Project description:The de novo DNA methyltransferase Dnmt3a is mutated in human acute myeloid leukemia, and suppresses tumorigenesis in murine models of leukemia and lung cancer. Conversely, deregulation of the other de novo DNA methyltransferase, Dnmt3b, predominantly promotes tumorigenesis. However, the molecular mechanisms underlying the roles of Dnmt3a and Dnmt3b in cancer remain poorly understood. Using conditional knockout mice, here we show that Dnmt3a -- but not Dnmt3b -- strongly protects epidermal stem cells from carcinogen-induced tumor initiation, without affecting the progression of benign lesions to aggressive carcinomas. Only upon combined deletion of Dnmt3a and Dnmt3b, squamous cell carcinomas acquired a more aggressive fate and even became metastatic, indicating that Dnmt3b is tumor-suppressive, rather than pro-tumorigenic, in epidermal neoplasia. Mechanistically, Dnmt3a promotes the expression of epidermal differentiation genes by interacting with their enhancers, and inhibits the expression of lipid metabolism and cell proliferation genes by directly methylating their promoters. Altogether, we demonstrate that Dnmt3a, but not Dnmt3b, is critical for suppressing epidermal tumor initiation, while both enzymes prevent tumor progression.
Project description:The de novo DNA methyltransferase Dnmt3a is mutated in human acute myeloid leukemia, and suppresses tumorigenesis in murine models of leukemia and lung cancer. Conversely, deregulation of the other de novo DNA methyltransferase, Dnmt3b, predominantly promotes tumorigenesis. However, the molecular mechanisms underlying the roles of Dnmt3a and Dnmt3b in cancer remain poorly understood. Using conditional knockout mice, here we show that Dnmt3a -- but not Dnmt3b -- strongly protects epidermal stem cells from carcinogen-induced tumor initiation, without affecting the progression of benign lesions to aggressive carcinomas. Only upon combined deletion of Dnmt3a and Dnmt3b, squamous cell carcinomas acquired a more aggressive fate and even became metastatic, indicating that Dnmt3b is tumor-suppressive, rather than pro-tumorigenic, in epidermal neoplasia. Mechanistically, Dnmt3a promotes the expression of epidermal differentiation genes by interacting with their enhancers, and inhibits the expression of lipid metabolism and cell proliferation genes by directly methylating their promoters. Altogether, we demonstrate that Dnmt3a, but not Dnmt3b, is critical for suppressing epidermal tumor initiation, while both enzymes prevent tumor progression.
Project description:The de novo DNA methyltransferase Dnmt3a is mutated in human acute myeloid leukemia, and suppresses tumorigenesis in murine models of leukemia and lung cancer. Conversely, deregulation of the other de novo DNA methyltransferase, Dnmt3b, predominantly promotes tumorigenesis. However, the molecular mechanisms underlying the roles of Dnmt3a and Dnmt3b in cancer remain poorly understood. Using conditional knockout mice, here we show that Dnmt3a -- but not Dnmt3b -- strongly protects epidermal stem cells from carcinogen-induced tumor initiation, without affecting the progression of benign lesions to aggressive carcinomas. Only upon combined deletion of Dnmt3a and Dnmt3b, squamous cell carcinomas acquired a more aggressive fate and even became metastatic, indicating that Dnmt3b is tumor-suppressive, rather than pro-tumorigenic, in epidermal neoplasia. Mechanistically, Dnmt3a promotes the expression of epidermal differentiation genes by interacting with their enhancers, and inhibits the expression of lipid metabolism and cell proliferation genes by directly methylating their promoters. Altogether, we demonstrate that Dnmt3a, but not Dnmt3b, is critical for suppressing epidermal tumor initiation, while both enzymes prevent tumor progression.
Project description:Here, we show that Dnmt3a and Dnmt3b show non-overlapping and unique patterns of genomic localization in human epidermal stem cells and their differentiated counterparts. Dnmt3a, but not Dnmt3b, binds to the TSSs of a cohort of genes required for the interaction of stem cells with their underlying stroma. Unexpectedly, TSSs bound by Dnmt3a are highly transcribed and are devoid of DNA-methylation. Conversely, Dnmt3b specifically decorates the genebody of genes that establish the stem cell and differentiated signatures. Genic occupation by Dnmt3b correlates with high levels of DNA-methylation, broad domains of histone H3K4me3 8, and robust transcription. Intriguingly, both proteins also bind to the most active subset of enhancers, and are required for the production of their associated bidirectional enhancer RNAs 9. We show that typical and super-enhancers are very dynamically regulated during the linear transition of epidermal stem cells to differentiated keratinocytes. Interestingly, Dnmt3a and Dnmt3b show a strong preference for the super-enhancers that define the ectodermal lineage, but importantly, that also establish the functional traits associated to the stem cell and differentiated states. These enhancers contain very low levels of DNA-methylation, but high amounts of DNA-hydroxymethylation. Depletion of either protein completely impairs human epidermal stem cell self-renewal by inducing their spontaneous differentiation.
Project description:The Nucleosome Remodeling and Deacetylase (NuRD) complex plays an important role in gene expression regulation, stem cell self-renewal, and lineage commitment. Yet little is known about the dynamics of NuRD during cellular differentiation. Here, we study these dynamics using genome-wide profiling and quantitative interaction proteomics in mouse embryonic stem cells (ESCs) and neural progenitor cells (NPCs). The genomic targets of NuRD are highly dynamic during differentiation, with most binding occurring at cell-type specific promoters and enhancers. We identify ZFP296 as a novel, ESC-specific NuRD interactor that also interacts with the SIN3A complex. ChIP-sequencing in Zfp296 knockout (KO) ESCs reveals decreased NuRD binding both genome-wide and at ZFP296 binding sites, although this has little effect on the transcriptome. Nevertheless, Zfp296 KO ESCs exhibit delayed induction of lineage-specific markers upon differentiation to embryoid bodies. In summary, we identify an ESC-specific NuRD interacting protein which regulates genome-wide NuRD binding and cellular differentiation.
Project description:FOXA1 is a pioneer factor that is important in hormone dependent cancer cells to stabilise nuclear receptors, such as estrogen receptor (ER) to chromatin. FOXA1 binds to enhancers regions that are enriched in H3K4mono- and dimethylation (H3K4me1, H3K4me2) histone marks and evidence suggests that these marks are requisite events for FOXA1 to associate with enhancers to initate subsequent gene expression events. However, exogenous expression of FOXA1 has been shown to induce H3K4me1 and H3K4me2 signal at enhancer elements and the order of events and the functional importance of these events is not clear. We performed a FOXA1 Rapid Immunoprecipitation Mass Spectrometry of Endogenous Proteins (RIME) screen in ERα-positive MCF-7 breast cancer cells in order to identify FOXA1 interacting partners and we found histone-lysine N-methyltransferase (MLL3) as the top FOXA1 interacting protein. MLL3 is typically thought to induce H3K4me3 at promoter regions, but recent findings suggest it may contribute to H3K4me1 deposition, in line with our observation that MLL3 associates with an enhancer specific protein. We performed MLL3 ChIP-seq in breast cancer cells and unexpectedly found that MLL3 binds mostly at non-promoter regions enhancers, in contrast to the prevailing hypothesis. MLL3 was shown to occupy regions marked by FOXA1 occupancy and as expected, H3K4me1 and H3K4me2. MLL3 binding was dependent on FOXA1, indicating that FOXA1 recruits MLL3 to chromatin. Motif analysis and subsequent genomic mapping revealed a role for Grainy head like protein-2 (GRHL2) which was shown to co-occupy regions of the chromatin with MLL3. Regions occupied by all three factors, namely FOXA1, MLL3 and GRHL2, were most enriched in H3K4me1. MLL3 silencing decreased H3K4me1 at enhancer elements, but had no appreciable impact on H3K4me3 at enhancer elements. We identify a complex relationship between FOXA1, MLL3 and H3K4me1 at enhancers in breast cancer and propose a mechanism whereby the pioneer factor FOXA1 can interact with a chromatin modifier MLL3, recruiting it to chromatin to facilitate the deposition of H3K4me1 histone marks, subsequently demarcating active enhancer elements.
Project description:In this study, we mapped modification of lysine 4 and lysine 27 of histone H3 genome-wide in a series of mouse embryonic stem cells (mESCs) varying in DNA methylation levels based on knock-out and reconstitution of DNA methyltransferases (DNMTs). We extend previous studies showing cross-talk between DNA methylation and histone modifications by examining a breadth of histone modifications, causal relationships, and direct effects. Our data shows a causal regulation of H3K27me3 at gene promoters as well as H3K27ac and H3K27me3 at tissue-specific enhancers. We also identify isoform differences between DNMT family members. This study provides a comprehensive resource for the study of the complex interplay between DNA methylation and histone modification landscape. Histone ChIP-seq of H3K4me3, H3K27me3, H3K4me1, and H3K27ac were performed on wild-type, Dnmt triple knock-out (Dnmt1/3a/3b; TKO), Dnmt double knock-out (Dnmt3a/3b; DKO), and respective reconstitution mouse embryonic stem cell lines