Downloaded from genome.cshlp.org on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press Ordoñez&Kulis et al., Chromatin activation in multiple myeloma, 2nd revision, submitted in July 2020 1 Chromatin activation as a unifying principle underlying pathogenic 2 mechanisms in multiple myeloma 3 Raquel Ordoñez1,2*, Marta Kulis3,4*, Nuria Russiñol3, Vicente Chapaprieta5, Arantxa Carrasco- 4 Leon1, Beatriz García-Torre4, Stella Charampopoulou4, Guillem Clot2,4, Renée Beekman2,4, 5 Cem Meydan6, Martí Duran-Ferrer4, Núria Verdaguer-Dot4, Roser Vilarrasa-Blasi4, Paula 6 Soler-Vila4, Leire Garate1,7, Estíbaliz Miranda1,2, Edurne San José-Enériz1,2, Juan R. Rodriguez- 7 Madoz1, Teresa Ezponda1, Rebeca Martínez-Turrilas1, Amaia Vilas-Zornoza1, David Lara- 8 Astiaso1, Daphné Dupéré-Richer8, Joost H.A. Martens9, Halima El-Omri10, Ruba Y Taha10, 9 Maria J. Calasanz1,2, Bruno Paiva1,2,7, Jesus San Miguel1,2,7, Paul Flicek11, Ivo Gut12, Ari 10 Melnick6, Constantine S. Mitsiades13, Jonathan D. Licht8, Elias Campo2,3,4,5, Hendrik G. 11 Stunnenberg9, Xabier Agirre1,2**, Felipe Prosper1,2,7** and Jose I. Martin-Subero2,4,5,14**. 12 13 * Shared first authorship 14 ** Shared senior authorship 15 16 1. Centro de Investigación Médica Aplicada (CIMA), IDISNA, Pamplona, Spain. 17 2. Centro de Investigación Biomédica en Red de Cáncer, CIBERONC. 18 3. Fundació Clínic per a la Recerca Biomèdica, Barcelona, Spain. 19 4. Institut d’Investigacions Biomèdiques August Pi I Sunyer (IDIBAPS), Barcelona, Spain. 20 5. Departamento de Fundamentos Clínicos, Universitat de Barcelona, Barcelona, Spain. 21 6. Division of Hematology/Oncology, Department of Medicine, Weill Cornell Medical College, New York, USA. 22 7. Clínica Universidad de Navarra, Pamplona, Spain. 23 8. Division of Hematology/Oncology, University of Florida Health Cancer Center, Gainesville, USA. 24 9. Radboud Institute for Molecular Life Sciences, Nijmegen, Netherlands. 25 10. Department of Hematology & BMT, Hamad Medical Corporation, NCCCR, Doha, Qatar. 26 11. European Molecular Biology Laboratory, European Bioinformatics Institute (EMBL-EBI), Wellcome Trust 27 Genome Campus, Hinxton, UK. 28 12. CNAG-CRG, Centre for Genomic Regulation (CRG), Barcelona Institute of Science and Technology (BIST), 29 Barcelona, Spain. 30 13. Department of Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical School, USA. 31 14. Institució Catalana de Recerca i Estudis Avançats (ICREA), Barcelona, Spain. 32 33 Correspondence: [email protected] (Xabier Agirre), [email protected] (Felipe Prosper) and [email protected] 34 (Jose I Martin-Subero). 35 Word count: 36 ·Abstract: 175 (1,213 characters) 37 ·Manuscript: 4,127 (27,175 characters) 38 ·Methods: 694 (4,901 characters) 39 ·Figure legends: 1,386 (8,812 characters) 40 Figure Count: 4 41 Supplemental figures: 18 42 Supplemental tables: 9 43 1 Downloaded from genome.cshlp.org on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press Ordoñez&Kulis et al., Chromatin activation in multiple myeloma, 2nd revision, submitted in July 2020 44 ABSTRACT 45 Multiple myeloma (MM) is a plasma cell neoplasm associated with a broad variety of 46 genetic lesions. In spite of this genetic heterogeneity, MMs share a characteristic malignant 47 phenotype whose underlying molecular basis remains poorly characterized. In the present 48 study, we examined plasma cells from MM using a multi-epigenomics approach and 49 demonstrated that when compared to normal B cells, malignant plasma cells showed an 50 extensive activation of regulatory elements, in part affecting co-regulated adjacent genes. 51 Among target genes upregulated by this process, we found members of the NOTCH, NF-kB, 52 MTOR signaling and TP53 signaling pathways. Other activated genes included sets involved 53 in osteoblast differentiation and response to oxidative stress, all of which have been shown 54 to be associated with the MM phenotype and clinical behavior. We functionally 55 characterized MM specific active distant enhancers controlling the expression of thioredoxin 56 (TXN), a major regulator of cellular redox status, and in addition identified PRDM5 as a novel 57 essential gene for MM. Collectively our data indicates that aberrant chromatin activation is a 58 unifying feature underlying the malignant plasma cell phenotype. 59 60 Keywords. Multiple myeloma, epigenomics, chromatin. 61 62 2 Downloaded from genome.cshlp.org on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press Ordoñez&Kulis et al., Chromatin activation in multiple myeloma, 2nd revision, submitted in July 2020 63 INTRODUCTION 64 MM is an aggressive hematological neoplasm characterized by the uncontrolled expansion 65 and accumulation of malignant plasma cells (PCs) in the bone marrow (San Miguel 2014; Kumar et al. 66 2017). Given the central role of genomic alterations in cancer development, the study of molecular 67 mechanisms underlying MM pathogenesis has mostly focused on the genetic aberrations of these 68 tumors (Stratton et al. 2009). Such studies have revealed that MM patients are genetically 69 heterogeneous, without a single and unifying genetic event identified in all patients (Robiou du Pont 70 et al. 2017), but nevertheless the central paradox of MM remains, i.e. terminally-differentiated 71 plasma cells normally do not divide. This indicates that the essential aspects of the gene regulatory 72 networks within the malignant plasma cell are dysfunctional and suggests that epigenetic 73 deregulation of gene expression may be a root cause of the disease. Over the last decades, a greater 74 understanding of the role of histone, DNA and other chromatin modifications in the epigenetic 75 control of gene expression has transformed our understanding of transcriptional patterns in normal 76 and neoplastic cells (Allis and Jenuwein 2016). In spite of the multifaceted nature of the epigenome 77 (Stricker et al. 2017), the cancer epigenomics field has been mostly focused on the association 78 between DNA methylation and transcription (Esteller 2008). However, recent whole-genome DNA 79 methylation studies indicate that this association is less clear than previously appreciated (Kulis et al. 80 2012; Kretzmer et al. 2015; Kulis et al. 2015), and that understanding gene deregulation in cancer 81 requires the integrative analysis of various epigenetic marks including histone modifications and 82 chromatin accessibility (Martens et al. 2010; Ooi et al. 2016; Beekman et al. 2018). In MM, although 83 several reports have identified alterations in the DNA methylome of malignant plasma cells (Walker 84 et al. 2011; Heuck et al. 2013; Kaiser et al. 2013; Agirre et al. 2015), their pathogenic impact has been 85 uncertain and the chromatin regulatory network underlying aberrant cellular functions in MM has 86 just started to be characterized (Agarwal et al. 2016; Jin et al. 2018). 87 3 Downloaded from genome.cshlp.org on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press Ordoñez&Kulis et al., Chromatin activation in multiple myeloma, 2nd revision, submitted in July 2020 88 RESULTS 89 Integrative analysis of multiple epigenetic layers in MM 90 We performed an integrative analysis of multiple epigenetic layers in purified plasma cells 91 from 3 MM patients, including whole-genome maps of six core histone modifications (i.e. H3K27ac, 92 H3K4me1, H3K4me3, H3K36me3, H3K27me3, and H3K9me3) by ChIP-seq, chromatin accessibility by 93 ATAC-seq, DNA methylation by whole-genome bisulfite sequencing (WGBS) and gene expression by 94 RNA-seq. The major limiting factor for such multi-omics characterization was the high amount of 95 starting material needed from each individual patient. To overcome this limitation, we also analyzed 96 the chromatin regulatory landscape of an extended series of MM patients, including additional 97 profiles for H3K27ac (n=11), H3K4me1 (n=7), ATAC-seq (n=14) and RNA-seq data (n=37). Normal 98 control consisted of reference epigenomes from B cell subpopulations at different maturation stages, 99 including naive B cells (both from peripheral blood, pb-NBCs, and tonsils, t-NBCs), germinal center B 100 cells (GCBCs), memory B cells (MBCs) and tonsillar PCs (t-PCs) (Beekman et al. 2018) (Fig. 1A, 101 Supplemental Tables S1 and S2). Additionally, we were able to characterize the H3K27ac and 102 transcriptional profiles of bone marrow plasma cells (bm-PCs), the normal counterpart of MM, from 103 which a complete reference epigenome could not be generated due to the scarcity of this cell 104 subpopulation in healthy bone marrow. 105 Unsupervised principal component analysis indicated that MM displays a distinct epigenetic 106 configuration than normal B cell subpopulations, which is reflected in every single layer of the 107 epigenome (Fig. 1B). As B cell differentiation entails modulation of all studied epigenetic layers, we 108 focused our analysis on the genome fraction changing in MM but showing stable chromatin profiles 109 across the normal B cell maturation program. This strategy shall allow us to identify chromatin 110 changes specifically altered during myelomagenesis (Fig. 1C and Supplemental Fig. S1 and S2). 111 Overall, we observed that each histone modification undergoes more gains than losses in MM, but at 112 different degrees, being the H3K27ac, H3K4me1 and H3K4me3 regulatory marks those with the 4 Downloaded from genome.cshlp.org on September 26, 2021 - Published by Cold Spring Harbor Laboratory Press Ordoñez&Kulis et al., Chromatin activation in multiple myeloma,