Epigenetic reprogramming and DNA demethylation

Abstract

Embryonic development begins with fertilization of the egg, a progressive process that gives rise to the zygote and subsequently to the formation of somatic tissues. Normally once cells acquire a fate, it is stably maintained. Conversion back to a multipotent state occurs rarely in-vivo, but can be achieved experimentally by inducing ‘reprogramming’. In this study I am looking at the epigenetic mechanisms that underlie reprogramming and, in particular, DNA methylation and demethylation. To address this I am taking advantage of the cellular fusion system. Fusion of pluripotent cells with differentiated cells results in the formation of transient heterokaryon and hybrid cells, where the somatic partner is efficiently reprogrammed. This gives me the opportunity to monitor early and late events in pluripotent conversion, in which global remodelling of chromatin and changes in DNA methylation occur. Here, I examine changes in DNA methylation that are induced at imprinted loci and pluripotency-associated genes when somatic cells are fused with either mouse embryonic stem (ES) or embryonic germ (EG) cells. I focus on defining the factors and order of events that accompany reprogramming. I show that acquisition of pluripotency is an early process occurring at the heterokaryon stage, and is followed by imprint erasure later in hybrids. However reprogramming of imprinting is only induced by EG, but not ES cells, and it requires sequential steps of 5-methylcytosine oxidation mediated by Tet proteins and nucleotide exchange upon several rounds of DNA synthesis. I provide evidence that Tet proteins are dispensable for pluripotent reprogramming using CRISPR-Cas9 genome editing to abrogate the expression of both Tet1 and Tet2. This result suggests that either DNA demethylation can occur without TET activity (implying a redundancy with other demethylating agents and routes), or that DNA demethylation is not required for inducing pluripotency. Finally, I describe how CRISPR/Cas9 approaches were used to demonstrate that non-canonical Wnt signalling components are downstream targets of JARID2 in ES cells.