Differentiation shifts from a reversible to an irreversible heterochromatin state at the DM1 locus.

Epigenetic defects caused by hereditary or de novo mutations are implicated in various human diseases. It remains uncertain whether correcting the underlying mutation can reverse these defects in patient cells. Here we show by the analysis of myotonic dystrophy type 1 (DM1)-related locus that in mutant human embryonic stem cells (hESCs), DNA methylation and H3K9me3 enrichments are completely abolished by repeat excision (CTG2000 expansion), whereas in patient myoblasts (CTG2600 expansion), repeat deletion fails to do so. This distinction between undifferentiated and differentiated cells arises during cell differentiation, and can be reversed by reprogramming of gene-edited myoblasts. We demonstrate that abnormal methylation in DM1 is distinctively maintained in the undifferentiated state by the activity of the de novo DNMTs (DNMT3b in tandem with DNMT3a). Overall, the findings highlight a crucial difference in heterochromatin maintenance between undifferentiated (sequence-dependent) and differentiated (sequence-independent) cells, thus underscoring the role of differentiation as a locking mechanism for repressive epigenetic modifications at the DM1 locus.

EpiTyping: analysis of epigenetic aberrations in parental imprinting and X-chromosome inactivation using RNA-seq.

Human pluripotent stem cells (hPSCs) hold a central role in studying human development, in disease modeling and in regenerative medicine. These cells not only acquire genetic modifications when kept in culture, but they may also harbor epigenetic aberrations, mainly involving parental imprinting and X-chromosome inactivation. Here we present a detailed bioinformatic protocol for detecting such aberrations using RNA sequencing data. We provide a pipeline designed to process and analyze RNA sequencing data for the identification of abnormal biallelic expression of imprinted genes, and thus detect loss of imprinting. Furthermore, we show how to differentiate among X-chromosome inactivation, full activation and aberrant erosion of X chromosome in female hPSCs. In addition to providing bioinformatic tools, we discuss the impact of such epigenetic variations in hPSCs on their utility for various purposes. This pipeline can be used by any user with basic understanding of the Linux command line. It is available on GitHub as a software container ( https://github.com/Gal-Keshet/EpiTyping ) and produces reliable results in 1-4 d.

Differentiation of uniparental human embryonic stem cells into granulosa cells reveals a paternal contribution to gonadal development.

Genomic imprinting underlies the mammalian requirement for sexual reproduction. Nonetheless, the relative contribution of the two parental genomes during human development is not fully understood. Specifically, a fascinating question is whether the formation of the gonad, which holds the ability to reproduce, depends on equal contribution from both parental genomes. Here, we differentiated androgenetic and parthenogenetic human pluripotent stem cells (hPSCs) into ovarian granulosa-like cells (GLCs). We show that in contrast to biparental and androgenetic cells, parthenogenetic hPSCs present a reduced capacity to differentiate into GLCs. We further identify the paternally expressed gene IGF2 as the most upregulated imprinted gene upon differentiation. Remarkably, while IGF2 knockout androgenetic cells fail to differentiate into GLCs, the differentiation of parthenogenetic cells supplemented with IGF2 is partly rescued. Thus, our findings unravel a surprising essentiality of genes that are only expressed from the paternal genome to the development of the female reproductive system.

Large-Scale Analysis of X Inactivation Variations between Primed and Naïve Human Embryonic Stem Cells.

X chromosome inactivation is a mammalian dosage compensation mechanism, where one of two X chromosomes is randomly inactivated in female cells. Previous studies have suggested that primed human embryonic stem cells (hESCs) maintain an eroded state of the X chromosome and do not express XIST, while in naïve transition, both XIST and the eroded X chromosome are reactivated. However, the pattern of chromosome X reactivation in naïve hESCs remains mainly unknown. In this study, we examine the variations in the status of X chromosome between primed and naïve hESCs by analyzing RNA sequencing samples from different studies. We show that most samples of naïve hESCs indeed reactivate XIST and there is an increase in gene expression levels on chromosome X. However, most of the naïve samples do not fully activate chromosome X in a uniform manner and present a distinct eroded pattern, probably as a result of XIST reactivation and initiation of re-inactivation of chromosome X. This large-scale analysis provides a higher-resolution description of the changes occurring in chromosome X during primed-to-naïve transition and emphasizes the importance of taking these variations into consideration when studying X inactivation in embryonic development.

Genome-wide analysis of haploinsufficiency in human embryonic stem cells.

Haploinsufficiency describes a phenomenon where one functioning allele is insufficient for a normal phenotype, underlying several human diseases. The effect of haploinsufficiency on human embryonic stem cells (hESC) has not been thoroughly studied. To establish a genome-wide loss-of-function screening for heterozygous mutations, we fuse normal haploid hESCs with a library of mutant haploid hESCs. We identify over 600 genes with a negative effect on hESC growth in a haploinsufficient manner and characterize them as genes showing less tolerance to mutations, conservation during evolution, and depletion from telomeres and X chromosome. Interestingly, a large fraction of these genes is associated with extracellular matrix and plasma membrane and enriched for genes within WNT and TGF-ß pathways. We thus identify haploinsufficiency-related genes that show growth retardation in early embryonic cells, suggesting dosage-dependent phenotypes in hESCs. Overall, we construct a unique model for studying haploinsufficiency and identified important dosage-dependent pathways involved in hESC growth and survival.