Introduction The degree to which embryonic stem cells

Introduction The degree to which embryonic stem glycosylase (ESCs) represent generic properties of pluripotent founder cells in mammalian embryos is unresolved (Nichols and Smith, 2012; Smith, 2001). Mouse ESCs are the paradigmatic model. However, it is increasingly clear that there are differences in early embryos and derivative stem cells among mammals (Nichols and Smith, 2009; Roode et al., 2012; Rossant, 2008). In this context, rat ESCs provide a useful comparator for interrogating pluripotency in vitro and seeking to extract generic principles. ESCs from mouse and rat can be derived and maintained using the cytokine leukemia inhibitory factor (LIF) in combination with two small molecule inhibitors (2i) that block the mitogen activated protein kinase (MAPK/ERK) pathway and reduce the activity of GSK3 (Buehr et al., 2008; Kawamata and Ochiya, 2010; Li et al., 2008; Nichols et al., 2009b; Ying et al., 2008). Rat ESCs can colonize chimeras and pass through the germline, thereby fulfilling the functional criteria for naive pluripotent stem cells. However, rat ESCs differ from mouse ESCs in a propensity to undergo unscheduled differentiation, which can lead to complete collapse of cultures (Blair et al., 2012). The interplay between extrinsic regulators and the transcriptional circuitry that governs pluripotent stem cell self-renewal is incompletely understood (Chen et al., 2008; Niwa, 2007). Mouse and rat ESCs appear to express similar core pluripotency factors that are central to establishing and maintaining the naive pluripotent state (Blair et al., 2011). However, rat ESCs also express lineage determination factors (Hong et al., 2013) that are suppressed in mouse ESCs cultured in 2i (Marks et al., 2012). Here, we explore the inappropriate expression of lineage-specifying transcription factors in undifferentiated rat ESCs. We reveal an underlying mechanism that can be counteracted to stabilize self-renewal.
Results
Discussion Rat ESCs can be derived with similar high efficiency to mouse ESCs using 2i with LIF, yet they are more prone to differentiation during expansion (Blair et al., 2011). The present findings indicate that differentiation is triggered by overinhibition of GSK3 and can be suppressed by fine-tuning the inhibitor concentration. The altered sensitivity of rat ESCs appears to arise from higher expression of Lef1. However, different thresholds for β-catenin/TCF3-mediated derepression of pluripotency factors versus β-catenin/LEF1-mediated upregulation of differentiation factors enable precise titration of GSK3 inhibition to favor self-renewal. Coexpression of CDX2 and OCT4 in rat ESCs is surprising. However, other trophectoderm lineage markers are not appreciably expressed and trophoblast-like cells are not seen. Therefore, CDX2 is not sufficient to activate a trophoblast differentiation program in rat ESCs, in contrast to findings from mouse ESC overexpression (Niwa et al., 2005). Furthermore, OCT4 levels are not significantly reduced by the presence of CDX2. This may be attributable to the presence of 2i. Alternatively, the reciprocal inhibition circuit between Oct4 and Cdx2 may be specific to mouse (Berg et al., 2011). We surmise that the inhibition of GSK3 should be precisely tuned such that intracellular β-catenin levels are sufficient to remove TCF3 from chromatin (Shy et al., 2013) but not to engage appreciably with other TCF/LEF factors. With optimal inhibition, TCF3 targets that contribute to ESC self-renewal, such as Esrrb, are fully derepressed but activation of lineage specification genes is minimal. This model is consistent with findings of differential effects of Tcf3 and Tcf1 in mouse ESCs (Yi et al., 2011). Indeed, mouse ESC self-renewal efficiency declines at CH concentrations higher than 3 μM (Ying et al., 2008), and in elevated concentrations of CH, mouse ESCs show upregulation of T, Cdx1, and Cdx2 (Figure S2). Notably, mouse ESCs totally deficient in GSK3 (Doble et al., 2007) can self-renew without CH but accompanied by continuous differentiation (Ying et al., 2008).