Preview

PDF, English Download (70MB) | Terms of use

Abstract

Epigenetic systems contribute to genome regulation in health and disease, and underpin chromatin-based memory. The epigenome is therefore tightly regulated but evidence has emerged that altered epigenetic states (epialleles) can be induced in response to perturbations or environmental stimuli. Moreover, these can be mitotically or meiotically heritable in yeast, worms and plants, driving phenotypes independent of the genotype. Nevertheless, the prevalence and significance of epialleles in mammals remains unclear. To investigate the potential for transmission of acquired chromatin states in mammals, I optimised a modular and releasable dCas9 system coupled with KRAB epigenetic repressor to programme epigenetic states to endogenous loci in an ESC model of development. With this tool, I was able to induce de novo heterochromatin domains comprising physiologically relevant levels of DNA methylation, H3K9me3, H4K20me3, with concurrent loss of H3K4me3, leading to absolute silencing of local genes at the single-cell level. Despite the extant paradigm predicting that such major heterochromatin regions are epigenetically transmitted, here I observed that deposited heterochromatin domains exhibit only transient memory function, which is rapidly reverted with time and DNA replications in pluripotent cells. By loss of function genome-wide CRISPR screening coupled with my epigenetic memory assay, I found that Dppa2 is specifically responsible for counteracting epigenetic memory of epialleles. DPPA2 is a small protein that together with DPPA4 binds to most GC-rich gene promoters and is exclusively expressed in pluripotent cells. I found that deletion of Dppa2 enables robust epigenetic memory of programmed heterochromatin in ESC, without influencing cell identity. Furthermore unlike ESC, I observed that epigenetic memory is maintained in wildtype lineage-restricted cells, which do not express Dppa2/4, under selective conditions that favour the epiallele. This includes stable inheritance of epigenetic silencing at the tumour suppressor gene p53 in cell subpopulations, in both in vitro and in vivo assays. This result provides a proof of principle that epimutation of genes that facilitate a selective advantage, such as p53, can be inherited during organogenesis in vivo, with implications for predisposition to diseases including cancer. I propose this reflects the synergistic influences of weak- acting epigenetic inheritance and positive epiallele selection. This may be relevant to multiple gene-environment contexts in mammals and has relevance to the concept of ‘soft inheritance’.