Impact of DNA hypomethylation on differentiation trajectories in mouse embryonic stem cell in vitro differentiation models

Abstract

Epigenetic chromatin modifications such as DNA methylation are important during mammalian embryonic development for reinforcing lineage-specific gene expression as cells differentiate. Parental DNA methylation patterns from the egg and sperm are erased following formation of the zygote, leading to low levels of methylation in the pluripotent embryonic stem cells of the blastocyst’s inner cell mass. When these mouse embryonic stem cells (mESCs) are grown in culture, they are surprisingly tolerant of perturbations to epigenetic regulatory systems. Constitutive DNA methylation as well as total abrogation of DNA methylation have been reported to have minimal impact on viability, self-renewal or pluripotency gene expression in mESCs. However, when cells start to differentiate, loss of DNA methylation becomes incompatible with viable embryonic development. Mouse embryos with catalytic mutations in the maintenance DNA methyltransferase Dnmt1 die at embryonic day 9.5, just after the gastrulation stage. The question remains both when and why do cells become dependent on DNA methylation during development.

Stem cell-derived in vitro models of early development have enabled a shift away from the requirement of real embryos to study these transitions – for example, three-dimensional gastruloid models mimic aspects of mouse development, recapitulating many features of axial tissue patterning. In my PhD, I aimed to use this system to investigate how the loss of DNA methylation impacts on developmental progression under transcriptional depletion or chemical inhibition of maintenance methyltransferase enzyme DNMT1. Like their in vivo counterparts, I found that hypomethylated gastruloids exhibit a delay similar to the developmental arrest around the gastrulation stage in DNMT1-defective embryos and fail to establish consistent symmetry-breaking and developmental patterning. Transcriptional analysis revealed persistent expression of pluripotency gene networks in hypomethylated gastruloids in addition to activation of DNA methylation-sensitive genes. Unexpectedly, I also found gene signatures indicative of both primitive and definitive erythropoiesis appearing at later stages of gastruloid development in hypomethylated conditions, with implications for the cellular origins of these two different blood lineages.

To validate these findings under a pharmaceutical means of inducing DNA hypomethylation I used a newly discovered inhibitor of DNMT1, GSK-3484862. Characterisation of this inhibitor in mESCs revealed faithful induction of DNA hypomethylation at a range of concentrations and treatment durations, with effects dependent on the presence of DNMT1. The inhibitor showed minimal impact on pluripotency in mESCs, however it appeared to have a greater impact on mESC proliferation than observed with genetic depletion of Dnmt1.

Gastruloids cultured with the DNMT1 inhibitor showed similar phenotypes to those generated by transcriptional depletion of Dnmt1. Furthermore, gene expression analysis indicated similar insufficient repression of genes associated with earlier stages of gastruloid development and ectopic formation of erythropoietic signatures at late stages.

Overall, my results support the view that DNA methylation is required to facilitate correct patterns of gene expression during post-blastocyst development, and in its absence this process is skewed. This provides insight into the developmental processes most tolerant of global DNA hypomethylation to understand more about the differentiation of these lineages and the shift towards dependence on DNA methylation in the embryo as a whole.