br Introduction Pluripotent stem cells can generate

Introduction Pluripotent stem cells can generate all cell types of the body and hold great potential for transplantation medicine and the study of early development. Pluripotency arises in the inner cell mass of blastocyst-stage embryos during formation of the epiblast, and both human and mouse blastocysts can give rise to pluripotent embryonic stem cells (ESCs). Differentiation of the pluripotent epiblast toward the primary germ layers occurs after implantation of the embryo during the process of gastrulation. Signaling proteins belonging to the BMP and WNT families are key gastrulation factors that mediate induction of the primitive streak in the embryo and can induce primitive streak derivatives in human ESCs (hESCs) and mouse ESCs (mESCs) (Bakre et al., 2007; Blauwkamp et al., 2012; Davidson et al., 2012; Drukker et al., 2012; Gadue et al., 2006; Lako et al., 2001; Lindsley et al., 2006; Nostro et al., 2008; Sumi et al., 2008; ten Berge et al., 2008). However, BMP4 additionally induces trophoblast (Drukker et al., 2012; Xu et al., 2002), complicating efforts to obtain single lineages. Furthermore, other reports show that both BMP and WNT signals support the self-renewal of mESCs instead (Hao et al., 2006; Ogawa et al., 2006; Singla et al., 2006; ten Berge et al., 2011; Ying et al., 2003). These conflicting reports may reflect the action of BMP and WNT signals on different pluripotent states, as the epiblast of post implantation mouse embryos can also give rise to a pluripotent cell type, the epiblast stem cell (EpiSC) (Brons et al., 2007; Tesar et al., 2007). EpiSCs are developmentally more advanced than mESCs and possess different morphology, growth factor requirements, gene expression profile, and epigenetic state (Brons et al., 2007; Tesar et al., 2007). They can generate teratomas, a measure of pluripotency, but unlike mESCs are not competent to contribute to acetylcholine chloride chimeras. EpiSCs express many differentiation factors present in the primitive streak (Brons et al., 2007; Tesar et al., 2007) and were found to comprise heterogeneous populations of cells with distinct potency (Bernemann et al., 2011; Tsakiridis et al., 2014). This suggests that EpiSCs are to some extent prespecified, and their pluripotent state has therefore been designated “primed,” as opposed to the unspecified “naïve” pluripotent state of mESCs (Nichols and Smith, 2009). Similar observations were made for hESCs, consistent with them occupying a primed pluripotent state (Blauwkamp et al., 2012; Davidson et al., 2012; Drukker et al., 2012; Stewart et al., 2006). Interestingly, for both EpiSCs and hESCs, it has been shown that endogenous WNT proteins, produced by the cells themselves, drive prespecification of the cells (Blauwkamp et al., 2012; Frank et al., 2012; Sumi et al., 2013; Tsakiridis et al., 2014).
Results
Discussion This work shows that endogenous WNT signals are major hidden factors in the differentiation of hESCs and EpiSCs and affect the outcome of directed differentiation protocols in hitherto unappreciated ways. We show that WNT signals induce the main gastrulation factors Nodal, Wnt3, and Fgf8 and are required and sufficient for the induction of mesoderm by the commonly used mesoderm inducer BMP4. A surprising finding is that BMP4 induces both mesoderm as well as trophoblast-committed cells from hESCs, but only mesoderm induction requires the activation of WNT genes by BMP4. These findings are summarized in Figure 6, and they have obvious ramifications for the guided differentiation of hESCs. For instance, to induce trophoblast one should stimulate with BMP4 in the presence of a WNT inhibitor to avoid induction of mesoderm. Conversely, mesoderm is best obtained using WNT3A in lieu of BMP4 to avoid trophoblast induction. Furthermore, WNT-inhibited hESCs differentiate more efficiently to the trophoblast lineage, suggesting that genuine EpiSCs and hESCs, maintained as homogeneous undifferentiated populations by WNT inhibition, are superior substrates for differentiation as they contribute fewer undesired lineages to the population. We also find that different cell lines and culture media display various tendencies for endogenous WNT-induced differentiation, affecting their suitability for specific purposes such as mesoderm or neural differentiation. This may also influence to what extent WNT inhibition supports their self-renewal or improves their subsequent differentiation. Another interesting observation is that spontaneous endogenous WNT signals induce endoderm in EpiSCs, consistent with the finding that low levels of WNT3A (25 ng/ml) induce definitive endoderm (D’Amour et al., 2006), whereas high levels of WNT3A (250 ng/ml) induce mesoderm. This may reflect a later function for WNT in redirecting primitive streak-specified cells from endoderm to mesoderm (Loh et al., 2014). Finally, ESCs are commonly aggregated into EBs for the derivation of mesendodermal lineages. We now show that EBs mediate the transition of mESCs into EpiSCs, which is followed by activation of an endogenous WNT gradient and primitive streak induction (ten Berge et al., 2008). A more controlled way of inducing mesendodermal lineages would be by directly inducing genuine EpiSCs with defined levels of WNT signals.