Keep the fate: how chromatin regulators safeguard embryonic stem cell identity

Segregation of cells that form the embryo from those that produce the surrounding extra‐embryonic tissues is critical for early mammalian development, but the regulatory layers governing these first cell fate decisions remain poorly understood. Recent work in The EMBO Journal identifies two chromatin regulators, Hdac3 and Dax1, that synergistically restrict the developmental potential of mouse embryonic stem cells and act as a lineage barrier to primitive endoderm formation.

Recent work reports histone deacetylase Hdac3 and nuclear receptor Dax1 as lineage barriers to primitive endoderm formation.

In the mammalian embryo, the first cell fate specification event segregates outer trophectoderm placental progenitors from the inner cell mass, forming a structure termed the blastocyst. The inner cell mass subsequently separates into the pluripotent epiblast (EPI) cells that will give rise to all cells of the embryo proper, and primitive endoderm (PE) cells which will generate the yolk sac (Fig 1A) (reviewed in Rossant & Cross, 2001).

Figure 1. Histone deacetylase Hdac3 and nuclear receptor Dax1 as lineage barriers to primitive endoderm formation in mESCs.

(A) The first three lineages in the mouse preimplantation embryo at the blastocyst stage. (B) The transition from mESCs to PE‐like cells can be induced by deleting a pluripotency‐associated gene, Nanog, or by overexpression of the endoderm‐associated gene Gata6. In addition, HDAC3 and DAX1 act as additional barriers for this transition, safeguarding the EPI identity of mESCs. (C) The genetically engineered reporter mESC lines used by Olivieri et al (2021). The EPI‐specific gene Nanog is tagged with a green fluorescent protein (GFP), and the PE‐specific gene Gata6 is coupled with a red fluorescent protein (mCherry). Cells in the EPI‐like state activate the GFP reporter only, while cells in the PE‐like state activate only mCherry. (D) The molecular mechanism by which HDAC3 safeguards the identity of mESCs. By binding to Gata6 enhancer region—enh^‐45, HDAC3 inhibits the activation of the Gata6 gene which is required to activate PE genes. Removal of HDAC3 allows Gata6 expression and its binding to cis‐regulatory elements of PE genes, thus enabling induction of PE‐like cell identity.

Notably, self‐renewing stem cell types can be used to study the transition between EPI and PE cells in vitro. For example, mouse embryonic stem cells (mESCs) isolated from the embryo inner cell mass can be maintained in culture and driven into a cellular identity that resembles the PE lineage. PE‐like cells can either arise spontaneously in mESC cultures or can be induced either by the deletion of EPI‐specific genes such as Nanog (Mitsui et al, 2003) or by the overexpression of endoderm‐associated genes such as Gata6 (Shimosato et al, 2007; Morgani et al, 2013; Wamaitha et al, 2015; lo Nigro et al, 2017) (Fig 1B). Importantly, because these stem cells capture the developmental potential of their counterparts in the embryo, they are a useful model to study early mammalian development and the principles underlying cell fate choice and maintenance.

Cell fate decisions are the result of the activity of specific sets of genes. Multiple layers of regulation ensure the proper execution of gene expression programmes that “induce and seal the fate” of a cell. These include transcriptional control at the chromatin level, from histone acetylation associated with active genes, down to binding of transcription factors on accessible cis‐regulatory elements, which in turn modulate the activity of individual genes (reviewed in Klemm et al, 2019). However, which chromatin regulators mediate the binary decision of stem cells between EPI and PE fate in the inner cell mass and prevent aberrant lineage acquisition is still unclear.

In this study, Olivieri et al (2021) set out to investigate a part of this complex transcriptional regulatory network that controls the transition between EPI‐like and PE‐like cells using mESCs as an in vitro model of mouse embryonic development. The authors engineered these cells so that key genes for each of the two lineages are coupled with a fluorescent reporter. The genes were Nanog, active in EPI‐like cells and tagged with a green fluorescent protein (GFP), and Gata6—active in PE‐like cells and joined to the red fluorescent protein mCherry (Fig 1C). This enabled the authors to identify the barriers for the emergence of PE‐like cells after subjecting the cells to different culture conditions and after performing CRISPR/Cas9 loss‐of‐function experiments.

With this system, Olivieri et al (2021) tested whether a chromatin regulator, the histone deacetylase HDAC3, acts as an epigenetic barrier on the transition from mESCs to PE‐like cells in mESC culture. The authors deleted the gene that encodes for HDAC3 in the reporter mESC line and switched the cells to culture conditions that promote heterogeneity or pluripotency exit and EPI priming. This revealed that in contrast to their wild‐type counterparts, cells with the deleted Hdac3 gene (Hdac3 ^−/−) transcriptionally resembled the PE of the embryo, activating known endoderm regulators such as Gata4, Gata6 and Sox17. In addition, Hdac3 ^−/− cells had the ability to form 3D structures called spheroids that contained EPI‐like cells surrounded by PE‐like cells—a composition similar to that present in the blastocyst. These results suggested that HDAC3 acts as a barrier for the activation of primitive endoderm‐like cell identity in mESCs.

In addition to the removal of Hdac3, the authors also confirmed that the perturbation of a previously reported inhibitor of PE—DAX1 (Niakan et al, 2006), also enables the transition to PE‐like cells. Interestingly, when both the Hdac3 and Dax1 genes were deleted in mESCs, the conversion rate to the PE‐like state increased considerably, suggesting that removing both factors reduces the threshold required for the lineage transition. Noteworthy, these factors are shown to act through independent pathways, HDAC3 cooperating with NCOR1/NCOR2 and DAX1 forming a complex together with NR5A2 and ESRRB.

To test whether the resulting PE‐like cells recapitulate the developmental potential of PE cells in the embryo, the authors injected both Hdac3‐ and Dax1‐depleted cells into blastocysts at embryonic day (E) 3.5. Both of the induced PE‐like cell lines localized together with Sox17‐expressing cells in chimeric embryos and expressed PE markers, showing that these cells behave similarly to the PE in the embryo. The importance of these findings is twofold: first, HDAC3 is a barrier that safeguards the lineage restriction of mESCs and the transition from EPI to PE cells can be modelled in vitro by removing this barrier, and second, there might be multiple, independent pathways promoting induction of PE identity.

To better understand how the two independent complexes elicit their lineage restricting function, the authors performed epigenomic analyses and identified cis‐regulatory elements that were bound by both HDAC3 and DAX1. The co‐regulated cis‐regulatory elements activated in mutant cells also contained transcription factor binding motifs for GATA factors, while regulatory elements inactivated in compound mutants contained motifs for OCT4‐SOX2. This suggests that Hdac3 and Dax1 share the regulatory landscape with core pluripotency factors. Following leads that implicate GATA6 in PE specification, the authors focused on a known enhancer element (enh^‐45) upstream of GATA6. This element is found among the most strongly activated cis‐regulatory elements in the double mutants, while in the embryo, this element is active in E6.5 visceral endoderm but not in the EPI. To test whether these cis‐regulatory elements are indeed implicated in the conversion to the PE state, the authors deleted the enh^‐45 region in Hdac3 ^−/−; Dax1 ^−/− mESCs, demonstrating that cells were no longer able to cross the lineage restriction boundary. Therefore, Hdac3 and Dax1 prevent the activation of the PE transcriptional programme by suppressing a specific cis‐regulatory element of Gata6 (Fig 1D).

The findings of Olivieri et al (2021) further enhance our understanding of how lineage specification during early mammalian development is regulated and how chromatin pathways may restrict the conversion from EPI to PE cells. Nevertheless, there are still outstanding questions regarding the EPI‐PE fate choice. What other chromatin remodelling factors are involved in these initial cell fate decisions? Although the authors examine the role of one member of the HDAC deacetylating enzymes family, other members might also play a role, which still needs to be elucidated. In addition, given that other repressive chromatin remodelling enzymes, such as the Polycomb repressive complex 2 (PRC2), are reported to restrict lineage priming towards PE (Illingworth et al, 2016), it would be interesting to investigate whether HDAC and PRC2 complexes exhibit synergistic effects in restricting the plasticity of mESCs. Although the role of HDAC3 and DAX1 in PE fate specification is well supported here, it remains unclear whether PE‐like cells generated in vitro by removing Hdac3 or Dax1 can give rise to PE‐derivative tissues that can be maintained at post‐implantation stages. This would serve as a definitive confirmation that the resulting PE‐like cells fully recapitulate the developmental potential of PE cells in the embryo.

Collectively, Olivieri et al (2021), with a remarkable tour‐de‐force, provide detailed evidence on the synergistic action of HDAC3 and DAX1 in controlling the segregation of embryonic EPI and extra‐embryonic endoderm cells in mESCs.

Conflict of interest

V.P. is an adviser to LifeSci Venture Partners.

The EMBO Journal (2021) 40: e108437.