Epigenetic control of cell-fate in mouse blastocysts: role of covalent histone modifications and chromatin remodeling

Summary

The first cell fate decision in mammalian preimplantation embryos is the segregation of the inner cell mass (ICM) and trophectoderm (TE) cell lineages. The ICM develops into the embryo proper, whereas the TE ensures embryo implantation and is the source of the extra-embryonic trophoblast cell lineages, which build the functional co-units of the placenta. The development of a totipotent zygote into a multilineage blastocyst is associated with the generation of distinct transcriptional programs. Several key transcription factors participate in the ICM and TE-specific transcriptional networks and recent studies indicate that post-translational histone modifications as well as ATP-dependent chromatin remodeling complexes converge with these transcriptional networks to regulate ICM and TE lineage specification. This review will discuss our current understanding and future perspectives related to transcriptional and epigenetic regulatory mechanisms that are implicated in first mammalian lineage commitment in mice.

Introduction

The window of preimplantation embryo development encompasses a continuum of molecular, cellular, and morphological events, which transform a totipotent zygote into a multilineage blastocyst that will implant and develop into an offspring. A functional blastocyst is comprised of a fluid filled cavity enclosed by three distinct cellular lineages: the trophectoderm epithelium (TE), the epiblast, and the primitive endoderm (PE). These lineages are established via two sequential cell-fate decisions that occur during blastocyst formation. The first cell-fate decision involves the differentiation of a 16–32 cell embryo into two cell populations: the outer TE and the inner cell mass. The second cell-fate comprises the segregation of the ICM into the epiblast and the PE. The TE is required for implantation into the uterus and formation of the placenta. The cells that make up the epiblast are “pluripotent” and have the capacity to give rise to all of the tissues and organs in the body, while, the cells in PE are important for establishment of the yolk sac. Each of these differentiation events are tightly regulated by transcriptional and epigenetic mechanisms. The proper execution of these early cell-fate decisions is critical for implantation, placentation, and successful development to term. The underlying cellular and molecular processes that accompany blastocyst formation, and the working models describing the origin of the ICM and TE cell populations have been extensively reviewed elsewhere (Cockburn and Rossant 2010; Oron and Ivanova 2012; Zernicka-Goetz et al. 2009).

In the present article we discuss the emerging body of literature describing epigenetic mechanisms (e.g., histone modifications and chromatin remodeling) that govern the first cell-fate decision, ICM and TE segregation, in a mouse model for preimplantation embryo development. Furthermore, relevant mechanistic studies in ES cells, a cell reference system for the pluripotent ICM, will be discussed. Finally, the mechanisms by which epigenetic regulatory machinery converge with pivotal transcriptional networks to modulate gene expression and facilitate specification of the ICM and TE lineages will be discussed.

Cellular and morphological events associated with segregation of the blastocyst ICM and TE

The precise stage of mouse preimplantation development when blastomeres begin to commit and adopt an ICM and TE cell-fate is still under debate (Dietrich and Hiiragi 2007; Piotrowska-Nitsche et al. 2005; Ralston and Rossant 2008; Torres-Padilla et al. 2007). The general consensus is that up until the 4- to 8-cell stage all blastomeres are totipotent. This view is based on the capacity of isolated blastomeres to contribute to both the ICM and TE lineage in chimeric embryos (Kelly 1977). At the 8-cell stage individual blastomeres begin to polarize and undergo compaction through cell-to-cell contact mediated by E-cadherin (CDH1) (Fig. 1) (De Vries et al. 2004; Ducibella et al. 1977; Larue et al. 1994; Riethmacher et al. 1995; Vestweber and Kemler 1984). The cell polarity proteins Partitioning defective homologue 3 (PARD3), PARD6B, and Protein kinase C zeta (PKCζ) are important for establishment of cell polarity and cell allocation during preimplantation development (Alarcon 2010; Plusa et al. 2005; Vinot et al. 2005). In mammalian cells these proteins form a complex and localize to the apical membranes where they interact with Cell division cycle 42 (CDC42) and tight junction (TJ) proteins (Hurd et al. 2003). Then around the 16- to 32-cell stage the blastomeres begin to adopt outside and inside cell-fates with the inside and outside cells eventually becoming ICM and TE, respectively (Fig. 1) (reviewed by (Johnson and McConnell 2004). At this stage the outer and inner cells are no longer totipotent and have become restricted to ICM or TE. Currently, there are two major working models for lineage commitment in mouse preimplantation embryos: the cell polarity model (Johnson and Ziomek 1981) and the inside/outside model (Fig. 1) (Tarkowski and Wroblewska 1967). The Cell polarity model proposes that blastomere polarization and asymmetric division into polar (outside cells) and non-polar (inside cells) drives segregation into ICM and TE [(Johnson and Ziomek 1981; Ziomek and Johnson 1980); reviewed by (Cockburn and Rossant 2010)]. Alternatively, the inside/outside model suggest that cell position within the embryo dictates an ICM and TE cell-fate [(Hillman et al. 1972; Tarkowski and Wroblewska 1967) reviewed by (Cockburn and Rossant 2010)]. The current notion is that both models are important for the segregation of the ICM and TE during blastocyst formation in mice. Nonetheless, further research is necessary in order to establish the precise mechanism by which blastomeres become committed to an ICM vs. TE cell-fate. The transcriptional and epigenetic underpinnings that mediate lineage specification and commitment during blastocyst development are discussed below.

Figure 1.

A. Different stages of mouse preimplantation development (from 1-cell to blastocyst). Zygotic gene expression is induced between the 1-cell and 8-cell stage. At the 8-cell stage blastomeres undergo compaction and begin to separate into inner and outer positions. The outer cells become polar with apical and basal poles. The morula-to-blastocyst transition is associated with the formation of a blastocoel cavity and segregation of the ICM and TE lineages. B. Timeline highlighting important research discoveries associated with ICM and TE-fate specification.

Reciprocal regulation of OCT4, NANOG, and CDX2 controls ICM and TE commitment in mice

Along with the morphological and cellular changes that facilitate blastocyst formation a subset of transcription factors are necessary for segregation of the ICM and TE (Fig. 1 and Table 1). Caudal type homeobox 2 (CDX2) is required for specification of the TE, while Octamer 3/4 (OCT4) and NANOG are essential for establishment of the pluripotent ICM and epiblast, respectively (Chambers et al. 2003; Mitsui et al. 2003; Nichols et al. 1998; Strumpf et al. 2005). During preimplantation development CDX2 is widely expressed at the 8- and 16-cell stage, however, during blastocyst formation it becomes restricted to the TE epithelium (Dietrich and Hiiragi 2007). Embryos that lack CDX2 develop to the blastocyst stage, but fail to maintain a blastocoel cavity and undergo implantation (Strumpf et al. 2005). Similarly, the pluripotency factors OCT4 and NANOG are broadly expressed at the 8- and 16-cell stages, but become restricted to the ICM during blastocyst formation (Dietrich and Hiiragi 2007; Nichols et al. 1998; Palmieri et al. 1994). Embryos deficient in OCT4 or NANOG arrest at the blastocyst stage and fail to establish a functional ICM (Mitsui et al. 2003; Nichols et al. 1998). Loss of function and gain of function studies in preimplantation embryos and ES cells have demonstrated a mutually antagonist relationship between CDX2, OCT4, and NANOG. For example, genetic ablation or RNAi-mediated knockdown of Cdx2 results in failure to repress Oct4 and Nanog expression in the TE lineage in blastocysts (Strumpf et al. 2005; Wang et al. 2010; Wu et al. 2010). In support of these observations studies in mouse ES cells have demonstrated that forced expression of Cdx2 or ablation of Oct4 can promote a TE cell-fate (Niwa et al. 2005). At the chromatin level CDX2 can interact with the Oct4 and Nanog promoters to facilitate transcriptional silencing (Chen et al. 2009; Niwa et al. 2005; Wang et al. 2010). In contrast, forced expression of Oct4 in mouse TS cells promotes an ES cell-fate (Wu et al. 2011). Whether OCT4-mediated repression of Cdx2 alone is sufficient to promote TS cell differentiation into ES-like cells is not known, however, work in mouse ES cells has revealed that OCT4 binds to the Cdx2 promoter and facilitates transcriptional repression of Cdx2 via epigenetic modifications (Yuan et al. 2009).

Table 1.

Summary of key factors implicated in proper development of the ICM and TE lineages during preimplantation development

TF/epigenetic regulator	Function	Manipulations	Loss of function phenotype	References
CDX2	Homeobox TF; Activator and repressor of transcription	Zygotic KO; RNAi	Lethality at blastocyst stage; Ectopic expression of Oct4 and Nanog in TE; Failure to maintain cavity; No outgrowths	Strumpf et al., 2005; Wang et al., 2010; Wu et al., 2010
OCT4	Homeobox TF; Activator and repressor of transcription	Zygotic KO	Lethality at blastocyst stage; Loss of pluripotency; Failure to derive ES cells; Excess trophectoderm	Nichols et al., 1998
NANOG	Homeobox TF; Activator and repressor of transcription	Zygotic KO	Lethality at blastocyst stage; Loss of pluripotency; Failure to form epiblast; In vitro, ICMs differentiate into parietal endoderm-like cells	Mitsui et al., 2003
GATA3	Zinc finger TF; Activator and repressor of transcription	RNAi	Inhibition of morula to blastocyst transition; repression of TE markers.	Home et al., 2009; Ralston et al., 2010
TEAD4	TF; Activator and repressor of transcription	Zygotic KO; RNAi	Arrest during morula-to-blastocyst transition; loss of TE-specific gene expression.	Yagi ei al., 2007, Nishioka et al., 2008, Home et al., 2012
TCFAP2C	TF; Activator and repressor of transcription	RNAi	Arrest during morula-to-blastocyst transition; Loss of cell polarity; Failure to form tight junctions; Cell-cycle defects; Absence of TE formation	Inchul et al., 2013
KLF5	Zinc finger TF; Activator and repressor of transcription	Zygotic KO	Developmental arrest at the blastocysts stage due to improper ICM and TE development	Lin et al., 2010
CARM1	Histone arginine methyltransferase (H3R26me)	Overexpression in single blastomeres	Increase in H3R26 methylation; Increased expression of Nanog and Sox2; Altered cell-fate; Propensity to contribute to ICM	Torres-Padilla et al., 2007.
ESET	Histone lysine methyltransferase (H3K9me3)	Zygotic KO; RNAi	Peri-implantation lethality due to defective ICM development.	Dodge et al. 2004, Yeap et al., 2009, Yuan et al., 2009.
SUV39H1	Histone lysine methyltransferase (H3K9me3)	RNAi	Loss of TE lineage plasticity due to altered H3K9Me3 modification.	Alder et al., 2010, Rugg-Gunn et al., 2010.
EED	PRC2 core component	Ectopic overexpression in the TE.	Loss of TE function and repression of TE-specific regulators.	Saha et al., 2012.
KDM6B	Lysine-specific histone demethylase	RNAi	Delayed blastocyst maturation and loss of TE function resulting in defective embryo implantation	Saha et al., 2012
BRG1	ATP-dependent chromatin remodeler	Zygotic KO; RNAi	Embryonic lethality at blastocyst stage; Ectopic expression of OCT4 and NANOG; Failure to form outgrowths; Failure to derive ES cells	Bultman et al., 2000; Kidder et al., 2009; Wang et al., 2010; Ho et al., 2011
BAF155	Subunit of BRG1 chromatin remodeling complexes	Zygotic KO; RNAi	Embryonic lethality at blastocyst stage; Failure to form outgrowths	Guidi et al., 2001

In stark contrast to the mouse preimplantation embryo, the existence of reciprocal co-regulation of OCT4 and CDX2 in the ICM and TE in bovine and human blastocysts is less clear. In these species OCT4 is expressed in both the ICM and TE (Cauffman et al. 2005; Kirchhof et al. 2000; Kurosaka et al. 2004), while, CDX2 expression is enriched in the TE (Adjaye et al. 2005; Berg et al. 2011). To better understand the differences between cattle, humans, and mice Berg et al analyzed the regulatory sequences upstream of the Oct4/OCT4 promoters (Berg et al. 2011). They found that the mouse Oct4/OCT4 distal enhancer (conserved region 4; CR4) contains a transcription factor AP-2γ (TCFAP2C) binding motif that is absent in the bovine and human OCT4 gene. Furthermore, the authors showed that a mouse Oct4-GFP transgene containing the CR4 region was restricted to the ICM in mouse blastocysts, but not bovine blastocysts. Reciprocally, a bovine OCT4-GFP transgene containing the CR4 region was not restricted to the ICM in mouse blastocysts, suggesting that TCFAP2C binding to the mouse Oct4 distal enhancer is necessary for Oct4 repression in the blastocyst TE. However, functional studies in mouse preimplantation embryos were not conducted in that study (Berg et al. 2011). To address the biological role of TCFAP2C in Oct4 regulation, Choi et al utilized an RNAi approach and found that neither TCFAP2C or its closely related family member TCFAP2A are necessary for Oct4 repression in the TE lineage of mouse blastocysts (Choi et al. 2013). These findings demonstrate the CDX2 is the predominant negative regulator of Oct4 expression in mice. In regards to the role of TCFAP2C in mice, a recent study demonstrated that TCFAP2C functions as a master regulator of blastocyst formation and acts upstream of CDX2 to regulate key cellular processes necessary for formation of the TE epithelium (see below; Choi et al., 2012). Collectively, these findings suggest that differences in the developmental timing of Oct4 silencing in the TE lineage in cattle, humans, and mice are not attributed to TCFAP2C function, but likely involve other intrinsic factors.

Other transcription factors and master regulators that control blastocyst formation and/or lineage commitment in mice

Studies in mice implicated several other transcription factors, like GATA3, TEAD4, KLF5 and TCFAP2C in blastocyst formation and transcriptional regulation of the ICM and TE lineages (Fig. 1 and Table 1). During preimplantation mouse development GATA3 expression is induced during 4–8 cell stage (Home et al. 2009; Ralston et al. 2010). However, during the morula-to-blastocyst transition GATA3 expression becomes restricted within the TE-lineage cells (Home et al. 2009; Ralston et al. 2010). Depletion of GATA3 results in partial inhibition of blastocyst development due to defective TE-specific transcriptional program (Home et al. 2009). In correlation to this function, it has been shown that in the developing mouse embryo GATA3 expression is restricted within the extraembryonic trophoblast lineage until embryonic day (E) 9.5. (George et al. 1994). However, Gata3^−/− embryos can survive until E11–E11.5 (Pandolfi et al. 1995), indicating that some other factors might compensate loss of GATA3 during pre-implantation development. The molecular analyses predicted that CDX2 and GATA3 function in parallel for proper development of the TE-lineage (Ralston et al. 2010). Thus, CDX2-mediated pathways could prevail and compensate preimplantation development in the absence of GATA3. Alternatively, other GATA factors, like GATA2 might compensate GATA3-loss in some context during preimplantation development. The induction of GATA2 upon GATA3 depletion in trophoblast cells (Ray et al. 2009) further supports this scenario.

During mouse preimplantation development, both GATA3 and CDX2 expression is directly regulated by TEAD4, a TEA domain containing transcription factor (Fig. 1 and Table 1). TEAD4-null mouse embryos do not form blastocysts due to lack of establishment of a TE-specific transcriptional program and is characterized by near complete loss of CDX2 and GATA3 expression (Nishioka et al. 2008; Ralston et al. 2010; Yagi et al. 2007). Interestingly, TEAD4 expression is conserved within the TE-lineage of other mammalian blastocysts, including human (Home et al. 2012). Thus, TEAD4 is considered to be a master regulator that establishes a TE-specific transcriptional program during preimplantation development.

Along with TEAD4 other transcription factors such as TCFAP2C may function as master regulators of blastocyst formation and/or TE specification in mice (Fig. 1 and Table 1) (Choi et al. 2012). Recently, it was demonstrated that TCFAP2C could bind and regulate a diverse group of genes that are crucial for blastocyst formation in mammals (Choi et al. 2012). These include genes important for regulation of cell polarity (e.g., Pard6b), TJ biogenesis (e.g., Cldn4 and Tjp2), fluid accumulation (e.g., ATPb1 and Aqp3), and cell-cycle (e.g., p21/WAF) (Adiga et al. 2007; Alarcon 2010; Barcroft et al. 2003; Madan et al. 2007; Moriwaki et al. 2007; Sheth et al. 2008). In support of these findings, cell allocation and TJ-mediated paracellular sealing are perturbed in TCFAP2C knockdown embryos (Choi et al. 2012). Consistent with the phenotype of TEAD4-null embryos (Yagi et al. 2007), the expression of Cdx2 is also downregulated in TCFAP2C knockdown morulae (Choi et al. 2012), suggesting that TCFAP2C may work in conjunction with TEAD4 to regulate Cdx2 expression. Because TEAD4 and TCFAP2C co-occupy a large number of genes in TS cells it is plausible that TEAD4 and TCFAP2C function as master regulators of blastocyst formation and/or cell-fate (Home et al. 2012).

In contrast to OCT4 and NANOG, or CDX2, GATA3, TEAD4 and TCFAP2C, which are important for ICM or TE-specification, respectively, KLF5 is implicated in the formation of both ICM and TE lineages (Fig. 1 and Table 1) (Lin et al. 2010). Interestingly, in Klf5 mutant embryos, cell polarity appeared to be normal. However, their development is arrested at the blastocyst stage due to defective TE development. Also, loss of Klf5 resulted in reduced expression of Oct4 and Nanog within the ICM. Thus, it is predicted that KLF5 functions in the TE and ICM lineage cells in a cell autonomous fashion and promotes proper transcriptional program in both lineages (Lin et al. 2010).

Covalent histone modifications and first mammalian lineage commitment

Analyses in mouse preimplantation embryos indicated that the DNA methylation patterns vary between the ICM and the TE (Santos et al. 2002). However, lack of preimplantation phenotype in mouse embryos deficient for Dnmt1, Dnmt3a, and Dnmt3b (Sakaue et al. 2010) suggests that DNA methylation is dispensable during ICM and TE specification. Thus, instead of DNA-methylation, post-translational modifications (PTMs) of histones within chromatin could be the most important epigenetic component for orchestrating transcriptional program during TE vs. ICM commitment.

Acetylation is the first reported PTM that was discovered about five decades ago (Allfrey et al. 1968). Since then several other PTMs, including methylation, ubiquitination, SUMOylation, and phosphorylation have been reported for different histone residues including lysine, arginine, serine and threonine. Histone PTMs have been implicated in multiple aspects of DNA metabolism, including replication, recombination, and repair (Bhaumik et al. 2007; Chi et al. 2010; Li 2002). However, modulation of transcriptional outcome in association with histone PTMs is probably the most studied field. The PTM’s within histone molecules, especially at histone H3 and H4 tails alter chromatin dynamics and accessibility at specific gene loci to activate or repress transcription. Despite of the presence of multitude of histone PTM’s and a plethora of knowledge about their roles in different tissue and cell systems, surprisingly, our understanding about the contribution of distinct histone modifications in the context of TE vs. ICM commitment is at its infancy. To date the most convincing studies that functionally relate histone modifications with TE vs. ICM segregation are mostly focused on methylations of histone lysine and arginine residues (Alder et al. 2010; Rugg-Gunn et al. 2010; Saha et al. 2013; Torres-Padilla et al. 2007; Yeap et al. 2009; Yuan et al. 2009).

Histone H3 Arginine methylation

The first convincing evidence to relate a specific histone PTM in TE vs. ICM lineage segregation came from a study (Torres-Padilla et al. 2007) showing that alteration of methylation at histone H3 arginine 26 residue (H3R26me) could alter TE vs. ICM commitments of a blastomere (Table 1). Through ectopic overexpression of coactivator-associated arginine methyltransferase 1 (CARM1), the study showed that higher levels of H3R26me facilitates ICM fate in a blastomere by up-regulating specific factors like Nanog and Sox2. Interestingly, although the study focused only on H3R26 methylation, CARM1 methylates other arginine residues including arginine’s 2, 17 at the N-terminus as well as arginine’s (128/129/131/134) at the C-terminus of histone H3 (Zhang and Reinberg 2001). Thus, it is a point of further study whether ectopic H3R26 methylation alone is sufficient to skew blastomeres towards ICM fate. Interestingly, Carm1-null mouse embryos do not show overt preimplantation phenotype, rather show perinatal death. These results indicate that a CARM1 function is dispensable during preimplantation development.

Histone H3 lysine methylation

Along with histone H3 arginine residue, methylations at histone H3 lysine 9 and lysine 27 have also been implicated in TE vs. ICM lineage commitment. Gene knockout studies in mice showed that absence of (ERG-associated protein with SET domain) (ESET, also known as SETDB1), a histone methyltransferase that specifically trimethylates H3K9 residue, results in peri-implantation lethality (Table 1) (Dodge et al. 2004). Blastocysts null for Eset showed defective growth of the ICM. This observation is further explained by recent studies (Yeap et al. 2009; Yuan et al. 2009), which showed that depletion of ESET promotes TE fate in blastomeres by induction of TE-specific genes including Cdx2 and Tcfap2a. Not surprisingly, depletion of ESET in mouse embryonic stem (ES) cells leads to induction of a trophoblast cell phenotype in vitro and in vivo (Yeap et al. 2009; Yuan et al. 2009). Collectively, these studies indicate that ESET-mediated H3K9Me3 modification is important to suppress TE-specific transcriptional program during ICM commitment.

Recent studies indicated that H3K9 methylation could also be important in repressing ICM-specific genes within the TE (Alder et al. 2010; Rugg-Gunn et al. 2010). These studies have indicated that function of suppressor of variegation 3–9 homolog 1 (SUV39H1), another histone methyltransferase that specifically trimethylates H3K9 is involved in incorporating repressive H3K9Me3 modification at ICM-specific as well as somatic lineage regulators within the TE-lineage or in TE-derived trophoblast stem cells (Table 1) (Alder et al. 2010; Rugg-Gunn et al. 2010). Thus, ICM vs. TE commitment is finely tuned via H3K9Me3 incorporation at specific-chromatin domains by distinct histone H3K9 methyl transferases.

In addition to H3K9Me3, regulation of repressive histone modification H3K27Me3 is also important during preimplantation lineage development. In recent years, a number of studies (Alder et al. 2010; Dahl et al. 2010; Rugg-Gunn et al. 2010) revealed a quantitative difference in the global H3K27Me3 level in the TE vs. ICM of a mouse blastocyst and indicated that; in general, the promoter regions of several TE-specific genes are enriched with repressive H3K27Me3 modification in the ICM lineage or in ICM-derived ES cells. However, H3K27Me3 modification is indicated to be dispensable for blastocyst maturation as both maternal and zygotic deletion of enhancer of zeste homolog 2 (EZH2), the H3K27 methyltransferase, showed no effect on ICM and TE lineages establishment (Table 1) (Terranova et al. 2008). Intriguingly, analyses of molecular mechanisms that control H3K27Me3 modification revealed that loss of H3K27Me3 incorporation at the TE-specific genes like Cdx2 and Gata3 is essential for proper development of the TE lineage and embryo implantation (Saha et al. 2013). It has been shown that, during ICM vs. TE commitment, the differential incorporation of H3K27me3 modification at the chromatin domains of TE regulators is coordinated via combinatorial regulation of embryonic ectoderm development (EED) and lysine-specific demethylase 6B (KDM6B) (Table 1) (Saha et al. 2013). Relative loss of EED expression (Rugg-Gunn et al. 2010; Saha et al. 2013) within the TE and function of KDM6B prevents polycomb repressor 2 (PRC2) complex recruitment and incorporation of H3K27Me3 mark at the chromatin domains of Cdx2 and Gata3 within the TE lineage, ensuring transcriptional activation (Rugg-Gunn et al. 2010). The functional relevance of this mechanism was confirmed through ectopic gain of EED along with depletion of KDM6B in preimplantation mouse embryos, which abrogated CDX2 and GATA3 expression in the nascent TE-lineage and resulted in improper TE development leading to failure in embryo implantation to the uterus (Rugg-Gunn et al. 2010). The importance of KDM6B is also validated during bovine blastocyst development. The depletion of KDM6B in bovine preimplantation embryos altered H3K27Me3 levels and reduced the rate of blastocyst development (Canovas et al. 2012). Collectively, these studies indicate that balance of H3K9Me3 vs. H3K27Me3 incorporation within specific chromatin domains is important to dictate TE vs. ICM fate in blastomeres (Fig. 2).

Figure 2.

The model proposes that repression of TE-specific genes via PRC2/H3K27Me3 and ESET/H3K9Me3 axis promotes ICM commitment. In contrast, repression of ICM-specific genes via SUB39H1/H3K9Me3 axis facilitates TE-fate. KDM6B promotes TE fate by preventing incorporation of H3K27Me3 mark at the TE-specific genes within the TE lineage.

Besides the above-discussed PTMs, other histone marks could also be involved in transcriptional regulation during preimplantation lineage commitment. Analyses in preimplantation embryos showed that several other histone modifications, that are associated with transcriptional activation including H3 lysine 4 methylation (H3K4me) and histone H3/H4 acetylation (Gupta et al. 2008; Ma and Schultz 2008; Sarmento et al. 2004; Wongtawan et al. 2011) also exist throughout the preimplantation development. However, functional consequences of those modifications in ICM vs. TE commitment are yet to be established.

Role of ATP-dependent chromatin remodeling complexes in early development

Along with the covalent modifications of histones, non-covalent modifications of chromatin structure are instrumental in regulation of gene expression during mammalian development. A growing body of evidence supports a fundamental role for ATP-dependent chromatin remodeling complexes in regulation of stem cell self-renewal and cellular differentiation during embryonic and fetal development [reviewed by (Ho and Crabtree 2010)]. ATP-dependent chromatin remodeling complexes function to assemble and/or displace nucleosomes at target gene promoters to either activate or repress transcription (Struhl and Segal 2013). Currently, there are four known families of ATP-dependent chromatin remodeling complexes in mammals: SWI/SNF, ISWI, CHD, and INO80. These complexes consist of multiple subunits which comprise a core catalytic subunit that confers chromatin remodeling activity (Ho and Crabtree 2010). Each family is based on the sequence and structure of the catalytic ATPase. For an extensive review of these chromatin remodeling families and their role in mammalian development please refer to these excellent reviews (de la Serna et al. 2006; Ho and Crabtree 2010). In this review we focus mainly on the biological role of the SWI/SNF chromatin remodeling family, whose roles in blastocyst development and ES cell pluripotency are better understood.

SWI/SNF chromatin remodeling family

The SWI/SNF family of chromatin remodeling proteins contain 9–12 subunits and utilize either Brahma (BRM) or Brahma-related gene 1 (BRG1) as the catalytic subunit (Wang et al. 1996a; Wang et al. 1996b). BRG1 and BRM contain several important domains necessary for its catalytic activity and interaction with chromatin (Trotter and Archer 2008). These include a DEXH-box helicase, a helicase, and a bromo domain. The bromo and helicase domains facilitate interactions with acetylated histones and unwinding of DNA, while the DEXH-box helicase domain contains a ATP binding site necessary for its catalytic activity. The non-catalytic subunits that assemble into SWI/SNF complexes are referred to as BRG1/BRM-associated factors (BAFs). There are at least 14 known BAFs in mammals. These include BAFs 45a-d, 47, 53a-b, 60a-c, 155, 170, 250A, and 250B [(Wang et al. 1996a); reviewed by (Ho and Crabtree 2010)]. In different cell-types SWI/SNF complexes can contain different combinations of BAFs that are thought to be important for biological specificity (Ho et al. 2009b; Lessard et al. 2007). In addition, BRG1 and BRM can interact with various transcription factors (Kadam and Emerson 2003; Trotter and Archer 2008) and covalent epigenetic modifiers such as histone methyltransferases (e.g., CARM1), histone deacetylases (e.g., HDAC1/2), and DNA methyltransferases (e.g., DNMT3A) to either activate or repress transcription in different cellular contexts (Datta et al. 2005; Underhill et al. 2000; Xu et al. 2004).

Functional studies in mice have delineated a crucial role for SWI/SNF remodeling complexes in cellular proliferation and differentiation. Gene knockout studies have demonstrated that BRG1 containing complexes are critical throughout development. However, loss of BRM does not disrupt embryonic and fetal development, albeit, adult mice are overweight (Reyes et al. 1998). The most interesting findings from these studies was when either Brg1 or Baf155 was genetically ablated (Table 1). Knockout of zygotic Brg1 causes embryonic lethality at the blastocyst stage (Bultman et al. 2000). Brg1^−/− blastocysts cultured in vitro fail to give rise to TE and ICM outgrowths (Bultman et al. 2000). Similarly, knockout of BAF155 results in embryonic lethality at the blastocyst stage and failure to form outgrowths (Guidi et al. 2001). These early findings suggested that BRG1/BAF155 containing chromatin complexes might be essential for normal blastocyst development in mice. Nonetheless, the etiology of the mutant phenotype and the underlying molecular mechanisms by which BRG1/BAF155 regulates blastocyst development and pluripotency were largely not known until recently.

Emerging role of Brg1/BAF155 complexes in pluripotency

In recent years, a handful of studies have elucidated the biological role of BRG1 and BAF155 in early embryonic development using mouse ES cells as a model system for the pluripotent ICM (Ho et al. 2009a; Ho et al. 2011; Ho et al. 2009b; Kidder et al. 2009; Wang et al. 2010). In ES cells genome-wide chromatin immunoprecipitation (ChIP) analyses of BRG1 binding was carried out (Ho et al. 2009b; Kidder et al. 2009). BRG1 was shown to bind to two major groups of genes: genes important for ES cell pluripotency and genes important for differentiation such as developmental transcription factors. In support of these findings depletion of BRG1 or BAF155 resulted in loss of pluripotency and spontaneous differentiation (Ho et al. 2009a; Ho et al. 2009b; Kidder et al. 2009), suggesting that BRG1 and BAF155 containing complexes are necessary for maintenance of self-renewal and prevention of differentiation in ES cells. Interestingly, ablation of Brg1 or BAF155 in ES cells results in elevated expression of Nanog before and during retinoic acid (RA) induced differentiation (Ho et al. 2009a; Schaniel et al. 2009), demonstrating that Brg1 is necessary for fine-tuning its expression in ES cells and silencing its expression during differentiation. To determine the molecular composition of BRG1 containing complexes in ES cells, a proteomic approach was used to establish the subunit composition in ES cells versus other cell-types (Ho et al. 2009a). In ES cells there is an ES cell specific BAF containing complex (esBAF) that contains BRG1 and BAF155, but lacks BAF170. Furthermore, through mass spectrometry and co-immunoprecipitation it was demonstrated that BRG1 could interact with OCT4 and NANOG in ES cells (Ho et al. 2009a; Liang et al. 2008). Consistent with this BRG1, NANOG, and OCT4 co-localize to a subset of target genes in ES cells (Ho et al. 2009b; Kidder et al. 2009). Collectively, these results demonstrate that a BRG1/BAF155 containing complex is essential for self-renewal and pluripotency in ES cells and suggest that BRG1 may regulate pluripotency via interactions with OCT4 and NANOG.

Maintenance of self-renewal and pluripotency in mouse ES cells is regulated by leukemia inhibitory factor (LIF) signaling (Smith et al. 1988; Williams et al. 1988). To establish a connection between esBAF complexes and LIF signaling Ho et al evaluated STAT3 target genes in ES cells (Ho et al. 2011). Through this analysis they determined that esBAF is required for the recruitment of STAT3 to target gene promoters. Interestingly, BRG1 was found to prime STAT3 target genes so that they could be actively expressed in ES cells. Loss of BRG1 resulted in gene silencing through recruitment of PcG complexes, suggesting that BRG1 acts in opposition to Polycomb group (PcG) complexes to maintain self-renewal in ES cells (Ho et al. 2011). Altogether, these studies demonstrate that esBAF is crucial for maintenance of self-renewal and pluripotency in ES cells.

Role of Brg1 in lineage commitment in preimplantation embryos

Apart from the role of BRG1 in ES cell pluripotency it is equally important to understand the biological function of BRG1 in lineage commitment in blastocysts. Recent studies have provided some tantalizing clues into the role of BRG1 in blastocyst development. The first evidence came from a study that utilized a combination of RNAi and transcriptome analysis (Kidder et al. 2009). The authors determined that approximately 2000 genes were deregulated in BRG1 knockdown blastocysts. Intriguingly, core regulators of pluripotency, such as OCT4 and NANOG, were ectopically expressed in the TE lineage of BRG1 knockdown blastocysts (Kidder et al. 2009). Consistent with the original Brg1 null phenotype, BRG1 knockdown blastocysts failed to form TE outgrowths when cultured in vitro. To establish a direct link between BRG1 and Oct4 regulation in blastocysts two major sets of experiments were carried out (Wang et al. 2010). First through ChIP and co-immunoprecipitation analysis BRG1 was shown to interact with CDX2 at the Oct4 distal enhancer. Second, combined depletion of BRG1 and CDX2, compared to depletion of BRG1 or CDX2 alone, resulted in elevated levels of Oct4 transcripts in blastocysts suggesting that BRG1 cooperates with CDX2 to regulate TE specification in mouse blastocysts. Importantly, these studies were the first to establish a direct role for an epigenetic modifier in CDX2-mediated silencing of Oct4 expression in mouse blastocysts. These findings in conjunction with earlier studies that showed OCT4 and ESET negatively regulate Cdx2 expression in pluripotent cells (Yeap et al. 2009; Yuan et al. 2009), suggest that lineage commitment in mouse blastocysts involves a combination of transcriptional and epigenetic mechanisms. Additional studies are necessary to identify the underlying mechanisms by which BRG1 and ESET converge with CDX2 and OCT4 to regulate specification of the ICM and TE. Moreover, it will be important to determine the molecular mechanism by which BRG1 represses Oct4 and Nanog transcription in the TE.

Future perspectives

As discussed above, recent studies have immensely increased our understanding regarding molecular mechanisms that contribute to cell-fate specification of ICM vs. TE lineages. However several questions remain unanswered. For example:

1. Although the importance of several transcription factors have been identified for the establishment of ICM and TE-specific transcriptional programs, it is still unknown how combinatorial functions of distinct transcription factors orchestrate nucleoprotein complexes at ICM and TE chromatin to fine tune gene expression patterns. The advancement of new technologies, specifically genome-wide analyses to identify transcription factor target genes and global gene expression analyses will facilitate to identify more detailed mechanistic components.

2. Regarding covalent histone modifications several questions needed to be answered. For example, (i) What mechanisms regulate expression of histone methylases and demethylases in the ICM and TE-lineage?, (ii) What mechanisms specify chromatin occupancy patterns of different histone modification enzymes? For example, both ESET and SUV39H1 methylates H3K9 residue. However, ESET is important for the suppression of TE-specific genes in nascent ICM lineage cells, whereas SUV39H1 is implicated in suppressing ICM-specific genes within the nascent TE-lineage. Though, the mechanisms that specify which genes will be regulated by these molecules are unknown.

3. The underlying epigenetic mechanisms by which SWI/SNF chromatin remodeling complexes converge with core transcriptional networks to regulate lineage commitment and ES cell pluripotency are largely not known. (i) Does OCT4 and NANOG recruit BRG1 to target gene promoters? Or alternatively does BRG1 facilitate open chromatin to allow OCT4 and NANOG to bind to target gene promoters? (ii) Does BRG1/esBAF associate with a unique set of epigenetic enzymes in ES cells and preimplantation embryos? (iii) Does BRG1/esBAF regulate gene expression in ES cells and embryos via a dual mechanism consisting of nucleosome remodeling and recruitment of enzymes that induce PTMs of histones? (iv) Finally, does BRG1/esBAF control cell-fate commitment via transcriptional regulation of Oct4 and Nanog in preimplantation embryos?

Future studies in these contexts will enhance our understanding regarding epigenetic controls underlying physiological processes for proper preimplantation development and could create opportunities for new scientific and therapeutic pursuits.

Abbreviations

ICM

inner cell mass

TE

trophectoderm

PE

primitive endoderm

CDH1

E-Cadherin

PARD3

Partitioning defective homologue 3

PARD6B

Partitioning defective homologue 6b

PKCζ

Protein kinase C zeta

CDC42

cell division cycle 42

TJ

tight junction

CLDN4

Claudin 4

TJP2

tight junction protein 2

ES cells

embryonic stem cells

TS cells

trophoblast stem cells

RNAi

RNA interference

CDX2

Caudal type homeobox 2

OCT4

Octamer 3/4

CR4

conserved region 4

GFP

green fluorescent protein

TCFAP2C

transcription factor AP-2γ

TEAD4

TEA domain family member 4

GATA3

GATA binding protein 3

GATA2

GATA binding protein 2

ATPB1

Na/K-ATPase β1

AQP3

Aquaporin 3

KLF5

Kruppel-like factor 5

PTM

post-translational modifications

H3R26me

histone H3 arginine 26

ESET

ERG-associated protein with SET domain

H3K9me

Histone H3 lysine 9

H3K27me

histone H3 lysine 27

EED

embryonic ectoderm development

KDM6B

lysine-specific demethylase 6B

SUV39H1

variegation 3–9 homolog 1

EZH2

enhancer of zeste homolog 2

PRC2

polycomb repressor 2 complex

H3K4

histone H3 lysine 4

BRM

Brahma

BRG1

Brahma related gene 1

BAF

BRG1-associated factor

CARM1

coactivator-associated arginine methyltransferase 1

HDAC

histone deacetylase

DNMT

DNA methyltransferase

LIF

leukemia inhibitory factor