aarhus, Denmark) (Online Supplementary Methods and Figure S1-S2).

Results An in-depth study was performed in PCM-EVW-ES,3 the Portuguese cohort,4 and HUVH patients to select previoushaematologica | 2019; 104(3)

ly undescribed VWF mutations and mutations with an unknown or controversial pathogenic mechanism. Eighteen mutations (15 patients) were selected: 8 intronic (4 in canonical, GT and AG, splice site sequences), 5 synonymous, 2 missense, and 3 delins. The patients’ phenotypic and molecular data are summarized in Table 1 and Online Supplementary Table S2. Total mRNA was obtained from platelets and leukocytes of all patients, with the exception of patients UMP08 and UMP14, in whom platelet RNA isolation failed due to blood lysis. All mutations were analyzed by both Sanger sequencing and NGS, and results were compared to the predictions generated by in silico analysis (Online Supplementary Table S3).

Mutations in canonical splice site sequences The c.1533+1G>A mutation (intron 13) was identified in a type 3 VWD carrier (UMP01). The exon 11-15 region was analyzed with a specific primer pair to avoid amplification of a prevalent alternative-splicing product (skipping of exons 14 and 15) in VWF from leukocytes.7 Whereas only the expected PCR product was observed in platelets, leukocyte amplification resulted in 4 PCR bands. Sanger analysis of leukocytes showed 2 mRNA aberrant transcripts: 1) lacking exon 13; and 2) lacking exon 13 and 14 (Online Supplementary Figure S3). By means of NGS, it was 589


N. BorrĂ s et al. A

B

C

D

Figure 1. Analysis of the c.5664+2T>C mutation in patient UMP03. A) RT-PCR products amplified with primers located in exons 30 and 35 in leukocyte (L) and platelet (P) RNA, separated on 1% agarose gel. B) Traditional Sanger sequencing of PCR products from patient leukocytes (L2) and platelets (P), and control leukocytes (L). Analysis of the band 2 from patient leukocyte on agarose gel demonstrates exon 33 skipping. In platelets, only the allele without c.5664+2T>C mutation could be amplified. C) NGS of PCR products from leukocytes showed exon 33 skipping, indicated by arrows depict aberrant transcripts. However, exon 33-34 skipping was also detected, but in a really low of transcripts. D) Schematic representation of the mutation in genomic DNA and its effect on the VWF mRNA sequence. M indicates a 100-bp DNA ladder.

determined that 19% were transcripts without exon 13, 45% transcripts without exons 13 and 14, and an additional aberrant transcript without exon 14 was detected in 8% transcripts (Online Supplementary Table S4). By both techniques, no effect was visible in platelets due to NMD. The c.3379+1G>A mutation (intron 25) was identified in a type 1H VWD patient (UMP02). We designed two new primers to analyze the exon 22-26 region. Amplification of leukocyte cDNA yielded two bands, one with the expected size and a smaller band, whereas in platelets only the expected PCR product was observed (Online Supplementary Figure S4). Sanger analysis of leukocytes showed that c.3379+1G>A caused exon 25 skipping, leading to a frameshift at position 1075 and adding 88 aberrant amino acids before a PTC was encountered (p.Pro1075ValfsTer88). On NGS, however, two aberrant 590

transcripts were detected: the major one, which lacked exon 25, was detected in 43% of all reads, and the minor one, which resulted from activation of a cryptic donor splice site (DSS) 31 nucleotides upstream from the WTDSS, was found in 1.4% of reads (p.Pro1117ValfsTer88) (Table 2 and Online Supplementary Table S4). In platelets, however, the mutated allele had undergone NMD, which precluded observation of aberrant transcripts. The c.5664+2T>C mutation was identified in a type 1 VWD patient (UMP03), in trans with the p.Leu2407Pro mutation. To analyze the PSSM in intron 33, the exon 3035 region was amplified. The PCR product had the expected size in platelets, whereas an additional smaller one was observed in leukocytes (Figure 1). Sanger sequencing of the leukocyte PCR product confirmed exon 33 skipping (p.Gly1874AlafsTer32) (Table 2). NGS detected 2 aberrant haematologica | 2019; 104(3)


Effect of VWF mutations on mRNA splicing

A

B

C

D

Figure 2. Analysis of the c.546G>A mutation, located at nucleotide 14 from the beginning of exon 6, in patient UMP08. A) RT-PCR products amplified with primers located in exons 4 and 7 in leukocyte RNA (L), separated on 1% agarose gel. B) Traditional Sanger sequencing of PCR product from patient L (1) showed the single nucleotide change. C) NGS of PCR products from leukocytes identified the single nucleotide variant, as well as exon 6 skipping. Arrows show aberrant transcripts. D) Schematic representation of the mutation in genomic DNA and its effect on the VWF mRNA sequence. M indicates a 100-bp DNA ladder.

transcripts: the major one (37% of reads) showed exon 33 skipping, and the minor (2% of reads) showed exons 33 and 34 skipping. Moreover, analysis of the p.Leu2407Pro mutation confirmed that NMD of the allele carrying the c.5664+2T>C mutation had occurred in platelets.

Candidate intronic mutations The c.7081+6G>T mutation, identified in a type 1 VWD patient (UMP04), generated a new GT dinucleotide in intron 41 (Online Supplementary Table S3). The exon 38-43 region showed the expected PCR product in both cell types (772 bp). To confirm these results, two informative SNPs (rs216321 in exon 20 and rs216902 in exon 35) were genotyped. Both were in heterozygous state, indicating that the allele carrying the mutation was unaffected by NMD. The c.7730-56C>T mutation (intron 45) was identified in a type 1 VWD patient (UMP05). The exon 43-49 region, examined by Sanger and NGS, showed no visible effect on mRNA splicing. To confirm the presence of the allele carrying the intronic mutation, two informative SNPs were haematologica | 2019; 104(3)

analyzed: rs216321 and rs1800380. Surprisingly, the sequence obtained in both cell types indicated that only one allele was expressed. To further explore these results, we tested whether intron 45 was retained within the mature mRNA (Online Supplementary Methods) and performed complete sequencing of VWF cDNA from leukocytes. Nonetheless, no changes were observed (data not shown). The c.7730-4C>G mutation (intron 45) was identified in a type 2A/2M patient (UMP06), combined with the p.Arg1374Cys mutation. The HSF predicted activation of a cryptic intronic acceptor splice site (ASS), but in the exon 43-49 region, splicing was not affected by the mutation. To confirm the presence of the allele carrying c.77304C>G, the p.Arg1374Cys mutation was analyzed, and both nucleotides were seen in platelets and leukocytes, suggesting expression of both alleles. We then performed the same experiment as was done in patient UMP05 to test intron 45 retention, but no changes in VWF cDNA were observed (data not shown). 591


N. BorrĂ s et al.

A

B

C

D

Figure 3. Analysis of the c.1109G>A (p.Cys370Tyr) mutation in patient UMP10, located in the last nucleotide of exon 9. A) RT-PCR products amplified with primers located in exons 6 and 12 in leukocyte (L) and platelet (P) RNA, separated on 1% agarose gel. B) Traditional Sanger sequencing of PCR product from patient L (2) showed exon 9 skipping. The L (3) band could not be purified in the agarose gel due to low concentration, thus it was not analyzed by Sanger. In platelets, only the non-mutated allele could be amplified. C) NGS of PCR products from leukocytes identified transcripts lacking exon 9, as well as transcripts lacking exons 8 and 9. D) Schematic representation of the mutation in genomic DNA and its effect on the VWF mRNA sequence. M indicates a 100-bp DNA ladder.

The c.8254-5T>G mutation (intron 51) was found in a type 1 VWD patient (UMP07). The exon 49-52 region was amplified, and the expected 546-bp size was observed in both cell types. Sequence analysis by the two techniques revealed no changes in VWF cDNA. Interestingly, this patientâ&#x20AC;&#x2122;s two children, heterozygous for the c.8254-5T>G mutation, had bleeding symptoms. Hence, we performed complete sequencing of VWF cDNA from leukocytes. However, no changes were identified.

Synonymous mutations The mutations c.546G>A (exon 6) and c.4866C>T (exon 28) were identified in trans in a type 1 VWD patient (UMP08). These mutations were studied only in leukocyte RNA. To investigate c.546G>A, the exon 4-7 region was amplified, which resulted in the expected band and a 592

slightly diffuse, smaller band. Sanger sequencing detected the c.546G>A change (Figure 2). However, NGS not only detected the nucleotide change, but also recognized a loss of 125 nucleotides corresponding to exon 6 (p.Thr179ProfsTer31) in small 2.3% of reads (Online Supplementary Table S4). The predicted impact of the c.546G>A mutation was discrepant and only two in silico algorithms predicted an effect on mRNA splicing (Online Supplementary Table S3) that was confirmed in vivo even though with a slight effect. To study c.4866C>T, the exon 25-31 region was amplified and sequenced, but no effect on mRNA processing was seen, as was predicted by in silico tools. Both nucleotides were identified in heterozygous state, indicating the presence of both alleles. Only offspring carrying the c.546G>A mutation (1 of 3 children) had lower levels of VWF:Ag and VWF:RCo compared to haematologica | 2019; 104(3)


Effect of VWF mutations on mRNA splicing

Table 2. Effect of VWF mutations on mRNA from leukocytes and platelets. Mutation type NT change Intronic

AA change

Exon

Intron

-

-

13

Exon 13 skipping (r.1433_1533del)

NMD

p.Gly478AlafsTer138

25

Exon 25 skipping (r.3223_3379del)

NMD

p.Pro1075ValfsTer88

Activation of a cryptic site at +126 nt

NMD

p.Pro1117ValfsTer88

-

c.1533+1G>A c.3379+1G>A*

-

c.5664+2T>C

-

c.7081+6G>T

-

c.7082-2A>G †

-

c.7730-4C>G

-

c.7730-56C>T

-

-

-

-

-

-

Synonymous c.546G>A †

p.Ser182

6

c.3291C>T

p.Cys1097

25

c.3426T>C

p.Cys1142

26

c.4866C>T

p.Asp1622

28

c.7437G>A

p.Ser2479

43

c.8254-5T>G

Missense

33

No visible effect

-

NMD

p.Ala2361GlyfsTer40 -

in exon 42 (r.7082_7088del) 45

No visible effect

No visible effect

45

No visible effect

No visible effect

51

No visible effect

No visible effect

-

NA

p.Thr179ProfsTer31

-

-

-

-

-

-

9

-

p.Pro1162Leu

26

-

25

-

38

p.Phe1875_Cys1948del

Activation of a cryptic site +7 nt

5

p.Cys2235ArgfsTer8

p.Gly1874AlafsTer32

NMD

41

p.Cys370Tyr

c.6699_6702dup

NMD

Exon 33+34 skipping (r.5621_5842del) ‡ No visible effect

p.Leu150Pro

GlnValAspProGluPro

Protein prediction

Exon 33 skipping (r.5621_5664del)

41

c.449T>C

c.3223-7_3236dup p.Pro1079_Tyr1080insLeu

Platelet effect

in exon 25 (r.3349_3379del) ‡

c.1109G>A DelIns c.3485_3486delinsTG

Leukocyte effect

-

Exon 6 skipping (r.533_657del) ‡

-

No visible effect

No visible effect

No visible effect

No visible effect

No visible effect

NA

-

Exon 43 skipping (r.7288_7437)

NA

p.Val2430GlyfsTer335

Activation of a cryptic site at +146 nt

NA

p.Ser2479AlafsTer23

in exon 43 (r.7434_7437del) No visible effect Exon 9 skipping (r.998_1109del) Exon 8+9 skipping (r.875_1109del)

No visible effect

-

NMD

p.Glu333AlafsTer87

NMD

p.Ser292ThrfsTer87

No visible effect

-

(r.3477_3478instttgcaggt

p.Pro1079_Tyr1080insLeu

ggaccccgagcc)

GlnValAspProGluPro

NMD

p.Cys2235ArgfsTer8

No visible effect (r.3477_3478instttgcaggtggaccccgagcc)

-

No visible effect

AA:amino acid; NA: not available; NMD: nonsense-mediated decay; NT:nucleotide. *Previous functional studies by Nesbitt et al.23 †Previous functional studies by Corrales et al.7 ‡Leukocyte effect observed in a really low % of reads by NGS.

the others. This may be explained by the mutation and/or other factors not explored in the current study. The c.3291C>T mutation (exon 25), was identified in patient UMP09 (female), as well as a novel mutation in F8 (c.1346C>G, p.Ala430Gly). Of note, her brother, who was not included in this study, also had these mutations. As both siblings experienced bleeding, and the HSF prediction regarding c.3291C>T interpreted creation of an exonic splicing silencer site that could potentially alter splicing, we hypothesized that this synonymous mutation could have an effect on VWF mRNA processing. Hence, the exon 22-26 region was amplified, but no change in the VWF cDNA sequence was observed. In addition, we tested whether intron 25 was retained within mature VWF mRNA (Online Supplementary Methods), but no differences compared to control leukocyte cDNA were found (data not shown). Finally, as both siblings presented similar FVIII:C levels, we investigated X chromosome inactivation in patient UMP09. The results demonstrated skewed inactivation of the WT X chromosome in this patient (Online Supplementary Figure S5).

Missense mutations The mutation c.1109G>A (p.Cys370Tyr, exon 9) was identified in a type 3 VWD carrier (UMP10). The exon 6haematologica | 2019; 104(3)

12 region was investigated. Whereas the expected size was observed in platelets, amplification from leukocytes resulted in 2 additional bands of ̴650 and ̴500 bp (Figure 3). Sanger sequencing of leukocyte mRNA confirmed that the 650-bp band corresponded to a transcript lacking exon 9. NGS of leukocyte mRNA revealed 3 different transcripts: the WT, a transcript skipping exon 9 (p.Glu333AlafsTer87) (13% of reads), and a transcript skipping exons 8 and 9 (p.Ser292ThrfsTer87) (7% of reads)(Table 2). Sequencing of PCR products from platelets showed no changes, suggesting that the mutated allele had experienced NMD.

Duplication mutation The c.3223-7_3236dup mutation, which generated an in-frame tandem duplication of 21 nucleotides, including 7 nucleotides from intron 24 and the first 14 from exon 25, was detected in a type 1 VWD patient (UMP11). Study of this mutation by the two techniques demonstrated a change in the native ASS of intron 24 in both cell types, which resulted in maintaining the 21-bp insert corresponding to 7 aberrant amino acids in mature VWF mRNA (p.Pro1079_Tyr1080insLeuGlnValAspProGluPro) (Table 2 and Online Supplementary Figure S6). Furthermore, NGS showed that the aberrant mRNA transcript was 593


N. BorrĂ s et al. present in Ě´ 14% of reads in both cell types (Online Supplementary Table S4).

Combined potential splice site mutations in individual patients The c.7082-2A>G mutation (intron 41) was identified in a type 3 VWD patient (UMP12) combined in trans with the p.Leu150Pro (c.449C>T, exon 5) mutation. To study c.7082-2A>G, the exon 38-43 region was amplified in platelets and leukocytes, which yielded a band of the expected size. However, Sanger and NGS sequencing of the PCR products revealed activation of a cryptic ASS

A

B

within exon 42. This led to deletion of the 7 initial nucleotides of exon 42 (r.7082_7088del), as was predicted by the 4 algorithms, and resulted in a frameshift and a PTC (p.Ala2361GlyfsTer40) (Figure 4). Additionally, NGS showed that the aberrant transcript was present in 24% of reads. On sequencing of platelet amplicons, there were no changes, a finding highly suggestive of NMD. As type 3 VWD is characterized by absent or non-functional expression of both VWF alleles, we postulated that the p.Leu150Pro missense mutation could have an effect on VWF mRNA. Nonetheless, no sequence change was identified. These results were confirmed by analysis of 2

Figure 4. Analysis of the c.7082-2A>G mutation in patient UMP12. Agarose gel electrophoresis results of RTPCR amplification of exon 38 to 43 using RNA from leukocytes and platelets were the same as those of healthy controls (data not shown). A) Traditional Sanger sequencing of PCR product from patient leukocyte (L) demonstrate activation of a cryptic splice site 7 nucleotides downstream of the native splice site within exon 42. In patient platelet (P), only the allele carrying the p.Leu150Pro mutation could be amplified. B) NGS of PCR products from leukocytes showed deletion of the 7 initial nucleotides of exon 42. Arrows show aberrant transcripts. C) Schematic representation of the mutation in genomic DNA and its effect on the VWF mRNA sequence. WO: without.

C

594

haematologica | 2019; 104(3)


Effect of VWF mutations on mRNA splicing

informative SNPs, rs1800375 and rs1800376, which were found in heterozygous state in leukocyte transcripts and homozygous state in platelets. The c.[3426T>C; 3485_3486delinsTG] mutations (exon 26) were found in a type 2A VWD patient (UMP13), combined in trans with p.Cys2773Ser. In silico analysis predicted no impact on the splice site for c. 3485_3486delinsTG, but 2 of the 4 algorithms predicted that c.3426T>C would have an effect on splicing. The exon 25-28 region was investigated, but no effect was found on mRNA processing. Study of the p.Cys2773Ser mutation and 2 informative SNPs (rs2228317 and rs4021576) confirmed that the 2 alleles were present, suggesting that exon 26 mutations do not have an effect on VWF splicing. The mutations (c.7437G>A exon 43), and c.6699_6702dup (p.Cys2235ArgfsTer8, exon 38) were identified in trans in a severe type 1 VWD patient (UMP14) and studied only in leukocyte RNA. For c.7437G>A, we amplified exons 41-45, and 2 bands emerged: one with the expected size, and a smaller band of Ě´ 350 bp (Figure 5). Sequencing analysis by both tech-

A

niques revealed 2 aberrant transcripts: one lacked exon 43 (p.Val2430GlyfsTer335) and the other showed activation of a cryptic DSS within exon 43, leading to deletion of its 4 final nucleotides (p.Ser2479AlafsTer23) (Table 2). Study of c.6699_6702dup in leukocyte mRNA showed no splicing changes, but the duplication introduced 4 nucleotides that changed the reading frame and generated a PTC (p.Cys2235ArgfsTer8). The c.546G>A (p.=, exon 6) and c.7082-2A>G (intron 41) mutations in trans with c.8155+3G>C (intron 50) were detected in a type 3 VWD patient (UMP15). This patient had been described in our previous article, but at that time we were only able to identify the effect of c.8155+3G>C (p.Gly2706ValfsTer24) on splicing. On reanalysis of this case with NGS, we were able to characterize the effect of c.546G>A (p.Thr179ProfsTer31) and c.7082-2A>G (p.Ala2361GlyfsTer40) in cis in leukocyte mRNA. The results are consistent with those observed in patients UMP08 and UMP12, who harbored each mutation individually. In addition, in leukocytes, transcripts resulting from c.546G>A were detected in 2% of reads, transcripts

B

C

D

Figure 5. Analysis of the c.7437G>A (p.=) mutation in patient UMP14, located in the last nucleotide of exon 43. A) RT-PCR products amplified with primers located in exons 41 and 45 in leukocyte RNA (L), separated on 1% agarose gel. B) Traditional Sanger sequencing of PCR product from patient leukocyte (L) shows two aberrant transcripts: activation of a cryptic splice site - 4 nucleotides upstream to WT-DSS - in exon 43 (1), and exon 43 skipping (2). C) NGS of PCR products gave the same results than Sanger sequencing. D) Schematic representation of the mutation in genomic DNA and its effect on the VWF mRNA sequence. M indicates a 100bp DNA ladder. WO: without.

haematologica | 2019; 104(3)

595


N. BorrĂ s et al.

from c.7082-2A>G in 6% and transcripts from c.8155+3G>C in 95% of reads. In platelet analyses, NGS detected transcripts without exon 6 resulting from the c.546G>A mutation in 11% of the total reads but no aberrant transcript derived from the mutation c.7082-2A>G indicating that NMD rate could differ between 5' and 3' regions of mRNA. These results indicate the c.[546G>A; 7082-2A>G] allele is under-represented compared to the allele carrying the c.8155+3G>C mutation, showing that this allele underwent NMD in both cell types, as was previously hypothesized.7

Discussion This study reports the results of in-depth analysis of 18 PSSM in samples from VWD patients. The novel and robust procedure used, combining two-step RT-PCR and NGS sequencing, was faster and more sensitive than the method used in our previous article.7 This new approach provides several advantages, such as allele-specific individual sequences, higher sensitivity to detect transcript variants present at low copy numbers, and simplified sample preparation. Moreover, the NGS data on total reads obtained for each transcript is additional information that can provide an approximation of the expression levels of each VWF mRNA. Studies are currently ongoing to confirm that the relative expression of each transcript obtained by NGS is comparable to that provided by realtime RT-PCR. The proven usefulness of leukocyte analysis to interpret mutations masked by NMD in platelets was seen in relation to the c.3379+1G>A mutation, which had been previously investigated by RT-PCR in platelets.23 In that study, the true pathogenic effect of this mutation could not be determined due to NMD. In the present study, using leukocyte mRNA and NGS technology, we found that the mutation induces two aberrant transcripts. Similarly, the pathogenic mechanism of 7 PSSM (c.1533+1G>A, c.5664+2T>C, c.7082-2A>G, c.546G>A, c.7437G>A, and p.Cys370Tyr) was determined in leukocytes, and only the c.3223-7_3236dup mutation effect was observed in both cell types. Because this mutation is located in the D3 domain implicated in multimerization, it is possible that the pathogenic mechanism leading to type 1 VWD may be related to impaired VWF secretion due to intracellular retention, as has been described for other D3 mutations such as p.Cys1130Phe.24 To confirm this hypothesis, expression studies by means of heterologous cell lines or blood outgrowth endothelial cells (BOECs) will remain the gold standard. As would be expected, mutations located in the consensus splicing sequence affected mRNA processing. These included c.1533+1G>A, c.5664+2T>C, leading to exon skipping, and c.7082-2A>G, activating a cryptic splice site. Both these molecular effects have been reported previously in splicing mutations causing VWD.7,25,26 The c.70822A>G in trans with p.Leu150Pro leads to the development of type 3 VWD. The p.Leu150Pro mutation does not affect splicing. However, other mutations located in the propeptide, such as p.Asp141Tyr, have been identified in type 3 patients,27 and, as has been described in these cases, we suspect that p.Leu150Pro may compromise propeptide folding and affect intracellular survival and the capacity to mediate multimerization. The c.1533+1G>A mutation in 596

leukocytes generated three aberrant transcripts. One of them, which leads to the inframe deletion of exons 13-14, should have been detected in platelets since it would not generate a PTC and would not be degraded. However, because the mutation is located in a region with weak splicing signal sequences, as previously demonstrated,7,25,26 we suggest that the results obtained in leukocytes could be an artifact that would not occur in the natural cellular type of VWF expression. Based on these observations, we propose that the real effect of c.1533+1G>A is the skipping of exon 13, which predicts a frameshift at position 478 and addition of 138 aberrant amino acids before a PTC is encountered (p.Gly478AlafsTer138). Two synonymous mutations included in our study, c.546G>A and c.7437G>A, affect splicing of VWF mRNA. To our knowledge, only three synonymous mutations affecting VWF mRNA processing have been reported: c.7056C>T,28 c.7464C>T,29 and c.3390C>T.26 Of note, the results of the present study have almost doubled the number of reported mutations of this type causing VWD. The pathogenic effect of c.546G>A could only be documented by NGS; it eluded Sanger detection because of the low expression in leukocytes and platelets. The c.7437G>A mutation located at the last nucleotide of exon 43 was found in a patient with severe type 1 VWD (UMP14) in trans with c.6699_6702dup (p.Cys2235ArgfsTer8). This synonymous mutation produced 2 splicing variants. First, creation of a new DSS 4 nucleotides upstream of the WTDSS in exon 43 that generated a PTC in exon 44 and would lead to NMD in platelets; and second, exon 43 skipping and generation of a PTC in exon 52. In this latter case, we would expect that NMD had been abolished in platelets, since this cellular mechanism is not effective when a PTC is encountered within 50 bp of the last exonexon junction.30 Study of the p.Cys370Tyr (c.1109G>A in exon 9) mutation, resulted in generation of 2 VWF transcripts, one of them identified only by NGS because of their low expression. Of note, only one other missense mutation, p.Gly1108Arg, has been reported to affect VWF splicing. Thus, these two mutations support the concept that missense mutations on the last exonic nucleotides can also have repercussions on the splicing process.31 Interestingly, the pathogenic effect of this mutation is the same as that described for c.1109+2T>C (intron 9).32 In addition to the patient reported here, the p.Cys370Tyr mutation was identified in 6 additional related patients included in the PCM-EVW-ES: in homozygous state in a type 3 VWD patient and in heterozygosis in 4 type 3 carriers and 1 patient with type 1 VWD, based on their phenotype levels. Certain genetic modifiers of VWF levels33 and interindividual variability in NMD efficiency between patients carrying identical mutations may lead to differences in the disease severity and clinical phenotype.34 Lastly, study of c.7081+6G>T, c.7730-4C>G, c.773056C>T, c.8254-5T>G, c.3291C>T, c.4866C>T and c.[3426T>C; 3485_3486delinsTG] showed no visible effect on mRNA processing, although this does not necessarily mean that they have no effect on splicing. For instance, mRNA SNP analysis in patient UMP05 with c.7730-56C>T showed an absence of 1 allele in leukocytes and platelets, suggesting that the allele may have experienced NMD in both cell types or lack expression due to a mutation that was not detected by our sequencing protocol. Of particular note, deep intronic mutations such as haematologica | 2019; 104(3)


Effect of VWF mutations on mRNA splicing

c.6599-20A>T have been found to cause VWD.35 In patient UMP07 carrying the c.8254-5T>G mutation, no informative SNP was identified and we were unable to determine whether there was a lack of expression of 1 allele, which could explain cosegregation of the mutation in the family with bleeding symptoms. Study of c.[3426T>C; 3485_3486delinsTG] and c.77304C>G showed no visible effect on splicing. However, these mutations have been identified in homozygous state in a type 3 patient (c.[3426T>C; 3485_3486delinsTG])3 and in a severe type 1 VWD patient (c.7730-4C>G)4 with VWF:Ag at 7% and VWF:RCo at 5%, combined with the heterozygous p.Ala631Val (previously reported in a healthy control).36 Based on these findings, there is substantial evidence that these mutations would have an effect on VWF levels. Therefore, to unequivocally determine the potential deleterious effect of these variants, functional studies remain essential. These studies are traditionally carried out by in vitro analyses performed using heterologous cell lines (COS7, AtT-20 and HEK293). However, the advent of the possibility of obtaining BOECs from patients represents a valuable alternative, since this is the functional expression site of VWF.37 Moreover, BOECs allows protein expression and mRNA studies simultaneously. In silico algorithms used to assess the impact of mutations on splicing are more sensitive and accurate in determining the putative effect of intronic mutations than that of syn-

References 1. Rodeghiero F, Castaman G, Dini E. Epidemiological investigation of the prevalence of von Willebrand's disease. Blood. 1987;69(2):454-459. 2. Batlle J, Perez-Rodriguez A, Corrales I, et al. Molecular and clinical profile of von Willebrand disease in Spain (PCM-EVWES): Proposal for a new diagnostic paradigm. Thromb Haemost. 2016;115(1):4050. 3. Borras N, Batlle J, Perez-Rodriguez A, et al. Molecular and clinical profile of von Willebrand disease in Spain (PCM-EVWES): comprehensive genetic analysis by next-generation sequencing of 480 patients. Haematologica. 2017;102(12): 2005-2014. 4. Fidalgo T, Salvado R, Corrales I, et al. Genotype-phenotype correlation in a cohort of Portuguese patients comprising the entire spectrum of VWD types: impact of NGS. Thromb Haemost. 2016;116(1):17-31. 5. Richards S, Aziz N, Bale S, et al. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genet Med. 2015;17(5):405-424. 6. Wang QY, Song J, Gibbs RA, Boerwinkle E, Dong JF, Yu FL. Characterizing polymorphisms and allelic diversity of von Willebrand factor gene in the 1000 Genomes. J Thromb Haemost. 2013;11(2):261-269.

haematologica | 2019; 104(3)

onymous or missense mutations, such as c.546G>A and p.Cys370Tyr. Therefore, the results here should not be considered definitive, and as with all analytical approaches, should form one aspect of a wider investigation.5 In conclusion, we present an extensive study reporting the effect of 18 candidate mutations on VWF mRNA processing. In vivo mRNA studies incorporating NGS technology together with traditional sequencing enabled us to determine the pathogenic effect of 8 PSSM (44%). Our study emphasizes the importance of examining selected mutations, including synonymous and missense mutations, to determine their pathogenic role in splicing. Taken together, our results add to the current knowledge about the molecular events leading to VWD. Acknowledgments We are indebted to Baxalta US Inc., now a part of Shire, for support of the PCM-EVW-ES (Grant H13-000845). Funding This study was also supported by the Spanish Ministry of the Economy and Competitiveness (MINECO, Ministerio de Economiá y Competitividad), Instituto de Salud Carlos III (ISCIII) (PI12/01494, PI15/01643 and RD12/0042/0053). We are very grateful for the kind collaboration of the participating patients and their families. CIBERCV is an initiative of ISCIII, co-financed by the European Regional Development Fund (ERDF), “A way to build Europe”.

7. Corrales I, Ramirez L, Altisent C, Parra R, Vidal F. The study of the effect of splicing mutations in von Willebrand factor using RNA isolated from patients' platelets and leukocytes. J Thromb Haemost. 2011;9(4):679-688. 8. Wang GS, Cooper TA. Splicing in disease: disruption of the splicing code and the decoding machinery. Nat Re Genet. 2007;8(10):749-761. 9. Eikenboom JC, Reitsma PH, Peerlinck KM, Briët E. Recessive inheritance of von Willebrand's disease type I. Lancet. 1993;341(8851):982-986. 10. Jaffe EA, Hoyer LW, Nachman RL. Synthesis of von Willebrand factor by cultured human endothelial cells. Proc Natl Acad Sci U S A. 1974;71(5):1906-1909. 11. Sporn LA, Chavin SI, Marder VJ, Wagner DD. Biosynthesis of von Willebrand protein by human megakaryocytes. J Clin Invest. 1985;76(3):1102-1106. 12. Kieffer N, Guichard J, Farcet JP, Vainchenker W, Breton-Gorius J. Biosynthesis of major platelet proteins in human blood platelets. Eur J Biochem. 1987;164(1):189-195. 13. Fink L, Holschermann H, Kwapiszewska G, et al. Characterization of platelet-specific mRNA by real-time PCR after laserassisted microdissection. Thromb Haemost. 2003;90(4):749-756. 14. McVey JH, Boswell EJ, Takamiya O, et al. Exclusion of the first EGF domain of factor VII by a splice site mutation causes lethal factor VII deficiency. Blood. 1998;92(3):920-926. 15. David D, Santos IM, Johnson K,

16.

17. 18.

19.

20.

21.

22.

Tuddenham EG, McVey JH. Analysis of the consequences of premature termination codons within factor VIII coding sequences. J Thromb Haemost. 2003;1(1): 139-146. Roberts RG, Bentley DR, Bobrow M. Infidelity in the structure of ectopic transcripts: a novel exon in lymphocyte dystrophin transcripts. Hum Mutat. 1993;2 (4):293-299. Byers PH. Killing the messenger: new insights into nonsense-mediated mRNA decay. J Clin Invest. 2002;109(1):3-6. Lewis BP, Green RE, Brenner SE. Evidence for the widespread coupling of alternative splicing and nonsense-mediated mRNA decay in humans. Proc Natl Acad Sci U S A. 2003;100(1):189-192. Martorell L, Luce E, Vazquez JL, et al. Advanced cell-based modeling of the royal disease: characterization of the mutated F9 mRNA. J Thromb Haemost. 2017;15(11): 2188-2197. Hojny J, Zemankova P, Lhota F, et al. Multiplex PCR and NGS-based identification of mRNA splicing variants: Analysis of BRCA1 splicing pattern as a model. Gene. 2017;637:41-49. Brunak S, Engelbrecht J, Knudsen S. Prediction of human mRNA donor and acceptor sites from the DNA sequence. J Mol Biol. 1991;220(1):49-65. Corrales I, Ramirez L, Altisent C, Parra R, Vidal F. Rapid molecular diagnosis of von Willebrand disease by direct sequencing. Detection of 12 novel putative mutations in VWF gene. Thromb Haemost. 2009;101(3):570-576.

597


N. BorrĂ s et al.

23. Nesbitt IM, Hampton KK, Preston FE, Peake IR, Goodeve AC. A common splice site mutation is shared by two families with different type 2N von Willebrand disease mutations. Thromb Haemost. 1999;82(3):1061-1064. 24. Tjernberg P, Vos HL, Castaman G, Bertina RM, Eikenboom JC. Dimerization and multimerization defects of von Willebrand factor due to mutated cysteine residues. J Thromb Haemost. 2004;2(2):257-265. 25. Gallinaro L, Sartorello F, Pontara E, et al. Combined partial exon skipping and cryptic splice site activation as a new molecular mechanism for recessive type 1 von Willebrand disease. Thromb Haemost. 2006;96(6):711-716. 26. Pagliari MT, Baronciani L, Garcia Oya I, et al. A synonymous (c.3390C>T) or a splicesite (c.3380-2A>G) mutation causes exon 26 skipping in four patients with von Willebrand disease (2A/IIE). J Thromb Haemost. 2013;11(7):1251-1259. 27. Baronciani L, Federici AB, Cozzi G, et al. Expression studies of missense mutations p.D141Y, p.C275S located in the propeptide of von Willebrand factor in patients

598

28.

29.

30.

31.

32.

with type 3 von Willebrand disease. Haemophilia. 2008;14(3):549-555. Daidone V, Gallinaro L, Grazia Cattini M, et al. An apparently silent nucleotide substitution (c.7056C>T) in the von Willebrand factor gene is responsible for type 1 von Willebrand disease. Haematologica. 2011;96(6):881-887. Yadegari H, Biswas A, Akhter MS, et al. Intron retention resulting from a silent mutation in the VWF gene that structurally influences the 5' splice site. Blood. 2016;128(17):2144-2152. Maquat LE. Nonsense-mediated mRNA decay: splicing, translation and mRNP dynamics. Nat Rev Mol Cell Biol. 2004;5(2): 89-99. James PD, O'Brien LA, Hegadorn CA, et al. A novel type 2A von Willebrand factor mutation located at the last nucleotide of exon 26 (3538G>A) causes skipping of 2 nonadjacent exons. Blood. 2004;104(9): 2739-2745. Castaman G, Plate M, Giacomelli SH, Rodeghiero F, Duga S. Alterations of mRNA processing and stability as a pathogenic mechanism in von Willebrand factor

33.

34. 35.

36.

37.

quantitative deficiencies. J Thromb Haemost. 2010;8(12):2736-2742. James PD, Lillicrap D. von Willebrand disease: clinical and laboratory lessons learned from the large von Willebrand disease studies. Am J Hematol. 2012;87 Suppl 1:S4-11. Miller JN, Pearce DA. Nonsense-mediated decay in genetic disease: friend or foe? Muta Res Rev Mutat Res. 2014;762:52-64. Hawke L, Bowman ML, Poon MC, Scully MF, Rivard GE, James PD. Characterization of aberrant splicing of von Willebrand factor in von Willebrand disease: an underrecognized mechanism. Blood. 2016;128(4): 584-593. Bellissimo DB, Christopherson PA, Flood VH, et al. VWF mutations and new sequence variations identified in healthy controls are more frequent in the AfricanAmerican population. Blood. 2012;119(9): 2135-2140. Wang JW, Bouwens EA, 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.

haematologica | 2019; 104(3)


ARTICLE

Coagulation & its Disorders

Factor VIII cross-matches to the human proteome reduce the predicted inhibitor risk in missense mutation hemophilia A

Ferrata Storti Foundation

Daniel P. Hart,1,2 Nazmiye Uzun,3 Stuart Skelton,1,3 Alison Kakoschke,3 Jacob Househam,3 David S. Moss3 and Adrian J. Shepherd3

Blizard Institute, Barts and The London School of Medicine and Dentistry, QMUL, London; 2The Royal London Hospital Haemophilia Centre, Barts Health NHS Trust, London; 3Department of Biological Sciences and Institute of Structural and Molecular Biology, Birkbeck, University of London, UK 1

Haematologica 2019 Volume 104(3):599-608

ABSTRACT

S

ingle missense mutations in the F8 gene encoding the coagulation protein factor VIII give rise predominantly to non-severe hemophilia A. Despite only a single amino acid sequence difference between the replacement, therapeutic factor VIII and the patient’s endogenous factor VIII, therapeutic factor VIII may still be perceived as foreign by the recipient’s immune system and trigger an immune response (inhibitor). Inhibitor formation is a life-long risk for patients with non-severe hemophilia A treated with therapeutic factor VIII, but remains difficult to predict. The aim of this study was to understand whether fortuitous, primary sequence cross-matches between therapeutic factor VIII and proteins in the human proteome are the reason why certain F8 mutations are not associated with inhibitor formation. We predicted which therapeutic factor VIII differences are potentially perceived as foreign by helper T cells – a necessary precursor to inhibitor development – and then scanned potentially immunogenic peptides against more than 100,000 proteins in the proteome. As there are hundreds of disease-causing F8 missense mutations and the human leukocyte antigen gene complex governing peptide presentation to helper T cells is highly polymorphic, these calculations pose a huge combinatorial challenge that we addressed computationally. We found that cross-matches between therapeutic factor VIII and the human proteome are commonplace and have a profound impact on the predicted risk of inhibitor development. Our results emphasize the importance of knowing both the F8 missense mutation and the human leukocyte antigen alleles of a patient with missense mutation hemophilia A if his underlying risk of inhibitor development is to be estimated.

Introduction Subjects with all severities of hemophilia A are at risk of an alloimmune response (inhibitor formation) against infused, therapeutic FVIII (tFVIII) concentrate. It is well recognized that the more disruptive the F8 mutation, the more severe the hemophilia and the more likely it is that inhibitors will arise.1 Consequently, severe hemophilia A has been the priority for inhibitor-related research, surveillance and intervention over the past decades.2–5 However, it is also clear that only a single amino acid difference between an endogenous F8 genotype and the wild-type tFVIII sequence is sufficient to induce an immune response that results in clinically relevant inhibitors6–8 and that this risk is life-long in the context of non-severe hemophilia A.8 Hemophilia A caused by a missense mutation is typically associated with a less severe bleeding phenotype than that caused by incomplete F8 transcripts. In contrast to boys and men with severe hemophilia A, those living with non-severe hemophilia A are more likely to remain hospital dependent for on-demand tFVIII administration throughout their lives in the event of injury or surgery. The treathaematologica | 2019; 104(3)

Correspondence: ADRIAN J. SHEPHERD a.shepherd@mail.cryst.bbk.ac.uk Received: April 20, 2018. Accepted: September 27, 2018. Pre-published: September 28, 2018. doi:10.3324/haematol.2018.195669 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/104/3/599 ©2019 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

599


D.P. Hart et al.

ment burden for this group is surprisingly high, with 44% of a large London cohort being reported to have received some hemostatic treatment in a 2-year observation window, 79% of whom received tFVIII concentrate.9 Consequently, inhibitor surveillance in non-severe hemophilia A requires adult treaters to be ever vigilant.8 In contrast to the systematic inhibitor screening in patients exposed early to tFVIII for severe hemophilia A,5 inhibitor screening in the setting of non-severe hemophilia A is currently more reactive and sporadic,9 but recognized to be of increasing importance given the aging population of those living with non-severe hemophilia A.10 Inhibitor occurrence in non-severe hemophilia A can be devastating, with neutralization of infused FVIII concentrate and potential cross-reactivity with endogenous FVIII. This cross-reactivity occurs in at least 50% of identified cases11 and results in loss of a patient’s previous non-severe FVIII activity baseline level (FVIII:C) resulting in a worsening bleeding phenotype, often in later decades of life.8 This, in turn, results in increased bleed rates and an increased risk of premature mortality.12 In this context, the early detection of inhibitor occurrence – or, better still, the ability to reliably predict an individual’s risk of developing inhibitors before any have formed – has the potential to influence subsequent clinical decisions in ways that would substantially improve the patient’s outcomes. The T-cell dependency of inhibitor generation is well described, with confirmed tFVIII-specific CD4+ T-cell responses13–15 and immunoglobulin class switching.16 The activation of CD4+ T cells depends on their interaction with “foreign” peptides – in this case, tFVIII-derived peptides spanning the location of the endogenous F8 missense mutation – presented by major histocompatibility complex (MHC) class II molecules. However, not all such foreign peptides are perceived to be immunologically different from self and, if the difference is undetected, there is presumed negligible risk of an immune response. There are two key mechanisms at work here. Firstly, not all peptides are capable of binding to an individual’s repertoire of MHC molecules and are therefore never presented to T cells. Secondly, not all binding peptides are distinguishable from self-peptides bound to the same MHC molecules; in such cases, T cells that are capable of binding to these MHC:peptide complexes are expected to have been removed from the T-cell repertoire by self-tolerance mechanisms.17 What is unclear, however, is whether an understanding of MHC presentation and self-tolerance can enable us to make useful predictions about the inhibitor risk of patients with missense mutation hemophilia A – for example, by accurately predicting whether individual patients have a negligible risk of developing inhibitors. The aim of this study was to address this point directly. Given that MHC molecules are encoded by genes that are among the most polymorphic in the human genome, and there are several hundred disease-causing F8 missense mutations, the first aim of this study was to predict inhibitor risk based on an analysis of tFVIII peptide presentation by MHC molecules. Such an investigation poses a huge combinatorial challenge – one that is arguably impractical to address using purely in vitro techniques. Building on the approach developed in an earlier study that we undertook using a much smaller dataset,18 we analyzed MHC:peptide complexes associated with 25 common human leukocyte antigen (HLA) class II alleles and 956 distinct F8 missense mutations, requiring over four 600

million peptide-HLA isoform combinations to be evaluated. However, this preliminary analysis did not take into account the possibility that fortuitous cross-matches between tFVIII-derived peptides and peptides at other locations – both within the FVIII protein sequence itself, and more generally to other proteins in the human proteome – may play a protective role by ensuring that T cells capable of triggering an immune response have been removed from the repertoire by self-tolerance mechanisms. This includes, but is far from limited to, the welldescribed homology between FVIII and factor V.19 Such human proteome cross-matching is the main focus of this study. Here we demonstrate that cross-matches between tFVIII and other parts of the proteome are commonplace and have a profound impact on the predicted inhibitor risk for individuals living with non-severe hemophilia A.

Methods Novel peptide-MHC surfaces In previous work, we developed a methodology for predicting which patients with non-severe hemophilia A are at risk of developing antibodies against tFVIII.18 Specifically, we predicted which F8 missense mutation/HLA isoform combinations would present novel peptide-MHC (pMHC) surfaces to CD4+ T cells, taking into account the reasonable assumption that T cells capable of binding to pMHC surfaces that are formed by endogenous FVIII peptides and presented by the same MHC molecules would have been removed from the T-cell repertoire by central tolerance mechanisms. Novelty arises when a tFVIII-derived peptide is an MHCbinder (most are not) and either (i) the equivalent endogenous FVIII-derived peptide is a non-binder (this may occur if the missense mutation is at an MHC-facing, peptide-anchoring position, as residues at such positions anchor the peptide to the MHC molecule) or (ii) the relevant amino-acid difference is at a T-cell receptor (TCR)-facing position (Figure 1A). Our earlier work focused exclusively on the location of the hemophilia-causing F8 missense mutation and HLA-DR presentation. Here we extend our analysis to include HLA-DP and -DQ presentation and, crucially, to take into account the possibility of fortuitous peptide cross-matches to other locations – both within FVIII itself, and more generally within other proteins in the human proteome (Figure 1B).

Peptide-MHC binding prediction We used the in silico tool NetMHCII20 to predict the strength with which a given peptide binds to an MHC molecule – specifically, the portable version of NetMHCII 2.2 (Technical University of Denmark, http://www.cbs.dtu.dk/services/NetMHCII/). If a peptide is predicted to bind with a 50% inhibitory concentration (IC50) of <1000 nmol/L – the conventional threshold used to indicate that peptide-MHC class II binding is of biological significance21 – it is a candidate for forming a novel pMHC surface. Deciding whether a given combination of tFVIII peptide and MHC molecule forms a novel pMHC surface in comparison to the corresponding endogenous FVIII-derived peptide requires the considerations outlined in Figure 1 to be made, given: (i) knowledge of the positions of the MHC anchoring pockets for the chosen HLA isoform (generally these are at positions 1, 4, 6 and 9); and (ii) the binding register of the peptide (predicted by NetMHCII), which specifies the stretch of nine consecutive amino-acid residues that form the preferred binding core within the MHC groove. Predictions were made for 25 common HLADR, -DP and -DQ isoforms with estimated worldwide population haematologica | 2019; 104(3)


Human proteome and hemophilia A inhibitor risk

coverages of >70%, >90% and >80%, respectively,22 using UniProt23 FVIII sequence P00451. For the full set of 25 HLA class II alleles considered for this research, over four million peptide-HLA isoforms combinations were evaluated in order to identify the combinations that are predicted to form a novel pMHC surface – plus a similar order of additional evaluations necessary to identify potential cross-matches to the human proteome. A detailed breakdown of the calculations performed is given in Online Supplementary Table S1.

Scanning the human proteome for cross-matches Our preliminary risk assessment focused exclusively on the location of the disease-causing F8 missense mutation – that is, we assessed whether one or more peptides would be perceived as foreign, given the binding properties and side-chain orientations of both tFVIII and endogenous FVIII peptides spanning the location of the missense mutation (Figure 1A). The innovative hypothesis explored here is that some pMHC surfaces that were identified as risk-associated by this preliminary approach may not be novel within the broader context of the whole human proteome. For this research, we compared pMHC surfaces formed by tFVIII peptides spanning the location of the F8 missense mutation with those formed by peptides from the human proteome. Following research showing that the “maximal representation of the ‘immunological self’”17 is made available for tolerance induction in the thymus, we used the complete human proteome from Ensembl24 containing over 100,000 proteins, including alternative isoforms that have an associated protein product.

As previously, we confined our analysis to 15-mers, this being the most common peptide length chosen for MHC class II binding experiments. The canonical proteome was sub-divided into all possible 9-mers. The resultant dataset consists of nearly 38 million 9-mers of which more than 11 million are non-identical. A summary of the computational pipeline used to identify novel pMHC surfaces is shown in Figure 2.

Statistical analysis We evaluated the accuracy of our method using F8 mutation data downloaded from the largest source of F8 mutation data in the public domain (The European Association for Haemophilia and Allied Disorders. The Factor VIII Gene (F8) Variant Database. http://www.factorviii-db.org. Accessed November 26, 2016). The dataset contained 956 distinct F8 missense mutations at 605 different loci from 3,243 individuals. Ninety of the missense mutations were associated with at least one reported case of inhibitor formation, with a total of 160 individuals (prevalence 4.9%) listed as having inhibitors. We tested the null hypothesis that our predicted rate and the database reported rate of inhibitor formation are independent. In the absence of HLA-typing information for the patients, we predicted whether a patient with a given missense mutation has a risk of inhibitor formation based on the combined predictions for our chosen set of 25 common HLA class II isoforms. We evaluated different IC50 binding thresholds for tFVIII peptides, on the grounds that binding strength is likely to be an important factor in inhibitor risk.25 However, when considering peptide-MHC binding in the

A

Figure 1. Schematic diagram showing how side-chain differences may, or may not, lead to novel peptide-MHC surfaces. (A) If a tFVIII peptide spanning the location of the F8 missense mutation is a non-binder, it poses no risk of forming a novel pMHC surface capable of inducing an immune response. Otherwise, one needs to consider the position of the F8 missense mutation within the MHC groove of the tFVIII peptide and the corresponding endogenous peptide. For most HLA class II isoforms, positions 1,4 6 and 9 are MHC-facing, and positions 2, 3, 5, 7 and 8 are TCR-facing. Where the F8 missense mutation is at a downward, MHC-facing position (top row of A, denoted by a diamond), there are two scenarios: both tFVIII and endogenous peptides are binders, implying no risk; or the tFVIII peptide is a binder and the endogenous peptide is a non-binder, implying a potential risk. Where the F8 missense mutation is at an upward, TCR-facing position (bottom row of A, denoted by a diamond) and both peptides are binders, there is a potential risk. (B) Where a tFVIII peptide is associated with a potential risk according to the preceding assessment, a peptide from elsewhere in the human proteome that (i) has the same TCR-facing residues and (ii) is a binder, will militate against this risk, as no novel pMHC surface will be formed. FVIII: factor VIII; tFVIII: therapeutic factor VIII; TCR: T-cell receptor; MHC: major histocompatibility complex; pMHC: peptide-major histocompatibility complex.

B

haematologica | 2019; 104(3)

601


D.P. Hart et al.

context of self-tolerance prediction (e.g. the binding potential of proteome cross-matches), we maintained a binding threshold of 1000 nM. For a given tFVIII-peptide binding threshold, a patient’s predicted risk of inhibitor formation was deemed to fall within one of the following three categories: “predicted low/negligible risk”, “predicted at risk”, or “unknown risk”. A patient was deemed to be at low/negligible risk if no novel pMHC surfaces were predicted to be formed with any of the 25 HLA isoforms. A patient was deemed to be “at risk” if novel pMHC surfaces were predicted to be formed by all, or all but one, of the HLA isoforms encoded by at least one HLA gene complex – HLA-DRB1, HLADRB3/4/5, HLA-DP, and/or HLA-DQ isoforms. (Here the “all, or all but one” condition was designed to rule out the possibility that

the patient was a heterozygous individual having a risk-free combination of common HLA isoforms, i.e. two or more low/negligible risk isoforms per gene complex.) The inhibitor risk for patients in neither of the preceding categories was considered “unknown”. Patients with unknown inhibitor risk were omitted from the statistical test. The statistical test was undertaken using the two-tailed Fisher exact test implemented in the R statistical programming language. The Fisher exact test is necessary because the sample size and background inhibitor rate are both relatively low; in such circumstances, the calculation of exact P values is important. Following standard convention, a P value of <0.05 was used to define statistical significance.

Figure 2. Flowchart for the assessment of human proteome cross-matching and missense mutation hemophilia A inhibitor risk. The flowchart shows the process by which the presence, or otherwise, of a novel peptide-MHC surface is determined given a specific combination of endogenous F8 missense mutation and human leukocyte antigen isoform. tFVIII: therapeutic factor VIII; FVIII: factor VIII; TCR: T-cell receptor.

602

haematologica | 2019; 104(3)


Human proteome and hemophilia A inhibitor risk

Results An example of proteome cross-matching Arg593Cys (R593C) (R612C using Human Genome Variation Society numbering) is a relatively common F8 missense mutation that has been identified as being associated with an “increased” risk of inhibitor formation – for example, 12/106 (11.3%) of R593C individuals in the INSIGHT cohort have been reported as having an inhibitor.8 Taking the common HLA-DR allele HLADRB1*01:01 as an example, our analysis of predicted inhibitor risk proceeded as follows. We began by using NetMHCII20 to predict whether any

of the tFVIII 15-mers spanning position 593 are binders. Several such 15-mers were predicted to bind to the MHC molecule associated with HLA-DRB1*01:01, and some had binding cores that span position R593. There were two such cores – IQRFLPNPA and YLTENIQRF – both of which were associated with multiple binding peptides, as shown in Figure 3A. The next step assessed whether either of these cores was associated with predicted pMHC surface novelty compared to their respective endogenous counterparts. Both of the endogenous cores – IQCFLPNPA and YLTENIQCF, respectively [with a Cys (C) in place of the Arg (R) of the equivalent tFVIII peptides] – were associat-

A

B

Figure 3. An example of information peptideMHC binding and novel surface: Arg593Cys for HLA-DRB1*0101. (A) NetMHCII predicts there are binding tFVIII 15-mers that have two cores – IQRFLPNPA and YLTENIQRF – containing Arg593 (R593). (B) These two cores form pMHC surfaces that are novel compared to the equivalent surfaces for endogenous FVIII. FVIII: factor VIII; tFVIII: therapeutic factor VIII; MHC: major histocompatibility complex; pMHC: peptide-major histocompatibility complex.

haematologica | 2019; 104(3)

603


D.P. Hart et al.

ed with predicted binding 15-mers. Hence the question of predicted novelty hinged on the position of R593 within the 9-mer: whether at a position involved in MHC-binding (and hence invisible to a TCR), or at a TCR-facing position. As the binding pockets in the MHC groove for HLA-DRB1*01:01 are at positions 1, 4, 6 and 9, both IQRFLPNPA (with R593 at TCR-facing position 3) and YLTENIQRF (with R593 at TCR-facing position 8) are associated with the formation of pMHC surfaces that are novel in comparison to those formed by endogenous FVIII, as shown in Figure 3B. Based on the preceding analysis alone, we would predict that a patient with the R593C mutation and HLA-DR allele HLA-DRB1*01:01 would be at risk of developing inhibitors. However, a tFVIII-derived pMHC surface that is novel with respect to an individualâ&#x20AC;&#x2122;s endogenous FVIII may not be novel in the wider context of his proteome. To evaluate this possibility of a proteome cross-match that reduces the risk of inhibitor development, we searched for a pattern for each of the 9-mer cores matching at T-cell facing positions 2, 3, 5, 7 and 8. For IQRFLPNPA and YLTENIQRF these patterns are XQRXLXNPX and XLTXNXQRX, respectively, where the letter X matches any amino-acid type. These patterns were scanned against a library containing all 11,272,502 unique 9-mers from the human proteome. In this case, the pattern XQRXLXNPX matched the 9-mer FQRELNNPL in human tubulin polyglutamylase (UniProt25 Q6ZT98), and pattern XLTXNXQRX matched both the 9-mer GLTENSQRD in dystrobrevin binding protein 1 (dysbindin) (UniProt D6RJC6) and the 9-mer ELTKNAQRA in the uncharacterized

human protein C2orf48 (UniProt Q96LS8), as shown in Figure 4A. The final step was to check whether a given crossmatching, proteome-derived 9-mer occurred as a binding core for HLA-DRB1*01:01, as only then would we hypothesize tolerance. In this case, NetMHCII predicted that both FQRELNNPL in tubulin polyglutamylase and ELTKNAQRA in C2orf48 form cores within 15-mers with a predicted IC50 <1000 nmol/L, as shown in Figure 4B. Hence, we ultimately predicted that the F8 missense mutation/HLA allele combination R593C/HLADRB1*01:01 confers no, or negligible, risk of inhibitor formation owing to fortuitous cross-matches to peptides in the human proteome.

Proteome cross-matches and inhibitor risk stratification In terms of individual combinations of F8 missense mutation and HLA-DR/DP/DQ isoforms, the impact of proteome cross-matches on predicted risk is shown in a comprehensive heat map (Online Supplementary Figure S1), with a subset of combinations shown in Figure 5. Each individual F8 missense mutation/HLA isoform combination is shown as a single square. An analysis of the full set of data indicates that the percentage of F8 missense mutation/HLA isoform combinations associated with predicted inhibitor risk falls appreciably when proteome crossmatches are taken into account: from 49% to 31% with a binding threshold of 1000 nM; and from 37% to 21% with a binding threshold of 500 nM. These predictions strongly suggest that the risk of

A

B

604

Figure 4. A proteome crossmatching example: Arg593Cys for HLA-DRB1*0101. This example concerns the two tFVIII binding cores from Figure 3, IQRFLPNPA and YLTENIQRF. (A) Both cores cross-match to 9-mers from proteins in the human proteome at TCR-facing positions 2, 3, 5, 7 and 9. (B) Both of the matched 9-mers (one each from tubulin polyglutamylase and C2orf48, but none from dysbindin) are predicted by NetMHCII to form binding cores within 15-mers derived from these proteins. Hence, we conclude that the F8 missense mutation/HLA isoform combination Arg593Cys/HLA-DRB1*0101 is associated with negligible risk of inhibitor formation. FVIII: factor VIII; tFVIII: therapeutic factor VIII; TCR: T-cell receptor; HLA: human leukocyte antigen.

haematologica | 2019; 104(3)


Human proteome and hemophilia A inhibitor risk

inhibitor formation is largely HLA-dependent and, in most cases, cannot be reliably predicted from knowledge of the F8 genotype alone. The proportion of reported F8 missense mutations that we predict to be risk-associated varies considerably between different HLA isoforms. For example, with a binding threshold of 1000 nM, it ranges between 21% (HLA-DQA1*03:01-DQB1*03:02 and HLADQA1*04:01-DQB1*04:02) to 86% (HLA-DRB1*01:01) without proteome cross-matches, and between 13% (HLA-DQA1*04:01-DQB1*04:02) and 48% (HLA-

DRB1*01:01) with proteome cross-matches. Values for all 25 HLA isoforms and multiple binding thresholds are given in Online Supplementary Table S2. The number of F8 missense mutations associated with no, or negligible, predicted risk (black columns in the heatmap) is only 69 out of 956 (7%). We consider any (human) protein that, in effect, contributes to a reduction in predicted inhibitor risk for one or more F8 missense mutation/HLA isoform combinations to be â&#x20AC;&#x153;protectiveâ&#x20AC;?. Of the 20,300 proteins in the human pro-

A

B

Figure 5. MHC-binding strengths of F8 peptides predicted to form novel pMHC surfaces with and without proteome scanning. Heatmap showing the predicted occurrence of novel pMHC surfaces and binding strengths for 25 HLA-DR/DP/DQ isoforms (y axis) covering the first 50 missense mutations in the Factor VIII Gene (F8) Variant Database (x axis). Black and gray squares indicate F8 missense mutation/HLA isoform combinations that are not predicted to form a novel pMHC surface. Otherwise the temperature color scale indicates the predicted binding strength of the strongest binding peptide with a novel pMHC surface for each remaining F8 missense mutation/HLA isoform combination. The full heatmap for all missense F8 mutations is given in Online Supplementary Figure S1. (A) MHC-binding strengths of F8 peptides predicted to form novel pMHC surfaces (colored squares), or not (black squares), without proteome scanning. (B) MHC-binding strengths of F8 peptides predicted to form novel pMHC surfaces with proteome scanning. Gray squares indicate F8 missense mutation/HLA isoform combinations that are no longer predicted to form a novel pMHC surface after cross-matches to the proteome are taken into account. MHC: major histocompatibility complex; pMHC: peptidemajor histocompatibility complex; HLA: human leukocyte antigen.

haematologica | 2019; 104(3)

605


D.P. Hart et al.

teome, 4,605 are protective. The protein that affords the most protection is factor V, which is recognized as having high homology with FVIII.19 The most protective proteins identified in our analysis are listed in Table 1. It is notable that FVIII and all the top entries in Table 1 (coagulation factor V, hephaestin-like protein 1, ceruloplasmin, and hephaestin) have copper-binding sites.

tion as the basis for identifying patients with missense mutation hemophilia A who are at risk of developing inhibitors against tFVIII. Our pipeline incorporates a novel strategy that we term proteome scanning â&#x20AC;&#x201C; the identification of fortuitous cross-matches between potentially antigenic tFVIII peptides and peptides arising else-

Evaluation of risk prediction accuracy

Table 1. Human proteins that afford the greatest proteome crossmatch protection.

Given the paucity of published data specifying the HLA profiles of patients with missense mutation hemophilia A, we based the evaluation of how accurate our approach is at predicting potential risk on data in the Factor VIII Gene Variant Database, focusing on F8 missense mutations that we predict to have low or negligible risk with any common HLA isoform. Arguably the most important performance indicator for our method is the number of false negatives: are there individuals with missense mutation hemophilia A and with data in the Factor VIII Gene Variant Database who we predict to have a zero, or negligible, inhibitor risk but have, in fact, developed inhibitors? The results in Table 2 for our chosen set of 25 HLA alleles show that the number of false negatives (column 3) is very low with conservative cut-offs for peptide-MHC binding affinity. The number of patients predicted to have a zero, or negligible risk of inhibitor development is considerably higher when crossmatches to the proteome are taken into account, but the false negative rate remains comparatively low; these factors taken together contribute to the higher prevalence of statistically significant P values with proteome crossmatching.

Discussion In this paper, we highlight the potential value of an in silico predictive model of HLA class II antigen presenta-

UniProt ID

Protein name

P12259 Q6MZM0 P00450 Q9BQS7 P00451

Coagulation factor V Hephaestin-like protein 1 Ceruloplasmin Hephaestin Coagulation factor VIII [match to different, but homologous, location within the protein] Usherin Zinc finger protein 345 Zinc finger protein 268 Dynein heavy chain domain-containing protein Cytosolic acyl coenzyme A thioester hydrolase Lysine-specific demethylase 3B Disintegrin and metalloproteinase domain-containing protein 30 Zinc finger protein 835 Retinol-binding protein 3 Uncharacterized protein C6orf118

O75445 Q14585 Q14587 Q96M86 O00154 Q7LBC6 Q9UKF2 Q9Y2P0 P10745 Q5T5N4

Protected peptide count 640 457 437 389 251 150 142 134 83 76 75 75 74 73 67

The protected peptide count for a given combination of human protein p, F8 missense mutation m and HLA isoform h is incremented by 1 every time a peptide in tFVIII that spans the location of m is (i) associated with a predicted risk of inhibitor development for h prior to considering cross-matches to the human proteome, and (ii) cross-matches to a peptide in p that is a predicted binder for h. Hence a peptide that cross-matches to multiple binding peptides at different locations within p will be counted multiple times. The final count for p is the aggregate of individual counts for all F8 missense mutation/HLA isoform combinations considered in this study.

Table 2. Evaluation of zero/negligible inhibitor risk prediction with and without proteome scanning.

IC50 binding threshold (nmol/L)

Patients predicted to have zero/negligible risk No inhibitors Inhibitors

Patients predicted to have an inhibitor risk Inhibitors No inhibitors

P

Without proteome scanning 1000 500 300 200 100 50

28 49 122 179 362 593

1 3 3 9 20 36

116 92 84 76 37 31

1344 985 787 660 338 228

0.72 0.62 5.84x10-4 0.02 0.02 2.01x103

With proteome scanning 1000 500 300 200 100 50

103 157 322 465 777 1,114

4 7 14 26 42 66

80 65 57 53 23 22

622 339 261 232 133 115

0.02 4.50x10-5 1.14x10-8 1.07x10-8 6.57x10-5 3.72x10-5

P values were calculated by applying the Fisher exact test to patients from the Factor VIII Gene Variant Database falling into the following categories: predicted to have zero/negligible risk, observed to have no inhibitors (column 2); predicted to have zero/negligible risk, observed to have inhibitors (column 3); predicted to have inhibitor risk, observed to have inhibitors (column 4); and predicted to have inhibitor risk, observed to have no inhibitors (column 5). P values <0.05 are deemed statistically significant and are shown in bold.

606

haematologica | 2019; 104(3)


Human proteome and hemophilia A inhibitor risk

where within the human proteome. Such cross-matching peptides are the basis for protection against inhibitor development because of presumed T-cell tolerance mechanisms. Here we focused on the surfaces formed by tFVIII peptides that span the locations of known disease-causing F8 missense mutations and are predicted to bind to the MHC molecules for 25 common HLA class II alleles. Given a conservative binding threshold of 1000 nM, the number of F8 missense mutation/HLA isoform combinations associated with a risk of developing inhibitors was predicted to fall by more than a third – from 49% to 31% – when cross-matches to the proteome are taken into account. These results were shown to be statistically significant with a dataset derived from the Factor VIII Gene Variant Database of patients with missense mutation hemophilia A. Although our proteome scanning approach reduces the number of patients predicted to be at risk of developing inhibitors, that number remains higher than, albeit closer to, the number of patients that have – at least to date – developed inhibitors. This is inevitable for a model based entirely on a consideration of MHC/TCR interactions, as a range of downstream factors may militate against inhibitor development. These include: the absence of sufficient T cells capable of binding to a given pMHC surface for reasons other than self-tolerance; the lack of co-stimulatory signaling; or the level of exposure to tFVIII being below the threshold necessary for inhibitor formation (only cursory information about a patient’s degree of exposure to tFVIII is available in the Factor VIII Gene Variant Database). There are a number of ways in which this analysis could be refined. Firstly, we took no account of potential F8 genotype mismatches between tFVIII products (derived from common F8 genotypes H1 and H2 that differ only in the B domain) and rare genotypes H3-8, such as the M2238V found in approximately 23% of black people.26 Nor did we consider the antigenic impact of different linkers used in B-domain-modified tFVIII products. Secondly, proteome scanning was performed against a single reference proteome. It is likely that additional cross-matches will be found if allelic variants are taken into account, adding further to the potential advantages of personalized risk assessment. Scanning against an individual’s own proteome would be the optimal predictive strategy. The impact of proteome variability will be assessed in future work using data from IGSR: The International Genome Sample Resource.27 There are several more challenging issues. Our model of peptide-MHC binding is imperfect, for example: we do not take into account the impact of cathepsin cleavage on the availability of FVIII peptides for MHC class II binding (there are no established computational methods for predicting cleavage by cathepsins, and different sets of cathepsins occur in different professional antigen-present-

References 1. Schwaab R, Brackmann HH, Meyer C, et al. Haemophilia A: mutation type determines risk of inhibitor formation. Thromb Haemost. 1995;74(6):1402–1406. 2. Gouw SC, van der Bom JG, Ljung R, et al.

haematologica | 2019; 104(3)

ing cells28); peptide differences at anchoring positions,29 or outside the binding core,30 are known to affect the formation of pMHC surface in specific cases (but the prevalence of such effects is poorly understood); and a given TCR may not be in contact with all TCR-facing residues (but the binding angle and register of individual TCR is currently unpredictable).31 Validating the accuracy of inhibitor risk prediction for patients with non-severe hemophilia is also particularly problematic. In practice, the current, clinical gold standard for inhibitor detection is a functional, clotting-based Bethesda assay. However, heat treatment modifications in the presence of residual FVIII:C (i.e. non-severe hemophilia A) are often omitted, resulting in reduced sensitivity of detection.32 More importantly, a purely “functional” clotting assay does not detect the totality of antibody responses against a protein therapeutic. The absence of a more “neutral” screening assay (e.g. based on enzyme-linked immunosorbence) to pick up any anti-tFVIII antibody response first, and subsequently for the functional assay (Bethesda) to determine its inhibitory potential and clinical relevance, compromises our knowledge of the totality of anti-tFVIII responses in our cohorts of patients. It is also evident that the screening practice for antibody responses in non-severe hemophilia A, in contrast to that for severe hemophilia A, is often opportunistic and passive, further reducing the likelihood of detecting the totality of antitFVIII antibody responses by missing the optimal immunological windows for screening after tFVIII exposure.9 Given the life-long risk of inhibitor formation in non-severe hemophilia A, we have concerns that true negatives (i.e. patients confirmed to have a zero risk of inhibitor development) are impossible to identify in a clinical study of nonsevere hemophilia A, even when factors such as age and exposure are taken into account. Notwithstanding these limitations, this study provides compelling evidence of the importance of HLA class II genotyping for analyzing the inhibitor risk of patients with missense mutation hemophilia A. Moreover, we have demonstrated that an innovative computational pipeline incorporating proteome scanning predicts that a large proportion of F8 missense mutation/HLA isoform combinations afford a negligible risk of inhibitor development, with a low error rate when evaluated using the largest available dataset of patients with F8 missense mutations and conservative MHC binding thresholds. This represents an important step forward, as it closes part of the gap between predicted/potential inhibitor risk and observed inhibitor rates. These insights may ultimately contribute to the design of future clinical studies (with HLA typing of missense mutation hemophilia A patients) that are of direct translational relevance. Acknowledgments DPH received funding from the British Society of Haematology.

Factor VIII products and inhibitor development in severe hemophilia A. N Engl J Med. 2013;368(3):231–239. 3. Gouw SC, van den Berg HM, Fischer K, et al. Intensity of factor VIII treatment and inhibitor development in children with severe hemophilia A: the RODIN study. Blood. 2013;121(20):4046–4055.

4. Collins PW, Palmer BP, Chalmers EA, et al. Factor VIII brand and the incidence of factor VIII inhibitors in previously untreated UK children with severe hemophilia A, 20002011. Blood. 2014;124(23):3389–3397. 5. Collins PW, Chalmers E, Hart DP, et al. Diagnosis and treatment of factor VIII and IX inhibitors in congenital haemophilia: (4th

607


D.P. Hart et al.

6.

7. 8.

9.

10.

11.

12.

13.

14.

608

edition). UK Haemophilia Centre Doctors Organization. Br J Haematol. 2013;160(2): 153–170. Fischer K, Iorio A, Lassila R, et al. Inhibitor development in non-severe haemophilia across Europe. Thromb Haemost. 2015;114 (4): 670–675. Hay CR. Factor VIII inhibitors in mild and moderate-severity haemophilia A. Haemophilia. 1998;4(4):558–563. Eckhardt CL, van Velzen AS, Peters M, et al. Factor VIII gene (F8) mutation and risk of inhibitor development in nonsevere hemophilia A. Blood. 2013;122(11):1954–1962. Batty P, Austin SK, Khair K, et al. Treatment burden, haemostatic strategies and real world inhibitor screening practice in nonsevere haemophilia A. Br J Haematol. 2017;176(5):796–804. Hvas A-M, Poulsen LH. Inhibitor screening in non-severe haemophilia patients; a major challenge. Br J Haematol. 2017;176(5):683– 685. van Velzen AS, Eckhardt CL, Hart DP, et al. Inhibitors in nonsevere haemophilia A: outcome and eradication strategies. Thromb Haemost. 2015;114(1):46–55. van Velzen AS, Eckhardt CL, Streefkerk N, et al. The incidence and treatment of bleeding episodes in non-severe haemophilia A patients with inhibitors. Thromb Haemost. 2016;115(3):543–550. James EA, Van Haren SD, Ettinger RA, et al. T-cell responses in two unrelated hemophilia A inhibitor subjects include an epitope at the factor VIII R593C missense site: T-cell responses to a FVIII missense site in hemophilia A. J Thromb Haemost. 2011;9(4):689– 699. Ettinger RA, James EA, Kwok WW, Thompson AR, Pratt KP. HLA-DR-restricted

15.

16.

17.

18.

19.

20.

21.

22.

23.

T-cell responses to factor VIII epitopes in a mild haemophilia A family with missense substitution A2201P. Haemophilia. 2010;16 (102):44–55. James EA, Kwok WW, Ettinger RA, Thompson AR, Pratt KP. T-cell responses over time in a mild hemophilia A inhibitor subject: epitope identification and transient immunogenicity of the corresponding self-peptide. J Thromb Haemost. 2007;5(12): 2399–2407. Whelan SFJ, Hofbauer CJ, Horling FM, et al. Distinct characteristics of antibody responses against factor VIII in healthy individuals and in different cohorts of hemophilia A patients. Blood. 2013;121(6):1039–1048. Derbinski J, Kyewski B. How thymic antigen presenting cells sample the body’s selfantigens. Curr Opin Immunol. 2010;22(5): 592–600. Shepherd AJ, Skelton S, Sansom CE, Gomez K, Moss DS, Hart DP. A large-scale computational study of inhibitor risk in non-severe haemophilia A. Br J Haematol. 2015;168(3): 413–420. Davidson CJ, Hirt RP, Lal K, et al. Molecular evolution of the vertebrate blood coagulation network. Thromb Haemost. 2003;89(3): 420–428. Nielsen M, Lund O. NN-align. An artificial neural network-based alignment algorithm for MHC class II peptide binding prediction. BMC Bioinformatics. 2009;10(1):296. Southwood S, Sidney J, Kondo A, et al. Several common HLA-DR types share largely overlapping peptide binding repertoires. J Immunol. 1998;160(7):3363–3373. Wang P, Sidney J, Kim Y, et al. Peptide binding predictions for HLA DR, DP and DQ molecules. BMC Bioinformatics. 2010;11 (1):568. UniProt Consortium. Reorganizing the pro-

24. 25.

26.

27.

28.

29.

30.

31. 32.

tein space at the Universal Protein Resource (UniProt). Nucleic Acids Res. 2012;40(D1): D71-75. Yates A, Akanni W, Amode MR, et al. Ensembl 2016. Nucleic Acids Res. 2016;44(D1):D710-716. Lazarski CA, Chaves FA, Jenks SA, et al. The kinetic stability of MHC class II:peptide complexes is a key parameter that dictates immunodominance. Immunity. 2005;23(1): 29–40. Viel KR, Ameri A, Abshire TC, et al. Inhibitors of factor VIII in black patients with hemophilia. N Engl J Med. 2009;360 (16):1618–1627. 1000 Genomes Project Consortium, Auton A, Brooks LD, et al. A global reference for human genetic variation. Nature. 2015;526 (7571):68–74. Hsing LC, Rudensky AY. The lysosomal cysteine proteases in MHC class II antigen presentation. Immunol Rev. 2005;207(1):229– 241. Kersh GJ, Miley MJ, Nelson CA, et al. Structural and functional consequences of altering a peptide MHC anchor residue. J Immunol. 2001;166(5):3345–3354. Deng L, Langley RJ, Brown PH, et al. Structural basis for the recognition of mutant self by a tumor-specific, MHC class II-restricted T cell receptor. Nat Immunol. 2007;8(4):398–408. Rudolph MG, Stanfield RL, Wilson IA. How TCRs bind MHCs, peptides, and coreceptors. Annu Rev Immunol. 2006;24:419–466. Jennings IA, Kitchen D, Kitchen S, et al. Variation in practice for FVIII and FIX inhibitor investigations - results from a UK NEQAS (Blood Coagulation) and UK HCDO multicentre exercise. Res Pract Thromb Haemost. 2017;1(S1):525.

haematologica | 2019; 104(3)


ARTICLE

Immunodeficiencies

Selective loss of function variants in IL6ST cause Hyper-IgE syndrome with distinct impairments of T-cell phenotype and function

Tala Shahin,1,2* Dominik Aschenbrenner,3* Deniz Cagdas,4,5* Sevgi Köstel Bal,1,2,6 Cecilia Domínguez Conde,1,2 Wojciech Garncarz,1,2 David Medgyesi,1,2 Tobias Schwerd,3,7 Betül Karaatmaca,4 Pınar Gur Cetinkaya,4 Saliha Esenboga,4 Stephen R. F. Twigg,8 Andrew Cant,9 Andrew O. M. Wilkie,8 Ilhan Tezcan,4,5 Holm H. Uhlig3,10 and Kaan Boztug1,2,11,12

Ludwig Boltzmann Institute for Rare and Undiagnosed Diseases, Vienna, Austria; CeMM Research Center for Molecular Medicine of the Austrian Academy of Sciences, Vienna, Austria; 3Translational Gastroenterology Unit, John Radcliffe Hospital, University of Oxford, UK; 4Section of Pediatric Immunology, Ihsan Doğramacı Children’s Hospital, Hacettepe University, Ankara, Turkey; 5Institute of Child Health, Hacettepe University, Ankara, Turkey; 6Department of Pediatric Allergy and Immunology, Ankara University School of Medicine, Cebeci, Turkey; 7Dr. von Hauner Children's Hospital, LudwigMaximilians-University of Munich, Germany; 8Clinical Genetics Group, MRC Weatherall Institute of Molecular Medicine, John Radcliffe Hospital, University of Oxford, UK; 9 Institute of Cellular Medicine, Newcastle University, Newcastle upon Tyne, UK; 10 Department of Paediatrics, University of Oxford, UK; 11Department of Pediatrics and Adolescent Medicine, Medical University of Vienna, Austria and 12St. Anna Kinderspital and Children’s Cancer Research Institute, Department of Pediatrics, Medical University of Vienna, Austria 1 2

Ferrata Storti Foundation

Haematologica 2019 Volume 104(3):609-621

*TSh, DA and DC contributed equally to this work. #HHU and KB are co-senior authors.

ABSTRACT

H

yper-IgE syndromes comprise a group of inborn errors of immunity. STAT3-deficient hyper-IgE syndrome is characterized by elevated serum IgE levels, recurrent infections and eczema, and characteristic skeletal anomalies. A loss-of-function biallelic mutation in IL6ST encoding the GP130 receptor subunit (p.N404Y) has very recently been identified in a singleton patient (herein referred to as PN404Y) as a novel etiology of hyper-IgE syndrome. Here, we studied a patient with hyper-IgE syndrome caused by a novel homozygous mutation in IL6ST (p.P498L; patient herein referred to as PP498L) leading to abrogated GP130 signaling after stimulation with IL-6 and IL-27 in peripheral blood mononuclear cells as well as IL-6 and IL-11 in fibroblasts. Extending the initial identification of selective GP130 deficiency, we aimed to dissect the effects of aberrant cytokine signaling on T-helper cell differentiation in both patients. Our results reveal the importance of IL-6 signaling for the development of CCR6-expressing memory CD4+ T cells (including T-helper 17-enriched subsets) and non-conventional CD8+ T cells which were reduced in both patients. Downstream functional analysis of the GP130 mutants (p.N404Y and p.P498L) have shown differences in response to IL-27, with the p.P498L mutation having a more severe effect that is reflected by reduced T-helper 1 cells in this patient (PP498L) only. Collectively, our data suggest that characteristic features of GP130-deficient hyper-IgE syndrome phenotype are IL-6 and IL-11 dominated, and indicate selective roles of aberrant IL-6 and IL-27 signaling on the differentiation of T-cell subsets.

Introduction Hyper-IgE syndromes (HIES) comprise a group of primary immunodeficiencies (PIDs) associated with recurrent pulmonary infections, eczema and skin abscesses. As a subtype of HIES, autosomal-dominant STAT3 deficiency also involves skeletal abnormalities including scoliosis, craniosynostosis and retained dentition.1-4 We haematologica | 2019; 104(3)

Correspondence: KAAN BOTZUG kaan.boztug@rud.lbg.ac.at HOLM H. UHLIG holm.uhlig@ndm.ox.ac.uk Received: March 26, 2018. Accepted: October 3, 2018. Pre-published: October 11, 2018. doi:10.3324/haematol.2018.194233 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/104/3/609 ©2019 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

609


T. Shahin et al.

recently identified a singleton patient (herein referred to as PN404Y) carrying a homozygous loss-of-function mutation in IL6ST (p.N404Y) encoding the cytokine receptor subunit GP130 with remarkable clinical manifestations resembling that of STAT3 HIES, whereby PN404Y experienced recurrent infections, eczema, elevated IgE, eosinophilia, impaired acute-phase response, scoliosis and craniosynostosis.5 The GP130 subunit binds to several cytokine receptors including IL-6 receptor alpha (IL-6RA), IL-11RA, IL-27RA, leukemia inhibitory factor (LIF) receptor, oncostatin M (OSM) receptor and ciliary neurotrophic factor (CNTF) receptor. Upon stimulation with the respective cytokine, several JAK/STAT pathways are activated downstream of these receptors.6 The importance of GP130-mediated signaling was shown in Il6st-/- mice that die embryonically due to myocardial and hematopoietic defects.7 Moreover, postnatal conditional inactivation of gp130 in mice leads to neurological, hepatic and immunological defects with impaired acute-phase response, increased susceptibility to infections, and development of lung emphysema.8 With regards to cytokines, the contribution of IL-6 signaling to immune responses against pathogens has been emphasized in Il-6-deficient mice as well as in children with autoantibodies against IL-6.9-12 The roles of other GP130dependent cytokines in the context of disease have been studied, whereby the importance of IL-27 signaling has been shown in mediating T-cell responses, IL-11 in bone development, LIF in hematopoiesis and thymocyte development, and CTNF in the maintenance of motor neurons.13-17 Here, we identify a patient harboring a novel biallelic mutation in IL6ST (herein referred to as PP498L), presenting with elevated IgE levels, recurrent infections, severe atopic dermatitis, and characteristic skeletal abnormalities. We observe complete or partial disruption of selected cytokine signaling pathways in various cell types from PP498L. We extend our initial findings on human GP130 deficiency by comparatively dissecting signaling defects and perturbations in T-cell differentiation in the PP498L patient and our previously reported patient (PN404Y).5 Finally, we pinpoint selective roles of impaired IL-6, IL-11 and IL-27 signaling in the presence of other functional cytokine signaling pathways such as IL-21 and IL-10.

Methods Subjects Patients and healthy controls were included with informed written consent and approval from the Institutional Review Boards of the Medical University of Vienna, Hacettepe University Medical School in Ankara, Oxfordshire Research Ethics Committee B, the London Riverside Research Ethics Committee, and Oxford Gastrointestinal Illness Biobank.5

GP130 expression analysis Patient and healthy donor fibroblasts were detached by incubation with 50 mM EDTA in PBS for 30 minutes (min) on ice followed by gentle scraping with a silicon blade (CytoOne®). Cells were immediately washed with PBS and stained with GP130BV421 or an isotype control, CD4-BV421 (both IgG1κ), for 35 min on ice. Cells were washed and resuspended in PBS for flow cytometry analysis.

610

p-STAT analysis by flow cytometry Peripheral blood mononuclear cells (PBMCs). Frozen PBMCs from patient and healthy donor controls were thawed and allowed to recover for four hours (h) at 37°C in complete media (RPMI-1640 with 10% FBS). Cells were subsequently stained with CD3-FITC, CD4-BV605, CD8-V450 and CD19-PECy7 for 20 min at 37°C. Five minutes through the extracellular staining, cells were stimulated for 15 min at 37°C with IL-6, IL-21, IL-27 (all 100 ng/mL) or IL-10 (50 ng/mL). Cells were then immediately fixed for 10 min at 37°C, washed and permeabilized for 35 min on ice. Cells were then stained with p-STAT3-AF647 for 1 h at room temperature, washed again, and resuspended in FACS buffer for flow cytometry analysis. T lymphoblasts and Epstein-Barr virus-transformed lymphoblastoid cell lines (EBV-LCLs). T lymphoblasts from patient and healthy donors were starved for 2 h in T-cell media (RPMI-1640 with 5% humanserum) without IL-2, while EBV-LCLs were starved for 2-3 h in serum-deprived media (RPMI-1640). Cells were then stimulated with the cytokines mentioned and p-STAT staining performed as described above. (See Online Supplementary Table S1 for all antibodies used.) Fibroblasts and HEK293 GP130-KO cell line. Fibroblasts and HEK293 GP130-KO cells5 were grown in complete media (DMEM with 10% FBS) in 12-well and 96-well plates, respectively. Fibroblasts and HEK293 GP130-KO cells were serum-starved overnight or for 2-3 h, respectively, followed by stimulation with increasing concentrations (0.01, 0.1, 1, 10, 100 ng/mL) of IL-6, IL11, IL-27, LIF and OSM cytokines for 15 min, immediate detachment using Trypsin-EDTA, and fixation for 10 min at 37°C. The p-STAT staining was performed as described above. HEK293 GP130-KO cells were transiently transfected with empty plasmid [pcDNA3.1(+)] or plasmid encoding WT or mutant GP130 using Lipofectamine 2000 (Thermo Fisher) before p-STAT3 analysis as previously described.5 Cells were co-transfected with either IL-6RA or IL-11RA to enhance the phosphorylation signal. GFP-coding plasmid was co-transfected to allow gating on successfully transfected cells.

Intracellular cytokine staining T-cell cytokine production was analyzed in 0.5-1x106 total PBMCs that were stimulated for 5 h with Phorbol 12-myristate 13-acetate (PMA, 0.2 mM) and Ionomycin (1 mg/mL) with the addition of Brefeldin A during the final 2.5 h. To identify T-cell subpopulations, surface staining was performed as indicated below. Subsequently, cells were fixed, permeabilized and stained for intracellular cytokines followed by flow cytometry analysis. The following antibodies were used: IL-4, IL-10, IL-13, IL-17A, IL-22, IFN-γ and TNF. Dead cells were excluded from analysis by staining with fixable viability dye eFluor780 (eBioscience).

Chemokine receptor profiling For the analysis of surface marker expression by flow cytometry, 0.5-1x106 total PBMCs were incubated with fluorochromeconjugated antibodies for 15 min in PBS supplemented with 0.5% human serum at 37°C or room temperature, as assessed by titration experiments at both temperatures. To exclude dead cells from the analysis, cells were stained with fixable viability dye eFluor780 (eBioscience). Antibodies used for cell surface immunophenotyping included: CD3, TCRαb, CD4, CD8, CD25, CD45RA, CD127, CCR4, CCR6, CCR7, CCR9, CCR10, CXCR3, CXCR5, CRTh2. All samples were acquired using the Beckman Coulter CytoFlex, BD LSRII or BD LSR Fortessa. Other methods and details of the antibodies used can be found in Online Supplementary Table S1.

haematologica | 2019; 104(3)


IL6ST variants lead to aberrant T-cell phenotype

Table 1. Immunological characterization, immunoglobulin and eosinophil values of PP498L.

Age at evaluation

1.8 mo

9.9 yr

12.7 yr

CD45RA+CCR7+, %

8000 (3500-13100) 6480 (2300-7000) 4640 (1700-5300) ND

3600 (1100-5900) 2772 (700-4200) 1584 (300-2000) ND

CD45RA-CCR7+, %

ND

ND

CD45RA-CCR7-, %

ND

ND

CD45RA+CCR7-, %

ND

ND

CD45RA+CCR7+, %

2080 (400-1700) ND

1008 (300-1800) ND

CD45RA–CCR7+, %

ND

ND

CD45RA–CCR7–, %

ND

ND

CD45RA+CCR7-, %

ND

ND

CD27–IgD+, %

2.23 (1.3-6.3) 480 (200-1400) 880 (549-1225) ND

1.57 (0.9-2.6) 324 (90-900) 216 (228-516) ND

3500 (1000-5300) 2555 (800-3500) 1610 (400-2100) 61.2 (57.4-84.9) 6.9 (11.3-26.7) 18.2 (3.3-15.2) 13.7 (0.4-2.6) 945 (200-1200) 29.7 (28.4-80.6) 5.5 (1.4-5) 12.3 (6.2-29.3) 29.7 (9.1-49.1) 1.7 (0.9-3.4) ND

CD27+IgD+, %

ND

ND

CD27+IgD, %

ND

ND

Lymphocyte subsets Absolute lymphocyte count, g/L CD3+ T cells, cells/mL CD4+ T cells, cells/mL

CD8+ T cells, cells/mL

CD4/CD8 CD56+CD16+ NK cells, cells/mL CD19+ B cells, cells/mL

Age at evaluation Immunoglobulin IgG (g/L) IgM (g/L) IgA (g/L) IgE (IU/L) Eosinophils (g/L)

1.8 mo

6.5 yr

9.9 yr

12.1 yr

56.7 (226-370) 90.5 (75.2-86.7) 6 (4.6-10.2) 5.1 (3.3-9.6) 12.7 yr

5.19 (7.5-15.5) 0.24 (0.12-0.87) 0.073 (0.06-0.58) 6.71 (2-34) 6.3 (0.02-0.85)

19.80 (6.5-14.1) 1.07 (0.55-2.1) 1.28 (0.83-2.17) 2433 (2-307) NA (-)

13.10 (7.3-13.5) 1.23 (0.8-1.5) 0.94 (0.70-2.22) 2788 (2-696) 0.6 (0.0-0.5)

12.60 (7.7-15.1) 1.61 (0.7-1.5) 1.12 (1.08-3.25) 2788 (2-696) 1.1 (0.0-0.5)

15.80 (7.7-15.1) 1.86 (0.7-1.5) 1.46 (1.08-3.25) 6974 (2-696) 2.5 (0.0-0.5)

Reference ranges for B-cell absolute counts and percentages of naïve/memory subsets were obtained from a previously published cohort.46 Values out of range are shown in bold. mo: months; yr: years; ND: not determined; NA: not available.

haematologica | 2019; 104(3)

611


T. Shahin et al. A

B

C

D

E

F

G

Figure 1. Identification of a novel IL6ST variant in PP498L. (A) Pedigree of PP498L showing consanguinity in the family. PP498L has 3 deceased siblings, a sister, and twin brothers [one who died in utero (III-3)], and a brother who has congenital blindness of unknown etiology but is unremarkable for his immune system. PP498L is homozygous for IL6ST c.1493C>T (p.P498L/p.P498L). (B) X-ray images of PP498L showing scoliosis (left), scaphocephaly (top right), edema in the right ankle (bottom center), and flexion contractures of the small joints in the hand (bottom right). (C) Flow cytometry plot illustrating a reduction in CD19+ B cells and normal CD3+ T cells in PBMCs of PP498L at the of age 12.4 years. (D) Sanger sequencing of the identified IL6ST variant (c.1493C>T) in PP498L and family members. (E) Linear representation of GP130 and crystal structure of the GP130-IL6-IL6RÎą complex (adapted and modified5) outlining the protein domains and the two mutations found in PP498L and the previously described patient (PN404Y ; p.N404Y). D: domains; TM: transmembrane domain; CT: cytoplasmic tail. (F) Conservation of the amino acid proline at position 498 across species including some adjacent amino acids. (G) Flow cytometry analysis of GP130 expression in fibroblasts of PP498L compared to that of a healthy donor. Average of mean fluorescence intensity (MFI) from 2-3 technical replicates is shown on the top right area of the graphs.

612

haematologica | 2019; 104(3)


IL6ST variants lead to aberrant T-cell phenotype

Results Clinical disease manifestation and immunological characterization of Patient PP498L Patient PP498L was born to consanguineous Turkish parents (first-degree cousins) (Figure 1A). He experienced diarrhea at one month of age, recurrent otitis media, bilateral keratitis, and recurrent respiratory infections including pneumonia (Serratia marcescens and Pseudomonas aeruginosa were isolated) complicated by empyema and pneumothorax. In addition, he was followed up for severe eczema and food allergy (milk, egg and wheat; class III).

At 12 years of age, he developed an aphthous tongue ulcer suggestive of an undefined fungal lesion. No neutropenia was recorded, yet he benefited from granulocyte colony stimulating factor. He is currently under monthly intravenous immunoglobulin (IVIG) and cyclosporine A therapy. Early after birth, the patient exhibited flexion contractures of the hand joints and presented with scaphocephaly (suggesting craniosynostosis) (Figure 1B) and ArnoldChiari type 1 malformation. He has thoracolumbar scoliosis (30°) (Figure 1B), clubbing, crowded teeth and mild macroglossia. Cartilage destruction and erosion were seen

Figure 2. Functional assessment of GP130P498L variant in primary cells. (A) Assessment of GP130 function by flow cytometry measurement of percentage of p-STAT3 positive cells after stimulation of primary T cells with IL-6 (100 ng/mL), IL-27 (100 ng/mL), IL-10 (50 ng/mL) or IL-21 (100 ng/mL) in PP498L (orange), mother of PP498L (green), and 4 healthy donors (HD: blue; age-matched HDs are represented by circles and adult HDs by squares). (B) Overlayed histograms showing shifts in p-STAT3 signal upon IL-27 stimulation (solid line) and baseline (dotted lines) in CD3+, CD4+ and CD8+ T cells of both PP498L (orange) and PN404Y (red) compared to a HD (black). Values represent percentage of p-STAT3 positive cells. (C and D) Percentage p-STAT3 assessed in (C) T lymphoblasts from PP498L and 2 HDs and in (D) Epstein-Barr virus-transformed lymphoblastoid cell lines (EBV-LCLs) from PP498L, mother of PP498L and 2 HDs after stimulation with IL-6 (100 ng/mL), IL-10 (50 ng/mL), IL-21 (100 ng/mL), or IL-27 (100 ng/mL). Data shown are from 2-3 independent experiments (shown by different data point shapes) with 2-3 replicates each. Statistical analysis on IL-6 stimulation of T lymphoblasts (3 independent experiments) was performed using an unpaired two-tailed Student t-test on the means of the technical replicates. **P<0.01, n=3.

A

B

C

D

haematologica | 2019; 104(3)

613


T. Shahin et al.

during an operation for left hip dislocation at the age of 4 years. He subsequently developed right ankle edema, followed by progressive difficulty in walking and, finally, knee edema. He did not experience pain, erythema, or any increase in acute-phase reactants. The degeneration of the joints was classified as destructive arthropathy. In addition, the patient had neurodevelopmental delay of 1-2

years at the age of 4 years. PP498L showed consistently elevated serum IgE levels (>2000 IU/mL) and eosinophilia (Table 1). Progressive decline in absolute B-cell counts and low central memory T cells were observed (Table 1 and Figure 1C). These findings indicated a high National Institute of Health Hyper IgE syndromes (NIH HIES) score of 57 (Online Supplementary Table S2), raising suspicion of

Table 2. Comparative analysis of clinical features presenting in patients with IL6ST variants compared to other patient cohorts with defined gene defects.

GP130 STAT3 ZNF341 DOCK8 CARD11 PGM3 defect HIES deficiency deficiency DN defect deficiency (n=2) (n=140)47,48 (n=19)24,25 (n=64)49 (n=12)50,51 (n=29)52–54

Gene defecta,b IL6ST Inheritance AR IgEc ↑ Eosinophilia + Eczema + NIH HIES score 40-57 (typical)d Infections Abscess -/+ Pneumonia ++ Sinusitis/otitis + Keratitis/ conjunctivitis ++ Candidiasis Viral infections Parenchymal lung abnormalities Bronchiectasis ++ Pneumatocele ++ Dysmorphia Prominent forehead ++ Cathedral palate Decidual teeth retention ++ Abnormal bone fractures ++ Craniosynostosis/ ++ abnormal skull shape Scoliosis ++ Allergies + Autoimmunity CNS involvement ++ Neoplasia Immunological features Immunoglobulins Normal (excluding IgE) Lymphopenia CD4 lymphopenia Th17 cells Low CD19+ B cells Normal/ low Switched-memory Normal/ B cells low

Loeys-Dietz syndrome (n=8)55

ERBIN (n=3)56

Wiskott ARPC1b Aldrich defect syndrome (n=6)58–60 (incidence: 1/250000)57

STAT3 AD ↑↑ + + >40

ZNF341 AR ↑↑ +/+ 11-62

DOCK8 AR ↑ + ++ 20-40

CARD11 AD ↑ + ++ ND

PGM3 AR ↑↑ + + 27-55

TGFBR1/TGFBR2 AD ↑↑ + ++ ND

ERBIN AD ↑ + + ND

WAS X-linked ↑ + + ND

ARPC1B AR ↑ + + ND

++ +++ + + ++ +

+ + + ++ -

+ + ++ + + +++

+ ++ + + +

++ ++ + + ++

++ ++ ++ -

++ + -

+ + + + +

+ + + +

++ ++

+ +

+ +

+ -

+ -

+ -

-

+ -

+ -

++ ++ ++ ++ +

+ + + + +

+ + + + -

+ +/-

+ -

++ -

-

-

-

++ + + + +

+/+ + +/-

+ +++ ++ ++ ++

+/+++ + -

+ + ++ ++ +

+ +++ -

++ + -

Normal

Normal / high IgG +/+/Low Normal

Normal / low +/+/Low Normal/ low Low

Normal / low Normal Normal/ low Normal/ low

Normal

Normal

Normal

Normal Normal

Normal Normal

Normal

Normal

Low Normal/ low Low

Low

+ + Normal Normal/ low Normal/ low

+ ++ ++

+ -

Variable / high IgA +/+/Normal Normal/ low Normal/ low

Normal / high IgA -/+ ND Normal Normal

Apart from STAT3 HIES, DOCK8 deficiency and WAS, the phenotypical comparisons have been made based on a limited number of published cases. Future studies and larger cohorts of patients will be needed to describe the full phenotypic spectrum of the individual disease entities. bThere are additional gene defects that can present with high IgE, e.g. immune dysregulation polyendocrinopathy enteropathy X-linked (IPEX) syndrome due to FOXP3 defects.61 c↑↑ refers to IgE levels >5000 IU/mL; ↑ refers to IgE levels 10005000 IU/mL. dThe National Institute of Health Hyper-IgE syndromes (NIH HIES) score18 may be highly variable amongst individuals. The numbers indicated here represent ranges of values which are typically seen in the respective diseases; however, these should be taken with caution, in particular for those disease entities for which so far only a few individuals have been identified and described in the scientific literature. DN: dominant negative; ND: not determined; CNS: central nervous system. a

614

haematologica | 2019; 104(3)


IL6ST variants lead to aberrant T-cell phenotype

the autosomal dominant form of HIES, but capillary sequencing performed prior to whole exome sequencing did not reveal a mutation in STAT3.18 Family history was remarkable for the early death of 3 siblings (Figure 1A). The first female child (III-1 on Figure 1A) had scaphocephaly, foot deformities, recurrent diarrhea, respiratory infections, keratitis, and retarded growth and development. She died at the age of 3.5 years due to intestinal perforation. The second gestation resulted in twin brothers: III-2 was born prematurely at 24 weeks with a foot deformity and died 2 h after birth, whereas III-3 died in utero at 20 weeks. The third gestation was PP498Lâ&#x20AC;&#x2122;s older brother (III4) who has congenital blindness, and the fourth was PP498L. Patient PN404Y bearing a GP130N404Y mutation has been described previously.5 The marked similarity of clinical

A

B

C

D

phenotypes between PP498L and PN404Y is illustrated in Online Supplementary Table S3.

Identification of a novel IL6ST mutation We performed whole exome sequencing in PP498L to identify the underlying molecular disease etiology, and identified a homozygous missense mutation in IL6ST (c.1493C>T, p.P498L) (Figure 1D) deemed disease-causing based on functional predictions (Online Supplementary Table S4) and phenotypic similarity with the recently reported IL6ST-mutant (p.N404Y) patient, PN404Y.5 These amino acid positions are within the fifth and fourth domains of GP130, respectively, forming crucial interactions with other residues to maintain the acute bend in the

E

F

haematologica | 2019; 104(3)

Figure 3. Functional assessment of GP130P498L variant in fibroblasts. (A-E) Dose-escalation curves showing relative mean fluorescence intensity (rMFI) of p-STAT3 after stimulation of PP498L and healthy donor (HD) fibroblasts as well as PP498L fibroblasts transduced with wild-type (WT) GP130 with (A) IL-6, (B) IL-11, (C) IL-27, (D) LIF, (E) OSM, following overnight starvation in serum-free media. Bar graphs (right) showing rMFI of fibroblasts upon stimulation with the highest concentration of the corresponding cytokine: 3 (HD), 6 (p.P498L) and 4 (p.P498L + WT GP130) replicates of 3 independent experiments; Mann-Whitney-test; *P<0.05, **P<0.01. (F) Percentage of p-STAT4 assessed in PP498L and HD-derived fibroblasts after stimulation with LIF (100 ng/mL) following a 3-hour starvation in serumfree media. Three independent experiments were performed; the shapes of the symbols represent the individual experiments.

615


T. Shahin et al.

stimulates T lymphoblasts) was aberrant in PP498L (Figure 2C and D). In addition, we tested the effect of the new p.P498L substitution on the activation of other STAT family transcription factors in T lymphoblasts. Stimulation with IL-6 mainly activated STAT3 in healthy donors with no compensatory increase in activation of STAT1 in the patient cells (Online Supplementary Figure S1A), whereas phosphorylation of STAT1, STAT3 and STAT4 was abolished in patient T lymphoblasts upon stimulation with IL-27 compared to healthy donor-derived cells (Online Supplementary Figure S1B). Activation of STATs by GP130-independent cytokines including IL-4, IL-21 and IFNb was unaffected in patient cells, except for STAT4 which showed increased phosphorylation upon IL-12 stimulation of PP498L T lymphoblasts (Online Supplementary Figure S1C-F). To further evaluate the spectrum of mutation-dependent signaling defects, we analyzed p-STAT3 responses in fibroblasts from PP498L and a healthy donor, after overnight starvation. IL-6 or IL-11 stimulation demonstrated significantly reduced p-STAT3 levels (Figure 3A and B), with the aberrant IL-11 signaling likely to underlie the majority of bone manifestations in PP498L.14 OSM stimulation resulted in a partial and statistically significant reduction in p-STAT3

protein structure that is conserved across species (Figure 1E and F).19 We found GP130 protein expressed in PP498L fibroblasts, albeit at lower levels (Figure 1G), and in a CRISPR-engineered HEK293-IL6ST knockout (KO) cell line with transient overexpression of the GP130N404Y mutant.5

Functional assessment of GP130P498L mutation in primary and patient-derived cells Phosphorylation of STAT3 (p-STAT3) is a direct downstream effect of GP130 activation. To assess the impact of the p.P498L substitution, we studied STAT3 phosphorylation in primary T cells from PP498L and observed markedly decreased p-STAT3 levels upon stimulation with IL-6 and IL-27 compared to stimulation with IL-10 or IL-21 and to healthy donors (Figure 2A). Our previous study5 on PN404Y showed that IL-6 signaling had been abolished but a smaller reduction in p-STAT3 after stimulating primary T cells with IL-27 (CD3+, CD4+, and CD8+ T cells) (Figure 2B), indicating a mutation-dependent effect on the severity of downstream signaling through selected cytokines. Furthermore, IL-6 signaling was shown to be defective in both EBV-LCLs and T lymphoblasts derived from PBMCs of PP498L, and IL-27 (that

A

B

C

D

E

Figure 4. Functional assessment of GP130P498L variant in GP130-KO HEK293 cell line. (A-E) Relative mean fluorescence intensity (rMFI) of p-STAT3 in GP130 CRISPRknockout HEK293 cells that were transfected with a plasmid coding for the GP130P498L (GP130 KO + p.P498L), wild-type GP130 (GP130 KO + WT GP130) or transfected with the empty plasmid (GP130 KO + eV), after stimulation with (A) IL-6, (B) IL-11, (C) IL-27, (D) LIF, (E) OSM. From left to right: dose-escalation curves, stacked histograms displaying shifts in p-STAT3 signals, and bar graphs showing rMFI of fibroblasts upon stimulation with the highest concentration of the corresponding cytokine. (6 replicates of 3 independent experiments are shown; Wilcoxon matched-pairs signed rank test; *P<0.05.)

616

haematologica | 2019; 104(3)


IL6ST variants lead to aberrant T-cell phenotype

responses in the patient-derived fibroblasts, whereas reductions in p-STAT3 upon LIF and IL-27 stimulation were not significant (Figure 3C-E). Unlike IL-27, LIF, and OSM, upon receptor binding, IL-6 and IL-11 form a hexameric complex requiring two GP130 proteins, possibly explaining the extremity of their disrupted signaling. Moreover, PP498L’s fibroblasts showed a reduced p-STAT1 response upon IL-27 or OSM treatment compared to healthy donor fibroblasts (Online Supplementary Figure S2). In addition to STAT3, stimulation of fibroblasts with LIF induces phosphorylation of STAT4.20 Interestingly, in PP498L fibroblasts, LIF-induced pSTAT4 was slightly reduced when cells were starved in

A

D

B

serum-free media for 3 h (Figure 3F). To demonstrate causality of the novel IL6ST mutation on the observed phenotype, we ectopically expressed WT-GP130 in GP130P498L fibroblasts. This rescued the defects in the IL-6 and IL-11 signaling pathways and compensated the partial reductions in LIF, IL-27 and OSM signaling (Figure 3A-E). Finally, overexpression of GP130P498L in HEK293-IL6ST KO cells mirrored the defects observed in patient fibroblasts with significant reductions upon stimulation with IL-6, IL-11, IL-27 and LIF, while STAT3 phosphorylation in response to OSM was reduced but did not reach statistical significance (P=0.0625) (Figure 4A-E).

C

E

F

Figure 5. Phenotypic characterization of CD4+ and CD8+ T-cell compartments. (A and D) Representative dot-plot and bar graph summary showing percentages of (A) CD4+ and (D) CD8+ naïve (CD45RA+CCR7+) and memory [CD45RA–CCR7+/– including CD45RA+CCR7– terminally differentiated effector memory (TEMRA)] cells shown as frequency of live CD3+CD4+CD25– or CD3+CD8+CD25– T cells, respectively. (B and E) Bar graph summary showing CXCR3+ and CCR6+ frequencies of live (B) CD3+CD4+ and (E) CD3+CD8+ cells, and as frequencies of live (B) CD3+CD4+CD25– and (E) CD3+CD8+CD25– memory T cells. (C and F) t-Distributed Stochastic Neighbor Embedding (TSNE)-based analysis was performed on the following parameters: CD45RA, CD25, CD127, CCR4, CCR6, CCR7, CCR9, CCR10, CXCR3, CXCR5 and CRTh2. Overlaid heatmap statistics indicate median CCR6 expression in (C) live CD3+CD4+CD8– and (F) CD3+CD8+CD4– T cells. Bar graph summaries: mean+Standard Deviation: healthy donor (HD) (adult): n=19, HD age-matched controls (9-14 years): n=10-11, HD age-matched controls (6-7 years): n=69, PP498L: n=5 independent replicates from peripheral blood mononuclear cells (PBMCs) isolated at 3 distinct time points, seven and four months apart, PN404Y: n=3 replicates from 2 independent experiments and PBMCs taken two months apart. Mann-Whitney test, *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001. Some HD control data shown have been published previously.5

haematologica | 2019; 104(3)

617


T. Shahin et al.

Effects of IL6ST mutations on T-cell differentiation and function As cells from PP498L and PN404Y exhibited particularly aberrant IL-6 and IL-27 signaling, we hypothesized an impact on T-cell differentiation. Despite a history of recurrent infections, both patients had rather reduced CD4+ memory T cells and lower CCR6 expression in CD4+ memory T cells and total CD4+ T cells (particularly in PP498L; P<0.05) (Figure 5A and B and Online Supplementary Figure S3A and B).5 However, CXCR3 expression was reduced only in CD4+ T cells of PP498L, while CD4+ memory T cells from PN404Y showed increased expression of CCR4 and CRTh2. To investigate changes in Th-cell phenotypes, we analyzed chemokine receptor-expression linked to Th-cell homing and function.21,22 t-Distributed Stochastic Neighbor Embedding (TSNE) revealed that predominantly CD4+CCR6+, but less so CD4+CCR6– Th-cells, were reduced in both patients (Figure 5C). Moreover, although both patients showed normal numbers of total CD8+ memory T cells (Figure 5D), CCR6 expression within CD8+ memory T cells was reduced (particularly in PP498L; P<0.05) (Figure 5E and F and Online Supplementary Figure S3C and D). Furthermore, both patients showed reduced frequencies of RORγt+ (particularly in PP498L; P<0.05) but not TBET+CD8+ T cells (Online Supplementary Figure S4A). The CD8+RORγt+ memory T cells co-expressed CCR6 and intermediate level of TBET (Online Supplementary Figure S4B and C). These results demonstrate that functional IL6 signaling in human T cells is required for normal development of both CCR6-expressing CD4+ and CD8+ T cells. In addition, PP498L showed reduced CD8+ CXCR3+ T-cell frequencies (Figure 5E). Both patients had normal frequencies of peripheral regulatory T cells (Online Supplementary Figure S5). However, in peripheral blood CD4+ memory cells, and more evidently in total CD4+ T cells, we noted reduced Th17enriched23 CCR6+ CCR4+ CXCR3– frequencies and lower Th1-enriched CCR6–CCR4–CXCR3+ frequencies only in PP498L (Figure 6A). On the other hand, Th2-enriched CCR6– CCR4+ CXCR3– frequencies were significantly increased in PN404Y. To confirm chemokine receptor-based enrichments, and to evaluate also the composition of T-cell phenotypes inside chemokine receptor-enriched compartments, we analyzed subset specific transcription factor expression at the protein level (Online Supplementary Figure S6). TBET, GATA3 and RORγt expression fell into normal ranges, with elevated expression of GATA3 inside Th2 and Th17-enriched subsets of PN404Y only (Online Supplementary Figure S7). Interestingly, at the functional level, PP498L showed reduced IL-17A-producing CD4+ memory T-cell numbers. Furthermore, PP498L showed reduced IFNγ-producing and low IL-10-producing (P=0.0667) CD4+ memory T-cell frequencies while IL-4 production was in the normal range (Figure 6B-E). Besides these aberrations in CD4+ T cells, PP498L had significantly less IFNγ-producing CD8+ memory T cells (Figure 6F and G) indicating that, in vivo, the IL-27-specific signaling defects shown to be more severe in PP498L affect both CD4+ and CD8+ T-cell composition and effector function.

Discussion In this study, we show an intriguingly high degree of phenotypic similarity between the clinical manifestations 618

exhibited by 2 unrelated patients with cytokine selective IL6ST-loss of function mutations. Both patients present with elevated IgE, eosinophilia, recurrent infections (including invasive infections, severe lung pathology, and keratitis), and skeletal abnormalities (abnormal skull form/craniosynostosis and scoliosis). This is strikingly reminiscent of key features of HIES due to STAT3 variants or the recently studied ZNF341 deficiency,24,25 whereas they are less in common with other forms including DOCK8-deficiency26,27 and other deficiencies leading to high levels of serum IgE28 (Table 2). While STAT3 is a downstream transcription factor for several signaling pathways, including IL-10, IL-21 and IL-23, the mutations in IL6ST encoding GP130, constrict the defect to selected signaling pathways and suggest that defects in IL-6 and IL11 signaling dominate the HIES phenotype, whereas additional defects in IL-27, LIF and OSM signaling might contribute to the individual immunopathology. The two mutations affecting GP130 (p.P498L and p.N404Y) are 95 amino acids apart, yet both are on the highly conserved membrane-proximal ectodomain of the protein, known to play a crucial role in signal transduction and downstream JAK activation.29-31 Both variants show stable GP130 surface expression, preserving a partially intact quaternary protein structure, allowing GP130 to bind its ligands, and maintaining downstream signaling of cytokines such as LIF. The effect of individual GP130 variants may depend on various factors, including the genetic variant itself, the cytokine levels, as well as GP130 expression and receptor/co-receptor stochiometries in different cell types, explaining the cell type specific responses. We have addressed these factors by plasmid transfection combined with cytokine stimulation assays that cover a range of concentrations, as well as by comparing diverse types of primary immune cells. We found that the GP130P498L variant has a significant impact on IL-6, IL-11, IL-27 and LIF signaling in the HEK293 transfection assay and impaired downstream signaling upon stimulation with IL-6 and IL27 in primary T cells, T lymphoblasts and EBV-LCLs, as well as IL-6 and IL-11 in fibroblasts. Yet, in fibroblasts, signaling upon OSM (and to some extent IL-27 and LIF) stimulation was reduced but not completely abrogated. STAT1 that was phosphorylated in response to IL-27 and OSM in healthy donor fibroblasts showed decreased phosphorylation in PP498L fibroblasts. Furthermore, in response to IL-27, STAT1, STAT3 and STAT4 phosphorylation was compromised in T lymphoblasts of PP498L, demonstrating a general defect in signal transduction. IL-6 is a key cytokine responsible for the activation and differentiation of both T and B cells, as well as pro-inflammatory cues, including the acute-phase response.12 Therefore, the absence of an IL-6 response in PP498L explains his high susceptibility to pulmonary infections with no fevers. In contrast, aberrant IL-11 signaling underlies the observed bone manifestations in both patients with supporting evidence from IL11RA-deficient patients that present with craniosynostosis and delayed tooth eruption.14,15 IL-27 promotes differentiation of CD4+ T cells towards the Th1 phenotype, promotes IFNγ+IL-10+FOXP3– T-helper cell differentiation, and enhances CD8+ T-cell responses by increasing proliferation and effector functions such as IFNγ production and cytolytic activity.13,32-36 Defects in these immune functions were particularly seen in PP498L who had completely aberrant IL-27 signaling in primary T haematologica | 2019; 104(3)


IL6ST variants lead to aberrant T-cell phenotype

A

B

D

E

F

G

H

haematologica | 2019; 104(3)

I

C

Figure 6. Functional characterization of CD4+ and CD8+ T cells with transcription factor expression analysis. (A) Flow cytometry analysis of frequencies of Th-cell subtype-enriched compartments based on chemokine receptor expression as percent of live CD3+CD4+CD25– T cells or CD3+CD4+CD25– memory T (Th1-enriched), cells: CXCR3+CCR4–CCR6– (Th2-enriched) and CCR4+CXCR3–CCR6– + + – CCR6 CCR4 CXCR3 (Th17-enriched). (B and C) Dotplot presentation (B) and summary (C) of intracellular cytokine staining (ICCS) for IFN-γ and IL-17A shown as percent of live CD3+CD4+CD25– memory T cells from PP498L. (D and E) Dot-plot presentation (D) and summary (E) of ICCS for IL-4 and/or IL-13 shown as percent of live CD3+CD4+CD25– memory T cells from PP498L and PN404Y. (F and G) Dot-plot presentation (F) and summary (G) of ICCS for IFN-γ and/or IL-10 shown as percent of live CD3+CD4+CD25– memory T cells from PP498L and PN404Y (H and I) Dot-plot presentation (H) and summary (I) of ICCS for IFN-γ shown as percent of live CD3+CD8+CD25– memory T cells from PP498L and PN404Y. Bar graph summaries: mean+Standard Deviation (SD): healthy donor (HD) (adult): n=19, HD agematched controls (9-14 years): n=7-11, HD agematched controls (6-7 years): n=6-9, PP498L: n=5 independent replicates from peripheral blood mononuclear cells (PBMCs) isolated at 3 distinct time points, seven and four months apart, PN404Y: n=3 replicates from 2 independent experiments and PBMCs taken five months apart. Mann-Whitney test; *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001. Some HD control data shown have been published previously.5

619


T. Shahin et al.

cells and T lymphoblasts compared to PN404Y who had some retained signaling. To investigate the effects of aberrant GP130 signaling pathways on T cells, we aimed to characterize T-cell phenotype and function in PP498L and PN404Y. Both patients had low CD8+CCR6+ T cells. CD8+CCR6+ T cells co-expressed RORγt and intermediate level of TBET reminiscent of non-conventional CD8+ T cells, including mucosal-associated invariant T cells (MAIT cells) that are enriched in the CD161+ and CCR6+ fraction of CD8+ T cells and express a high level of RORγt but an intermediate level of TBET.37-39 These data support previous observations of the critical role of STAT3 signaling for the development of non-conventional T cells including MAIT cells.36 In addition, both patients presented with low CCR6+CCR4+CXCR3– Th17enriched T cells and IL-17A production that were more prominent in PP498L, possibly explaining the development of a tongue fungal lesion only in PP498L. Reminiscent of low Th17 cells and IL-17A production in HIES patients,24,40-42 these findings highlight the importance of IL-6 in the development of CCR6-expressing and IL-17A-producing human Th-cells despite the presence of functional IL-1b, IL-23 and IL-21 signaling pathways that are critical for human Th17 differentiation.43,44 While reduced CD8+CCR6+ T cell and Th17 cell frequencies are common to both patients, we also identified non-shared phenotypic aberrations. PN404Y showed increased Th2 frequencies similar to that observed in STAT3 and DOCK8-deficient patients. The PN404Y patient also presented with greater GATA3 expression in both Th2 and Th17-enriched subsets, pointing towards a Th2-biased polarization to the detriment of classical Th1 and Th17 cells. On the other hand, PP498L showed lower CXCR3 expression with reduced CCR6–CCR4–CXCR3+ Th1-enriched T cells and low IFNγ production by CD4+ and CD8+ memory T cells. This can also be compared to STAT3 LOF patients with normal levels of Th1 cells with a defective IL-27/STAT3 axis but functional IL-27/STAT1 axis, leading to a normal IFNγ response upon IL-27 stimulation.45 Hence, we speculate that the characteristic T-cell features of PP498L might be due to complete loss of IL-27 signaling that has been shown to play a role in Th1 and CD8+ T-cell memory development and effector responses. In summary, by characterizing a novel mutation in

References 1. Grimbacher B, Holland SM, Gallin JI, et al. Hyper-IgE Syndrome with Recurrent Infections — An Autosomal Dominant Multisystem Disorder. N Engl J Med. 1999; 340(9):692-702. 2. Grimbacher B, Holland SM, Puck JM. Hyper-IgE syndromes. Immunol Rev. 2005; 203(1):244-250. 3. Holland SM, DeLeo FR, Elloumi HZ, et al. STAT3 Mutations in the Hyper-IgE Syndrome. N Engl J Med. 2007;357(16): 1608-1619. 4. Minegishi Y, Saito M, Tsuchiya S, et al. Dominant-negative mutations in the DNAbinding domain of STAT3 cause hyper-IgE syndrome. Nature. 2007;448(7157):10581062. 5. Schwerd T, Twigg SRF, Aschenbrenner D, et

620

IL6ST and comparing phenotypic and functional features of patients with two independent mutations, we define IL-6 and IL-11 signaling as the dominant defects in GP130/STAT3 HIES. Our data suggest that a shared GP130-STAT3 signaling module is the functional basis of the striking phenotypic similarities of patients with pathogenic IL6ST and STAT3 variants. IL-6 signaling plays a critical role in vivo in the development of human CD4+CCR6+ helper T cells including Th17 and contributes to the development of human CD8+CCR6+ T cells. The discovery of two IL6ST non-synonymous mutations among hundreds of patients with overlapping immune and skeletal problems supports a model whereby only a limited number of non-synonymous combinatorial defects that affect selected signaling cascades are viable, despite severe systemic pathology due to embryological effects, but have sufficient pathogenicity to drive an immunopathology of combined immunodeficiency, elevated IgE levels, and skeletal anomalies. Acknowledgments We are grateful to all patients, their families and the healthy donors who have given their blood samples for this study. We also thank the nurses including Feride Özkan, Meliha Erol. We are grateful that Tatjana Hirschmugl and Raúl Jiménez Heredia and Jasmin Dmytrus performed exome sample preparation and assisted in data interpretation. We acknowledge the contribution of the Oxford Gastrointestinal Illness Biobank, which is supported by the NIHR Biomedical Research Centre, Oxford. Funding Supported by an ERC Starting grant (agreement 310857), a doctoral fellowship program (Cell Communication in Health and Disease (CCHD) from the Medical University of Vienna (both to KB), the contribution of the Oxford Gastrointestinal biobank, which is supported by the NIHR Oxford Biomedical Research Centre, the Leona M. and the Harry B. Helmsley Charitable Trust (to HHU), the Department of Health, UK, Quality, Improvement, Development and Initiative Scheme (grant to AOMW), the Wellcome Trust (project grant 093329 to AOMW and SRFT), Investigator Award 102731 and grant 090532/Z/09/Z supporting the Wellcome Trust Centre for Human Genetics both to AOMW and the Deutsche Forschungsgemeinschaft (grant SCHW1730/1-1 to TSc).

al. A biallelic mutation in IL6ST encoding the GP130 co-receptor causes immunodeficiency and craniosynostosis. J Exp Med. 2017;214(9):2547-2562. 6. O’Shea JJ, Plenge R, Mohseni M, et al. JAK and STAT signaling molecules in immunoregulation and immune-mediated disease. Immunity. 2012;36(4):542-550. 7. Yoshida K, Taga T, Saito M, et al. Targeted disruption of gp130, a common signal transducer for the interleukin 6 family of cytokines, leads to myocardial and hematological disorders. Proc Natl Acad Sci U S A. 1996;93(1):407-411. 8. Betz UA, Bloch W, van den Broek M, et al. Postnatally induced inactivation of gp130 in mice results in neurological, cardiac, hematopoietic, immunological, hepatic, and pulmonary defects. J Exp Med. 1998; 188(10):1955-1965.

9. Kopf M, Baumann H, Freer G, et al. Impaired immune and acute-phase responses in interleukin-6-deficient mice. Nature. 1994;368(6469):339-342. 10. van der Poll T, Keogh C V, Guirao X, Buurman WA, Kopf M, Lowry SF. Interleukin-6 gene-deficient mice show impaired defense against pneumococcal pneumonia. J Infect Dis. 1997;176(2):439444. 11. Puel A, Picard C, Lorrot M, et al. Recurrent staphylococcal cellulitis and subcutaneous abscesses in a child with autoantibodies against IL-6. J Immunol 2008;180(1):647-54. 12. Hunter CA, Jones SA. IL-6 as a keystone cytokine in health and disease. Nat Immunol. 2015;16(5):448-457. 13. Yoshida H, Hunter CA. The Immunobiology of Interleukin-27. Annu Rev Immunol. 2015;33:417–443.

haematologica | 2019; 104(3)


IL6ST variants lead to aberrant T-cell phenotype

14. Keupp K, Li Y, Vargel I, et al. Mutations in the interleukin receptor IL11RA cause autosomal recessive Crouzon-like craniosynostosis. Mol Genet Genomic Med. 2013;1(4):223-237. 15. Nieminen P, Morgan NV, Fenwick AL, et al. Inactivation of IL11 Signaling Causes Craniosynostosis, Delayed Tooth Eruption, and Supernumerary Teeth. Am J Hum Genet. 2011;89(1):67-81. 16. Escary J-L, Perreau J, Duménil D, Ezine S, Brûlet P. Leukaemia inhibitory factor is necessary for maintenance of haematopoietic stem cells and thymocyte stimulation. Nature. 1993;363(6427):361-364. 17. Masu Y, Wolf E, Holtmann B, Sendtner M, Brem G, Thoenen H. Disruption of the CNTF gene results in motor neuron degeneration. Nature. 1993;365(6441):27-32. 18. Grimbacher B, Schäffer AA, Holland SM, et al. Genetic linkage of hyper-IgE syndrome to chromosome 4. Am J Hum Genet. 1999; 65(3):735-744. 19. Xu Y, Kershaw NJ, Luo CS, et al. Crystal structure of the entire ectodomain of gp130: insights into the molecular assembly of the tall cytokine receptor complexes. J Biol Chem. 2010;285(28):21214-21218. 20. Nguyen HN, Noss EH, Mizoguchi F, et al. Autocrine loop involving IL-6 family member LIF, LIF receptor, and STAT4 drives sustained fibroblast production of inflammatory mediators. Immunity. 2017; 46(2):220-232. 21. Sallusto F, Lanzavecchia A. Heterogeneity of CD4+ memory T cells: functional modules for tailored immunity. Eur J Immunol. 2009;39(8):2076-2082. 22. Brodie T, Brenna E, Sallusto F. OMIP-018: Chemokine receptor expression on human T helper cells. Cytometry A. 2013; 83A(6):530-532. 23. Acosta-Rodriguez EV, Rivino L, Geginat J, et al. Surface phenotype and antigenic specificity of human interleukin 17–producing T helper memory cells. Nat Immunol. 2007;8(6):639-646. 24. Béziat V, Li J, Lin J-X, et al. A recessive form of hyper-IgE syndrome by disruption of ZNF341-dependent STAT3 transcription and activity. Sci Immunol. 2018;3(24): eaat4956. 25. Frey-Jakobs S, Hartberger JM, Fliegauf M, et al. ZNF341 controls STAT3 expression and thereby immunocompetence. Sci Immunol. 2018;3(24). 26. Zhang Q, Davis JC, Lamborn IT, et al. Combined Immunodeficiency Associated with DOCK8 Mutations. N Engl J Med. 2009;361(21):2046-2055. 27. Engelhardt KR, McGhee S, Winkler S, et al. Large deletions and point mutations involving the dedicator of cytokinesis 8 (DOCK8) in the autosomal-recessive form of hyperIgE syndrome. J Allergy Clin Immunol. 2009;124(6):1289-1302.e4. 28. Zhang Q, Boisson B, Béziat V, Puel A, Casanova J-L. Human hyper-IgE syndrome: singular or plural? Mamm Genome. 2018;29(7-8):603-617. 29. Kurth I, Horsten U, Pflanz S, et al. Importance of the membrane-proximal extracellular domains for activation of the signal transducer glycoprotein 130. J Immunol. 2000;164(1):273-282. 30. Hammacher A, Wijdenes J, Hilton DJ, Nicola NA, Simpson RJ, Layton JE. Ligandspecific utilization of the extracellular membrane-proximal region of the gp130-related signalling receptors. Biochem J. 2000;345 Pt

haematologica | 2019; 104(3)

1(Pt 1):25-32. 31. Timmermann A, Küster A, Kurth I, Heinrich PC, Müller-Newen G. A functional role of the membrane-proximal extracellular domains of the signal transducer gp130 in heterodimerization with the leukemia inhibitory factor receptor. Eur J Biochem. 2002;269(11):2716-2726. 32. Yoshimura T, Takeda A, Hamano S, et al. Two-sided roles of IL-27: induction of Th1 differentiation on naive CD4+ T cells versus suppression of proinflammatory cytokine production including IL-23-induced IL-17 on activated CD4+ T cells partially through STAT3-dependent mechanism. J Immunol. 2006;177(8):5377-5385. 33. Schneider R, Yaneva T, Beauseigle D, ElKhoury L, Arbour N. IL-27 increases the proliferation and effector functions of human naïve CD8+ T lymphocytes and promotes their development into Tc1 cells. Eur J Immunol. 2011;41(1):47-59. 34. Kamiya YS, Mizuguchi J, Morishima TN, et al. Generation with enhanced granzyme B T cell + augmentation of effector CD8 augmentation of effector CD8 T cell generation with enhanced granzyme B expression by IL-27 1. J Immunol. 2005; 17516861751693. 35. Owaki T, Asakawa M, Morishima N, et al. STAT3 is indispensable to IL-27-mediated cell proliferation but not to IL-27-induced Th1 differentiation and suppression of proinflammatory cytokine production. J Immunol. 2008;180(5):2903-2911. 36. Batten M, Kljavin NM, Li J, Walter MJ, de Sauvage FJ, Ghilardi N. Cutting edge: IL-27 is a potent inducer of IL-10 but not FoxP3 in murine T cells. J Immunol. 2008; 180(5):2752-2756. 37. Wilson RP, Ives ML, Rao G, et al. STAT3 is a critical cell-intrinsic regulator of human unconventional T cell numbers and function. J Exp Med. 2015;212(6):855-864. 38. Turtle CJ, Delrow J, Joslyn RC, et al. Innate signals overcome acquired TCR signaling pathway regulation and govern the fate of human CD161(hi) CD8 + semi-invariant T cells. Blood. 2011;118(10):2752-2762. 39. Dias J, Sandberg JK, Leeansyah E. Extensive phenotypic analysis, transcription factor profiling, and effector cytokine production of human MAIT Cells by flow cytometry. Methods Mol Biol. 2017;1514:241-56. 40. Ma CS, Chew GYJ, Simpson N, et al. Deficiency of Th17 cells in hyper IgE syndrome due to mutations in STAT3. J Exp Med. 2008;205(7):1551-1557. 41. Tangye SG, Pillay B, Randall KL, et al. Dedicator of cytokinesis 8–deficient CD4 + T cells are biased to a T H 2 effector fate at the expense of T H 1 and T H 17 cells. J Allergy Clin Immunol. 2017;139(3):933-49. 42. Keles S, Charbonnier LM, Kabaleeswaran V, et al. Dedicator of cytokinesis 8 regulates signal transducer and activator of transcription 3 activation and promotes TH17 cell differentiation. J Allergy Clin Immunol. 2016;138(5):1384-1394.e2. 43. Zhou L, Ivanov II, Spolski R, et al. IL-6 programs TH-17 cell differentiation by promoting sequential engagement of the IL-21 and IL-23 pathways. Nat Immunol. 2007;8(9): 967-974. 44. Korn T, Bettelli E, Gao W, et al. IL-21 initiates an alternative pathway to induce proinflammatory TH17 cells. Nature. 2007;448(7152):484-487. 45. Ma CS, Avery DT, Chan A, et al. Functional

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

STAT3 deficiency compromises the generation of human T follicular helper cells. Blood. 2012;119(17):3997-4008. Morbach H, Eichhorn EM, Liese JG, Girschick HJ. Reference values for B cell subpopulations from infancy to adulthood. Clin Exp Immunol. 2010;162(2):271-279. Chandesris M-O, Melki I, Natividad A, et al. Autosomal Dominant STAT3 Deficiency and Hyper-IgE Syndrome. Medicine (Baltimore). 2012;91(4):e1-e19. Gernez Y, Freeman AF, Holland SM, et al. Autosomal Dominant Hyper-IgE Syndrome in the USIDNET Registry. J allergy Clin Immunol Pract. 2018;6(3):996-1001. Engelhardt KR, Gertz ME, Keles S, et al. The extended clinical phenotype of 64 patients with dedicator of cytokinesis 8 deficiency. J Allergy Clin Immunol. 2015;136(2):402-412. Dadi H, Jones TA, Merico D, et al. Combined immunodeficiency and atopy caused by a dominant negative mutation in caspase activation and recruitment domain family member 11 (CARD11). J Allergy Clin Immunol. 2018;141(5):1818-1830.e2. Ma CA, Stinson JR, Zhang Y, et al. Germline hypomorphic CARD11 mutations in severe atopic disease. Nat Genet. 2017;49(8):11921201. Ben-Khemis L, Mekki N, Ben-Mustapha I, et al. A founder mutation underlies a severe form of phosphoglutamase 3 (PGM3) deficiency in Tunisian patients. Mol Immunol. 2017;90:57-63. Zhang Y, Yu X, Ichikawa M, et al. Autosomal recessive phosphoglucomutase 3 (PGM3) mutations link glycosylation defects to atopy, immune deficiency, autoimmunity, and neurocognitive impairment. J Allergy Clin Immunol. 2014; 133(5):1400-1409.e5. Sassi A, Lazaroski S, Wu G, et al. Hypomorphic homozygous mutations in phosphoglucomutase 3 (PGM3) impair immunity and increase serum IgE levels. J Allergy Clin Immunol. 2014;133(5):14101419.e13. Felgentreff K, Siepe M, Kotthoff S, et al. Severe eczema and Hyper-IgE in Loeys– Dietz-syndrome. Contribution to new findings of immune dysregulation in connective tissue disorders. Clin Immunol. 2014;150 (1):43-50. Lyons JJ, Liu Y, Ma CA, et al. ERBIN deficiency links STAT3 and TGF- pathway defects with atopy in humans. J Exp Med. 2017;214(3):669-680. Candotti F. Clinical Manifestations and Pathophysiological Mechanisms of the Wiskott-Aldrich Syndrome. J Clin Immunol. 2018;38(1):13-27. Somech R, Lev A, Lee YN, et al. Disruption of thrombocyte and T lymphocyte development by a mutation in ARPC1B. J Immunol. 2017;199(12):4036-4045. Kahr WHA, Pluthero FG, Elkadri A, et al. Loss of the Arp2/3 complex component ARPC1B causes platelet abnormalities and predisposes to inflammatory disease. Nat Commun. 2017;8:14816. Kuijpers TW, Tool ATJ, van der Bijl I, et al. Combined immunodeficiency with severe inflammation and allergy caused by ARPC1B deficiency. J Allergy Clin Immunol. 2017;140(1):273-277.e10. Barzaghi F, Passerini L, Bacchetta R. Immune dysregulation, polyendocrinopathy, enteropathy, x-linked syndrome: a paradigm of immunodeficiency with autoimmunity. Front Immunol. 2012;3:211.

621


ARTICLE Ferrata Storti Foundation

Stem Cell Transplantation

T-cell receptor-α repertoire of CD8+ T cells following allogeneic stem cell transplantation using next-generation sequencing Cornelia S. Link-Rachner,1,2 Anne Eugster,2 Elke Rücker-Braun,1 Falk Heidenreich,1,3 Uta Oelschlägel,1 Andreas Dahl,2,4 Christian Klesse,3 Matthias Kuhn,5 Jan Moritz Middeke,1 Martin Bornhäuser,1,2 Ezio Bonifacio2 and Johannes Schetelig1,3

Haematologica 2019 Volume 104(3):622-631

Medizinische Klinik und Poliklinik I, Universitätsklinikum Carl Gustav Carus, TU Dresden; 2DFG Research Center for Regenerative Therapies Dresden, TU Dresden; 3 DKMS Clinical Trials Unit, Dresden; 4BIOTEChnology Center, TU Dresden and 5Institut für Medizinische Informatik und Biometrie (IMB), Medizinische Fakultät der TU Dresden, Germany 1

ABSTRACT

A

Correspondence: CORNELIA S. LINK-RACHNER cornelia.link@uniklinikum-dresden.de Received: June 22, 2018. Accepted: September 25, 2018. Pre-published: September 27, 2018. doi:10.3324/haematol.2018.199802 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/104/3/622

lloreactivity or opportunistic infections following allogeneic stem cell transplantation are difficult to predict and contribute to posttransplantation mortality. How these immune reactions result in changes to the T-cell receptor repertoire remains largely unknown. Using next-generation sequencing, the T-cell receptor alpha (TRα) repertoire of naïve and memory CD8+ T cells from 25 patients who had received different forms of allogeneic transplantation was analyzed. In parallel, reconstitution of the CD8+/CD4+ T-cell subsets was mapped using flow cytometry. When comparing the influence of anti-Tcell therapy, a delay in the reconstitution of the naïve CD8+ T-cell repertoire was observed in patients who received in vivo T-cell depletion using antithymocyte globulin or post-transplantation cyclophosphamide in case of haploidentical transplantation. Sequencing of the TRα identified a repertoire consisting of more dominant clonotypes (>1% of reads) in these patients at 6 and 18 months post transplantation. When comparing donor and recipient, approximately 50% and approximately 80% of the donors' memory repertoire were later retrieved in the naïve and memory CD8+ T-cell receptor repertoire of the recipients, respectively. Although there was a remarkable expansion of single clones observed in the recipients' memory CD8+ TRα repertoire, no clear association between graft-versus-host disease or cytomegalovirus infection and T-cell receptor diversity was identified. A lower TRα diversity was observed in recipients of a cytomegalovirus-seropositive donor (P=0.014). These findings suggest that CD8+ T-cell reconstitution in transplanted patients is influenced by the use of T-cell depletion or immunosuppression and the donor repertoire. Introduction

©2019 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

622

Allogeneic stem cell transplantation (SCT) remains an essential component in the treatment of hematologic malignancies such as acute leukemia, lymphoma and myelodysplastic syndrome.1,2 Despite advances in transplantation regimens and supportive care, alloreactivity following SCT remains difficult to predict, and a large proportion of post-transplantation mortality is due to severe graft-versus-host reactions.3,4 T cells play an important role in the initiation of graft-versus-host disease (GvHD),5 in which the T-cell receptor (TCR) is a key element in the generation of an immune response against presented antigens. The heterodimeric TCR consists of an alpha (TRα) and beta (TRb) chain. Quantitative analyses of the TCR repertoire by next-generation sequencing (NGS) allow a more sophisticated assessment of the adaptive immune system.6,7 The TCR repertoire has been proposed to modulate post-SCT outcomes. So far, TCR sequencing has largely been performed by sequencing the TRb in patients folhaematologica | 2019; 104(3)


T-cell receptors following SCT

lowing allogeneic SCT.8 For example, assessing the recovery of the TRb repertoire following allogeneic transplantation revealed a restricted TCR repertoire diversity in patients affected by viral infections.9 We have previously shown that the TRα repertoire becomes oligoclonal and is dominated by a few expanded clonotypes following cytomegalovirus (CMV) infections in transplanted patients.10 Furthermore, GvHD and the relapse of acute myeloid leukemia (AML) correlate with lower TRb diversity,11 and the expansion of individual TCR clonotypes was observed in GvHD patients in selected studies.12 Formation of the TCR repertoire starts within months of transplantation, resulting in a more donor-like repertoire after the first year following transplantation.13 Little is known about how TCR changes are influenced by the choice of transplant regimen, especially by different forms of T-cell depletion. One study revealed lower TRb diversity in patients receiving an in vitro T-cell-depleted stem cell graft than in patients who received a non-Tcell-depleted cord blood graft.9 GvHD prophylaxis using post-transplantation cyclophosphamide (PTCy) on day +3 following SCT is an established therapeutic option in patients receiving haploidentical transplantation.14 Furthermore, the application of antithymocyte globulin (ATG) prior to transplantation as part of the conditioning therapy has become a common procedure to prevent GvHD, especially in patients with a mismatched donor.15 However, there have been no studies comparing these different regimens of in vivo T-cell depletion (ATG or PTCy) and their impact on the TCR repertoire. Here we analyzed the TRα repertoire of naïve and memory CD8+ T cells in 25 patients following different forms of allogeneic transplantation. This study addressed the question of whether the recipient TRα repertoire is influenced by anti-T-cell therapies such as ATG or PTCy, and if there are differences in response to haploidentical transplantation between fully matched or mismatched donor transplants. Furthermore, we analyzed to what extent the donor TRα repertoire is transferred to the recipient. Finally, the correlation between TRα repertoire diversity and the clinical manifestation of GvHD or CMV reactivation were addressed.

T-cell depletion were recruited (SIB-noATG), and samples from the 5 patient donors were analyzed in parallel. Acute GvHD (aGvHD) was defined as GvHD diagnosed within the first 100 days following SCT. In contrast, chronic GvHD (cGvHD) was diagnosed in cases with GvHD after the first 100 days or the typical clinical presentation of cGvHD features.16 CMV reactivation was determined by detection of CMV virus load in the peripheral blood.

Immunophenotyping by flow cytometry Routine assessment of differential blood counts was used to define the engraftment of neutrophil leukocytes and reconstitution of whole lymphocytes. Samples for immunophenotyping were taken on day 60, day 120 and day 180 following transplantation. Twenty healthy stem cell donors were analyzed to define normal ranges (controls). CD4+ and CD8+ T cells were characterized according to the expression of CCR7 and CD45RA as naïve (CCR7+CD45RA+), central memory (CM, CCR7+CD45RA-), effector memory (EM, CCR7-CD45RA-), and terminally differentiated effector memory (TEMRA, CCR7- CD45RA+) T cells.17 Staining was performed using the following antibodies as previously described:18 CD45V500, CD3-PerCP-Cy5.5, CD8-APCH7, CCR7-FITC, CD45RAPE (all BD Biosciences, San Jose, CA, USA) and CD4-eFluor450 (eBioscience, San Diego, CA, USA) (Online Supplementary Figure S1). Immunophenotyping was performed using a BD FACSCanto II (BD Bioscience, San Jose, CA, USA).

Sorting for T-cell receptor-α sequencing For TRα sequencing, naïve and memory cells were sorted from CD45+CD3+CD8+ T cells as previously described18 on day 60 and day 180 following transplantation. Sorted memory cells contained all memory subsets (CM, EM, TEMRA). Both T-cell subsets were sorted to a purity of more than 99% as checked by flow cytometry after sorting. For naïve CD8+ T cells and memory CD8+ T cells, a maximum of 1,000,000 cells were sorted, with a mean of 133,590 and 808,726 sorted cells for naïve and memory CD8+ T cells, respectively.

Library preparation for T-cell receptor-α sequencing Library preparation was performed as previously described.10,18,19 The final PCR product contained the nucleotide sequence for the variable region of the TRα (V and J segments), including the complementary determining region (CDR3).

Methods Next-generation sequencing Patients Patients (n=25) and donors were recruited after obtaining written informed consent and the approval of the local ethical review board (EK-279072013). To qualify, patients needed to have received their first SCT for an underlying hematologic malignancy. Patients who suffered a relapse during the observation period were excluded from the study. All SCTs were performed at Dresden University Hospital. Patients' characteristics are shown in Table 1. Patients were stratified into different groups according to their SCT protocol. In the first group, 5 patients received matched unrelated donor transplants and ATG (UD-ATG) as an addition to conditioning chemotherapy. The second group contained 5 patients who received mismatched unrelated donor transplants (9/10 allele match) and ATG (mmUD-ATG). Group three contained 5 patients who received transplants from matched unrelated donors without the application of ATG (UD-noATG), whereas the fourth group (Haplo-PTCy) was made up of patients who underwent haploidentical transplantation and the use of PTCy. Lastly, 5 patients with matched related donors without any use of haematologica | 2019; 104(3)

Next-generation sequencing was performed using an Illumina HiSeq 2500 (Illumina, San Diego, CA, USA), generating 150-basepair reads. Extraction of CDR3 sequences was performed using MiTCR as previously described.19,20 MiTCR used error correction with the highest stringency. Sequencing was performed, calculating 20 reads per cell individually for each sample for normalization. Samples that failed quality control were excluded from the analyses and indexed as “not available”. TRα chains with identical CDR3 region amino acid (AA) sequences were defined as clonotypes. Clonotypes with a TRα read frequency of more than 1% of reads were defined as “dominant” and more than 10% as “highly dominant” clonotypes. Sequence data are available at VDJServer under project UUID 3544765015263285736-242ac11d-0001-012 (https://vdjserver.org).

Statistical analysis Data analysis was performed using GraphPad Prism (v.5.01, La Jolla, CA, USA), KNIME 2.5.2 and R (v.2.15.2 and Studio v.0.98.945, Boston, MA, USA). Immunophenotyping results were 623


C.S. Link-Rachner et al. Table 1. Patients’ characteristics. Group Patient Disease Age ID# ID#

Sex CMVStatus

UD-ATG UD-ATG UD-ATG UD-ATG UD-ATG mmUD-ATG mmUD-ATG mmUD-ATG mmUD-ATG mmUD-ATG UD-noATG UD-noATG UD-noATG UD-noATG UD-noATG Haplo-PTCy Haplo-PTCy Haplo-PTCy Haplo-PTCy Haplo-PTCy SIB-noATG SIB-noATG SIB-noATG SIB-noATG SIB-noATG

m m m m m f f m m f f m m m f m m f f m f f m m m

TCR_001 TCR_003 TCR_023 TCR_041 TCR_056 TCR_013 TCR_038 TCR_048 TCR_054 TCR_055 TCR_005 TCR_008 TCR_017 TCR_019 TCR_027 TCR_012 TCR_014 TCR_026 TCR_036 TCR_049 TCR_002 TCR_011 TCR_024 TCR_040 TCR_063

ALL T-ALL AML CLL AML B-ALL AML AML AML AML B-NHL CLL AML MDS AML AML AML MDS AML AML AML AML MDS T-NHL AML

47 34 54 71 68 63 49 63 68 63 61 73 46 64 72 62 21 65 51 65 53 50 63 49 59

Treatment status at SCT

Transplant Stem T-cell Conditioning regimen cell depletion regimen allele-matching source

UD-SCT 10/10 neg 1st CR pos 1st CR UD-SCT 10/10 st neg 1 CR UD-SCT 10/10 pos PR UD-SCT 10/10 pos 1st CR UD-SCT 10/10 pos 2nd CR UD-SCT 9/10 neg 1st CR UD-SCT 9/10 pos 1st CR UD-SCT 9/10 neg 1st CR UD-SCT 9/10 neg Induction failure UD-SCT 9/10 pos 2nd relapse UD-SCT 10/10 pos stable UD-SCT 10/10 st neg 1 CR UD-SCT 10/10 neg no response UD-SCT 10/10 pos 1st CR UD-SCT 10/10 neg > 2nd relapse Haplo-SCT pos Induction failure Haplo-SCT pos 2nd relapse Haplo-SCT neg 2nd CR Haplo-SCT pos Induction failure Haplo-SCT pos 1st CR SIB-SCT 10/10 st pos 1 CR SIB-SCT 10/10 pos no response SIB-SCT 10/10 pos 3rd relapse SIB-SCT 10/10 neg 1st relapse SIB-SCT 10/10

PBSC PBSC PBSC PBSC PBSC PBSC PBSC PBSC PBSC PBSC PBSC PBSC PBSC PBSC PBSC BM BM BM BM BM PBSC PBSC PBSC BM PBSC

ATG ATG ATG ATG ATG ATG ATG ATG ATG ATG no no no no no PTCy PTCy PTCy PTCy PTCy no no no no no

TBI 12 Gy/Eto TBI 12 Gy/Eto Flu/Bu 8 Flu/Bu 8 Flu/Treo Flu/TBI 8 Gy Bu/Cy Flu/TBI 2Gy Flu/Bu 8 Flu/Mel Flu/Bu 8 Flu/Bu 8 Flu/Bu 8 Flu/Bu 8 Flu/Bu 8 Flu/Cy/TBI 2Gy Flu/Cy/TBI 2Gy Flu/Cy/TBI 2Gy Flu/Cy/TBI 2Gy Flu/Cy/TBI 2Gy Flu/TBI 8Gy Flu/Bu 8 Flu/Bu 8 Bu/Cy Flu/Bu 8

Donor Donor Donor Sex CMV-Status Age m m m f m f m m m f f m m m m m f m m m f f f m f

neg pos neg pos pos pos pos pos pos neg neg neg neg neg pos pos neg pos neg pos pos neg neg neg pos

23 34 23 55 51 32 19 50 48 46 27 23 42 20 41 33 44 29 25 38 60 49 67 45 64

Post grafting immunosuppression Mtx, CsA Mtx, CsA Mtx, CsA CsA, MMF Mtx, CsA Mtx, CsA Mtx, CsA Mtx, CsA Mtx, CsA Mtx, CsA Mtx, CsA Mtx, CsA Mtx, CsA Mtx, CsA Mtx, CsA Tac, MMF Tac, MMF Tac, MMF Tac, MMF Tac, MMF CsA Mtx, CsA Mtx, CsA Mtx, CsA Mtx, CsA

ALL: acute lymphoblastic leukemia; T-ALL: T-cell acute lymphoblastic leukemia; B-ALL: B-cell acute lymphoblastic leukemia; AML: acute myeloid leukemia; CLL: chronic lymphocytic leukemia; MDS: myelodysplastic syndrome; NHL: non-Hodgkin lymphoma; T-NHL: T-cell non-Hodgkin's lymphoma; B-NHL: B-cell non-Hodgkin lymphoma; m: male; f: female; CR: complete remission; PR: partial remission; UD-SCT: unrelated donor stem cell transplant; Haplo-SCT: haploidentical donor stem cell transplant; SIB-SCT: identical sibling donor stem cell transplant; PBSC: peripheral blood stem cells; BM: bone marrow; ATG: antithymocyte globulin; PTCy: post-transplantation cyclophosphamide; TBI: total body eradication; Eto: etoposide; Flu: fludarabine; Bu: busulfan; Mtx: methotrexate; CsA: cyclosporine; Tac: tacrolimus; MMF: mycophenolate mofetil: pos: positive; neg: negative. .

analyzed using ANOVA and Dunnett's multiple comparison test. Simpson’s diversity index (Ds) was used as a general measure of diversity, as previously described; this index is known to be insensitive to differences in sample sizes.18,21 Differences in Ds between groups were tested with Wilcoxon tests or Kruskal-Wallis rank sum tests. P=0.05 was considered statistically significant for all tests.

Results Engraftment On day 60 following transplantation, 19 out of 25 patients had recovered leukocyte counts. The median day of neutrophil granulocyte engraftment (defined as 0.5x109 neutrophils/L) was day 20 post transplantation. In contrast, lymphocytes remained significantly below control levels on day 60 (P<0.001) and day 120 (P<0.01), and recovered to control levels only by day 180 (Online Supplementary Figure S2A). 624

Suppression of naïve T cells in patients with in vivo T-cell depletion Assessment of CD4+ T cells revealed suppressed numbers on days 60, 120 and 180 compared to controls (all P<0.001). Notably, patients who received ATG had significantly lower CD4+ counts on day 180 than those who did not receive ATG as part of their conditioning regimen (P<0.001). Analyses of CD4+ T-cell subsets revealed the sustained suppression of naïve T cells compared to controls (P<0.001). The lowest mean naïve CD4+ numbers were found in both groups receiving ATG and PTCy at all time points (P<0.001). In all groups, the CD4+ T-cell compartment was dominated by CCR7–CD45RA– EM cells with significantly higher values (P<0.001) for patients than for controls (Figure 1A). In contrast to the low CD4+ counts in transplanted patients, the CD8+ T-cell numbers from all samples were increased on days 120 and 180 compared to those of controls (P<0.05 and P<0.01, respectively) (Online Supplementary Figure S2B). The proportion of haematologica | 2019; 104(3)


T-cell receptors following SCT

A

B

Figure 1. CD4+ and CD8+ T-cell subsets. Immune reconstitution of T-cell subsets of CD4+ T cells (A) and CD8+ T cells (B) on day 60 (d60), day 120 (d120) and day 180 (d180) post transplantation. Patient groups with 5 patients were analyzed separately and compared to the control population (C, n=20).

CCR7+CD45RA+ naïve CD8+ T cells was also suppressed within all groups compared to those of controls (P<0.001). Again, the lowest counts were seen in patients receiving ATG and PTCy, but these counts were not statistically significant. CD8+ T cells were primarily composed of EM cells and TEMRA cells for all groups. Both the EM and TEMRA fractions were comparable to or higher than those in the reference group, with differentiated cells becoming the dominant population over time (EM, P<0.01 on day 60; TEMRA, P<0.01 on day 120 and P<0.001 on day 180); no significant differences were seen between the transplantation groups (Figure 1B).

The recipients’ repertoires were heavily shaped by the memory repertoires of the donors. We found that 77.2% and 80.0% of the TRα reads from the donors' memory repertoires were populated by clonotypes shared by donors’ and patients’ memory repertoires on days 60 and 180, respectively. Notably, 41.5% (on day 60) and 61.0% (on day 180) of the donors’ memory repertoires were also found in the naïve repertoires of the recipients. In contrast, only 8.9% and 10.1% of the donors’ naïve repertoires were recovered in the recipients’ memory repertoires, and 6.0% and 18.6% were recovered in the recipients’ naïve repertoires on days 60 and 180, respectively (Figure 2A).

T-cell receptor-α repertoire composition and diversity are shaped by the memory T-cell receptor-α repertoire of the donor

Clinical context of T-cell receptor-α repertoire composition in SIB-noATG patients

We had access to blood samples from 5 donors at the time of transplant donation (Online Supple