FGF/MAPK signaling sets the switching threshold of a bistable circuit controlling cell fate decisions in ES cells

Abstract

Intracellular transcriptional regulators and extracellular signaling pathways together regulate the allocation of cell fates during development, but how their molecular activities are integrated to establish the correct proportions of cells with particular fates is not known. Here we study this question in the context of the decision between the epiblast (Epi) and the primitive endoderm (PrE) fate that occurs in the mammalian preimplantation embryo. Using an embryonic stem (ES) cell model, we discover two successive functions of FGF/MAPK signaling in this decision. First, the pathway needs to be inhibited to make the PrE-like gene expression program accessible for activation by GATA transcription factors in ES cells. In a second step, MAPK signaling levels determine the threshold concentration of GATA factors required for PrE-like differentiation, and thereby control the proportion of cells differentiating along this lineage. Our findings can be explained by a simple mutual repression circuit modulated by FGF/MAPK signaling. This may be a general network architecture to integrate the activity of signal transduction pathways and Fate Decisions in ES Cells 2 transcriptional regulators, and serve to balance proportions of cell fates in several contexts.

Dynamics of lineage commitment revealed by single-cell transcriptomics of differentiating embryonic stem cells

“…Addition of MEKi alone, however, led to a specific reduction of the XEN-like subpopulation ( Supplementary Fig. 7c-e), in agreement with previous results 26,27 . This effect was unlikely due to interference with RA signaling since increasing RA concentration did not reverse the effect ( Supplementary Fig.…”

“…The epiblast lineage then further develops to neuroectoderm-like cells by 96 h. A recently published study by Klein et al used single-cell RNA-seq to characterize mESC differentiation by LIF withdrawal 41 and also found a small XEN-like subpopulation. That and other studies 42,27 show that XEN-like cells occur more generally in in vitro differentiation of mESCs and are not an idiosyncratic artefact of exposure to RA. We also found that mESCs grown in 2i/L (but not in serum and LIF) efficiently generate XEN cells under RA exposure ( Supplementary Fig.…”

“…5d). A similar GRN has been used successfully before in a report by Schröter et al, studying induction of the XEN lineage by exogenous Gata4 expression 27 . Importantly, our model does not strictly require an ectoderm bias.…”

“…5a-c). To explore the role of an intrinsic epiblast or ectoderm bias we developed a simple phenomenological model based on a minimal gene regulatory network (GRN) 27,32 . Briefly, the GRN is comprised of two lineage-specific, auto-activating expression programs that mutually repress each other (Fig.…”

“…7a, b). Similarly, Schröter et al have observed, for a different differentiation assay, that pre-culture in 2i/L greatly increases the ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/s41467-017-01076-4 number of XEN-like cells generated from mESCs 27 . Together with those results our observations thus support a model in which mESCs grown in 2i/L functionally correspond to a slightly earlier developmental stage than mESCs grown in serum and LIF 42 .…”