HomeResearchEpigenetic Mechanisms of Pluripotency and Lineage Commitment

Our Scientists

Epigenetic Mechanisms of Pluripotency and Lineage Commitment

Research Summary

Bradley Bernstein applies genomic approaches to study chromatin structure and function, with the goal of understanding how epigenetic regulation contributes to mammalian development and human disease.

My laboratory studies chromatin, the higher-order structure that organizes and regulates the genome. We seek to understand how histone modifications and associated proteins contribute to the epigenetic regulation of gene expression during mammalian development. We are also interested in how these regulatory mechanisms go awry in cancer and other diseases. Our studies provide comprehensive views of the chromatin landscape in mammalian cells (the epigenome) and identify signature chromatin structures that appear to underlie the epigenetic regulation of key genes in developing stem cells.

Although they carry essentially identical genetic information, the different cell types in a multicellular organism exhibit markedly different phenotypes that can be maintained over long periods. These cellular states reflect distinct underlying transcriptional programs. Chromatin structure, which is closely linked to transcription, both influences and is influenced by this process. Increasing evidence suggests that chromatin plays particularly important roles in maintaining lineage-specific transcriptional programs and facilitating their faithful transmission to daughter cells. This epigenetic gene regulation, which is critical to mammalian development, is poorly understood.

My lab applies an integrative approach involving biochemistry, genetics, and genomics to study chromatin in stem cells and disease models. We have combined cell biology and biochemical techniques with microarrays and high-throughput sequencing technologies to acquire comprehensive views of chromatin state across the genome in diverse mammalian cells. To gain insight into how chromatin influences genome function in distinct settings, we have applied these approaches to a wide range of modifications in many cell types. We are also analyzing specific modifications in developing stem cells, with the goal of elucidating chromatin and epigenetic regulatory mechanisms that underlie development.

Our studies have revealed a complex and dynamic chromatin landscape. We have identified large regions, or domains, continuously enriched for modified histones. These typically coincide with loci that encode transcription factors or other developmental regulators. In lineage-committed cells, domains of the "activating" histone H3 lysine 4 trimethylation mark overlay active developmental genes, while "repressive" lysine 27 trimethylation correlates with silent ones. We proposed that these domains contribute to the epigenetic maintenance of gene expression patterns through cell division. The implicated modifications are intimately linked to trithorax (lysine 4) and Polycomb (lysine 27) protein complexes, which are essential for epigenetic regulation of Hox genes and other developmental regulators in multicellular organisms. The large size of the domains may help ensure faithful transmission of the chromatin modifications as histones segregate to daughter strands of the DNA during mitosis.

Unexpectedly, extension of our studies to pluripotent embryonic stem (ES) cells revealed a novel type of chromatin domain enriched for both lysine 4 and lysine 27 trimethylation. We hypothesized that the opposing modifications of these "bivalent domains" help keep developmental genes silent but poised for activation in ES cells. We subsequently traced the fates of bivalent domains during differentiation, leveraging increasingly high-throughput genomic tools. We found that loci that are bivalent in ES cells tend to resolve to a univalent state upon ES cell differentiation. Induced genes become further enriched for H3K4me3 and lose H3K27me3; many noninduced genes retain H3K27me3 but lose H3K4me3. The resolution of chromatin marks at promoters that are initially bivalent appears closely related to developmental potential; for example, genes that retain H3K4me3 in neural progenitors exhibit higher expression in neural tissues.

Although bivalent domains are uniquely pervasive in ES cells, they are also observed with varying frequencies in other cell types. For example, ~10–-20 percent of genes that are bivalent in ES cells remain bivalent in neural progenitors, and many of these encode proteins with functions in neurogenesis. Our findings suggest that the relative levels of lysine 4 and lysine 27 trimethylation at key developmental loci may help balance potency and commitment throughout development.

Understanding the structure and function of bivalent domains is a major focus of our group. We have combined molecular and genetic approaches to identify the Polycomb proteins at bivalent domains in ES cells and to examine their roles in the epigenetic regulation of target loci during differentiation. Essentially all bivalent domains are occupied by Polycomb-repressive complex 2 (PRC2), which contains the lysine 27 methyltransferase. However, less than half are also occupied by PRC1, an additional Polycomb complex that catalyzes histone ubiquitinylation. Bivalent genes bound by both complexes are exquisitely enriched for developmental regulators and highly conserved between human and mouse ES cells. Furthermore, these double-positive bivalent genes retain lysine 27 trimethylation through differentiation much more efficiently than ones that just carry PRC2. This suggests that both Polycomb-repressive complexes are required to confer robust epigenetic repression of a gene locus through differentiation.

Our current research goals are (1) to determine the mechanisms that underlie the initial establishment and the resolution of bivalent domains during development, (2) to elucidate the chromatin complexes that associate with and regulate these domains, and (3) to define their higher-order organization. An increasingly detailed picture of the associated complexes and their structures will enable us to develop genetic assays and models to test more fully our hypotheses on the roles of bivalent and univalent domains in guiding transcriptional changes throughout development.

Grants from the National Human Genome Research Institute, the National Institutes of Health Roadmap Epigenomics Program, the Burroughs Wellcome Fund and the Starr Cancer Consortium provided partial support for these projects.

As of May 30, 2012

Scientist Profile

Early Career Scientist
Massachusetts General Hospital
Cell Biology