The molecular mechanisms controlling the establishment of im

The molecular mechanisms controlling the establishment of imprinting at the GNAS cluster and leading to the methylation defects in PHP1B are mostly unknown, in part because of a paucity of suitable animal models and lack of accessible Gsα-imprinted human tissues. During the murine embryonic development, the differential methylation of exon 1A (A/B in humans) and Nespas/Gnasxl (AS and XL in humans) DMRs is established during the oogenesis (germline DMRs) whereas the differential methylation of Nesp DMR occurs postfertilization (somatic DMR), with a key role played by Nesp transcription in establishing the specific-allele methylation at the Gnas locus (Chotalia et al., 2009; Coombes et al., 2003; Liu et al., 2000). A recent study analyzing a large number of human fetal gonads from gestational weeks 6.5–22 suggested that epigenetic reprogramming in human primordial germ ptio (hPGCs) probably involves, as observed in mice but with a different timing, two distinct periods: an early wave of genome-wide demethylation before 7 weeks of gestation and a later wave of imprint erasure and changes in chromatin modifications after 9 weeks of gestation (Gkountela et al., 2013; Laird 2013). Studies in hESCs and hPGCs indicated that allelic silencing of A/B is established during the gametogenesis (Frost et al., 2011) and that of XLsα already established at 5 weeks postfertilization (supporting the gametic specific-allele methylation of both A/B and XL DMRs as observed in the mice) (Crane et al., 2009). The complete allelic silencing of the NESP55 transcript occurs during implantation 5–11 weeks after fertilization (Crane et al., 2009; Rugg-Gunn et al., 2005a, b), in agreement with a somatic DMR. Tissue-specific silencing of paternal Gsα most likely takes place after 11 weeks postfertilization and after tissue differentiation (Turan et al., 2014; Zheng et al., 2001). A genome-wide DNA methylation revealed the maintenance of GNAS methylation in hiPSCs with culture, although hypermethylation and hypomethylation were also observed (Nazor et al., 2012). In an effort to document imprinting at the GNAS locus and contribute to the development of models allowing its dynamic study and tissue-specific silencing of paternal Gsα in (patho)physiological conditions in humans, we studied methylation at the four GNAS DMRs in hESCs and hiPSCs and their progenies. We also examined the expression of four GNAS transcripts (Gsα, A/B, XLsα, and NESP55) in hiPSCs and derivatives.
Results
Discussion To assess if methylation marks at the GNAS locus were maintained in hESCs and hiPSCs or subjected to variation upon derivation technique and subsequent culture, we quantified and compared methylation at the GNAS locus in hESCs and hiPSCs (four cell lines each). Our results showed that methylation at the four DMRs was similar in hESCs and hiPSCs. These results are consistent with a whole-genome single-base resolution DNA methylome study by Lister et al. (2011) reporting globally similar methylation comparing hESC and hiPSC methylomes. In addition, we found that methylation at the paternally imprinted NESP (maternal expression of the transcript) and maternally imprinted XL (paternal expression of the transcript) DMRs was maintained in all PSCs (hESCs and hiPSCs) and similar to that in controls and parental fibroblasts, in contrast to the two maternally imprinted AS and A/B DMRs. Two main conclusions can be drawn from these observations. First, previous studies have indicated that epigenetic instability is a rare occurrence in hESCs but, in contrast, that the differential methylation that marks imprinted loci could be erased during nuclear reprogramming of somatic cells (Frost et al., 2011). Analysis of germline methylation imprints in human PSCs has revealed some instability and this independently of the parental origin of the imprint (Lund et al., 2012; Nazor et al., 2012; Rugg-Gunn et al., 2007; Takikawa et al., 2013; Tobin and Kim 2012). In this regard, aberrant DNA methylation at the maternally imprinted H19 and paternally imprinted KCNQOT1 genes in iPSCs has been reported (Lister et al., 2011). Our results indicate that the control of GNAS genomic methylation imprinting stability does not vary specifically as function of the PSC type (hESCs versus hiPSCs) and is independent of the reprogramming procedure. This is further supported by the similar methylation pattern observed for two clones obtained from the same parental fibroblast either by retroviral or episomal reprogramming methods. Second, our results indicate that the control of methylation at the NESP and XL DMRs (paternally and maternally imprinted, respectively) is more stringent than that at AS and A/B DMR (both maternally imprinted). NESP DMR methylation analyzed in two studies was reported differentially methylated in the majority of hESC lines with exceptional loss or gain of methylation (Frost et al., 2011; Huntriss et al., 2011). Methylation of XL DMR reported in only two human in vitro fertilization blastocysts was variable (4.8% and 77.1%) (Huntriss et al., 2011). Our results further document and enrich these observations. In all cases of DMR methylation instability, we observed demethylation and not hypermethylation, indicating that whatever the underlying mechanism, AS and A/B are prone to demethylation during PSC derivation or maintenance. Why the control of methylation at the AS and A/B DMRs is less stringent than that at the NESP and XL DMRs in PSCs is not explained. Methylation at the A/B DMR was low in the polyclonal iPSC04603_polyF cell line and normal in the monoclonal iPSC04603_c27 line, both derived from the same parental fibroblasts, raising the possibility that clonality may affect methylation results.