Omics Views of Mechanisms for Cell Fate Determination in Early Mammalian Development

During mammalian preimplantation development, a totipotent zygote undergoes several cell cleavages and two rounds of cell fate determination, ultimately forming a mature blastocyst. Along with compaction, the establishment of apicobasal cell polarity breaks the symmetry of an embryo and guides subsequent cell fate choice. Although the lineage segregation of the inner cell mass (ICM) and trophectoderm (TE) is the first symbol of cell differentiation, several molecules have been shown to bias the early cell fate through their inter-cellular variations at much earlier stages, including the 2- and 4-cell stages. The underlying mechanisms of early cell fate determination have long been an important research topic. In this review, we summarize the molecular events that occur during early embryogenesis, as well as the current understanding of their regulatory roles in cell fate decisions. Moreover, as powerful tools for early embryogenesis research, single-cell omics techniques have been applied to both mouse and human preimplantation embryos and have contributed to the discovery of cell fate regulators. Here, we summarize their applications in the research of preimplantation embryos, and provide new insights and perspectives on cell fate regulation.


Introduction
During mammalian preimplantation embryonic development, a fertilized egg goes through 2-cell, 4-cell, 8-cell, and morula (referring to the compacted 16-to 32-cell embryos) stages to form a hollow sphere blastocyst embryo, during which sequential events occur, including zygotic genome activation (ZGA), embryo compaction, and two rounds of cell fate determination [1] (Figure 1).Cells at the blastocyst stage comprise three cell lineages: trophectoderm (TE) for placenta, primitive endoderm (PE; a predecessor of the yolk sac), and epiblast (EPI; the progenitor for the fetus).TE cells are localized in the outer layers of the blastocyst with apical polarity, whereas cells in the center, referred to as inner cell mass (ICM) cells, consist of PE and EPI with strong cell pluripotency [2][3][4][5].The location of ICM cells delineates the ''embryonic pole" of blastocyst and lies on the opposite of the ''ab-embryonic pole," demonstrating cell polarity in the blastocyst [6].Compared with mouse embryos, lineage allocations of human preimplantation embryos are completed a little later, until cell fate decisions emerge after the cavitation of blastocysts (Figure 1).Currently, two models have been proposed to explain early human cell fate determination: one is a two-step model in which the formation of TE-ICM and TE-PE-EPI occurs in order [7], and the other is a one-step model which suggests the simultaneous occurrence of TE, EPI, and PE lineages [8].A more recent study has reported that during human blastocyst formation, a type of unspecified cell, similar to ICM cells, emerges at the Blastocyst 2 (B2) to B3 stages, which is indicated by the EPI marker Ifi16 and the PE marker Gata4.This finding provides new evidence for a two-step model [9].
Although the biological functions of transcription factors and signaling pathways at the morula and blastocyst stages have been revealed, what and how other omics events control the first two rounds of cell fate determination remain largely unknown.Currently, early cell fate determination has been reported to occur much earlier than the morula stage when morphological heterogeneity emerges.For example, the molecular basis of cell polarity at the 8-cell stage [10] and cell-to-cell differences (cellular heterogeneity) at the 2-and 4-cell stages [11,12] have been shown to regulate cell fate decisions by functioning in downstream networks.
In this review, we summarize the molecular events that occur during early embryogenesis, as well as the current understanding of their regulatory roles in cell fate decisions.Singlecell omics techniques have been employed as powerful tools in early embryogenesis research to dissect programmed molecular events in mouse and human preimplantation embryos, and many discoveries of cell fate determination have been reported.Here, we have summarized their applications in the research field of preimplantation embryos, and new insights and future perspectives based on current omics data are also included.

Early cell fate determination in mammalian preimplantation embryos
The first two waves of cell fate determination lead to the formation of three cell lineages.Each wave is indicated by the heterogeneous distribution of specific transcription factors and the distinctly activated states of signaling pathways.

The first cell fate determination
Both TE and ICM cell lineages are achieved by the first cell fate determination (Figure 1), in which the representative transcription factors are activated, such as SOX2, OCT4, and NANOG in ICM and CDX2 in TE [5,[13][14][15][16].Sox2 and Oct4 are highly expressed at the morula and blastocyst stages and are detectable at the 2-and 4-cell stages [13,14,[17][18][19], indicating that these two genes potentially regulate lineage  separation at a much earlier stage.Blastocyst cavitation is not affected by lack of SOX2 but fails to survive shortly after implantation [13].Nanog displays a unique expressional pattern with a predominant distribution in the ICM fate cells of morula embryos versus a randomly scattered distribution in the TE [14,20].Cdx2 is a specific marker gene of TE in early embryogenesis and is predominantly expressed in the outer cells of morula and blastocyst than in the inner cells.Loss of Cdx2 fails to maintain blastocoels and induces embryonic death prior to implantation [5].Additionally, Id2 has been identified as a specific TE cell marker by single-cell expression analysis [21].
The Hippo and Notch signaling pathways drive the first cell fate determination in mammals (Figure 2A), whereas Hippo also serves as a tumor suppressor signaling pathway conserved in both mice and humans [22].In mouse blastocysts, Hippo is activated in ICM cells but inactive in TE cells through changes in the state of its key components, TEAD4 and YAP.As an important effector of the Hippo pathway, the phosphorylation of YAP activates the Hippo pathway and favors ICM fate [23,24].First, NF2 (also known as Merlin) in the Hippo pathway acts upstream of LATS (LATS1/2) kinases; second, another effector of Hippo, junction-associated scaffolding angiomotin (Amot), is phosphorylated by activated LATS; third, the activated NF2, LATS, and Amot form a regulatory complex, further phosphorylating YAP for its retention in the cytoplasm, which results in the high expression of Sox2 and Oct4 and facilitates ICM lineage specificity [25,26].In TE cells, Hippo is inactive; thus, YAP and Amot are in an unphosphorylated state.Unphosphorylated Amot localizes to the apical domain without functioning in YAP [27].Unphosphorylated YAP translocates from the cytoplasm to the nucleus and interacts with TEAD4 to up-regulate the expression of Cdx2 and Gata3, promoting cell differentiation into the TE lineage [28,29].The Notch signaling pathway controls the first cell fate determination by cooperating with YAP and TEAD4 [30] (Figure 2A).In TE cells, both the YAP-TEAD4 complex of Hippo and the NICD-RBPJ complex of Notch enter the nucleus and up-regulate the expression of TE-specific genes, such as Cdx2, to promote TE fate [31,32].

The second cell fate determination
After the first cell fate determination, ICM cells further differentiated into EPI and PE cells in the second cell fate determination.Nanog and Gata6, serving as the specific markers for EPI and PE, respectively, are required for this process [33,34].Although both begin to be expressed at the 8-cell stage [35], the balance of expression is not tipped until ICM formation.In blastocyst embryos, Nanog + /Gata6 À results in EPI fate, whereas Gata6 + /Nanog À leads to PE fate, referring to a typical ''salt-and-pepper" distribution model.Additionally, other transcription factor-coding genes, such as the downstream genes of Gata6 including Sox17, Gata4, and Sox7, were Hippo and Notch signaling pathways involved in the ICM and TE cell fate allocations.In the Hippo pathway, unphosphorylated YAP enters the nucleus to interact with TEAD4, and thus activates the expression of Cdx2 and Gata3 in TE cells (bottom panel); phosphorylated YAP maintains in cytoplasm, resulting in the up-regulated expression of Oct4 and Sox2 in ICM cells (top panel).Amot regulates the activation of YAP.In the Notch pathway, NICD and RBPJ form a complex in the nucleus of TE cells, activating the expression of Cdx2 and Gata3.B. FGF signaling pathway involved in the PE and EPI cell fate allocations.In EPI cells (bottom panel), NANOG coordinates with OCT4 and SOX2 to up-regulate Fgf4 expression and repress Gata6 expression.In PE cells (top panel), FGF4 secreted from EPI cells activates the FGF/MAPK signaling, and further up-regulates GATA6 and OCT4 which activate the expression of Sox17, Gata4, and Sox7.p, phosphorylation.also found to be intimately correlated with the maturation of PE [36].
The FGF signaling pathway regulates the second cell fate determination in mammalian embryos [1,27,37] (Figure 2B).When the second cell fate determination is initiated, ICM cells express Nanog and Gata6 in a salt-and-pepper manner, in which relatively high expression levels of Gata6 and Nanog separately prompt the expression of Fgfr [a gene encoding fibroblast growth factor receptor (FGFR)] and Fgf4 [a gene encoding fibroblast growth factor 4 (FGF4)] [38].NANOG collaborates with OCT4 and SOX2 to up-regulate the expression of Fgf4 and represses the expression of Gata6, leading to the secretion of FGF4 and bias of EPI [1].FGF4 specifically interacts with FGFR and activates the FGF/MAPK signaling cascade in PE cells, resulting in up-regulation of Gata6 and inhibition of Nanog [1,33].Subsequently, activated FGF signaling and GATA6 cooperatively up-regulate the downstream genes of GATA6 and OCT4, including Sox17, Gata4, and Sox7 [27].After separation of PE and EPI, specified PE progenitors express apical polarity proteins and form an epithelial layer to prevent further cell mixing [35,39].
Additionally, based on single-cell RNA sequencing (scRNA-seq) analysis in 16-cell embryos, BMP signaling was shown to regulate the correct development of TE and PE fates.In blastocysts, the BMP ligands, Bmp4 and Bmp7, are specifically expressed in ICM cells, whereas the BMP receptor Bmpr2 is predominantly distributed in TE cells.Repressing BMP activity by blocking Smad4 or Bmpr2 results in a significant reduction in the cell numbers of TE and PE, but not embryonic EPI.Depletion of Bmp4 and Bmp7 significantly decreases the cell numbers of TE and PE, but not EPI [40].This discovery indicates that BMP signaling is important for the development of extra-embryonic TE and PE lineages, expanding the knowledge of key signaling pathways in early cell fate decisions.
Most of the transcription factors interact or influence each other and coordinate with the signaling pathways of Hippo, Notch, and FGF, to form a functional network modulating the early cell fate decisions.

Establishment of apicobasal cell polarity: symmetrybreaking
When the compaction process is initiated, the blastomeres of the embryo break the symmetry of cell morphology (symmetry-breaking).And owing to the increased cell-cell adhesion and altered cortex tension, both obviously different cell morphology and cell polarization emerge at the 8-cell stage.Molecule-based cell polarity interacts with signaling pathways to regulate cell lineage allocation, including spindle assembly and glycolysis.

Molecular basis of cell polarity establishment
Cell polarity is axially established from the center to the surface, dividing blastomeres into polar cells with non-polar cells residing in the outer and central areas [7,41,42].After compaction, most cells of mid-or late 8-cell embryos give rise to a polar cell and a non-polar cell at the 16-cell stage or two polar cells in which one is internalized into the inner cell by the apical constriction driven by cortical tension [43].Inner cells remain unpolarized and express pluripotency-associated factors, generating ICM cells in blastocysts [4].The outer cells polarize their apical cortex and establish a cortical F-actin ring consisting of apical polarity proteins, such as PAR6B, ezrin, and keratin, which further triggers the differential regulation of the Hippo effector YAP and the transcription factorcoding gene Cdx2, benefiting TE cell fate.Thus, this internalization event interacts with the Hippo signaling pathway to regulate cell fate [25,[44][45][46].
Based on scRNA-seq and RNA interference (RNAi) experiments, transcription factor-coding genes Tfap2c and Tead4 were found to be essential for embryo polarization at the 8to 16-cell stages.Mechanically, TFAP2C and TEAD4 recruit ezrin, promoting the polarization growth of apical protein clusters, which eventually leads to apical protein centralization and subsequent regulation of apical formation and lineage allocation through positive feedback with RhoA [10].
The regulation of cell polarity on cell fate determination seems to imply that components in polarity cells also participate in cell fate control by asymmetric inheritance during cell division, even though lineage specification driven by asymmetrically inherited fate determinants is more widely accepted in non-mammalian embryos than in mammalian embryos [47].Keratin participates in the regulation of cell polarity, Hippo signaling pathway, and mechanisms in epithelial tissues [48].In mouse early 8-cell embryos, keratin-containing cells are asymmetrically inherited during cell division, leading to the formation of inner daughter cells with fewer keratins and outer daughter cells with more keratins.As a TE cell marker [49], more keratins anchor to the apical domain and promote apical polarization and YAP-dependent expression of Cdx2, prompting cells with more keratins to differentiate into TE cells [50].

Downstream regulatory network of cell polarity establishment
Spindle assembly plays a role in regulating cell fate through cell polarity and keratin.Because of the lack of a microtubule-organizing center (MTOC) to generate microtubule asters in mammalian preimplantation embryos [51,52], whether spindle organization regulates early mammalian lineage segregation remains elusive.A recent study reported that cellular heterogeneities of cell polarity in early mouse 8-cell embryos activate an asymmetric spindle assembly, which forms a single microtubule aster (monoastral spindle) from the apically localized non-centrosomic MTOC in an unusual manner.The entirely assembled spindle attaches to the residual apparatus and activates a spatially asymmetrical pattern of division, separating cells inside and outside locations.After mitosis, pulling toward the cortex in the apical regions of the monoastral spindle triggers a burst of F-actin and myosin II, delivering microtubules to the cortex in which the F-actin ring is established and apical polarity proteins are enriched, which retain TE-specified cells in the peripheral position.In contrast, basal regions with an anastral spindle do not clear F-actin or form a ring, internalizing into the central position.Moreover, cellular heterogeneity in keratin assembly differentially triggers the formation of monoastral spindles by stabilizing the cell cortex in some blastomeres of 8-cell embryos.The outer cells with rings and monoastral spindles displayed a higher level of YAP protein, and disruption of F-actin reduced Cdx2 expression, fully proving that the monoastral spindle regulates TE/ICM lineage segregation [53].These results suggest that asymmetric spindle organization modulates the choice of ICM-TE fate.
Glucose and glycolysis also modulate TE and ICM separation at the polarized morula stage.In TE fate cells of morula embryos, the hexosamine biosynthetic pathway (HBP) responsible for glucose metabolism allows YAP1 to localize to the nucleus.The pentose phosphate pathway (PPP), which participates in glucose-dependent nucleotide synthesis, combines with sphingolipid (S1P) signaling to activate the mTOR pathway and allow the translation of Tfap2c.Ultimately, YAP1, TEAD4, and TFAP2C form a heterotrimer that functions as a transcription factor to activate TE-specific markers such as Cdx2, implicating that glucose signaling specifically controls TE cell fate rather than ICM fate specification in early embryos [54].

Cellular heterogeneity: before symmetry-breaking
As mentioned in the section above, the expression patterns of transcription factor-coding genes such as Sox2, Nanog, Cdx2, and Gata6 display significant heterogeneity with distinct cell lineages of ICM-TE or PE-EPI-TE during early cell fate decision processes [55].Cellular heterogeneities at the earlier developmental stages (including the 2-and 4-cell stages) also participate in subsequent cell fate decisions by their functions on those specific transcription factors [11,12,56].The regulatory relationship between cell fate decisions and cellular heterogeneities, including heterogeneous histone H3 methylation at arginine 26 (H3R26me) and heterogeneous transcription factor kinetics, is discussed in the following section.

Heterogeneous H3R26me
The epigenetic reprogramming process is essential for sustaining pluripotency in early mouse embryogenesis [57,58].Many cell fate regulators reported in earlier stages are closely related to a type of histone methylation: H3R26me.
H3R26me and its methyltransferase CARM1 display significant cell-to-cell variations in mouse 4-cell embryos with equatorial and meridional division of the zygote [56].Overexpression of Carm1 increases the abundance of H3R26me and the expression of Nanog, Sox2, and Sox21, indicating that CARM1-H3R26me instructs cell fate decisions by promoting the expression of transcription factors [56,59].Interestingly, CARM1 localizes to a nuclear paraspeckle consisting of p54nrb, PSPC1, PSF, and LncNEAT1.The number of paraspeckles is differentially accumulated in nuclei of 2to 4-cell embryos as well [60].Depletion of p54nrb or LncNEAT1 leads to failed blastocyst cavitation and elevated expression of Cdx2, promoting cell differentiation into the TE fate [56,60].All aforementioned studies reveal a CARM1-H3R26me-mediated mechanism of epigenetic manipulation on cell fate determination: in mouse 4-cell embryos, cells with high levels of CARM1 and H3R36me favor ICM fate, and cells with low levels of CARM1 and H3R26me direct developmental fate to TE.In addition, CARM1 also regulates cell fate by affecting cell polarity and keratin allocation [61].
Similar to Carm1, heterogenous Prdm14 controls cell lineage in 4-cell embryos.PRDM14, a PR-domain and zinc finger protein, is only distributed in early embryonic tissues and during reprogramming events [62][63][64].Using single-cell quantitative combinatorial expression profiling of chromatin modifiers, Prdm14 was observed to have heterogeneous expression at the mouse 4-cell stage.Mechanistically, PRDM14 interacts with CARM1 to drive progenies toward pluripotent cells by increasing H3R26me in mouse 4-cell embryos [65].
The aforementioned cellular heterogeneities in CARM1 and CARM1-paraspeckles occur at the 4-cell stage; however, the key questions are what induces these differences and whether there is cellular heterogeneity associated with early cell fate decision before the 4-cell stage.A long non-coding RNA, LincGET, provides an answer [11].LincGET, serving at the upstream of the CARM1-H3R26me axis, presents asymmetric expression starting at the late 2-cell stage and ending at the 8-cell stage in mice.Depletion of LincGET leads to developmental arrest at the 2-cell stage in early mouse embryos.Overexpression of LincGET or Carm1 increases chromatin accessibility and expression of ICM-specific genes.Mechanistically, LincGET interacts and forms a complex with CARM1 to increase the level of H3R26me, which then activates the chromatin accessibility of ICM-specific genes and biases blastomeres with higher LincGET levels toward ICM fate.Conversely, blastomeres with lower LincGET levels preferentially differentiate into TE cells [11].
Thus, LincGET-CARM1 (in paraspeckle)-H3R26me forms a robust regulatory axis to modulate the subsequent cell fate decision at the 2-to 4-cell stages.

Heterogeneous kinetics of transcription factors
Cellular differences in the dynamic kinetics of transcription factors lead to their differential accessibility to DNA targets, which in turn are controlled by differential epigenetic regulation and participate in early cell fate determination.
Transcription factor-coding genes, Oct4 and Sox2, are highly expressed in the ICM lineage at the morula and blastocyst stages and also exhibit detectable expression in earlier stages, implying that these two transcription factors are involved in early cell fate regulation at an earlier time in a way that is different from their expression.The biological activity of transcription factors in vivo strongly depends on their kinetic behaviors, which effectively modulates gene expression and cell fate [66,67].Using a fluorescence decay after photoactivation (FDAP) assay, two distinct kinetic behaviors of OCT4 resulting from differences in OCT4 accessibility to its DNA targets were observed in mouse 4-cell embryos.These results further point to two distinct cell lineages by lineage tracing: cells with slower OCT4 kinetics are more likely to give rise to the pluripotent cell lineage that contributes to ICM fate.Conversely, cells with faster OCT4 kinetics mostly differentiate into TE fate [68].Similar to OCT4, the binding kinetics of SOX2 to DNA is also found to participate in lineage specificity and cell fate choice by an FDAP assay [17].Long-lived binding of SOX2 corresponds to higher CARM1 and H3R26me levels in the same blastomeres of 4cell embryos.Carm1 deficiency significantly decreases the long-lived bound fraction of SOX2 and the expression of Sox2, Nanog, Oct4, and Sox2 targets.Thus, blastomeres in 4cell embryos with long-lived SOX2 binding mode preferentially bias their progenies to ICM cells [17].
Taken together, we describe a mechanistic model of cellular heterogeneities to explain the cell fate decision during the developmental process from 2-cell to morula stages (Figure 3).
(1) The expression of LincGET and the number of CARM1containing nuclear paraspeckles are asymmetrically distributed in 2-and 4-cell embryos.(2) LincGET coupling with heterogeneous CARM1 and CARM1-containing paraspeckles differentially regulates H3R26me in 4-cell embryos, leading to distinct chromatin accessibility at the 4-cell stage.(3) The opened chromatin benefits the long-lived binding of OCT4 and SOX2 to DNA in 4-cell embryos, resulting in the subsequent increased transcription of ICM-related target genes, including Nanog, Oct4, Sox2, and Sox21, biasing cells to ICM fate.Conversely, lower expression of LincGET, together with fewer CARM1 and CARM1-containing paraspeckles, decreases H3R26me levels and further results in the short-lived DNA binding of OCT4/SOX2, leading to the low expression levels of their targets and activation of TE-specific markers, which predominantly direct cell differentiation into TE cells.

Application of single-cell omics to early cell fate determination
Many of the aforementioned molecular events involved in early cell fate determination were discovered by single-cell transcriptomics or single-cell expression profiles, such as the outer cell-specific marker Id2 at the morula stage, differentiated BMP signaling at the morula and blastocyst stages, regulation of Tfap2c and Tead4 on cell polarity at the 8-to 16-cell stages, heterogeneous Prdm14 at the 4-cell stage, and asymmetric LincGET at the 2-cell stage [10,11,40,65].Thus, singlecell omics is a feasible and robust tool to study cell fate determination during early mammalian embryogenesis.

Single-cell transcriptome
Benefiting from the application of next-generation sequencing platform-based scRNA-seq in mouse preimplantation embryos, dozens of protein-coding genes showed bimodal expression patterns in mouse 2-and 4-cell embryos, and their co-expression exhibited high association with TE and ICM cell fate commitment [69].Similarly, analyses based on the singleblastomere transcriptome data of mouse and human preimplantation embryos also revealed that the earliest symmetrybreaking and initial cell-to-cell biases emerge at the first embryonic cleavage division with a binomial distribution pattern [70].Cell-to-cell transcriptional variations were further elevated with the subsequent development of embryos.During the preimplantation process, some genes with the function of lineage specifiers showed ever-increasing asymmetry between blastomeres, whereas others showed a minimized inter-cell difference, suggesting that transcriptional regulation of symmetry-breaking and early cell fate determination contains complicated positive and negative feedbacks by the differential allocation of two types of genes in early mammalian embryos [70].In mouse 4-cell embryos, scRNA-seq revealed that the target genes of OCT4 and SOX2 displayed highly heterogeneous expression patterns, in which Sox21 is the one showing the most heterogeneous expression profile [71,72].Depletion of Sox21 contributes to greater TE fate than ICM fate by up-regulating the expression and protein levels of Cdx2, indicating that Sox21 benefits ICM cell fate.Intriguingly, the expression of Sox21 and other ICM-related genes, such as Nanog and Esrrb, is influenced by CARM1.The increased or inhibited CARM1 corresponds to up-regulation or downregulation of Sox21, indicating that CARM1-containing epigenetic regulation acts upstream of these genes in cell fate determination [59].
In short, scRNA-seq analysis uncovered inter-cell transcriptional variabilities appearing at the 2-to 4-cell stages in early mouse embryos [11,59,65,69,70].These cellular heterogeneities affect the first cell fate decision in mouse preimplantation development by inducing other biological processes such as cell division, gene expression, and epigenetic modification.

Other single-cell omics techniques
Except for single-cell transcriptomics, omics at different layers, including DNA methylation, chromatin architecture, RNA methylation, and RNA translation, is yet to be explored in their regulatory roles in cellular heterogeneity and early cell fate determination during early mammalian embryogenesis, despite the fact that their dynamic landscapes and regulations in epigenetic remodeling have already described.The singlecell omics and multi-omics techniques available for research on mammalian preimplantation embryos are summarized in Table 1.Considering the similarities in both size and volume between mature mammalian oocytes and single cells of early preimplantation embryos, single-cell sequencing techniques applied to oocytes are included in this section as well.
DNA methylation, usually referred to as 5-methylcytosine (5mC) methylation, plays a vital role in early mammalian embryogenesis [73][74][75][76].The acquisition of a whole-scale DNA methylome is frequently based on bisulfite treatment fol-lowed by sequencing (BS-seq) using a large amount of input material [77].Advances in some techniques, such as singlecell reduced representation bisulfite sequencing (scRRBS), single-cell post-bisulfite adaptor tagging DNA methylome sequencing (scPBAT-seq), single-cell BS-seq (scBS-seq), and single-cell MspJI-based strand-specific 5mC sequencing method (scMspJI-seq), have overcome the inherent difficulty of input cells, through which single-cell dynamic DNA methylation landscapes in early mouse and human embryogenesis have been successfully obtained [73][74][75][76].In mouse 4-cell embryos, genetic lineage can be traced by DNA 5mC [75].As an oxidized derivative of 5mC, 5-formylcytosine (5fC) regulates gene expression in maternal and paternal genomes in early mammalian embryos [78].Single-cell chemical-labelingenabled C-to-T conversion sequencing (scCLEVER-seq) was developed to reveal the single-cell landscape of DNA 5fC and provides important resources for functional exploration of epigenetic reprogramming in single cells and studies related to ten-eleven translocation protein family (TET)-dependent active demethylation during early mammalian embryonic development [79,80].
long-range interaction loci with deep sequencing [81].Singlenucleus or single-cell Hi-C (snHi-C/scHi-C) reveals highorder chromatin and chromatin conformation profiles in mouse oocytes, zygotes, and blastomeres of preimplantation embryos [82][83][84], which have provided us with a powerful tool for deciphering cell-to-cell variability of chromatin architecture and cell fate determination.RNA N 6 -methyladenosine (m 6 A), the most abundant RNA methylation, plays a critical role in cellular regulation and function.Benefiting from the single-cell m 6 A sequencing (scm 6 A-seq) technique, single-cell landscapes of m 6 A methylome and transcriptome have been profiled simultaneously in mouse oocytes and single blastomeres of early mouse embryos.Interestingly, m 6 A-dependent asymmetries emerge in the blastomeres of 2-cell embryos, possibly indicating the potential significance of m 6 A in early cell fate determination [85].
Although single-cell transcriptomics methods were developed over a decade ago, translatomics at single-cell resolution was not achieved until 2021, when Vanlnsberghe et al. and Brannan et al. invented single-cell ribosome sequencing (scRibo-seq) and surveying targets by APOBEC-mediated profiling of ribosomes (Ribo-STAMP) separately [86,87].These two methods are high-throughput and are available for cell lines rather than for preimplantation embryonic cells.More recently, dual-omics methodology of single oocyte, transcriptome and translatome sequencing (T&T-seq), successfully characterized the single-cell profiles of the transcriptome and translatome simultaneously during mouse and human oocyte maturation [88].These single-cell techniques for the translatome offer robust tools for future investigation of both cell-to-cell translational differences and the mechanisms of translational regulation during early mammalian embryogenesis.
Multi-omics profiling, which integrates the genome, transcriptome, epigenome, and proteome, provides a powerful approach to simultaneously capture multi-layer profiles.The integrated multi-omics database DevOmics also provides a convenient tool for investigators to study the molecular regulators and relative mechanisms in early mouse and human embryos [89].Single-cell chromatin overall omic-scale landscape sequencing (scCOOL-seq) can simultaneously analyze chromatin state, nucleosome positioning, DNA methylation, copy number variation, and ploidy in the same individual mammalian cell.Using scCOOL-seq, the dynamics and heterogeneity of DNA methylation have been described in the preimplantation embryos of mice and humans [90,91].Improved scCOOL-seq (iscCOOL-seq) also dissects complex epigenetic alterations during mouse oocyte growth [92].The combined approach of the whole-scale transcriptome with DNA methylome, single-cell genome-wide methylome and transcriptome sequencing (scM&T-seq), enables the identification of susceptibility to glucocorticoids in human blastocyst embryos [93].Single-cell nucleosome occupancy, DNA methylation, and RNA expression sequencing (scNOMeRe-seq) [94] and single-cell chromatin accessibility, RNA barcoding, and DNA methylation sequencing (scChaRM-seq) [95] are tripartite-omics techniques, which are capable of simultaneously acquiring the profiles of the DNA methylome, transcriptome, and chromatin accessibility in single blastomeres in mouse preimplantation embryos (where genetic lineages were remodeled by DNA methylation in 4-and 8-cell embryos) and in single human oocytes, respectively.All these techniques provide important resources for comprehensively understanding the functional regulatory landscape in early mammalian embryos.
Although single-cell omics and single-cell multi-omics techniques, including DNA methylome, chromatin architecture, chromatin accessibility, RNA methylome, and transcriptome, have been utilized for analyzing dynamic landscapes of early mammalian embryos, few studies so far have focused on early cell-to-cell heterogeneity and cell fate determination, except transcriptome, which are worthy of further investigation.

Concluding remarks and perspectives
Spatial and temporal accuracy in early embryogenesis is crucial for subsequent fetal development; thus, early cell fate decisions are a matter of widespread interest.In mice, the first two rounds of cell fate are determined by transcription factors and signaling pathways at the morula and blastocyst stages, which are also affected by the establishment of cell polarity at 8-cell stage.Cell fate regulation at the 2-to 4-cell stages is characterized by cellular heterogeneities in RNA transcription, histone modification, and transcription factor kinetics.Although numerous studies have extended our understanding of cell fate determination during early embryonic development, many questions remain unanswered.For example, is there other important regulatory layers, such as RNA translation, asymmetrically distributed at the 2-to 4-cell stages or during an earlier stage and associated with early cell fate segregation?Do other types of histone modifications, such as histone acetylation, mediate distinct regulatory axes to determine cell fate?What is the origin of cell-to-cell heterogeneity, and whether this type of heterogeneity occurs randomly or procedurally?As the development of early embryos is punctual and regular, what mechanisms drive the switching between different lineage patterns and adjust the mismatch between the position and fate of a certain cell?In view of the powerful cell totipotency and embryonic plasticity in mouse 2-and 4-cell embryos [96][97][98], whether the molecular asymmetry can be inherited or reconstructed in new 2-and 4-cell embryos developed from ''half"-and ''one-quarter"-cleaving embryos is an exciting scientific question waiting to be explored.The answers to these questions will provide a clear view of the detailed mechanisms for early cell fate regulation.
Currently, single-cell omics/multi-omics techniques enable the accessibility of genetic and epigenetic profiles, which facilitates a comprehensive understanding of the development of early embryos.Except for single-cell transcriptomes, more single-cell technologies at different layers, such as RNA modifications, RNA translation, and proteome, are waiting to be innovated, integrated, and applied to explore the cell fate decision of early embryos.Advanced techniques for capturing subcellular structures may pave the way for parsing intra-cell heterogeneity caused by asymmetric molecular distribution.
Additionally, despite the fact that mouse and human preimplantation embryonic development are relatively conserved in many mechanisms and remarkably similar in morphogenesis, the understanding of early human embryogenesis is still rare.Most mechanisms and molecular events involved in early cell fate determination have been studied in mouse models.However, human lineage specification and blastocyst formation using contradictory models remain elusive.Moreover, com-

Figure 1
Figure 1 Preimplantation embryo development in mouse and human Schematic of the early stages of mouse (top panel) and human (bottom panel) embryonic development at different time points.The timeline in the middle indicates the embryonic days (E) from E0 to E6.0.Color legends at the top and bottom indicate the respective embryonic cells and lineages in mouse (top) and human (bottom).TE, trophectoderm; ICM, inner cell mass; PE, primitive endoderm; EPI, epiblast; ZGA, zygotic genome activation.

Figure 2
Figure 2  Signaling pathways involved in ICM-TE and PE-EPI fates A. Hippo and Notch signaling pathways involved in the ICM and TE cell fate allocations.In the Hippo pathway, unphosphorylated YAP enters the nucleus to interact with TEAD4, and thus activates the expression of Cdx2 and Gata3 in TE cells (bottom panel); phosphorylated YAP maintains in cytoplasm, resulting in the up-regulated expression of Oct4 and Sox2 in ICM cells (top panel).Amot regulates the activation of YAP.In the Notch pathway, NICD and RBPJ form a complex in the nucleus of TE cells, activating the expression of Cdx2 and Gata3.B. FGF signaling pathway involved in the PE and EPI cell fate allocations.In EPI cells (bottom panel), NANOG coordinates with OCT4 and SOX2 to up-regulate Fgf4 expression and repress Gata6 expression.In PE cells (top panel), FGF4 secreted from EPI cells activates the FGF/MAPK signaling, and further up-regulates GATA6 and OCT4 which activate the expression of Sox17, Gata4, and Sox7.p, phosphorylation.

Figure 3
Figure 3 Cellular heterogeneities of cell fate regulators control differential fate decision A schematic of the heterogeneous levels of fate regulators at the 2-and 4-cell stages contribute to the heterogeneous expression of lineage transcription factors at the morula stage.Left: asymmetric LincGET expression and differential CARM1-containing paraspeckles are present in blastomeres of 2-cell embryos.Middle: heterogeneous CARM1, H3R26me, and OCT4/SOX2 kinetics are present in the blastomeres of 4-cell embryos.Right: regulations by transcription factors are presented at the blastomeres of morula embryos.H3R26me, histone H3 methylation at arginine 26.

Table 1
Single-cell sequencing techniques applied to mouse and human preimplantation embryos