@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix skos: . vivo:departmentOrSchool "Science, Faculty of"@en, "Microbiology and Immunology, Department of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Lorzadeh, Alireza"@en ; dcterms:issued "2020-01-13T16:47:19Z"@en, "2020"@en ; vivo:relatedDegree "Doctor of Philosophy - PhD"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description """Epigenetic modifications including reversible and non-uniform chemical marks to chromatin support activation and silencing of gene transcription. Alterations in normal epigenetic states are associated with transformation across a wide range of cancer types including myeloid malignancies. To understand the role of epigenetic regulation of normal human hematopoietic progenitors and its dysfunction in myeloid transformation, I developed a low-cell input chromatin immunoprecipitation method that, when combined with an analytical framework enables a simultaneous assessment of chromatin accessibility and histone modification state. This method enabled a comparative analysis of the epigenomic states of normal and malignant human blood cell compartments. Application of this methodology to highly purified, phenotypically defined subsets of primitive and terminally differentiating normal human cord blood cells showed that multiple human hematopoietic progenitor phenotypes display a common H3K27me3 signature. This signature includes many large organized H3K27me3 domains co-marked by H3K9me3 also found in the mature lymphoid cells in cord blood (CB) but not in co-isolated monocytes or erythroblasts. These results indicate a marked difference in the epigenomic changes primitive human neonatal hematopoietic cells undergo when they initiate terminal differentiation of the lymphoid and myeloid lineages. Further evidence that this differential H3K27me3 contraction directly impacts hematopoietic differentiation was obtained by manipulating H3K27me3 regulators in cell line models of inducible neutrophil differentiation in vitro. These methodologies were then used to explore epigenomic dysfunction found in the leukemic cells obtained from patients presenting with acute myeloid leukemia (AML) whose blasts differed in their content of neomorphic isocitrate dehydrogenase (IDH) mutations. Comparison of the methylation landscape in the AML cells with and without IDH mutations revealed a higher fractional DNA methylation level at active enhancers in the IDH mutant cells. However, there was no significant difference in global occupancy of histone modifications between the leukemic cells from the two patient groups. Collectively, these findings reveal previously unknown relationships of epigenetic modifications in normal and malignant human blood cells."""@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/73311?expand=metadata"@en ; skos:note """Epigenetic Regulation in Normal Hematopoiesis and Its Dysfunction in Leukemia by Alireza Lorzadeh B.Sc., McGill University, 2012 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Microbiology and Immunology) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) January 2020 © Alireza Lorzadeh, 2020 ii The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, the dissertation entitled: Epigenetic Regulation in Normal Hematopoiesis and Its Dysfunction in Leukemia submitted by Alireza Lorzadeh in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Microbiology and Immunology Examining Committee: Dr. Martin Hirst Supervisor Dr. Connie Eaves Supervisory Committee Member Dr. Carolyn Brown University Examiner Dr. LeAnn Howe University Examiner Additional Supervisory Committee Members: Dr. Ninan Abraham Supervisory Committee Member Dr. Aly Karsan Supervisory Committee Member iii Abstract Epigenetic modifications including reversible and non-uniform chemical marks to chromatin support activation and silencing of gene transcription. Alterations in normal epigenetic states are associated with transformation across a wide range of cancer types including myeloid malignancies. To understand the role of epigenetic regulation of normal human hematopoietic progenitors and its dysfunction in myeloid transformation, I developed a low-cell input chromatin immunoprecipitation method that, when combined with an analytical framework enables a simultaneous assessment of chromatin accessibility and histone modification state. This method enabled a comparative analysis of the epigenomic states of normal and malignant human blood cell compartments. Application of this methodology to highly purified, phenotypically defined subsets of primitive and terminally differentiating normal human cord blood cells showed that multiple human hematopoietic progenitor phenotypes display a common H3K27me3 signature. This signature includes many large organized H3K27me3 domains co-marked by H3K9me3 also found in the mature lymphoid cells in cord blood (CB) but not in co-isolated monocytes or erythroblasts. These results indicate a marked difference in the epigenomic changes primitive human neonatal hematopoietic cells undergo when they initiate terminal differentiation of the lymphoid and myeloid lineages. Further evidence that this differential H3K27me3 contraction directly impacts hematopoietic differentiation was obtained by manipulating H3K27me3 regulators in cell line models of inducible neutrophil differentiation in vitro. These methodologies were then used to explore epigenomic dysfunction found in the leukemic cells obtained from patients presenting with acute myeloid leukemia (AML) whose blasts differed in their content of neomorphic isocitrate dehydrogenase (IDH) mutations. Comparison of iv the methylation landscape in the AML cells with and without IDH mutations revealed a higher fractional DNA methylation level at active enhancers in the IDH mutant cells. However, there was no significant difference in global occupancy of histone modifications between the leukemic cells from the two patient groups. Collectively, these findings reveal previously unknown relationships of epigenetic modifications in normal and malignant human blood cells. v Lay Summary The mechanisms that control the production of normal blood cells at normal levels are known to be very complex and still poorly understood. However, it is clear that leukemia in humans involves a perturbation of this control system that results in the production of too many cells that cannot function properly. A key player in this control system is the DNA inside each cell and certain proteins that are bound to it to form a complex referred to as chromatin. The goal of this thesis was to understand how the composition of the chromatin of normal blood cell precursors changes during normal blood cell production and how this might be altered in certain types of human leukemia. This was previously difficult because the normal cells of interest were too rare to be studied. I was able to overcome this barrier by developing a more sensitive method and then use it to reveal novel molecular processes that regulate the production of normal and malignant blood cells. vi Preface The Introduction chapter is a version of a published review manuscript [Li, L., Lorzadeh, A., and Hirst, M. (2014). Regulatory variation: An emerging vantage point for cancer biology. Wiley Interdiscip. Rev. Syst. Biol. Med. 6, 37–59.]. It was written by me, Luolan Li and Dr. Martin Hirst. Chapter 3 represents a work that was published in Cell Reports in 2016 [Lorzadeh, A., Bilenky, M., Hammond, C., Knapp, D.J.H.F., Li, L., Miller, P.H., Carles, A., Heravi-Moussavi, A., Gakkhar, S., Moksa, M., et al. (2016). Nucleosome Density ChIP-Seq Identifies Distinct Chromatin Modification Signatures Associated with MNase Accessibility. Cell Rep. 17, 2112–2124.]. A detailed protocol of the method described in this chapter is published in the Journal of Visualized Experiments in 2017 [Lorzadeh, A., Gutierrez, R.L., Jackson, L., Moksa, M., and Hirst, M. (2017). Generation of native chromatin immunoprecipitation sequencing libraries for nucleosome density analysis. Journal of Visualized Experiments 2017.]. The Cell Reports manuscript was written by me, Drs. Hirst and Eaves with approval and input form all co-authors. The JoVE Manuscript was written by me, Michelle Moksa in Dr. Hirst’s lab and Dr. Hirst with approval and input form all co-authors. I designed the study, perfomed ChIP-seq and Sequential ChIP-seq experiments and analysed all data under Dr. Martin Hirst’s supervision. Anonymized CB cells used here were initially obtained with informed consent by Dr. Connie Eaves at the Terry Fox Laboratory (TFL) of BC Cancer according to protocols approved by the Research Ethics Board (REB) of the University of British Columbia (UBC) (certificate# H13-03536) and the Women and Children’s Hospital of British Columbia. All samples were from normal full-term deliveries. CB subset isolations were performed by technicians in the TFL and Dr. David Knapp and Colin Hammond in Dr. Eaves’ lab. H1 and H9 human embryonic stem cells (hESCs) were a gift from STEMCELL Technologies (STEMCELL). RNA sequencing (RNAseq) and whole vii genome bisulfite sequencing (WGBS) libraries were generated by the epigenomics group at Canada’s Michael Smith Genome Sciences Centre at BC Cancer. As a part of this Chapter, I developed a R package for quality control of ChIP-seq libraries using qPCR which is published in the Comprehensive R Archive Network (CRAN) repository Chapter 4 is based on work not yet submitted for publication. The CB cells used in that study were obtained and isolated as in Chapter 3. CB-derived and adult naive T-cells were obtained with informed consent according to UBC REB approved protocols and isolated by Zohreh Sharafian in Dr. Lavoie’s lab. RNA-seq, WGBS and ChIP-seq data for the T cells were generated by the epigenomics group at Canada’s Michael Smith Genome Sciences Centre of BC Cancer. Fangwu Wang in Dr. Eaves lab performed the CB cell culturing experiments. Michelle Moksa generated the RNA-seq libraries. Marcus Wong and Qi Cao in Dr. Hirst’s lab helped me generate the PBAL libraries. Qi Cao and Colin Hammond also assisted me in performing the HL-60 experiments and the parallel experiments performed with primary CB subsets were performed by Fangwu Wang in Dr. Eaves’ lab. I designed the study overall, performed the ChIP-seq experiments and data analysis under the supervision of Drs. Hirst and Eaves. The manuscript on which this chapter is based was written by myself, and Drs. Hirst and Eaves with approval and input from all co-authors. Chapter 5 is based on work not yet submitted for publication. Samples of consented, anonymized normal adult bone marrow (BM) were obtained by Dr. Eaves from the UBC REB-approved (certificate# H13-03536) Hematology Cell Bank and the desired subsets isolated by Colin Hammond. AML patients’ cells were similarly acquired from the UBC REB-approved (certificate# H04-61292 and H13-02687) Hematology Cell Bank of BC, isolated by Jeremy Parker and Deborah Deng in Dr. Aly Karsan’s lab and genotyped by the BC Cancer Centre for Clinical viii Genomics. Michelle Moksa and Marcus Wong generated RNA-seq and PBAL libraries, respectively. I designed the study, performed the ChIP-seq experiments and analyzed the data under the supervision of Dr. Hirst. ix Table of Contents Abstract ........................................................................................................................................... iii Lay Summary .................................................................................................................................. v Preface ............................................................................................................................................ vi Table of Contents ........................................................................................................................... ix List of Tables ................................................................................................................................. xv List of Figures ............................................................................................................................... xvi List of Symbols ............................................................................................................................. xiv List of Abbreviations .................................................................................................................... xix Acknowledgements .................................................................................................................... xxiv Chapter 1: Introduction ................................................................................................................ 1 1.1 Epigenetics .................................................................................................................. 1 1.2 DNA methylation ........................................................................................................ 3 1.2.1 Writers of DNA methylation: the DNMT family ................................................ 5 1.2.2 Readers of DNA methylation: the MBD family .................................................. 9 1.2.3 Erasers of DNA methylation: the TET family .................................................. 10 1.3 Histone modifications ................................................................................................ 11 1.3.1 Histone acetylation ............................................................................................ 15 1.3.2 Histone deacetylation ........................................................................................ 19 1.3.3 Histone methylation ........................................................................................... 21 1.3.4 Histone demethylation ....................................................................................... 25 1.4 Epigenetic Assays ...................................................................................................... 26 1.4.1 Chromatin immunoprecipitation sequencing .................................................... 26 x 1.4.2 Whole genome bisulfite sequencing .................................................................. 28 1.5 Epigenomic dysfunction in cancer ............................................................................ 29 1.5.1 DNA methylation in cancer ............................................................................... 29 1.5.2 Histone Modification in Cancer ........................................................................ 31 1.6 The hematopoietic system as a model to understand epigenetic regulation of normal tissue homeostasis and its malfunction in AML .................................................................... 35 1.7 Rationale and Specific Aims ..................................................................................... 39 Chapter 2: Materials and Methods ............................................................................................ 45 2.1 Cells ........................................................................................................................... 45 2.2 Low Input Native ChIP-seq ....................................................................................... 46 2.3 Nucleosome Density(nd) ChIP-seq Integrative Analysis .......................................... 47 2.4 ChIP-seq .................................................................................................................... 48 2.5 Sequential ChIP-seq .................................................................................................. 49 2.6 WGBS Library Construction ..................................................................................... 50 2.7 RNA-seq .................................................................................................................... 51 2.8 Ribo-depleted RNA-seq ............................................................................................ 52 2.9 Post Bisulfite Adaptor Ligation (PBAL) ................................................................... 53 2.10 HL60 treatment with ATRA, EPZ-6438 and GSK-J4 ............................................... 54 2.11 Quantitative measurement of CD11b+ HL60 cells ................................................... 54 2.12 Stromal cell-containing cultures of CB cells ............................................................. 55 2.13 Chapter 3 integrative analysis ................................................................................... 56 2.14 Chapter 4 integrative analysis ................................................................................... 57 2.15 Chapter 5 Integrative analysis ................................................................................... 58 xi Chapter 3: Nucleosome density ChIP-Seq identifies distinct chromatin modification signatures associated with MNase accessibility ......................................................................................... 62 3.1 Introduction ............................................................................................................... 62 3.2 Results ....................................................................................................................... 63 3.2.1 Immunoprecipitated DNA fragment sizes correlate with different patterns of histone modification .......................................................................................................... 63 3.2.2 Nucleosome density ChIP-seq (ndChIP-seq) classifies promoters based on nucleosome density ........................................................................................................... 71 3.2.3 A majority of bivalent promoters in CD34+ CB cells identified by ndChIP-seq are heterogeneous ........................................................................................................................ 72 3.2.4 MNase accessibility and histone modification status is associated with RNA expression .......................................................................................................................... 83 3.2.5 MNase accessibility reveals a bimodal relationship between H3K27me3 and CpG methylation states ...................................................................................................... 84 3.2.6 Promoter MNase accessibility profiles show cell type differences ................... 85 3.2.7 Human CD34+ CB cells and ESCs exhibit different MNase accessibility at CpG-island-containing promoters ..................................................................................... 90 3.2.8 Modeling chromatin states differences using nucleosome density data ............ 91 3.3 Discussion .................................................................................................................. 95 Chapter 4: Polycomb contraction differentially regulates terminal human hematopoietic differentiation programs ............................................................................................................ 98 4.1 Introduction ............................................................................................................... 98 4.2 Results ..................................................................................................................... 104 xii 4.2.1 Functionally distinct human hematopoietic progenitors share a common polycomb signature ......................................................................................................... 104 4.2.2 Terminal differentiation is differentially associated with a genome-wide contraction of H3K27me3 density ................................................................................... 110 4.2.3 H3K27me3 LOCKs are enriched in lamina-associated domains (LADs) ....... 116 4.2.4 H3K27me3 contraction is associated with reduced BMI1 expression ............ 121 4.2.5 Hematopoietic progenitor subsets share lineage-specific enhancers marked by H3K27ac 124 4.2.6 EZH2 inhibition differentially alters the production of different hematopoietic lineages consistent with their acquired epigenomic features. .......................................... 135 4.3 Discussion ................................................................................................................ 139 Chapter 5: Analysis of the function of neomorphic IDH mutations in human AML .............. 143 5.1 Introduction ............................................................................................................. 143 5.2 Results ..................................................................................................................... 147 5.2.1 CpG Island methylation signatures are not IDH mutant-dependent in AML cells ………………………………………………………………………………..148 5.2.2 Increase in genomic occupancy of histone modifications in AML is independent of a mutant IDH genotype ........................................................................... 157 5.2.3 Mutant IDH AML cells exhibit a unique hypermethylation pattern at non-promoter regulatory regions ............................................................................................ 157 5.2.4 The AML active enhancer landscape is poised in normal adult BM CD34+CD38- cells .......................................................................................................... 167 xiii 5.2.5 Heterochromatin in IDH mutant leukemic cells is influenced by co-occurring mutations. ........................................................................................................................ 173 5.2.6 AML cells retain a progenitor H3K27me3 structure ....................................... 176 5.3 Discussion ................................................................................................................ 179 Chapter 6: Conclusion ............................................................................................................. 182 6.1 Interpretation and significance ................................................................................ 182 6.1.1 ndChIP-seq is a robust assay for the simultaneous measurement of histone modification and nucleosome density in rare cell populations. ...................................... 182 6.1.2 Polycomb contraction differentially regulates lymphoid and myeloid differentiation .................................................................................................................. 183 6.1.3 DNA hypermethylation driven by IDH mutant is at intergenic regulatory regions ………………………………………………………………………………..186 6.2 Limitations and future directions ............................................................................. 188 6.2.1 Limitation of ndChIP-seq ................................................................................ 188 6.2.2 Identification of mechanisms regulating contraction of H3K27me3 in differentiated myeloid cells ............................................................................................. 189 6.2.3 TF switching events are mediated by hypermethylated active enhancers to lead to abnormal transcription profiles ................................................................................... 190 6.3 Concluding remarks ................................................................................................. 192 Bibliography ................................................................................................................................ 195 xiv List of Symbols W = Gaussian Mixture Distribution weight µ = Mean s = Standard Deviation xv List of Tables Table 1. DNA methylation proteins in cancer ............................................................................... 41 Table 2. Histone modification proteins in cancer .......................................................................... 42 Table 3. A table of cell types used in this thesis ........................................................................... 60 Table 4. Correlation (Pearson) of histone modifications ............................................................... 97 xvi List of Figures Figure 1. Cytosine methylation and demethylation. ........................................................................ 4 Figure 2. DNA Methylation and Transcriptional Regulation. ......................................................... 6 Figure 3. Histone modifications and transcriptional regulation. ................................................... 13 Figure 4. H3K27ac correlates with TF binding site, evolutionary conserved regions and active transcription. .................................................................................................................................. 17 Figure 5. Super-enhancer. .............................................................................................................. 20 Figure 6. ndChIP-seq from 10,000 cells shows concordance with ChIP-seq performed on 1 million cells. .................................................................................................................................. 65 Figure 7. MNase accessibility is chromatin state-dependent.. ...................................................... 67 Figure 8. MNase accessibility at ChromHMM derived chromatin states is modification dependent ....................................................................................................................................... 69 Figure 9. ndChIP-seq identifies heterogeneously marked promoters in individual cells within bulk populations ............................................................................................................................ 73 Figure 10. Gene promoters show reproducible fragment size distributions associated with distinct RNA expression and DNA methylation profiles. .......................................................................... 76 Figure 11. Nucleosome density profiles are non-overlapping and correlate between replicates...78 Figure 12. Bivalent promoters in ESCs and heterogeneous promoters marked with H3K27me3 and H3K4me3 in human cord blood CD34+ cells are enriched in developmental pathways ....... 81 Figure 13. Distinct patterns of MNase accessibility across cord blood CD34+ and ES cell types are not due to difference in MNase digestion ................................................................................ 86 Figure 14. Modified promoters in ESCs show distinct patterns of MNase accessibility. ............. 88 xvii Figure 15. H3K27me3-enriched promoters in human ESCs and CD34+ cord blood cells show reciprocal MNase accessibility profiles. ........................................................................................ 93 Figure 16. H3K27me3 signatures are shared across functionally distinct CD34+ hematopoietic progenitor populations ................................................................................................................. 100 Figure 17. Sorting strategy for hematopoietic populations profiled in this study ....................... 102 Figure 18. Each progenitor populations possesses a unique expression profile .......................... 105 Figure 19. H3K27me3 signal is stable at promoters across progenitor populations ................... 108 Figure 20. Myeloid differentiation is associated with genome-wide H3K27me3 contraction .... 111 Figure 21. H3K27me3 contraction during myeloid differentiation coincides with a loss of progenitor LOCKs ....................................................................................................................... 114 Figure 22. H3K27me3 LOCKs lost during myeloid differentiation are associated with LADs in CD34+ progenitor cells ............................................................................................................... 117 Figure 23. ChromHMM identified polycomb repressed regions are enriched in LADs in progenitor cells. ........................................................................................................................... 119 Figure 24. . BMI1 is expressed at a lower level in erythroid precursors, monocytes and ESCs compare to other blood cells ........................................................................................................ 122 Figure 25. A majority of active enhancers in mature myeloid cells are marked by H3K27ac in CD34+ progenitor cell populations ............................................................................................. 125 Figure 26. Total number of enhancers decreased in erythroid precursor, monocyte and differentiated lymphoid cells ....................................................................................................... 127 Figure 27. Lineage-specific enhancers are marked by H3K27ac in hematopoietic progenitor subsets .......................................................................................................................................... 130 xviii Figure 28. The majority of active enhancers first appearing in differentiated cell populations contribute to the establishment of super enhancers ..................................................................... 133 Figure 29. EZH2 inhibition arrests HL60 cell growth and lymphoid production in vitro ........... 136 Figure 30. IDH mutants show heterogenous expression profile. ................................................ 145 Figure 31. RNA expression does not segregate mutant IDH from wild type. ............................. 149 Figure 32. Hypermethylation in mutant IDH occurs mainly outside CGIs. ................................ 152 Figure 33. Genome wide occupancy of histone modifications are similar between mutant IDH and wild type. .............................................................................................................................. 155 Figure 34. IDH mutant have unique methylation phenotype at active enhancers. ...................... 159 Figure 35. Enhancer methylation is prominent in mutant IDH. .................................................. 162 Figure 36. IDH mutant AML methylated enhancers are enriched in PU.1 binding site. ............ 165 Figure 37. Leukemic transformation is accompanied by increase in active enhancers ............... 168 Figure 38. AML active enhancer landscape is poised in normal adult bone marrow. ................ 171 Figure 39. H3K27me3 density at promoters separates mutant IDH based on NPM1 and DNMT3A mutation status………………………………………………………………………174 Figure 40. Progenitor H3K27me3 LOCKs are retained in AML blasts………………………..177 Figure 41. TF switching at hyper-methylated enhancers in mutant IDH ..………………….....193 xix List of Abbreviations 2-OGDD 2-oxoglutarate-dependent dioxygenases 5caC 5-carboxylcytosine 5fC 5-formylcytosine 5hmC 5-hydroxymethylcytosine 5mC 5-methylcytosine A Adenine ADD ATRX-DNMT3-DNMT3L AML Acute myeloid leukemia ASH2L ATRA absent/small/homeotic-2 like All-trans retinoic acid BER Base excision repair BM Bone marrow BMI1 C B cell-specific Moloney murine leukemia virus integration site 1 Cytosine CB Cord blood CD Cluster of differentiation CGI CpG islands ChIP ChIP-seq Chromatin immunoprecipitation Chromatin immunoprecipitation sequencing CIMP CMML CGI methylator phenotype Chronic myelomonocytic leukemia xx CMP CREAM Common myeloid progenitors Clustering of genomic Regions Analysis Method D2HG D-2-hydroxyglutarate DMAP1 DNMT1-associated protein 1 DMR Differentially methylated regions DNA Deoxyribonucleic acid DNAP1 DNMT1-associated protein 1 DNMT DNA methyltransferases EED Embryonic ectoderm development ESC Embryonic stem cell ESCC Esophageal squamous cell carcinoma ETS E26 transformation-specific EZH2 EPO FACS FDR G Enhancer of zest 2 Erythropoietin Fluorescent activating cell sorting False discovery rate Guanine GATA1 GATA binding factor 1 GMP Granulocyte-macrophage progenitors GNAT GREAT Gcn5 related N-acetyltransferase Genomic region enrichment of annotation tool CREBBP CREB binding protein xxi H3K27ac Histone 3 Lysine 27 acetylation H3K27me3 Histone 3 Lysine 27 tri-methylation H3K36me3 Histone 3 Lysine 36 tri-methylation H3K4me1 Histone 3 Lysine 4 mono-methylation H3K4me3 Histone 3 Lysine 4 tri-methylation H3K9me3 Histone 3 Lysine 9 tri-methylation HAT Histone acetyl transferases HDAC HSC Histone deacetylases Hematopoietic stem cell IDH Iso-citrate dehydrogenase IP Immunoprecipitation JmjC Jumonji C KDM MACS2 Lysine demethylase Model-based analysis of ChIP-seq MBD Methyl-CpG-binding domain MDS Myelodysplastic Syndrome MeCP2 Methyl-CpG-binding protein 2 MEP Megakaryocyte-erythroid progenitors MLL Mixed-lineage leukemia MNase MPN Micrococcal Nuclease Myeloproliferative neoplasms ndChIP Nucleosome density Chromatin immunoprecipitation xxii NPM1 Nucleophosmin 1 NSD Nuclear receptor binding SET domain NuRD Nucleosome remodeling deacetylase PB Peripheral blood PBAL Post bisulfite adapter ligation PCAF P300/CBP-associated factor PCNA P-NML Proliferating cell nuclear antigen CD45highCD34highCD38midCD71-CD10- PRC1 Polycomb repressive complex 1 PRC2 Polycomb repressive complex 2 PWWP Pro-Trp-Trp-Pro qPCR quantitative polymerase chain reaction RbBP retinoblastoma-binding protein RFTS Replication foci targeting sequence RNA Ribonucleic acid SAH S-adenosyl homocysteine SAM S-adenosyl-L-methionine SET Su3-9, enhancer of zest and trithorax SETMAR SET domain and mariner transposase fusion protein SIRT Sirtuins SOM Self-organizing maps SUV39H1 Suppressor of variegation 3-9 homolog 1 xxiii SUZ12 Suppressor of zeste 12 SYMD2 T SET and MYND domain containing 2 Thymine TAL1 T cell acute lymphocytic leukemia protein 1 TCA Tricarboxylic acid cycle TCGA The Cancer Genome Atlas TDG Thymine DNA glycosylase TET Ten-eleven-translocation TF Transcription factor UHRF1 Ubiquitin like with PHD and ring finger domains 1 WDR5 WD repeat domain 5 α-KG α-ketoglutarate xxiv Acknowledgements I offer my enduring gratitude to the faculty, staff and my fellow students at UBC, my family and funding agencies who have supported me during my PhD training period. I would like to sincerely thank Dr. Martin Hirst, my PhD supervisor, for giving me the opportunity to complete my PhD training in his laboratory, for his unwavering support and guidance and his contagious enthusiasm for science. Thank you for being a great “Cheer Leader”. I would like to extend my gratitude to my thesis committee members, Drs. Connie Eaves, Aly Karsan and Ninan Abraham for their inputs and support during my training. Collaborations with Drs. Eaves and Karsan were critical to my research projects. I also would like to specially thank Dr. Eaves and the phenomenal cast of trainees in her lab for being so welcoming and helpful. It was an absolute privilege to work with Dr. Eaves. Special thanks to Colin for his effort and patience while working with me. Many thanks to amazing people at Hirst lab. Michelle for helping me with my experiments and travel arrangements. Our great technicians: Angela, Qi and Marcus for helping me with my experiments, Sherry for dealing with paperwork, and Misha and Ali for their help with data analysis. I owe special thanks to Anniack for helping me with data processing. Special shout out to all my amazing colleagues at the Hirst lab. I have always valued their input and I have learned a great deal from them, especially Gloria for her contributions and lifesaving scripts; Alice, Jasper, Alvin and Rashedul for great conversations in and outside the lab; and Linda for keeping me company on weekends with fond memories. I am most grateful for my amazing parents and their unconditional and unlimited love and support throughout my whole life. Their perseverance and resilience in the face of adversity has been a motivating force during difficult times for me. They will be forever my role models. xxv To my parents 1 Chapter 1: Introduction Heritable genetic information is encoded in deoxyribonucleic acid (DNA) (Avery et al., 1944; Gall, 2016). DNA is a double helix structure composed of two strands containing purine (Adenine and Guanine) and pyrimidine (Thymine and Cytosine) nitrogenous bases in addition to a sugar-phosphate backbone (Gall, 2016; Watson and Crick, 1953). Chromatin refers to the compact structure of DNA strands wrapped 1.6 times around histone octamer complexes known as nucleosomes (Li et al., 2014). Until the mid 20th century, the concept that an identical DNA sequence could be differentially interpreted to give rise to all cell types (embryonic and somatic) within a multicellular organism was unknown. However, a series of experiments in the mid-20th century showed that a phenotypically normal embryo resulted from the introduction of nuclei extracted from both embryonic (Gurdon et al., 1958) and somatic (Gurdon and Laskey, 1970) cells. These early experiments were followed up by series of similar findings (Hoppe and Illmensee, 1982; Illmensee and Hoppe, 1981) which collectively suggested that all cells in an organism (except for certain immune cells) share the same genetic information (Chan and Gurdon, 1996). This discovery raised an important question: given that all cells in an organism share the same genetic information, despite showing completely different phenotypes, how do cells selectively interpret the same genetic sequence? 1.1 Epigenetics In 1940s Conrad Waddington described the “epigenetic landscape” in Drosophila Melanogaster embryo development as a ball running down an uneven hill with saddle points representing differentiation decision points (Creighton and Waddington, 2006; Waddington, 2 2006). In the mid 20th century the relationship between the genome and development was not clear and epigenetic mechanisms were considered by some an unnecessary complication to genetics. In the late 20th century, depiction of the role of DNA methylation in X chromosome inactivation (Mohandas et al., 1981), the molecular and structural characterization of nucleosomes (Gall and Murphy, 1998; Kornberg and Thomas, 1974; Luger et al., 1997; Luzzati and Nicolaïeff, 1963), along with the identification of the relationship of histone acetylation in the activation of gene transcription (Grunstein, 1997) reignited interest in the role of chromatin structure and its covalent modification in transcriptional regulation. Epigenetics, in the broadest terms, describes mitotically heritable information beyond the DNA sequence itself. Mammalian genomes carry non-uniform chemical marks that are covalently bound to DNA bases or to chromatin proteins and thereby regulate local gene transcription activity, either positively or negatively (Boyes and Bird, 1991; Local et al., 2018; Müller, 1995; Tompkins et al., 2012). Examples of such modifications are the covalent modifications of DNA bases (e.g., DNA methylation and hydroxy-methylation) and RNA bases (e.g., RNA methylation and acetylation) and modifications of amino acids on histone tails (e.g., histone methylation and acetylation). With the exception of some specialized immune cells (i.e., B and T cells), the genome remains largely unaltered during development and during cell differentiation; whereas the epigenome undergoes significant alteration during these same processes (Tompkins et al., 2012). Environmental factors can significantly impact the epigenome of an organism as demonstrated in studies of toxicant exposure in non-pathological and tumour tissues (Kile et al., 2012; Pilsner et al., 2007; Ramirez et al., 2008). 3 1.2 DNA methylation DNA methylation at the fifth carbon of the cytosine ring (5mC) is the most comprehensively studied epigenetic modifications in vertebrates (Figure 1) (Boyes and Bird, 1991; Ehrlich et al., 1982; Tsumura et al., 2006; Watt and Molloy, 1988). The mammalian genome is reduced in CpG content due to the spontaneous deamination of methylated cytosine to thymine (Ooi and Bestor, 2008; Tahiliani et al., 2009). In normal somatic mammalian cells 60% to 80% of cytosines are methylated (5mC) when found in CpGs, and to a lesser extent in non-CpG contexts in specific cell types, such as neuronal cells (Lee et al., 2017; Rizzardi et al., 2019; Shirane et al., 2013; Smith and Meissner, 2013). Symmetrical CpG methylation is not essential for cell viability but is required for differentiation and plays important roles in the regulation of gene transcription, maintenance of genome stability, genetic imprinting, repression of endogenous retroviral transcripts, as well as in X chromosome inactivation (Encode Consortium et al., 2013; Tsumura et al., 2006). Regions of the genome that have retained the ancestral CpG density, called CpG islands (CGIs), are largely unmethylated at the majority of the promoters of the genes with which they are associated (Deaton and Bird, 2011). In general, methylation of CpGs in promoter regions is associated with repression of transcription initiation through the inhibition of transcriptional activators, or through the recruitment of other repressive regulatory complexes (Myant et al., 2011; Watt and Molloy, 1988; Wiench et al., 2011). DNA methylation can directly block transcription factor (TF) binding by altering either the structure of TF motifs, or the accessibility of chromatin (Wiench et al., 2011). However, other classes of TFs will bind their corresponding motif regardless of DNA methylation or even show greater affinity to their binding site in the presence of DNA methylation (Wang et al., 2018). Furthermore, TFs can also mediate DNA methylation at 4 Figure 1. Cytosine methylation and demethylation. DNMTs and TETs carry out cytosine methylation and demethylation, respectively. 5 promoters of genes by directly recruiting de novo DNMTs to gene promoters (Gu et al., 2011; Stadler et al., 2011). Promoter CpG methylation also occurs in combination with other epigenetic silencing machinery, such as the activity of Histone 3 lysine 9 (H3K9) methyltransferases (e.g. G9a) in embryonic cell types during the epigenetic reprogramming that accompanies mammalian development (Dong et al., 2008; Epsztejn-Litman et al., 2008). This redundant inhibition is also thought to maintain gene retroviral silencing. In contrast to methylation at CpG island promoters, methylation in gene bodies is positively correlated with transcription activation and H3K36 trimethylation (H3K36me3) – a histone modification highly associated with actively transcribed genes (Hahn et al., 2011). Taken together it is clear that the relationship between DNA methylation and transcription is context specific. 1.2.1 Writers of DNA methylation: the DNMT family DNA methylation is added and maintained by a family of DNA methyltransferases (the DNMT family) (Figure 1, 2A and Table 1). DNMTs transfer a methyl group to carbon 5 of cytosine from the methyl donor, S-adenosyl-L-methionine (SAM). In this reaction, SAM is converted to S-adenosyl homocysteine (SAH), and a hydrogen group on carbon 5 of deoxycytidine is substituted by methyl group to yield 5-methyl-cytidine. The human genome encodes five members in this protein family: DNMT1, DNMT2, DNMT3A, DNMT3B and DNMT3L. DNMT1 is responsible for the maintenance of methylation during DNA replication by recognition of hemi-methylated regions (Kishikawa et al., 2003; Lee et al., 2001). DNMT1 is recruited to DNA by proliferating cell nuclear antigen (PCNA) and Ubiquitin Like With PHD And Ring Finger Domains 1 (UHRF1). The complex binds to mCpG on the parental strand and interacts with the newly synthesized strand through its CXXC zing-finger domain to methylate the opposing CpG 6 a) DNA methylation and transcriptional regulation in normal cellsEnhancer DNMTTFActivator TFPromoterTET5hmCRNA pol IICGIC 5mCAcHDACMeCP2DNMT3ADNMT3ADNMT3L DNMT3LLTR (long terminal repeat)RNA pol IIb) DNA methylation and transcriptional regulation in cancer cellsEnhancerTF TFRNA pol IILTR (long terminal repeat)DNMT3LDNMT3L Mutant DNMT3AMutant DNMT3APromoterTetRNA pol IICGIMutant TET2HG 7 Figure 2. DNA Methylation and Transcriptional Regulation. A) DNA methylation and transcriptional regulation in normal cells. DNMT3A and DNMT3L form a tetramer to catalyze DNA methylation. MeCP2, member of MBD family, binds to methylated cytosine and recruits histone modifiers and chromatin remodelers to the site resulting in a more compact chromatin structure. These two mechanisms together keep the long terminal repeats (LTRs) silenced in the genome. Tet proteins catalyze 5mC (black circles) into 5hmC (gray circles) and lead to loss of methylation state in CGIs of promoters, thus activate transcription. TF binding at enhancer sites depletes DNMTs resulting in hypomethylation of the region. It may also recruit other activators that interact with the promoters and regulate transcription initiation. B) DNA methylation and transcriptional regulation in cancer cells. Mutations in DNMT3A interrupt the tetramerization of DNMT3A-DNMT3L and cause hypomethylation and aberrant activation of LTRs. Loss-of-function mutations or inhibition by oncometabolite 2HG of Tet will lead to loss of 5hmC and global hypermethylation, resulting in an altered methylation status and disrupted transcription initiation. 8 on the nascent DNA strand (Arita et al., 2008; Bostick et al., 2007; Chuang et al., 1997). DNMT1 targets replication foci by way of a replication foci targeting sequence (RFTS) (Arita et al., 2008). DNMT1 also interacts with transcriptional repressors and histone deacetylases through DNMT1-associated protein 1 (DMAP1) binding domain (Rountree et al., 2000). DMAP1 directly interacts with DNMT1 at replication foci, binds to TSG101, a co-repressor (Sun et al., 1999), and maintains a transient repressive state of genes (Rountree et al., 2000). In contrast to DNMT1, DNMT3A and DNMT3B are responsible for symmetrical de novo CpG methylation (Okano et al., 1999). The mechanism of de novo methyltransferases is less understood compared to that of DNMT1. Studies have suggested that de novo methyltransferases recognize specific DNA sequences and interact with TFs to suppressed transcription (Deaton and Bird, 2011; Klose and Bird, 2006; Li et al., 2014; Robertson, 2005). DNMT3A/B interacts with modified and unmodified histones through a Pro-Trp-Trp-Pro (PWWP) and an ATRX-DNMT3-DNMT3L (ADD) domain (Du et al., 2015). The PWWP domain recognizes di- and tri-methylated H3K36 and the ADD domain directs DNMT3A/B to sites with unmethylated H3K4. This interaction at CpG-dense regions is prohibited by H3K4me3, or by way of other regulatory proteins, such as FBXL10 (Boulard et al., 2015) that binds unmethylated polycomb-bound CpG islands promoters to block de novo methylation. DNMT3L (Dnmt3-like) is a cofactor of both DNMT3A/B and acts to stimulate their enzymatic activities. DNMT3L lacks the PWWP and C-terminal catalytic domain found in DNMT3A/B (Jia et al., 2007; Lyko, 2018). DNMT3L stimulates DNMT3A methyltransferase activity by increasing its binding affinity by catalyzing a conformation change in the context of a heterotetrametric complex (2 DNMT3L surrounding 2 DNMT3A at the center of the complex) (Holz-Schietinger et al., 2011; Jia et al., 2007). 9 DNMT2 is a non-canonical DNMT and its function in mammalian genome regulation is poorly understood. DNMT2 is the only member of the human DNMT family that consists only of a catalytic domain with a unique CFT binding domain (Goll et al., 2006). DNMT2 is thought to play a role in post-transcription regulation of tRNAs (Schaefer et al., 2010) where it catalyzes methylation of tRNAs to block endonuclease cleavage. 1.2.2 Readers of DNA methylation: the MBD family 5mC acts to repress gene transcription in part through the recruitment of methyl-CpG-binding domain (MBD)-containing proteins (Figure 2A and Table 1). There are five members of the MBD family of proteins and among those, MeCP2 and MBD1-4 have been shown to bind to methylated CpGs (Lewis et al., 1992; Meehan et al., 1989). Once bound to DNA, MBD-containing proteins act by recruiting histone modifiers and chromatin remodeling complexes resulting in chromatin compaction and transcriptional repression. MBD-containing proteins have different levels of affinity for modified and unmodified CpGs (Fraga et al., 2003; Hameed et al., 2014). For example, MBD1 recruits histone modifying enzymes to methylated CpGs and results in silencing of transcription (Hameed et al., 2014). MBD2 and 3 are highly similar in sequence (77% identity) and recruit chromodomain-helicase-DNA-binding proteins (e.g. Mi-2) and nucleosome remodeling deacetylase (NuRD) chromatin remodeling complexes to methylated CpGs (Ginder and Williams, 2018; Gnanapragasam et al., 2011). In contrast to the other members of MBD family, MBD4 contains a glycosylase domain and is involved in identifying G.T and G.U mismatches originating from spontaneous deamination of methylated and hydroxymethylated cytosines, respectively, to activate a base excision repair (BER) process (Rai et al., 2008). 10 1.2.3 Erasers of DNA methylation: the TET family The ten-eleven-translocation family (TET1-3) of proteins catalyze the sequential oxidation of 5-methylcytosine (5mC) to 5-hydroxymethylcytosine (5hmC) in an α-ketoglutarate (α-KG)-dependent manner (Figure 1 and 2A) (Hu et al., 2013; Iyer et al., 2009; Tahiliani et al., 2009). α-KG acts as an oxidizing agent to yield succinate from oxidation reactions catalyzed by TET enzymes. 5hmC is thought to be an intermediate product in the pathway that actively or passively demethylates 5mC. 5hmC is found enriched at promoters and regulatory sites marked with H3K4me1 and is positively correlated with active transcription (Johnson et al., 2016; Lin et al., 2017; Wu et al., 2011b). In contrast to 5mC, the abundance of 5hmC varies significantly with tissue type (Spiers et al., 2017; Sun et al., 2014). Interestingly, 5hmC (along with other oxidized form of 5mC) has been shown to be more abundant in post mitotic cell types that comprise the mammalian brain than any other cell type (Kriaucionis and Heintz, 2009)(Globisch et al., 2010). 5hmC comprises ~0.15-0.6% in brain tissues and ~0.01-0.05% in other tissues (Globisch et al., 2010; Kriaucionis and Heintz, 2009; Skvortsova et al., 2017). Regulated demethylation through generation of 5hmC offers an important mechanism for a dynamic reprogramming of the mammalian methylome, and in effect, transcriptional regulation. 5hmC itself is further oxidized by TET family proteins to generate 5-formylcytosine (5fC) and 5-carboxylcytosine (5caC) (Figure 1). The thymine DNA glycosylase (TDG) protein, along with the mechanism of BER, actively excise 5fC or 5caC, ultimately resulting in demethylation of methylated cytosine. There are three members of TET family of enzymes: TET1, TET2, and TET3. All three have the capacity to oxidize methylated cytosine via a C-terminal cysteine-rich catalytic domain (Pastor et al., 2013; Wu and Zhang, 2017). Unlike the TET2 protein, full-length TET1 and TET3 proteins also contain a CXXC domain and thus are able to bind CpGs. TET enzymes have high 11 affinity for methylated CpG (Pastor et al., 2013) and higher affinity for 5mC compared to 5hmC and 5fC (Crawford et al., 2016; Weber et al., 2016). Conversion of 5mC to 5hmC by TET enzymes is faster compared to other oxidation reactions catalyzed by these enzymes (i.e. 5hmc to 5fC). TET enzymes can be regulated through post-translational modifications, such as monoubiquitylation and acetylation (Chen et al., 2014; Zhang et al., 2017). For example, acetylation of lysine 111 of TET2 increases its catalytic activity and protein stability during oxidative stress (Zhang et al., 2017). TET enzymes are recruited by a variety of TFs at their specific regulatory sites (Costa et al., 2013; Neri et al., 2013a; Wang et al., 2015). For example, NANOG physically interacts with TET1 and TET2 in mouse embryonic stem cells (ESCs) (Costa et al., 2013). Polycomb repressive complex 2 (PRC2), a histone modifier, has also been reported to interact with TET1 in mouse ESCs (Neri et al., 2013a). In AML and human monocytes, WT1 and PU.1 have also been shown to facilitate the recruitment of TET2 to their target gene promoters and enhancers (Mingay et al., 2018; Wang et al., 2015). 1.3 Histone modifications Chromatin dynamics control the accessibility of DNA transcription complexes (Li et al., 2014). Nucleosomes consists of two copies of each of the histone proteins (e.g. H2A, H2B, H3 and H4) and roughly 146 pairs of DNA bases. The H1 protein connects individual nucleosomes to form a higher order compacted chromatin structure (Hergeth and Schneider, 2015). Covalent chemical modification of histones is another form of epigenetic modification that influences transcriptional regulation and chromatin structure (Figure 3A) (Li et al., 2014). Post-translational covalent chemical modifications of amino acids on the amino terminal (N-terminal) tail of 12 histones, in particular, regulate local transcriptional activity (Jaenisch et al., 2010; Local et al., 2018; Petruk et al., 2017). 13 14 Figure 3. Histone modifications and transcriptional regulation. A) Regulation of gene expression by suppressive and permissive histone markers and associated regulatory factor in nontransformed cells. B) Upregulation of suppressive markers and their associated factors or downregulation of permissive markers and their associated protein leads to aberrant expression of genes, such as tumor suppressor genes in transformed cells. C) Upregulation of permissive histone marker and their associated regulatory factors or downregulation of suppressive histone markers and their associated regulatory factors leads to aberrant expression of genes in transformed cells. Variation in histone associated genes (red histone) can lead to abnormal expression of genes through disruption of regulatory factors. 15 Examples of such modifications are acetylation, methylation, phosphorylation, ubiqutylation, sumoylation and hydroxylation (Jiang et al., 2012, 2007; Shiio and Eisenman, 2003; Unoki et al., 2013; Wang et al., 2001; Weake and Workman, 2008). These functionally relevant histone modifications play important roles in transcriptional regulation by generating a ‘histone code’ that underlines active or repressive chromatin structure (Dunham et al., 2012; Jaenisch et al., 2010; Jenuwein and Allis, 2001; Local et al., 2018; Petruk et al., 2017). The type and location of the modifications dictate the consequence of a chromatin state. For instance, H3K27me3 and H3K27ac are repressive and active marks, respectively (Jaenisch et al., 2010; Lewis, 2007). As expected, these histone modifiers are type- and location-specific (Bannister and Kouzarides, 2011). Here I will focus on the regulation and transcriptional consequences of the most commonly studied histone modifications, acetylation and methylation. 1.3.1 Histone acetylation Histone acetylation is one of the major post-translational histone modifications that correlates with active transcription (Jaenisch et al., 2010; Roadmap Epigenomics Consortium et al., 2015). Histone acetylation predominantly occurs on lysine residues (Grunstein, 1997). Histone acetylation plays a major role in the unfolding of silenced chromatin by adding a negatively charged acetyl group on the lysine residues of histone proteins (Garcia-Ramirez et al., 1995). Histone acetyl transferases (HAT) carry out this modification by transferring the acetyl group from acetyl-coenzyme A to a lysine residue within a histone’s N-terminal tail. The addition of an acetyl group creates repulsion of the histone octamer from DNA, and as a result, reduces their electrostatic affinity (Brownell and Allis, 1995; Brownell et al., 1996; Cano and Pstana, 1979; Kouzarides, 2007; López-Rodas et al., 1991) The unfolding of DNA from histones provides 16 opportunity for transcriptional machinery to bind to DNA. As a result, histone acetylation is considered a permissive modification. HATs can be classified into five families of enzyme subgroups: MYST (MOZ, Ybf2/Sas3, Sas2, Tip60), Gcn5 related N-acetyltransferase (GNAT), p300/CBP, steroid receptor co-activator family proteins, and cytoplasmic HATs. These enzymes are also capable of acetylating non-histone proteins inside or outside the nucleus (Bannister and Kouzarides, 1996; Champagne et al., 2001; Dyda et al., 2000; Hilfiker et al., 1997; Kawasaki et al., 2000; Lee and Workman, 2007; Mizzen et al., 1996; Ogryzko et al., 1996). Histone H3 lysine acetylation is found to be associated with enhancer regions and active transcriptional start sites (Dunham et al., 2012; Jaenisch et al., 2010). HATs are site- and substrate-specific (Lee and Workman, 2007). For example, Gcn5/PCAF is required for the acetylation of H3K9, whereas p300/CBP is required for acetylation of H3K18 and H3K27 (Jin et al., 2011). H3K27 acetylation (H3K27ac) is positively correlated with the expression of nearby genes and thus is commonly used to identify active regulatory regions (Jaenisch et al., 2010; Roadmap Epigenomics Consortium et al., 2015). H3K27ac is enriched in evolutionarily conserved regions of the genome and transcription binding elements (Figure 4) (Roadmap Epigenomics Consortium et al., 2015). Furthermore, H3K27ac enriched regions overlap directly with regions bound by TFs as well as open chromatin regions as measured by DNAse1 hyper-sensitivity and ATAC-seq (Dunham et al., 2012; Fu et al., 2018; Pan et al., 2019). Taken together, these observations are supportive of the permissive qualities of H3K27ac and its close relationship with TFs. For this reason, H3K27ac density genome-wide is broadly used for TF analysis and network prediction (e.g. Pellacani et al., 2016). 17 38 161 TF binding siteDNA accessible regionH3K27acExpression level 18 Figure 4. H3K27ac correlates with TF binding site, evolutionary conserved regions and active transcription. Tracks from top to bottom: Refseq genes, Gene expression across 53 tissues (ENCODE consortium), H3K27ac in 7 cell lines (ENCODE consortium), DNase hypersensitivity sites in 125 cell types (ENCODE consortium) and 161 transcription factors’ binding sites (ENCODE consortium). 19 The location, pattern and strength of the H3K27ac signal can be linked to the expression of nearby genes in order to identify cell-type specific gene signatures (Hnisz et al., 2013). Large domains of H3K27ac with high ranking H3K27ac density are used to define a chromatin state termed ‘super-enhancer’ (Figure 5) (Hnisz et al., 2013). Super-enhancers coincide with the binding of clusters of transcription co-activators collectively known as mediators (Hnisz et al., 2017; Madani Tonekaboni et al., 2019). These regions have been associated with genes responsible for cell identity in both normal and malignant cells and expression of super-enhancer-associated genes can segregate cells of different types (Hnisz et al., 2013). Super-enhancers differ significantly in cancers of different origins and are enriched in disease-specific single nucleotide polymorphisms (SNPs) (Hnisz et al., 2013). Recent studies have leveraged super-enhancers to identify therapeutic targets, for example in primary ependymoma (Mack et al., 2018). 1.3.2 Histone deacetylation Histone acetylation levels are determined by the concerted actions of both HATs and histone deacetylases (HDAC). HDACs remove the acetyl group from histone tails, causing chromatin contraction and compaction, and thereby leading to gene repression (Leipe and Landsman, 1997). Removal of acetyl groups drives interaction between negatively charged DNA with the positively charged histone. The first successful purification of a protein with HDAC activity was performed in 1969 from a calf thymus extract (Inoue and Fujimoto, 1969). Today,18 mammalian HDACs classified into four groups have been described (Grozinger et al., 1999; Kao et al., 2000; Seto and Yoshida, 2014; Taunton et al., 1996; Yang et al., 1996, 1997). Group I, II, and IV consist of Zn2+ dependent HDACs, and are named HDAC1-11. The zinc ion is required for the hydrolysis of the acetylated amide bond, subsequently causing deacetylation of the N-terminal 20 Figure 5. An example of Super-enhancer. Genome browser view of chr1:85,346,571-85,946,750 showing H3K27ac density and profile within a super-enhancer boundaries (red box) in K562 cell line (ENCODE consortium). Super enhancerH3K27acchr1:85,346,571-85,946,750 21 domain of the histone (Gennip et al., 2003; Verdin et al., 2003). In contrast, Group III comprises of the NAD+ dependent Sirtuins, SIRT1-7. Histone deacetylase classes HDAC I, II and IV share a similar three-dimensional structure and HDAC I and II classes are more closely related to one another than to any other groups. On the other hand, HDAC III is evolutionarily unrelated to the other classes (Haberland et al., 2009). As a result, HDACs refer to class I, II and IV exclusively in this introduction. HDACs are rarely present as monomeric complexes; they often appear as multi-protein complexes, and interact with transcription regulating and chromatin modifying enzymes, such as NuRD (Yang and Seto, 2003). HDACs are generally associated with transcriptional repression (Laporte et al., 2017; Liu et al., 2014; Xie et al., 2012) and their inhibition leads to activation of genes, for example the tumour suppressor CDKN2A in synovial sarcoma (Laporte et al., 2017). In addition, HDACs have been implicated in the in regulation of cell cycle progression and apoptosis (Laporte et al., 2017; Ropero and Esteller, 2007) and their dysregulation is associated with tumouur progression in multiple malignancies (Haberland et al., 2009; Ropero and Esteller, 2007). Consequently, HDAC inhibition is extensively studied as a therapeutic option for certain tumours (Dubois et al., 2009; Laporte et al., 2017; Marks et al., 2001; Ropero and Esteller, 2007). 1.3.3 Histone methylation Histone methylation occurs at both arginine and lysine residues of histone H3 and H4, first reported in a calf thymus extract in 1964 (Allfrey and Mirsky, 1964). Histone methylation plays a critical role in normal cell development and regulation of gene expression (Bernstein et al., 2002; Heintzman et al., 2007; Santos-Rosa et al., 2002; Schulze et al., 2009). The transcriptional outcome of histone lysine methylation depends not only on the degree of methylation (such as, mono-, di-, 22 and trimethylation), but also its location (Figure 3A) (Local et al., 2018; Santos-Rosa et al., 2002). For instance, trimethylation of histone 3 lysine 4 (H3K4me3), as well as of H3K36me3 and H3K79me2, are associated with active transcription (Kolasinska-Zwierz et al., 2009; Santos-Rosa et al., 2002). In contrast, methylation of H4K20me3, H3K27 and H3K9 are associated with repressive chromatin states (Becker et al., 2016; Nelson et al., 2016; Petruk et al., 2017). Both permissive and repressive histone modifications have been reported to exist concurrently on a nucleosome (Bernstein et al., 2006; Roadmap Epigenomics Consortium et al., 2015) with the so-called ‘bivalent’ state containing both H3K27me3 and H3K4me3 being the most well studied (Bernstein et al., 2006). In addition, antagonistic relationships between histone modifications, like H3K27me3 and H3K36me3, can also be observed. 1.3.3.1 Methyl transferases Lysine methyltransferases (KMT) are responsible for methylating histone lysine residues (Hyun et al., 2017). KMTs contain a catalytic SET domain which catalyzes the transfer of a methyl group from S-adenosylmethionine to the ε-amine on the side chain of lysine residue (Dillon et al., 2005). It is worth noting that KMTs use the same substrate as DMNTs. Therefore, disruption in S-adenosylmethionine production will have consequences on both DNA and histone methylation. Here I introduce H3K4me3, H3K4me1, H3K27me3, H3K9me3 and H3K36me3 modifications and their corresponding methyl transferases. In mammals there are six H3K4 methyltransferases: SET1A, SET1B, mixed-lineage leukemia 1 (MLL1), MLL2, MLL3, and MLL4 (Højfeldt et al., 2013; Hyun et al., 2017; Karatas et al., 2013; Li et al., 2016; Lu et al., 1999; Yang et al., 2015). These methyltransferases act as a scaffold protein and form a complex with WD repeat domain 5 (WDR5), retinoblastoma- 23 binding protein (RbBP), absent/small/homeotic-2 like (ASH2L), DPY30, and other complex specific subunits (Hyun et al., 2017). H3K4 methyltransferases show variable affinity for various degrees of H3K4 methylation. For example, MLL1/2 exhibits strong affinity towards H3K4me1/2; but MLL3/4 show specificity for H3K4me1 (Dorighi et al., 2017; Duncan et al., 2015). Polycomb repressive complex 2 (PRC2), but not PRC1, is responsible for the methylation of H3K27me2/3 (Margueron and Reinberg, 2011). Canonical PRC2 is comprised of three main subunits encoded by the evolutionarily conserved genes: embryonic ectoderm development (EED), suppressor of zeste 12 (SUZ12), and enhancer of zest 2 (EZH2) (Margueron and Reinberg, 2011). The catalytic domain of PRC2 is encoded in EZH2 (Kasinath et al., 2018). Additionally, EED is an allosteric domain, while SUZ12 is involved in binding of the complex to histones (Kasinath et al., 2018). Suppressor of variegation 3-9 homolog 1 (SUV39H1), SUV39H2, SETDB1, G9a, and the PRDM family are responsible for the methylation of H3K9 (Brower-Toland et al., 2009; Collins et al., 2005; Loyola et al., 2009; Rice et al., 2003; Tachibana et al., 2002). These complexes exhibit both site and H3K9 methylation degree specificity. For example, G9a has been shown to methylate H3K9me1/2 at euchromatin, ultimately suppressing gene expression in early embryogenesis (Tachibana et al., 2002). Multiple methyltransferases, such as nuclear receptor binding SET domain protein (NSD) 1/2/3, SETD3, SET domain and mariner transposase fusion protein (SETMAR) and SET and MYND domain containing 2 (SYMD2) are responsible for the methylation of H3K36 (Hyun et al., 2017). Surprisingly, only SET2 can catalyze methylation of H3K36me3 (Yuan et al., 2009). Together, these relationships attest to the complex role of histone methylation in regulating chromatin state and transcription. 24 1.3.3.2 Histone Lysine Methylation International consortia have applied ChIP-seq based methodologies to generate reference epigenome resources across a broad array of normal and diseased cell types (Dunham et al., 2012; Roadmap Epigenomics Consortium et al., 2015; Stunnenberg et al., 2016). These and related studies have identified genomic features associated to, and correlated the occupancy of commonly studied histone methylation states (H3K4me1, H3K4me3, H3K27me3, H3K9me3, and H3K36me3) with gene transcription and DNA methylation (Dunham et al., 2012; Roadmap Epigenomics Consortium et al., 2015). For example, H3K4me3 is found to be unusually stable across different cell types relative to other modifications (Roadmap Epigenomics Consortium et al., 2015). Moreover, H3K4me3 is enriched at active promoters and correlates positively and negatively with transcription and DNA methylation, respectively (Santos-Rosa et al., 2002). In addition, different degrees of H3K4 methylation are associated with unique genomic feature. As an example, sites enriched with H3K4me1 alone are associated with poised regulatory regions (Heintzman et al., 2007; Local et al., 2018). H3K4 methylation can co-occur with other permissive and repressive histone modifications, such as H3K27ac and H3K27me3 (Jaenisch et al., 2010). The concurrence of permissive and repressive marks can thereby reinforce or weaken their positive relationship with transcription, respectively (Bernstein et al., 2006; Roadmap Epigenomics Consortium et al., 2015). Another well studied epigenomic modification is the repressive mark H3K27me3 whose presence in a gene promoter is negatively correlated with transcription (Petruk et al., 2017). The relationship between DNA methylation and H3K27me3 is more complex than with H3K4me3. Genome wide analyses of H3K27me3 and DNA methylation show hypermethylation at H3K27me3 enriched sites (Roadmap Epigenomics Consortium et al., 2015). Promoters enriched 25 with both H3K4me3 and H3K27me3, referred to as bivalent promoters, show lower transcription than promoters enriched with H3K4me3 alone and are generally hypomethylated. Bivalent promoters are thought to regulate developmentally important genes in ESCs (Bernstein et al., 2006). They are also present in terminally differentiated populations (Pellacani et al., 2016). Co-occupied regions with H3K4me1 and H3K27me3, commonly referred to as bivalent enhancers, are also hypomethylated, and correlate negatively with transcription (Roadmap Epigenomics Consortium et al., 2015; Zentner et al., 2011). Unlike and H3K4me3 and H3K27me3 which alone are exclusively associated with permissive and repressed chromatin states respectively, H3K9me3 occupancy has been correlated with both. H3K9me3 was initially described as mark associated with repressed heterochromatin (Becker et al., 2016; Peters et al., 2001). However, co-occurrence of H3K9me3 and H3K36me3 in the gene body of actively transcribed gene is a hallmark of active zinc finger gene clusters suggesting a more complex relationship between H3K9me3 and transcription (Hahn et al., 2011; Roadmap Epigenomics Consortium et al., 2015). H3K36me3 is associated with actively transcribed genes and is enriched in the gene body of actively transcribed genes and depleted at promoters (Kolasinska-Zwierz et al., 2009). Together, these relationships attest to the complex role of histone methylation in regulating chromatin state and transcription. 1.3.4 Histone demethylation Analogous to histone acetylation, histone methylation levels are defined by the combined actions of histone methyltransferases and histone demethylases. Histone lysine demethylases can be divided into two broad classes: the lysine demethylase 1 (KDM1) family and the Jumonji C (JmjC) domain family (Allis et al., 2007; Hyun et al., 2017; Shi et al., 2004; Whetstine et al., 2006). 26 The KDM1 family includes members that use flavin as a co-factor and carry out demethylation of H3K4 and H3K9 through oxidative demethylation (Cloos et al., 2008). The KDM1 family is only able to remove methyl groups from di- and mono- methylated lysines. For example, KDM1A interacts with CoREST and removes methyl group from H3K4me1 and H3K4me2 through oxidation, thereby resulting in transcriptional repression (Yang et al., 2006). The second and largest class of demethylases is the JmjC domain family. This family of histone demethylases include KDM2/3/4/5/6 and are able to demethylate all degrees of histone lysine methylation (Cloos et al., 2008). JmjC domain KDMs are 2-oxoglutarate-dependent dioxygenases and thus require Fe2+ and oxygen for their activity. Coordinated action between methyltransferase and demethylases is a part of a complex and intricate network of epigenetic modifiers and TFs, which ultimately regulates transcription. 1.4 Epigenetic Assays 1.4.1 Chromatin immunoprecipitation sequencing One of the first uses of nascent massively parallel sequencing was chromatin immunoprecipation sequencing (Barski et al., 2007; Johnson et al., 2007; Mikkelsen et al., 2007; Robertson et al., 2007) Subsequently, numerous innovative experiment approaches have been developed for both the qualitative and quantitative measurement of epigenomic features including assays for whole genome mapping of DNA methylation, histone modifications, open chromatin and 3 dimensional structure (Barski et al., 2007; Buenrostro et al., 2013; Lieberman-Aiden et al., 2009; Lister et al., 2009; Song and Crawford, 2010) Chromatin immunoprecipitation followed by high throughput sequencing (ChIP-seq) is a standard method for genome-wide mapping of histone modifications and TF biding sites (Barski 27 et al., 2007; Johnson et al., 2007). ChIP-seq uses specific antibodies to immunoprecipitate DNA fragments occupied by TFs or histone modifications. There are multiple iterations of the ChIP-seq assay, but all of them can be categorized into two broad groups - cross-linked and native ChIP-seq (Barski et al., 2007; Johnson et al., 2007; Maunakea et al., 2010). Cross-linked ChIP consists of covalently attaching proteins of interest to DNA prior to chromatin fragmentation and immunoprecipitation (Solomon et al., 1988). This covalent linkage is chemical or UV induced (Gilmour and Lis, 1984; Jackson, 1978). After immunoprecipitation, protein and DNA are separated by reverse crosslinking. The reverse crosslinking reduces the final yield of nucleic acid. For this reason, many low input ChIP-seq methodologies avoid the cross linking step (Brind’Amour et al., 2015). Native ChIP utilizes enzymatic digestion of the chromatin using Micrococcal Nuclease (MNase) (Hebbes et al., 1988). MNase digests the linker DNA between the nucleosomes. Native ChIP is more suitable as a low input assay because it does not need to covalently attach proteins of interest to DNA, hence there is no DNA loss during the reverse crosslinking step. However, unlike sonication, MNase digestion is concentration-sensitive and it can lead to over or under digestion if there is too little or too much chromatin, respectively. Optimal digestion of chromatin where majority of fragments are mono-nucleosome size is crucial in a production of high-quality Native ChIP-seq libraries. Commonly employed ChIP-seq methods require hundreds of thousands or millions of cells, although recent advances has reduced the number of required cells to hundreds of, and even single cells (Brind’Amour et al., 2015; Grosselin et al., 2019). However, these methods are challenging to implement and may require case by case optimization. 28 1.4.2 Whole genome bisulfite sequencing A now commonly used method for whole genome mapping of DNA methylation is bisulfite sequencing (BS-seq) (Clark et al., 2017; Hayatsu et al., 1970; Lister et al., 2009). Bisulfite treatment of single-stranded DNA converts cytosine, but not methylated cytosine, to uracil, which subsequently converts to thymidine following PCR. Therefore, BS-seq will read unmethylated C as T without affecting the methylated Cs. BS-seq maps genome wide methylation at base pair resolution. However, a caveat to BS-seq is its inability to distinguished hydroxy methylation of cytosine from methylation of cytosine. In the first iteration of BS-seq, extracted DNA is fragmented with sonication and bisulfite conversion is done after the ligation of Illumina sequencing adaptors to fragmented DNA. Bisulfite conversion of single stranded DNA causes breakage in DNA strand and leads to a lower yield. To circumvent this issue, a new iteration of BS-seq simultaneously fragments the DNA and converts unmethylated cytosine to uracil through bisulfite conversion before Illumina adapter ligation (Miura et al., 2012). This method increases the yield and reduces the minimal input DNA requirement. Further advances have now made single-cell mapping of DNA methylation genome wide possible (Hui et al., 2018). Since there is no local enrichment in BS-seq, it requires high depth sequencing for accurate detection of DNA methylation. The advent of massively parallel sequencing methods has spurred the development of all quantitative genome-wide approaches to analyze chromatin changes in normal and diseased states of cellular growth and development described in the next section. 29 1.5 Epigenomic dysfunction in cancer Cancer is a disease of the genome that leads to aberrant gene transcription and subsequent abnormal function of the cell. Large-scale genome re-sequencing efforts have identified a heterogeneous collection of genetic mutations within the coding regions of genes encoding transcriptional regulatory proteins in human cancer (Zhang and Wang, 2015). Further functional genomic characterization of select tumours has revealed that specific regulatory mutations can disrupt transcription patterns that lead to abnormal silencing of tumour suppressor genes and the expression of oncogenes. Emerging research is now beginning to explore and characterize genetic and epigenetic variations in non-coding regulatory regions. 1.5.1 DNA methylation in cancer Gross alterations to DNA methylation have been observed in nearly every cancer examined to date. The genomes of cancer cells are characterized by genome wide hypomethylation combined paradoxically with localized regions of hypermethylation (Gama-sosa et al., 1983). Hypermethylation is frequently found within promoter associated CpG islands (CGI) and these differ between and within cancer subtypes (Graff et al., 1995; Nguyen et al., 2001). This is perhaps a consequence of a combination of unique environmental exposures and genetic differences between tumour cells. Loss-of-function mutations in the DNMT gene family have been identified in a wide variety of human cancers (Figure 2B and Table 1). Somatic inactivating mutations in DNMT1 have been found in colorectal cancer and acute myeloid leukemia (AML) (Dolnik et al., 2012; Kanai et al., 2003). Mutations in DNMT3A have been identified in approximately 20% of AML and myelodysplastic (MDS) patient genomes (Walter et al., 2011; Yan et al., 2011). These driver 30 mutations are thought to disrupt tetramerization of the Dnmt3a-Dnmt3l complex, resulting in hypomethylation and deregulation of critical control pathways (Holz-Schietinger et al., 2012). Surprisingly, cancer genomes harboring DNMT3A mutations also contain blocks of hypermethylation in tumour suppressor associated CG islands (CGIs). Recent studies have demonstrated direct interactions between DNMT3A and certain TFs responsible for these seemingly contradictory events (Gu et al., 2011). Emerging evidence suggests that the deregulation of DNMT expression is also associated with disease progression in cancer (Chen et al., 2012; Ben Gacem et al., 2012; Hayette et al., 2012; Zhang et al., 2012). Contrary to the loss-of-function mutations mentioned above, upregulation of all three active members of the DNMT family has been reported in various cancers (Ben Gacem et al., 2012; Wu et al., 2011a). Despite the robust evidence in favor of an association between the upregulation of DNMT and cancer, its exact role in tumorigenesis and whether it plays a similar role to DNMT mutations remains unknown. Similarly, deregulation of MBD family members, primarily through upregulation, has been identified in cancer cells. For example, MeCP2 has been shown to be overexpressed in prostate cancer cells. It has also been found to facilitate cell growth by transcriptionally repressing critical apoptotic pathways (Bernard et al., 2006). A broad range of non-synonymous mutations disrupting TET family members in cancer genomes have been described (Dolnik et al., 2012; Tefferi et al., 2009). The majority of these mutations have unknown functional impact (Forbes et al., 2011), but approximately one-quarter are annotated as missense, frameshift and nonsense mutations that inactivate TET, leading to lower levels of 5hmC (Ko et al., 2010), and deregulating transcriptional responses in cell signaling and differentiation pathways (Gaidzik et al., 2012). Furthermore, downregulation of TET2 and the 31 subsequent loss of 5hmC has recently been discovered to be a hallmark of melanomas (Lian et al., 2012). 1.5.2 Histone Modification in Cancer Histone modifications and deregulation of the pathways which control them collude with other factors to determine etiologies of human diseases, including cancer (Figure 3B, C and Table 2) (Li et al., 2013; Love et al., 2012; Maiques-Diaz et al., 2012; Schwartzentruber et al., 2012; Wang et al., 2013; Wei et al., 2013). HATs are both site- and substrate- specific and their deregulation and aberrant histone H3 and H4 acetylation have been observed in a range of tumour types (Love et al., 2012)(Wang et al., 2013)(Jin et al., 2011)(Pfister et al., 2008). For example, overexpression of GCN5 has been observed in lung cancer and its abundance correlated with tumour size (Chen et al., 2013a). CREBBP, which encodes creb binding protein (CBP), has been shown to undergo mutation in diffuse large B-cell lymphoma and found to decrease the global level of H3K18ac (Pasqualucci et al., 2011a). Similarly, deletion of 3p24, a region that harbours P300/CBP-associated factor (PCAF), and hypermethylation at the promoter region of PCAF, are both seen in esophageal squamous cell carcinomas (ESCCs) (Pasqualucci et al., 2011a). Some HATs have been shown to be both downregulated and upregulated in different cancer types. Downregulation or upregulation of hMOF, a Myst family member responsible for acetylation of H4K16, has been shown to correlate with tumorigenesis in human (Pfister et al., 2008; Wang et al., 2013; Zhao et al., 2013). Downregulation of hMOF and loss of H4K19 are involved in human renal carcinoma, breast cancer, and medulloblastoma (Pfister et al., 2008; Wang et al., 2013). In contrast, hMOF and H4K16ac are upregulated in human non-small cell lung cancer (Zhao et al., 2013). H3K27ac, an active enhancer marker (Jaenisch et al., 2010), exhibits distinct 32 signatures in primary matched normal and transformed cells, and even within a population of transformed cells (Akhtar-Zaidi et al., 2012; Wang et al., 2011a). Mutation and downregulation or upregulation of HATs are thought to contribute to the genesis and progression of cancer through aberrant gene expression. Diverse patterns of misregulation of HATs in oncogenesis dictate the involvement of HATs in a complex network of regulatory elements. Further investigation is needed to better understand the role of HATs in the pathology of cancer and the networks of regulators in which HATs are involved. Unlike HATs which show both upregulation and downregulation in cancer cells, HDACs are found to be exclusively upregulated in cancer cells. This suggests a common mechanism of action (Jung et al., 2012; Maiques-Diaz et al., 2012; Marshall et al., 2010; Stypula-Cyrus et al., 2013). Furthermore, HDAC upregulation has been observed in pre-cancerous lesions, suggesting that it is an early event in tumour development (Stypula-Cyrus et al., 2013). Aberrant recruitment of HDACs has also been observed in cancer (Maiques-Diaz et al., 2012). In (8;21) AML, HDAC-containing complexes are recruited by the AML1/ETO fusion protein to deacetylate H4 at genes associated with differentiation and thereby thought to promote the acquisition of malignant properties. Numerous cases of deregulation and/or mutation to histone methyl transferases have been observed in cancer cells (Kondo et al., 2008; McCabe et al., 2012; De Rooij et al., 2013). However, no common pattern of disruption for methyltransferases in cancer has emerged to date. For instance, MLL2, a H3K4 methyltransferase, is upregulated in both colon and breast tumors, and its abundance correlates with tumour progression (Natarajan et al., 2010). In contrast, MLL2 mutations in diffuse large B-cell lymphoma are largely inactivating (Pasqualucci et al., 2011b). NSD2, a member of the nuclear SET domain containing family of H3K36 methyl-transferases, is 33 overexpressed in human leukemias, breast, bladder, and prostate cancers (Hudlebusch et al., 2011; Kassambara et al., 2009). Deregulation of repressive HMTs is also associated with cancer (Khan et al., 2013; McCabe et al., 2012). A neomorphic mutation in the catalytic domain of EZH2, a member of the PRC2 complex, leads to an increase in H3K27me3 and its inhibition prevents proliferation in large B-cell lymphoma (Ganji et al., 2012; McCabe et al., 2012; Morin et al., 2010) In contrast, inactivating mutations of EZH2, observed in some myelodyplastic patient genomes, lead to a decrease in H3K27me3 levels (Khan et al., 2013). On the other hand, overexpression of EZH2 is seen in human prostate and non-small cell lung cancers (Kikuchi et al., 2012; Li et al., 2013). In addition, G9a, a H3K9 methyltransferase, participates in abnormal chromatin methylation and is associated with aberrant gene expression in human lung, head and neck, and prostate cancers (Chen et al., 2010; Kondo et al., 2008; Tan et al., 2012). Finally, the NUP98-NSD1 fusion gene drives upregulation of the HOXA gene cluster by elevating H3K36me3 levels at these gene loci in human AML (Wang et. al., 2007; Shiba et. al., 2013). Both up- and down-regulation of histone demethylases have been associated with cancer (Ding et al., 2013; De Rooij et al., 2013; Wang et al., 2009b). For example, LSD1 downregulation correlates with metastasis in breast cancer in part through the activation of the transforming growth factor beta 1 (TGFβ1) gene (Wang et al., 2009b). In contrast, inhibition of LSD1 has been shown to improve the therapeutic efficiency of all-trans-retinoic acid in human AML (Schenk et. al. 2012). Additionally, selective inhibition of LSD1 in AML, pluripotent teratocarcinomas and embryonic carcinomas, inhibits cellular proliferation (Schenk et al., 2012; Wang et al., 2011b, 2009b). Histone demethylase JMJD1A is upregulated in tumours under hypoxic and nutrient-starved conditions (Osawa et al., 2013). Knockdown of the demethylases JMJD2A, JMJD2C, 34 KDM2B, KDM6A, KDM8, JMJD6, JMJD7, and JMJD8 inhibits metastasis in squamous cell carcinoma (SCC) (Ding et al., 2013). Furthermore, inactivating mutations in JARID1C and global increases in H3K4me3 abundance have been described in clear cell renal cell carcinoma (ccRCC) in humans (Dalgliesh et al., 2010; Niu et al., 2012). Inactivating mutations of UTX, a H3K27 demethylase, have been described in multiple myeloma, esophageal squamous cell carcinoma, and renal cell carcinoma (Van Haaften et al., 2009). On the other hand, JMJD3, a H3K27 demethylase, is upregulated in human myelodysplasic cell lines (Wei et al., 2013). The oncometabolite 2-hydoxy glutarate (D2HG) produced by mutant isocitrate dehydrogenases (IDH) can also inhibit JMJC histone demethylases, suggesting both DNA methylation and histone modification reprogramming can be driven by a common mechanism (Chowdhury et al., 2011; Lu et al., 2012). Deregulation of histone interpreters, such as proteins harboring PHD domains, the inhibitor of growth (ING) family, as well as PRC1 complex members, are also observed in cancer cells (Kumamoto et al., 2009; De Rooij et al., 2013; Wang et al., 2009a). For example, a chimeric protein NUP98-PHF23, which is a result of the cryptic translocation of t(11;17)(p15;p13) has been observed in patients’ AML cells (Reader et al., 2007)(Wang et al., 2009a). NUP98-PHF23 recognizes H3K4me3 through a PHD domain in NUP98 and plays the role of a ‘boundary factor’ that prevents removal of H3K4me3 and addition of H3K27me3 at select genes to promote their transcription (Wang et al., 2009a). ING2, a member of the ING family and a H3K4me3 reader, is upregulated in colon cancer (Kumamoto et al., 2009). Members of the PRC1 complex, which recognizes the PRC2 modified H3K27me3, have also been shown to be upregulated in cancer (Lohse et al., 2013; Majewski et al., 2010). Upregulation of BMI1, a subunit of the PCR1 complex, also correlates with the enhancement of pancreatic adenocarcinoma (PA) tumour growth as well as metastasis (Proctor et al., 2013). Together, these findings suggest the importance of deregulated 35 epigenetic modification in cancer and highlight the lack commonalities in mis-regulation of epigenetic modifiers in cancer (with the exception HDAC). 1.6 The hematopoietic system as a model to understand epigenetic regulation of normal tissue homeostasis and its malfunction in AML Hematopoiesis describes the process by which mature blood cells are produced (Eaves, 2015). Understanding of the cellular changes involved and their control has relied on two types of analyses. The first to be developed relied on a description of the morphological changes erythroid, granulocytes, monocytes and megakaryocytes undergo during the last series of three to five amplifying divisions that separate the morphologically indistinguishable “blasts” from the final mature blood cells. An appreciation of a hierarchical structure within the blast population was later developed from a combination of clonal assays of mature blood cell output activities from isolated and phenotypically separable subpopulations of the blasts (Laurenti and Göttgens, 2018). This has led to the current view of hematopoiesis as a multilayer hierarchy in which self-renewing hematopoietic stem cells (HSCs) with enormous regenerative potential in transplantation assays in vivo generate various types of progenitors with reduced self-sustaining ability and eventually single-lineage-restricted differentiation capacity when similarly stimulated. To distinguish the final series of changes in which these progenitors give rise to single types of fully functional blood cell types from the earlier steps involved in their production, the late stages are referred to as terminally differentiating cells. More recent single cell transcriptomic studies (Buenrostro et al., 2018), lineage tracing studies of blood cell production in the unperturbed mouse ( Rodriguez-Fraticelli et al., 2018), and in human cells in vitro and in xenografted immunodeficient mice (Haas et al., 2018; Knapp et al., 36 2018, 2019; Laurenti and Göttgens, 2018) now suggest a more complex model of the initial stages of hematopoiesis. This model incorporates several important new conceptual modifications: (i) that cells defined as HSCs in transplant assays may have intrinsically specified propensities for generating particular lineages; (ii) that multiple types of primitive hematopoietic cells may vary in the types and duration of lineage outputs they can sustain according to whether they are stimulated under homeostatic versus highly stimulated conditions, (iii) that even under the same conditions, they may pursue different molecular trajectories to reach different lineage-restricted progenitor cell types, and (iv) that all of these parameters are altered during the development of the organism from the embryo to the adult (Babovic and Eaves, 2014; Notta et al., 2016). The accessibility of hematopoietic cells and the ability to isolate defined subsets with different output activities has made them attractive for epigenetic analyses of potential underlying mechanisms regulating these changes (Álvarez-Errico et al., 2015; Bock et al., 2012; Bröske et al., 2009; Deaton et al., 2011; Hodges et al., 2011; Stewart et al., 2015). Comprehensive genome wide DNA methylation studies of Dnmt1 knockout mice demonstrated the critical role DNA methylation plays during hematopoietic differentiation (Bröske et al., 2009; Hogart et al., 2012; Ji et al., 2010). Reduced activity of Dnmt1 was then found to lead to a selective loss of lymphoid differentiation in vivo (Bröske et al., 2009) and lymphoid progenitors showed hypermethylation at TF binding sites known to be important for neutrophil-macrophage production (Ji et al., 2010). Taken together, these results demonstrated the importance of DNA methylation in preserving the production of lymphoid progenitors as well as those of the neutrophil and macrophage lineages in cells that normally share both capabilities. There is also evidence that DNA methylation plays a role at later stages of differentiation (Lee et al., 2001; Makar and Wilson, 2014). For instance, 37 conditional knockout of Dnmt1 in mouse CD8 T cells was found to lead to an increased expression of cytokines normally produced by CD4 T cells (Makar and Wilson, 2014). Unlike DNA methylation, analyses of the histone modifications in different types of primitive hematopoietic cells have been studied less extensively due to the continuing requirement for large numbers of rare cells. However, recent comprehensive profiling of H3K4me1 and H3K4me2 marks throughout the mouse hematopoietic hierarchy has revealed the appearance of lineage-related regulatory enhancers in normal adult progenitor populations (Lara-Astiaso et al., 2014). This suggests that lineage restriction and initiation of regulatory programs actually precede biologically relevant RNA expression. Although, RNA evidence of lineage “priming” was noted in cells with inferred multi-lineage potential had been noted many years earlier (Billia et al., 2001; Hu et al., 1997). In addition to primed enhancers in early progenitors, the study of Lara-Astiaso et. al. (Lara-Astiaso et al., 2014) also suggested that some lineage specific enhancers are also established de novo in later stages of hematopoiesis. Other chromatin modifications have also been shown to have relevance to the regulation of hematopoietic cell differentiation. For example, bivalent promoters in human hematopoietic progenitors (CD34+CD133+) were shown to generally lose most H3K4me3 and retain most H3K27me3 modifications (Cui et al., 2009). However, bivalent promoters identified in progenitor populations using conventional ChIP-seq methodologies may also reflect heterogeneously marked regions especially when annotating chromatin states in a highly heterogeneous population (like that of the bulk CD34+ CD45+ cells obtained from normal human CB, or adult PB or BM). Collectively, these studies highlight the importance of the epigenetic landscape in lineage specific gene regulation during hematopoiesis. Nevertheless, results currently being accrued are thought to suggest that regulatory programs are activated prior to lineage restriction, followed by chromatin 38 remodeling and localized silencing of lineage-specific genes to enact the restriction process. At the same time, the dynamics of when and where permissive and repressive chromatin modifications are established during the early stages of human hematopoiesis remain poorly understood. AML is a malignancy in which there is a deregulated clonal overproduction of hematopoietic cells that fail to differentiate normally and also suppress the production of normal blood cells, although the primitive normal cells are not eliminated by the disease itself. It is also the most common acute leukemia affecting adults and remains poorly treated. Disruption in the chromatin state and DNA methylation of hematopoietic cells due to deregulation of epigenetic modifiers sets the stage for the development of AML (Ley et al., 2013). Recurrent somatic genetic lesion in the genomes of AML blasts are thought to widely affect signal transduction genes, TFs, and epigenetic modifiers (Breitenbuecher et al., 2009; Khasawneh and Abdel-Wahab, 2014; Zeisig et al., 2011). Recent advances in sequencing technology have led to the identification of a series of highly recurrent gain and loss of function mutations in epigenetic modifiers in AML blast genomes (Abdel-Wahab et al., 2011). Such regulators include polycomb-related protein additional sex combs like transcription regulator 1 (ASXL1), DNMT3A, MLL, TET2, IDH1 and IDH2 (Ley et al., 2013). In addition, somatic genetic lesions affecting epigenetic modifiers (for example DNMT3A and TET2) are observed in clonal expansion of hematopoietic progenitors in patients above 65 year of age (Genovese et al., 2014). These individuals are predisposed to develop AML, suggesting that mutation in epigenetic modifiers and subsequent rewiring of epigenome play a critical role in early stages of AML development. IDH mutations are recurrent genetic lesions in blood cell transformation, suggesting that there are unique features in the epigenomes of blood cells that make them more susceptible to IDH 39 mutations than other tissue types beside brain (Dang et al., 2009; Ley et al., 2013). Mutant IDH enzymes directly inhibit of 2-oxygluterate dependent di-oxygenases (2-OGDD) class of epigenetic modifiers (i.e. TETs and JmjC histone modifiers) by producing D2HG, a competitive inhibitor to these modifiers. Increases in DNA methylation, H3K9me3 and H3K72me3 has been associated with presence of D2HG (Lu et al., 2012; Noushmehr et al., 2010; Turcan et al., 2012). In addition, a unique characteristic of IDH mutant AML is the CGI methylator phenotype (CIMP) (Figueroa et al., 2010). This suggests that IDH neomorphic mutations lead to localized reprograming of the epigenetic landscape of primary AML through inhibition of 2-OGDD epigenetic modifiers, creating in turn, a unique DNA methylation landscape in AML. 1.7 Rationale and Specific Aims An orchestrated interplay of TFs, epigenetic modifications and extrinsic factors controls cellular differentiation and development (Stricker et al., 2016). The critical role of epigenetic modifications during development, lineage restriction and the activation, and differentiation of HSC dictates the importance of epigenetic modifiers (Ji et al., 2010; Kamminga et al., 2006; Karantanos et al., 2016; Li et al., 2014; Petruk et al., 2017; Tenen, 2003). Manipulation of epigenetic modifiers influence the differentiation capacity of hematopoietic progenitors (Karantanos et al., 2016; Mochizuki-Kashio et al., 2011; Petruk et al., 2017) and somatic mutation of several epigenetic regulators create expanded clones of “normal” hematopoietic cells that predispose, albeit weakly to the development of AML. Current evidence in mouse models points to the interaction of specific TFs with so-called lineage-specific regulatory regions of the DNA to promote the restriction of primitive cells and activate lineage-specific gene expression programs in their progeny. 40 However, the dynamics of epigenetic modifications that take place during the early stages of human hematopoiesis and their disruption during blood cell malignant transformation are still poorly understood. In this thesis, I sought to test the hypothesis that different epigenetic landscape changes occur when primitive human hematopoietic progenitors become restricted to the lymphoid and myeloid lineages and that their perturbation of these can be associated with altered normal differentiation trajectories. To address this gap, I first sought to improve the sensitivity of available methods that have been a barrier to obtaining detailed epigenome data on rare cells (Aim 1, Chapter 3). I then used the improvements obtained to generate a rich resource of transcriptome, methylome and histone modifications for many of the progenitor and terminally differentiating types of normal human hematopoietic cells isolated from primary sources of normal CB and adult BM (Aim 2, Chapter 4). These findings were then compared with parallel data generated from leukemic blasts obtained from AML patients with and without neomorphic IDH mutations (Aim 3, Chapter 5). 41 Table 1. DNA methylation proteins in cancer Protein Function Alteration Cancer Types Reference DNMT1 Maintenance of DNA methylation Mutation Colorectal cancers, AML (Dolnik et al., 2012; Kanai et al., 2003) Upregulation Bladder cancer, gastric cancer, breast cancer (Ben Gacem et al., 2012; Wu et al., 2011a; Yang et al., 2011) DNMT3A de novo DNA methylation Mutation AML,MDS,CMML,ALL,Lung cancer (Jankowska et al., 2011; Kim et al., 2013; Walter et al., 2011; Yan et al., 2011) Upregulation Ovarian cancer, gastric cancer, pancreatic cancer, breast cancer (Ben Gacem et al., 2012; Samudio-Ruiz and Hudson, 2012; Yang et al., 2011; Zhang et al., 2012) DNMT3B de novo DNA methylation Upregulation Gastric cancer, pancreatic cancer, breast cancer, AML (Ben Gacem et al., 2012; Hayette et al., 2012; Yang et al., 2011; Zhang et al., 2012) MeCP2 Methylcytosine binding Upregulation Prostate cancer (Bernard et al., 2006) MBD1 Methylcytosine binding Upregulation Pancreatic cancer (Xu et al., 2013) MBD2 Methylcytosine binding Upregulation Glioblastoma (Zhu et al., 2011) TET1 DNA hydroxylase Mutation, translocation AML (Dolnik et al., 2012; Kanai et al., 2003; Lorsback et al., 2003) TET2 DNA hydroxylase Mutation AML, CMML, MPN, MDS (Abdel-Wahab et al., 2009; Dolnik et al., 2012; Kanai et al., 2003; Tefferi et al., 2009) Downregulation Melanomas, CML (Albano et al., 2011; Lian et al., 2012) IDH1 Isocitrate dehydrogenases Mutation AML, CMML,Gliomas, Cholangiocarcinoma, Prostate cancer (Borger and Zhu, 2012; Ghiam et al., 2012; Mardis et al., 2009; Parsons et al., 2008; Ward et al., 2010) IDH2 Isocitrate dehydrogenases Mutation AML, CMML, Gliomas, Cholangiocarcinoma (Borger and Zhu, 2012; Marcucci et al., 2010; Ward et al., 2010) Downregulation Melanomas (Lian et al., 2012) 42 Table 2. Histone modification proteins in cancer Modification Protein Function Alteration Cancer types Effect Reference Histone acetylation (H3K9ac) GCN5 Acetyltransferase Overexpression Lung cancer, resistance breast cancer Increase in tumor size, facilitates growth (Chen et al., 2013a; Toth et al., 2012) Histone acetylation (H3K9ac) PCAF Acetyltransferase Down-regulation, hypermethylation at promoter, partial deletion esophageal squamous cell carcinoma Suppression of p21 (Zhu et al., 2009) Histone acetylation (H3K18/27ac) P300 Acetyltransferase Upregulation Melanoma Enrichment of H3 acetylation at promoter of cell cycle regulatory genes (Yan et al., 2013) Histone acetylation (H3K18/27ac) CBP Acetyltransferase Mutation Large B-cell lymphoma Disruption of acetylation of p53 and global reduction in H3K18ac (Pasqualucci et al., 2011a) Histone acetylation (H4K16ac) hMOF Acetyltransferase Downregulation/upregulation Renal carcinoma, breast cancer, medulloblastoma, non-small cell lung cancer Induce entry to S phase (Pfister et al., 2008; Wang et al., 2013; Zhao et al., 2013) Histone de-acetylation HDAC2 Deacetylase Upregulation Colon cancer, lung cancer, neuroblastoma Initiation of tumorogenesis, disruption of apoptosis (Jung et al., 2012; Marshall et al., 2010; Stypula-Cyrus et al., 2013) Histone de-acetylation HDAC1 Deacetylase Aberrant recruitment AML Disrupts differentiation (Maiques-Diaz et al., 2012) Histone acetylation reader BRD4 Histone Acetylation recognition Downregulated, heypermethylation at promoter region Colon cancer Facilitation of tumor growth (Rodriguez et al., 2012) 43 H3K4 methylation MLL1 Histone methyl transferase Fusion AML Upregulation of HOX genes (Grembecka et al., 2012; Méreau et al., 2013; Thiel et al., 2010) H3K79 methylation DOT1L Histone methyl transferase AML It is required for MLL-AF10, MLL-AF6 (Chen et al., 2013b) H3K27 methylation EZH2 Histone methyl transferase Gain of function mutation, Inactivating mutation/ Upregulation Diffuse large B-cell lymphoma, MDS, Non-small lung cell cancer, Prostate cancer stem cell Increase in H3K27me3, Decrease in H3K27me3 and upregulation of HOXA genes, enhancement of tumor growth (Ganji et al., 2012; Khan et al., 2013; Kikuchi et al., 2012; McCabe et al., 2012) H3K9 methylation G9a Histone methyltransferase Upregulation Prostate cancer Stabilize the chromatin and promote tumor growth (Kondo et al., 2008) H3K36 methylation NSD1 Histone methyl transferase Fusion AML Upregulation of HOXA genes (Wang et al., 2007) H3K4/9 demethylase LSD1 Histone demethylase Downregulation/upregulation Breast cancer/non-AML(t(15;17)) Activation of TGFB1/Downregulation of genes associated with myeloid differentiation (Schenk et al., 2012; Wang et al., 2009b) H3K4 demethylase JARID1A Histone demethylase Fusion acute megakaryoblastic leukemia Upregulation of HOXA and B (De Rooij et al., 2013) H3K4 demethylase JARID1C Histone demethylase Inactivating mutation Clear cell renal carcinoma Global increase in H3K4me3 (Dalgliesh et al., 2010; Niu et al., 2012) H3K9/36 demethylase JMJD2A Histone demethylase Upregulation Squamous cell carcinoma Induce metastasis (Ding et al., 2013) H3K27 demethylase JMJD3 Histone demethylase Upregulation MDS upregulation of NF-kB and (Wei et al., 2013) H3K27 demethylase UTX Histone demethylase Mutation Multiple myeloma, squamous cell carcinoma, renal carcinoma Inactivating mutation (Van Haaften et al., 2009) 44 Histone methylation readers PHF Reader Fusion AML Protects H3K4me3 at HOX gene promoter (Wang et al., 2009a) Histone methylation readers ING2 Reader Upregulation Colon cancer Upregulation of MMP13 and metastasis (Kumamoto et al., 2009) Histone methylation readers BMI1 Reader Upregulation Pancreatic adenocarcinoma Tumor growth and metastasis (Proctor et al., 2013) 45 Chapter 2: Materials and Methods 2.1 Cells Purified subsets of viable normal human CB cells were provided by Dr. C. Eaves who obtained the original CBs with informed consent from women undergoing normal full-term delivery according to UBC REB and the Women and Children’s Hospital of British Columbia. The CB cells used for the work described in this thesis were from 8 different large pools of CB samples. Samples of consented, anonymized normal adult bone marrow (BM) and AML patients’ leukemic blasts were obtained and processed, respectively by Drs. Eaves and Karsan from the UBC REB-approved Hematology Cell Bank, Purified subsets of viable normal cells (DAPI-negative and >90%purity) were isolated by established procedures (Knapp et al Blood 2019) using cell surface antigens as itemized in Table 3. AML samples were genotyped (Refer to Chapter 5, Figure 30C) using the clinical myeloid panel at the Centre for Clinical Genomics (CCG) of BC Cancer. Cryopreserved AML cells were partially thawed on ice, then resuspended in PBS+20% FBS and viable (propidium iodide-negative) cells and snap-frozen. H1 and H9 cells that had been maintained in mTeSRTM1 and then dissociated in Gentle Cell Dissociation reagent followed by Accutase were a gift from STEMCELL. HL60 cells were originally obtained from Dr. Eaves’ lab and maintained in RPMI medium (StemCell, 36750) with 10% fetal bovine serum (FBS, Sigma, F1051) with 1X Penicillin-Streptomycin (Gibco, Life Technologies, Fisher Thermo. 15140122) and 1X GlutaMAX (Gibco, Life Technologies, Fisher Thermo. 35050061). 46 2.2 Low Input Native ChIP-seq Cells were lysed in mild non-ionic detergents (0.1% Triton X-100 and Deoxycholate) and protease inhibitor cocktail (Calbiochem) in order to preserve the integrity of histones harboring epitopes of interest during cell lysis. Cells were then digested by Microccocal nuclease (MNase) at room temperature for 5 minutes and 0.25 mM EDTA added to stop the reaction. Antibodies, H3K4me1 (Diagenode: Catalogue# pAb-037-050, lot# A1657D), H3K27me3 (Diagenode: Catalogue# pAb-069-050, lot# A1811-001P), H3K9me3 (Diagenode: Catalogue# pAb-056-050, lot# A1675-001P), H3K4me3 (Cell Signaling Technology: Catalogue# 9751S, lot# 8), and H3K36me3 (Abcam: Catalogue# Ab9050, lot# GR220921-1) were incubated with anti-IgA magnetic Dynabeads (Invitrogen) for 2 hours. Digested chromatin was incubated with magnetic beads alone for 1.5 hours. Digested chromatin was then separated from the beads and incubated with antibody-bead complex overnight in IP buffer (20mM Tris-HCl pH 7.5, 2mM EDTA, 150mM NaCl, 0.1% Triton X-100, 0.1% Deoxycholate). Immunoprecipitates were washed twice in a low (20mM Tris-HCl pH 8.0, 2mM EDTA, 150mM NaCl, 1% Triton X-100, 0.1% SDS) and high salt (20 mM Tris-HCl pH 8.0, 2 mM EDTA, 500 mM NaCl, 1% Triton X-100, 0.1% SDS) wash buffers. Immunoprecipitates were then eluted in 1% SDS in a 100 mM sodium bicarbonate solution for 1.5 hour at 65oC. Histones were digested by Protease (Invitrogen) for 30 min at 50oC and DNA fragments purified using Sera Mag magnetic beads in 30% polyethylene glycol (PEG). Illumina sequencing libraries were generated by end repair, 3’ A-addition, and Illumina sequencing adaptor ligation (New England BioLabs, E6000B- 10). Libraries were then indexed and PCR amplified (10 cycles) and sequenced on Illumina HiSeq 2500 sequencing platform following the manufacture’s protocols (Illumina, Hayward CA.). Sequence reads were aligned to GRCh37- lite using Burrows-Wheeler Aligner (BWA) 0.5.7. (Li and Durbin, 2010) and converted 47 to bam format by SAMtools (version 0.1.13). Sequence reads with BWA mapping quality scores <5 were discarded and reads that aligned to the same genomic coordinates were counted only once in the profile generation. The derived sequence data are available at The European Genome-phenome Archive under accession EGAS00001001681. 2.3 Nucleosome Density(nd) ChIP-seq Integrative Analysis Statistically significantly enriched regions were identified by Model-based analysis of ChIP-seq version 2 (MACS2) peak caller (Feng et al., 2012) using a corrected q-value of 0.01 and 0.05 for narrow and broad peaks, respectively. Immunoprecipitated fragment size distributions at these enriched regions and use of a Gaussian mixture distribution algorithm enabled a weighted distribution at each promoter to be calculated using the Mclust version 3.0 R-statistical package (Fraley and Raftery, 2002). The optimized MNase digestion led to a Gaussian mixture distribution with multiple components for which the weights (w) and Gaussian distributions yielded mean values corresponding to mono-, di-, tri- and etc. fragment lengths, assuming a single nucleosome would be represented by 100-220 bp fragments and two nucleosomes by 280-600 bp fragments. The area under all of these distributions was set equal to the area under the total distribution so that the total distribution had the weight of 1 and w1+ w2+w3+...+wn=1, with w1 representing the weight of a mono-nucleosome fragment distribution and w2 is for di-nucleosome distribution, etc. Therefore, the dominant chromatin structure in the majority of the population would be represented by its w value. Promoters with a mono-nucleosome weighted distribution >0.6 were assigned as mono-nucleosome dominated promoters, and those with di-nucleosome weighted distribution >0.6 were assigned as di-nucleosome dominated promoters, etc. Promoters with weighted distribution greater than 0.5 and smaller than 0.6 were excluded from subsequent analysis. 48 2.4 ChIP-seq The ChIP-seq data were generated as a part of the Canadian Epigenetics, Environment and Health Research Consortium (CEEHRC) activity. Standard operating procedures for ChIP-seq library construction are available (http://www.epigenomes.ca/protocols-and-standards) or by request. In brief, frozen cell pellets were cross linked by formaldehyde (1% final concentration) and washed by PBS. Cells were lysed and chromatin was sheared by Sonic Dismembrator 550 (Fisher Scientific). Chromatin fragments were incubated with protein G and A Sepharose beads (GE Healthcare) for two hours at 4oC to eliminate non-specific binding. Unbound chromatin fragments were separated from the beads and incubated with beads and the following antibodies; H3K4me1 (Diagenode: Catalogue# BP140, lot# A1863-001P), H3K27me3 (Diagenode: Catalogue# BP50, lot# A1811-001P), H3K9me3 (Diagenode: Catalogue# pAb-056-050, lot# A1675-001P), H3K4me3 (Diagenode: Catalogue# BP1, lot# A5051-001P), and H3K36me3 (Abcam: Catalogue# BP41, lot# A1847-001P) at 4oC overnight in IP buffer (10mM Tris-HCl pH 7.5, 2mM EDTA, 90mM NaCl, 0.1% Triton X-100, 0.1% Deoxycholate, 0.1% SDS). IPs were washed twice with ChIP wash buffer (20mM Tris-HCl pH 8.0, 2mM EDTA, 150mM NaCl, 1% Triton X-100, 0.1% SDS) and once with Final ChIP wash buffer (20 mM Tris-HCl pH 8.0, 2 mM EDTA, 500 mM NaCl, 1% Triton X-100, 0.1% SDS). Immunoprecipitates were eluted in elution buffer (1% SDS, 100 mM sodium bicarbonate) and incubated at 68oC for 2 hours. DNA fragments were stripped from histones and purified using QIAquick PCR Purification Kit (Qiagen). Illumina sequencing libraries were generated as previously described in Section 2.2. Libraries were PCR-amplified and sequenced on Illumina HiSeq 2000/2500 sequencing platforms following the manufacture’s protocols (Illumina, Hayward CA.). Sequence reads were aligned to GRCh37-lite 49 using Burrows-Wheeler Aligner (BWA) 0.5.7. (Li and Durbin, 2010) and converted to bam format by SAMtools (version 0.1.13). Duplicate reads were marked using Picard Tools’ MarkDuplicates.ja (version 1.71). Data are available through the following accessions and resources: Reference Epigenome Registry Accession, IHECRE00000237.1 and The European Genome-phenome Archive Accession EGAS00001000552, http://www.epigenomes.ca). Processed datasets are available through http://www.epigenomes.ca/data-release/ and http://epigenomesportal.ca/ihec/ under accession CEMT32). 2.5 Sequential ChIP-seq 600,000 human CD34+ CB cells were lysed in mild non-ionic detergents (0.1% Triton X-100 and Deoxycholate) and protease inhibitor cocktail (Calbiochem). Cells were digested by Microccocal nuclease (MNase) at 37oC for 9 minutes and 0.25 mM EDTA used to stop the reaction. A first round of IP was performed as described before (see Section 2.2) using the antibody against H3K4me3. After the first IP, chromatin was eluted in a solution of 30 mM DTT, 500 mM NaCl, and 0.1% SDS at 37° for 30 minutes. Eluted chromatin was diluted 30-fold with IP buffer (20mM Tris-HCl pH 7.5, 2mM EDTA, 150mM NaCl, 0.1% Triton X-100, 0.1% Deoxycholate) and subjected to a second IP using the antibody against H3K27me3. A second round of IP and elution were performed as described before (see Section 2.2). Sequential IP was also performed with the first IP against H3K27me3 and second IP against H3K4me3. The derived ChIP-seq datasets are available at The European Genome-phenome Archive under accession EGAS00001001681. 50 2.6 WGBS Library Construction WGBS data were generated as a part of CEEHRC and are available through the following accessions and resources; Reference Epigenome Registry Accession, IHECRE00000237.1 and The European Genome-phenome Archive Accession EGAS00001000552, http://www.epigenomes.ca). To track the efficiency of bisulfite conversion, 1 ng of lambda DNA (Promega) was spiked into 1 μg genomic DNA quantified using Qubit fluorometric and arrayed in a 96-well microtiter plate. DNA was sheared to a target size of 300 bp using Covaris sonication and the fragments used for sequencing library construction by end repair (New England BioLabs, E6000B-10), 3’ A-addition (New England BioLabs, E6000B- 10), and cytosine methylated paired-end adapters (5’- AmCAmCTmCTTTmCmCmCTAmCAmCGAmCGmCTmCTTmCmCGATmCT-3’ and 3’- GAGmCmCGTAAGGAmCGAmCTTGGmCGAGAAGGmCTAG-5’) ligation. Bisulfite conversion of the methylated adapter-ligated DNA fragments was achieved using the EZ Methylation-Gold kit (Zymo Research) following the manufacturer’s protocol. Five cycles of PCR using Kapa HiFi U+ polymerase (Kapa Biosystems) was used to enrich for the bisulfite converted DNA and introduce fault tolerant hexamer barcode sequences. Post-PCR purification and size-selection of bisulfite converted DNA was performed from pre-cast 8% TBE gels (Invitrogen), extracting the 350-500 bp fraction, or the 275-425bp fraction if the former was of weak intensity. Gel slurries were added to Spin-X filter tubes (Fisher) and the eluate ethanol precipitated and resuspended in EB for sequencing. Libraries were sequenced on Illumina Hiseq 2000/2500 sequencing platforms following the manufacture’s protocols (Illumina, Hayward CA.). Three lanes of Illumina HiSeq 2500 data were aligned to GRCh37-lite reference using Novoalign (version 3.01.00). Following merging, a weighted methylation mean was calculated as total 51 unconverted reads divided by the total number of reads within a promoter region using custom in house scripts. CpG density was calculated as a ratio of CpG to total number of bases within a promoter. Methylation data for H1 ESCs (GSM432685 and GSM432685) and H9 ESCs (GSM706059, GSM706060, and GSM706061) was obtained from the Roadmap Epigenomics Consortium (Kundaje et al., 2015). 2.7 RNA-seq RNA-seq data were generated as a part of CEEHRC and are available through the following accessions and resources: Reference Epigenome Registry Accession, IHECRE00000237.1 and The European Genome-phenome Archive Accession EGAS00001000552, http://www.epigenomes.ca). Standard operating procedures for RNA-seq library construction are available (http://www.epigenomes.ca/protocols-and- standards) or by request. Libraries were sequenced on a HiSeq 2500 platform following the manufacture’s protocols (Illumina, Hayward CA.). Sequence reads were aligned to a transcriptome reference generated by JAGuaR (version 2.0.2) (Butterfield et al., 2014) using the GRCh37-lite reference genome supplemented by read-length-specific exon–exon junction sequences (Ensemble v69 gene annotations) and bam files were generated using SAMtools (version 0.1.13). The resulting bam files were repositioned to GRCh37-lite by JAGuaR (Butterfield et al., 2014). To quantify exon and gene expression reads per kilobase per million mapped reads (RPKM) metrics was calculated. Processed datasets are available at http://www.epigenomes.ca/data-release/ and http://epigenomesportal.ca/ihec/ under accession CEMT32). RNAseq data for H1 (GSM438361) and H9 (GSM706044) were obtained from the Roadmap Epigenomics Consortium (Kundaje et al., 2015). 52 2.8 Ribo-depleted RNA-seq Total RNA was rRNA depleted using NEBNext rRNA Depletion Kit (New England BioLabs, E6310L). 1st strand cDNA was generated using Maxima H minus First Strand cDNA Synthesis Kit (Thermo Scientific, K1652) with the addition of 1 ug of Actinomycin D (Sigma, A9415). The product was purified using in-house prepared 20% PEG, 1M NaCL Sera-Mag bead solution at 1.8X ratio and eluted in 35 µL of Qiagen EB buffer. Second Strand cDNA was synthesized in a 50 µL volume using SuperScript Choice System for cDNA Synthesis (Life Technologies, 18090-019) with 12.5 mM GeneAmp dNTP Blend with dUTP. Double stranded cDNA was then purified with 20% PEG, 1M NaCL Sera-Mag bead solution at 1.8X ratio, eluted in 40 µL of Qiagen EB buffer, and fragmented using Covaris E220 (55 seconds, 20% duty factor, 200 cycles per burst). Sheared cDNA was End Repaired/Phosphorylated, single A-tailed, and Adapter Ligated using custom reagent formulations (New England BioLabs, E6000B-10) and in-house prepared Illumina forked small adapter. 20% PEG, 1M NaCl Sera-Mag bead solution was used to purify the template in-between each of the enzymatic steps. To complete the process of generating strand directionality, adapter-ligated template was digested with 5 U of AmpErase Uracil N-Glycosylase (Life Technologies, N8080096). Libraries were then PCR-amplified and indexed using Phusion Hot Start II High Fidelity Polymerase (Thermo Scientific, F 549-L). An equal molar pool of each library was sequenced on HiSeq2500 (Illumina) PE75. Libraries were sequenced on a HiSeq 2500 platform following the manufacture’s protocols (Illumina, Hayward CA.). Sequence reads were aligned to a transcriptome reference generated by JAGuaR (version 2.0.2)(Butterfield et al., 2014) using the GRCh37-lite reference genome supplemented by read-length-specific exon–exon junction sequences (Ensemble v75/v69 gene annotations) and bam files generated using Sambamba (version 0.5.5)(Tarasov et al., 2015). The resulting bam files were 53 repositioned to GRCh37-lite by JAGuaR. To quantify exon and gene expression, reads per kilobase per million,mapped read (RPKM) metrics were calculated. A standardized analytical pipeline introduced by CEEHRC and the International Human Epigenomic Consortium (IHEC) (http://ihec-epigenomes.org/) was applied to qualify the resulting data. Pairwise differential expression analysis was carried out using an inhouse MATLAB script, DEFine, on RPKM values that are corrected for GC-biased transcripts. MetaScape version 2.0 (Tripathi et al., 2015) (http://metascape.org) was used for genome enrichment analysis of differentially expressed genes. All figures were generated by R statistical software (Team and R Development Core Team, 2016). 2.9 Post Bisulfite Adaptor Ligation (PBAL) PBAL libraries were constructed as previously described (Hui et al., 2018). In brief, genomic DNA was sheered and subjected to bisulfite conversion using the MethylEdge Bisulfite Conversion kit (Promega, N1301) using a bead-based automated protocol. Bisulfite-converted DNA was mixed with 180 μl of MethylEdge Binding Buffer and 1.8 μl of 20 mg/ml decontaminated MagSi-DNA all-round silica beads (MagnaMedics, MD02018) and left at room temperature for 15 minutes and washed twice with 220 μl of 80% ethanol for 30 seconds. 60 μl of MethylEdge desulfonation buffer was added to the beads and incubated at room temperature for 15 minutes, then washed twice with 100 μl of 80% ethanol and dried for 1 minute. To elute the DNA, 20 μl 10 mM Tris-HCL, pH 8.5 (Qiagen, 19086) was added to the DNA-bead mix and incubated in a Thermomixer C (Eppendorf, 5382000015) at 56°C while being centrifuged at 2,000 rpm for 15 minutes. The bisulfite-converted single-stranded DNA was converted to double stranded DNA through 1 cycle of PCR with random hexamers as previously described (Hui et al., 2018) followed by standard illumina library construction. Libraries were aligned using Novoalign 54 V3.02.10 (www.novocraft.com) to human genome assembly GRCh37 (hg19). Duplicate reads were marked by Picard V1.31 (http://picard.sourceforge.net). and discarded (http://picard.sourceforge.net). Methylation calls were generated using Novomethyl V1.01 (www.novocraft.com). 2.10 HL60 treatment with ATRA, EPZ-6438 and GSK-J4 HL60 cells were cultured at 500,000 cells/mL in RPMI (StemCell, 36750) with 10% FBS (Sigma, F1051) with 1X Penicillin-Streptomycin (Gibco, Life Technologies, Fisher Thermo. 15140122) and 1X GlutaMAX (Gibco, Life Technologies, Fisher Thermo. 35050061) in the presence of 1 µM ATRA (Sigma, R2625)± 20 µM EPZ-6438 (Selleckchem, S7128) or 1 µM GSK-J4 (Sigma, 420205) in 96-well U-shape bottom plates for 9 days. Cells were treated with each compound every 3 days and reset to original seeding density. Concentration and viability were assessed every 3 days using a Countess II automated cell counter (Thermo Finsher, AMQAX100). 2.11 Quantitative measurement of CD11b+ HL60 cells HL60 cells were cultured at 1,000,000 cells/mL in RPMI (StemCell, 36750) with 10% FBS (Sigma, F1051) with 1X Penicillin-Streptomycin (Gibco, Life Technologies, Fisher Thermo. 15140122) and 1X GlutaMAX (Gibco, Life Technologies, Fisher Thermo. 35050061) in presence or absence of 0.1 µM ATRA (Sigma, R2625), 20uM Tazemetostat (Selleckchem, S7128) ± 1 µM GSK-J4 (Sigma, 420205) in 96 well U-shaped bottom plates for 48 and 72 hours. The concentration of DMSO in control wells was adjusted to the maximum level (1.1%) used in treated wells. After 48 and 72 hours, all cells were harvested, suspended in Hanks' Balanced Salt Solution (HBSS; STEMCELL) supplemented with 2% FBS and 1.5 µg/mL anti-human CD32 antibody 55 (Clone IV.3; STEMCELL). Cells were stained for 20-30 minutes on ice with a 1:200 dilution of anti-CD11b-BV711 (Clone M1/70, Biolegend) prior to analysis by flow cytometry. Flow cytometry data were analyzed in R using the package flowCore (Hahne et al., 2009) and custom scripts. 2.12 Stromal cell-containing cultures of CB cells CB cells pooled from 3 donors were thawed, stained with antibodies. For “bulk” cultures, 50 cells with a CD45highCD34highCD38midCD71-CD10- (P-NML) phenotype isolated by fluorescent activated cell sorting (FACS) were deposited into each well of a 96 well flat-bottomed plate preloaded with 100,000 MS-5 stromal cells and alpha-MEM medium with 2  mM glutamine, 7.5% FBS and 10−4 M β-mercaptoethanol (Sigma) plus 50 ng/ml human Stem Cell Factor (SCF, a gift from Novartis), 10 ng/ml human FLT3-Ligand (FLT3L, a gift from Immunex), 1 ng/ml human Inteleukin-3 (IL-3, a gift from Novartis) , and 3 units/mL of human erythropoietin (EPO, a gift from STEMCELL) in the presence of 1 µM GSK-J4 or 0.4 µM EPZ-6438, as indicated. The same dose of inhibitor was added to each half medium change performed after 4 and 7 days, respectively. Cells were cultured with 50 ng/ml SCF, 10 ng/mL FLT3L, 10 ng/ml IL-7 (R&D Systems), and 3 units/ml EPO for the second week, and 10 ng/ml IL-7 and 3 units/ml EPO only for the third week. After 3 weeks, all cells were harvested, stained with antibodies, and assessed by FACS to detect cultures containing >10 cells belonging to one or more of the following cell populations: monocytes (CD45+CD33+CD14+ cells), neutrophils (CD45+CD33+CD15+ cells) and B cells (CD45+CD33-CD14-CD15-CD19+ cells). For clonal cultures, single P-NML cells (Knapp et. al. Blood 2019) were deposited into each well of a 96-well plate preloaded with 9,000 MS-5 cells and 333 each of M210B4 mouse BM 56 fibroblasts expressing human IL-3 and G-CSF, and sl/sl mouse fibroblasts expressing human SCF and IL-3, and human FLT3L, with alpha-MEM medium with 2 mM glutamine, 7.5% FBS and 10−4 M β-mercaptoethanol (Sigma) plus 50 ng/mL, SCF (Novartis), 10 ng/mL FLT3L (Immunex) added for the first 2 weeks. Weekly half-medium changes were performed. After 3 weeks, all cells were harvested, stained with antibodies and assessed by FACS to detect clones of >10 monocytes, neutrophils or B cells using the same FACS analysis protocol as for the bulk cultures. Clones containing >10 CD45+ events in the absence of any mature cells were classified as “undefined”. Clones with <10 CD45+ events were classified as “negative”. 2.13 Chapter 3 integrative analysis MACS2 was employed to identify enriched regions with a false discovery rate (FDR) value of ≤0.05 for broad peaks and ≤0.01 for narrow peaks in the ChIP-seq data. Genomic Region Enrichment of Annotations Tool (GREAT) version 2.0 was used to annotate enriched biological processes associated with promoters. Concordance between sequential ChIP-seq and ndChIP-seq was calculated at promoters with a sequence read coverage of >4. ChromHMM was employed to identify nucleosome density states at promoters marked with H3K4me3 and/or H3K27me3 across ESCs and human CD34+ CB cells. H3K4me3 one nucleosome and two nucleosome and H3K27me3 one nucleosome and two nucleosome promoters in H1, H9, and human CD34+ CB cells were treated as separate tracks in the ChromHMM model (Weighted distribuition value ≥0.65). The ChromHMM LearnModel was used to generate 15 chromatin states across ESCs and human CD34+ CB cells. ChromHMM emissions below 0.2 were omitted from further analysis. All figures in this chapter were generated by R statistical software (Team and R Development Core Team, 2016). 57 2.14 Chapter 4 integrative analysis Genome browser tracks were generated by converting bam files to wiggle files using a custom script (http://www.epigenomes.ca/tools-and-software). Wiggle files then converted to bigwig for display on UCSC genome browser by UCSC tool, Wig2Bigwig script. MACS2 was employed to identify regions enriched in H3K4me3, H3K4me1 and H3K27ac with a false discovery rate (FDR) value of ≤0.01 for H3K4me3, H3K4me1 and H3K27ac peaks in ChIP-seq data. Finder 2.0 (http://www.epigenomes.ca/tools-and-software/finder) was used to identify regions enriched for H3K27me3 and H3K9me3 in the ChIP-seq data after eliminating those found to overlap with ENCODE blacklist regions. ChIP-seq signals were calculated as tag densities generated using HOMER v4.10(Heinz et al., 2010) and normalized to the total number of tags within enriched regions. Heat maps of fragment distribution at promoters were generated by deeptools (Ramírez et al., 2016). ChromHMM (Ernst and Kellis, 2012) was used to identify 18 chromatin states based on MACS2-identified enriched regions in each cell type, as previously described (Ernst and Kellis, 2017). To identify active enhancers, we first found MACS2-identified H3K27ac-enriched regions in all populations (CD34+CD38-, CMP, GMP, MEP, Monocyte, Erythroid, B cell and T cell) and created an enhancer catalogue of these cells. Any enriched region that overlapped within -2Kb of a transcription start site (TSS) defined by Ensemble v75 human gene annotation was discarded. All H3K27ac libraries were subsampled to the depth of the library with the lowest depth. Next, enhancers with >40 H3K27ac tags in at least one cell type were selected. The normalized signal H3K27ac tag density was then calculated across all regions in the enhancer catalogue for each cell type. Enhancers with a signal >25 were marked as active in the corresponding cell type. The same strategy was used to identify poised enhancers with regions enriched in H3K4me1 with the exception of a signal threshold of 40. SOM analysis was carried 58 by oposSOM (Löffler-Wirth et al., 2015) R package on ranked normalized signals of H3K27ac at the union of regions enriched in H3K27ac identified in all cell types analyzed. Enhancer gene association was done using bedtools (Quinlan and Hall, 2010) closestBed command with default parameters using Ensemble v75 human gene annotation. Promoters (sites within 2kb of the TSS of protein coding regions) at a distance >50 Kb from enhancer regions were eliminated. GREAT was used for gene enrichment analysis of regulatory regions. To identify H3K27me3 LOCKs, MACS2 broad mode was used to call peaks for H3K27me3 with an FDR cut off of 0.05. Resulting peaks were then entered into Clustering of genomic Regions Analysis Method (CREAM) (Madani Tonekaboni et al., 2019) with a WScutoff of 1.5, MinLength of 1000 and peakNumMin of 2 to identify H3K27me3 LOCKs. All figures in this chapter were generated by R statistical software (Team and R Development Core Team, 2016). 2.15 Chapter 5 Integrative analysis Differential methylated regions (DMRs) were identified using an inhouse script. First, to reduce stochastic sampling bias, coverage from both strands were combined. Next, differential methylated CpGs were identified in a pairwise comparison with 0.6 difference in fractional methylation, minimum fractional methylation for hyper and hypo methylated CpG of 0.75 and 0.25, respectively, minimum coverage of 3 and p-value of 0.005 (f-test). Lastly, DMRs were identified with following criteria: minimum 3 CpG within a region, adjacent CpG < 500bp apart and adjacent CpG must have the same differential status. Fractional methylation at ChromHMM defined chromatin states were calculated by averaging fractional methylation of CpGs within that particular region. ChIP-seq analysis and ChromHMM state calling was done as previously 59 described in Sections 2.13 and 2.14. All figures in this chapter were generated by R statistical software (Team and R Development Core Team, 2016). 60 Table 3. A table of human cell types used in this thesis Cell Source Surface Markers Reference Bulk CB CD34+ Cord Blood CD34+ (Knapp et al., 2019) CB CD34+CD38- Cord Blood CD34+ CD38- (Conneally et al., 1997) Common Myeloid Progenitor (CMP) Cord Blood CD34+CD38+CD10-CD7-CD135+CD45RA- (Knapp et al., 2019) Granulocytes and Macrophages Progenitor (GMP) Cord Blood CD34+CD38+CD10-CD7-CD135+CD45RA+ (Knapp et al., 2019) Megakaryocytes and Erythrocytes Progenitor (MEP) Cord Blood CD34+CD38+CD10-CD7-CD135-CD45RA- (Knapp et al., 2019) Monocyte Cord Blood CD45+CD34-CD11b+CD33+CD14+ (Marimuthu et al., 2018; Ziegler-Heitbrock et al., 2010) Erythroid Cord Blood CD45-CD34-GPA+ (Knapp et al., 2019) B Cell Cord Blood CD45+CD34-CD11b-CD33-CD19+CD7- (Knapp et al., 2019) T Cell Cord Blood CD3+CD4+CCR7+CD45RO-CD25-CD235- (Van Den Broek et al., 2018) Bulk BM CD34+ Adult Bone Marrow CD34+ - BM CD34+CD38- Adult Bone Marrow CD34+CD38- - AML total Blast (>70%) Adult Bone Marrow NA - 61 HL60 Promyelocytic leukemia cell line NA - H1 Human Embryonic Cell line NA - H9 Human Embryonic Cell line NA - 62 Chapter 3: Nucleosome density ChIP-Seq identifies distinct chromatin modification signatures associated with MNase accessibility 3.1 Introduction The spectrum of histone modifications present in different regions establish chromatin states that, in turn, control access of the molecular complexes that allow activation of transcription (Dunham et al., 2012; Jenuwein and Allis, 2001; Roadmap Epigenomics Consortium et al., 2015). Histone modifications can occur as singular alterations (Allfrey and Mirsky, 1964), but multiple modifications have also been found on the tail of a single histone molecule (Bernstein et al., 2006; Matsumura et al., 2015; Sachs et al., 2013). One example of co-occurring histone modifications is the combination of H3K27me3 and H3K4me3 at promoters of genes associated with developmentally important genes in ESCs (Bernstein et al., 2006; Voigt et al., 2013), most of which (~75%) become monovalent in their differentiated progeny (Mikkelsen et al., 2007). However, bivalency is not a unique feature of ESCs as they have also been identified in chromatin immunoprecipitation-sequencing (ChIP-seq) profiles of other cell types (Mikkelsen et al., 2007; Roadmap Epigenomics Consortium et al., 2015). In addition, sophisticated cell-sorting and single molecule imaging techniques have indicated that at least a portion of bivalent promoters identified in populations of cells by ChIP-seq are monovalent in different component cell subsets (Hong et al., 2011; Shema et al., 2016). However, the extent to which such heterogeneity in histone modifications of promoter nucleosomes exists at the single-cell level remains generally unknown, and the range of functional consequences of variability in promoter nucleosome modification has also remained undefined. 63 DNA methylation can be both associated with and mutually exclusive to specific histone modifications. For example, DNA methylation is not found on genomic regions marked by H3K4me3 (Ooi et al., 2007). On the other hand, de novo methylation is preferentially targeted to genomic regions with elevated H3K36me3 levels (Baubec et al., 2015). The relationship between DNA methylation and H3K27me3 is also complex and not fully understood. Direct interaction of the catalytic subunit of the PRC2 complex, EZH2 with DNMT3A has been reported (Rush et al., 2009), as has the recruitment of this complex to H3K27me3-marked promoters in the absence of DNA methylation (Jermann et al., 2014). More recently, epigenomic profiles of a collection of primary human cell types revealed H3K27me3-enriched chromatin states associated with both hyper- and hypomethylated DNA, suggesting that additional factors contribute to the methylation state of H3K27me3-marked regions of DNA (Roadmap Epigenomics Consortium et al., 2015). Given that both nucleosome density and histone modification play a role in the local control of gene transcription, I sought here to investigate the relationship between bivalent histone modifications and nucleosome density utilizing a native ChIP approach. 3.2 Results 3.2.1 Immunoprecipitated DNA fragment sizes correlate with different patterns of histone modification I first modified aspects of previously described ChIP-seq protocols to enable MNase accessibility to be combined with histone modification profiling of reduced numbers of cells (Barski et al., 2007; Maunakea et al., 2010). This protocol utilizes micrococcal nuclease (MNase) digestion to digest DNA preferentially between nucleosomes (Noll and Kornberg, 1977), thereby producing nucleosome-enriched fragments of chromatin that can be leveraged by an analytical 64 framework that exploits their size distributions (Section 2.3). We first optimized the MNase treatment of purified CD34+ cells isolated from a pool of human CB to produce a DNA fragment size distribution profile anticipated to predominantly contain DNA fragments associated with single nucleosomes (Figure 6A; 146-150 bp). We then performed IP on this MNase-digested chromatin using validated antibodies specific for five histone modifications selected for profiling by International Human Epigenome Consortium (IHEC). This revealed a histone modification-associated shift in the sizes of the DNA fragments (Figure 7A and B, single nucleosome peak at ~270 bp due to the addition of adaptors required for sequencing) not previously observed in DNA immunoprecipitated from formaldehyde-crosslinked chromatin (data not shown). Sequencing of the resulting libraries as per IHEC recommendations showed a significant correlation (average Spearman rho = 0.80, Figure 6B, C, D, and Table 4) with the ChIP-seq tracks across all 5 marks in an independently generated reference epigenome from a different, larger pool (106 cells) of similarly isolated normal human CD34+ CB cells by Canadian Epigenetics, Environment and Health Research Consortium (CEEHRC). Correlating immunoprecipitated fragment size, determined by paired-end read boundaries, with ChromHMM-derived chromatin states derived from these cell types (Ernst and Kellis, 2012) revealed unique fragment size profiles for each type of histone modification (Figure 7C, D, E, and 8A). H3K27me3, H3K36me3, H3K4me3 and H3K9me3 profiles showed chromatin state-specific fragment length distributions for each profile (Figure 8A), whereas the distribution of fragments with H3K4me1 remained consistent across all chromatin states (Figure 8A). When observed, the chromatin state-specific fragment lengths were enriched in regions with high signal to noise ratios (and were thus selected for inclusion in the ChromHMM model). In contrast, genome wide distributions were dominated by the input fragment length distribution regardless of the histone modification present (Figure 8B). 65 66 Figure 6. ndChIP-seq from 10,000 cells shows concordance with ChIP-seq performed on 1 million cells. (A) Agilent bioanalyzer traces of micrococcal nuclease (MNase) digestion of 10,000 human cord blood CD34+ cells in triplicate (Replicate1: red, Replicate2: blue, Replicate3: green). (B) UCSC browser view of the HOXA gene cluster showing visual concordance between ndChIP-seq from 10,000 human CD34+ CB cells per IP (10K, lower track in pair) and reference epigenome tracks (CEEHRC) generated from 1 million human CD34+ CB cells per IP (1M, upper track in pair) (C) Heatmap showing the Pearson correlation of H3K4me3 signal between two ndChIP-seq replicates calculated in the promoters of coding genes (TSS+/-1.5Kb). (D) Heatmap showing the Pearson correlation of H3K4me3 signal between ndChIP-seq and the reference epigenome H3K4me3 dataset. 67 68 Figure 7. MNase accessibility is chromatin state-dependent. Shift in H3K27me3 (A) and H3K4me3 (B) fragment sizes (DNA fragment + sequencing adaptor of 125bp) before (blue) and following IP (red) of MNase digested chromatin from human CD34+ cord blood cells. (C) CEEHRC generated ChromHMM chromatin state definitions and histone mark probabilities in CD34+ cells CB. H3K27me3 and Input (red panel) (D) and H3K4me3 and Input (green panel) (E) DNA fragment size distributions at ChromHMM derived chromatin states in CD34+ cord blood cells. (F) Ratio of single- (pink) and two- (red) nucleosomes marked with H3K27me3 at gene promoters. (G) Ratio of single (dark green) and two (green) nucleosomes marked with H3K4me3 at gene promoters. (H) Genome wide densities of DNA fragments containing one or two nucleosomes marked with H3K27me3 (left) and at promoters (right). (I) Genome wide densities of DNA fragments containing one or two nucleosomes marked with H3K4me3 (left) and at promoters (right). 69 70 Figure 8. MNase accessibility at ChromHMM derived chromatin states is modification dependent. (A) H3K27me3, H3K4me3, H3K4me1, H3K9me3, H3K36me3 immunoprecipitated fragment length distributions at ChromHMM determined chromatin states in human CD34+ CB cells. (B) H3K27me3, H3K4me3, H3K4me1, H3K9me3, H3K36me3 immunoprecipitated fragment length distributions genome wide in human CD34+ CB cells. Immunoprecipitated fragment length distribution across ChromHMM determined chromatin states for H3K4me3 (C) and H3K27me3 (D) CEEHRC reference epigenome ChIP-seq datasets derived from crosslinked and sheared chromatin extracted from 1 million human CD34+ CB cells. (E) H3K4me3 (green) and H3K27me3 (red) immunoprecipitated fragment length distributions derived from ndChIP-seq within MACS2-identified regions enriched in human cord blood CD34+ cells. (F) Input DNA fragment length distributions at MACS2-identified regions enriched for H3K4me3 (green) and H3K27me3 (red) immunoprecipitations shown in (E). 71 The variable distribution of fragment sizes associated with specific chromatin states was not evident in the CEEHRC CD34+ CB cell dataset derived from formaldehyde cross-linked and sonicated ChIP-seq (Figure 8C and D). 3.2.2 Nucleosome density ChIP-seq (ndChIP-seq) classifies promoters based on nucleosome density To explore the functional implications of observed differences in immunoprecipitated fragment sizes, we next focused on promoter regions in the CD34+ CB cell ChIP-seq data. We first analyzed the distribution of H3K4me3 and H3K27me3 marks in fragments identified by their size as enriched in either one- (100-220 bp) or two- (280-600 bp) nucleosomes within regions identified as enriched by MACS2 (Figure 8E and F). This analysis revealed a negative correlation between the inferred ratios of fragments associated with one or two nucleosomes selectively at H3K4me3- and H3K27me3-marked promoters (Figure 7F and G). This finding suggests that MNase accessibility is stable and regionally specific; hence likely to reflect variations in nucleosome density associated with specifically modified nucleosomes. We also found a majority of H3K27me3-marked promoters to be dominated by DNA fragment sizes indicative of the presence of two nucleosomes (Figure 7H). In contrast, we found H3K4me3-marked promoters to be dominated by fragment sizes expected to contain a single nucleosome (Figure 7I). To examine these relationships at all promoters, we considered the fragment distribution to be a mixture of two independent Gaussian distributions predicted for DNA fragments associated with one (mean length of ~150 bp) or two nucleosomes (mean length of ~320 bp) (Figure 9A). A weighted value was assigned to each distribution based on the following formula: 𝑤" ∗𝑛(𝑥; 𝜇", 𝜎") + 𝑤. ∗ 𝑛(𝑥; 𝜇., 𝜎.) = 1; where w1 and w2 represent the contributions of the one- 72 nucleosome and two-nucleosome fragment distributions, respectively (Fraley and Raftery, 1999, 2002; Verbeek et al., 2003). We then assigned promoters as predominantly associated with one- or two-nucleosomes based on their prevalence in the total distribution (cut-off at w >0.6). Mixed promoters with w values between 0.5 and 0.6 for either single- or two-nucleosome-associated fragment distributions were excluded from subsequent analysis. 3.2.3 A majority of bivalent promoters in CD34+ CB cells identified by ndChIP-seq are heterogeneous Promoters identified as marked by both H3K4me3 and H3K27me3 were originally found in ESCs, but have since been described in a variety of progenitor and differentiated cell populations (Roadmap Epigenomics Consortium et al., 2015). However, the degree to which such bivalent patterns reflect variations in histone modifications in different cells within the population analyzed (cellular heterogeneity) has not generally been examined (Brookes et al., 2012; Shema et al., 2016). We reasoned that the extent to which bivalent states are indicative of H3K27me3 and H3K4me3 marks on the same nucleosomes would be reflected in the similarity of the distributions of fragments obtained from dually labeled chromatin immunoprecipitates (Figure 9B). Examination of H3K4me3 and H3K27me3 fragment distributions at bivalently marked promoters in our CD34+ CB data revealed distributions that differed significantly (Kolmogorov-Smirnov p-value <3.44 x 10-12, Figure 9C). This finding suggests that most of the promoters in this population identified as bivalent reflect differences in the marks present in individual cells in the population, with promoters in any given single cell marked by only one of the two histone modifications. 73 74 Figure 9. ndChIP-seq identifies heterogeneously marked promoters in individual cells within bulk populations. (A) Schematic of the process of assigning one or two nucleosome dominated promoters. A Gaussian mixture distribution model was used to determine the contributions of DNA fragment sizes to the total fragment distribution expected from one and two nucleosomes at each enriched promoter (B) Distinguishing between bivalent and heterogeneous chromatin states based on immunoprecipitated DNA fragment length distributions from MNase digested chromatin. (C) H3K4me3 (green) and H3K27me3 (red) immunoprecipitated and input (blue) fragment distributions at co-marked promoters. 75 To undertake a broader examination of bivalent promoters in our dataset, we classified all apparently marked (H3K4me3- and/or H3K27me3-enriched) promoters into 4 states: a transcriptionally active state in which either one or two nucleosomes are marked by H3K4me3 (Figure 10A); a repressive state in which either one- or two-nucleosomes are marked by H3K27me3 (Figure 10B); a heterogeneous state consisting of one-nucleosome marked by H3K4me3 and two-nucleosomes marked by H3K27me3 (Figure 10C); and a bivalent state in which two-nucleosomes are present and marked by H3K4me3 and H3K27me3 (Figure 10D). The w values obtained within promoters correlated well between the replicates and the degree of overlap in promoters defined as having single nucleosomes in one replicate and two in another replicate was not significant (Figure 10E, F, and 11A, B, C). Heterogeneous states, defined as likely containing different fragment lengths for H3K4me3 and H3K27me3 marks, were found predominantly at promoters (e.g., FOXC1). A smaller, but nevertheless significant, set of promoters was found to be in a bivalent state and associated with similar fragment length distributions for both marks (e.g., ETV7). This approach did not identify any gene promoters containing two nucleosomes marked by H3K4me3 and also containing a single H3K27me3-marked nucleosome. On average, only 6 gene promoters per replicate contained single nucleosome sized fragments marked by both H3K4me3 and H3K27me3 (Figure 10G left panel; Bivalent Single-Nucleosome). A majority of promoters marked with both H3K4me3 and H3K27me3 (>60% across all replicates) appeared to be associated with a single nucleosome marked by H3K4me3, or alternatively, with two nucleosomes marked by H3K27me3 (Figure 10G left panel; Heterogeneous). 76 77 Figure 10. Gene promoters show reproducible fragment size distributions associated with distinct RNA expression and DNA methylation profiles. Fragment length distributions at promoters (defined as +1000 bp and -300 bp from the transcriptional start site (TSS) are shown as a circle with a black line indicating the TSS) at a two-nucleosome dominated H3K4me3 promoter (HN1L) (A), at a one-nucleosome dominated H3K27me3 promoter (APOH) (B), at a heterogeneous promoter (FOXC1) (C), and at a bivalent promoter (ETV7) (D). Correlation between two nucleosomes in weighted fragment size distributions (W2) of two replicates of H3K27me3 ndChIP-seq profiles from human CD34+ CB cells (Pearson correlation of 0.85). (F) Correlation between W2 of two replicates of H3K4me3 ndChIP-seq profiles from human CD34+ CB cells (Pearson correlation of 0.83). (G) Number of each class of promoters identified in human CD34+ CB cells (left panel). *Indicates an insignificant number of promoters (<10 promoters). RNA expression distribution at each identified promoter class (middle panel). Fractional DNA methylation distribution across classes promoter (right panel). (H) Genome browser shot of ETV7 and FOXC1 promoters. H3K27me3 (red), H3K4me3 (green), H3K27me3-H3K4me3 (blue), H3K4me3-H3K27me3 (cyan) and sequential IPs Input DNA (black). 78 Figure S3. Nucleosome density profiles are non-overlapping and correlate between replicates (Relates to Figure 3). (A) Overlap of one nucleosome dominated H3K27me3 marked promoters (green) and two nucleosomes dominated H3K27me3 marked promoters (red) between human cord blood CD34+ replicates. (B) Correlation between two nucleosomes fragment size weighted distributions (W2) of H3K27me3 ndChIP-seq profiles from human cord blood CD34+ cells in replicates 1 and 2 (Pearson correlation of 0.72). (C) Correlation between W2 of H3K4me3 ndChIP-seq profiles from human cord blood CD34+ cells in replicates 1 and 2 (Pearson correlation of 0.63). (D) Heatmap showing Pearson correlation of H3K4me3 (right) and H3K27me3 (left) signal between ndChIP-seq and sequential IPs calculated in the promoters identified as bivalent and heterogeneous (TSS -1Kb/300bp). A B C D 30244412492616 51031644012532615 1931 79 Figure 11. Nucleosome density profiles are non-overlapping and correlate between replicates. A) Overlap of one nucleosome dominated H3K27me3 marked promoters (green) and two nucleosomes dominated H3K27me3 marked promoters (red) between human CD34+ CB replicates. (B) Correlation between two nucleosomes fragment size weighted distributions (W2) of H3K27me3 ndChIP-seq profiles from human CD34+ CB cells in replicates 1 and 2 (Pearson correlation of 0.72). (C) Correlation between W2 of H3K4me3 ndChIP-seq profiles from human CD34+ CB cells in replicates 1 and 2 (Pearson correlation of 0.63). (D) Heatmap showing Pearson correlation of H3K4me3 (right) and H3K27me3 (left) signal between ndChIP- seq and sequential IPs calculated in the promoters identified as bivalent and heterogeneous (TSS -1Kb/300bp). 80 Gene enrichment (McLean et al., 2010) analysis of the genes associated with heterogeneous promoter profiles revealed they were generally associated with genes related to development and stem cell differentiation, like bivalent promoters in human ESCs (Figure 12A and B). Among those associated with promoters identified as bivalent (containing two nucleosomes marked simultaneously by both H3K4me3 and H3K27me3, on average 270 across all replicates, gene enrichment corrected p-value of <3.9 x 10-17) were genes for key TFs implicated in hematopoietic cell differentiation and self-renewal. These included GATA3 (Frelin et al., 2013), CEBPA (Rebel et al., 2002), EBF2 (Kieslinger et al., 2010), and the ETS TF gene, ETV7 (Geltink and Grosveld, 2013). As a test of our classifications, we performed an experiment in which MNase-treated CD34+ CB cell chromatin was subjected first to ChIP with H3K4me3 followed by H3K27me3, and the inverse (Section 2.5). As predicted by our ndChIP-seq classifications, the results showed an enrichment at the ETV7 (bivalent) promoter, but not at the (heterogeneous) FOXC1 promoter (Figure 10H). To assess the ndChIP-seq classification of heterogeneous and bivalent promoters genome wide, we compared H3K4me3 and H3K27me3 signals at identified bivalent and heterogeneous promoters with that of sequential IPs signals (H3K4me3 followed by H3K27me3 and vice versa) (Figure 11D). The results from H3K4me3 and H3K27me3 IPs correlated (Pearson; 0.775 and 0.323, respectively) with sequential IPs at ndChIP-seq classified bivalent promoters, indicating that this signal is maintained after a second IP. In contrast, low or negative correlations were observed between H3K4me3 and H3K27me3 IPs (Pearson; -0.198 and 0.026, respectively) and sequential IPs at ndChIP-seq classified heterogeneous promoters. 81 single-organism cellular processbiological regulat ionregulat ion of cellular processm ult icellular organism al processsingle-m ult icellular organism processdevelopm ental processcell com m unicat ionm ult icellular organism al developm entsingle organism signalinganatom ical st ructure developm entsystem developm entcellular com ponent organizat ion or biogenesiscellular com ponent organizat ionsingle-organism developm ental processsignal t ransduct ioncellular developm ental processcell different iat ionorgan developm entanatom ical st ructure m orphogenesisregulat ion of m ult icellular organism al processnervous system developm entregulat ion of signal t ransduct ionregulat ion of localizat ioncell developm entt issue developm entcellular com ponent m ovem entneurogenesiscell adhesionbiological adhesionorgan m orphogenesisgenerat ion of neuronsregulat ion of m ult icellular organism al developm entlocom ot ionneuron different iat ioncell project ion organizat ionem bryo developm entcardiovascular system developm entcell m orphogenesisneuron developm entcell m ot ilitycellular com ponent m orphogenesiscell m igrat ionepithelium developm entsensory organ developm entext racellular m at rix organizat ionext racellular st ructure organizat ioncell m orphogenesis involved in different iat ionneuron project ion developm entregulat ion of cellular com ponent m ovem entregulat ion of anatom ical st ructure m orphogenesiscell project ion m orphogenesiscell part m orphogenesiswound healingvasculature developm entregulat ion of locom ot ionneuron project ion m orphogenesiscell-cell adhesionregulat ion of cell m igrat ionrenal system developm entregulat ion of cell m ot ilityurogenital system developm entheart developm entblood vessel developm entblood circulat ionem bryonic m orphogenesiscirculatory system processorganelle localizat ionposit ive regulat ion of cell m igrat ionposit ive regulat ion of cell m ot ilityposit ive regulat ion of cellular com ponent m ovem entblood vessel m orphogenesisposit ive regulat ion of locom ot ionkidney developm entem bryonic organ developm entstem cell different iat ionglycosam inoglycan m etabolic processam inoglycan m etabolic processconnect ive t issue developm entear developm entregulat ion of Wnt receptor signaling pathwayregulat ion of wound healingrenal system processstem cell developm entregulat ion of coagulat ionregulat ion of exocytosisodontogenesisregulat ion of canonical Wnt receptor signaling pathwaym esenchym e developm entinner ear developm entregulat ion of blood coagulat ionregulat ion of organ growthnegat ive regulat ion of Wnt receptor signaling pathwayglom erular filt rat ionendocardial cushion developm entsynapt ic vesicle exocytosisat riovent ricular valve m orphogenesisat riovent ricular valve developm entvasculogenesiscardiac epithelial to m esenchym al t ransit ionm etanephros developm enthyaluronan biosynthet ic processposit ive regulat ion of odontogenesisG-protein coupled receptor signaling pathway coupled to cGMP nucleot ide second m essengerm etanephric glom erular m esangial cell different iat ionparacrine signalingsingle-organism process0 50 100 150 200 250 300305.86293.85268.70212.78206.68173.56169.65166.99160.88159.48150.14149.74149.36144.50141.63124.35120.68113.82106.6781.5080.4176.8275.2571.6866.4064.4462.7557.8657.7957.4156.9856.4851.0250.4449.0148.2245.4041.2241.0240.8840.3940.0737.3435.5735.4935.4735.3335.2734.7733.9533.3833.2030.9330.5930.0828.3328.2227.5827.5527.2827.0726.3026.0225.9125.8925.8625.6725.5625.4125.2425.2025.1623.1622.3920.0018.1217.8617.5517.4817.4416.1716.1115.8014.8614.3613.7113.3113.2513.0112.9412.7011.9610.6310.5710.369.899.579.368.878.697.827.627.455.955.950GO Biological Process-log10(Binom ial p value)Job ID: 20151130-public-3.0.0-46iOjJDisplay nam e: CD34_K27_K4_di_m ono_rep3B A 82 Figure 12. Bivalent promoters in ESCs and heterogeneous promoters marked with H3K27me3 and H3K4me3 in human CD34+ CB cells are enriched in developmental pathways. (A) Gene enrichment analysis of heterogeneous promoters marked with H3K27me3 and H3K4me3 in human CD34+ CB cells. (B) Gene enrichment analysis of bivalent promoters in human ESCs (H1 and H9). 83 NdChIP-seq analysis thus showed that a majority of bivalently marked promoters in individual human CD34+ CB cells are associated with nucleosomes that are separately marked either by H3K4me3 or H3K27me3. However, they also suggest there is a small subset of promoters controlling genes encoding factors implicated in hematopoietic development that can be bivalent at the single cell level. 3.2.4 MNase accessibility and histone modification status is associated with RNA expression To further investigate the functional significance of combining nucleosome accessibility and histone modification profiles in CD34+ CB cells, we correlated the ndChIP-seq states identified with RNA-seq data for these (Section 2.7). Strikingly, we found that each class of promoter exhibited a unique relationship to RNA expression (Figure 10G middle panel). As expected, active states characterized by H3K4me3 on single nucleosomes correlated with active transcription of associated genes. Promoters enriched in two H3K4me3-marked nucleosomes were also found to be associated with active genes, but transcription levels from the latter were significantly lower than for genes with promoters enriched in single H3K4me3-marked nucleosomes (two-sided t-test p-value = 0.003). These results support a model where changes in MNase accessibility in combination with changes in histone modification contribute to the control of gene expression (Jin et al., 2015). As expected, promoters marked by H3K27me3, regardless of their MNase accessibility, were associated with low levels of expression of the genes they regulate. However, we noted that promoters with single H3K27me3-marked nucleosomes were associated with significantly reduced expression of their downstream genes, compared to genes whose promoters appeared 84 enriched in their content of two H3K27me3-marked nucleosomes (two-sided t-test p-value = 1.7x10-5). Genes with promoters identified as heterogeneous showed intermediate levels of expression relative to active and repressed genes, possibly reflecting the known biological heterogeneity of cell types within the CD34+ CB compartment. Promoters classified as bivalent (Figure 10G; Bivalent Two-Nucleosome) showed similar levels of expression of their downstream genes compared to promoters with two H3K27me3-marked nucleosomes (two-sided t-test, p-value = 0.231). Genes associated with bivalent promoters were also found to be significantly repressed compared to heterogeneously marked promoters (two-sided t-test, p-value = 0.005) further supporting the ndChIP-seq based classification. 3.2.5 MNase accessibility reveals a bimodal relationship between H3K27me3 and CpG methylation states Assessment of CpG methylation at promoters identified as enriched in their content of H3K4me3- and H3K27me3-marked nucleosomes showed a majority were hypomethylated (Figure 10G right panel) as expected (Roadmap Epigenomics Consortium et al., 2015). In agreement with previous reports (Neri et al., 2013b), bivalent promoters were found to be hypomethylated (median fractional methylation <0.2). However, promoters enriched in their content of a single H3K27me3-marked nucleosome were found to be hypermethylated (median fractional methylation >0.8) and promoters enriched in their content of two H3K27me3-marked nucleosomes were hypomethylated. 85 3.2.6 Promoter MNase accessibility profiles show cell type differences Application of the ndChIP-seq protocol to two well-characterized human ESC lines (H1 and H9, (Hong et al., 2011; Meissner, 2010)) again showed a modification-dependent fragment length distribution shift following IP of fragments containing H3K27me3 and/or H3K4me3, indicative of an enrichment of DNA fragments associated with one- or two-nucleosomes in these cells (Figure 13A and B). However, in contrast to human CD34+ CB cells, the ESC results did not show enrichment for two-nucleosome fragment sizes at bivalent promoters for H3K27me3 across the 6 replicates tested (Figure 14A, B, and C). Rather, the ESC promoter distributions were dominated by single-nucleosome length fragments for both H3K4me3 and H3K27me3, perhaps reflective of a more generally open chromatin structure than may be present in more restricted cell types (West et al., 2014). Analysis of the ESC nucleosome fragment size distributions also showed that the majority (>80%) of promoters containing H3K4me3- and H3K27me3-marked nucleosomes were present together on single nucleosomes and hence reflective of a bivalent state at the single cell level (Figure 13C and 14D left panel) (Shema et al., 2016). Input size distributions (Figure 13D) and the number of two-nucleosome H3K4me3-marked promoters were similar in the CD34+ CB and human ESC datasets (average of 302 and 240 promoters across all replicates, respectively), arguing against differences in MNase digestion. The human ESCs, like the CD34+ CB cells, showed levels of expression of genes with bivalent promoters that were significantly lower than those with promoters marked by a single H3K4me3 modified nucleosome (two-sided t-test, p-value < 7.3 x 10-17). On the other hand, bivalent promoters in the human ESCs identified genes with similar levels of expression as genes with promoters containing two H3K4me3-marked nucleosomes (two-sided t-test, p-value = 0.235) (Figure 13C and 14D middle panel). In contrast, in the CB CD34+ cells, this similarity was 86 Fragment Length (bp) Fragment Length (bp) B A Fluorescent Unit (FU) B 0 10000 -10 5000 -5 0 5 0.5 1.00 0 0 250 500 750 1000 C D 87 Figure 13. Distinct patterns of MNase accessibility across CD34+ CB and ESC types are not due to difference in MNase digestion. (A) Shift in H3K27me3 (left) and H3K4me3 (middle) immunoprecipitated fragment size compared to input control DNA (Input, right) libraries in H1 human ESCs. Red arrow indicates mono-nucleosome fragment size. (B) Shift in H3K27me3 (left) and H3K4me3 (middle) immunoprecipitated fragment size compared to input control DNA (Input, right) libraries in H9 human embryonic cells. (C) Number of promoters identified in each promoter class in H9 human embryonic cells (left panel). * Indicates <10 promoters. Log of RNA expression level of genes whose promoters were found in each identified state (middle panel). Fractional DNA methylation of promoters in each state (right panel). (D) Input DNA control fragment length distributions genome wide in human cord blood CD34+ and human ESCs, H1 and H9. 88 89 Figure 14. Modified promoters in ESCs show distinct patterns of MNase accessibility. H3K4me3 (green) and H3K27me3 (red) immunoprecipitated fragment distributions at co-marked promoters in human CD34+ CB cells (A), and human H1 (B) and H9 (C) ESCs. (D) Number of promoters identified in each class of promoters identified in human H1 ESCs (left panel). * Indicates <10 promoters. RNA expression distribution at each identified promoter class (middle panel). Fractional DNA methylation distribution across promoter classes (right panel). 90 associated with promoters with two H3K27me3-marked nucleosomes (two-sided t-test, p-value 0.231) (Figure 10G middle panel). These results extend to human ESCs the finding that bivalent promoters are prominent features in this cell type (Bernstein et al., 2006; Hong et al., 2011) where they are generally more accessible to MNase digestion than promoters associated with bivalency in CD34+ CB cells. 3.2.7 Human CD34+ CB cells and ESCs exhibit different MNase accessibility at CpG-island-containing promoters Examination of the relationship between H3K27me3-marked promoters and CpG methylation in human ESCs revealed a reciprocal bimodal relationship compared to that seen in human CD34+ CB cells (Figure 13C and 14D right panel). Thus, in the ESCs, promoters with two H3K27me3-marked nucleosomes were enriched in methylated CpGs whereas those with predominantly one such nucleosome were hypomethylated. In addition, we found a similar reciprocal relationship in ESCs and CD34+ CB cells in the CpG island density associated with single and double H3K27me3-marked nucleosome-containing promoters (ESCs showing more CpG islands in single H3K27me3-marked nucleosome-containing promoters and CD34+ CB cells showing more in double H3K27me3-marked nucleosome-containing promoters, Figure 15A). This finding raises the possibility that CpG islands in promoters may influence nucleosome accessibility of H3K27me3-marked promoters to MNase digestion in a cell type-specific manner. Comparison of the median methylation levels in all regions containing H3K27me3-marked nucleosomes in human ESCs and CB CD34+ cells (Figure 15B) showed these to be hypermethylated in both cell types, consistent with observations of the Roadmap Epigenomics Consortium (Roadmap Epigenomics Consortium et al., 2015). However, CpG density was 91 inversely correlated with single nucleosome marked promoters in ESCs and those containing two nucleosomes in CD34+ CB cells (Figure 15C). Assessment of the fractional methylation levels at H3K27me3 marked promoters lacking CpG islands showed hypermethylation of their DNA (average fractional methylation = 0.77 across all cell types) regardless of their nucleosome content (Figure 15D). Two nucleosomes dominated H3K27me3 marked CGI-containing promoters showed a higher median fractional methylation level compared to that of single nucleosome dominated promoters in human ESCs (average fractional methylation of 0.60 vs. 0.18, respectively); a relationship not observed in CB CD34+ cells (Figure 15D). Together these results suggest changes in chromatin structure with prevalent H3K27me3 marking of nucleosomes at CpG-rich regions in the earliest stages of human embryo development, represented by ESCs, with later acquisition of a more compacted chromatin structure at CpG-rich H3K27me3-marked promoters despite their retention of their original DNA methylation status. 3.2.8 Modeling chromatin states differences using nucleosome density data Incorporation of nucleosome density data as an independent feature in ChromHMM was then used to generate chromatin states for human ESCs and CD34+ CB cells that incorporated nucleosome accessibility (Section 2.13). Chromatin states generated for CD34+ CB cells independently confirmed these effectively segregated bivalent and heterogeneous promoters and gave classifications supported by sequential ChIP-seq data (average concordance = 0.70; Figure 15E). Incorporation of both human ESCs and CD34+ CB cells into the model identified expected and unexpected chromatin states and revealed transitions between them (Figure 15F). The expected transition of a bivalent to heterogeneous state was seen in 497 bivalent ESC promoters (State 1, Figure 15F). The inverse relationship was also observed, where 235 bivalent promoters 92 in CD34+ CB cells were identified as heterogeneous in ESCs (state 3, Figure 15F). Enrichment analysis of genes associated with state 1 promoters revealed that they were enriched in genes that regulate embryonic and anatomical structure development (q-value < 2.61x10-85) (Figure 15G). Conversely, in addition to general developmental terms, genes associated with state 3 promoters were enriched in genes implicated in the regulation of cell differentiation and lineage commitment (q-value < 8.91x10-35) (Figure 15G). Also identified were a small number of promoters (84; state 7) that maintained bivalency across human ESCs and CD34+ CB cells (no significant gene enrichment found). 93 94 Figure 15. H3K27me3-enriched promoters in human ESCs and CD34+ CB cells show reciprocal MNase accessibility profiles. (A) Fraction of H3K27me3-enriched promoters associated with one- or two-nucleosomes containing CpG islands in human ESCs and CD34+ CB cells. (B) Median fractional DNA methylation at H3K27me3-enriched regions genome wide and at one- and two- nucleosome promoters in human ESCs and CD34+ CB cells. (C) Median CpG density, defined as the ratio of CpG to the total number of base pairs in a promoter, at two or one nucleosome dominated promoters. (D) Median fractional DNA methylation at one (red) and two (black) H3K27me3 marked nucleosome CpG islands and non-CpG island promoters in human CD34+ CB cells (circle) and ESCs (H1, triangle; H9, square). (E) ChromHMM nucleosome density states at promoters marked with either H3K4me3 or H3K27me3 in human CD34+ CB cells. (F) ChromHMM nucleosome density states at promoters marked with either H3K4me3 and H3K27me3 in human CD34+ CB and ESCs (H1 and H9 cells). (G) Gene enrichment analysis of ChromHMM-derived state 1 (green) and state 3 (orange). 95 3.3 Discussion Complex interactions between TFs, epigenetic modifiers, and chromatin structure regulate gene expression during development and later enable the sustained production of short-lived specialized cell types throughout adult life. The importance of epigenomic regulation in these processes is underscored by accumulating evidence of their perturbation in the generation of malignant populations (Abdel-Wahab et al., 2011; Karnezis et al., 2016; Shen et al., 2011; Sigauke et al., 2006; Wang et al., 2011c; Wiegand et al., 2010). Information about the roles of different chromatin modifications in controlling patterns of transcription that specify cell types is also rapidly accruing. Until recently, our understanding of these combinatorial states was limited to assessments of DNA and histone modifications (Bernstein et al., 2006; Roadmap Epigenomics Consortium et al., 2015). However, these fail to distinguish between heterogeneity at the level of cellular versus individual nucleosomes. Current single molecule technologies have attempted to address this issue, but still lack the comprehensiveness required to discriminate differences genome wide (Shema et al., 2016). The methodology described here overcomes this barrier. Importantly it enables nucleosome density profiles of promoters appearing to be associated with histones bivalently marked with H3K27me3 and H3K4me3 modifications to be definitively shown to be co-associated or not with the same promoters in individual cells. Interestingly, in CD34+ CB cells <40% of bivalent promoters were bivalent at the single-cell level, whereas the corresponding value in human ESCs was >80%. These observations provide important evidence of variability in the cell type- and promoter-specific sources of heterogeneity contributing to the profile of bivalent histones detected from ChIP-seq data obtained from bulk cell analyses. 96 This study also provides insights as to the impact of nucleosome density and chromatin structure on gene expression and DNA methylation. Classification of promoters based on the density and modification of their associated nucleosomes into four major groups (active, repressed, heterogeneous, and bivalent) revealed corresponding unique cell type-specific gene expression and DNA methylation profiles. We observed a negative correlation of promoter DNA methylation with H3K27me3 at CGIs. These findings suggest the existence of protective protein complexes that prevent DNA methylation and assist in H3K27me3 deposition on nucleosomes at CpG island-containing promoters (Boulard et al., 2015; Jermann et al., 2014; Wu et al., 2013). An example of such a protein is FBXl10, which prevents DNA methylation at CpG-dense PRC2-bound regions (Boulard et al., 2015). Finally, our analysis has revealed a more compact chromatin structure at CpG islands marked by H3K27me3 in human CD34+ CB cells as compared to ESCs. In contrast, DNA methylation at these promoters is conserved, suggesting the dynamic changes in histone modification and chromatin structure that accompany development stand in contrast to a much greater stability of DNA methylation (Herlofsen et al., 2013; Sørensen et al., 2010). The dynamic change of nucleosome density observed in H3K27me3-enriched regions between human CD34+ CB cells and ESCs may be explained by differential expression and engagement of linker histone H1 variants during development and differentiation (Terme et al., 2011). In support of this model, we find H1X, a replication independent H1 variant, to be upregulated five-fold in CD34+ CB cells compared to human ESCs (RPKM of 583 vs. 117). Given the role of H3K27me3 in regulating crucial genes in early development, and the unique content of chromatin remodelers and H3K27me3 binding proteins in human ESCs (Agger et al., 2007; Boyer et al., 2006), their DNA may be more accessible to MNase digestion than derivative tissue cell types. 97 Table 4. Correlation (Pearson) of histone modifications (H3K4me1, H3K27me3, H3K9me3, and H3K36me3) signals at MACS2 identified enriched regions between ndChIP-seq from 10,000 human cord blood cells per IP and ChIP-seq generated from 1 million human cord blood CD34+ cells per IP. 98 Chapter 4: Polycomb contraction differentially regulates terminal human hematopoietic differentiation programs 4.1 Introduction Epigenetic modifications govern local chromatin activity and support gene activation or silencing through the regulation of chromatin structure and DNA accessibility (Li et al., 2014; Stricker et al., 2016). Numerous cell differentiation processes are known to be accompanied by obligatory changes in chromatin structure mediated by proteins that specifically modify histones and thereby establish and maintain defined transcriptional regulatory states (Dunham et al., 2012; Jaenisch et al., 2010; Local et al., 2018; Pott and Lieb, 2014; Roadmap Epigenomics Consortium et al., 2015). However, the details of epigenomic programing changes that accompany lineage restriction in normal tissues, and if, and how these may impact the activation and execution of downstream terminal differentiation programs remain poorly defined. Previous studies using ChIP-seq and transposase-accessible chromatin sequencing (ATAC-seq) to identify sites of permissive histone modifications have provided insights into the dynamics of active regulatory regions in the genomes of variously defined primitive subsets of mouse and human hematopoietic cells (Buenrostro et al., 2018; Corces et al., 2016; Lara-Astiaso et al., 2014). Together, the results of these analyses have suggested a model in which certain enhancers are initially “primed” during early stages of hematopoietic cell differentiation by an acquisition of “permissive” histone 3 lysine 4 mono-methylation (H3K4me1) modification that later gain additional H3K27ac histone modifications to enable terminal differentiation programs to become activated. However, the chromatin modifications that take place within different subsets of different early stages of human hematopoiesis have remained uncharacterized. To address this gap, we then used the low input ChIP-seq protocol described in Chapter 3 99 to identify H3K4me3, H3K4me1, H3K27me3, H3K27ac, H3K36me3, and H3K9me3 sites genome-wide, plus whole genome bisulfite sequencing and RNA-seq protocols following IHEC standards (Stunnenberg et al., 2016) (Figure 16A and 17A) in a series of phenotypically defined subsets of normal human cord blood (CB) cells (Chapter 2, Table 3). To identify changes associated with different stages in the differentiation process as currently inferred from functional assays (Knapp et al., 2019; Notta et al., 2016), 8 phenotypically defined subsets that are selectively enriched in cells with different hematopoietic proliferative and differentiation properties were chosen for characterization (Figure 16A). Four of these phenotypes identify cells with progenitor activity but predominantly different lineage options and comprise the bulk of the total CD34+ CB population (Knapp et al., 2019). The other four represent different downstream mature or maturing blood cell types. All 8 populations were isolated by fluorescent activated cell sorting (FACS) from CB samples based on their differential expression of historically defined surface marker profiles (see Chapter 2, Table 3). They consisted of the CD38- subset, which contains the most primitive cells and all HSCs and 3 phenotypically distinguished subsets of the CD38+ cells, historically referred to as common myeloid progenitors (CMPs), granulocyte-macrophage progenitors (GMPs) and megakaryocyte-erythroid progenitors (MEPs). The four more differentiated cell types also present in the light-density fraction of CB cells were circulating erythroblasts (hereafter referred to here as “erythroid precursors”), monocytes, mature B cells and mature T cells. 100 F H3K27me3CACD34.CD38.GMPCMPMEPErythroidMonocyteBCellTCellCD34.CD38.GMPCMPMEPErythroidMonocyteBCellTCellCellCellCellCB CD34+ CD38-CMPGMPMEPMonocyteErythroidB-CellT-Cell0.20.40.60.81ErythroidGMPCD34.CD38.CMPMEPMonocyteBCellTCellErythroidGMPCD34.CD38.CMPMEPMonocyteBCellTCellCellCellCellCB CD34+ CD38-CMPGMPMEPMonocyteErythroidB-CellT-Cell0.60.70.80.91CD34.CD38.GMPCMPMEPErythroidMonocyteBCellTCellCD34.CD38.GMPCMPMEPErythroidMonocyteBCellTCellCellCellCellCB CD34+ CD38-CMPGMPMEPMonocyteErythroidB-CellT-Cell0.20.40.60.81BEG HErythroidCD34+CD38-CMPGMPMEPTCellMonocyteBCellErythroidCD34+CD38-CMPGMPMEPTCellMonocyteBCellCellCellCellCB CD34+ CD38-CMPGMPMEPMonocyteErythroidB CellT Cell0.750.80.850.90.951ErythroidMonocyteCD34.CD38.GMPCMPMEPBCellTCellErythroidMonocyteCD34.CD38.GMPCMPMEPBCellTCellCellCellCellCB CD34+ CD38-CMPGMPMEPMonocyteErythroidB CellT Cell0.50.60.70.80.91ErythroidCD34.CD38.CMPMEPBCellTCellGMPMonocyteErythroidCD34.CD38.CMPMEPBCellTCellGMPMonocyteCellCellCellCB CD34+ CD38-CMPGMPMEPMonocyteErythroidB CellT Cell0.850.90.951ErythroidCD34+CD38-CMPGMPMEPTCellMonocyteBCellErythroidCD34+CD38-CMPGMPMEPTCellMonocyteBCellCellCellCellCB CD34+ CD38-CMPGMPMEPMonocyteErythroidB CellT Cell0.750.80.850.90.951ErythroidCD34.CD38.CMPMEPBCellTCellGMPMonocyteErythroidCD34.CD38.CMPMEPBCellTCellGMPMonocyteCellCellCellCB CD34+ CD38-CMPGMPMEPMonocyteErythroidB CellT Cell0.850.90.951ErythroidMonocyteCD34.CD38.GMPCMPMEPBCellTCellErythroidMonocyteCD34.CD38.GMPCMPMEPBCellTCellCellCellCellCB CD34+ CD38-CMPGMPMEPMonocyteErythroidB CellT Cell0.50.60.70.80.91ErythroidGMPCD34.CD38.CMPMEPMonocyteBCellTCellErythroidGMPCD34.CD38.CMPMEPMonocyteBCellTCellCellCellCellCB CD34+ CD38-CMPGMPMEPMonocyteErythroidB-CellT-Cell0.60.70.80.91H3K4me3 H3K4me3H3K27me3CD34+;38-CD7−;34+;38+;10−;RA−;135+ (CMP)CD7−;34+;38+;10−;45RA+;135+ (GMP)CD7−;34+;38+;10−;45RA-;135- (MEP)CD45+;34-;33+;11b+;14+ (Monocyte)CD45-;GPA+ (Erythroid)CD45+;34-;33-;11b-;19+;7- (B-Cell)CD3+, 4+, 45RO-, 25-, CCR7+ (T-Cell)RNA-seqChIP-seqH3K4me3H3K4me1H3K27me3H3K27acH3K9me3H3K36me3WGBSPermissiveRepressiveCEBPa GATA1 GATA1CEBPaD0.000.250.500.751.00H1 CD34+CD38− CMP GMP MEP Monocyte Erythroid B−Cell T−Cella$Cella$nCellH1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−Cell 101 Figure 16. H3K27me3 signatures are shared across functionally distinct CD34+ hematopoietic progenitor populations A) Schematic of experimental design and colour legend for cell types analyzed. Unsupervised hierarchical clustering and heatmap of pairwise Spearman correlations for protein coding gene RPKM values (B), H3K4me3 coding gene promoter (± 2 Kb) density (C), and H3K27me3 coding gene promoter density (D) across cell types indicated by colour. Genome browser view of H3K27me3 (left panel) and H3K4me3 (right panel) density tracks across cell populations indicated by track colour at the CEBPa (E) and GATA1 (F) locus. Unsupervised hierarchical clustering and heatmap showing pairwise spearman correlation of H3K4me3 (G) and H3K27me3 (H) density genome-wide. 102 CD34.CD38.CMPGMPMEPMobilized_CD34_Primary_Cells_FemalePeripheral_Blood_Mononuclear_Primary_CellsTCell1TCell2CD4_Memory_Primary_CellsCD4_Naive_Primary_CellsCD8_Naive_Primary_CellsMonocyteBCellErythroidGM12878K562Adult_LiverThymusSpleenOvaryPsoas_MuscleRight_AtriumLeft_VentricleRight_VentricleAortaSigmoid_ColonSmall_IntestineEsophagusGastricLungFetal_Intestine_LargeFetal_Intestine_SmallPancreatic_IsletsPancreasBrain_Hippocampus_MiddleFetal_Brain_FemaleBrain_Germinal_MatrixNeurosphere_Cultured_Cells_Cortex_DerivedNeurosphere_Cultured_Cells_Ganglionic_Eminence_DerivedhESC_Derived_CD184+_Endoderm_Cultured_CellsH1_Derived_Neuronal_Progenitor_Cultured_CellshESC_Derived_CD56+_Ectoderm_Cultured_CellsH1_Cell_Line4starH1_BMP4_Derived_Mesendoderm_Cultured_CellsHUES64_Cell_LineH1_BMP4_Derived_Trophoblast_Cultured_CellshESC_Derived_CD56+_Mesoderm_Cultured_CellsHEPG2A549HELABreast_Myoepithelial_CellsPenis_Foreskin_Keratinocyte_Primary_Cells_skin02Penis_Foreskin_Keratinocyte_Primary_Cells_skin03HMECBreast_vHMECNHEKUniversal_Human_ReferenceHSMMNHLFH1_Derived_Mesenchymal_Stem_CellsPenis_Foreskin_Fibroblast_Primary_Cells_skin01Penis_Foreskin_Fibroblast_Primary_Cells_skin02HUVECPenis_Foreskin_Melanocyte_Primary_Cells_skin01Penis_Foreskin_Melanocyte_Primary_Cells_skin03Source SourceHirstREMC0.60.70.80.91BACD4-BV605CD235-FITCCD3-PE CD45RO-AF700CCR7-AF647CD25-PECy7CD4 T cellsCD34.CD38.CMPGMPMEPMobilized_CD34_Primary_Cells_FemalePeripheral_Blood_Mononuclear_Primary_CellsTCell1TCell2CD4_Memory_Primary_CellsCD4_Naive_Primary_CellsCD8_Naive_Primary_CellsMonocyteBCellErythroidGM12878K562Adult_LiverThymusSpleenOvaryPsoas_MuscleRight_AtriumLeft_VentricleRight_VentricleAortaSigmoid_ColonSmall_IntestineEsophagusGastricLungFetal_Intestine_LargeFetal_Intestine_SmallPancreatic_IsletsPancreasBrain_Hippocampus_MiddleFetal_Brain_FemaleBrain_Germinal_MatrixNeurosphere_Cultured_Cells_Cortex_DerivedNeurosphere_Cultured_Cells_Ganglionic_Eminence_DerivedhESC_Derived_CD184+_Endoderm_Cultured_CellsH1_Derived_Neuronal_Progenitor_Cultured_CellshESC_Derived_CD56+_Ectoderm_Cultured_CellsH1_Cell_Line4starH1_BMP4_Derived_Mesendoderm_Cultured_CellsHUES64_Cell_LineH1_BMP4_Derived_Trophoblast_Cultured_CellshESC_Derived_CD56+_Mesoderm_Cultured_CellsHEPG2A549HELABreast_Myoepithelial_CellsPenis_Foreskin_Keratinocyte_Primary_Cells_skin02Penis_Foreskin_Keratinocyte_Primary_Cells_skin03HMECBreast_vHMECNHEKUniversal_Human_ReferenceHSMMNHLFH1_Derived_Mesenchymal_Stem_CellsPenis_Foreskin_Fibroblast_Primary_Cells_skin01Penis_Foreskin_Fibroblast_Primary_Cells_skin02HUVECPenis_Foreskin_Melanocyte_Primary_Cells_skin01Penis_Foreskin_Melanocyte_Primary_Cells_skin03SourceSourceHirstREMC0.60.70.80.91CD34.CD38.CMPGMPMEPMobilized_CD34_Primary_Cells_FemalePeripheral_Blood_Mononuclear_Primary_CellsTCell1TCell2CD4_Memory_Primary_CellsCD4_Naive_Primary_CellsCD8_Naive_Primary_CellsMonocyteBCellErythroidGM12878K562Adult_LiverThymusSpleenOvaryPsoas_MuscleRight_AtriumLeft_VentricleRight_VentricleAortaSigmoid_ColonSmall_IntestineEsophagusGastricLungFetal_Intestine_LargeFetal_Intestine_SmallPancreatic_IsletsPancreasBrain_Hippocampus_MiddleFetal_Brain_FemaleBrain_Germinal_MatrixNeurosphere_Cultured_Cells_Cortex_DerivedNeurosphere_Cultured_Cells_Ganglionic_Eminence_DerivedhESC_Derived_CD184+_Endoderm_Cultured_CellsH1_Derived_Neuronal_Progenitor_Cultured_CellshESC_Derived_CD56+_Ectoderm_Cultured_CellsH1_Cell_Line4starH1_BMP4_Derived_Mesendoderm_Cultured_CellsHUES64_Cell_LineH1_BMP4_Derived_Trophoblast_Cultured_CellshESC_Derived_CD56+_Mesoderm_Cultured_CellsHEPG2A549HELABreast_Myoepithelial_CellsPenis_Foreskin_Keratinocyte_Primary_Cells_skin02Penis_Foreskin_Keratinocyte_Primary_Cells_skin03HMECBreast_vHMECNHEKUniversal_Human_ReferenceHSMMNHLFH1_Derived_Mesenchymal_Stem_CellsPenis_Foreskin_Fibroblast_Primary_Cells_skin01Penis_Foreskin_Fibroblast_Primary_Cells_skin02HUVECPenis_Foreskin_Melanocyte_Primary_Cells_skin01Penis_Foreskin_Melanocyte_Primary_Cells_skin03SourceSourceHirstREMC0.60.70.80.91Blood Cluster 103 Figure 17. Sorting strategy for hematopoietic populations profiled in this study. A) Representative examples of sorting strategies for CD34+CD38-, CMP, GMP, MEP, monocyte, erythroid precursors, B cell and T cell isolated from pool of cord blood. B) Unsupervised hierarchical clustering and heatmap of pairwise spearman correlations for protein coding gene RPKM values across blood cell types profiled in this study in the context of all cell types profiled by NIH Epigenome RoadMap Consortium (Roadmap Epigenomics Consortium et al., 2015). The cluster of blood cell types is indicated by the shaded box. 104 4.2 Results 4.2.1 Functionally distinct human hematopoietic progenitors share a common polycomb signature RNA expression profiles among the progenitor populations analyzed (CD34+CD38- cells, CMPs, GMPs and MEPs) were more highly correlated with each other (Spearman R >0.92) than with any of the more mature CB cell types examined; i.e., erythroid precursors, monocytes and B- and T-cells (average Spearman R = 0.76, 0.83, 0.84 and 0.84, respectively, Figure 16B and 17B), confirming relationships seen in previously published datasets (Roadmap Epigenomics Consortium et al., 2015). Expression signatures derived for each progenitor subset were also in agreement with their respective published features (Figure 18A and B). For example, pathway and gene enrichment analysis of uniquely upregulated transcripts in GMPs showed these were enriched in terms related to leukocyte differentiation, inflammation response and regulation of immune response (Benjamini q-value <10e-20, Figure 18C) and included the CD135 cell surface marker. Transcripts uniquely up-regulated in MEPs were enriched in terms related to myeloid lineage differentiation (Benjamini q-value <10e-3) (Fig. 18D). Genes up-regulated in CD34+CD38- cells and CMPs as compared to the more differentiated cells were enriched in terms related to hematopoiesis regulation and differentiation (Benjamini q-value <10e-3) (Figure 18E and F). 105 GYPA ITGAM PTPRCCD38 CD7 FLT3CD14 CD19 CD3405010015002040600100200300020406005010002040608005010015020025002040600500100015002000RPKMCD34.CD38.CMPGMPMEPMonocyteErythroidBCellTCell1blackRPKMTAL1MPO-10123CD34+CD38- CMPGMPMEPMonocyteErythroidB-CellT-Celllog10(RPKM) MarkH3K27me3H3K36me3H3K4me3Log10(RPKM)CMPGMPMEPMonocyteErythroidCD34+CD38-320-1-10123CD34+CD38- CMPGMPMEPMonocyteErythroidB-CellT-Celllog10(RPKM) MarkH3K27me3H3K36me3H3K4me3B CellT CellAG Log10(q-value)CDEFMPLTFRCVWFTAL1EPORGATA1GATA3HOXB5ID2NR4A2IL3RAELANECSF2RACEBPAMPOCD33SPI1CellGroupCellCB CD34+ CD38-CMPGMPMEPGroupErythroid/MegakaryocyteMonocyte/GranulocyteCD34+CD38- Compartment-1-0.500.51HB0.000.250.500.751.00H1 CD34+CD38− CMP GMP MEP Monocyte Erythroid B−Cell T−Cella$Cella$nCellH1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−Cell0.000.250.500.751.00H1 CD34+CD38− CMP GMP MEP Monocyte Erythroid B−Cell T−Cella$Cella$nCellH1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−Cell 106 Figure 18. Each progenitor populations possesses a unique expression profile. A) Expression (RPKM) of cell type specific cell surface marker genes across cell types as indicated by the colour legend on the bottom right. B) Expression of previously identified progenitor population specific genes across CD34+CD38-, CMP, GMP and MEP. Gene ontology analysis of GMP (C), MEP (D), CD34+CD38- (E), and CMP (F) up-regulated genes identified by DEFine (FDR > 0.01). G) Genome browser view of H3K4me3 density in progenitor populations at the TAL1 and MPO locus across CD34+CD38-, CMP, GMP and MEP. H) Expression of genes marked with H3K4me3, H3K27me3 or H3K36me3 across each cell type. 107 Examination of the chromatin state of the same subsets showed H3K4me3 and H3K36me3 densities correlated with expected transcript levels and cell type-specific signatures (Figure 16C, E-G, 18G and H). H3K4me3 densities at promoters of genes differentially expressed between different progenitor subsets, or between them and the later cell types, also showed expected relationships (Figure 19A). In contrast, H3K27me3 occupancy was nearly identical across all of the four phenotypically defined progenitor subsets (Figure 16D and H; promoter Spearman R >0.97), and cell-type specific signatures were apparent only in the mature cell types (Figure 16E and F). Amongst the more mature populations, the lymphoid cells showed a significantly higher correlation of H3K27me3 density at promoters with progenitors as compared to the corresponding data for the erythroid precursors and the monocytes (Figure 16D). This was most readily evident for the erythroid precursors in which RNA expression, and H3K4me3 and H3K27me3 signatures at promoters showed the lowest correlation (average Spearman R < 0.56) with the results for the four progenitor populations (including MEPs) (Figure 16D). There was also no significant difference in H3K27me3 density between promoters of differentially expressed genes in a comparison of GMPs and MEPs (2-sided t-test, p-value > 0.05, Figure 19B). In contrast, genes whose expression appeared down-regulated in monocytes and erythroid precursors compared to GMPs and MEPs, respectively, showed a significant gain of H3K27me3 density at the promoters of these down-regulated genes (2-sided t-test, p-value <2.2 x 10-16). 108 H3K4em3 tag densityH3K27em3 tag densityBCD34CD38CMPGMPMEPMonocyteErythriodBcellTcell10 25 50 75 100Promoter (<=1kb)Promoter (1-2kb)5' UTR3' UTR1st ExonOther Exon1st IntronOther IntronDownstream (<=3kb)Distal IntergenicPercentage100CD34CD38CMPGMPMEPMonocyteErythriodBcellTcell10 25 50 75 100Promoter (<=1kb)Promoter (1-2kb)5' UTR3' UTR1st ExonOther Exon1st IntronOther IntronDownstream (<=3kb)Distal Intergenic0.000.250.500.751.00H1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−Cella$Cella$nCellH1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−CellPercentage75502501 0755025E 1 CD34CD382 CMP3 GMP4 MEP5 Monocyte6 Erythriod7 Bcell8 Tcell15040302010TSS1020304050Binding sites (%) (5'->3')Feature0-1kb1-3kb3-5kb5-10kb10-100kb>100kbDistribution of transcription factor-binding loci relative to TSSD FPercentage of enriched regions1 CD34CD382 CMP3 GMP4 MEP5 Monocyte6 Erythriod7 Bcell8 Tcell15040302010TSS1020304050Binding sites (%) (5'->3')Feature0-1kb1-3kb3-5kb5-10kb10-100kb>100kbDistribution of transcription factor-binding loci relative to TSSMEPGMPCMPPercentage of enriched regionsTSS204040200.000.250.500.751.00H1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−Cella$Cella$nCell H1 CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−CellTSS20404020GCD34CD38CMPGMPMEPMonocyteErythriodBcellTcell10 25 50 75 100Promoter (<=1kb)Promoter (1-2kb)5' UTR3' UTR1st ExonOther Exon1st IntronOther IntronDownstream (<=3kb)Distal IntergenicCD34CD38CMPGMPMEPMonocyteErythriodBcellTcell10 25 50 75 100Promoter (<=1kb)Promoter (1-2kb)5' UTR3' UTR1st ExonO her Exon1st IntronOther IntronDownstream (<=3kb)Distal IntergenicCD34CD38CMPGMPMEPMonocyteErythriodBcellTcell10 25 50 75 100Promoter (<=1kb)Promoter (1-2kb)5' UTR3' UTR1st ExonOther Exon1st IntronOther IntronDownstream (<=3kb)Distal IntergenicCGMP vs ErythroidGMP up-regulatedGMP vs MEPGMP up-regulatedGMP vs MEPMEP up-regulatedMEP vs MonocyteMEP up-regulatedCMPGMPMEPMonocyteErythroid CMPGMPMEPMonocyteErythroid CMPGMPMEPMonocyteErythroid CP MP MEPMonocyteErythroid02040Normalized H3K27me3 tagsCMPGMPMEPMonocyteErythroid3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4AT A + EPZ✱✱✱ns✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱ns✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱ ✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4 + EPZ✱✱✱ns✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱GMP vs ErythroidGMP up-regulatedGMP vs MEPGMP up-regulatedGMP vs MEPMEP up-regulatedMEP vs MonocyteMEP up-regulatedCMPGMPMEPMonocyteErythroidCMPGMPMEPMonocyteErythroidCMPGMPMEPMonocyteErythroidCMPGMPMEPMonocyteErythroid0100200300Normalized H3K4me3 tagsCMPGMPMEPMonocyteErythroid3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZns✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱ns✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱A0.000.250.500.751.00H1 CD34+CD38− CMP GMP MEP Monocyte Erythroid B−Cell T−Cella$Cella$nCellH1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−Cell0.000.250.500.751.00H1 CD34+CD38− CMP GMP MEP Monocyte Erythroid B−Cell T−Cella$Cella$nCellH1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−CellH3K27me3 Occupancy (MB) 109 Figure 19. H3K27me3 signal is stable at promoters across progenitor populations. H3K4me3 (A) and H3K27me3 (B) tag density ±2 Kb of transcription start sites of genes up-regulated in GMP and MEP across CMP, GMP, MEP, monocyte and erythroid precursor as indicated by the colour legend on the bottom right. C) Plot of the cumulative number of base pairs marked by H3K27me3 across cell types indicated by colour legend on the bottom right. Percentage of H3K27me3- (D) and H3K4me3-(E) enriched regions at genomic features. Percentage of H3K27me3 (F) and H3K4me3 (G) enriched regions with respect to their binned distance to the TSS of coding genes across cell types indicated by colours. Distance from TSS for each bin indicated in the bottom panel (*** p < 0.001). 110 4.2.2 Terminal differentiation is differentially associated with a genome-wide contraction of H3K27me3 density We also identified different patterns of H3K27me3 occupancy in the 4 different mature cell types examined. Both monocytes and erythroid precursors showed significantly fewer (30-52%) H3K27me3–marked histones genome-wide (Figure 19C) with a pronounced contraction of the broad H3K27me3 domains shared by all four of the progenitor populations examined (Figure 20A). The remaining H3K27me3 in the erythroid precursors and monocytes took on a more punctate structure reminiscent of that seen in pluripotent cells (Hawkins et al., 2010) (Figure 20A-E). IP fragment distributions at promoters within 2 Kb of TSSs were also significantly different in the monocyte and erythroid precursor populations in comparison to the four progenitor subsets or to either of the mature lymphoid cell types analyzed (Kolmogorov–Smirnov test, p-value < 7x10-12; Figure 20C). Examination of H3K27me3 promoter distributions in all of these cell phenotypes also revealed an increase in the proportion of H3K27me3-marked histones at promoters in the monocytes and erythroid precursors as compared to the progenitors from which they are thought to have derived in vivo (Figure 19D and F), without a measurable change in H3K4me3 promoter occupancy (Figure 19E and G). To examine the functional consequence of the global alteration in H3K27me3 occupancy we measured H3K27me3 within gene bodies and related this to transcript expression across all of the eight CB cell types profiled (Figure 20F). Consistent with the genome-wide patterns, we observed an increased number of genes with a loss in H3K27me3 density in monocytes and erythroid precursors (1418 and 1980, respectively) compared to the two types of mature lymphoid cells (942 in B cells and 873 in T cells) when compared to CD34+CD38- cells (Figure 20F). 111 CD34+CD38-MonocyteErythroidB CellT CellChr1:47,000,000-48,500,0001MBH3K27me3 DensityA BEF0.000.250.500.751.00H1 CD34+CD38− CMP GMP MEP Monocyte Erythroid B−Cell T−Cella$Cella$nCellH1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−Cell123CMPGMPMEPMonocyteErythroidB CellT CellCD34+CD38-4Kb 4KbTSSCMax Normalized Coverage0.000.250.500.751.00-2000 0 2000Max Normalized CoverageCD34CD38CMPGMPMEPMonocyteErythriodB-CellT-CellH1TSS1.000.750.500.250.-2Kb +2Kb0.000.250.500.751.00-2000 0 2000Normalized CoverageCD34CD38CMPGMPMEPMonocyteErythriodBcellTcell1H1-2Kb +2KbTSSDHCD7 EPB42CD14 CD19 CD3605010015020002040600300600900050100150200250050100150RPKMblackCD34.CD38.CMPGMPMEPMonocyteErythroidBCellTCellEPB42H3K27me3CD14CD36H3K27me3H3K27me3CD7 EPB42CD14 CD19 CD3605010015020002040600300600900050100150200250050100150RPKMblackCD34.CD38.CMPGMPMEPMonocyteErythroidBCellTCellCD7 EPB42CD14 CD19 CD3605010015020002040600300600900050100150200250050100150RPKMblackCD34.CD38.CMPGMPMEPMonocyteErythroidBCellTCellIJRPKMRPKMRPKM196 132970 20 40 60RPKMMonocyteErythroidB CellablackPercentage overlappedG0.000.250.500.751.00H1 CD34+CD38− CMP GMP MEP Monocyte Erythroid B−Cell T−Cella$Cella$nCellH1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−Cell87324299422360141818841980132233020.000.250.500.751.00H1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−Cella$Cella$nCellH1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−Cell0.000.250.500.751.00H1 CD34+CD38− CMP GMP MEP Monocyte Erythroid B−Cell T−Cella$Cella$nCellH1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−CellH3K27me3 marked regions > 100Kb0100200300400CD34+CD38-CMPGMPMEPMonocyteErythroidB-CellT-Cell# Regions with > 100Kb occupancyCD34+CD38-CMPGMPMEPMonocyteErythroidB-CellT-Cell 112 Figure 20. Myeloid differentiation is associated with genome-wide H3K27me3 contraction. A) Number of H3K27me3-marked genomic regions with widths >100 Kb across cell types indicated by the colour legend on the bottom right. B) Genome browser view of H3K27me3 density profiles over a 2 Mb window of chromosome 1 highlighting contraction of H3K27me3 in the monocyte (brown) and erythroid precursor (purple) cell types. Maximum value normalized H3K27me3 (C) and H3K4me3 (D) density at coding gene promoters (± 2 Kb of TSS). E) Heatmap of H3K27me3 densities coloured by cell type at gene promoters and flanks (TSS ± 4 Kb) for regions marked by H3K27me3 in monocytes but not erythroid precursors (panel 1), in erythroid precursor populations but not monocyte (panel 2) or in both monocyte and erythroid precursor (panel 3). Shading indicates increased density. F) Sankey diagram of number of genes retaining or losing H3K27me3 at their gene body in monocyte, erythroid precursor, B cell and T cell with respect to CD34+CD38- population. (Dark shade: loss of H3K27me3, light shade retention of H3K27me3). G) Percentage of overlap of genes that lost H3K27me3 within up-regulated gene bodies in monocyte, erythroid precursor and B cells with respect to CD34+CD38- (absolute number of overlap genes is shown on top of each bars). Genome browser view of H3K27me3 density at CD14 (H), EPB42 (I) and CD36 (J) locus. Expression (RPKM) of each gene across all profiled cell types are shown in the right panel. 113 Loss of H3K27me3 also correlated with a greater proportion of up-regulated genes in monocytes and erythroid precursors compared to mature lymphoid cells (Figure 20G). Among the genes that lost H3K27me3 and were upregulated are well-known markers of erythroid and monocyte lineage differentiation, including CD36, EPB42 and CD14 (Figure 20H-J). In addition to these differential gene-specific losses of H3K27me3 in the erythroid cells and monocytes, we noted an associated loss of many large organized chromatin K27me3 domains (LOCKs, (Wen et al., 2009)) (Figure 21A-C). Interestingly, these LOCKs were generally retained in the B and T cells (Figure 21A, D and E). Moreover, a majority of the few remaining LOCKs identified in the monocytes and erythroid precursors (102 and 48, respectively) overlapped with LOCKs present in the four progenitor populations (91 and 41, respectively, Figure 21B and C). However, even within the LOCKs retained in the monocytes and erythroid precursors, their H3K27me3 densities showed evidence of contraction, with an average width of 52 Kb compared to 268 Kb seen in the progenitor subsets, including those (GMPS and MEPs) already largely restricted to the neutrophil/monocyte and the erythroid/megakaryocyte lineages, respectively (Figure 21F-H). These data thus provide evidence of a genome-wide contraction of H3K27me3 density in both gene bodies and within LOCKs during the process by which certain myeloid-restricted CB progenitors activate their terminal differentiation programs (Figure 21I) that appears lacking in the process that results in the production of mature lymphoid cells. 114 3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱0.70.80.91.0-4000-2000 020004000CD34CD38CMPGMPMEPMonocyteB_CellT_Cell_rep1G0.70.80.91.0-4000-2000 020004000CD34CD38CMPGMPMEPErythroidB_CellT_CellB CD EIAH1CD34+CD38-CMPGMPMEPMonocyteErythroidB CellT CellHL60F111 91 11MonocyteProgenitor populations16141 7ErythroidProgenitor populations14953 22 187 3115B CellProgenitor populationsT CellProgenitor populationsMax Normalized Coverage1.00.90.80.7-4Kb +4Kb0 -4Kb +4Kb0HChr20:49,500,000-50,300,000500KbLOCKHSCLymphoidMyeloid0.000.250.500.751.00H1 CD34+CD38− CMP GMP MEP Monocyte Erythroid B−Cell T−Cella$Cella$nCellH1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−Cell0.000.250.500.751.00H1 CD34+CD38− CMP GMP MEP Monocyte Erythroid B−Cell T−Cella$Cella$nCellH1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−Cell 115 Figure 21. H3K27me3 contraction during myeloid differentiation coincides with a loss of progenitor LOCKs. A) Genome browser view of H3K27me3 density on chromosome 20 across cell types as indicated by the colour legend on the bottom right. A H3K27me3 LOCK (FDR <0.05) present in progenitor cells but absent in monocyte and erythroid precursors is indicated by the shaded box. Venn diagram of LOCKS common to progenitors (CD34+CD38-; MEP; CMP; GMP) compared to those called in monocytes (B), erythroid precursors (C), B cells (D), and T cells (E). F) Violin plot of base pairs marked by H3K27me3 within LOCKs across cell types as indicated (*** p <0.001). Maximum value normalized H3K27me3 density-enriched within LOCKs called in erythroid precursors (G) and monocytes (H). I) Cartoon illustrating contraction of H3K27me3 (red) during myeloid differentiation. 116 4.2.3 H3K27me3 LOCKs are enriched in lamina-associated domains (LADs) Use of ChromHMM to generate an 18-state model based on H3K4me1, H3K4me3, H3K27me3, H3K27ac, H3K36me3, H3K9me3 occupancy across all of the 8 different hematopoietic subsets examined (Figure 22A) showed H3K9me3 to be the most highly enriched mark within H3K27me3-defined LOCKS (Figure 22B and C). Polycomb-repressed regions with or without H3K9me3 were found to be enriched in lamina associated domains (LADs) and in intergenic regions in both the progenitor subsets and in the B and T cells but, again, not in monocytes or erythroid precursors (Figure 22A, D, E and 23A). In contrast, regions enriched in H3K9me3 alone were strongly associated with LADs in all cell types (Figure 22A and 23A). Monocytes and erythroid precursors showed reduced DNA methylation at LOCKs compared to the four progenitor subsets (2-sided t-test p-value < 5.5 x 10-8; Figure 22F and 23B). Erythroid precursors particularly, but also the monocytes showed reduced DNA methylation more broadly in chromatin states identified as polycomb-repressed and/or heterochromatin-repressed in the ChromHMM model (Figure 23C-F). These findings are consistent with a previously reported reduced global level of DNA methylation in monocytes and neutrophils relative to different progenitor populations (Farlik et al., 2016). They also corroborate a previously reported relationship between DNA methylation and H3K9me3 densities (Dong et al., 2008; Freitag and Selker, 2005; Lehnertz et al., 2003) and, in the present context, suggest that CD34+ hematopoietic progenitors share a higher order chromatin structure that is associated strongly with LADs and enriched in sites of H3K27me3 and H3K9me3. 117 3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱ ✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱0123H3K27me3H3K9me3H3K4me1H3K4me3H3K36me3H3K27acHistone modifcation occupancy (MB)CD34+CD348−CMPGMPMEPMonocyteErythroidB−CellT−CellA BFD EGenome %CpGIslandExonGeneTESTSSTSS+/-2KbLADBiavlent EnhancerPoised EnhancerPoised Genic EnhnacerStrong Active Flanking PromotersWeak flanking prmoterActive Flanking PromotersStrong Active PromotersActive PromotersBivalent promotersPolycomb repressedRepressed regionsHeterochromatinZinc FingerOtherTranscribed regions 1Transcribed regions 2Active Genic Enhnacer 1Active Genic Enhnacer 20.20.40.60.81FractionalMethylationGenome %CpGIslandExonGeneTESTSSTSS+/-2KbLADCellCell CD34+CD38-CMPGMPMEPMonocyteErythroidB CellT CellAML-2-1012Genome %CpGIslandExonGeneTESTSSTSS+/-2KbLADCellCell CD34+CD38-CMPGMPMEPMonocyteErythroidB CellT CellAML-2-1012LADTSS+/- bTSSTESGenExonCpGIsl dGeno e %Enrichment Z-ScoreG en o me %C pG I sl a ndE x onG en eT EST SST SS +/ - 2 KbL ADC el lCellCD34+CD38-CMPGMPMEPMonocyteErythroidB CellT CellAML-2-10122-2Occupancy (10MB)0.000.250.500.751.00H1 CD34+CD38− CMP GMP MEP Monocyte Erythroid B−Cell T−Cella$Cella$nCellH1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−Cell0.000.250.500.751.00H1 CD34+CD38− CMP GMP MEP Monocyte Erythroid B−Cell T−Cella$Cella$nCellH1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−CellCH3K27me3H3K9me3H3K36me3H3K27acH3K4me3H3K4me1Biavlent EnhancerPoised EnhancerPoised Genic EnhnacerStrong Active Flanking PromotersWeak flanking prmoterActive Flanking PromotersStrong Active PromotersActive PromotersBivalent promotersPolycomb repressedRepressed regionsHeterochromatinZinc FingerOtherTranscribed regions 1Transcribed regions 2Active Genic Enhnacer 1Active Genic Enhnacer 200.20.40.60.81MAXMINGenome %CpGIslandExonGeneTESTSSTSS+/-2KbLADBiavlent EnhancerPoised Enhanc rPoise Genic EnhnacerStrong Active Flanking PromotersWeak flanking prmoterActive Flanking PromotersStrong Active PromotersActive PromotersBivalent promotersPolycomb repressedRepressed regionsHeterochromatinZinc FingerOtherTranscribed regions 1Transcribed regions 2Active Genic Enhnacer 1Active Genic Enhnacer 20.20.40.60.81Genome %CpGIslandExonGeneTESTSSTSS+/-2KbLADBiavlent EnhancerPoised EnhancerPoised Genic EnhnacerStrong Active Flanking PromotersWeak flanking prmoterActive Flanking PromotersStrong Active PromotersActive PromotersBivalent promotersPolycomb repressedRepressed regionsHeterochromatinZinc FingerOtherTranscribed regions 1Transcribed regions 2Active Genic Enhnacer 1Active Genic Enhnacer 200.20.40.60.81 MAXMINMAXErythroidMEPGenome %CpGIslandExonGeneTESTSSTSS+/-2KbLADBiavlent EnhancerPoised EnhancerPoised Genic EnhnacerStrong Active Flanking PromotersWeak flanking prmoterActive Flanking PromotersStrong Active PromotersActive PromotersBivalent promotersPolycomb repressedRepressed regionsHeterochromatinZinc FingerOtherTranscribed regions 1Transcribed regions 2Active Genic Enhnacer 1Active Genic Enhnacer 20.20.40.60.81Genome %CpGIslandExonGeneTESTSSTSS+/-2KbLADBiavlent EnhancerPoised EnhancerPoised Genic EnhnacerStrong Active Flanking PromotersWeak flanking prmoterActive Flanking PromotersStrong Active PromotersActive PromotersBivalent promotersPolycomb repressedRepressed regionsHeterochromatinZinc FingerOtherTranscribed regions 1Transcribed regions 2Active Genic Enhnacer 1Active Genic Enhnacer 200.20.40.60.81G nome %CpGIslandExonGeneTESTSSTSS+/-2KbLADBiavlent EnhancerPoised EnhancerPoised Genic EnhnacerStrong Active Flanking PromotersWeak flanking prmoterActive Flanking PromotersStrong Active PromotersActive PromotersBivalent promotersPolycomb repressedRepressed regionsHeterochromatinZinc FingerOtherTranscribed regions 1Transcribed regions 2Active Genic Enhnacer 1Active Genic Enhnacer 20.20.40.60.81H3K27me3H3K9me3H3K36me3H3K27acH3K4me3H3K4me1Biavlent EnhancerPoised EnhancerPoised Genic EnhnacerStrong Active Flanking PromotersWeak flanking prmoterActive Flanking PromotersStrong Active PromotersActive PromotersBivalent promotersPolycomb repressedRepressed regionsHeterochromatinZinc FingerOtherTranscribed regions 1Transcribed regions 2Active Genic Enhnacer 1Active Genic Enhnacer 200.20.40.60.813K4me1H3K4me33K27ac3K36me33K9me33K27me3DS+/-2KbSESenexonGIslandenome %L DSS+/-2KbTSSTESenexonpGIslandenome %LADTSS+/-2KbTSSTESeneExonpGIslandenome %10101010B Cell 118 Figure 22. H3K27me3 LOCKs lost during myeloid differentiation are associated with LADs in CD34+ progenitor cells. A) Emission probabilities by histone mark for 18 states ChromHMM chromatin state model across all cell types (upper panel). Enrichment of chromatin states within genomic features for MEPs (red), erythroid precursors (purple) and B cells (blue). B) Cumulative number of bases marked by indicated histone modification within progenitor population’s LOCKS as indicated by colour legend on bottom right. C) Genome browser view of H3K27me3 and H3K9me3 density on chromosome 20 across cell types as indicated by colour legend on bottom left. A H3K27me3 LOCK (FDR < 0.05) present in progenitor cells but lost in monocyte and erythroid precursor is indicated by the shaded box. Heatmap of z-score enrichment of ChromHMM identified polycomb repressed (D) and dually repressed (E) chromatin states at genomic features. F) Violin plot of fractional CpG methylation levels within progenitor’s LOCKs (*** p < 0.001). 119 BAGenome %CpGIslandExonGeneTESTSSTSS+/-2KbLADBiavlent EnhancerPoised EnhancerPoised Genic EnhnacerStrong Active Flanking PromotersWeak flanking prmoterActive Flanking PromotersStrong Active PromotersActive PromotersBivalent promotersPolycomb repressedRepressed regionsHeterochromatinZinc FingerOtherTranscribed regions 1Transcribed regions 2Active Genic Enhnacer 1Active Genic Enhnacer 20.20.40.60.81Genome %CpGIslandExonGeneTESTSSTSS+/-2KbLADBiavlent EnhancerPoised EnhancerPoised Genic EnhnacerStrong Active Flanking PromotersWeak flanking prmoterActive Flanking PromotersStrong Active PromotersActive PromotersBivalent promotersPolycomb repressedRepressed regionsHeterochromatinZinc FingerOtherTranscribed regions 1Transcribed regions 2Active Genic Enhnacer 1Active Genic Enhnacer 20.20.40.60.81Genome %CpGIslandExonGeneTESTSSTSS+/-2KbLADBiavlent EnhancerPoised EnhancerPoised Genic EnhnacerStrong Active Flanking PromotersWeak flanking prmoterActive Flanking PromotersStrong Active PromotersActive PromotersBivalent promotersPolycomb repressedRepressed regionsHeterochromatinZinc FingerOtherTranscribed regions 1Transcribed regions 2Active Genic Enhnacer 1Active Genic Enhnacer 20.20.40.60.81Genome %CpGIslandExonGeneTESTSSTSS+/-2KbLADBiavlent EnhancerPoised EnhancerPoised Genic EnhnacerStrong Active Flanking PromotersWeak flanking prmoterActive Flanking PromotersStrong Active PromotersActive PromotersBivalent promotersPolycomb repressedRepressed regionsHeterochromatinZinc FingerOtherTranscribed regions 1Transcribed regions 2Active Genic Enhnacer 1Active Genic Enhnacer 20.20.40.60.81Genome %CpGIslandExonGeneTESTSSTSS+/-2KbLADBiavlent EnhancerPoised EnhancerPoised Genic EnhnacerStrong Active Flanking PromotersWeak flanking prmoterActive Flanking PromotersStrong Active PromotersActive PromotersBivalent promotersPolycomb repressedRepressed regionsHeterochromatinZinc FingerOtherTranscribed regions 1Transcribed regions 2Active Genic Enhnacer 1Active Genic Enhnacer 20.20.40.60.81Genome %CpGIslandExonGeneTESTSSTSS+/-2KbLADBiavlent EnhancerPoised EnhancerPoised Genic EnhnacerStrong Active Flanking PromotersWeak flanking prmoterActive Flanking PromotersStrong Active PromotersActive PromotersBivalent promotersPolycomb repressedRepressed regionsHeterochromatinZinc FingerOtherTranscribed regions 1Transcribed regions 2Active Genic Enhnacer 1Active Genic Enhnacer 20.20.40.60.81Genome %CpGIslandExonGeneTESTSSTSS+/-2KbLADBiavlent EnhancerPoised EnhancerPoised Genic EnhnacerStrong Active Flanking PromotersWeak flanking prmoterActive Flanking PromotersStrong Active PromotersActive PromotersBivalent promotersPolycomb repressedRepressed regionsHeterochromatinZinc FingerOtherTranscribed regions 1Transcribed regions 2Active Genic Enhnacer 1Active Genic Enhnacer 20.20.40.60.81Genome %CpGIslandExonGeneTESTSSTSS+/-2KbLADBiavlent EnhancerPoised EnhancerPoised Genic EnhnacerStrong Active Flanking PromotersWeak flanking prmoterActive Flanking PromotersStrong Active PromotersActive PromotersBivalent promotersPolycomb repressedRepressed regionsHeterochromatinZinc FingerOtherTranscribed regions 1Transcribed regions 2Active Genic Enhnacer 1Active Genic Enhnacer 20.20.40.60.81Genome %CpGIslandExonGeneTESTSSTSS+/-2KbLADBiavlent EnhancerPoised EnhancerPoised Genic EnhnacerStrong Active Flanking PromotersWeak flanking prmoterActive Flanking PromotersStrong Active PromotersActive PromotersBivalent promotersPolycomb repressedRepressed regionsHeterochromatinZinc FingerOtherTranscribed regions 1Transcribed regions 2Active Genic Enhnacer 1Active Genic Enhnacer 20.20.40.60.81Genome %CpGIslandExonGeneTESTSSTSS+/-2KbLADBiavlent EnhancerPoised EnhancerPoised Genic EnhnacerStrong Active Flanking PromotersWeak flanking prmoterActive Flanking PromotersStrong Active PromotersActive PromotersBivalent promotersPolycomb repressedRepressed regionsHeterochromatinZinc FingerOtherTranscribed regions 1Transcribed regions 2Active Genic Enhnacer 1Active Genic Enhnacer 20.20.40.60.81LADTSS +/-2KbTSSTESGeneExonCpGIslandGenome %1 0LADTSS +/-2KbTSSTESGeneExonCpGIslandGenome %LADTSS +/-2KbTSSTESGeneExonCpGIslandGenome %LADTSS +/-2KbTSSTESGeneExonCpGIslandGenome %LADTSS +/-2KbTSSTESGeneExonCpGIslandGenome %1 0 1 0 1 0 1 0olycomb RepressedterochromatinDually RepressedLOCKSC D E F1000Chr20:49,650,000-50,050,000200Kb0.000.250.500.751.00H1 CD34+CD38− CMP GMP MEP Monocyte Erythroid B−Cell T−Cella$Cella$nCellH1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−Cell1.00.750.50.250FractionalMethylation 120 Figure 23. ChromHMM identified polycomb repressed regions are enriched in LADs in progenitor cells. A) Enrichment of ChromHMM chromatin states within genomic features for T cells (navy), monocyte (brown), GMP (yellow), CMP (orange) and CD34+CD38- cells (green). B) Genome browser view of Fractional DNA methylation on chromosome 20 across cell types as indicated by colour legend. A H3K27me3 LOCK (FDR < 0.05) present in progenitor cells but lost in monocyte and erythroid precursor is indicated by the shaded box. Fractional CpG methylation signal within ChromHMM defined H3K9me3 (C), H3K27me3 (D) and H3K9me3/H3K27me3 (E) marked regions and genome-wide (F) across cell types as indicated by colour legend. 121 4.2.4 H3K27me3 contraction is associated with reduced BMI1 expression The punctate pattern of H3K27me3 observed in monocyte and erythroid precursors has been extensively described in both human and murine ESCs but not in other cell types (Hawkins et al., 2010). It was therefore of interest to determine whether there were any similarities in the expression of PcG proteins unique to these cell types. BMI1, a component of PRC1 was the only PcG gene that demonstrated consistently lower transcript expression in all 3 cell types compared to other blood cell types profiled (Figure 24A). Reduced BMI1 expression in myeloid and erythroid cells compared to other blood cell types also observed in mouse hematopoietic cells (Figure 24B) (Lara-Astiaso et al., 2014). Transcriptional repression of BMI1 has previously been associated with loss of both H3K27me3 and H3K9me3 and a concomitant reduction in heterochromatin and its inhibition in vivo compromises hematopoietic progenitor self-renewal activity (Hyland et al., 2011; Park et al., 2003). Interestingly we also found that, similar to H3K27me3 (Fig 20A), there is a directional loss of H3K9me3 occupancy specifically in monocyte and erythroid precursors (Fig 24C). Together, these observations suggest that the contraction of H3K27me3 observed in monocyte and erythroid precursors that is shared with ESCs, is associated with a reduced expression of BMI1, a PRC1 complex member previously implicated in the H3K27me3 maintenance. 122 PCGF1PRC1PCGF2PRC1PCGF3PRC1BMI1PRC1CBX6PRC1CBX7PRC1CBX8PRC1ZNF84PRC1ZNF17PRC1KDM7ADemethylaseUTYDemethylaseRING1PRC1CBX2PRC1CBX4PRC1SUZ12PRC2RBBP4PRC2RBBP7PRC2KDM6ADemethylaseKDM6BDemethylaseEZH1PRC2EZH2PRC2AEBP2PRC2JARID2PRC2EEDPRC20240501001500306090024605010015020002040020406001234501020300102005101520020400123010203005101520250102030051015200102001020010203005101520250100200300010203001234RPKMH1CD34.CD38.CMPGMPMEPMonocyteErythroidBCellTCell1blackB ITPMRPKMAB C04080120LT-HSCHSCMPPCLPCMPGMPMFGranulocyteMono BCD4CD8 NKMEPEryAEryBTPMblackBlackred0.000.250.500.751.00H1 CD34+CD38− CMP GMP MEP Monocyte Erythroid B−Cell T−Cella$Cella$nCellH1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−Cell0.000.250.500.751.00H1 CD34+CD38− CMP GMP MEP Monocyte Erythroid B−Cell T−Cella$Cella$nCellH1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−Cell0.000.250.500.751.00H1 CD34+CD38− CMP GMP MEP Monocyte Erythroid B−Cell T−Cella$Cella$nCellH1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−Cell0100200300400500CD34+CD38-CMPGMPMEPMonocyteErythroidB-CellT-Cell# Regions with > 100Kb occupancyCD34+CD38-CMPGMPMEPMonocyteErythroidB-CellT-CellH3K9me3 marked regions > 100Kb 123 Figure 24. BMI1 is expressed at a lower level in erythroid precursors, monocytes and ESCs compare to other blood cells. A) Expression (RPKM) of PRC1 and PRC2 components and H3K27me3 demethylase across cell types as indicated by the colour legend on the bottom right. B) Expression (TPM) of Bmi1 in mouse hematopoietic cells. C) Plot of the cumulative number of base pairs marked by H3K9me3 across cell types indicated by colour legend on the bottom right. 124 4.2.5 Hematopoietic progenitor subsets share lineage-specific enhancers marked by H3K27ac Examination of a catalogue of the active (H3K27ac and H3K4me1) and primed (H3K4me1) enhancers identified in each CB population revealed a relatively consistent presence of primed enhancers in all four progenitor populations (Spearman, average R >0.8; Figure 25A) with active enhancers showing a higher degree of progenitor specificity (Spearman, average R >0.52; Figure 25B and 26C). The progenitor populations also showed a consistently higher number of total enhancers, as measured by the sum of both the active and primed enhancer states, by comparison to the four more mature cell types (Figure 26A). In contrast, the number of active enhancers was higher in 3 of the 4 more mature cell types (2-sided t-test, p-value =0.026), the exception being the erythroid precursors that had the lowest number of active enhancers as compared with all other cell types (Figure 26B). When the gain or loss of H3K27ac and H3K4me1 from the most primitive CD34+CD38- cells to each differentiated cell type according to predicted trajectories was tracked, a directional loss of total enhancers during progenitor differentiation was noted, with the erythroid precursors showing the greatest overall enhancer loss (Figure 25C). However, this directional loss was largely restricted to primed enhancers with active enhancers staying either the same or increasing with progressive differentiation (Figure 25C). The majority of primed enhancers (>90%) and surprisingly active enhancers (>72%) found in mature cells were already evident in the progenitor populations (Figure 25D, E and 26D). 125 GMPCD34.CD38.CMPMEPProErythMonocyteB_CellT_Cell_1GMPCD34.CD38.CMPMEPProErythMonocyteB_CellT_Cell_1CellCellCellCB CD34+ CD38-CMPGMPMEPMonocyteErythroidB CellT Cell00.20.40.60.81ProErythCMP.xMEP.xCD34.CD38..xGMP.xMonocyte.xB_CellT_CellProErythCMP.xMEP.xCD34.CD38..xGMP.xMonocyte.xB_CellT_CellCellCellCellCB CD34+ CD38-CMPGMPMEPMonocyteErythroidB CellT Cell00.20.40.60.81A B CGMPCD34.CD38.CMPMEPProErythMonocyteB_CellT_Cell_1GMPCD34.CD38.CMPMEPProErythMonocyteB_CellT_Cell_1CellCellCellCB CD34+ CD38-CMPGMPMEPMonocyteErythroidB CellT Cell00.20.40.60.81D0 25 50 75100TcellBcellErythroidMonocyteProportion of Active EnhancersFirst Appear in CD34+CD38-First Appear in Commited Myeloid ProgenitorsSpecificblack0 25 50 75100TcellBcellErythroidMonocyteProportion of EnhancersFirst Appear in CD34+CD38-First Appear in Commited Myeloid ProgenitorsSpecificblackMonocyteErythroidB CellT CellPercentageMonocyteErythroidB CelT Cel10075502500255075100TcellBcellErythroidMonocyteProportion of Active EnhancersPoised in CD34+CD38-Poised in Commited Myeloid ProgenitorsSpecificblackFir t appear in CD34+CD38-Fir t appear in myeloid progenitorscificE0.000.250.500.751.00H1 CD34+CD38− CMP GMP MEP Monocyte Erythroid B−Cell T−Cella$Cella$nCellH1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−CellSPI1EPB41EPB41SPI1CD34+CD38- Corresponding Progenitors0-2240-224ErythroidMonocyteLog10(RPKM)F GH EPB41 SPI1CD34.CD38.CMPGMPMEPMonocyteErythroidCD34.CD38.CMPGMPMEPMonocyteErythroid01002003000100020003000bps x 10e6blackCD34.CD38.CMPGMPMEPMonocyteErythroid300201000EPB41 SPI1CD34.CD38.CMPGMPMEPMonocyteErythroidCD34.CD38.CMPGMPMEPMonocyteErythroid01002003000100020003000bps x 10e6blackCD34.CD38.CMPGMPMEPMonocyteErythroid3020100Loss| GainLoss | GainRPKMNumber of enhancersNumber of active enhancers3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days01234DMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱0.000.250.500.751.00H1 CD34+CD38− CMP GMP MEP Monocyte Erythroid B−Cell T−Cella$Cella$nCellH1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−Cell- 126 Figure 25. A majority of active enhancers in mature myeloid cells are marked by H3K27ac in CD34+ progenitor cell populations. Unsupervised hierarchical clustering and heatmap of pairwise Spearman correlations of H3K4me1 (A) and H3K27ac (B) densities across cell types as indicated by the colour legend on the bottom left. C) Number of total enhancers (top panel) and active enhancers (bottom panel) that are gained or lost in comparison to CD34+CD38- cells in each cell type. Percentage of enhancers (D) and active enhancers (E) present in differentiated cells that are first detected in CD34+CD38- cells, myeloid progenitor cells, or uniquely in monocytes, erythroid precursors, or lymphoid cells as indicated. F) Expression of associated genes with active enhancers in monocytes and erythroid precursors that first appear in CD34+CD38- cells or their corresponding progenitor populations. G) Genome browser view of H3K27ac density at SPI1 (top panel) and EPB41 (bottom panel) locus. H) Expression (RPKM) of SPI1 (left panel) and EPB4 (right panel). *** p <0.001. 127 T Cell0255075100TcellBcellErythroidMonocyteProportionblackGMP & MEPCMP & MEPCMP & GMPCMP & GMP & MEPMEP UniqueGMP UniqueCMP UniqueSpecific0255075100TcellBcellErythroidMonocyteProportionblackGMP & MEPCMP & MEPCMP & GMPCMP & GMP & MEPMEP UniqueGMP UniqueCMP UniqueSpecificMonocyteErythroidB CellT CellMonocyteErythroidB CellMEPGMPCMP0255075100MEPGMPCMPProportionblackGMP & MEPCMP & MEPCMP & GMPCMP & GMP & MEPMEP UniqueGMP UniqueCMP UniqueSpecific10075502500255075100TcellBcellErythroidMonocyteProportionblackGMP & MEPCMP & MEPCMP & GMPCMP & GMP & MEPMEP UniqueGMP UniqueCMP UniqueSpecific0500010000Number of Active Enhancers CD34+ CD38-CMPGMPMEPMonocyteErythroidB CellT CellNumber of Active Enhancers0200004000060000Number of EnhancersCD34+ CD38-CMPGMPMEPMonocyteErythroidB CellT CellNumber of EnhancersA BC D EPercentage0255075100TcellBcellErythroidMonocyteProportion of Active EnhancersPoised in CD34+CD38-Poised in Commited Myeloid ProgenitorsSpecificblackFir t poised in CD34+CD38-Fir t poised in myeloid progenitorsecific0255075100TcellBcellErythroidMonocyteProportion of Active EnhancersPoised in CD34+CD38-Poised in Commited Myeloid ProgenitorsSpecificblackMonocyteErythroidB CellT Cell107550250PercentageF0.000.250.500.751.00H1 CD34+CD38− CMP GMP MEP Monocyte Erythroid B−Cell T−Cella$Cella$nCellH1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−Cell 128 Figure 26. Total number of enhancers decreased in erythroid precursor, monocyte and differentiated lymphoid cells. A) Total number of enhancers (active and primed enhancers) across cell types as indicated by the colour legend. B) Total number of active enhancers across cell types as indicated by the colour legend. C) Percentage of active enhancers that are shared between or unique to each progenitor populations as indicated by the color legend. D) Percentage of active enhancers that are present in progenitor populations or de novo (specific) in differentiated cells as indicated by the colour legend. E) Percentage of active enhancers that are poised in progenitor populations or de novo in differentiated cells as indicated by the colour legends. F) Percentage of active enhancers present in differentiated cells that are first poised in CD34+CD38- cells, progenitor cells, or appear de novo in erythroid precursors, monocytes or lymphoid cells as indicated by the colour legend. 129 A significant fraction (>80%) of the active enhancers in the differentiated cell types were also already primed across the progenitor subsets, as noted earlier for similar subsets of mouse hematopoietic cells (Lara-Astiaso et al., 2014) (Figure 26E and F). Interestingly, genes associated with enhancers found to be active in monocytes and erythroid precursors, but already active in progenitors also produced significantly higher levels of transcripts in the mature cells compared to their inferred parental progenitors (pairwise Wilcoxon signed-rank test Bonferroni corrected p-value <2x10-5, Figure 25F). They were also enriched in terms related to specific hematopoietic differentiation programs (Benjamini corrected p-value <10e-30, Figure 27A and B). For example, SPI1 and EPB41 were associated with an active enhancer in progenitor and in lineage-restricted cells, with their expression increasing significantly during the full course of the differentiation process (Figure 25G and H). To further understand how active enhancer states differ between progenitors and terminally differentiated cells, self-organizing maps (SOM) were generated using ranked normalized H3K27ac signals across all hematopoietic enhancers. MEP and GMP enhancer clusters thus identified (Section 2.14; Figure 27C) showed progression towards, and enrichment in, enhancers belonging to erythroid precursor and monocyte enhancer clusters, respectively compared to the CD34+CD38- population (Figure 27D). Notably, the differentiated enhancer clusters enriched in the MEPs and GMPs suggested a pre-existing H3K27ac signature at cell type-specific enhancers in progenitors. 130 Dblood coagulationcellular response to DNA damage stimuluschromatin remodelingchromosome organizationerythrocyte developmenterythrocyte differentiationerythrocyte homeostasisheme metabolic processhemopoiesishemostasismegakaryocyte differentiationmyeloid cell developmentmyeloid cell homeostasisporphyrin-containing compound biosynthetic processporphyrin-containing compound metabolic processregulation of erythrocyte differentiationtetrapyrrole biosynthetic processtransition metal ion homeostasisubiquitin-dependent protein catabolic process-40-20 0 20First Appear in CD34+CD38-First Appear in Committed Myeloid Progenitors150100 50 0 50CCD34+CD38-MEPMonocyteCMP B-CellT-CellErythroidGMPCD34+CD38-MEPMonocyteCMP B-CellT-CellErythroidGMPMonocyteErythroid0 30 60 900 10 20 30erythrocyte differentiationerythrocyte homeostasisheme biosynthetic processheme metabolic processmyeloid cell homeostasisporphyrin-containing compound biosynthetic processporphyrin-containing compound metabolic processprotoporphyrinogen IX metabolic processtetrapyrrole biosynthetic processtetrapyrrole metabolic processactivation of immune responseendocytosisimmune responseimmune response-activating signal transductioninnate immune responsepositive regulation of immune responsepositive regulation of immune system processregulation of cytokine productionregulation of defense responseregulation of innate immune responseErythroidMonocyteE3020100 MonocyteErythroid0 30 60 900 10 20 30erythrocyte differ ntiationerythr c te homeostasiheme biosynthetheme metab lic procesmyeloid cell homeostasiporphyrin-containi g compound biosynthetporphyrin-c ntaining compoundprotoporphyrinogen IX metaboltetrapyrrole biosynthettetrapyrrole metabolic processactivation of immune responseendocyto ismmune responseimmune response-activating signal transductioninnatepositive regulation of immune respon epositive regulation of immune system pro essregulation of cytoki e p oductionregulation of defensregulation of innate immune responseErythroidMonocyte9060300EnrichedDepletedNo ChangeHighLowFBA-Log10(q-value)-Log10(q-value)0255075100TcellBcellErythroidMonocyteProportion of Active EnhancersPoised in CD34+CD38-Poised in Commited Myeloid ProgenitorsSpecificblackFir t appear in CD34+CD38-Fir t appear in myeloid progenitors 131 Figure 27. Lineage-specific enhancers are marked by H3K27ac in hematopoietic progenitor subsets. GREAT pathway enrichment analysis of active enhancers in erythroid precursor (A) and monocytes (B) that are first apparent in CD34+CD38- or other progenitor subsets. C) SOM plot of rank normalized H3K27ac signal contained within the hematopoietic enhancer catalogue across cell types profiled in this study. D) SOM plot of rank normalized H3K27ac signal within hematopoietic enhancers with respect to CD34+CD38- for each cell type. Gene enrichment analysis of super enhancers in erythroid precursor (E) and monocyte (F). 132 Interestingly, CMPs showed enrichment in both lymphoid and monocyte enhancer clusters consistent with recent data suggesting that CMPs are quite heterogeneous and could thus appear to share some similarity in enhancer landscape with lymphoid cells as compared to the MEPs and GMPs in which cells with lymphoid potential appear largely eliminated (Knapp et al., 2019; Notta et al., 2016). To further compare the enhancer landscape across progenitor and mature cell type, ROSE RANK algorithm (Hnisz et al., 2013) was used to identify the high amplitude active enhancers and clusters of H3K27ac (super-enhancers) in each of the 8 cell types analyzed. The results revealed a >2-fold increase in the number of super-enhancers in the mature cell types compared to the numbers of these in the progenitor populations (Figure 28A). Surprisingly, a majority (>70%) of the active enhancers seen only in the mature cells were located within the boundaries of super enhancers, including the human counterparts of enhancers previously identified in differentiating mouse hematopoietic cells (Figure 28B and C). Pathway analysis of genes associated with monocyte and erythroid precursor super-enhancers were enriched in terms related to leukocyte activity and erythroid differentiation, respectively (Figure 27E and F). Taken together these data suggest a model in which a majority of active enhancers are initially marked in progenitor populations and then further reinforced to form super-enhancers during the final processes of terminal differentiation. 133 A0.000.250.500.751.00MonocyteErythroidB CellT CellNumber of Super EnhancersOut side Super-EnhancerWithin Super-Enhancer0.000.250.500.751.00MonocyteErythroidB CellT CellNumber of Super EnhancersOut side Super-EnhancerWithin Super-EnhancerGYPAS100A7/8 BCL11bEBF1BCMonocyteErythroidB CellT CellNumber of super enhancers0.000.250.500.751.00Proportion of enhancersOutside of super e hancersithin of super enhancers20001500100050005101520CD34+ CD38-CMPGMPMEPMonocyteErythroidB CellT CellNumber of Super Enhancers CD34+ CD38-CMPGMPMEPMonocyteErythroidB CellT Cell0.000.250.500.751.00H1 CD34+CD38− CMP GMP MEP Monocyte Erythroid B−Cell T−Cella$Cella$nCellH1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−Cell0.000.250.500.751.00H1 CD34+CD38− CMP GMP MEP Monocyte Erythroid B−Cell T−Cella$Cella$nCellH1CD34+CD38−CMPGMPMEPMonocyteErythroidB−CellT−Cell 134 Figure 28. The majority of active enhancers first appearing in differentiated cell populations contribute to the establishment of super enhancers. A) Total number of super enhancers across cell types as indicated by color legend on the bottom right. B) Fraction of cell type-specific active enhancers that reside within super enhancers. C) Genome browser view of H3K27ac density across cell types as indicated by color at lineage-specific super enhancers indicated by the shaded box. 135 4.2.6 EZH2 inhibition differentially alters the production of different hematopoietic lineages consistent with their acquired epigenomic features. The broad H3K27me3 domains shared by CB progenitors that are no longer present in monocyte and erythroid precursors but persist in mature lymphoid cells (Figure 22I) suggested that these differences may reflect different functional roles in the processes that allow these different types of mature cells to be produced. To test this possibility, 2 inhibitors that target complexes implicated in the maintenance of repressive chromatin (i.e., EPZ-6438, an inhibitor of EZH2, a major component of PRC2, and GSK-J4, a H3K27me3 demethylase inhibitor) were used to alter the ability of HL-60 cells to activate a granulopoietic differentiation program (Breitman et al., 1980). After 3 days of treatment with EPZ-6438, the HL60 cells showed the same growth-arrest expected as a consequence of their differentiation as is obtained with ATRA (Figure 29A) and a subsequent expected loss of viability (Figure 29B). In contrast, in the presence of GSK-J4, growth arrest and subsequent loss of viability were indistinguishable from control cells exposed only to DMSO. Confirmation of induced granulopoietic differentiation in the presence of ATRA or EPZ-6438 was obtained by FACS detection of the appearance of increased numbers of CD11b+ cells after 48 hours of exposure to these treatments compared to controls (p-value <0.001, Figure 29C). Thus, the inferred EPZ-6438-mediated H3K27me3 removal in HL60 cells appears to promote the same phenotypic alterations as ATRA in HL60 cells, consistent with the importance of H3K27me3 removal to enable the terminal granulopoietic differentiation to be activated (Fiskus et al., 2009). 136 D EA BC3 days 6 days 9 days01234Number of Cells in MillionsDMSO ControlATRA GSK-J4EPZATRA + GSK-J4ATRA + EPZ✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱ns✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱✱3 days 6 days 9 days020406080100Percentage ViabilityDMSO ControlATRAGSK-J4EPZATRA + GSK-J4ATRA + EPZ3 days 6 days 9 days020406080100Percentage ViabilityDMSO ControlATRAGSK-J4EPZATRA + GSK-J4ATRA + EPZ48 hr 72 hr020406080100%CD11bDMSO ControlATRA GSK-J4EPZCD19+ B cellsCD14+ monocyteCD15+ neutrophil010000200003000040000Number of Cells✱nsns48 hr 72 hr020406080100%CD11bDMSO ControlATRA GSK-J4EPZFCD19+ B cellsCD14+ monocyteCD15+ neutrophil010000200003000040000Number of Cellsnsns nsDMSO EPZ050100150200Number of ClonesNegativeNMBNM+BUndefined 137 Figure 29. EZH2 inhibition arrests HL60 cell growth and lymphoid production in vitro. A) Total number of HL60 cells in culture after 3, 6 and 9 days of treatment with ATRA, GSK-J4 (K27me demethylase inhibitor) or EPZ (EZH2 inhibitor). B) Percentage of viable cells in (A). C) Percentage of CD11b+ cells assessed by FACS after 48 and 72 hours of treatment with ATRA, GSK-J4 or EPZ. Total cell number of CD19+ B cells, CD14+ monocytes, and CD15+ neutrophils after 3 weeks of culturing CB-derived CD45highCD34highCD38midCD71-CD10- (P-NML) cells with lymphoid and neutrophil/monocyte lineage-stimulatory cytokines in the presence or absence of EPZ (D) or GSK-J4 (E). F) Bar plot showing the number of clones with different contents. M = CD14+ monocytes, N= CD15+ neutrophils, B = B cells, negative = <10 CD45+ events, undefined = no detectable mature cells. (* p < 0.05, ** p < 0.01 and *** p < 0.001). 138 A final experiment was then undertaken to determine whether these inhibitors would differentially affect the ability of human CB progenitors with dual granulopoietic and B-lymphopoietic differentiation potential to activate one or both of their primed terminal differentiation programs when maintained in vitro under conditions optimized to support the production of both of these lineages. Accordingly, a series of cultures were initiated with the CD38midCD71-CD10- subset of CD34+ CB cells previously shown to be highly and selectively enriched in cells restricted to B and neutrophil/macrophage production (Knapp et al., 2019). Assessment of the mature cell types generated after 3 weeks in these cultures showed a selectively decreased content of CD19+ B lineage cells when EPZ-6438 was presence (2-sided t-test p-value <0.014), with no effects of either inhibitor on the output of cells expressing phenotypic markers of monocyte/macrophages or neutrophils (Figure 29D and E). To further investigate this effect of EZH2 inhibition on B cell differentiation, we analyzed the outputs of this same CD34+CD38midCD71-CD10- CB cell population under similarly optimized conditions in microcultures initiated with single input cells. Analyses of their clonal outputs 3 weeks later showed that the EPZ-6438 treatment had reduced the proportion of clones that contained CD19+ B-lineage cells (27% vs 42% of 192 cells tested in the EPZ vs the control cultures) without any evidence of any toxicity to the input cells (21-day frequency of clone formation = 118/192 and 128/192, respectively) (Figure 29F). Together, these results suggest that continued maintenance of H3K27me3 by PRC2 is required for progenitors with lymphoid potential to differentiate into mature lymphoid cells. 139 4.3 Discussion Epigenetic mechanisms play a central role in hematopoietic cell fate decisions (Álvarez-Errico et al., 2015; Bock et al., 2012; Bröske et al., 2009; Deaton et al., 2011; Hodges et al., 2011; Stewart et al., 2015). Of particular interest here is the repressive H3K27me3 chromatin modification that has been implicated in the regulation of hematopoiesis in vivo and in vitro (Kamminga et al., 2006; Petruk et al., 2017; Xie et al., 2014; Xu et al., 2015; Yu et al., 2017) and how manipulation of H3K27me3 may impact the steps progenitors undergo to produce mature myeloid and lymphoid cell types (Mochizuki-Kashio et al., 2011; Petruk et al., 2017). Both overexpression and inactivation of PRC2 components, responsible for the methylation of H3K27 have been reported in hematopoietic malignancies, further reinforcing the importance of H3K27me3 in regulating normal hematopoietic differentiation events (Ernst et al., 2010; Ganji et al., 2012). Here, analysis of the epigenomic states of immunophenotypically defined, FACS-purified subpopulations of CD34+ and mature hematopoietic cells isolated from normal human CB samples revealed a broad and stable repressive H3K27me3 landscape across all progenitor subsets analyzed; i.e., those defined phenotypically as CD34+CD38-, CMPs, GMPs and MEPs. This contrasts markedly with the changing cell-type specific landscape defined by the active histone modifications in these same populations. A striking and lineage-selective genome-wide H3K27me3 signature evident in 2 mature myeloid cell types (monocytes and erythroid precursors) and not seen in 2 mature lymphoid cell types (B cells and T cells) was also identified. Intriguingly, the H3K27me3 signature common to the monocytes and erythroid precursors displayed a punctuated structure reminiscent of the H3K27me3 signature associated with ESCs (Zhu et al., 2013). 140 In the classical Waddington (Waddington, 2006) view, the chromatin of very primitive cells is envisaged to exist in a privileged state allowing for the dynamic remodeling required for lineage restriction and later activation of a differentiation program. One early finding in support of such privileged chromatin in ESCs is their bivalent state (Bernstein et al., 2006; Hawkins et al., 2010; Mikkelsen et al., 2007; Zhu et al., 2013) in which the nucleosomes in the promoters of developmentally important genes are marked by both permissive and repressive histone modifications (H3K4me3 and H3K27me3, respectively) that subsequently resolve to an homogenously active or repressed chromatin state as differentiation occurs (Bernstein et al., 2006; Zhu et al., 2013). However, subsequent epigenomic studies across a broad range of primary tissues have revealed that bivalent promoters are not unique to ESC chromatin and can be found in many cell types including fully functional, terminally differentiated cell types (Lorzadeh et al., 2016; Roadmap Epigenomics Consortium et al., 2015). Another prevailing model of H3K27me3 occupancy during ESC differentiation has posited that, upon differentiation, H3K27me3 genomic occupancy spreads outwards from discrete focal regions to occupy large genomic regions (Hawkins et al., 2010), in a fashion that correlates with their differentiation state. Here, an opposite picture is revealed; i.e. H3K27me3 profiles revert back to a punctuated structure in cells that have differentiated into certain types of mature blood cells. This unexpected observation suggests that at some point in the differentiation of these cells there is an overall loss of repressive chromatin rather than a further compaction as might have been predicted from studies of ESC differentiation (Hawkins et al., 2010). In addition, H3K27me3 contraction in erythroid precursors, monocyte and ESC cells was found to correlate with a marked reduction in expression of the PRC1 complex member BMI1. Why B and T cells should retain a broader H3K27me3 landscape also raises interesting questions. One possibility is that 141 this could be related to their known ability to activate a “stem-like” state upon antigenic stimulation and then execute a large expansion of their progeny (Goldrath et al., 2006). The H3K27me3 contraction evident in erythroid precursors and monocytes includes a specific loss of H3K27me3 marked LOCKs that in the progenitors and mature lymphoid cells show co-occupancy with another suppressive histone modification, H3K9me3. The lineage-specific restructuring of H3K27me3 specifically during myeloid differentiation highlights the importance of higher order chromatin structure in the differentiation and establishment of cellular identity in the hematopoietic system. Of equivalent interest is the finding that H3K27me3- and H3K27me3/K9me3-enriched regions that are common to progenitor and lymphoid cells and that are lost in different mature myeloid cell types are strongly enriched in LADs. This finding reinforces the concept that myeloid cells can be distinguished from B and T cells as well as myeloid progenitors based on their lamin distribution and nuclear rigidity (Swift et al., 2013). Taken together these observations provide molecular support of the observation that manipulation of lamin expression can specifically modulate myeloid, but not lymphoid, cell differentiation (Swift et al., 2013). In contrast to H3K27me3, we found the active enhancer mark H3K27ac is highly variable across all progenitor populations analyzed. In addition, the majority of active enhancers identified in the mature cells were first detected in their historically defined progenitor populations despite recent evidence of their considerable heterogeneity in many functional and molecular properties (Knapp et al., 2019). Key lineage-specific regulators were also identified among the genes associated with active enhancers in the progenitor populations. During differentiation, these active enhancers increased in width and amplitude with an associated increase in expression of the genes they regulate. Taken together, these results suggest that priming of key regulatory regions during 142 hematopoietic differentiation occurs in the context of traditionally described active enhancers whose activity is reinforced and increases as lineage restriction is finalized and terminal differentiation is activated. This would thus refine the currently proposed primed (H3K4me1) to an active enhancer model (Lara-Astiaso et al., 2014) and suggest that additional features beyond the presence of H3K27ac, for example formation of lineage specific phase separated condensates (Sabari et al., 2018), may be required for full enhancer activity. The appearance of lineage-specific active enhancers in myeloid progenitor cells prior to the acquisition by their mature progeny of a global H3K27me3 contraction is particularly intriguing. This finding suggests that active enhancers already present in progenitor cells are constrained by H3K27me3 and fully activate their target genes only after large scale chromatin structure reorganization. Thus, the broad H3K27me3 domains observed in both progenitors and mature lymphoid cells may act as a general block for the generation of multiple mature myeloid cell types whose differentiation is normally completed within 3-5 cell divisions. These results reveal a previously unknown relationship of H3K27 modification in myeloid differentiation and further reinforce the importance of chromatin structure in the regulation of normal hematopoietic differentiation processes. 143 Chapter 5: Analysis of the function of neomorphic IDH mutations in human AML 5.1 Introduction There are three IDH proteins encoded in the human genome, IDH1, 2 and 3. IDH1 is located in the cytoplasm and in peroxisomes and takes part in lipid metabolism. In juxtaposition, IDH2 localizes in the mitochondria and is part of the tricarboxylic acid cycle (TCA) cycle. Both IDH1 and 2 require NADP+ as a cofactor. In contrast IDH3, which is located in the mitochondria, uses NAD+ as a cofactor. Neomorphic mutations to IDH and IDH2 are recurrent in AML (Dang et al., 2010; Ley et al., 2013) and occur at arginine 132 (R132) in IDH1, and arginine 140 (R140) and arginine 172 (R172) in IDH2 (Losman and Kaelin, 2013). Neomorphic IDH mutations are typically heterozygous and disrupt IDH substrate binding and ability to convert isocitrate to α-KG (Dang et al., 2009; Losman and Kaelin, 2013) (Figure 30A, The Cancer Genome Atlas (TCGA)). These mutations thus drive the accumulation of the metabolite D-2- hydroxyglutarate (D2HG) (Dang et al., 2009), an inhibitor of the 2-oxoglutarate-dependent dioxygenases (2- OGDD) family of enzymes that include a class of epigenetic modifiers that are responsible for both histone and DNA demethylation (Xu et al., 2011). Examples of such enzymes are the TET and JmjC histone demethylase families (Figueroa et al., 2010; Kats et al., 2014; Lu et al., 2012). Inhibition of these modifiers through D2HG accumulation potentially drives pathogenic chromatin states. Interestingly, with the exception of low grade glioma and AML, IDH mutations are rare in other malignancies (Dang et al., 2010) suggesting potentially unique roles for the epigenetic modifiers sensitive to D2HG in brain and blood cell transformation. 144 IDH mutations are associated with a generally favourable survival in gliomas, whereas clinical outcomes of IDH mutations in AML appear mixed with differences in clinical outcomes associated with IDH1 and IDH2 mutations (Abbas et al., 2010; Hartmann et al., 2010; Paschka et al., 2010). However, the level of D2HG measured in the urine samples of patients suffering from AML harbouring IDH1 or IDH2 mutations is found to be similar (DiNardo et al., 2013). Mutant IDH proteins have been reported to increase the level of DNA methylation, H3K9me3 and H3K27me3 genome wide (Figueroa et al., 2010; Lu et al., 2012). As in glioma, mutant IDH AML is associated with a CGIs methylator phenotype (CIMP) (Figueroa et al., 2010). However, a recent report indicated that there are two distinct CIMP signatures in AML, a mutant IDH-dependent one and an independent one (Kelly et al., 2017) with the IDH independent CIMP being associated with a more favorable outcome. To gain a better understanding of the role of IDH gain of function mutations in AML, the low input ChIP-seq methodology described in Chapter 3 was applied to three AML samples without IDH mutations and six AML samples with IDH mutations. The transcriptional and epigenetic profiles obtained were then compared between these two groups as well as with parallel data obtained on CD34+CD38- and bulk CD34+ cells isolated from normal adult BM samples (Figure 30B). IDH mutations included IDH1R132H, IDH1R132C, IDH2R172K. The ChIP-seq data included genome wide profiles for H3K4me3, H3K4me1, H3K27me3, H3K27ac, H3K36me3, and H3K9me3. In addition, matching RNA-seq and WGBS data were generated for all samples. A standardized analytical pipeline was applied to quantify and analyze the resulting data (Section 2.15). 145 696470433 4293561230200400600800Number of Overlapped GenesIDH mutant up-regulatedIDH mutant down-regulatedIDH wild type down-regulatedIDH wild type up-regulated03006009001200Number of GenesACCD34+; CD38- normal bone marrowCD34+ normal bone marrowAML IDH blast x3AML IDH mutant blast x6NGS assaysRNA expressionHistone modificationsH3K4me3H3K4me1H3K27me3H3K27acH3K9me3H3K36me3DNA methylationSH2B3SETD2ETV6PTPN11NRASCEBPAFLT3−ITDIDH2 R172KIDH1 R132CIDH1 R132HNPM1DNMT3A R882H230_60 95_81215_82230_55353_21231_98357_17354_80 76_51110111223456789NADEIDH mutant vs IDH wild type IDH mutant vs Normal bone marrow progenitors IDH wild type vs Normal bone marrow progenitors-4 -2 0 2 4 -10 -5 0 5 10 -10 -5 0 5 10020400501001500.02.55.07.510.012.5log2 fold change-log10 adjusted p-valueDown-regulatedInsignificantUp-regulatedIDH mutant vs IDH wild type IDH mutant vs Normal bone marrow progenitors IDH wild type vs Normal bone marrow progenitors-4 -2 0 2 4 -10 -5 0 5 10 -10 -5 0 5 10020400501001500.02.55.07.510.012.5log2 fold change-log10 adjusted p-valueDown-regulatedInsignificantUp-regulatedIDH mutant vs IDH wild type IDH mutant vs Normal bone marrow progenitors IDH wild type vs Normal bone marrow progenitors-4 -2 0 2 4 -10 -5 0 5 10 -10 -5 0 5 10020400501001500.02.55.07.510.012.5log2 fold change-log10 adjusted p-valueDown-regulatedInsignificantUp-regulatedGFPMLRARAIDH1IDH2TP53TET2RUNX1NPM1DNMT3AFLT30 10 20Mutation RateGenesBGroupDNMT3ANPM1GroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3A0.850.90.951oupMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38-Bone marrow CD34+IDH1 R132HIDH1 R132CIDH2 R172RIDH WT0.60.70.80.917851166863125 593CommonIDH mutant uniqueIDH wild type uniquecellular defense responsecytokine biosynthetic processcytokine productioncytokine-mediated signaling pathwaydefense response to other organismdendritic cell differentiationGraft-versus-host diseaseHematopoietic cell lineageimmune response-regulating signaling pathwayImmunoregulatory interactions between a Lymphoid and a non-Lymphoid cellleukocyte activation involved in immune responseleukocyte differentiationleukocyte homeostasisleukocyte mediated cytotoxicityleukocyte migrationlymphocyte activationlymphocyte mediated immunitymacrophage activationMalariamyeloid cell differentiationmyeloid leukocyte differentiationNatural killer cell mediated cytotoxicitynegative regulation of immune system processOsteoclast differentiationphagocytosisPID IL4 2PATHWAYpositive regulation of response to external stimulusregulation of innate immune responseregulation of interleukin-10 productionresponse to bacteriumT cell activationT cell costimulationTNFs bind their physiological receptors5101520CommonIDH mutant uniqueIDH wild type uniqueactin cytoskeleton organizationAdherens junctioncell junction organizationcell morphogenesis involved in differentiationcellular response to growth factor stimuluschemical synaptic transmissioncirculatory system processestablishment or maintenance of cell polarityHippo signaling pathwayregulation of epithelial cell proliferationresponse to woundingsmall GTPase mediated signal transductiontransmembrane receptor protein tyrosine kinase signaling pathwayurogenital system development33.544.55CommonIDH mutant uniqueIDH wild type uniquecellular defense responsecytokine biosynthetic processcytokine productioncytokine-mediated signaling pathwaydefense response to other organismdendritic cell differentiationGraft-versus-host diseaseHematopoietic cell lineageimmune response-regulating signaling pathwayImmunoregulatory interactions between a Lymphoid and a non-Lymphoid cellleukocyte activation involved in immune responseleukocyte differentiationleukocyte homeostasisleukocyte mediated cytotoxicityleukocyte migrationlymphocyte activationlymphocyte mediated immunitymacrophage activationMalariamyeloid cell differentiationmyeloid leukocyte differentiationNatural killer cell mediated cytotoxicitynegative regulation of immune system processOsteoclast differentiationphagocytosisPID IL4 2PATHWAYpositive regulation of response to external stimulusregulation of innate immune responseregulation of interleukin-10 productionresponse to bacteriumT cell activationT cell costimulationTNFs bind their physiological receptors5 10 15 20HCommonIDH mutant uniqueIDH wild type uniqueactin cytoskeleton organizationAdherens junctioncell junction organizationcell morphogenesis involved in differentiationcellular response to growth factor stimuluschemical synaptic transmissioncirculatory system processestablishment or maintenance of cell polarityHippo signaling pathwayregulation of epithelial cell proliferationresponse to woundingsmall GTPase mediated signal transductiontransmembrane receptor protein tyrosine kinase signaling pathwayurogenital system development3 3.5 4 4.5 5GenesMutation frequency 146 Figure 30. IDH mutants show heterogenous expression profile. A) Frequency of top ten mutated genes in 200 patients in TCGA cohort. B) Schematic of experimental design and available data set. C) Mutation status of primary AML blasts in this study. D) Unsupervised hierarchical clustering and heatmap of pairwise Spearman correlations for protein coding gene RPKM values indicated by the colour legend. E) Volcano plots of DESeq2 defined differentially expressed genes in pair wise comparison across IDH mutant, IDH wild type AML and CD34+CD38- progenitor cells. F) Upset plot showing overlap of up regulated genes from comparisons in E. G) Gene enrichment analysis of up regulated genes from comparisons in E. H) Gene enrichment analysis of down regulated genes from comparisons in E. 147 5.2 Results Primary AML sample datasets were analyzed based on their IDH status irrespective of co-occurring mutations (Figure 30C). Consistent with TCGA data (Ley et al., 2013), DNMT3A and NPM1 were the two most commonly mutated genes in all of the nine AML samples (6 and 5, respectively; Figure 30A and C). Unsupervised hierarchical clustering of protein-coding gene expression from the cohort of normal and leukemic samples segregated the normal CD34+CD38- (Spearman correlation > 0.97) and bulk CD34+ populations from the AML cells from all nine patients. However, clustering of the primary AML samples was not influenced by the status of the IDH genes nor any of the commonly mutated genes present in these samples (Figure 30D). Variability in RNA expression across mutant and wild type IDH replicates was similar (Spearman correlation range of 0.8 to 0.96 and 0.86 to 0.93, respectively). Furthermore, differential expression analysis across all IDH mutant and wild type cells revealed 5 and 12 down- and up-regulated genes respectively (adjusted p-value < 0.01). In contrast, normal CD34+CD38- and both IDH mutant and wild type AML cells showed significant transcriptional differences with transcripts of 1166 and 593 genes up-regulated in the IDH mutant and wild type cells, respectively, compared to the normal CD34+CD38- BM cells (adjusted p-value <0.01, Figure 30E) 470 of which were shared (Figure 30F). Up-regulated genes common to both IDH mutant and wild type cells were enriched in term related to leukocyte regulation and hematopoiesis (Figure 30G, adjusted p value < 0.001). Up-regulated genes unique to the AML cells with wild type IDH genes were not significantly enriched in any biological terms. Corresponding values for transcripts of down-regulated genes were 862 and 785 (adjusted p-value <0.01) and were highly enriched in terms related to cellular morphogenesis and differentiation both for those that were shared and those that were genes unique to the IDH mutant AML cells (Figure 30H, adjusted p-value <0.001), whereas the IDH wild type 148 unique down-regulated genes which were enriched in circulatory system process. These observations thus did not uncover evidence of any consistent genome-wide expression alterations specifically attributable to the presence or absence of a mutant IDH allele. 5.2.1 CpG Island methylation signatures are not IDH mutant-dependent in AML cells Gain of function IDH mutations have been associated with increases in DNA methylation via D2HG inhibition of the TET family of enzymes in low grade glioma and AML (Figueroa et al., 2010). For example, unsupervised clustering of 450K array datasets generated from genomic DNA extracted from 200 AML samples revealed a cluster that was enriched but not exclusive to patients harbouring IDH mutations (Ley et al., 2013; TCGA, Figure 31A). This contrasts to a similar analysis of low-grade glioma samples in which CpG island methylation was found to be sufficient to classify patient samples as IDH mutant or wild type (Figure 31B). As in the case of AML, however, RNA expression alone was insufficient to classify glioma samples by their IDH status (Figure 31C and D). Previous CpG methylation studies of AML cell genomes utilized reduced representation strategies enriched for CpG dense regions (e.g., 450k arrays). In contrast, datasets generated as part of the present study utilized a WGBS assessment strategy enabling the measurement of an average of 25.3 million CpGs across the sample sets. Surprisingly, a comparison of the genome wide fractional DNA methylation levels at single CpGs revealed no consistent increase in DNA methylation levels in the IDH mutant AML cells compared to the IDH wild type AML cells or the normal adult BM CD34+CD38- and CD34+ cells (Figure 32A). As previously reported, fractional DNA methylation was consistently higher at CGIs across all AML samples regardless of their IDH 149 150 Figure 31. RNA expression does not segregate mutant IDH from wild type. A) Unsupervised hierarchical clustering and heatmap of pairwise Pearson correlations for DNA methylation (TCGA AML cohort). B) Unsupervised hierarchical clustering and heatmap of pairwise Pearson correlations for DNA methylation (TCGA glioma cohort). C) Unsupervised hierarchical clustering and heatmap of pairwise Spearman correlations for protein coding gene TPM values (TCGA AML cohort). D) Unsupervised hierarchical clustering and heatmap of pairwise Spearman correlations for protein coding gene TPM values (TCGA glioma cohort). 151 status when compared to the normal BM subsets (Figure 32B). However, DNA methylation levels were not consistently higher in the mutant compared to the wild type IDH samples and fractional DNA methylation across all CGIs did not segregate the IDH mutant and wild type AML cells (Figure 32C). Partitioning the CGIs into promoter-associated and non-promoter CGIs also did not separate the IDH mutant and wild type AML cells based on fractional DNA methylation (Figure 32D and E). This result differs from the previous report by Figueroa et al. (Figueroa et al., 2010), but is in agreement with the recent report supporting an independence of CIMP from IDH mutation status in AML cells (Kelly et al., 2017). Pairwise comparisons of differentially methylated regions (Section 2.15) in AML cells with and without IDH mutations did, however identify a directional hyper-methylation associated with a mutant IDH genotype in 14 of the 18 comparisons (Figure 32F and G, FDR <0.001). Across the group of AML cells harbouring IDH mutations, 3/6 replicates consistently demonstrated a hyper-methylated phenotype compared to those harbouring a wild type IDH genotype (Figure 32H), possibly because of accompanying NPM1 frameshift mutations in the remaining mutant IDH samples, given the recently reported association of NPM1 mutations with a demethylator phenotype in AML (Kelly et al., 2018). Taken together these observations demonstrate that the genetic context in which the IDH mutation arises plays a role in defining DNA methylation levels in AML cells. 152 0.000.250.500.751.00CBCD34CD38_1CBCD34CD38_2CBCD34IDH1_1IDH1_2IDH1_3IDH1_CIDH2_1IDH2_2IDHwt_1IDHwt_2IDHwt_3Fractional MethylationCBCD34CD38_1CBCD34CD38_2CBCD34IDH1_1IDH1_2IDH1_3IDH1_CIDH2_1IDH2_2IDHwt_1IDHwt_2IDHwt_30.250.500.751.00Bone Marrow CD34+CD38- rep1Bone Marrow CD34+CD38- rep2Bone Marrow CD34+IDH1 rep1IDH1 rep2IDH1 rep3IDH1CIDH2 rep1IDH2 rep2IDHwt rep1IDHwt rep2IDHwt rep3Fractional MethylationBone Marrow CD34+CD38- rep1Bone Marrow CD34+CD38- rep2Bone Marrow CD34+IDH1 rep1IDH1 rep2IDH1 rep3IDH1CIDH2 rep1IDH2 rep2IDHwt rep1IDHwt rep2IDHwt rep30.250.500.751.00Bone Marrow CD34+CD38- repBone Marrow CD34+CD38- repBone Marrow CD34+IDH1 repIDH1 repIDH1 repIDH1IDH2 repIDH2 repIDHwt repIDHwt repIDHwt repFractional thylationBone Marrow CD34+CD38- rep1Bone Marrow CD34+CD38- rep2Bone Marrow CD34+IDH1 rep1IDH1 rep2IDH1 rep3IDH1CIDH2 rep1IDH2 rep2IDHwt rep1IDHwt rep2IDHwt rep30.000.250.500.751.00Hyper HypoPercentage non−promoter CGIout side promoters and CGIspromoterpromoter CGIABGroupDNMT3ANPM1GroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38-Bone marrow CD34+0.90.920.940.960.981GroupDNMT3ANPM1GroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWT0.880.90.920.940.960.981GroupDNMT3ANPM1GroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWT0.80.850.90.951C DEIDHwt rep3IDHwt rep2IDHwt rep1−10000 01000020000IDH1 rep1IDH1 rep2IDH1 rep3IDH1 rep1IDH1 rep2IDH1 rep3IDH1 rep1IDH1 rep2IDH1 rep3NumberHyperHypoIDHwt rep3IDHwt rep2IDHwt rep102500050000IDH1CIDH2 rep1IDH2 rep2IDH1CIDH2 rep1IDH2 rep2IDH1CIDH2 rep1IDH2 rep2NumberHyperHypoF GIDH1 rep1IDH1 rep2IDH1 rep3IDH1CIDH2 rep1IDH2 rep2025005000750010000NumberHyperHypoI0.000.250.500.751.00Hyper HypoPercentage non−promoter CGIout side promoters and CGIspromoterpromoter CGI1730 477HypoHyper0 5 10 15 200 1 2 3myeloid leukocyte differentiationimmune response−activating signal transductionFc receptor signaling pathwaynegative regulation of carbohydrate metabolic processmyeloid cell differentiationcalcium−dependent cell−cell adhesioncell adhesionbiological adhesioncell−cell adhesionhomophilic cell adhesion−log10 q−valueHyperHypoIDH1 rep1IDH1 rep2IDH1 rep3IDH1CIDH2 rep1IDH2 rep2025005000750010000NumberHyperHypoH JHypoHyper0 2 4 60 10 20 30 40ETV1SpiBRUNXELF5PU.1SF1ZEB1SlugNr5a2REST−NRSF−log10(q−Value)HyperHypoKHyper Hypo−2−101−202log10(RPKM)variableBM_CD3438_rep1BM_CD3438_rep2BM_CD3438_rep3BM_CD34IDH1_rep1IDH1_rep2IDH1_rep3IDH1CIDH2_rep1IDH2_rep2IDHwt_rep1IDHwt_rep2IDHwt_rep3GroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38-Bone marrow CD34+IDH1 R132HIDH1 R132CIDH2 R172RIDH WT0.60.70.80.91L0.00.10.2IDH1 rep1IDH1 rep2IDH1 rep3IDH1CIDH2 rep1IDH2 rep2IDHwt rep1IDHwt rep2IDHwt rep3PercentageIDH1 rep1IDH1 rep2IDH1 rep3IDH1CIDH2 rep1IDH2 rep2IDHwt rep1IDHwt rep2IDHwt rep3MGroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38-Bone marrow CD34+IDH1 R132HIDH1 R132CIDH2 R172RIDH WT0.60.70.80.91Fractional Methylation 153 Figure 32. Hypermethylation in mutant IDH occurs mainly outside CGIs. Fractional methylation genome wide (A) and at CGIs (B). Unsupervised hierarchical clustering and heatmap of pairwise Pearson correlations for DNA methylation at all CGIs (C), at CGI promoters (D) and non-promoter CGIs (E) indicated by the colour legend. Number of pairwise hyper- (black bar) and hypo-methylated (grey bar) regions across IDH mutants and wild type. (H) Number of hyper- (black bar) and hypo-methylated (grey bar) regions across each IDH mutant and all IDH wild type. I) Distribution of DMRs common in at least 3 IDH mutants across genomic features. J) Gene enrichment analysis of hyper (black bars) and hypo-methylated (grey bar) regions common in at least 3 IDH mutants. K) Transcription binding site enrichment at hyper (black bars) and hypo-methylated (grey bar) regions common in at least 3 IDH mutants. L) Expression (RPKM) of nearest genes (<50Kb from TSS) to hyper and hypo-methylated regions common in at least 3 IDH mutants. M) Percentage overlap of hyper and hypo-methylated regions common in at least 3 IDH mutants and MASC2 identified H3K27ac enriched regions in each sample. 154 Analysis of differentially methylated regions (DMRs) that were common to 3 or more AML samples harbouring IDH mutations as compared to the wild type IDH AMLs identified 1730 and hyper- and 477 hypo-methylated regions with 88% of the hyper-methylated regions located outside of CGIs and promoters (within 2kb of TSSs, Figure 32I). In contrast, only 16% of hypo-methylated regions were similarly annotated. Pathway enrichment analysis of genes associated with the hyper-methylated regions revealed terms related to myeloid differentiation (Figure 32J). On the other hand, genes associated to hypo-methylated regions were enriched in non-specific terms such as cell adhesion. In support of our previous observation in a murine IDH1R132H model (Mingay et al., 2018), mutant IDH-specific hypermethylated regions were enriched in PU.1 binding elements (Figure 32K); however no significant directional difference in the expression of genes associated with hyper-methylated DMRs was observed in the IDH mutant AML cells (Figure 32L). As a majority of hyper-methylated DMRs were located outside promoter CGIs, their overlap with enhancer regions defined by presence of H3K27ac was assessed. The percentage overlap of hyper-methylated DMRs with MACS2 identified H3K27ac marked regions varied from 8-30% across AML samples (Figure 32M) with no enrichment observed in the IDH mutant AML samples. Taken together these results suggest that one consequence of the IDH mutation in AML cells is a gain in methylation at enhancer regions defined by the presence of H3K27ac (Figure 33A and B). However, these methylation gains are not associated with de novo acquisition of H3K27ac. 155 FH3K4me3 H3K36me3 H3K27me3H3K9me3 H3K27ac H3K4me10e+001e+082e+083e+084e+080.0e+002.5e+085.0e+087.5e+081.0e+090e+001e+072e+070e+002e+084e+086e+088e+080.0e+002.5e+085.0e+087.5e+080.0e+002.5e+075.0e+077.5e+071.0e+08bpsblackBone Marrow CD34+Bone Marrow CD34+CD38-IDH mutantIDH wild typeH3K4me3 H3K36me3 H3K27me3H3K9me3 H3K27ac H3K4me10e+001e+082e+083e+084e+080.0e+002.5e+085.0e+087.5e+081.0e+090e+001e+072e+070e+002e+084e+086e+088e+080.0e+002.5e+085.0e+087.5e+080.0e+002.5e+075.0e+077.5e+071.0e+08bpsblackone arro 34+Bone Marrow CD34+CD38-I mutantIDH wild type0 10 20 30H3K27acH3K4me1H3K4me3H3K27me3H3K36me3Percentage transitionblackH3K27acH3K4me1H3K4me3H3K27me3H3K36me3Percentage TransitionEX1 X2 X3 X4 X5 X6 X7 X8 X9 X10X11X12X13X14X15X16X17X1812345678910111213141516171800.050.10.150.20.250.3H3K27acH3K4me3H3K4me1H3K27me3H3K9me3H3K36me31 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 1800.20.40.60.81H3K27acH3K4me3H3K4me1H3K27me3H3K9me3H3K36me312345678910111213141516171800.20.4.60.813 27ac3 4 e33 4 e13 27 e33 9 e33 36 e31 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 1800.20.40.60.81X1X2X3X4X5X6X7X8X9X10X11X12X13X14X15X16X17X182345678910111213141516171800.050.10.150.20.250.3C DIDH1 R132H IDH1 R132CIDH2 R172RIDH2 R172RIDHwt rep1IDHwt rep2IDHwt rep3IDH1 R132H IDH1 R132CIDH2 R172RIDH2 R172RIDHwt rep1IDHwt rep2IDHwt rep3Wild TypeMutantA BH3K27acFractional MethylationChr1:24454000-24456500CR +/-2kbbps 156 Figure 33. Genome wide occupancy of histone modifications are similar between mutant IDH and wild type. A) Genome browser view of H3K27ac density profiles over a 2.5kb window of chromosome 1 highlighting a hypermethylation at active enhancer in mutant IDH. B) Heatmap showing RPKM normalized H3K27ac density over 4kb window from the center of hyper-methylated regions common in at least 3 IDH mutant (CR = Center of Region). C) Barplot showing number of bps covered defined by MACS2 enriched regions for each histone modification profiled in this study. D) Heatmap showing enrichment value for each histone modification at each state of 18 states ChromHMM model (Top panel). Transition probability between ChromHMM defined 18 states (Bottom panel). E) Percentage of histone modification at most dynamic states. 157 5.2.2 Increase in genomic occupancy of histone modifications in AML is independent of a mutant IDH genotype As the 2-oxydependent dioxygenases family of enzymes include histone demethylases, the potential impact of a mutant IDH allele on histone modifications was also investigated. For this, the total genomic space occupied by regions enriched for each mark was compared in the AML samples with and without IDH mutations. In contrast to previous reports of global increases in H3K27me3 and H3K9me3 as measured by immunohistochemistry (Chowdhury et al., 2011), no significant increase or decrease in global occupancy of histone modifications was observed in the AML samples studied here (Figure 33C, two-sided t.test p-value > 0.05). In order to investigate specific alterations to epigenetic states associated with the presence of a mutant IDH allele, ChromHMM (Ernst and Kellis, 2012) was used in an 18-state model based on occupancy of the 6 profiled histone modifications to partition the genome (Figure 33D). Of the marks profiled, H3K27ac occupancy demonstrated the largest degree of variation across the nine AML samples regardless of their IDH status (Figure 33E). Indeed, 37% of all state transitions across any two samples involved a state that included H3K27ac. These observations imply that the active enhancers present in AML blasts are not associated directly with common genetic alterations observed in AML, including gain of function IDH mutations. 5.2.3 Mutant IDH AML cells exhibit a unique hypermethylation pattern at non-promoter regulatory regions To investigate the local remodeling of DNA methylation, the fractional methylation in all ChromHMM defined states was compared in both the normal adult BM subsets and all AML samples analyzed, regardless of their IDH status. This revealed an increase in fractional 158 methylation at active enhancers defined by the presence of both H3K4me1 and H3K27ac but lacking H3K4me3, in mutant IDH as compared to wild type AML cells (Figure 34A). The median fractional methylation was also greater than that observed for normal adult BM CD34+CD38- cells. A similar pattern was observed for genic enhancers defined by presence of H3K4me1, H3K27ac and H3K36me3. AML cells harbouring IDH1R132C or IDH2R172K mutations without a co-occurring NPM1 frameshift exhibited the largest average fractional methylation gains at primed enhancers defined by presence of H3K4me1 alone. However, this relationship was not upheld for AML samples harbouring IDH1R132H and NPM1 mutations, again highlighting the combinatorial impact of IDH and NPM1 mutations in shaping the DNA methylation landscape. In addition, AML cells generally showed higher fractional methylation levels compared to adult BM CD34+CD38- cells at bivalent enhancers, previously annotated as hypomethylated across 111 normal primary tissues and cells (Roadmap Epigenomics Consortium et al., 2015). Unsupervised clustering of fractional DNA methylation (Distance = 1- Pearson correlation) at active, poised, active genic and bivalent enhancers, segregated mutant IDH from wild type only at active enhancer regions (Figure 34B-E). Interestingly, two stable clusters among AML cells harbouring IDH mutations were observed, one contained IDH1R132H and NPM1 mutations and the other was IDH1R132C and IDH2R172K in the absence of a NPM1 mutation (Figure 34B). These results highlight the combinatorial effect of highly recurrent genetic lesions with IDH mutations. 159 GroupDNMT3ANPM1GroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38-Bone marrow CD34+0.750.80.850.90.951AGroupDNMT3ANPM1GroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38-Bone marrow CD34+0.750.80.850.90.951GroupDNMT3ANPM1GroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38-Bone marrow CD34+0.750.80.850.90.951GroupDNMT3ANPM1GroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38-Bone marrow CD34+0.70.80.91BCDEGroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38-Bone marrow CD34+IDH1 R132HID 1 R132CIDH2 R172RIDH WT0.60.70.80.910.000.250.500.751.00Active Promoter1Strong Active promoter1Active Enhancer1Poised EnhancerPoised Genic EnhancerActive Genic EnhancerFlanking Active PromoterTranscribed RegionsZing Finger GenesHeterochromatinSuppressedPRC2 SuppressedBivalent EnhancersBivalent PromotersActive Promoter2Strong Active promoter2OtherActive Enhancer2Fractional MethylationCBCD34CD38_1CBCD34CD38_2CBCD34IDH1_1IDH1_2IDH1_3IDH1_CIDH2_1IDH2_2IDHwt_1IDHwt_2IDHwt_3 160 Figure 34. IDH mutant have unique methylation phenotype at active enhancers. A) Fractional methylation at ChromHMM defined chromatin states (Figure 33D). Unsupervised hierarchical clustering and heatmap of pairwise Pearson correlations for DNA methylation at active enhancers (B), poised enhancers (C), genic active enhancers (D) and bivalent enhancers (E) indicated by the colour legend on bottom right. 161 To further investigate the relationship of H3K27ac and DNA methylation in AML cells with respect to their IDH status, we examined the relationship of DNA methylation and H3K27ac at strong active promoters defined by the presence of H3K4me3 and H3K27ac, active enhancers (H3K27ac and H3K4me1), and active enhancers within super enhancers (Hnisz et al., 2013). H3K27ac marked promoters were hypo-methylated in normal adult BM CD34+CD38- cells and in all AML samples irrespective of their IDH status (Figure 35A). At active enhancers there was an increase in fractional methylation levels in AML blasts compared to normal adult BM CD34+CD38 cells, which was increased in AML cells harbouring a mutant IDH allele (up to 46% and 84% in IDH wild type and mutant, respectively). However, an increase in fractional methylation at super-enhancers was exclusively observed in AML cells harbouring a mutant IDH allele. Hyper-methylated enhancers in IDH mutant AML cells were CpG poor compared to CpG islands (two-sided t.test p-value <2.2e-16, Figure 35B). Hyper-methylated active enhancers present in AML cells harbouring a mutant IDH allele that overlapped IDH wild type enhancers showed lower fractional DNA methylation in the IDH wild type context, (Figure 35C-F) further supporting a unique methylation signature at active enhancers in AML cells with IDH mutations. Taken together, these observations imply that IDH mutant-dependent hypermethylation predominately occurs at enhancers enriched in H3K27ac. 162 ADIDH1 R132H IDH1 R132CIDH2 R172RIDH2 R172RIDHwt rep1IDHwt rep2IDHwt rep3IDH1 R132H IDH1 R132CIDH2 R172RIDH2 R172RIDHwt rep1IDHwt rep2IDHwt rep3Chr11:61791575-61795466EIDH1 R132H IDH1 R132CIDH2 R172RIDH2 R172RIDHwt rep1IDHwt rep2IDHwt rep3IDH1 R132H Rep1 IDH1 R132H Rep2IDH2 R132CIDH2 R172RIDHwt rep1IDHwt rep2IDHwt rep3Chr20:52562770-52568739F0.00.10.20.30.4CGIIDH1_1IDH1_2IDH1_3IDH1_CIDH2_1IDH2_2CpG densityCellCGIIDH1_1IDH1_2IDH1_3IDH1_CIDH2_1IDH2_2CB0 25 50 75100IDHwt rep1IDHwt rep2IDHwt rep3Percentagenon−overlappingoverlapping0.000.250.500.751.00IDHwt_1IDHwt_2IDHwt_3Fractional Methylationregionnon−overlappingoverlapping0.000.250.500.751.00IDHwt_1IDHwt_2IDHwt_3Fractional Methylationregionnon−overlappingoverlappingNormal IDH mutant IDH wild type 0.000.250.500.751.00IDHwt_1IDHwt_2IDHwt_3Fractional Methylationregionnon−overlappingoverlapping 163 Figure 35. Enhancer methylation is prominent in mutant IDH. A) H3K27ac intensity and fractional DNA methylation at ChromHMM active promoter and enhancers, and super enhancers. B) CpG density at CGIs and active enhancers in mutant IDH leukemic blasts. C) Percentage overlap of hyper methylated enhancers in mutant IDH with active enhancers in IDH wild type (bars are indicative of standard deviation across 6 mutant IDH replicates). Genome browser view of H3K27ac density and fractional methylation profiles at Chr11:61791575-61795466 (D) and Chr20:52562770-52568739 (E). F) Fractional methylation in IDH wild type at overlapping and non-overlapping IDH mutant hyper methylated active enhancer with IDH wild type active enhancer. 164 To further classify hyper-methylated enhancers associated with IDH mutations in AML cells we searched for enriched TF binding elements. PU.1 was significantly enriched in hyper-methylated enhancers (top five TFs, q-value <10-20) (Figure 36A). PU.1 has previously been shown to bind its consensus binding sequence with the same affinity regardless of the methylation state of the embedded CpG, unlike other ETS family members, like ETS1 (Stephens and Poon, 2016). In contrast, DNA sequences predicted to bind FLI and SPDEF were significantly enriched in hypo-methylated enhancers (top five TF, q-value <10-5) (Figure 36B). Expression of genes associated with mutant IDH -associated hyper-methylated enhancers (≤50Kb away) successfully segregated normal adult BM CD34+CD38- cells from the AML cells, but gave no clear separation between the mutant and wild type IDH AML samples (Figure 36C). Genes associated with hyper methylated enhancers did not show significant down regulation in IDH mutant AML cells (Figure 36D). This observation supports a context specific relationship between CpG methylation and TF affinity and regulation, and highlights the importance of TFs that are insensitive to the presence of methylated CpGs or may show increased affinity when methylated CpGs are present. Thus, the aberrant methylation of active enhancers seen in AML cells with a mutant IDH genotype may rewire enhancer occupancy allowing for the binding of TFs with higher affinity for methylated CpGs, and at the same time prevent the binding of TF with reduced affinity for methylated CpGs. This is in particularly true in the case of ETS family of TFs which were enriched at hyper-methylated enhancers. Members of this family share similar motifs but their affinity changes upon CpG methylation (Stephens and Poon, 2016). Therefore, hyper-methylation of enhancers may promote binding of members with higher affinity to CpG methylation and in turn lead to abnormal transcription. 165 BA CGroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38−Bone marrow CD34+IDH1 R132HIDH1 R132CIDH2 R172RIDH WT−3−2−10123GroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38−Bone marrow CD34+IDH1 R132HIDH1 R132CIDH2 R172RIDH WT−3−2−10123GroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38−Bone marrow CD34+IDH1 R132HIDH1 R132CIDH2 R172RIDH WT−3−2−101230.18< 2.2e-16< 2.2e-160.673.1e-06< 2.2e-160.0187.9e-090.0082< 2.2e-168.6e-050.45IDH2_2 IDHwt_1 IDHwt_2 IDHwt_3IDH1_2 IDH1_3 IDH1_C IDH2_1CBCD34_ CBCD34CD38_1 CBCD34CD38_2 IDH1_1HighMedium Low HighMedium Low HighMedium Low HighMedium Low-101234-101234-101234methlog10(Expression + 0.1)HighMediumLowDlog10(RPKM) 166 Figure 36. IDH mutant AML methylated enhancers are enriched in PU.1 binding site. Transcription biding site enrichment at hyper (A) and hypo-methylated (B) mutant IDH enhancers. C) Unsupervised hierarchical clustering and heatmap of methylated enhancers associated genes expression Z-score indicated by the colour legend. D) Expression of genes associated to enhancers with > 75% (high), 35%-75% (medium) and <35% (low) fractional methylation. 167 Examination of H3K27ac density at ChromHMM defined active enhancers showed H3K27ac signals did not segregate AML cells based on their IDH status (Figure 37A) indicating that the ranking of H3K27ac density varies from that of DNA methylation. Common enhancers in IDH mutant AML cells showed enrichment of ETS TFs, including PU.1 in comparison to IDH wild type AML cells (Figure 37B). Genes associated with enhancers unique to IDH mutant AML cells were enriched in terms related to leukocyte proliferation and regulation, in comparison those unique to wild type AML cells that were enriched in terms related to regulation of cell adhesion and metabolism (Figure37C and D). Our results suggest that IDH mutant-dependent hypermethylation enhancers contributes to a disruption of hematopoietic proliferation and differentiation pathways. 5.2.4 The AML active enhancer landscape is poised in normal adult BM CD34+CD38- cells Examination in the normal BM and AML samples of their active and poised enhancer landscape, as defined by their content of genomic regions with H3K27ac and H3K4me1 or only H3K4me1 signals, respectively, revealed significantly more active enhancers in the leukemic cells than in the normal adult BM CD34+CD38- cells (IDH wild type: p-value = 0.039, IDH mutant: p-value = 0.009; Figure 37E). A significantly greater occupancy of just H3K4me1 was also seen in the leukemic cells (IDH wild type: p-value = 0.023, IDH mutant: p-value = 0.014; Figure 37F). However, adult bone marrow derived CD34+CD38- cells show a high correlation (spearman R value > 0.8) in H3K27ac and H3K4me1 signal at active and primed enhancers, respectively. A feature that was not observed among IDH mutant or wild type leukemic blasts where IDH status was also not discriminative (Figure 37G and H). 168 050001000015000Number of enhancersBM CD34+BM CD34+CD38−Mutant IDHWild Type IDH010000200003000040000Number of enhancersBM CD34+BM CD34+CD38−Mutant IDHWild Type IDHIDH wild type uniqueIDH mutant uniqueCommon0.0 0.5 1.0 1.5 2.00 2 40.0 2.5 5.0 7.5 10.012.5Pathways in cancercellular response to hormone stimulusdevelopmental growthresponse to woundinglymphocyte activationregulation of GTPase activitymyeloid leukocyte activationactin cytoskeleton organizationchemotaxisregulation of leukocyte proliferationdephosphorylationcarbohydrate derivative biosynthetic processlipid biosynthetic processchondrocyte differentiation involved in endochondral bone morphogenesisregulation of cell adhesion-log(q-Value)black795 893866IDHwtIDHmuA DBAP-1EHFGABPAFli1ETV4ETS1PU.1ERGEtv2ETV10 10 20 30-log10(q-Value)blackGroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38-Bone marrow CD34+IDH1 R132HIDH1 R132CIDH2 R172RIDH WT0.60.70.80.91roupT311Fra e hiftTT3882Troupone arro 34+ 38-one arro 34+I 1 132I 1 132I 2 172I T0.60.70.80.91GroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38-Bone marrow CD34+IDH1 R132HIDH1 R132CIDH2 R172RIDH WT0.60.70.80.91CEFGroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38−Bone marrow CD34+IDH1 R132HIDH1 R132CIDH2 R172RIDH WT0.50.60.70.80.91GroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38-Bone marrow CD34+IDH1 R132HIDH1 R132CIDH2 R172KIDH WT0.20.40.60.81050001000015000Number of enhancersBM CD34+BM CD34+CD38−Mutant IDHWild Type IDHG HGroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38−Bone marrow CD34+IDH1 R132HIDH1 R132CIDH2 R172RIDH WT0.20.40.60.81GroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38-Bone marrow CD34+IDH1 R132HIDH1 R132CIDH2 R172RIDH WT0.60.70.80.91GroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38-Bone marrow CD34+IDH1 R132HIDH1 R132CIDH2 R172RIDH WT0.60.70.80.91GroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38-Bone marrow CD34+IDH1 R132HIDH1 R132CIDH2 R172RIDH WT0.60.70.80.91 169 Figure 37. Leukemic blast shows increase in active enhancers. A) Unsupervised hierarchical clustering and heatmap of pairwise Spearman correlations for H3K27ac density at ChromHMM defined active enhancers indicated by the colour legend. B) Transcription binding site enrichment at active enhancers common in at least 3 IDH mutant with respect to active enhancers common in at least 2 IDH wild type. C) Venn-diagram of genes associated to active enhancers common in at least 3 IDH mutant and active enhancers common in at least 2 IDH wild type. D) Biological pathway analysis of genes associated to active enhancers common in at least 3 IDH mutant and active enhancers common in at least 2 IDH wild type. Number of active (E) and primed enhancers (F) across cell types as indicated by color legend. G) Unsupervised hierarchical clustering and heatmap of pairwise Spearman correlations for H3K27ac density at all enhancers marked with H3K27ac. H) Unsupervised hierarchical clustering and heatmap of pairwise Spearman correlations for H3K4me1 density at all enhancers marked with H3K4me1. 170 To investigate the change in enhancer status that accompanies leukemic transformation, we examined the transition of primed to active enhancers in a comparison of these in the normal adult BM CD34+CD38- cells and the AML cells. A majority of active enhancers present in all of the AML samples regardless of their IDH status were found to be marked by H3K4me1 in the CD34+CD38- population ( >93% , Figure 38A, C and D). Similarly, when comparing the IDH mutant to wild type AML cells, a majority of active enhancers in either group were primed with a strong H3K4me1 signal in the other group (Figure 38B and D). To assess if the AML specific enhancer landscape could be associated with a unique gene expression profile, enhancers that were present in the AML cells and not seen in the in adult BM CD34+ cells were identified, and the expression of the nearest genes examined. Interestingly, the resulting gene expression profiles successfully separated the normal from the leukemic cells, although a clear separation between the mutant and wild type IDH AML cells was not obtained (Figure 38E and F). These results support a model where AML blasts with different epigenetic landscapes may converge at transcription level. 171 0255075100IDH mutantIDH wild typePercentage denovoprimed in IDH wild typeprimed in mutant IDH0255075100IDH mutantIDH wild typePercentagedenovoprimed in CD34+CD38−A CWild TypeMutantCD34+CD34+CD38-H3K4me1H3K27acCommonMutantWild-Type0255075100IDH mutantIDH wild typePercentagedenovoprimed in CD34+CD38−0255075100IDH mutantIDH wild typePercentage denovoprimed in IDH wild typeprimed in mutant IDHBD E FGroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38−Bone marrow CD34+IDH1 R132HIDH1 R132CIDH2 R172RIDH WT−3−2−10123GroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38−Bone marrow CD34+IDH1 R132HIDH1 R132CIDH2 R172RIDH WT−3−2−10123GroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38−Bone marrow CD34+IDH1 R132HIDH1 R132CIDH2 R172RIDH WT−3−2−10123WTMutH3K27acH3K4me1FMBMchr18:60770000-60800000GroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38−Bone marrow CD34+IDH1 R132HIDH1 R132CIDH2 R172RIDH WT−3−2−10123 172 Figure 38. AML active enhancer landscape is poised in normal adult BM. A) Percentage of active enhancers that are marked with H3K4me1 in CD34+CD38- cells. B) Percentage of active enhancers that are marked with H3K4me1 in IDH mutant or wild type. C) Heatmap of H3K4me1 and H3K27ac densities coloured by cell type (Green : Normal ; Black : AML) at active enhancers and flanks (center of region ± 4 Kb) for common IDH mutant and wild type active enhancer (top panel), IDH mutant unique (meddle panel) or IDH wild type unique (bottom panel). Shading indicates increased density. D) Genome browser view of H3K27ac and H3K4me1 density and fractional methylation profiles at Chr18:60770000-60800000. Shaded area highlights active enhancer in IDH mutant or IDH wild type. E) Unsupervised hierarchical clustering and heatmap of expression Z-score for IDH mutant unique active enhancers associated genes indicated by the colour legend on bottom left. F) Unsupervised hierarchical clustering and heatmap of expression Z-score for IDH wild type unique active enhancers associated genes indicated by the colour legend on bottom left. 173 5.2.5 Heterochromatin in IDH mutant leukemic cells is influenced by co-occurring mutations. Despite the lack of difference in the extent of genomic occupancy of H3K27me3 and H3K9me3 between mutant and wild type IDH AML cells, genome-wide H3K9me3 and H3K27me3 signal intensity segregated the mutant and wild type IDH AML cells, irrespective of their NPM1 mutation status (Figure 39A and B). Examination of the signal intensity of the suppressive H3K27me3 mark and the active H3K4me3 mark at gene promoters (within 2 Kb of the TSS), revealed H3K4me3 signal can effectively segregate the leukemic cells from the normal adult BM CD34+CD38- cells, but cannot not separate the IDH mutant and wild type AML samples (Figure 39C). In contrast, H3K27me3 promoter density separated clearly all of these (Figure 39D). Within the IDH mutant branch of the dendrogram, a similar relationship to that previously shown using DNA methylation at active enhancers (Figure 34B) and the H3K9me3 signal genome wide (Figure 39A). Multidimensional scaling (MDS) analysis further clarified the separation of the IDH mutant and wild type AMLs based on their presence of a NPM1 mutation (Figure 39E). MDS thus grouped the AML samples into 3 independent clusters: the IDH mutant AML cells also with a mutant NPM1 (IDH mutant Cluster 1), the IDH mutant AML cells without a mutant NPM1 (IDH mutant cluster 2), and the AML cells with wild type IDH. The IDH mutant AML cells also showed greater overlap of H3K27me3 marked promoters compared to the IDH wild type AML cells (IDH mutant cluster 1: 4444, IDH mutant cluster 2: 4025 and IDH wild type: 3633) reflecting a greater similarity in the H3K27me3 landscape among the 2 groups of AML cells with a mutant IDH (Figure 39F-H). 174 Biv H3K27me3 H3K4me3030006000900001000200030004000010002000Number of promoters blackBone Marrow CD34+ CD38-Bone Marrow CD34+IDH mutant cluster 1IDH mutant cluster 2IDH wild typeH3K4me3H3K27me3Biv-1000 0 10002000IDH mutant cluster 1IDH mutant cluster 2IDH wild typeIDH mutant cluster 1IDH mutant cluster 2IDH wild typeIDH mutant cluster 1IDH mutant cluster 2IDH wild typeNumberGainLossIDH mutant cluster 1IDH mutant cluster 2IDH wild typeGroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38-Bone marrow CD34+IDH1 R132HIDH1 R132CIDH2 R172RIDH WT0.80.850.90.951FJ KBiv H3K27me3 H3K4me3030006000900001000200030004000010002000Number of promoters blackBone Marrow CD34+ CD38-Bone Marrow CD34+IDH mutant cluster 1IDH mutant cluster 2IDH wild typeBiv H3K27me3 H3K4me3030006000900001000200030004000010002000Number of promoters blackBone Marrow CD34+ CD38-Bone Marrow CD34+IDH mutant cluster 1IDH mutant cluster 2IDH wild typeGroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38-Bone marrow CD34+IDH1 R132HIDH1 R132CIDH2 R172RIDH WT0.60.70.80.91H3K4me3H3K27 e3Biv-1000 0 10002000IDH mutant cluster 1IDH mutant cluster 2IDH wild typeIDH mutant cluster 1IDH mutant cluster 2IDH wild typeIDH mutant cluster 1IDH mutant cluster 2IDH wild typeNumberGainLossIDH mutant cluster 1IDH mutant cluster 2IDH wild type633 392 6623874444452268IDH1_rep1 IDH1_rep2IDH1_rep3IDH cluster 1583 688 3264094025175429IDHC IDH2_rep1IDH2_rep2IDH cluster 2416 261 389134536333521469IDHwt_rep1 IDHwt_rep2IDHwt_rep3IDH wild typeG H IGroDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38-Bone marrow CD34+IDH1 R132HIDH1 R132CIDH2 R172RIDH WT0.60.70.80.91EGroupDNMT3ANPM1GroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38-Bone marrow CD34+0.20.40.60.81GroupDNMT3ANPM1GroupDNMT3ANPM1NPM1Frame ShiftWTDNMT3AR882HWTGroupBone marrow CD34+ CD38-Bone marrow CD34+0.20.40.60.81BM_CD3438_rep1BM_CD3438_rep2BM_CD3438_rep3BM_CD34IDH1_rep1IDH1_rep2IDH1_rep3IDH2_rep1IDH1CIDHwt_rep2IDHwt_rep1IDH2_rep2 IDHwt_rep3-0.50-0.250.000.25-0.5 0.0 0.5Coordinate 1Coordinate 2A B CD 175 Figure 39. H3K27me3 density at promoters separates mutant IDH based on NPM1 and DNMT3A mutation status. A) Unsupervised hierarchical clustering and heatmap of pairwise Spearman correlations for H3K9me3 density at H3K9me3 marked regions (A) and for H3K27me3 density at H3K27me3 marked regions (B) indicated by the colour legend. Unsupervised hierarchical clustering and heatmap of pairwise Spearman correlations for H3K4me3 (C) and H3K27me3 (D) density at promoters of coding genes (+/- 2Kb TSS) indicated by the colour legend. E) Multidimensional scaling of H3K27me3 density at promoters of coding genes (+/- 2Kb TSS). Venn-diagram of overlap of H3K27me3 marked promoter across IDH1R132H (F), IDH1R132C/IDH2R172K (G) and IDH wild type (H) replicates. I) Expression of genes with bivalent, H3K4me3 and H3K27me3 marked promoters across normal bulk CD34+ BM and AML cells. J) Number of bivalent, H3K4me3 and H3K27me3 marked promoters. K) Number of total bivalent, H3K4me3 and H3K27me3 marked promoters that are gained or lost in comparison to normal bulk CD34+ BM cells in IDH mutant and wild type. 176 Examination of bivalently marked promoters (as defined by the presence of both H3K4me3 and H3K27me3) showed these were associated with low levels of gene expression in all cell types and no significant increase in bivalent promoters was observed in association with the IDH status of the AML cells (Figure 39I and J). There was also a directional gain of H3K27me3 in AML cells regardless of their IDH status compare to adult bone marrow bulk CD34+ cells (Figure 39J and K). However, there was no significant difference in gain or loss of H3K27me3 between IDH mutant and wild type AML cells. Together, these results suggest that IDH mutant may share similar suppressive chromatin landscape and increase in H3K27me3 and/or H3K9me3 are independent of IDH mutation status. 5.2.6 AML cells retain a progenitor H3K27me3 structure In Chapter 4, a genome wide contraction of H3K27me3 in differentiated myeloid cells compared to progenitors isolated from CB was described. Interestingly, the primary AML blasts, regardless of their IDH status appeared to retain a progenitor-like H3K27me3 structure (Figure 40A). They also displayed a broad IP fragment distribution at promoters (Figure 40B) and LOCKs lost in differentiated normal CB monocytes and erythroid precursors were retained in the AML blasts (Figure 40C). In addition, genomic regions enriched in H3K27me3 containing LOCKs in AML blasts showed a broad distribution similar to that observed in the CB CD34+ progenitor populations and differentiated lymphoid cells (Figure 40D-F). This similarity between the polycomb structure in the AML blasts and normal CB progenitor populations suggest a potential link in their epigenomic features as well as their lack of morphological evidence of differentiation (Figure 40G). 177 HSCMyeloidAMLAIDH1 R132H rep1IDH1 R132H rep2IDH1 R132H rep3IDH1 R132CIDH2 R172R rep1IDH2 R172R rep2IDHwt rep1IDHwt rep2IDHwt rep3CD34+ CD34+CD38- rep1446158Progenitor population Primary AML< 2.2e-16< 2.2e-166.8e-115.4e-08< 2.2e-16< 2.2e-162468MonocyteErythroidAML rep1AML rep2AML rep3log10(# of bps covered)MonocyteErythroidAML rep1AML rep2AML rep3Erythroid-4000-2000 0200040000.60.70.80.91.0Max Normalized CoverageAML.Rep1AML.Rep2AML.Rep3ErythroidMonocyte-4000-2000 0200040000.60.70.80.91.0Max Normalized CoverageAML.Rep1AML.Rep2AML.Rep3MonocyteH3K27me3 H3K4me3-2000 0 2000-2000 0 20000.000.250.500.751.00Max Normalized CoverageAML rep1AML rep2AML rep3BFC DEGCD34+CD38- rep2CD34+CD38- rep3Chr20:49,000,000-50,500,000 178 Figure 40. Hematopoietic progenitor H3K27me3 LOCKs are retained in AML blasts. A) Genome browser view of H3K27me3 density on chromosome 20 across cell types as indicated by the colour legend on the bottom right. A H3K27me3 LOCK (FDR <0.05) present in progenitor cells but absent in CB monocyte and erythroid precursors is indicated by the shaded box. B) Maximum value normalized H3K27me3 (left panel) and H3K4me3 (right panel) density at coding gene promoters (± 2 Kb of TSS). C) Venn diagram of LOCKS common to progenitors compared to those present in AML blasts. D) Violin plot of base pairs marked by H3K27me3 within LOCKs across cell types as indicated. Maximum value normalized H3K27me3 density at erythroid precursors (E) and monocytes (F) H3K27me3 enriched regions within LOCKs identified in AML blasts. G) Cartoon illustrating contraction and retention of H3K27me3 (red) during myeloid differentiation and AML transformation, respectively. 179 5.3 Discussion Perturbations in the epigenetic control of the genome, in part driven by gain and loss of function mutations to epigenetic regulators, have emerged as recurrent observations in human hematologic malignancies including AML (Holz-Schietinger et al., 2012; Martin-Perez et al., 2010). The genomes of over 40% of primary AMLs harbor mutations that effect epigenetic factors that maintain DNA methylation and histone homeostasis, including neomorphic mutations in IDH (Ley et al., 2013). Mutations in IDH have been associated with increases in H3K9me3 and H3K27me3, DNA methylation and a CIMP phenotype in AML (Figueroa et al., 2010). Interrogation of changes in the epigenomic state that distinguish AMLs with neomorphic gain of function IDH mutations from those with a normal IDH genotype and/or normal adult BM CD34+CD38- cells revealed numerous differences. Comparisons of an extensive dataset of genome wide histone modifications, DNA methylation and transcriptome datasets revealed hyper-methylation at non-promoter regulatory regions and a specific heterochromatin signature associated with gain of function IDH mutations. AML blasts harbouring a mutant IDH allele were found to possess a unique epigenetic landscape that allowed their distinction from AML cells with a wild type IDH genotype, despite a convergence at the level of gene transcription. This is consistent with the concept that a heterogenous epigenetic landscape in different AML clones nevertheless ultimately drives a similar transcriptome profile to achieve a growth advantage and a blockade of terminal myeloid differentiation. In agreement with previous reports, DNA methylation alterations driven by neomorphic IDH mutations in the AML samples examined here occurred mainly outside CGIs (Kelly et al., 2017). However, the results also suggest that these IDH mutations do not drive CIMP in AML contrary to the widely accepted notion that CIMP is mutant IDH-dependent (Figueroa et al., 2010). 180 By partitioning the genome by histone modification states and integrating genome-wide single CpG methylation datasets, I observed that consistent hyper-methylation associated with neomorphic IDH mutations was primarily found at non-promoter regulatory regions marked with H3K27ac. Hyper-methylation of enhancers would be predicted to disrupt the interplay of TFs and epigenetic modifications that control hematopoiesis in two ways. It would prevent binding of TFs with low affinity to methylated CpGs and it would also promote the binding of TFs which are promiscuous or have higher affinity to methylated CpGs. Such disruptions would thus be expected to drive both increases and decreases in transcription of affected genes depending on the nature of the altered TF binding to particular enhancers. This concept is supported by the observation of an inconsistent alteration in the transcription of genes and a change in the fractional methylation of their corresponding enhancers. In recent studies, most cancer methylation profiling efforts have focused almost exclusively on hyper-methylation of CGIs (Arechederra et al., 2018; Brinkman et al., 2019; Chen et al., 2016; Saghafinia et al., 2018). The present results highlight the importance of aberrant methylation of sparse CpGs at CpG-poor non-promoter regulatory regions in AML cells. This study also highlights the effects of NPM1 mutations with neomorphic IDH mutations in the same cells. DNA methylation, H3K9me3 and H3K27me3 signal consistently grouped the mutant IDH based on the presence or absence of a mutant NPM1 genotype. IDH mutant AML cells with a mutant NPM1 allele showed a lower number of hyper-methylated DMRs compared to IDH mutant cells that lacked a NPM1 mutation. NPM1 has previously been associated with a hypo-methylation phenotype in AML (Kelly et al., 2018). Discrepancy among the mutant IDH replicates maybe the result of antagonistic effects of NPM1 and IDH mutations on DNA methylation and 181 H3K27me3. It seems also possible that NPM1 mutations may correlate with a lower level of H3K27me3 as it does with a hypo-methylation phenotype. 182 Chapter 6: Conclusion 6.1 Interpretation and significance A complex choreographed dance between a network of TFs and chromatin controls cellular differentiation and specification. Too many faulty steps in this dance would be incompatible with cell viability, but lesser perturbations would be expected to disrupt the transcriptional balance that regulates normal cellular proliferation and differentiation and result in a self-sustaining transformed state. To identify pathogenic disruptions to transcriptional control and further our understanding of their consequences, sensitive technologies that can accurately measure relevant molecular features in normal and diseased cellular states are required. 6.1.1 ndChIP-seq is a robust assay for the simultaneous measurement of histone modification and nucleosome density in rare cell populations. Given the combinatorial nature of chromatin modification and nucleosome positioning in transcriptional regulation, a method that enables simultaneous measurements of these features is likely to provide new insights into epigenetic regulation (Bernstein et al., 2002, 2006; Henikoff, 2008; Maehara and Ohkawa, 2016; Voong et al., 2016). The ndChIP-seq protocol presented here is a native ChIP-seq protocol optimized to enable simultaneous interrogation of histone modification and nucleosome density in lower numbers of cells than have been previously possible. This was key to enabling ChIP-seq to be applied to primitive subsets of primary sources of hematopoietic cells which cannot be accessed in the numbers historically required. The method described in Chapter 3 utilizes enzymatic digestion of chromatin that, when coupled to paired-end massively parallel sequencing and a Gaussian mixture distribution model, allows for the investigation of histone modifications at the single nucleosome level and the deconvolution of 183 epigenomic profiles driven by heterogeneity within a population of cells. Application of the ndChIP-seq method to generate H3K4me3 and H3K27me3 immunoprecipitated DNA fragment size distributions from a rare subset (CD34+) of normal human CB cells (10,000 cells per sample) was shown to enable an analysis of gene promoters enriched in single versus two nucleosomes and their association with expression, chromatin and methylation states. NdChIP-seq revealed four classes of promoter-specific profiles that effectively segregated apparent bivalently marked promoters into those likely to be distributed in different cells, and a small subset likely to be bivalent promoters at the single cell level. Correlation of these different promoter profiles with matching gene expression profiles also indicated that each state was associated with a unique transcriptional profile. Strikingly, extension of the use of the ndChIP-seq method to human ESC cells revealed an altered relationship between chromatin modification state and the nucleosome content in promoters as compared to that obtained for CB CD34+ cells. The ndChIP-seq method also made it possible to investigate genome wide histone modifications in combination with nucleosome density assessments to obtain new insights into mechanisms that regulate the unique transcriptomes of early and late stages of normal human hematopoietic cell populations not previously accessible to such analyses. 6.1.2 Polycomb contraction differentially regulates lymphoid and myeloid differentiation H3K27me3 plays an important role in maintaining multipotency and in regulating differentiation in various tissues and physiological niches (Margueron and Reinberg, 2011; Zhu et al., 2013). The importance of H3K27me3 in the regulation of normal human and mouse hematopoiesis has been demonstrated (Kamminga et al., 2006; Petruk et al., 2017; Xie et al., 2014; Xu et al., 2015; Yu et al., 2017; Mochizuki-Kashio et al., 2011). In addition, both overexpression 184 and inactivation of components of the PRC2 complex, which is responsible for the methylation of H3K27, have been reported in hematopoietic malignancies, emphasizing an important role of H3K27me3 in maintaining normal hematopoiesis (Ernst et al., 2010; Ganji et al., 2012). Building on these findings, the results presented in Chapter 4 provide new insights into the role and orchestration of H3K27me3 changes during normal hematopoiesis. These included the first measurements of differences in genome-wide histone features of different subsets on early and late stages of normal hematopoietic cells isolated from pooled samples of human cord blood (i.e., CD34+CD38- cells and CMPs, GMPs and MEPs, and terminally differentiated erythroid precursors, monocytes, B cells and T cells). Comparisons of repressive H3K27me3 landscapes across these cell types revealed a strikingly stable polycomb signature in contrast to the highly cell type specific signatures for active histone modifications in the progenitor populations and a striking contraction of H3K27me3 occupancy unique to the differentiated myeloid cells with a structure reminiscent of a H3K27me3 signature previously associated with pluripotent stem cells (Hawkins et al., 2010; Mikkelsen et al., 2007; Zhu et al., 2013). During ESC differentiation punctate H3K27me3 foci expand to form large (10-100s Kb) blocks (Hawkins et al., 2010), whereas our results indicate a reverse trend during hematopoietic differentiation with large H3K27me3 domains present in the progenitors that contract to a punctate state upon their differentiation to myeloid cells, but not lymphoid cells. This finding suggests differentiation of myeloid cells is accompanied by a loss of repressive chromatin rather than a further compaction as might have been expected (Hawkins et al., 2010). The similarity of the H3K27me3 punctate structure across differentiated myeloid and ESCs is intriguing and questions the previous association that this structure is reflective of an undifferentiated state. 185 H3K27me3 LOCKs identified in progenitor and lymphoid lineage cells show co-occupancy with another suppressive histone modification, H3K9me3, which is also lost in H3K27me3 LOCKs in myeloid cells. Furthermore, H3K9me3 shows the same pattern of genome wide reduction as H3K27me3. The drastic restructuring of H3K27me3 and H3K9me3 at the terminal stage of differentiation in myeloid cells suggests the importance of higher order chromatin structure in differentiation. Intergenic H3K27me3 and H3K27/K9me3 marked regions are strongly enriched in lamina-associated domains (LADs) only in progenitor and lymphoid cells. Myeloid cells can also be distinguished from lymphoid cells and progenitors based on lamin distribution and nucleus rigidity (Swift et al., 2013). As in the case of H3K27me3 regulation, manipulation of lamin expression modulates the terminal differentiation of myeloid cells (Swift et al., 2013). The coordinated global reduction in DNA methylation in myeloid cells as compared to both progenitor and lymphoid cells, and their accompanying loss of H3K27me3 LOCKs, a genome wide reduction in H3K27me3 and H3K9me3 occupancy and visible nuclear diversity point to a highly important lineage-specific change in chromatin structure in generation and establishing cellular identity of myeloid and lymphoid lineage cells. Furthermore, loss of H3K27me3 has been associated with the onset of senescence (Ito et al., 2018). In addition, a global loss of mega base pairs of H3K9me3 has been annotated in an in vitro model of Werner Syndrome, which is a premature aging disorder representative of accelerated senescence (Zhang et al., 2015). These observations raise a fascinating possibility that there are parallels in chromatin restructuring during senescence and myeloid cell maturation. In contrast to the apparent stable H3K27me3 signature shared by the different progenitor populations, the H3K27ac profile was quite variable. The data presented here suggest that the priming of regulatory regions of DNA in progenitors occur concomitantly with the creation of an 186 open chromatin state. The appearance of lineage-specific active enhancers in progenitor populations before genome wide H3K27me3 rearrangements take place suggests that activation of myeloid lineage-specific enhancers is not sufficient to initiate terminal myeloid differentiation events and massive restructuring of the polycomb signature is also needed. Thus an orchestrated interplay between H3K27 modifiers appears to play a critical role in human hematopoietic differentiation. 6.1.3 Mutant IDH dependent DNA hypermethylation occurs at intergenic regulatory regions Histone modifications and DNA methylation are gate keepers of chromosome. They facilitate the binding of TFs to DNA and in turn regulate transcription. Abruption in the function of epigenetic modifiers leads to disruptions of intricate network of TFs and epigenetic modifications which ultimately leads to abnormal transcription. Epigenetic modifiers are recurrently mutated in AML (Ley et al., 2013). However, consequences of these mutation on epigenetic landscape is poorly understood. To further study the abnormal epigenome that allows normal blood cells to acquire leukemic cell properties, I focused here on a comparison of the epigenome of primary AML cells with and without neomorphic IDH mutations. Neomorphic IDH mutations are recurrent genetic lesions observed in AML that drive the accumulation of D2HG (Dang et al., 2009), an inhibitor of 2-OGDD enzymes, a family that includes a class of histone and DNA demethylases (Xu et al., 2011). Interestingly, an increase in DNA methylation, and the acquisition of H3K9me3 and H3K27me3 are associated with presence of D2HG based on immunohistochemistry assays (Figueroa et al., 2010; Kats et al., 2014; Lu et al., 2012). Moreover, CIMP signature in AML has come to be associated with mutant IDH AMLs (Figueroa et al., 2010) 187 based on a study of DNA methylation in a cohort of 200 patients using 450K array technology. Here this association was compared with the histone modifications and DNA methylation landscape at single base pair resolution across normal CD34+CD38- bone marrow cells and leukemic blasts with and without IDH mutations. Contrary to accepted view, I found that CIMP AML is IDH-independent. Moreover, I was able to define a novel hyper-methylation phenotype at active non-promoter regulatory regions enriched in H3K27ac and H3K4me1 histone modifications and a fractional methylation signal successfully segregated the mutant IDH AMLs from the wild type AMLs only at these active regulatory regions. Current cancer methylome studies involving primary tissues focus almost exclusively on annotating CGI methylation. However, the findings in this thesis serve to highlight the importance of sparse CpGs and consequences that they may have on TF binding and transcription regulation. In addition, they further emphasize the importance of annotation of methylation fraction of all CpGs in cancer methylome studies for full appreciation of the aberrant methylome consequences. Hypermethylated active enhancers are highly enriched in the binding sites of ETS family TFs. This is intriguing, since the members of this family have similar motifs but exhibit different affinities for methylated DNA. This suggests that there is a possibility of binding of incorrect TF with similar motif but higher affinity for methylated DNA at these regions. For example, binding sites of PU.1 and ETS1 were found to be enriched at hypermethylated enhancers. PU.1 has a higher affinity for methylated DNA compare to ETS1. Here, it is proposed that DNA methylation at active enhancers influences transcription in two fashions. First, methylation of enhancers prevents binding of TFs and subsequently causes downregulation of associated genes. Second, it causes alteration in TF binding at these regulatory regions where there is a selective advantage for promiscuous TFs to bind methylated DNA over those whose binding is blocked when the DNA is 188 methylated (Figure 41). The combination of these two mechanisms would then be expected to underlie the abnormal transcription profile of mutant IDH AML cells. In addition, the CpG-poor nature of these non-promoter regulatory regions implies that methylation of a few sparse CpGs is sufficient to alter TF binding and effect transcription as in the case of promoters (Lioznova et al., 2019; Medvedeva et al., 2014), further emphasizing the importance of CpG methylation outside CGIs. Among IDH mutant AMLs, a consistent grouping was observed based on the genome wide H3K9me3 and H3K27me3 density in the cells, and the fractional DNA methylation at their active enhancers. This subgrouping of the AMLs with a mutant IDH genotype may be a result of other background mutations as suggested by the associations seen in those with and without NPM1 mutations. This is intriguing since NPM1 mutations in AML are associated with a hypo-methylation signature (Kelly et al., 2018). Thus the unique signature identified here maybe a result of antagonistic consequences of NPM1 and IDH mutations on DNA methylation and histone modifications. This highlights the importance of background mutations and their likely synergetic or antagonistic interactions with mutant IDH that could be of great importance in the considerations of epigenetic modifiers as a therapeutic option. 6.2 Limitations and future directions 6.2.1 Limitation of ndChIP-seq Unlike previous iterations of native ChIP-seq protocols, ndChIP-seq provides a means to investigate the combinatorial effect of chromatin structure and histone modification by utilizing fragment size post immunoprecipitation to integrate nucleosome density, determined by MNase accessibility, with histone modification measurements. Application of ndChIP-seq to primary cells 189 and tissues can thus provide novel insights into the integrative nature of epigenetic regulation and permit identification of epigenetic signatures due to heterogeneity within the population. However, there are also limitations to its use. ndChIP requires a minimum sequencing depth of 100 million paired-reads (50 million fragments) for histone marks and input, as the Gaussian mixture distribution algorithm will not perform optimally on libraries that have not been sequenced to this depth. Implementation of Cleavage Under Targets and Tagmentation (CUT&Tag) can address this short coming. CUT&Tag is a new iteration of ChIP-seq in which the Protein-A/G-Tn5 transposase complex recognizes the antibody at the site of tagmentation and reduces the required library depth to <10 million fragments with reduced background noise (Kaya-Okur et al., 2019). Furthermore, ndChIP-seq will not classify promoters with little separation between the weighted distribution value for mono- and di-nucleosome fragment lengths into mono- or di-nucleosome dominated promoters. Therefore, these promoters must be removed in the subsequent analysis. In addition, biological replicates are required to increase confidence in predicted distributions and identify technical variability in the MNase digestion and library construction. Application of a version of our analytical frame work has recently provided new insight into TF affinity in motif binding at accessible DNA and DNA wrapped around nucleosome (Meers et al., 2019). 6.2.2 Identification of mechanisms regulating contraction of H3K27me3 in differentiated myeloid cells H3K27me3 plays a major role in hematopoiesis. In this thesis, it is shown that contraction of polycomb differentially regulates myeloid and lymphoid differentiation. However, the mechanism(s) that maintain or remove H3K27me3 at LOCKs have not been identified. Comparison of expression of PRC1 and PRC2 components across hematopoietic cells profiled in 190 this work and human H1 ESCs revealed BMI1 as the only gene that showed a lower expression in monocytes or erythroid precursors and H1 ESCs as compared to other cell types. This expression pattern was corroborated in a mouse hematopoietic model (Lara-Astiaso et al., 2014). BMI1 is a component of PRC1, however, its downregulation has been associated with loss of H3K27me3 and H3K9me3 (Hyland et al., 2011). In addition, inhibition of BMI1 in vivo disrupts the self-renewal capacity of hematopoietic progenitors (Park et al., 2003). Collectively, these results suggest involvement of BMI1 in the maintenance of H3K27me3 spreading in the progenitor and lymphoid populations. To define the role of BMI1 in maintenance of LOCKs, I propose investigation of effect of BMI1 inhibitor on granulopoietic and/or B-lymphopoietic differentiation capacity of hematopoietic progenitors. Understanding this mechanism has clinical importance, since the spreading of H3K27me3 is also seen in primary AML samples regardless of their World Health Organization (WHO) classification. Inhibition of both EZH2 and BMI1 compromises the self-renewal capacity of hematopoietic progenitors (Kamminga et al., 2006; Park et al., 2003). In addition, inhibition of EZH2 in myeloproliferative cell line (HL-60) results in a proliferation halt. Moreover, an EZH2 inhibitor is currently used as a therapeutic option for various cancer types. Together, the results presented in Chapter 4 raise the question of whether a combination of EZH2 and BMI1 inhibitors might be a potent therapeutic option in AML. 6.2.3 TF switching events are mediated by hypermethylated active enhancers to lead to abnormal transcription profiles This work provides a high quality, complete and rich resource of genome wide methylome, histone modification profile and transcriptome of mutant and wild type IDH AML cells. However, a shortcoming of this study is the number of replicates used (6 mutant IDH AML samples and 3 191 wild type IDH AML samples). Primary AML samples have high degree of heterogeneity in their mutational, epigenetic and transcription profiles; therefore, greater numbers of primary samples are required to establish the consistency of the observations made in Chapter 5. IDH mutants show a unique hyper-methylation signature at intergenic active enhancers. These regions are significantly enriched in motifs belonging to the ETS family of TFs. Members of this family, such as PU.1 and ETS1, are thought to play a role in regulating hematopoietic differentiation (Ciau-Uitz et al., 2013; Kastner and Chan, 2008; Laslo et al., 2006; Lulli et al., 2006). These TFs have similar motifs but exhibit different affinities for methylated DNA (Stephens and Poon, 2016). These results raise the question of whether hyper-methylation of active enhancers promotes binding of DNA methylation-insensitive TFs over those that are sensitive to DNA methylation. Genome wide profiling of DNA methylation, H3K4me1, H3K27ac and ETS family of TFs in an in vitro model of leukemia using HOXA9 immortalized mouse BM cells with and without a IDH1R132C mutation, help to address this question. Profiling H3K4me1, H3K27ac and DNA methylation genome wide would identify enhancers that are methylated in presence of the IDH1R132C mutation so that these regions could then be assessed for TF switching events using ETS family TFs ChIP-seq. RNA-seq along with Chromatin interaction analysis by paired end tag sequencing (ChIA-PET) experiments can identify genes whose activity is affected by the hyper methylated enhancers. Lastly, synergistic and antagonistic effects of recurrent mutations in AML with mutant IDH suggested in this work need to be further investigated in a larger cohort of AML samples. 192 6.3 Concluding remarks The chromatin state dictated through epigenetic modifications modulates transcriptome and in turns controls proliferative and differentiation capacity of a cell. The epigenome of a progenitor cell undergoes dramatic rearrangement during differentiation to accommodate a new cellular identity with reduced differentiation capacity. Hematopoiesis is an excellent model for studying how epigenetic modifications may regulate self-renewal and differentiation activities. This thesis is a collection of research carried out to further our understanding of epigenetic regulation of hematopoiesis. Application of the ndChIP-seq protocol and other low input assays to primary normal human hematopoietic cells and AML cells revealed a novel model whereby a spreading and contraction of repressive histone modifications (i.e. H3K27me3 and H3K9me3) occupancy regulates proliferation and differentiation capacity of hematopoietic progenitors. This major observation highlights the importance of chromatin structure in regulating self-renewal and differentiation. Extension of this type of analysis to AML cells unveiled a unique CpG methylation phenotype at intergenic regulatory regions as a result of the inhibition of epigenetic modifiers. In light of the recent clinical application of inhibitors of epigenetic modifiers in various cancer types, the present findings may be exploited to design novel therapeutics to combat AML. At the same time, questions still remain with respect to specific epigenetic events reflecting lineage commitments. I anticipate that future work will focus on epigenome measurements at the single-cell level to address these issues through the provision of a more complete epigenetic landscape for all progenitor cell types. 193 DNA Methylation insensitive TFDNA Methylation sensitive TFMethylated CpGUnmethylated CpGH3K27acX methylation sensitive TF binding inhibitionTF switching 194 Figure 41. Transcription factor switching at hyper-methylated enhancers in mutant IDH. Cartoon illustrating two models of transcription modulation by hyper-methylation of active enhancers: (i) inhibition of binding of TF sensitive to DNA methylation (top row) and (ii) promotion of binding of TF promiscuous to DNA methylation over TF sensitive to DNA methylation with similar motif (bottom row). 195 Bibliography Abbas, S., Lugthart, S., Kavelaars, F.G., Schelen, A., Koenders, J.E., Zeilemaker, A., Van Putten, W.J.L., Rijneveld, A.W., Löwenberg, B., and Valk, P.J.M. (2010). Acquired mutations in the genes encoding IDH1 and IDH2 both are recurrent aberrations in acute myeloid leukemia: Prevalence and prognostic value. Blood 116, 2122–2126. Abdel-Wahab, O., Mullally, A., Hedvat, C., Garcia-Manero, G., Patel, J., Wadleigh, M., Malinge, S., Yao, J.J., Kilpivaara, O., Bhat, R., et al. (2009). Genetic characterization of TET1, TET2, and TET3 alterations in myeloid malignancies. Blood 114, 144–147. Abdel-Wahab, O., Patel, J., and Levine, R.L. (2011). Clinical Implications of Novel Mutations in Epigenetic Modifiers in AML. Hematol. Oncol. Clin. North Am. 25, 1119–1133. Agger, K., Cloos, P.A.C., Christensen, J., Pasini, D., Rose, S., Rappsilber, J., Issaeva, I., Canaani, E., Salcini, A.E., and Helin, K. (2007). UTX and JMJD3 are histone H3K27 demethylases involved in HOX gene regulation and development. Nature 449, 731–734. Akhtar-Zaidi, B., Cowper-Sallari, R., Corradin, O., Saiakhova, A., Bartels, C.F., Balasubramanian, D., Myeroff, L., Lutterbaugh, J., Jarrar, A., Kalady, M.F., et al. (2012). Epigenomic enhancer profiling defines a signature of colon cancer. Science. 336, 736–739. Albano, F., Anelli, L., Zagaria, A., Coccaro, N., Minervini, A., Rossi, A.R., and Specchia, G. (2011). Decreased TET2 gene expression during chronic myeloid leukemia progression. Leuk. Res. 35. Allfrey, V.G., and Mirsky, A.E. (1964). Structural modifications of histones and their possible role in the regulation of RNA synthesis. Science. 144, 559. Allis, C.D., Berger, S.L., Cote, J., Dent, S., Jenuwien, T., Kouzarides, T., Pillus, L., Reinberg, D., Shi, Y., Shiekhattar, R., et al. (2007). New Nomenclature for Chromatin-Modifying 196 Enzymes. Cell 131, 633–636. Álvarez-Errico, D., Vento-Tormo, R., Sieweke, M., and Ballestar, E. (2015). Epigenetic control of myeloid cell differentiation, identity and function. Nat. Rev. Immunol. 15, 7–17. Arechederra, M., Daian, F., Yim, A., Bazai, S.K., Richelme, S., Dono, R., Saurin, A.J., Habermann, B.H., and Maina, F. (2018). Hypermethylation of gene body CpG islands predicts high dosage of functional oncogenes in liver cancer. Nat. Commun. 9. Arita, K., Ariyoshi, M., Tochio, H., Nakamura, Y., and Shirakawa, M. (2008). Recognition of hemi-methylated DNA by the SRA protein UHRF1 by a base-flipping mechanism. Nature 455, 818–821. Avery, O.T., Macleod, C.M., and McCarty, M. (1944). Studies on the chemical nature of the substance inducing transformation of pneumococcal types: Induction of transformation by a desoxyribonucleic acid fraction isolated from pneumococcus type iii. J. Exp. Med. 79, 137–158. Babovic, S., and Eaves, C.J. (2014). Hierarchical organization of fetal and adult hematopoietic stem cells. Exp. Cell Res. 329, 185–191. Bannister, A.J., and Kouzarides, T. (1996). The CBP co-activator is a histone acetyltransferase. Nature 384, 641–643. Bannister, A.J., and Kouzarides, T. (2011). Regulation of chromatin by histone modifications. Cell Res. 21, 381–395. Barski, A., Cuddapah, S., Cui, K., Roh, T.Y., Schones, D.E., Wang, Z., Wei, G., Chepelev, I., and Zhao, K. (2007). High-Resolution Profiling of Histone Methylations in the Human Genome. Cell 129, 823–837. Baubec, T., Colombo, D.F., Wirbelauer, C., Schmidt, J., Burger, L., Krebs, A.R., Akalin, A., and Schübeler, D. (2015). Genomic profiling of DNA methyltransferases reveals a role for DNMT3B 197 in genic methylation. Nature 520, 243–247. Becker, J.S., Nicetto, D., and Zaret, K.S. (2016). H3K9me3-Dependent Heterochromatin: Barrier to Cell Fate Changes. Trends Genet. 32, 29–41. Bernard, D., Gil, J., Dumont, P., Rizzo, S., Monté, D., Quatannens, B., Hudson, D., Visakorpi, T., Fuks, F., and De Launoit, Y. (2006). The methyl-CpG-binding protein MECP2 is required for prostate cancer cell growth. Oncogene 25, 1358–1366. Bernstein, B.E., Humphrey, E.L., Erlich, R.L., Schneider, R., Bouman, P., Liu, J.S., Kouzarides, T., and Schreiber, S.L. (2002). Methylation of histone H3 Lys 4 in coding regions of active genes. Proc. Natl. Acad. Sci. U. S. A. 99, 8695–8700. Bernstein, B.E., Mikkelsen, T.S., Xie, X., Kamal, M., Huebert, D.J., Cuff, J., Fry, B., Meissner, A., Wernig, M., Plath, K., et al. (2006). A Bivalent Chromatin Structure Marks Key Developmental Genes in Embryonic Stem Cells. Cell 125, 315–326. Billia, F., Barbara, M., McEwen, J., Trevisan, M., and Iscove, N.N. (2001). Resolution of pluripotential intermediates in murine hematopoietic differentiation by global complementary DNA amplification from single cells: confirmation of assignments by expression profiling of cytokine receptor transcripts. Blood 97, 2257–2268. Bock, C., Beerman, I., Lien, W.H., Smith, Z.D., Gu, H., Boyle, P., Gnirke, A., Fuchs, E., Rossi, D.J., and Meissner, A. (2012). DNA Methylation Dynamics during In Vivo Differentiation of Blood and Skin Stem Cells. Mol. Cell 47, 633–647. Borger, D.R., and Zhu, A.X. (2012). IDH mutations: New genetic signatures in cholangiocarcinoma and therapeutic implications. Expert Rev. Anticancer Ther. 12, 543–546. Bostick, M., Jong, K.K., Estève, P.O., Clark, A., Pradhan, S., and Jacobsen, S.E. (2007). UHRF1 plays a role in maintaining DNA methylation in mammalian cells. Science (80-. ). 317, 1760– 198 1764. Boulard, M., Edwards, J.R., and Bestor, T.H. (2015). FBXL10 protects Polycomb-bound genes from hypermethylation. Nat. Genet. 47, 479–485. Boyer, L.A., Plath, K., Zeitlinger, J., Brambrink, T., Medeiros, L.A., Lee, T.I., Levine, S.S., Wernig, M., Tajonar, A., Ray, M.K., et al. (2006). Polycomb complexes repress developmental regulators in murine embryonic stem cells. Nature 441, 394–353. Boyes, J., and Bird, A. (1991). DNA methylation inhibits transcription indirectly via a methyl-CpG binding protein. Cell 64, 1123–1134. Breitenbuecher, F., Schnittger, S., Grundler, R., Markova, B., Carius, B., Brecht, A., Duyster, J., Haferlach, T., Huber, C., and Fischer, T. (2009). Identification of a novel type of ITD mutations located in nonjuxtamembrane domains of the FLT3 tyrosine kinase receptor. Blood, 113,4074–44077. Breitman, T.R., Selonick, S.E., and Collins, S.J. (1980). Induction of differentiation of the human promyelocytic leukemia cell line (HL-60) by retinoic acid. Proc. Natl. Acad. Sci. U. S. A. 77, 2936–2940. Brind’Amour, J., Liu, S., Hudson, M., Chen, C., Karimi, M.M., and Lorincz, M.C. (2015). An ultra-low-input native ChIP-seq protocol for genome-wide profiling of rare cell populations. Nat. Commun. 6. Brinkman, A.B., Nik-Zainal, S., Simmer, F., Rodríguez-González, F.G., Smid, M., Alexandrov, L.B., Butler, A., Martin, S., Davies, H., Glodzik, D., et al. (2019). Partially methylated domains are hypervariable in breast cancer and fuel widespread CpG island hypermethylation. Nat. Commun. 10. Van Den Broek, T., Borghans, J.A.M., and Van Wijk, F. (2018). The full spectrum of human 199 naive T cells. Nat. Rev. Immunol. 18, 363–373. Brookes, E., De Santiago, I., Hebenstreit, D., Morris, K.J., Carroll, T., Xie, S.Q., Stock, J.K., Heidemann, M., Eick, D., Nozaki, N., et al. (2012). Polycomb associates genome-wide with a specific RNA polymerase II variant, and regulates metabolic genes in ESCs. Cell Stem Cell 10, 157–170. Bröske, A.M., Vockentanz, L., Kharazi, S., Huska, M.R., Mancini, E., Scheller, M., Kuhl, C., Enns, A., Prinz, M., Jaenisch, R., et al. (2009). DNA methylation protects hematopoietic stem cell multipotency from myeloerythroid restriction. Nat. Genet. 41, 1207–1215. Brower-Toland, B., Riddle, N.C., Jiang, H., Huisinga, K.L., and Elgin, S.C.R. (2009). Multiple SET methyltransferases are required to maintain normal heterochromatin domains in the genome of Drosophila melanogaster. Genetics 181, 1303–1319. Brownell, J.E., and Allis, C.D. (1995). An activity gel assay detects a single, catalytically active histone acetyltransferase subunit in Tetrahymena macronuclei. Proc. Natl. Acad. Sci., 92, 6364–6368. Brownell, J.E., Zhou, J., Ranalli, T., Kobayashi, R., Edmondson, D.G., Roth, S.Y., and Allis, C.D. (1996). Tetrahymena histone acetyltransferase A: A homolog to yeast Gcn5p linking histone acetylation to gene activation. Cell 84, 843–851. Buenrostro, J.D., Giresi, P.G., Zaba, L.C., Chang, H.Y., and Greenleaf, W.J. (2013). Transposition of native chromatin for fast and sensitive epigenomic profiling of open chromatin, DNA-binding proteins and nucleosome position. Nat. Methods 10, 1213–1218. Buenrostro, J.D., Corces, M.R., Lareau, C.A., Wu, B., Schep, A.N., Aryee, M.J., Majeti, R., Chang, H.Y., and Greenleaf, W.J. (2018). Integrated Single-Cell Analysis Maps the Continuous Regulatory Landscape of Human Hematopoietic Differentiation. Cell 173, 1535-1548.e16. 200 Butterfield, Y.S., Kreitzman, M., Thiessen, N., Corbett, R.D., Li, Y., Pang, J., Ma, Y.P., Jones, S.J.M., and Birol, İ. (2014). JAGuaR: Junction Alignments to Genome for RNA-Seq Reads. PLoS One 9, e102398. Cano, A., and Pstana, A. (1979). Purification and Properties of a Histone Acetyltransferase from Artemia salina, Highly Efficient with H1 Histone. Eur. J. Biochem., 97, 65–72 Champagne, N., Pelletier, N., and Yang, X.J. (2001). The monocytic leukemia zinc finger protein MOZ is a histone acetyltransferase. Oncogene 20, 404–409. Chan, A.P.Y., and Gurdon, J.B. (1996). Nuclear transplantation from stably transfected cultured cells of Xenopus. Int. J. Dev. Biol., 2, 441–451 Chen, L., Wei, T., Si, X., Wang, Q., Li, Y., Leng, Y., Deng, A., Chen, J., Wang, G., Zhu, S., et al. (2013a). Lysine acetyltransferase GCN5 potentiates the growth of non-small cell lung cancer via promotion of E2F1, cyclin D1, and cyclin E1 expression. J. Biol. Chem. 288, 14510–14521. Chen, L., Deshpande, A.J., Banka, D., Bernt, K.M., Dias, S., Buske, C., Olhava, E.J., Daigle, S.R., Richon, V.M., Pollock, R.M., et al. (2013b). Abrogation of MLL-AF10 and CALM-AF10-mediated transformation through genetic inactivation or pharmacological inhibition of the H3K79 methyltransferase Dot1l. Leukemia 27, 813–822. Chen, M.F., Lu, M.S., Lin, P.Y., Chen, P.T., Chen, W.C., and Lee, K. Der (2012). The role of DNA methyltransferase 3b in esophageal squamous cell carcinoma. Cancer, 16, 4074–4089. Chen, M.W., Hua, K.T., Kao, H.J., Chi, C.C., Wei, L.H., Johansson, G., Shiah, S.G., Chen, P.S., Jeng, Y.M., Cheng, T.Y., et al. (2010). H3K9 histone methyltransferase G9a promotes lung cancer invasion and metastasis by silencing the cell adhesion molecule Ep-CAM. Cancer Res. 70, 7830–7840. Chen, X., Yu, Y., Yu, C., D’Alessio, A.C., Xiong, Y., Nakayama, K., Nakagawa, M., Fan, H.-Y., 201 Lv, L., and Nakagawa, T. (2014). CRL4VprBP E3 Ligase Promotes Monoubiquitylation and Chromatin Binding of TET Dioxygenases. Mol. Cell. 57, 247–260. Chen, Y., Breeze, C.E., Zhen, S., Beck, S., and Teschendorff, A.E. (2016). Tissue-independent and tissue-specific patterns of DNA methylation alteration in cancer. Epigenetics Chromatin 9, 10. Chowdhury, R., Yeoh, K.K., Tian, Y.M., Hillringhaus, L., Bagg, E.A., Rose, N.R., Leung, I.K.H., Li, X.S., Woon, E.C.Y., Yang, M., et al. (2011). The oncometabolite 2-hydroxyglutarate inhibits histone lysine demethylases. EMBO Rep. 12, 463–469. Chuang, L.S.H., Ian, H.I., Koh, T.W., Ng, H.H., Xu, G., and Li, B.F.L. (1997). Human DNA-(cytosine-5) methyltransferase-PCNA complex as a target for p21(WAF1). Science. 277, 1996–2000. Ciau-Uitz, A., Wang, L., Patient, R., and Liu, F. (2013). ETS transcription factors in hematopoietic stem cell development. Blood Cells, Mol. Dis. 51, 248–255. Clark, S.J., Smallwood, S.A., Lee, H.J., Krueger, F., Reik, W., and Kelsey, G. (2017). Genome-wide base-resolution mapping of DNA methylation in single cells using single-cell bisulfite sequencing (scBS-seq). Nat. Protoc. 12, 534–547. Cloos, P.A.C., Christensen, J., Agger, K., and Helin, K. (2008). Erasing the methyl mark: Histone demethylases at the center of cellular differentiation and disease. Genes Dev. 22, 1115–1140. Collins, R.E., Tachibana, M., Tamaru, H., Smith, K.M., Jia, D., Zhang, X., Selker, E.U., Shinkai, Y., and Cheng, X. (2005). In vitro and in vivo analyses of a Phe/Tyr switch controlling product specificity of histone lysine methyltransferases. J. Biol. Chem. 280, 5563–5570. Conneally, E., Cashman, J., Petzer, A., and Eaves, C. (1997). Expansion in vitro of 202 transplantable human cord blood stem cells demonstrated using a quantitative assay of their lympho-myeloid repopulating activity in nonobese diabetic-scid/scid mice. Proc. Natl. Acad. Sci. U. S. A. 94, 9836–9841. Corces, M.R., Buenrostro, J.D., Wu, B., Greenside, P.G., Chan, S.M., Koenig, J.L., Snyder, M.P., Pritchard, J.K., Kundaje, A., Greenleaf, W.J., et al. (2016). Lineage-specific and single-cell chromatin accessibility charts human hematopoiesis and leukemia evolution. Nat. Genet. 48, 1193–1203. Costa, Y., Ding, J., Theunissen, T.W., Faiola, F., Hore, T.A., Shliaha, P. V., Fidalgo, M., Saunders, A., Lawrence, M., Dietmann, S., et al. (2013). NANOG-dependent function of TET1 and TET2 in establishment of pluripotency. Nature 495, 370–374. Crawford, D.J., Liu, M.Y., Nabel, C.S., Cao, X.J., Garcia, B.A., and Kohli, R.M. (2016). Tet2 Catalyzes Stepwise 5-Methylcytosine Oxidation by an Iterative and de novo Mechanism. J. Am. Chem. Soc. 138, 730–733. Creighton, H., and Waddington, C.H. (2006). The Strategy of the Genes. AIBS Bull. Cui, K., Zang, C., Roh, T.Y., Schones, D.E., Childs, R.W., Peng, W., and Zhao, K. (2009). Chromatin Signatures in Multipotent Human Hematopoietic Stem Cells Indicate the Fate of Bivalent Genes during Differentiation. Cell Stem Cell 4, 80–93. Dalgliesh, G.L., Furge, K., Greenman, C., Chen, L., Bignell, G., Butler, A., Davies, H., Edkins, S., Hardy, C., Latimer, C., et al. (2010). Systematic sequencing of renal carcinoma reveals inactivation of histone modifying genes. Nature 463, 360–363. Dang, L., White, D.W., Gross, S., Bennett, B.D., Bittinger, M.A., Driggers, E.M., Fantin, V.R., Jang, H.G., Jin, S., Keenan, M.C., et al. (2009). Cancer-associated IDH1 mutations produce 2-hydroxyglutarate. Nature, 462, 739-744. 203 Dang, L., Jin, S., and Su, S.M. (2010). IDH mutations in glioma and acute myeloid leukemia. Trends Mol. Med. 16, 387–397. Deaton, A.M., and Bird, A. (2011). CpG islands and the regulation of transcription. Genes Dev. 25, 1010–1022. Deaton, A.M., Webb, S., Kerr, A.R.W., Illingworth, R.S., Guy, J., Andrews, R., and Bird, A. (2011). Cell type-specific DNA methylation at intragenic CpG islands in the immune system. Genome Res. 21, 1074–1086. Dillon, S.C., Zhang, X., Trievel, R.C., and Cheng, X. (2005). The SET-domain protein superfamily: Protein lysine methyltransferases. Genome Biol. 6. DiNardo, C.D., Propert, K.J., Loren, A.W., Paietta, E., Sun, Z., Levine, R.L., Straley, K.S., Yen, K., Patel, J.P., Agresta, S., et al. (2013). Serum 2-hydroxyglutarate levels predict isocitrate dehydrogenase mutations and clinical outcome in acute myeloid leukemia. Blood 121, 4917–4924. Ding, X., Pan, H., Li, J., Zhong, Q., Chen, X., Dry, S.M., and Wang, C.Y. (2013). Epigenetic activation of AP1 promotes squamous cell carcinoma metastasis. Sci. Signal. 6. Dolnik, A., Engelmann, J.C., Scharfenberger-Schmeer, M., Mauch, J., Kelkenberg-Schade, S., Haldemann, B., Fries, T., Krönke, J., Kühn, M.W.M., Paschka, P., et al. (2012). Commonly altered genomic regions in acute myeloid leukemia are enriched for somatic mutations involved in chromatin remodeling and splicing. Blood 120. Dong, K.B., Maksakova, I.A., Mohn, F., Leung, D., Appanah, R., Lee, S., Yang, H.W., Lam, L.L., Mager, D.L., Schübeler, D., et al. (2008). DNA methylation in ES cells requires the lysine methyltransferase G9a but not its catalytic activity. EMBO J. 27, 2691–2701. Dorighi, K.M., Swigut, T., Henriques, T., Bhanu, N. V., Scruggs, B.S., Nady, N., Still, C.D., 204 Garcia, B.A., Adelman, K., and Wysocka, J. (2017). Mll3 and Mll4 Facilitate Enhancer RNA Synthesis and Transcription from Promoters Independently of H3K4 Monomethylation. Mol. Cell 66, 568-576.e4. Du, J., Johnson, L.M., Jacobsen, S.E., and Patel, D.J. (2015). DNA methylation pathways and their crosstalk with histone methylation. Nat. Rev. Mol. Cell Biol. 16, 519–532. Dubois, F., Caby, S., Oger, F., Cosseau, C., Capron, M., Grunau, C., Dissous, C., and Pierce, R.J. (2009). Histone deacetylase inhibitors induce apoptosis, histone hyperacetylation and up-regulation of gene transcription in Schistosoma mansoni. Mol. Biochem. Parasitol. 168, 7–15. Duncan, E.M., Chitsazan, A.D., Seidel, C.W., and Alvarado, A.S. (2015). Set1 and MLL1/2 target distinct sets of functionally different genomic loci in vivo. Cell Rep. 13, 2741–2755. Dunham, I., Kundaje, A., Aldred, S.F., Collins, P.J., Davis, C.A., Doyle, F., Epstein, C.B., Frietze, S., Harrow, J., Kaul, R., et al. (2012). An integrated encyclopedia of DNA elements in the human genome. Nature 489, 57–74. Dyda, F., Klein, D.C., and Hickman, A.B. (2000). GCN5-related N-acetyltransferases: A structural overview. Annu. Rev. Biophys. Biomol. Struct. 29, 81–103. Eaves, C.J. (2015). Hematopoietic stem cells: concepts, definitions, and the new reality. Blood 125, 2605 LP – 2613. Ehrlich, M., Gama-Sosa, M.A., Huang, L.H., Midgett, R.M., Kuo, K.C., Mccune, R.A., and Gehrke, C. (1982). Amount and distribution of 5-methylcytosine in human DNA from different types of tissues or cells. Nucleic Acids Res. 10, 2709–2721. Encode Consortium, Carolina, N., and Hill, C. (2013). For Junk DNA. Nature 489, 57–74. Epsztejn-Litman, S., Feldman, N., Abu-Remaileh, M., Shufaro, Y., Gerson, A., Ueda, J., Deplus, R., Fuks, F., Shinkai, Y., Cedar, H., et al. (2008). De novo DNA methylation promoted by G9a 205 prevents reprogramming of embryonically silenced genes. Nat. Struct. Mol. Biol. 15, 1176–1183. Ernst, J., and Kellis, M. (2012). ChromHMM: Automating chromatin-state discovery and characterization. Nat. Methods 9, 215–216. Ernst, J., and Kellis, M. (2017). Chromatin-state discovery and genome annotation with ChromHMM. Nat. Protoc. 12, 2478–2492. Ernst, T., Chase, A.J., Score, J., Hidalgo-Curtis, C.E., Bryant, C., Jones, A. V., Waghorn, K., Zoi, K., Ross, F.M., Reiter, A., et al. (2010). Inactivating mutations of the histone methyltransferase gene EZH2 in myeloid disorders. Nat. Genet. 42, 722–726. Farlik, M., Halbritter, F., Müller, F., Choudry, F.A., Ebert, P., Klughammer, J., Farrow, S., Santoro, A., Ciaurro, V., Mathur, A., et al. (2016). DNA Methylation Dynamics of Human Hematopoietic Stem Cell Differentiation. Cell Stem Cell 19, 808–822. Figueroa, M.E., Abdel-Wahab, O., Lu, C., Ward, P.S., Patel, J., Shih, A., Li, Y., Bhagwat, N., Vasanthakumar, A., Fernandez, H.F., et al. (2010). Leukemic IDH1 and IDH2 Mutations Result in a Hypermethylation Phenotype, Disrupt TET2 Function, and Impair Hematopoietic Differentiation. Cancer Cell 18, 553–567. Fiskus, W., Wang, Y., Sreekumar, A., Buckley, K.M., Shi, H., Jillella, A., Ustun, C., Rao, R., Fernandez, P., Chen, J., et al. (2009). Combined epigenetic therapy with the histone methyltransferase EZH2 inhibitor 3-deazaneplanocin A and the histone deacetylase inhibitor panobinostat against human AML cells. Blood 114, 2733–2743. Forbes, S.A., Bindal, N., Bamford, S., Cole, C., Kok, C.Y., Beare, D., Jia, M., Shepherd, R., Leung, K., Menzies, A., et al. (2011). COSMIC: Mining complete cancer genomes in the catalogue of somatic mutations in cancer. Nucleic Acids Res. 39. Fraga, M.F., Ballestar, E., Montoya, G., Taysavang, P., Wade, P.A., and Esteller, M. (2003). The 206 affinity of different MBD proteins for a specific methylated locus depends on their intrinsic binding properties. Nucleic Acids Res. 31, 1765–1774. Fraley, C., and Raftery, A.E. (1999). MCLUST: Software for Model-Based Cluster Analysis. J. Classif. 16, 297–306. Fraley, C., and Raftery, A.E. (2002). Model-based clustering, discriminant analysis, and density estimation. J. Am. Stat. Assoc, 97, 611–631. Freitag, M., and Selker, E.U. (2005). Controlling DNA methylation: Many roads to one modification. Curr. Opin. Genet. Dev. 15, 191–199. Frelin, C., Herrington, R., Janmohamed, S., Barbara, M., Tran, G., Paige, C.J., Benveniste, P., Zuñiga-Pflücker, J.C., Souabni, A., Busslinger, M., et al. (2013). GATA-3 regulates the self-renewal of long-term hematopoietic stem cells. Nat. Immunol. 14, 1037–1044. Fu, S., Wang, Q., Moore, J.E., Purcaro, M.J., Pratt, H.E., Fan, K., Gu, C., Jiang, C., Zhu, R., Kundaje, A., et al. (2018). Differential analysis of chromatin accessibility and histone modifications for predicting mouse developmental enhancers. Nucleic Acids Res. 46, 11184–11201. Ben Gacem, R., Hachana, M., Ziadi, S., Ben Abdelkarim, S., Hidar, S., and Trimeche, M. (2012). Clinicopathologic significance of DNA methyltransferase 1, 3a, and 3b overexpression in Tunisian breast cancers. Hum. Pathol. 43, 1731–1738. Gaidzik, V.I., Paschka, P., Späth, D., Habdank, M., Köhne, C.H., Germing, U., Von Lilienfeld-Toal, M., Held, G., Horst, H.A., Haase, D., et al. (2012). TET2 mutations in Acute Myeloid Leukemia (AML): Results from a comprehensive genetic and clinical analysis of the AML study group. J. Clin. Oncol. 30, 1350–1357. Gall, J.G. (2016). DNA replication and beyond. Nat. Rev. Mol. Cell Biol. 17, 464. 207 Gall, J.G., and Murphy, C. (1998). Assembly of lampbrush chromosomes from sperm chromatin. Mol. Biol. Cell 9, 733–747. Gama-sosa, M.A., Slagel, V.A., Trewyn, R.W., Oxenhandler, R., Kuo, K.C., Gehrke, C.W., and Ehrlich, M. (1983). The 5-methylcytosine content of DNA from human tumors. Nucleic Acids Res. 11, 6883–6894. Ganji, G., Gilbert, S.A., Diaz, E., Chen, S.B., Pappalardi, M.B., Jiang, Y., Allen, K.E., Kruger, R.G., Creasy, C.L., Tummino, P.J., et al. (2012). Mutation of A677 in histone methyltransferase EZH2 in human B-cell lymphoma promotes hypertrimethylation of histone H3 on lysine 27 (H3K27). Proc. Natl. Acad. Sci, 109, 2989–2994. Garcia-Ramirez, M., Rocchini, C., and Ausio, J. (1995). Modulation of chromatin folding by histone acetylation. J. Biol. Chem. 270, 17923–17928. Geltink, R.K., and Grosveld, G. (2013). The ETS Transcription Factor ETV7 Exhausts Hematopoietic Stem Cells By Enhancing The Cell Cycle Entry and Cell Proliferation. Blood 122, 733 LP – 733. Genovese, G., Kähler, A.K., Handsaker, R.E., Lindberg, J., Rose, S.A., Bakhoum, S.F., Chambert, K., Mick, E., Neale, B.M., Fromer, M., et al. (2014). Clonal hematopoiesis and blood-cancer risk inferred from blood DNA sequence. N. Engl. J. Med. 371, 2477–2487. Ghiam, A.F., Cairns, R.A., Thoms, J., Dal Pra, A., Ahmed, O., Meng, A., Mak, T.W., and Bristow, R.G. (2012). IDH mutation status in prostate cancer. Oncogene 31, 3826. Gilmour, D.S., and Lis, J.T. (1984). Detecting protein-DNA interactions in vivo: Distribution of RNA polymerase on specific bacterial genes. Proc. Natl. Acad. Sci. U. S. A. 81, 4275–4279. Ginder, G.D., and Williams, D.C. (2018). Readers of DNA methylation, the MBD family as potential therapeutic targets. Pharmacol. Ther. 184, 98–111. 208 Globisch, D., Münzel, M., Müller, M., Michalakis, S., Wagner, M., Koch, S., Brückl, T., Biel, M., and Carell, T. (2010). Tissue Distribution of 5-Hydroxymethylcytosine and Search for Active Demethylation Intermediates. PLoS One 5, e15367. Gnanapragasam, M.N., Scarsdale, J.N., Amaya, M.L., Webb, H.D., Desai, M.A., Walavalkar, N.M., Wang, S.Z., Zu Zhu, S., Ginder, G.D., and Williams, D.C. (2011). p66 -MBD2 coiled-coil interaction and recruitment of Mi-2 are critical for globin gene silencing by the MBD2-NuRD complex. Proc. Natl. Acad. Sci, 108, 7487-7492. Goldrath, A.W., Bhattacharya, D., Weissman, I.L., Mathis, D., Benoist, C., and Luckey, C.J. (2006). Memory T and memory B cells share a transcriptional program of self-renewal with long-term hematopoietic stem cells. Proc. Natl. Acad. Sci, 103, 3304–3309. Goll, M.G., Kirpekar, F., Maggert, K.A., Yoder, J.A., Hsieh, C.L., Zhang, X., Golic, K.G., Jacobsen, S.E., and Bestor, T.H. (2006). Methylation of tRNAAsp by the DNA methyltransferase homolog Dnmt2. Science, 311, 395–398. Graff, J.R., Herman, J.G., Lapidus, R.G., Chopra, H., Xu, R., Pitha, P.M., Davidson, N.E., Baylin, S.B., Jarrard, D.F., and Isaacs, W.B. (1995). E-Cadherin Expression Is Silenced by DNA Hypermethylation in Human Breast and Prostate Carcinomas. Cancer Res, 55, 5195–5199. Grembecka, J., He, S., Shi, A., Purohit, T., Muntean, A.G., Sorenson, R.J., Showalter, H.D., Murai, M.J., Belcher, A.M., Hartley, T., et al. (2012). Menin-MLL inhibitors reverse oncogenic activity of MLL fusion proteins in leukemia. Nat. Chem. Biol. 8, 277–284. Grosselin, K., Durand, A., Marsolier, J., Poitou, A., Marangoni, E., Nemati, F., Dahmani, A., Lameiras, S., Reyal, F., Frenoy, O., et al. (2019). High-throughput single-cell ChIP-seq identifies heterogeneity of chromatin states in breast cancer. Nat. Genet. 51, 1060–1066. Grozinger, C.M., Hassig, C.A., and Schreiber, S.L. (1999). Three proteins define a class of 209 human histone deacetylases related to yeast Hda1p. Proc. Natl. Acad. Sci. U. S. A, 96, 4868–4873. Grunstein, M. (1997). Histone acetylation in chromatin structure and transcription. Nature 389, 349–352. Gu, P., Xu, X., Le Menuet, D., Chung, A.C.K., and Cooney, A.J. (2011). Differential recruitment of methyl CpG-binding domain factors and DNA methyltransferases by the orphan receptor germ cell nuclear factor initiates the repression and silencing of Oct4. Stem Cells 29, 1041–1051. Gurdon, J.B., and Laskey, R.A. (1970). The transplantation of nuclei from single cultured cells into enucleate frogs’ eggs. J. Embryol. Exp. Morphol. 24, 227–248. Gurdon, J.B., Elsdale, T.R., and Fischberg, M. (1958). Sexually mature individuals of Xenopus laevis from the transplantation of single somatic nuclei. Nature 182, 64–65. Van Haaften, G., Dalgliesh, G.L., Davies, H., Chen, L., Bignell, G., Greenman, C., Edkins, S., Hardy, C., O’Meara, S., Teague, J., et al. (2009). Somatic mutations of the histone H3K27 demethylase gene UTX in human cancer. Nat. Genet. 41, 521–523. Haas, S., Trumpp, A., and Milsom, M.D. (2018). Causes and Consequences of Hematopoietic Stem Cell Heterogeneity. Cell Stem Cell 22, 627–638. Haberland, M., Montgomery, R.L., and Olson, E.N. (2009). The many roles of histone deacetylases in development and physiology: Implications for disease and therapy. Nat. Rev. Genet. 10, 32–42. Hahn, M.A., Wu, X., Li, A.X., Hahn, T., and Pfeifer, G.P. (2011). Relationship between Gene Body DNA Methylation and Intragenic H3K9me3 and H3K36me3 Chromatin Marks. PLoS One 6, e18844. Hahne, F., LeMeur, N., Brinkman, R.R., Ellis, B., Haaland, P., Sarkar, D., Spidlen, J., Strain, E., 210 and Gentleman, R. (2009). flowCore: a Bioconductor package for high throughput flow cytometry. BMC Bioinformatics 10, 106. Hameed, U.F.S., Lim, J., Zhang, Q., Wasik, M.A., Yang, D., and Swaminathan, K. (2014). Transcriptional repressor domain of MBD1 is intrinsically disordered and interacts with its binding partners in a selective manner. Sci. Rep. 4. Hartmann, C., Hentschel, B., Wick, W., Capper, D., Felsberg, J., Simon, M., Westphal, M., Schackert, G., Meyermann, R., Pietsch, T., et al. (2010). Patients with IDH1 wild type anaplastic astrocytomas exhibit worse prognosis than IDH1-mutated glioblastomas, and IDH1 mutation status accounts for the unfavorable prognostic effect of higher age: Implications for classification of gliomas. Acta Neuropathol. 120, 707–718. Hawkins, R.D., Hon, G.C., Lee, L.K., Ngo, Q., Lister, R., Pelizzola, M., Edsall, L.E., Kuan, S., Luu, Y., Klugman, S., et al. (2010). Distinct epigenomic landscapes of pluripotent and lineage-committed human cells. Cell Stem Cell 6, 479–491. Hayatsu, H., Wataya, Y., Kai, K., and Iida, S. (1970). Reaction of Sodium Bisulfite with Uracil, Cytosine, and Their Derivatives. Biochemistry, 9, 2858–2865. Hayette, S., Thomas, X., Jallades, L., Chabane, K., Charlot, C., Tigaud, I., Gazzo, S., Morisset, S., Cornillet-Lefebvre, P., Plesa, A., et al. (2012). High DNA Methyltransferase DNMT3B Levels: A Poor Prognostic Marker in Acute Myeloid Leukemia. PLoS One 7. Hebbes, T.R., Thorne, A.W., and Crane-Robinson, C. (1988). A direct link between core histone acetylation and transcriptionally active chromatin. EMBO J. 7, 1395–1402. Heintzman, N.D., Stuart, R.K., Hon, G., Fu, Y., Ching, C.W., Hawkins, R.D., Barrera, L.O., Van Calcar, S., Qu, C., Ching, K.A., et al. (2007). Distinct and predictive chromatin signatures of transcriptional promoters and enhancers in the human genome. Nat. Genet. 39, 311–318. 211 Heinz, S., Benner, C., Spann, N., Bertolino, E., Lin, Y.C., Laslo, P., Cheng, J.X., Murre, C., Singh, H., and Glass, C.K. (2010). Simple Combinations of Lineage-Determining Transcription Factors Prime cis-Regulatory Elements Required for Macrophage and B Cell Identities. Mol. Cell 38, 576–589. Henikoff, S. (2008). Nucleosome destabilization in the epigenetic regulation of gene expression. Nat. Rev. Genet. 9, 15–26. Hergeth, S.P., and Schneider, R. (2015). The H1 linker histones: multifunctional proteins beyond the nucleosomal core particle. EMBO Rep, 16, 1439–1453. Herlofsen, S.R., Bryne, J.C., Høiby, T., Wang, L., Issner, R., Zhang, X., Coyne, M.J., Boyle, P., Gu, H., Meza-Zepeda, L.A., et al. (2013). Genome-wide map of quantified epigenetic changes during in vitro chondrogenic differentiation of primary human mesenchymal stem cells. BMC Genomics 14, 105. Hilfiker, A., Hilfiker-Kleiner, D., Pannuti, A., and Lucchesi, J.C. (1997). mof, a putative acetyl transferase gene related to the Tip60 and MOZ human genes and to the SAS genes of yeast, is required for dosage compensation in Drosophila. EMBO J. 16, 2054–2060. Hnisz, D., Abraham, B.J., Lee, T.I., Lau, A., Saint-André, V., Sigova, A.A., Hoke, H.A., and Young, R.A. (2013). XSuper-enhancers in the control of cell identity and disease. Cell 155, 934. Hnisz, D., Shrinivas, K., Young, R.A., Chakraborty, A.K., and Sharp, P.A. (2017). A Phase Separation Model for Transcriptional Control. Cell 169, 13–23. Hodges, E., Molaro, A., Dos Santos, C.O., Thekkat, P., Song, Q., Uren, P.J., Park, J., Butler, J., Rafii, S., McCombie, W.R., et al. (2011). Directional DNA methylation changes and complex intermediate states accompany lineage specificity in the adult hematopoietic compartment. Mol. Cell 44, 17–28. 212 Hogart, A., Lichtenberg, J., Ajay, S.S., Anderson, S., Margulies, E.H., and Bodine, D.M. (2012). Genome-wide DNA methylation profiles in hematopoietic stem and progenitor cells reveal overrepresentation of ETS transcription factor binding sites. Genome Res. 22, 1407–1418. Højfeldt, J.W., Agger, K., and Helin, K. (2013). Histone lysine demethylases as targets for anticancer therapy. Nat. Rev. Drug Discov. 12, 917–930. Holz-Schietinger, C., Matje, D.M., Harrison, M.F., and Reich, N.O. (2011). Oligomerization of DNMT3A controls the mechanism of de novo DNA methylation. J. Biol. Chem. 286, 41479–41488. Holz-Schietinger, C., Matje, D.M., and Reich, N.O. (2012). Mutations in DNA methyltransferase (DNMT3A) observed in acute myeloid leukemia patients disrupt processive methylation. J. Biol. Chem. 287, 30941–30951. Hong, S.H., Rampalli, S., Lee, J.B., McNicol, J., Collins, T., Draper, J.S., and Bhatia, M. (2011). Cell fate potential of human pluripotent stem cells is encoded by histone modifications. Cell Stem Cell 9, 24–36. Hoppe, P.C., and Illmensee, K. (1982). Full-term development after transplantation of parthenogenetic embryonic nuclei into fertilized mouse eggs. Proc. Natl. Acad. Sci. U. S. A. 79, 1912–1916. Hu, L., Li, Z., Cheng, J., Rao, Q., Gong, W., Liu, M., Shi, Y.G., Zhu, J., Wang, P., and Xu, Y. (2013). Crystal Structure of TET2-DNA Complex: Insight into TET-Mediated 5mC Oxidation. Cell 155, 1545–1555. Hu, M., Krause, D., Greaves, M., Sharkis, S., Dexter, M., Heyworth, C., and Enver, T. (1997). Multilineage gene expression precedes commitment in the hemopoietic system. Genes Dev. 11, 774–785. 213 Hudlebusch, H.R., Santoni-Rugiu, E., Simon, R., Ralfkiær, E., Rossing, H.H., Johansen, J.V., Jørgensen, M., Sauter, G., and Helin, K. (2011). The histone methyltransferase and putative oncoprotein MMSET is overexpressed in a large variety of human tumors. Clin. Cancer Res. 17, 2919–2933. Hui, T., Cao, Q., Wegrzyn-Woltosz, J., O’Neill, K., Hammond, C.A., Knapp, D.J.H.F., Laks, E., Moksa, M., Aparicio, S., Eaves, C.J., et al. (2018). High-Resolution Single-Cell DNA Methylation Measurements Reveal Epigenetically Distinct Hematopoietic Stem Cell Subpopulations. Stem Cell Reports 11, 578–592. Hyland, P.L., McDade, S.S., McCloskey, R., Dickson, G.J., Arthur, K., McCance, D.J., and Patel, D. (2011). Evidence for Alteration of EZH2, BMI1, and KDM6A and Epigenetic Reprogramming in Human Papillomavirus Type 16 E6/E7-Expressing Keratinocytes. J. Virol. Hyun, K., Jeon, J., Park, K., and Kim, J. (2017). Writing, erasing and reading histone lysine methylations. Exp. Mol. Med. 49. Illmensee, K., and Hoppe, P.C. (1981). Nuclear transplantation in mus musculus: Developmental potential of nuclei from preimplantation embryos. Cell 23, 9–18. Inoue, A., and Fujimoto, D. (1969). Enzymatic deacetylation of histone. Biochem. Biophys. Res. Commun, 36, 146–50. Ito, T., Teo, Y.V., Evans, S.A., Neretti, N., and Sedivy, J.M. (2018). Regulation of Cellular Senescence by Polycomb Chromatin Modifiers through Distinct DNA Damage- and Histone Methylation-Dependent Pathways. Cell Rep. 22, 3480–3492. Iyer, L.M., Tahiliani, M., Rao, A., and Aravind, L. (2009). Prediction of novel families of enzymes involved in oxidative and other complex modifications of bases in nucleic acids. Cell Cycle 8, 1698–1710. 214 Jackson, V. (1978). Studies on histone organization in the nucleosome using formaldehyde as a reversible cross-linking agent. Cell 15, 945–954. Jaenisch, R., Young, R.A., Kooistra, T., Lodato, M.A., Carey, B.W., Frampton, G.M., Creyghton, M.P., Cheng, A.W., Steine, E.J., Sharp, P.A., et al. (2010). Histone H3K27ac separates active from poised enhancers and predicts developmental state. Proc. Natl. Acad. Sci, 107, 21931–21936. Jankowska, A.M., Makishima, H., Tiu, R. V., Szpurka, H., Huang, Y., Traina, F., Visconte, V., Sugimoto, Y., Prince, C., O’Keefe, C., et al. (2011). Mutational spectrum analysis of chronic myelomonocytic leukemia includes genes associated with epigenetic regulation: UTX,,EZH2,and DNMT3A. Blood 118, 3932–3941. Jenuwein, T., and Allis, C.D. (2001). Translating the histone code. Science. 293, 1074–1080. Jermann, P., Hoerner, L., Burger, L., and Schübeler, D. (2014). Short sequences can efficiently recruit histone H3 lysine 27 trimethylation in the absence of enhancer activity and DNA methylation. Proc. Natl. Acad. Sci. U. S. A. 111. Ji, H., Ehrlich, L.I.R., Seita, J., Murakami, P., Doi, A., Lindau, P., Lee, H., Aryee, M.J., Irizarry, R.A., Kim, K., et al. (2010). Comprehensive methylome map of lineage commitment from haematopoietic progenitors. Nature, 467, 338–342. Jia, D., Jurkowska, R.Z., Zhang, X., Jeltsch, A., and Cheng, X. (2007). Structure of Dnmt3a bound to Dnmt3L suggests a model for de novo DNA methylation. Nature 449, 248–251. Jiang, J., Lu, J., Lu, D., Liang, Z., Li, L., Ouyang, S., Kong, X., Jiang, H., Shen, B., and Luo, C. (2012). Investigation of the Acetylation Mechanism by GCN5 Histone Acetyltransferase. PLoS One 7, e36660. Jiang, T., Zhou, X., Taghizadeh, K., Dong, M., and Dedon, P.C. (2007). N-formylation of lysine 215 in histone proteins as a secondary modification arising from oxidative DNA damage. Proc. Natl. Acad. Sci, 104, 60–65. Jin, Q., Yu, L.R., Wang, L., Zhang, Z., Kasper, L.H., Lee, J.E., Wang, C., Brindle, P.K., Dent, S.Y.R., and Ge, K. (2011). Distinct roles of GCN5/PCAF-mediated H3K9ac and CBP/p300-mediated H3K18/27ac in nuclear receptor transactivation. EMBO J. 30, 249–262. Jin, W., Tang, Q., Wan, M., Cui, K., Zhang, Y., Ren, G., Ni, B., Sklar, J., Przytycka, T.M., Childs, R., et al. (2015). Genome-wide detection of DNase i hypersensitive sites in single cells and FFPE tissue samples. Nature 528, 142–146. Johnson, D.S., Mortazavi, A., Myers, R.M., and Wold, B. (2007). Genome-wide mapping of in vivo protein-DNA interactions. Science (80-. ). 316, 1497–1502. Johnson, K.C., Houseman, E.A., King, J.E., Von Herrmann, K.M., Fadul, C.E., and Christensen, B.C. (2016). 5-Hydroxymethylcytosine localizes to enhancer elements and is associated with survival in glioblastoma patients. Nat. Commun. 7. Jung, K.H., Noh, J.H., Kim, J.K., Eun, J.W., Bae, H.J., Xie, H.J., Chang, Y.G., Kim, M.G., Park, H., Lee, J.Y., et al. (2012). HDAC2 overexpression confers oncogenic potential to human lung cancer cells by deregulating expression of apoptosis and cell cycle proteins. J. Cell. Biochem. 113, 2167–2177. Kamminga, L.M., Bystrykh, L. V., De Boer, A., Houwer, S., Douma, J., Weersing, E., Dontje, B., and De Haan, G. (2006). The Polycomb group gene Ezh2 prevents hematopoietic stem cell exhaustion. Blood 107, 2170–2179. Kanai, Y., Ushijima, S., Nakanishi, Y., Sakamoto, M., and Hirohashi, S. (2003). Mutation of the DNA methyltransferase (DNMT) 1 gene in human colorectal cancers. Cancer Lett. 192, 75–82. Kao, H.Y., Downes, M., Ordentlich, P., and Evans, R.M. (2000). Isolation of a novel histone 216 deacetylase reveals that class I and class II deacetylases promote SMRT-mediated repression. Genes Dev. 14, 55–66. Karantanos, T., Chistofides, A., Barhdan, K., Li, L., and Boussiotis, V.A. (2016). Regulation of T cell differentiation and function by EZH2. Front. Immunol, 7. Karatas, H., Townsend, E.C., Cao, F., Chen, Y., Bernard, D., Liu, L., Lei, M., Dou, Y., and Wang, S. (2013). High-affinity, small-molecule peptidomimetic inhibitors of mll1/wdr5 protein-protein interaction. J. Am. Chem. Soc. 135, 669–682. Karnezis, A.N., Wang, Y., Ramos, P., Hendricks, W.P.D., Oliva, E., D’Angelo, E., Prat, J., Nucci, M.R., Nielsen, T.O., Chow, C., et al. (2016). Dual loss of the SWI/SNF complex ATPases SMARCA4/BRG1 and SMARCA2/BRM is highly sensitive and specific for small cell carcinoma of the ovary, hypercalcaemic type. J. Pathol. 238, 389–400. Kasinath, V., Faini, M., Poepsel, S., Reif, D., Feng, X.A., Stjepanovic, G., Aebersold, R., and Nogales, E. (2018). Structures of human PRC2 with its cofactors AEBP2 and JARID2. Science, 359, 940–944. Kassambara, A., Klein, B., and Moreaux, J. (2009). MMSET is overexpressed in cancers: Link with tumor aggressiveness. Biochem. Biophys. Res. Commun. 379, 840–845. Kastner, P., and Chan, S. (2008). PU.1: A crucial and versatile player in hematopoiesis and leukemia. Int. J. Biochem. Cell Biol. 40, 22–27. Kats, L.M., Reschke, M., Taulli, R., Pozdnyakova, O., Burgess, K., Bhargava, P., Straley, K., Karnik, R., Meissner, A., Small, D., et al. (2014). Proto-oncogenic role of mutant IDH2 in leukemia initiation and maintenance. Cell Stem Cell 14, 329–341. Kawasaki, H., Schiltz, L., Chiu, R., Itakura, K., Taira, K., Nakatani, Y., and Yokoyama, K.K. (2000). ATF-2 has intrinsic histone acetyltransferase activity which is modulated by 217 phosphorylation. Nature 405, 195–200. Kaya-Okur, H.S., Wu, S.J., Codomo, C.A., Pledger, E.S., Bryson, T.D., Henikoff, J.G., Ahmad, K., and Henikoff, S. (2019). CUT&Tag for efficient epigenomic profiling of small samples and single cells. Nat. Commun. 10. Kelly, A.D., Kroeger, H., Yamazaki, J., Taby, R., Neumann, F., Yu, S., Lee, J.T., Patel, B., Li, Y., He, R., et al. (2017). A CpG island methylator phenotype in acute myeloid leukemia independent of IDH mutations and associated with a favorable outcome. Leukemia 31, 2011–2019. Kelly, A.D., Madzo, J., Madireddi, P., Kropf, P., Good, C.R., Jelinek, J., and Issa, J.P.J. (2018). Demethylator phenotypes in acute myeloid leukemia. Leukemia 32, 2178–2188. Khan, S.N., Jankowska, A.M., Mahfouz, R., Dunbar, A.J., Sugimoto, Y., Hosono, N., Hu, Z., Cheriyath, V., Vatolin, S., Przychodzen, B., et al. (2013). Multiple mechanisms deregulate EZH2 and histone H3 lysine 27 epigenetic changes in myeloid malignancies. Leukemia 27, 1301–1309. Khasawneh, M.K., and Abdel-Wahab, O. (2014). Recent discoveries in molecular characterization of acute myeloid leukemia. Curr. Hematol. Malig. Rep. 9, 93–99. Kieslinger, M., Hiechinger, S., Dobreva, G., Consalez, G.G., and Grosschedl, R. (2010). Early B cell factor 2 regulates hematopoietic stem cell homeostasis in a cell-nonautonomous manner. Cell Stem Cell 7, 496–507. Kikuchi, J., Takashina, T., Kinoshita, I., Kikuchi, E., Shimizu, Y., Sakakibara-Konishi, J., Oizumi, S., Marquez, V.E., Nishimura, M., and Dosaka-Akita, H. (2012). Epigenetic therapy with 3-deazaneplanocin A, an inhibitor of the histone methyltransferase EZH2, inhibits growth of non-small cell lung cancer cells. Lung Cancer 78, 138–143. Kile, M.L., Baccarelli, A., Hoffman, E., Tarantini, L., Quamruzzaman, Q., Rahman, M., 218 Mahiuddin, G., Mostofa, G., Hsueh, Y.M., Wright, R.O., et al. (2012). Prenatal arsenic exposure and DNA methylation in maternal and umbilical cord blood leukocytes. Environ. Health Perspect. 120, 1061–1066. Kim, M.S., Kim, Y.R., Yoo, N.J., and Lee, S.H. (2013). Mutational analysis of DNMT3A gene in acute leukemias and common solid cancers. APMIS 121, 85–94. Kishikawa, S., Murata, T., Ugai, H., Yamazaki, T., and Yokoyama, K.K. (2003). Control elements of Dnmt1 gene are regulated in cell-cycle dependent manner. Nucleic Acids Res. Suppl. 307–308. Klose, R.J., and Bird, A.P. (2006). Genomic DNA methylation: The mark and its mediators. Trends Biochem. Sci. 31, 89–97. Knapp, D.J.H.F., Hammond, C.A., Hui, T., Van Loenhout, M.T.J., Wang, F., Aghaeepour, N., Miller, P.H., Moksa, M., Rabu, G.M., Beer, P.A., et al. (2018). Single-cell analysis identifies a CD33 + subset of human cord blood cells with high regenerative potential. Nat. Cell Biol. 20, 710–720. Knapp, D.J.H.F., Hammond, C.A., Wang, F., Aghaeepour, N., Miller, P.H., Beer, P.A., Pellacani, D., VanInsberghe, M., Hansen, C., Bendall, S.C., et al. (2019). A topological view of human CD34+ cell state trajectories from integrated single-cell output and proteomic data. Blood blood-2018-10-878025. Ko, M., Huang, Y., Jankowska, A.M., Pape, U.J., Tahiliani, M., Bandukwala, H.S., An, J., Lamperti, E.D., Koh, K.P., Ganetzky, R., et al. (2010). Impaired hydroxylation of 5-methylcytosine in myeloid cancers with mutant TET2. Nature 468, 839–843. Kolasinska-Zwierz, P., Down, T., Latorre, I., Liu, T., Liu, X.S., and Ahringer, J. (2009). Differential chromatin marking of introns and expressed exons by H3K36me3. Nat. Genet. 41, 219 376–381. Kondo, Y., Shen, L., Ahmed, S., Boumber, Y., Sekido, Y., Haddad, B.R., and Issa, J.P.J. (2008). Downregulation of histone H3 lysine 9 methyltransferase G9a induces centrosome disruption and chromosome instability in cancer cells. PLoS One, 3. Kornberg, R.D., and Thomas, J.O. (1974). Chromatin structure: Oligomers of the histones. Science, 184, 865–868. Kouzarides, T. (2007). Chromatin Modifications and Their Function. Cell 128, 693–705. Kriaucionis, S., and Heintz, N. (2009). The nuclear DNA base 5-hydroxymethylcytosine is present in purkinje neurons and the brain. Science (80-. ). 324, 929–930. Kumamoto, K., Fujita, K., Kurotani, R., Saito, M., Unoki, M., Hagiwara, N., Shiga, H., Bowman, E.D., Yanaihara, N., Okamura, S., et al. (2009). ING2 is upregulated in colon cancer and increases invasion by enhanced MMP13 expression. Int. J. Cancer 125, 1306–1315. Laporte, A.N., Poulin, N.M., Barrott, J.J., Wang, X.Q., Lorzadeh, A., Werff, R.V., Jones, K.B., Underhill, T.M., and Nielsen, T.O. (2017). Death by HDAC inhibition in synovial sarcoma cells. Mol. Cancer Ther. 16. Lara-Astiaso, D., Weiner, A., Lorenzo-Vivas, E., Zaretsky, I., Jaitin, D.A., David, E., Keren-Shaul, H., Mildner, A., Winter, D., Jung, S., et al. (2014). Chromatin state dynamics during blood formation HHS Public Access. Science (80-. ). Laslo, P., Spooner, C.J., Warmflash, A., Lancki, D.W., Lee, H.J., Sciammas, R., Gantner, B.N., Dinner, A.R., and Singh, H. (2006). Multilineage Transcriptional Priming and Determination of Alternate Hematopoietic Cell Fates. Cell, 126, 755–766. Laurenti, E., and Göttgens, B. (2018). From haematopoietic stem cells to complex differentiation landscapes. Nature, 553, 418–426. 220 Lee, K.K., and Workman, J.L. (2007). Histone acetyltransferase complexes: One size doesn’t fit all. Nat. Rev. Mol. Cell Biol. 8, 284–295. Lee, J.H., Park, S.J., and Nakai, K. (2017). Differential landscape of non-CpG methylation in embryonic stem cells and neurons caused by DNMT3s. Sci. Rep. 7. Lee, P.P., Fitzpatrick, D.R., Beard, C., Jessup, H.K., Lehar, S., Makar, K.W., Pérez-Melgosa, M., Sweetser, M.T., Schlissel, M.S., Nguyen, S., et al. (2001). A critical role for Dnmt1 and DNA methylation in T cell development, function, and survival. Immunity 15, 763–774. Lehnertz, B., Ueda, Y., Derijck, A.A.H.A., Braunschweig, U., Perez-Burgos, L., Kubicek, S., Chen, T., Li, E., Jenuwein, T., and Peters, A.H.F.M. (2003). Suv39h-mediated histone H3 lysine 9 methylation directs DNA methylation to major satellite repeats at pericentric heterochromatin. Curr. Biol. 13, 1192–1200. Leipe, D.D., and Landsman, D. (1997). Histone deacetylases, acetoin utilization proteins and acetylpolyamine amidohydrolases are members of an ancient protein superfamily. Nucleic Acids Res. 25, 3693–3697. Lewis, E.B. (2007). A gene complex controlling segmentation in Drosophila. In Genes, Development, and Cancer: The Life and Work of Edward B. Lewis, p. Lewis, J.D., Meehan, R.R., Henzel, W.J., Maurer-Fogy, I., Jeppesen, P., Klein, F., and Bird, A. (1992). Purification, sequence, and cellular localization of a novel chromosomal protein that binds to Methylated DNA. Cell, 69, 905–914 Ley, T.J., Miller, C., Ding, L., Raphael, B.J., Mungall, A.J., Robertson, G., Hoadley, K., Triche, T.J., Laird, P.W., Baty, J.D., et al. (2013). Genomic and epigenomic landscapes of adult de novo acute myeloid leukemia. N. Engl. J. Med. 368, 2059–2074. Li, K., Liu, C., Zhou, B., Bi, L., Huang, H., Lin, T., and Xu, K. (2013). Role of EZH2 in the 221 growth of prostate cancer stem cells isolated from LNCaP cells. Int. J. Mol. Sci. 14, 11981–11993. Li, L., Lorzadeh, A., and Hirst, M. (2014). Regulatory variation: An emerging vantage point for cancer biology. Wiley Interdiscip. Rev. Syst. Biol. Med. 6, 37–59. Li, T., Zheng, Q., An, J., Wu, M., Li, H., Gui, X., Pu, H., and Lu, D. (2016). SET1A cooperates with CUDR to promote liver cancer growth and hepatocyte-like stem cell malignant transformation epigenetically. Mol. Ther. 24, 261–275. Lian, C.G., Xu, Y., Ceol, C., Wu, F., Larson, A., Dresser, K., Xu, W., Tan, L., Hu, Y., Zhan, Q., et al. (2012). Loss of 5-hydroxymethylcytosine is an epigenetic hallmark of Melanoma. Cell 150, 1135–1146. Lieberman-Aiden, E., Van Berkum, N.L., Williams, L., Imakaev, M., Ragoczy, T., Telling, A., Amit, I., Lajoie, B.R., Sabo, P.J., Dorschner, M.O., et al. (2009). Comprehensive mapping of long-range interactions reveals folding principles of the human genome. Science (80-. ). 326, 289–293. Lin, I.H., Chen, Y.F., and Hsu, M.T. (2017). Correlated 5-hydroxymethylcytosine (5hmC) and gene expression profiles underpin gene and organ-specific epigenetic regulation in adult mouse brain and liver. PLoS One 12. Lioznova, A. V., Khamis, A.M., Artemov, A. V., Besedina, E., Ramensky, V., Bajic, V.B., Kulakovskiy, I. V., and Medvedeva, Y.A. (2019). CpG traffic lights are markers of regulatory regions in human genome. BMC Genomics 20, 102. Lister, R., Pelizzola, M., Dowen, R.H., Hawkins, R.D., Hon, G., Tonti-Filippini, J., Nery, J.R., Lee, L., Ye, Z., Ngo, Q.M., et al. (2009). Human DNA methylomes at base resolution show widespread epigenomic differences. Nature 462, 315–322. 222 Liu, X., Yang, S., Zhao, M., Luo, M., Yu, C.W., Chen, C.Y., Tai, R., and Wu, K. (2014). Transcriptional repression by histone deacetylases in plants. Mol. Plant 7, 764–772. Local, A., Huang, H., Albuquerque, C.P., Singh, N., Lee, A.Y., Wang, W., Wang, C., Hsia, J.E., Shiau, A.K., Ge, K., et al. (2018). Identification of H3K4me1-associated proteins at mammalian enhancers. Nat. Genet. 50, 73–82. Löffler-Wirth, H., Kalcher, M., and Binder, H. (2015). OposSOM: R-package for high-dimensional portraying of genome-wide expression landscapes on bioconductor. Bioinformatics 31, 3225–3227. Lohse, B., Helgstrand, C., Kristensen, J.B.L., Leurs, U., Cloos, P.A.C., Kristensen, J.L., and Clausen, R.P. (2013). Posttranslational Modifications of the Histone 3 Tail and Their Impact on the Activity of Histone Lysine Demethylases In Vitro. PLoS One 8, e67653. López-Rodas, G., Georgieva, E.I., Sendra, R., and Loidl, P. (1991). Histone acetylation in Zea mays I: Activities of histone acetyltransferases and histone deacetylases. J. Biol. Chem, 266, 18745–18750. Lorsback, R.B., Moore, J., Mathew, S., Raimondi, S.C., Mukatira, S.T., and Downing, J.R. (2003). TET1, a member of a novel protein family, is fused to MLL in acute myeloid leukemia containing the t(10;11)(q22;23) [3]. Leukemia 17, 637–641. Lorzadeh, A., Bilenky, M., Hammond, C., Knapp, D.J.H.F., Li, L., Miller, P.H., Carles, A., Heravi-Moussavi, A., Gakkhar, S., Moksa, M., et al. (2016). Nucleosome Density ChIP-Seq Identifies Distinct Chromatin Modification Signatures Associated with MNase Accessibility. Cell Rep. 17, 2112–2124. Losman, J.A., and Kaelin, W.G. (2013). What a difference a hydroxyl makes: Mutant IDH, (R)-2-hydroxyglutarate, and cancer. Genes Dev. 27, 836–852. 223 Love, I.M., Sekaric, P., Shi, D., Grossman, S.R., and Androphy, E.J. (2012). The histone acetyltransferase PCAF regulates p21 transcription through stress-induced acetylation of histone H3. Cell Cycle 11, 2458–2466. Loyola, A., Tagami, H., Bonaldi, T., Roche, D., Quivy, J.P., Imhof, A., Nakatani, Y., Dent, S.Y.R., and Almouzni, G. (2009). The HP1α-CAF1-SetDB1-containing complex provides H3K9me1 for Suv39-mediated K9me3 in pericentric heterochromatin. EMBO Rep. 10, 769–775. Lu, C., Ward, P.S., Kapoor, G.S., Rohle, D., Turcan, S., Abdel-Wahab, O., Edwards, C.R., Khanin, R., Figueroa, M.E., Melnick, A., et al. (2012). IDH mutation impairs histone demethylation and results in a block to cell differentiation. Nature 483, 474–478. Lu, P.J., Sundquist, K., Baeckstrom, D., Poulsom, R., Hanby, A., Meier-Ewert, S., Jones, T., Mitchell, M., Pitha-Rowe, P., Freemont, P., et al. (1999). A novel gene (PLU-1) containing highly conserved putative DNA/chromatin binding motifs is specifically up-regulated in breast cancer. J. Biol. Chem. 274, 15633–15645. Luger, K., Mäder, A.W., Richmond, R.K., Sargent, D.F., and Richmond, T.J. (1997). Crystal structure of the nucleosome core particle at 2.8 Å resolution. Nature, 389, 251–260. Lulli, V., Romania, P., Morsilli, O., Gabbianelli, M., Pagliuca, A., Mazzeo, S., Testa, U., Peschle, C., and Marziali, G. (2006). Overexpression of Ets-1 in human hematopoietic progenitor cells blocks erythroid and promotes megakaryocytic differentiation. Cell Death Differ. 13, 1064–1074. Luzzati, V., and Nicolaïeff, A. (1963). The structure of nucleohistones and nucleoprotamines. J. Mol. Biol. 7, 142–163. Lyko, F. (2018). The DNA methyltransferase family: A versatile toolkit for epigenetic regulation. Nat. Rev. Genet. 19, 81–92. 224 Mack, S.C., Pajtler, K.W., Chavez, L., Okonechnikov, K., Bertrand, K.C., Wang, X., Erkek, S., Federation, A., Song, A., Lee, C., et al. (2018). Therapeutic targeting of ependymoma as informed by oncogenic enhancer profiling. Nature, 553, 101–105. Madani Tonekaboni, S.A., Mazrooei, P., Kofia, V., Haibe-Kains, B., and Lupien, M. (2019). Identifying clusters of cis-regulatory elements underpinning TAD structures and lineage-specific regulatory networks. Genome Res. 29, 1733–1743. Maehara, K., and Ohkawa, Y. (2016). Exploration of nucleosome positioning patterns in transcription factor function. Sci. Rep. 6. Maiques-Diaz, A., Chou, F.S., Wunderlich, M., Gómez-López, G., Jacinto, F. V., Rodriguez-Perales, S., Larrayoz, M.J., Calasanz, M.J., Mulloy, J.C., Cigudosa, J.C., et al. (2012). Chromatin modifications induced by the AML1-ETO fusion protein reversibly silence its genomic targets through AML1 and Sp1 binding motifs. Leukemia 26, 1329–1337. Majewski, I.J., Ritchie, M.E., Phipson, B., Corbin, J., Pakusch, M., Ebert, A., Busslinger, M., Koseki, H., Hu, Y., Smyth, G.K., et al. (2010). Opposing roles of polycomb repressive complexes in hematopoietic stem and progenitor cells. Blood 116, 731–739. Makar, K.W., and Wilson, C.B. (2014). DNA Methylation Is a Nonredundant Repressor of the Th2 Effector Program. J. Immunol, 173, 4402–4406. Marcucci, G., Maharry, K., Wu, Y.Z., Radmacher, M.D., Mrózek, K., Margeson, D., Holland, K.B., Whitman, S.P., Becker, H., Schwind, S., et al. (2010). IDH1 and IDH2 gene mutations identify novel molecular subsets within de novo cytogenetically normal acute myeloid leukemia: A cancer and leukemia group B study. J. Clin. Oncol. 28, 2348–2355. Mardis, E.R., Ding, L., Dooling, D.J., Larson, D.E., McLellan, M.D., Chen, K., Koboldt, D.C., Fulton, R.S., Delehaunty, K.D., McGrath, S.D., et al. (2009). Recurring mutations found by 225 sequencing an acute myeloid leukemia genome. N. Engl. J. Med. 361, 1058–1066. Margueron, R., and Reinberg, D. (2011). The Polycomb complex PRC2 and its mark in life. Nature 469, 343–349. Marimuthu, R., Francis, H., Dervish, S., Li, S.C.H., Medbury, H., and Williams, H. (2018). Characterization of human monocyte subsets by whole blood flow cytometry analysis. J. Vis. Exp. 2018. Marks, P.A., Rifkind, R.A., Richon, V.M., Breslow, R., Miller, T., and Kelly, W.K. (2001). Histone deacetylases and cancer: Causes and therapies. Nat. Rev. Cancer 1, 194–202. Marshall, G.M., Gherardi, S., Xu, N., Neiron, Z., Trahair, T., Scarlett, C.J., Chang, D.K., Liu, P.Y., Jankowski, K., Iraci, N., et al. (2010). Transcriptional upregulation of histone deacetylase 2 promotes Myc-induced oncogenic effects. Oncogene 29, 5957–5968. Martin-Perez, D., Piris, M.A., and Sanchez-Beato, M. (2010). Polycomb proteins in hematologic malignancies. Blood, 116, 5465–5475. Matsumura, Y., Nakaki, R., Inagaki, T., Yoshida, A., Kano, Y., Kimura, H., Tanaka, T., Tsutsumi, S., Nakao, M., Doi, T., et al. (2015). H3K4/H3K9me3 Bivalent Chromatin Domains Targeted by Lineage-Specific DNA Methylation Pauses Adipocyte Differentiation. Mol. Cell 60, 584–596. Maunakea, A.K., Nagarajan, R.P., Bilenky, M., Ballinger, T.J., Dsouza, C., Fouse, S.D., Johnson, B.E., Hong, C., Nielsen, C., Zhao, Y., et al. (2010). Conserved role of intragenic DNA methylation in regulating alternative promoters. Nature 466, 253–257. McCabe, M.T., Ott, H.M., Ganji, G., Korenchuk, S., Thompson, C., Van Aller, G.S., Liu, Y., Pietra, A. Della, LaFrance, L. V., Mellinger, M., et al. (2012). EZH2 inhibition as a therapeutic strategy for lymphoma with EZH2-activating mutations. Nature 492, 108–112. 226 McLean, C.Y., Bristor, D., Hiller, M., Clarke, S.L., Schaar, B.T., Lowe, C.B., Wenger, A.M., and Bejerano, G. (2010). GREAT improves functional interpretation of cis-regulatory regions. Nat. Biotechnol. 28, 495–501. Medvedeva, Y.A., Khamis, A.M., Kulakovskiy, I. V., Ba-Alawi, W., Bhuyan, M.S.I., Kawaji, H., Lassmann, T., Harbers, M., Forrest, A.R.R., and Bajic, V.B. (2014). Effects of cytosine methylation on transcription factor binding sites. BMC Genomics 15. Meehan, R.R., Lewis, J.D., McKay, S., Kleiner, E.L., and Bird, A.P. (1989). Identification of a mammalian protein that binds specifically to DNA containing methylated CpGs. Cell 58, 499–507. Meers, M.P., Janssens, D.H., and Henikoff, S. (2019). Pioneer Factor-Nucleosome Binding Events during Differentiation Are Motif Encoded. Mol. Cell, 75, 562–575. Meissner, A. (2010). Epigenetic modifications in pluripotent and differentiated cells. Nat. Biotechnol. 28, 1079–1088. Méreau, H., De Rijck, J., Čermáková, K., Kutz, A., Juge, S., Demeulemeester, J., Gijsbers, R., Christ, F., Debyser, Z., and Schwaller, J. (2013). Impairing MLL-fusion gene-mediated transformation by dissecting critical interactions with the lens epithelium-derived growth factor (LEDGF/p75). Leukemia, 27, 1245–1253. Mikkelsen, T.S., Ku, M., Jaffe, D.B., Issac, B., Lieberman, E., Giannoukos, G., Alvarez, P., Brockman, W., Kim, T.K., Koche, R.P., et al. (2007). Genome-wide maps of chromatin state in pluripotent and lineage-committed cells. Nature 448, 553–560. Mingay, M., Chaturvedi, A., Bilenky, M., Cao, Q., Jackson, L., Hui, T., Moksa, M., Heravi-Moussavi, A., Humphries, R.K., Heuser, M., et al. (2018). Vitamin C-induced epigenomic remodelling in IDH1 mutant acute myeloid leukaemia. Leukemia 32, 11–20. 227 Miura, F., Enomoto, Y., Dairiki, R., and Ito, T. (2012). Amplification-free whole-genome bisulfite sequencing by post-bisulfite adaptor tagging. Nucleic Acids Res. 40. Mizzen, C.A., Yang, X.J., Kokubo, T., Brownell, J.E., Bannister, A.J., Owen-Hughes, T., Workman, J., Wang, L., Berger, S.L., Kouzarides, T., et al. (1996). The TAF(II)250 subunit of TFIID has histone acetyltransferase activity. Cell, 87, 1261–1270. Mochizuki-Kashio, M., Mishima, Y., Miyagi, S., Negishi, M., Saraya, A., Konuma, T., Shinga, J., Koseki, H., and Iwama, A. (2011). Dependency on the polycomb gene Ezh2 distinguishes fetal from adult hematopoietic stem cells. Blood, 118, 6553-6561. Mohandas, T., Sparkes, R.S., and Shapiro, L.J. (1981). Reactivation of an inactive human X chromosome: Evidence for X inactivation by DNA methylation. Science (80-. ). 211, 393–396. Morin, R.D., Johnson, N.A., Severson, T.M., Mungall, A.J., An, J., Goya, R., Paul, J.E., Boyle, M., Woolcock, B.W., Kuchenbauer, F., et al. (2010). Somatic mutations altering EZH2 (Tyr641) in follicular and diffuse large B-cell lymphomas of germinal-center origin. Nat. Genet. 42, 181–185. Müller, J. (1995). Transcriptional silencing by the Polycomb protein in Drosophila embryos. EMBO J. 14, 1209–1220. Myant, K., Termanis, A., Sundaram, A.Y.M., Boe, T., Li, C., Merusi, C., Burrage, J., De Las Heras, J.I., and Stancheva, I. (2011). LSH and G9a/GLP complex are required for developmentally programmed DNA methylation. Genome Res, 21, 83–94. Natarajan, T.G., Kallakury, B. V., Sheehan, C.E., Bartlett, M.B., Ganesan, N., Preet, A., Ross, J.S., and FitzGerald, K.T. (2010). Epigenetic regulator MLL2 shows altered expression in cancer cell lines and tumors from human breast and colon. Cancer Cell Int. 10. Nelson, D.M., Jaber-Hijazi, F., Cole, J.J., Robertson, N.A., Pawlikowski, J.S., Norris, K.T., 228 Criscione, S.W., Pchelintsev, N.A., Piscitello, D., Stong, N., et al. (2016). Mapping H4K20me3 onto the chromatin landscape of senescent cells indicates a function in control of cell senescence and tumor suppression through preservation of genetic and epigenetic stability. Genome Biol. 17, 158. Neri, F., Incarnato, D., Krepelova, A., Rapelli, S., Pagnani, A., Zecchina, R., Parlato, C., and Oliviero, S. (2013a). Genome-wide analysis identifies a functional association of Tet1 and Polycomb repressive complex 2 in mouse embryonic stem cells. Genome Biol. 14. Neri, F., Krepelova, A., Incarnato, D., Maldotti, M., Parlato, C., Galvagni, F., Matarese, F., Stunnenberg, H.G., and Oliviero, S. (2013b). XDnmt3L antagonizes DNA methylation at bivalent promoters and favors DNA methylation at gene bodies in ESCs. Cell 155, 121. Nguyen, C., Liang, G., Nguyen, T.D.T., Tsao-Wei, D., Groshen, S., Lübbert, M., Zhou, J.H., Benedict, W.F., and Jones, P.A. (2001). Susceptibility of nonpromoter CpG islands to de novo methylation in normal and neoplastic cells. J. Natl. Cancer Inst. 93, 1465–1472. Niu, X., Zhang, T., Liao, L., Zhou, L., Lindner, D.J., Zhou, M., Rini, B., Yan, Q., and Yang, H. (2012). The von Hippel-Lindau tumor suppressor protein regulates gene expression and tumor growth through histone demethylase JARID1C. Oncogene 31, 776–786. Noll, M., and Kornberg, R.D. (1977). Action of micrococcal nuclease on chromatin and the location of histone H1. J. Mol. Biol. 109, 393–404. Notta, F., Zandi, S., Takayama, N., Dobson, S., Gan, O.I., Wilson, G., Kaufmann, K.B., McLeod, J., Laurenti, E., Dunant, C.F., et al. (2016). Distinct routes of lineage development reshape the human blood hierarchy across ontogeny. Science (80-. ). 351. Noushmehr, H., Weisenberger, D.J., Diefes, K., Phillips, H.S., Pujara, K., Berman, B.P., Pan, F., Pelloski, C.E., Sulman, E.P., Bhat, K.P., et al. (2010). Identification of a CpG Island Methylator 229 Phenotype that Defines a Distinct Subgroup of Glioma. Cancer Cell 17, 510–522. Ogryzko, V. V., Schiltz, R.L., Russanova, V., Howard, B.H., and Nakatani, Y. (1996). The transcriptional coactivators p300 and CBP are histone acetyltransferases. Cell 87, 953–959. Okano, M., Bell, D.W., Haber, D.A., and Li, E. (1999). DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development. Cell 99, 247–257. Ooi, S.K.T., and Bestor, T.H. (2008). The Colorful History of Active DNA Demethylation. Cell 133, 1145–1148. Ooi, S.K.T., Qiu, C., Bernstein, E., Li, K., Jia, D., Yang, Z., Erdjument-Bromage, H., Tempst, P., Lin, S.P., Allis, C.D., et al. (2007). DNMT3L connects unmethylated lysine 4 of histone H3 to de novo methylation of DNA. Nature 448, 714–717. Osawa, T., Tsuchida, R., Muramatsu, M., Shimamura, T., Wang, F., Suehiro, J.I., Kanki, Y., Wada, Y., Yuasa, Y., Aburatani, H., et al. (2013). Inhibition of histone demethylase JMJD1A improves anti-angiogenic therapy and reduces tumor-associated macrophages. Cancer Res. 73, 3019–3028. Pan, J., McKenzie, Z.M., D’Avino, A.R., Mashtalir, N., Lareau, C.A., St. Pierre, R., Wang, L., Shilatifard, A., and Kadoch, C. (2019). The ATPase module of mammalian SWI/SNF family complexes mediates subcomplex identity and catalytic activity–independent genomic targeting. Nat. Genet. 51, 618–626. Park, I.K., Qian, D., Kiel, M., Becker, M.W., Pihalja, M., Weissman, I.L., Morrison, S.J., and Clarke, M.F. (2003). Bmi-1 is required for maintenance of adult self-renewing haematopoietic stem cells. Nature 423, 302–305. Parsons, D.W., Jones, S., Zhang, X., Lin, J.C.H., Leary, R.J., Angenendt, P., Mankoo, P., Carter, H., Siu, I.M., Gallia, G.L., et al. (2008). An integrated genomic analysis of human glioblastoma 230 multiforme. Science (80-. ). 321, 1807–1812. Paschka, P., Schlenk, R.F., Gaidzik, V.I., Habdank, M., Krönke, J., Bullinger, L., Späth, D., Kayser, S., Zucknick, M., Götze, K., et al. (2010). IDH1 and IDH2 mutations are frequent genetic alterations in acute myeloid leukemia and confer adverse prognosis in cytogenetically normal acute myeloid leukemia with NPM1 mutation without FLT3 internal tandem duplication. J. Clin. Oncol. 28, 3636–3643. Pasqualucci, L., Dominguez-Sola, D., Chiarenza, A., Fabbri, G., Grunn, A., Trifonov, V., Kasper, L.H., Lerach, S., Tang, H., Ma, J., et al. (2011a). Inactivating mutations of acetyltransferase genes in B-cell lymphoma. Nature 471, 189–196. Pasqualucci, L., Trifonov, V., Fabbri, G., Ma, J., Rossi, D., Chiarenza, A., Wells, V.A., Grunn, A., Messina, M., Elliot, O., et al. (2011b). Analysis of the coding genome of diffuse large B-cell lymphoma. Nat. Genet. 43, 830–837. Pastor, W.A., Aravind, L., and Rao, A. (2013). TETonic shift: Biological roles of TET proteins in DNA demethylation and transcription. Nat. Rev. Mol. Cell Biol. 14, 341–356. Pellacani, D., Bilenky, M., Kannan, N., Heravi-Moussavi, A., Knapp, D.J.H.F., Gakkhar, S., Moksa, M., Carles, A., Moore, R., Mungall, A.J., et al. (2016). Analysis of Normal Human Mammary Epigenomes Reveals Cell-Specific Active Enhancer States and Associated Transcription Factor Networks. Cell Rep. 17, 2060–2074. Peters, A.H.F.M., O’Carroll, D., Scherthan, H., Mechtler, K., Sauer, S., Schöfer, C., Weipoltshammer, K., Pagani, M., Lachner, M., Kohlmaier, A., et al. (2001). Loss of the Suv39h histone methyltransferases impairs mammalian heterochromatin and genome stability. Cell 107, 323–337. Petruk, S., Mariani, S.A., De Dominici, M., Porazzi, P., Minieri, V., Cai, J., Iacovitti, L., 231 Flomenberg, N., Calabretta, B., and Mazo, A. (2017). Structure of Nascent Chromatin Is Essential for Hematopoietic Lineage Specification. Cell Rep. 19, 295–306. Pfister, S., Rea, S., Taipale, M., Mendrzyk, F., Straub, B., Ittrich, C., Thuerigen, O., Sinn, H.P., Akhtar, A., and Lichter, P. (2008). The histone acetyltransferase hMOF is frequently downregulated in primary breast carcinoma and medulloblastoma and constitutes a biomarker for clinical outcome in medulloblastoma. Int. J. Cancer 122, 1207–1213. Pilsner, J.R., Liu, X., Ahsan, H., Ilievski, V., Slavkovich, V., Levy, D., Factor-Litvak, P., Graziano, J.H., and Gamble, M. V. (2007). Genomic methylation of peripheral blood leukocyte DNA: Influences of arsenic and folate in Bangladeshi adults. Am. J. Clin. Nutr. 86, 1179–1186. Pott, S., and Lieb, J.D. (2014). What are super-enhancers? Nat. Genet. 47, 8. Proctor, E., Waghray, M., Lee, C.J., Heidt, D.G., Yalamanchili, M., Li, C., Bednar, F., and Simeone, D.M. (2013). Bmi1 Enhances Tumorigenicity and Cancer Stem Cell Function in Pancreatic Adenocarcinoma. PLoS One 8. Quinlan, A.R., and Hall, I.M. (2010). BEDTools: A flexible suite of utilities for comparing genomic features. Bioinformatics 26, 841–842. Rai, K., Huggins, I.J., James, S.R., Karpf, A.R., Jones, D.A., and Cairns, B.R. (2008). DNA Demethylation in Zebrafish Involves the Coupling of a Deaminase, a Glycosylase, and Gadd45. Cell 135, 1201–1212. Ramirez, T., Brocher, J., Stopper, H., and Hock, R. (2008). Sodium arsenite modulates histone acetylation, histone deacetylase activity and HMGN protein dynamics in human cells. Chromosomai, 117, 147–157. Ramírez, F., Ryan, D.P., Grüning, B., Bhardwaj, V., Kilpert, F., Richter, A.S., Heyne, S., Dündar, F., and Manke, T. (2016). deepTools2: a next generation web server for deep- 232 sequencing data analysis. Nucleic Acids Res. 44, W160–W165. Reader, J.C., Meekins, J.S., Gojo, I., and Ning, Y. (2007). A novel NUP98-PHF23 fusion resulting from a cryptic translocation t(11;17)(p15;p13) in acute myeloid leukemia [17]. Leukemia 21, 842–844. Rebel, V.I., Kung, A.L., Tanner, E.A., Yang, H., Bronson, R.T., and Livingston, D.M. (2002). Distinct roles for CREB-binding protein and p300 in hematopoietic stem cell self-renewal. Proc. Natl. Acad. Sci. U. S. A. 99, 14789–14794. Rice, J.C., Briggs, S.D., Ueberheide, B., Barber, C.M., Shabanowitz, J., Hunt, D.F., Shinkai, Y., and Allis, C.D. (2003). Histone Methyltransferases Direct Different Degrees of Methylation to Define Distinct Chromatin Domains. Mol. Cell 12, 1591–1598. Rizzardi, L.F., Hickey, P.F., Rodriguez DiBlasi, V., Tryggvadóttir, R., Callahan, C.M., Idrizi, A., Hansen, K.D., and Feinberg, A.P. (2019). Neuronal brain-region-specific DNA methylation and chromatin accessibility are associated with neuropsychiatric trait heritability. Nat. Neurosci. 22, 307–316. Roadmap Epigenomics Consortium, Kundaje, A., Meuleman, W., Ernst, J., Bilenky, M., Yen, A., Heravi-Moussavi, A., Kheradpour, P., Zhang, Z., Wang, J., et al. (2015). Integrative analysis of 111 reference human epigenomes. Nature 518, 317–329. Robertson, K.D. (2005). DNA methylation and human disease. Nat. Rev. Genet. 6, 597–610. Robertson, G., Hirst, M., Bainbridge, M., Bilenky, M., Zhao, Y., Zeng, T., Euskirchen, G., Bernier, B., Varhol, R., Delaney, A., et al. (2007). Genome-wide profiles of STAT1 DNA association using chromatin immunoprecipitation and massively parallel sequencing. Nat. Methods 4, 651–657. Rodriguez-Fraticelli, A.E., Wolock, S.L., Weinreb, C.S., Panero, R., Patel, S.H., Jankovic, M., 233 Sun, J., Calogero, R.A., Klein, A.M., and Camargo, F.D. (2018). Clonal analysis of lineage fate in native haematopoiesis. Nature 553, 212–216. Rodriguez, R.M., Huidobro, C., Urdinguio, R.G., Mangas, C., Soldevilla, B., Domínguez, G., Bonilla, F., Fernandez, A.F., and Fraga, M.F. (2012). Aberrant epigenetic regulation of bromodomain Brd4 in human colon cancer. J. Mol. Med. 90, 587–595. De Rooij, J.D.E., Hollink, I.H.I.M., Arentsen-Peters, S.T.C.J.M., Van Galen, J.F., Berna Beverloo, H., Baruchel, A., Trka, J., Reinhardt, D., Sonneveld, E., Zimmermann, M., et al. (2013). NUP98/JARID1A is a novel recurrent abnormality in pediatric acute megakaryoblastic leukemia with a distinct HOX gene expression pattern. Leukemia 27, 2280–2288. Ropero, S., and Esteller, M. (2007). The role of histone deacetylases (HDACs) in human cancer. Mol. Oncol. 1, 19–25. Rountree, M.R., Bachman, K.E., and Baylin, S.B. (2000). DNMT1 binds HDAC2 and a new co-repressor, DMAP1, to form a complex at replication foci. Nat. Genet. 25, 269–277. De Ruijter, A.J.M., Van Gennip, A.H., Caron, H.N., Kemp, S., and Van Kuilenburg, A.B.P. (2003). Histone deacetylases (HDACs): Characterization of the classical HDAC family. Biochem. J. 370, 737–749. Rush, M., Appanah, R., Lee, S., Lam, L.L., Goyal, P., and Lorincz, M.C. (2009). Targeting of EZH2 to a defined genomic site is sufficient for recruitment of Dnmt3a but not de novo DNA methylation. Epigenetics 4, 404–414. Sabari, B.R., Dall’Agnese, A., Boija, A., Klein, I.A., Coffey, E.L., Shrinivas, K., Abraham, B.J., Hannett, N.M., Zamudio, A. V., Manteiga, J.C., et al. (2018). Coactivator condensation at super-enhancers links phase separation and gene control. Science (80-. ). 361, eaar3958. Sachs, M., Onodera, C., Blaschke, K., Ebata, K.T., Song, J.S., and Ramalho-Santos, M. (2013). 234 Bivalent Chromatin Marks Developmental Regulatory Genes in the Mouse Embryonic Germline InVivo. Cell Rep. 3, 1777–1784. Saghafinia, S., Mina, M., Riggi, N., Hanahan, D., and Ciriello, G. (2018). Pan-Cancer Landscape of Aberrant DNA Methylation across Human Tumors. Cell Rep. 25, 1066-1080.e8. Samudio-Ruiz, S.L., and Hudson, L.G. (2012). Increased DNA methyltransferase activity and DNA methylation following epidermal growth factor stimulation in ovarian cancer cells. Epigenetics 7, 216–224. Santos-Rosa, H., Schneider, R., Bannister, A.J., Sherriff, J., Bernstein, B.E., Emre, N.C.T., Schreiber, S.L., Mellor, J., and Kouzarides, T. (2002). Active genes are tri-methylated at K4 of histone H3. Nature 419, 407–411. Schaefer, M., Pollex, T., Hanna, K., Tuorto, F., Meusburger, M., Helm, M., and Lyko, F. (2010). RNA methylation by Dnmt2 protects transfer RNAs against stress-induced cleavage. Genes Dev. 24, 1590–1595. Schenk, T., Chen, W.C., Göllner, S., Howell, L., Jin, L., Hebestreit, K., Klein, H.U., Popescu, A.C., Burnett, A., Mills, K., et al. (2012). Inhibition of the LSD1 (KDM1A) demethylase reactivates the all-trans-retinoic acid differentiation pathway in acute myeloid leukemia. Nat. Med. 18, 605–611. Schulze, J.M., Jackson, J., Nakanishi, S., Gardner, J.M., Hentrich, T., Haug, J., Johnston, M., Jaspersen, S.L., Kobor, M.S., and Shilatifard, A. (2009). Linking Cell Cycle to Histone Modifications: SBF and H2B Monoubiquitination Machinery and Cell-Cycle Regulation of H3K79 Dimethylation. Mol. Cell 35, 626–641. Schwartzentruber, J., Korshunov, A., Liu, X.Y., Jones, D.T.W., Pfaff, E., Jacob, K., Sturm, D., Fontebasso, A.M., Quang, D.A.K., Tönjes, M., et al. (2012). Driver mutations in histone H3.3 235 and chromatin remodelling genes in paediatric glioblastoma. Nature 482, 226–231. Seto, E., and Yoshida, M. (2014). Erasers of histone acetylation: The histone deacetylase enzymes. Cold Spring Harb. Perspect. Biol, 6. Shema, E., Jones, D., Shoresh, N., Donohue, L., Ram, O., and Bernstein, B.E. (2016). Single-molecule decoding of combinatorially modified nucleosomes. Science (80-. ). 352, 717–721. Shen, Y., Zhu, Y.M., Fan, X., Shi, J.Y., Wang, Q.R., Yan, X.J., Gu, Z.H., Wang, Y.Y., Chen, B., Jiang, C.L., et al. (2011). Gene mutation patterns and their prognostic impact in a cohort of 1185 patients with acute myeloid leukemia. Blood 118, 5593–5603. Shi, Y., Lan, F., Matson, C., Mulligan, P., Whetstine, J.R., Cole, P.A., Casero, R.A., and Shi, Y. (2004). Histone demethylation mediated by the nuclear amine oxidase homolog LSD1. Cell 119, 941–953. Shiio, Y., and Eisenman, R.N. (2003). Histone sumoylation is associated with transcriptional repression. Proc. Natl. Acad. Sci. U. S. A. 100, 13225–13230. Shirane, K., Toh, H., Kobayashi, H., Miura, F., Chiba, H., Ito, T., Kono, T., and Sasaki, H. (2013). Mouse Oocyte Methylomes at Base Resolution Reveal Genome-Wide Accumulation of Non-CpG Methylation and Role of DNA Methyltransferases. PLoS Genet. 9, e1003439. Sigauke, E., Rakheja, D., Maddox, D.L., Hladik, C.L., White, C.L., Timmons, C.F., and Raisanen, J. (2006). Absence of expression of SMARCB1/INI1 in malignant rhabdoid tumors of the central nervous system, kidneys and soft tissue: An immunohistochemical study with implications for diagnosis. Mod. Pathol. 19, 717–725. Skvortsova, K., Zotenko, E., Luu, P.L., Gould, C.M., Nair, S.S., Clark, S.J., and Stirzaker, C. (2017). Comprehensive evaluation of genome-wide 5-hydroxymethylcytosine profiling approaches in human DNA. Epigenetics and Chromatin 10. 236 Smith, Z.D., and Meissner, A. (2013). DNA methylation: Roles in mammalian development. Nat. Rev. Genet. 14, 204–220. Solomon, M.J., Larsen, P.L., and Varshavsky, A. (1988). Mapping proteinDNA interactions in vivo with formaldehyde: Evidence that histone H4 is retained on a highly transcribed gene. Cell 53, 937–947. Song, L., and Crawford, G.E. (2010). DNase-seq: A high-resolution technique for mapping active gene regulatory elements across the genome from mammalian cells. Cold Spring Harb. Protoc. 5. Sørensen, A.L., Jacobsen, B.M., Reiner, A.H., Andersen, I.S., and Collas, P. (2010). Promoter DNA methylation patterns of differentiated cells are largely programmed at the progenitor stage. Mol. Biol. Cell 21, 2066–2077. Spiers, H., Hannon, E., Schalkwyk, L.C., Bray, N.J., and Mill, J. (2017). 5-hydroxymethylcytosine is highly dynamic across human fetal brain development. BMC Genomics 18. Stadler, M.B., Murr, R., Burger, L., Ivanek, R., Lienert, F., Schöler, A., Wirbelauer, C., Oakeley, E.J., Gaidatzis, D., Tiwari, V.K., et al. (2011). DNA-binding factors shape the mouse methylome at distal regulatory regions. Nature 480, 490–495. Stephens, D.C., and Poon, G.M.K. (2016). Differential sensitivity to methylated DNA by ETS-family transcription factors is intrinsically encoded in their DNA-binding domains. Nucleic Acids Res. 44, 8671–8681. Stewart, M.H., Albert, M., Sroczynska, P., Cruickshank, V.A., Guo, Y., Rossi, D.J., Helin, K., and Enver, T. (2015). The histone demethylase Jarid1b is required for hematopoietic stem cell self-renewal in mice. Blood 125, 2075–2078. 237 Stricker, S.H., Köferle, A., and Beck, S. (2016). From profiles to function in epigenomics. Nat. Rev. Genet. 18, 51–66. Stunnenberg, H.G., Abrignani, S., Adams, D., de Almeida, M., Altucci, L., Amin, V., Amit, I., Antonarakis, S.E., Aparicio, S., Arima, T., et al. (2016). The International Human Epigenome Consortium: A Blueprint for Scientific Collaboration and Discovery. Cell 167, 1145–1149. Stypula-Cyrus, Y., Damania, D., Kunte, D.P., Cruz, M. Dela, Subramanian, H., Roy, H.K., and Backman, V. (2013). HDAC Up-Regulation in Early Colon Field Carcinogenesis Is Involved in Cell Tumorigenicity through Regulation of Chromatin Structure. PLoS One 8. Sun, W., Zang, L., Shu, Q., and Li, X. (2014). From development to diseases: The role of 5hmC in brain. Genomics 104, 347–351. Sun, Z., Pan, J., Hope, W.Q.X., Cohen, S.N., and Balk, S.P. (1999). Tumor susceptibility gene 101 protein represses androgen receptor transactivation and interacts with p300. Cancer, 86, 689–696. Swift, J., Spinler, K.R., Discher, D.E., Mohandas, N., Chasis, J.A., and Shin, J.-W. (2013). Lamins regulate cell trafficking and lineage maturation of adult human hematopoietic cells. Proc. Natl. Acad. Sci, 110, 18892–18897. Tachibana, M., Sugimoto, K., Nozaki, M., Ueda, J., Ohta, T., Ohki, M., Fukuda, M., Takeda, N., Niida, H., Kato, H., et al. (2002). G9a histone methyltransferase plays a dominant role in euchromatic histone H3 lysine 9 methylation and is essential for early embryogenesis. Genes Dev. 16, 1779–1791. Tahiliani, M., Koh, K.P., Shen, Y., Pastor, W.A., Bandukwala, H., Brudno, Y., Agarwal, S., Iyer, L.M., Liu, D.R., Aravind, L., et al. (2009). Conversion of 5-methylcytosine to 5-hydroxymethylcytosine in mammalian DNA by MLL partner TET1. Science (80-. ). 324, 930– 238 935. Tan, C.-T., Ko, J.-Y., and Tsai, H.-H. (2012). The Role of G9a in Tumorigenesis in Head and Neck Cancer. Otolaryngol. Neck Surg. 147, P183–P184. Tarasov, A., Vilella, A.J., Cuppen, E., Nijman, I.J., and Prins, P. (2015). Sambamba: Fast processing of NGS alignment formats. Bioinformatics, 31, 2032–2034 Taunton, J., Hassig, C.A., and Schreiber, S.L. (1996). A mammalian histone deacetylase related to the yeast transcriptional regulator Rpd3p. Science (80-. ). 272, 408–411. Team, R.D.C., and R Development Core Team, R. (2016). R: A Language and Environment for Statistical Computing. R Found. Stat. Comput. Tefferi, A., Lim, K.H., Abdel-Wahab, O., Lasho, T.L., Patel, J., Patnaik, M.M., Hanson, C.A., Pardanani, A., Gilliland, D.G., and Levine, R.L. (2009). Detection of mutant TET2 in myeloid malignancies other than myeloproliferative neoplasms: CMML, MDS, MDS/MPN and AML. Leukemia 23, 1343–1345. Tenen, D.G. (2003). Disruption of differentiation in human cancer: AML shows the way. Nat. Rev. Cancer 3, 89–101. Terme, J.M., Sesé, B., Millán-Ariño, L., Mayor, R., Belmonte, J.C.I., Barrero, M.J., and Jordan, A. (2011). Histone H1 variants are differentially expressed and incorporated into chromatin during differentiation and reprogramming to pluripotency. J. Biol. Chem. 286, 35347–35357. Thiel, A.T., Blessington, P., Zou, T., Feather, D., Wu, X., Yan, J., Zhang, H., Liu, Z., Ernst, P., Koretzky, G.A., et al. (2010). MLL-AF9-Induced Leukemogenesis Requires Coexpression of the Wild-Type Mll Allele. Cancer Cell 17, 148–159. Tompkins, J.D., Hall, C., Chen, V.C. -y., Li, A.X., Wu, X., Hsu, D., Couture, L.A., and Riggs, A.D. (2012). Epigenetic stability, adaptability, and reversibility in human embryonic stem cells. 239 Proc. Natl. Acad. Sci. 109, 12544-12549. Toth, M., Boros, I.M., and Balint, E. (2012). Elevated level of lysine 9-acetylated histone H3 at the MDR1 promoter in multidrug-resistant cells. Cancer Sci. 103, 659–669. Tripathi, S., Pohl, M.O., Zhou, Y., Rodriguez-Frandsen, A., Wang, G., Stein, D.A., Moulton, H.M., Dejesus, P., Che, J., Mulder, L.C.F., et al. (2015). Meta- and Orthogonal Integration of Influenza “oMICs” Data Defines a Role for UBR4 in Virus Budding. Cell Host Microbe 18, 723–735. Tsumura, A., Hayakawa, T., Kumaki, Y., Takebayashi, S.I., Sakaue, M., Matsuoka, C., Shimotohno, K., Ishikawa, F., Li, E., Ueda, H.R., et al. (2006). Maintenance of self-renewal ability of mouse embryonic stem cells in the absence of DNA methyltransferases Dnmt1, Dnmt3a and Dnmt3b. Genes to Cells 11, 805–814. Turcan, S., Rohle, D., Goenka, A., Walsh, L.A., Fang, F., Yilmaz, E., Campos, C., Fabius, A.W.M., Lu, C., Ward, P.S., et al. (2012). IDH1 mutation is sufficient to establish the glioma hypermethylator phenotype. Nature 483, 479–483. Unoki, M., Masuda, A., Dohmae, N., Arita, K., Yoshimatsu, M., Iwai, Y., Fukui, Y., Ueda, K., Hamamoto, R., Shirakawa, M., et al. (2013). Lysyl 5-hydroxylation, a novel histone modification, by jumonji domain containing 6 (JMJD6). J. Biol. Chem. 288, 6053–6062. Verbeek, J.J., Vlassis, N., and Kröse, B. (2003). Efficient greedy learning of gaussian mixture models. Neural Comput. 15, 469–485. Verdin, E., Dequiedt, F., and Kasler, H.G. (2003). Class II histone deacetylases: Versatile regulators. Trends Genet. 19, 286–293. Voigt, P., Tee, W.W., and Reinberg, D. (2013). A double take on bivalent promoters. Genes Dev. 27, 1318–1338. 240 Voong, L.N., Xi, L., Sebeson, A.C., Xiong, B., Wang, J.P., and Wang, X. (2016). Insights into Nucleosome Organization in Mouse Embryonic Stem Cells through Chemical Mapping. Cell 167, 1555-1570.e15. Waddington, C.H. (2006). Genetic Assimilation of the Bithorax Phenotype. Evolution (N. Y). Walter, M.J., Ding, L., Shen, D., Shao, J., Grillot, M., McLellan, M., Fulton, R., Schmidt, H., Kalicki-Veizer, J., O’Laughlin, M., et al. (2011). Recurrent DNMT3A mutations in patients with myelodysplastic syndromes. Leukemia 25, 1153–1158. Wang, D., Garcia-Bassets, I., Benner, C., Li, W., Su, X., Zhou, Y., Qiu, J., Liu, W., Kaikkonen, M.U., Ohgi, K.A., et al. (2011a). Reprogramming transcription by distinct classes of enhancers functionally defined by eRNA. Nature 474, 390–397. Wang, G., Luo, X., Wang, J., Wan, J., Xia, S., Zhu, H., Qian, J., and Wang, Y. (2018). MeDReaders: A database for transcription factors that bind to methylated DNA. Nucleic Acids Res. 46, D146–D151. Wang, G.G., Cai, L., Pasillas, M.P., and Kamps, M.P. (2007). NUP98-NSD1 links H3K36 methylation to Hox-A gene activation and leukaemogenesis. Nat. Cell Biol. 9, 804–812. Wang, G.G., Song, J., Wang, Z., Dormann, H.L., Casadio, F., Li, H., Luo, J.L., Patel, D.J., and Allis, C.D. (2009a). Haematopoietic malignancies caused by dysregulation of a chromatin-binding PHD finger. Nature 459, 847–851. Wang, H., Huang, Z.Q., Xia, L., Feng, Q., Erdjument-Bromage, H., Strahl, B.D., Briggs, S.D., Allis, C.D., Wong, J., Tempst, P., et al. (2001). Methylation of histone H4 at arginine 3 facilitating transcriptional activation by nuclear hormone receptor. Science (80-. ). 293, 853–857. Wang, J., Lu, F., Ren, Q., Sun, H., Xu, Z., Lan, R., Liu, Y., Ward, D., Quan, J., Ye, T., et al. (2011b). Novel histone demethylase LSD1 inhibitors selectively target cancer cells with 241 pluripotent stem cell properties. Cancer Res. 71, 7238–7249. Wang, K., Kan, J., Yuen, S.T., Shi, S.T., Chu, K.M., Law, S., Chan, T.L., Kan, Z., Chan, A.S.Y., Tsui, W.Y., et al. (2011c). Exome sequencing identifies frequent mutation of ARID1A in molecular subtypes of gastric cancer. Nat. Genet, 43. 1219–1223. Wang, Y., Zhang, H., Chen, Y., Sun, Y., Yang, F., Yu, W., Liang, J., Sun, L., Yang, X., Shi, L., et al. (2009b). LSD1 Is a Subunit of the NuRD Complex and Targets the Metastasis Programs in Breast Cancer. Cell 138, 660–672. Wang, Y., Zhang, R., Wu, D., Lu, Z., Sun, W., Cai, Y., Wang, C., and Jin, J. (2013). Epigenetic change in kidney tumor: Downregulation of histone acetyltransferase MYST1 in human renal cell carcinoma. J. Exp. Clin. Cancer Res. 32. Wang, Y., Xiao, M., Chen, X., Chen, L., Xu, Y., Lv, L., Wang, P., Yang, H., Ma, S., Lin, H., et al. (2015). WT1 recruits TET2 to regulate its target gene expression and suppress leukemia cell proliferation. Mol. Cell 57, 662–673. Ward, P.S., Patel, J., Wise, D.R., Abdel-Wahab, O., Bennett, B.D., Coller, H.A., Cross, J.R., Fantin, V.R., Hedvat, C. V., Perl, A.E., et al. (2010). The Common Feature of Leukemia-Associated IDH1 and IDH2 Mutations Is a Neomorphic Enzyme Activity Converting α-Ketoglutarate to 2-Hydroxyglutarate. Cancer Cell 17, 225–234. Watson, J.D., and Crick, F.H.C. (1953). Molecular structure of nucleic acids: A structure for deoxyribose nucleic acid. Nature 171, 737–738. Watt, F., and Molloy, P.L. (1988). Cytosine methylation prevents binding to DNA of a HeLa cell transcription factor required for optimal expression of the adenovirus major late promoter. Genes Dev. 2, 1136–1143. Weake, V.M., and Workman, J.L. (2008). Histone Ubiquitination: Triggering Gene Activity. 242 Mol. Cell 29, 653–663. Weber, A.R., Krawczyk, C., Robertson, A.B., Kusnierczyk, A., Vågbø, C.B., Schuermann, D., Klungland, A., and Schar, P. (2016). Biochemical reconstitution of TET1-TDG-BER-dependent active DNA demethylation reveals a highly coordinated mechanism. Nat. Commun. 7. Wei, Y., Chen, R., Dimicoli, S., Bueso-Ramos, C., Neuberg, D., Pierce, S., Wang, H., Yang, H., Jia, Y., Zheng, H., et al. (2013). Global H3K4me3 genome mapping reveals alterations of innate immunity signaling and overexpression of JMJD3 in human myelodysplastic syndrome CD34+ cells. Leukemia 27, 2177–2186. Wen, B., Wu, H., Shinkai, Y., Irizarry, R.A., and Feinberg, A.P. (2009). Large organized chromatin K9-modifications (LOCKs) distinguish differentiated from embryonic stem cells. Nat. Genet, 41, 246–250. West, J.A., Cook, A., Alver, B.H., Stadtfeld, M., Deaton, A.M., Hochedlinger, K., Park, P.J., Tolstorukov, M.Y., and Kingston, R.E. (2014). Nucleosomal occupancy changes locally over key regulatory regions during cell differentiation and reprogramming. Nat. Commun. 5. Whetstine, J.R., Nottke, A., Lan, F., Huarte, M., Smolikov, S., Chen, Z., Spooner, E., Li, E., Zhang, G., Colaiacovo, M., et al. (2006). Reversal of Histone Lysine Trimethylation by the JMJD2 Family of Histone Demethylases. Cell 125, 467–481. Wiegand, K.C., Shah, S.P., Al-Agha, O.M., Zhao, Y., Tse, K., Zeng, T., Senz, J., McConechy, M.K., Anglesio, M.S., Kalloger, S.E., et al. (2010). ARID1A Mutations in Endometriosis-Associated Ovarian Carcinomas. N. Engl. J. Med. 363, 1532–1543. Wiench, M., John, S., Baek, S., Johnson, T.A., Sung, M.H., Escobar, T., Simmons, C.A., Pearce, K.H., Biddie, S.C., Sabo, P.J., et al. (2011). DNA methylation status predicts cell type-specific enhancer activity. EMBO J. 30, 3028–3039. 243 Wu, X., and Zhang, Y. (2017). TET-mediated active DNA demethylation: Mechanism, function and beyond. Nat. Rev. Genet. 18, 517–534. Wu, C. Te, Wu, C.F., Lu, C.H., Lin, C.C., Chen, W.C., Lin, P.Y., and Chen, M.F. (2011a). Expression and function role of DNA methyltransferase 1 in human bladder cancer. Cancer 117, 5221–5233. Wu, H., D’Alessio, A.C., Ito, S., Wang, Z., Cui, K., Zhao, K., Sun, Y.E., and Zhang, Y. (2011b). Genome-wide analysis of 5-hydroxymethylcytosine distribution reveals its dual function in transcriptional regulation in mouse embryonic stem cells. Genes Dev. 25, 679–684. Wu, X., Johansen, J.V., and Helin, K. (2013). Fbxl10/Kdm2b Recruits Polycomb Repressive Complex 1 to CpG Islands and Regulates H2A Ubiquitylation. Mol. Cell 49, 1134–1146. Xie, H., Xu, J., Hsu, J.H., Nguyen, M., Fujiwara, Y., Peng, C., and Orkin, S.H. (2014). Polycomb repressive complex 2 regulates normal hematopoietic stem cell function in a developmental-stage-specific manner. Cell Stem Cell 14, 68–80. Xie, H.J., Noh, J.H., Kim, J.K., Jung, K.H., Eun, J.W., Bae, H.J., Kim, M.G., Chang, Y.G., Lee, J.Y., Park, H., et al. (2012). HDAC1 Inactivation Induces Mitotic Defect and Caspase-Independent Autophagic Cell Death in Liver Cancer. PLoS One 7, e34265. Xu, J., Zhu, W., Xu, W., Yao, W., Zhang, B., Xu, Y., Ji, S., Liu, C., Long, J., Ni, Q., et al. (2013). Up-Regulation of MBD1 Promotes Pancreatic Cancer Cell Epithelial-Mesenchymal Transition and Invasion by Epigenetic Down-Regulation of E-Cadherin. Curr. Mol. Med. 13, 387–400. Xu, J., Shao, Z., Li, D., Xie, H., Kim, W., Huang, J., Taylor, J.E., Pinello, L., Glass, K., Jaffe, J.D., et al. (2015). Developmental control of polycomb subunit composition by GATA factors mediates a switch to non-canonical functions. Mol. Cell 57, 304–316. 244 Xu, W., Yang, H., Liu, Y., Yang, Y., Wang, P., Kim, S.H., Ito, S., Yang, C., Wang, P., Xiao, M.T., et al. (2011). Oncometabolite 2-hydroxyglutarate is a competitive inhibitor of α-ketoglutarate-dependent dioxygenases. Cancer Cell 19, 17–30. Yan, G., Eller, M.S., Elm, C., Larocca, C.A., Ryu, B., Panova, I.P., Dancy, B.M., Bowers, E.M., Meyers, D., Lareau, L., et al. (2013). Selective inhibition of p300 HAT blocks cell cycle progression, induces cellular senescence, and inhibits the DNA damage response in melanoma cells. J. Invest. Dermatol. 133, 2444–2452. Yan, X.J., Xu, J., Gu, Z.H., Pan, C.M., Lu, G., Shen, Y., Shi, J.Y., Zhu, Y.M., Tang, L., Zhang, X.W., et al. (2011). Exome sequencing identifies somatic mutations of DNA methyltransferase gene DNMT3A in acute monocytic leukemia. Nat. Genet. 43, 309–317. Yang, X.J., and Seto, E. (2003). Collaborative spirit of histone deacetylases in regulating chromatin structure and gene expression. Curr. Opin. Genet. Dev. 13, 143–153. Yang, J., Wei, X., Wu, Q., Xu, Z., Gu, D., Jin, Y., Shen, Y., Huang, H., Fan, H., and Chen, J. (2011). Clinical significance of the expression of DNA methyltransferase proteins in gastric cancer. Mol. Med. Rep. 4, 1139–1143. Yang, M., Gocke, C.B., Luo, X., Borek, D., Tomchick, D.R., Machius, M., Otwinowski, Z., and Yu, H. (2006). Structural Basis for CoREST-Dependent Demethylation of Nucleosomes by the Human LSD1 Histone Demethylase. Mol. Cell 23, 377–387. Yang, W.M., Inouye, C., Zeng, Y., Bearss, D., and Seto, E. (1996). Transcriptional repression by YY1 is mediated by interaction with a mammalian homolog of the yeast global regulator RPD3. Proc. Natl. Acad. Sci. U. S. A, 92, 12845–12850. Yang, W.M., Yao, Y.L., Sun, J.M., Davie, J.R., and Seto, E. (1997). Isolation and characterization of cDNAs corresponding to an additional member of the human histone 245 deacetylase gene family. J. Biol. Chem. 272, 28001–28007. Yang, Y., Yin, X., Yang, H., and Xu, Y. (2015). Histone demethylase LSD2 Acts as an E3Ubiquitin ligase and inhibits cancer cell growth through promoting proteasomal degradation of OGT. Mol. Cell 58, 47–59. Yu, W., Zhang, F., Wang, S., Fu, Y., Chen, J., Liang, X., Le, H., Pu, W.T., and Zhang, B. (2017). Depletion of polycomb repressive complex 2 core component EED impairs fetal hematopoiesis. Cell Death Dis, 8. Yuan, W., Xie, J., Long, C., Erdjument-Bromage, H., Ding, X., Zheng, Y., Tempst, P., Chen, S., Zhu, B., and Reinberg, D. (2009). Heterogeneous nuclear ribonucleoprotein L is a subunit of human KMT3a/set2 complex required for H3 Lys-36 trimethylation activity in vivo. J. Biol. Chem. 284, 15701–15707. Zeisig, B.B., Arteaga, M.F., Thirant, C., and So, C.W.E. (2011). Collaboration between PcG Proteins and MLL Fusions in Leukemogenesis: An Emerging Paradigm. Cancer Cell 20, 551–553. Zentner, G.E., Tesar, P.J., and Scacheri, P.C. (2011). Epigenetic signatures distinguish multiple classes of enhancers with distinct cellular functions. Genome Res. 21, 1273–1283. Zhang, J.J., Zhu, Y., Zhu, Y., Wu, J.L., Liang, W.B., Zhu, R., Xu, Z.K., Du, Q., and Miao, Y. (2012). Association of increased DNA methyltransferase expression with carcinogenesis and poor prognosis in pancreatic ductal adenocarcinoma. Clin. Transl. Oncol. 14, 116-124. Zhang, W., Li, J., Suzuki, K., Qu, J., Wang, P., Zhou, J., Liu, X., Ren, R., Xu, X., Ocampo, A., et al. (2015). A Werner syndrome stem cell model unveils heterochromatin alterations as a driver of human aging. Science (80-. ). 348, 1160–1163. Zhang, Y.W., Wang, Z., Xie, W., Cai, Y., Xia, L., Easwaran, H., Luo, J., Yen, R.W.C., Li, Y., 246 and Baylin, S.B. (2017). Acetylation Enhances TET2 Function in Protecting against Abnormal DNA Methylation during Oxidative Stress. Mol. Cell 65, 323–335. Zhao, L., Wang, D.L., Liu, Y., Chen, S., and Sun, F.L. (2013). Histone acetyltransferase hMOF promotes S phase entry and tumorigenesis in lung cancer. Cell. Signal. 25, 1689–1698. Zhu, C., Qin, Y.-R., Xie, D., Chua, D.T.T., Fung, J.M., Chen, L., Fu, L., Hu, L., and Guan, X.-Y. (2009). Characterization of tumor suppressive function of P300/CBP-associated factor at frequently deleted region 3p24 in esophageal squamous cell carcinoma. Oncogene 28, 2821–2828. Zhu, D., Hunter, S.B., Vertino, P.M., and Van Meir, E.G. (2011). Overexpression of MBD2 in glioblastoma maintains epigenetic silencing and inhibits the antiangiogenic function of the tumor suppressor gene BAI. Cancer Res. 71, 5859–5870. Zhu, J., Adli, M., Zou, J.Y., Verstappen, G., Coyne, M., Zhang, X., Durham, T., Miri, M., Deshpande, V., De Jager, P.L., et al. (2013). Genome-wide chromatin state transitions associated with developmental and environmental cues. Cell 152, 642–654. Ziegler-Heitbrock, L., Ancuta, P., Crowe, S., Dalod, M., Grau, V., Hart, D.N., Leenen, P.J.M., Liu, Y.J., MacPherson, G., Randolph, G.J., et al. (2010). Nomenclature of monocytes and dendritic cells in blood. Blood 116, 74–80. """@en ; edm:hasType "Thesis/Dissertation"@en ; vivo:dateIssued "2020-05"@en ; edm:isShownAt "10.14288/1.0388290"@en ; dcterms:language "eng"@en ; ns0:degreeDiscipline "Microbiology and Immunology"@en ; edm:provider "Vancouver : University of British Columbia Library"@en ; dcterms:publisher "University of British Columbia"@en ; dcterms:rights "Attribution-NonCommercial-NoDerivatives 4.0 International"@* ; ns0:rightsURI "http://creativecommons.org/licenses/by-nc-nd/4.0/"@* ; ns0:scholarLevel "Graduate"@en ; dcterms:title "Epigenetic regulation in normal hematopoiesis and its dysfunction in leukemia"@en ; dcterms:type "Text"@en ; ns0:identifierURI "http://hdl.handle.net/2429/73311"@en .