Whereas we found that Blimp

Whereas we found that Blimp-1 deletion induced pluripotency in PGCs, the effects of Blimp-1 overexpression in pluripotent order Staurosporine appear to be more complicated. Forced expression of Blimp-1 in ESCs induces growth retardation (Nagamatsu et al., 2011). During PGC induction from ESCs, induction of an intermediate cell state, namely epiblast-like cells (EpiLCs), is important (Hayashi et al., 2011). The combination of transcription factors Prdm14, Blimp-1, and Tfap2c is critical to induce PGCs from EpiLCs (Nakaki et al., 2013). However, forced induction of these three factors cannot induce PGCs directly from ESCs. Furthermore, whereas Prdm14 alone can induce PGCs with low efficiency, Blimp-1 alone cannot even induce PGCs from EpiLCs. Overexpression of Prdm14 in ESCs enhances pluripotency maintenance but does not induce PGC differentiation (Okashita et al., 2014). These facts reveal that there are important differences between ESCs and EpiLCs in relation to PGC induction. One such difference is the existence of a suppressive mechanism in ESCs. Recently, it was reported that inactivation of Myc induces upregulation of germ cell marker genes in ESCs (Maeda et al., 2013). This indicates that PGC induction is suppressed in ESCs. However, in that report, Vasa was expressed much earlier than in vivo, and early markers, such as Blimp-1, were not activated efficiently. Therefore, it is unclear whether Myc inactivation induces functional differentiation. It would be intriguing to analyze differences between ESCs and EpiLCs in relation to the prerequisites for PGC differentiation. To our surprise, overexpression of Dnd enhanced downstream targets of pluripotency (Figure 3). When Dnd is inactivated, the number of PGCs is decreased but basal PGCs give rise to teratomas in vivo (Youngren et al., 2005). Teratomas contain three germ layers generated by pluripotent cells. Therefore, deletion of Dnd leads to pluripotency in PGCs. However, in the present work, Dnd did not appear to mediate the suppression of pluripotency in ESCs. Dnd is not a target gene of either BLIMP-1 or AKT (Dataset S1), indicating that Dnd is regulated differently from BLIMP-1 or AKT. The mechanisms underlying pluripotency acquisition upon Dnd deletion, and the relationship between Dnd and Blimp-1 or Akt in PGCs is an important issue to be investigated. In this study, we found a synergistic effect of AKT activation in the presence of bFGF and 2i+A83 on the acquisition of pluripotency. AKT activation suppressed MBD3-regulated genes in PGCs (Figure 7B). Furthermore, AKT activation downregulated MBD3 in MEFs (Figure 7G). Both AKT activation and MBD3 inactivation have been shown to prevent differentiation of ESCs in the absence of LIF (Watanabe et al., 2006; Kaji et al., 2006). It is conceivable that AKT downregulates MBD3 and thereby maintains pluripotency in the absence of LIF. Mbd3 is a roadblock of pluripotency (Rais et al., 2013). However, the regulation of Mbd3 expression is poorly understood. It would be interesting to analyze how AKT signaling downregulates Mbd3. It has been reported that a histone deacetylase (HDAC) inhibitor had a positive effect on EGC formation. We therefore analyzed AKT activation and HDAC target genes. For this purpose, we determined the genes that were upregulated in Hdac-deficient ESCs (Jamaladdin et al., 2014). GSEA analysis was performed at day 2 of the culture with or without AKT activation (Figure S3F). Whereas bFGF alone activated HDAC target genes, AKT activation suppressed this gene set, indicating that AKT enhances the formation of EGCs in a manner distinct from that of the HDAC pathway. In this study, we used target gene set analysis. We considered that this approach would allow us to understand the gene network and epigenetic state, which are difficult to analyze based on the individual gene expressions. Whereas both PGCs and EGCs express Oct3/4, the downstream targets of OCT3/4 were repressed from days 1 to 10 (Figures 2 and S1A). This indicated that the region downstream of OCT3/4 might differ between PGCs and EGCs. It was previously shown that Oct3/4 plays a critical role in PGC specification (Okamura et al., 2008). It would thus be of interest to investigate the molecular interaction between Oct3/4 and the factors that are important for PGC specification, such as Blimp-1 and Prdm14. We also applied this approach to epigenetic modifications. First, we analyzed the histone modification-associated active genes (H3K36me3, H3K79me2, and H3K4me3) (Marson et al., 2008; Mikkelsen et al., 2007). The targets of these modifications were also gradually upregulated (Figure S1B). On the other hand, another set of modification targets consisting of H3K4me3, H3K27me3, or both (i.e., the Bivalent domain targets) showed intriguing change (Figure S1C). Whereas the targets of the Bivalent domain were upregulated in EGCs compared with early culture periods, the H3K27me3 targets were repressed. The targets of H3K4me3 were maintained in an active state. With respect to the targets of DNA methylation, three different gene sets of 5hmC and two different gene sets of 5meC were collected from three different papers (Pastor et al., 2011; Borgel et al., 2010; Guibert et al., 2012). Except for one time point (EGCs of Figure S1E), both the 5hmC and 5mC targets were activated from the early phase of the culture (Figures S1D–S1F). PGCs contain DNA with a low level of methylation (Seki et al., 2005). Therefore, it is feasible that the targets of DNA methylation in ESCs are already hypomethylated in PGCs and so easy for the early activation.