Epigenetic regulation of heterochromatin structure and tumour progression

Since the discovery of DNA packaging into chromatin, and McClintock’s (1951) work on position-effect variegation providing evidence of non-mendelian inheritance, the principal of a genome maintaining ‘on’ and ‘off’ states has been widely adopted. However, the underlying mechanisms that regulate these dynamic chromatin states and their effect on disease are still poorly understood. DNA methylation and histone trimethylation at H3K9 and H4K20 are the core hallmarks of the heterochromatic constitutively ‘off’ state. Constitutive heterochromatin is predominantly comprised of repetitive satellite containing pericentromeric regions and telomeres and in mouse heterochromatin clusters into large chromocenters. These regions are cytologically more compact and generally transcriptionally silent across embryonic and differentiated mouse cell types. However, in addition to increased genomic instability, mouse tumour cells sustain increased satellite expression suggesting constitutive heterochromatin is disrupted. Therefore how constitutive heterochromatin is maintained has important implications for genome regulation and disease, and remains poorly understood. While satellite DNA sequences are not evolutionarily conserved, pericentromeric and telomeric heterochromatin occurs across species. Heterochromatin formation is therefore independent of the underlying DNA sequence, supporting the hypothesis that epigenetic components can regulate chromatin structure. DNA methylation is generally thought to be associated with transcriptional silencing and chromatin compaction. However, Gilbert et al (2007) showed that the complete loss of DNA methylation did not affect the compaction at heterochromatin or global genome compaction. The role of H3K9me3 in regulating heterochromatin has also been an area of keen interest. H3K9me3 patterns are established by suppressor of variegation 3-9 homologues and provide the binding site for heterochromatic protein 1 [HP1] which can in turn recruit Suv39h1. This Suv3-9h-HP1-H3K9 axis enables its propagation throughout heterochromatin. Peters et al (2001) demonstrated that in mice loss of suv39 homologues 1 and 2 caused a loss of H3K9me3 at constitutive heterochromatic domains. These Suv39h null mice demonstrated decreased genome stability, and an increased prevalence of oncogenesis. However cytological chromocenters are still present in the absence of H3K9me3. Therefore the function of H3K9me3 as a causative agent in heterochromatin formation is still debated. Broadly the aim was to investigate the phenotypic role of heterochromatic epigenetic components in cancer progression, and address whether H3K9me3 effects large scale chromatin structure. To identify heterochromatic gene silencing components, an inhibitor screen was performed in an artificial silenced reporter system. The reporter fluorophore was silenced by the presence of centromeric arrays from yeast/bacterial artificial chromosomes and human alpha satellite repeats enriched for H3K9me3. To address the function of the de-silencing components identified in cancer, the fitness of colon cancer cells [HCT116] was investigated before and after the development of resistance to the MEK inhibitor trametinib. The most intriguing result was that BET protein inhibition resulted in derepression of the reporter construct and trametinib resistant HCT116 cells were more sensitive to BET inhibitors, while subsequent investigation showed HP1 protein levels were altered. Analysis of publically available datasets of tumour drug resistance, showed elevated BET protein binding at HP1 promoters in resistant cell lines suggesting an indirect role in gene silencing. To investigate the consequence of H3K9me3 loss on chromatin structure, mouse embryonic stem cells that lacked both Suv39 homologues were used. Microccocal nuclease digestion and sucrose sedimentation demonstrated a global decompaction of large-scale chromatin fibres whilst re-expression of suv39h1 rescued H3K9me3 at chromocenters and global chromatin decompaction. Loss of Suv39h also increased chromatin associated RNA levels that were also rescued by Suv39h1 re-expression. This suggests that H3K9me3 has a role chromatin fibre compaction globally as well as at constitutive heterochromatin, potentially mediated by chromatin associated RNA. To conclude, multiple components were identified that are involved in transcriptional silencing. Evaluating their function in tumour progression demonstrated a possible role of BET proteins in the development of MEKi resistance that may be mediated through HP1 proteins. H3K9me3 and its binding partner HP1 affect global chromatin compaction. The global decompaction after Suv39h loss correlates with an increase in chromatin associated RNA, suggesting a possible mechanism for changes in chromatin compaction beyond H3K9me3