Abstract
In brief
Human embryogenesis still remains largely unexplored. This review helps identify some of our current gaps in knowledge pertaining to preimplantation development, which may have implications for understanding fundamental aspects of human development, assisted reproductive technologies, and stem cell biology.
Abstract
Preimplantation development is arguably one of the most critical stages of embryogenesis. Beginning with the formation of the totipotent zygote post-fertilization, a series of cell divisions, and a complex coordination of physical cues, molecular signals and changes in gene expression lead to the formation of the blastocyst, a structure capable of implanting into the uterine wall. The blastocyst is composed of more specified cellular lineages, which will give rise to every tissue of the developing organism as well as the extra-embryonic lineages which support fetal growth. While the mouse has been used as a model to understand the events of preimplantation development for decades, in recent years, an expanding body of work has been conducted using the human embryo. These studies have identified some crucial species differences, particularly in the transcriptional and spatio-temporal expression of lineage markers and responses to cell signaling perturbations. This review compares recent findings on preimplantation development in mouse and human, with a focus on the specification of the first cellular lineages. Highlighting differences and noting mechanisms that require further examination in the human embryo is of critical importance for both the accurate translation of results from the mouse model and our overall understanding of mammalian development. We further highlight the latest advancement in reproductive research, the development of the 3D stem cell-based models known as ‘blastoids’. The knowledge discussed in this review has major clinical implications for assisted reproductive technologies such as in vitro fertilization and for applications in stem cell biology.
Introduction
Preimplantation development, the period extending from fertilization to implantation, is arguably one of the most critical stages of embryogenesis. Common major developmental milestones among mammalian species that take place during this time have been extensively studied in animal models. After fertilization, the newly formed zygote switches its genomic control from maternally derived to embryo-specific transcription in the process of embryonic genome activation (EGA). This is followed by a few rounds of cell division until embryonic cells undergo an increase in cellular adhesion known as compaction, where they lose clear boundaries with neighboring cells and the embryo becomes a morula. Polarization also takes place within cells around this time, leading to the formation of the fluid-filled blastocyst cavity and the blastocyst via cavitation. Coinciding with blastocyst formation, molecular and spatio-temporal mechanisms are employed such that the previously totipotent blastomeres begin to acquire genetic signatures of more specified cell lineages. It is generally considered that the first of these lineage decisions results in the formation of the inner cell mass (ICM; prospective embryo proper and yolk sac) and the trophectoderm (TE; prospective placenta). The second lineage decision further specifies the cells of the ICM into the pluripotent epiblast (EPI; prospective embryo proper and mesoderm of prospective yolk sac) and the primitive endoderm (PE; endoderm of prospective yolk sac).
Much of what we have learned about these critical developmental milestones have come from studies conducted on mice. The mouse has been used as a model for the early human embryo for decades, as embryo morphologies between the species are quite similar (Blakeley et al. 2015). Furthermore, the mouse is a small, tractable model, which allows researchers to overcome the limitations encountered with larger animal models, such as technical difficulties, cost, and facility housing needs. The mouse model also circumvents the limited accessibility, ethical and legal restrictions associated with the use of human embryos. Additionally, the genome and epigenome are well established in the mouse, allowing for detailed molecular assessment and the design of transgenic models to better understand the underlying mechanism(s) driving preimplantation development. Further, the mouse is a high-throughput model as protocols for superovulation are well established, providing researchers with numerous embryos per mouse. With advancements in biotechnology, increased access to human embryos, and the development of new methodologies enabling single-cell sequencing, studies using the human embryo have become more feasible. As a result of the increased number of human-focused studies, species differences have been identified over the years in the transcriptional and spatio-temporal expression and regulation of lineage markers at both the protein and mRNA levels (Niakan & Eggan 2013, Blakeley et al. 2015, Petropoulos et al. 2016), cell signaling responses (Kuijk et al. 2012, Roode et al. 2012), X-Chromosome inactivation (XCI), and epigenetic regulation of development (Niakan & Eggan 2013, Blakeley et al. 2015, Petropoulos et al. 2016, Posfai et al. 2017, Boroviak et al. 2018, Gerri et al. 2020). Taking these differences into consideration, a more comprehensive understanding of human preimplantation and cross-species development is needed, which will enable more accurate translation and extrapolation of results toward human biology, infertility, and stem cell derivation. A better understanding of human preimplantation development will also effectively allow researchers to choose the most appropriate model, depending on the aspect of development they are interested in elucidating. Of note, one major limitation of human embryo research is the lack of human embryos sourced in vivo, and as such, we cannot effectively compare in vivo vs in vitro and some of the phenomena reported may be artifacts of the ex vivo environment. Nonetheless, these embryos do give rise to viable offspring, suggesting a close similarity to their in vivo counterparts. Knowledge of how these processes occur in human development has major clinical implications for assisted reproductive technologies (ART) such as in vitro fertilization and for applications in stem cell biology. In this review, recent findings pertaining to the processes of preimplantation development in the mouse and human are compared side-by-side, with the purpose of highlighting differences and noting mechanisms that require further examination in the human embryo.
Morula dynamics: compaction and polarization
In the mouse, preimplantation development spans 5 days encompassing the period from zygote formation (E0.5, where E is the embryonic day) to the mature implanting blastocyst (E4.5-E5) (Chazaud & Yamanaka 2016, Molè et al. 2020). After fertilization, the oocyte completes meiosis with the release of the second polar body. The first embryonic cleavage occurs 16–20 hours after fertilization (two-cell stage, E1.5), and the subsequent cleavage cycles take place at roughly 12-h intervals (Chazaud & Yamanaka 2016). After completing three cleavage rounds, the mouse embryo is at the 8-cell stage where compaction and polarization occur (Chazaud & Yamanaka 2016). In humans, the timing of these dynamics differs from that of the mouse (Fig. 1). Preimplantation development is generally considered to span 7 days, and the embryo does not reach the 8-cell stage until a slightly later timepoint, at E3 (Niakan & Eggan 2013, Petropoulos et al. 2016, Gerri et al. 2020). Despite this difference in developmental timing, compaction and polarization remain essential in both species for proper embryo development.
Compaction
Compaction is evident in mouse embryos at the 8-cell stage (Fig. 1) (Chazaud & Yamanaka 2016). It was initially proposed that cell–cell adhesion mediated by the homophilic interactions of E-cadherin between two neighboring cells might decrease surface energy at cell–cell interfaces thereby directly driving the compaction of the mouse embryo (Foty & Steinberg 2005). However, more recent findings suggest that the tensile force powering compaction is generated by the contractility of the actomyosin that starts at the 8-cell stage in mouse embryos (Maître et al. 2015). According to this study, cadherin adhesion molecules are essential to mechanically couple the cortices of contacting cells, and their role in compaction is to keep contractility low at cell–cell contact. While those two models disagree on the component producing the force necessary for compaction, they both suggest that E-cadherin is essential for that process.
In humans, the majority of research regarding compaction has primarily been focused on timing. Compaction in the human occurs later than in the mouse, at approximately the 16-cell stage, coinciding with E4.0 (Fig. 1) (Molè et al. 2020, Zhu et al. 2021). The reason for this delay is unknown, although one study suggests that it may be related to differences in the EGA timing between these species (Zhu et al. 2020). Time-lapse imaging of human embryos showed an elongated 8-cell-to-compaction transition compared to the mouse, with multiple cell divisions that investigators termed ‘pre-compaction’ (Gerri et al. 2020). The difference in timing could be an indication of different mechanisms driving compaction between the species, but further studies are required to better understand this process in humans.
Polarization
Polarization in mouse embryos is initiated at the 8-cell stage (Chazaud & Yamanaka 2016) and is completed by the end of the fourth cleavage round (16-cell, E3.0) (Molè et al. 2020). This event is characterized by the rapid acquisition of distinct characteristics by the cells of the embryo, with the apical regions facing outward and the basolateral regions facing inward. Actin accumulates apically, along with the polarity complex composed of partitioning defective proteins 3 and 6 (PAR3 and PAR6), atypical PKC (aPKC) and ERM proteins (ezrin, radixin, moesin), all enclosed by an actomyosin ring (Zhu et al. 2020). The basolateral domain, on the other hand, becomes enriched with cell adhesion proteins (Zhu et al. 2020). Three factors have been identified to trigger polarization in the mouse embryo: TFAP2C, TEAD4, and RhoA, which is accomplished in two phases: centralization and expansion (Zhu et al. 2020). During mouse embryonic development, the centralization phase is initiated when TFAP2C and TEAD4 reach a threshold level of expression after EGA. This then leads to the activation of genes associated with actin cytoskeleton remodeling, which allows polarity proteins (such as ezrin) to anchor to the apical surface by an unknown mechanism. Subsequently, activation of RhoA GTPase, which completes the proper apical domain formation, promotes the expansion of the centralized polarity proteins (Zhu et al. 2020). Polarization is important in the initiation of the TE program, as supported by embryo manipulation studies. Injection of Tfap2c and Tead4 mRNA into one blastomere at the two-cell stage resulted in precocious polarization at the four-cell stage and with premature activation of RhoA GTPase leading to the premature expression of TE transcription factors (CDX2 and GATA3) (Zhu et al. 2020). These studies provide evidence that the timing of EGA and cell polarization in the mouse embryo are connected and required for the initiation of the TE program in the outer cells.
Polarization in the human embryo, similar to compaction, is slightly delayed compared to the mouse and is initiated at E3.25 and completed by E4.0 (Zhu et al. 2021). Similar to the mouse, polarization in the human occurs in two phases. The first step is the polarization of F-actin, which takes place during compaction (Zhu et al. 2021). The second step is the polarization of the Par complex, which has been demonstrated by immunofluorescence detection of the PARD6 protein, part of the Par complex (Zhu et al. 2021). The PLC–PKC pathway may regulate polarization of the Par complex in the same way as in the mouse, as inhibition of this pathway at the morula stage resulted in a lower proportion of polarized cells and a reduced apical enrichment of PARD6 compared to controls (Zhu et al. 2021). In contrast to mouse embryos, where polarization and asymmetric reassembly of cell polarity components trigger the TE program in outer cells, in humans, blastomeres initiate the expression of TE factors independently of the polarity machinery, although polarization is thought to reinforce the TE fate (Zhu et al. 2021). Seminal studies in mice established our understanding of the mechanisms controlling polarization and its link with lineage segregation, on which similar research in humans has been based. While studies investigating polarization in human embryos remain sparse, the limited existing studies suggest that mechanisms of embryo polarization are mostly conserved between the species. However, more human-specific studies are needed to confirm this preliminary conclusion, particularly the role of polarization in lineage segregation.
The first cell fate decision (ICM/TE specification)
As mentioned earlier, studies performed in mice have led to the identification of two lineage segregation events that take place during preimplantation development. The first of these events involves the segregation of the TE from the pluripotent cell population, the ICM. Molecular differences amongst the cells that may distinguish these cell types are detectable as early as the 8-cell stage, with compaction and polarization playing an important function in driving those differences (Marikawa & Alarcón 2009, Chazaud & Yamanaka 2016). In the mouse, cells become fully committed to either the TE or ICM lineages at approximately the 32-cell stage coinciding with cavitation, and maturation of the TE lineage happening shortly after, at the late 32-cell stage (Posfai et al. 2017). The ICM lineage matures slower, committing around the time of the second cell fate decision, at the 64-cell stage (Posfai et al. 2017).
In humans, the timing of lineage segregation is believed to occur after cavitation has been initiated, with a mutually exclusive expression of key lineage markers initiating at late E5, suggesting a disconnect between morphology and lineage specification (Niakan & Eggan 2013, Petropoulos et al. 2016, Meistermann et al. 2021). The precise timing of cell fate specification in the human embryo and whether the human embryo contains a distinct ICM population or gene signature associated with an ICM lineage has been a topic of debate (Petropoulos et al. 2016, Meistermann et al. 2021, Radley et al. 2022). Using a time-course approach combined with immunostaining of key lineage markers identified from the mouse, a seminal study profiled the spatio-temporal expression of key lineage marker genes in the human embryo from zygote to blastocyst (E6) (Niakan & Eggan 2013). This study identified ubiquitous expression of POU5F1 at E5 (39–79 cell embryos) and the emergence of CDX2 expression at E5 (54 cell embryo). Further, to determine the timing of PE formation, they stained for SOX17 and observed at E5 (89 cell embryo) that SOX17 was largely co-expressed with POU5F1 (similar to the co-expression observed at mid-E5 in the Petropoulos et al. study). However, another embryo at E5 was identified with cells solely expressing SOX17, which could be interpreted as the emergence of a PE, though the majority of cells displayed distinct expression from E6 (181 cell embryo) onward (Niakan & Eggan 2013). They concluded that embryos display heterogeneous expression of SOX17 in the ‘ICM’ at E5 with distinct restriction to the putative PE by E7 (Niakan & Eggan 2013). This study highlights the variability in human embryo development within a developmental timepoint as indicated by the large variation in cell number per embryo (e.g. E5 ranged from 39 to 113 cells and E6 ranged from 181 to >256 cells) as well as the limitations associated when utilizing a limited number of selected markers. Nonetheless, it provided some of the first frameworks for understanding cell fate and preimplantation development in humans.
The increase in access to human embryos paired with the advancement of scRNA-seq technologies has provided the opportunity to investigate previously identified lineage markers throughout preimplantation development and the simultaneous expression of multiple markers, termed gene signatures to tease out the cell fate events. The most extensive human embryo single-cell transcriptome study to date profiled embryos from E3 to E7 (Petropoulos et al. 2016). As one of the goals was to determine cell fate specification, they categorized the E5 embryos into early-, mid-, and late-E5, which were based on gene signatures, developmental time, and morphology. In this study, cells from the early- (around the time of cavitation) and mid-E5 (blastocyst cavity starting to expand) embryos were determined to co-express markers for all cell fates, suggesting that there is a disconnect between embryo morphology and cell fate in the human, unlike the mouse where fate is initiated in the morula in a cell polarization and positional-dependent manner (Cockburn & Rossant 2010, Petropoulos et al. 2016). By late-E5 (expanded blastocyst cavity), a distinct TE and a population of inner cells were identified (Petropoulos et al. 2016). Further, a well-defined ICM signature, like in the mouse, could not be determined; however, distinct populations expressing markers for EPI and PE were identified. From this, the authors concluded that the human embryo lacks a distinct ICM population, and instead the three primary lineages are simultaneously specified. However, sequential, systematic sampling of embryos at E5 was not performed in this study, and as such, a key embryo stage may have been missed or combined with the other timepoints and as such have the gene signatures masked. Further, the availability of algorithms to handle scRNA-seq data was far less advanced and the authors relied on successive rounds of principal component analysis to determine the lineage populations. This is of great importance as one of the limitations of scRNA-seq data is the ability to discern noise which results from dropout and capture efficiency. Finally, the determination of cell lineage was guided based on previously established markers from the mouse and it is not far fetching, given differences in naïve/primed pluripotency and differences in identified lineage markers from comparisons to the mouse, that perhaps the human ‘ICM’ may be composed of different markers than the mouse (Blakeley et al. 2015).
Some of the challenges in staging embryos from previous work have recently been addressed by incorporating time-lapse imaging of human embryos. This allows for systematic sampling and provides the authors with the ability to associate developmental stages with molecular events (Meistermann et al. 2021). The resulting observations used the ESHRE naming convention for human blastocysts, whereby the developmental progression is indicated by a number in the format of ‘B#’ (Alpha Scientists in Reproductive Medicine and ESHRE Special Interest Group of Embryology 2011). The earliest blastocyst is denominated as B1, a slightly more progressed blastocyst as B2, and so on and so forth and the authors relied on the expansion of the cavity, the thinning of the zona and hatching to classify their embryos. They leveraged previously published scRNA-seq datasets and incorporated 25 new embryos (150 cells) spanning the morula stage to the ‘B5’ expanded blastocyst (~E6-E7) from their time-lapse imaging. The scRNA-seq data of individual cells from the same embryos was analyzed together. They then projected their data onto the newly generated pseudotime, which was created using previously published datasets (Yan et al. 2013, Petropoulos et al. 2016, Fogarty et al. 2017, Meistermann et al. 2021). Additionally, more advanced tools were incorporated for analysis, including uniform manifold approximation and projection, RNA velocity, and pseudotime (Meistermann et al. 2021). Meistermann et al. noted that inner cells either co-expressed markers of EPI/PE or only EPI. Similar to what Petropoulos et al. noted, an ambiguity with distinct genes associated with EPI/ICM fates was observed and thus they also concluded that a distinct ICM population was not present in the human embryo. Instead, this paper proposed that distinct TE/EPI lineages emerge during the transition between the B2 and B3 blastocyst stages (40–85 cells) and that by B4 (~130 cells), all three lineages are present. They further proposed that perhaps the PE lineage differentiates from the EPI. This was supported by results pertaining to IFI16, which Meistermann and others have identified as a potential marker of human EPI (Petropoulos et al. 2016, Boroviak et al. 2018, Meistermann et al. 2021). IFI16 was found to be restricted to the ICM (in cells positive for both EPI and PE markers) at the B4 blastocyst stage, but expression became specific to the EPI when cultured to post-implantation stages (Meistermann et al. 2021). This study provided evidence of a two-step model for fate specification in the human, similar though not identical to the mouse. One limitation of this study is the low number of additional cells incorporated in the analysis. The authors report the addition of 25 new embryos that were assessed with time-lapse; however, this corresponded to only 150 new cells and the identification of only 1 PE cell. It is therefore not far-fetched to wonder if additional cells related to the PE are present and perhaps were missed.
The latest study claims to have identified, for the first time, a distinct ICM gene signature in the human embryo. Here the authors develop a new mathematical framework termed entropy sorting (ES) to circumvent one of the major limitations of scRNA-seq analysis which is the ability to discern technical and biological noise and to accurately extract features (genes) associated with cell types (Radley et al. 2022). They first validated ES by curating synthetic data to assess the performance of their algorithm’s functional feature amplification via entropy sorting and entropy sort feature weighting, which encode the mathematical framework of ES. They then leveraged the data fromMeistermann et al. (2021), which included data generated by Yan et al. (2013), Petropoulos et al. (2016), and Fogarty et al. (2017). Unlike the scRNA-seq studies mentioned earlier, whose attempt to identify an ICM was slightly misled by mouse work, Radley et al. used the 3700 genes generated by ES and were able to identify a handful of genes expressed in a sub-population of cells at E5 which did not co-exist with either the EPI or PE. This is the first work to identify a distinct gene signature associated with the ICM (e.g. FGF1, S100Z, SPIC, PIMREG, BHMT, LAMA4, PRSS3), providing further support of a two-step model in cell fate specification. Further, in contrast to Meistermann, connectivity between the EPI and ICM as well as the PE and ICM was identified using nearest neighbor analysis, concluding that both EPI and PE emerge from the ICM. Importantly, the protein expression of LAMA4 was validated in human embryos (E5–E7) by immunostaining and only observed in inner cells. One limitation of this study is that the mRNA expression of some of the other genes identified as indicators of an ICM (e.g. S100Z, SPIC, PIMREG, BHMT) appear to overlap with the ‘pre-lineage’ or morula staged embryos. Another drawback of this study is that cells from all E5 embryos have been grouped together, making it difficult to align with previous datasets and determine whether it is at the early- (B2), mid- (B3), or late-(B3–B4) E5 embryo where this ICM population emerges. This is particularly important as it could help better define the specific window during E5 and as such more precisely guide researchers on where to focus efforts in order to elucidate the mechanism(s) driving cell fate specification. This more detailed timing would also provide clarity as to whether the disconnect between morphology and cell fate still pertains. Finally, functional studies to validate the presence of a bona fide ICM remain to be conducted. Perhaps experiments such as manipulation of LAMA4 or the other ICM signature genes identified could be performed to observe the impact on ICM formation. Additionally, examining interactions with ICM and known pluripotency genes, or lineage tracing experiments to determine whether these ICM cells do in fact give rise to EPI and PE populations would be informative.
In general, many of these works in the human have heavily relied on the scRNA-seq analysis to determine the temporal emergence of cell fate. Further, there are numerous inconsistencies in the literature in terms of the terminology used and a lack of transparency as to the precise day post fertilization or the number of cells present in the embryos assessed. This is particularly concerning when data is being leveraged and ‘blastocysts’, which include E5, E6, and E7, are being merged together. Finally, the use of fresh vs frozen embryos may add additional noise to the data generated, not only via technical artifact related to the mRNA expression but also on the embryo staging as frozen embryos may not have been vitrified at the exact ‘window’ during E5 and have a slight developmental time lag associated with thawing. Finally, the development of individual human embryos is more variable compared to the mouse, and as such, it is important to consider multiple ‘benchmarking’ parameters such as developmental time, morphokinetics, and cell number for embryo staging similar to Meistermann et al. Care should be applied when incorporating previous data as references and an attempt to properly align the developmental labels from these will be helpful. Although we are inching closer toward a better understanding of fate specification in humans, many gaps still remain.
Figure 1 provides a summary of lineage specification in both the human and mouse embryo, depicting the three proposed models in the human and the spatio-temporal expression of key fate genes that have been validated at the protein level. While the formation of a distinct ICM/TE cell population is the most intuitive, given the lack of functional studies to support this new finding, we have kept all three models.
Comparative schematic of mouse (top) and human (bottom) preimplantation development, depicting the number of cells, the timing of the various morphological stages, the temporal protein expression of known lineage markers, and proposed models for human lineage specification. The center axis (purple) depicts the embryonic day (E) of the associated event.
Citation: Reproduction 165, 4; 10.1530/REP-22-0144
Mechanisms of mouse ICM/TE specification
In mouse embryos, the Hippo signaling pathway is the major signaling cascade implicated in the first lineage segregation, which has been reviewed extensively (Chazaud & Yamanaka 2016). Beginning in the outer cells of the 8-cell embryo, the apical polarity complex sequesters LATS 1/2 and angiomotin (AMOT) to the apical domain, holding them inactive (Hirate & Sasaki 2014). This inactivation of Hippo allows YAP, or a redundant protein known as WW domain-containing transcription regulator I (WWTR1, also known as TAZ), to translocate to the nucleus and bind to TEAD4. The TEAD4-YAP/TAZ dimer activates CDX2 and Gata binding protein 3 (GATA3) (Guo et al. 2010), while simultaneously repressing the expression of SRY-Box Transcription Factor 2 (SOX2) (Frum et al. 2018). In inner cells, the lack of an apical domain allows AMOT and LATS1/2 to interact with E-cadherin at adherent junctions, activating the Hippo pathway (Hirate & Sasaki 2014). Active LATS1/2 is then able to phosphorylate the transcription factors YAP and TAZ, preventing their translocation to the nucleus and TEAD4-mediated regulation of transcription factors. Therefore, activated Hippo in the inner cells prevents the expression of CDX2 and GATA3, while permitting the expression of SOX2, allowing the establishment of a transcription network supporting pluripotency, ICM stabilization, and expansion (Wicklow et al. 2014). It has also been shown that AMOT can bypass the classical Hippo pathway to regulate YAP, implying that a combined Hippo-dependent and Hippo-independent mechanism may control YAP localization in the preimplantation mouse embryo (Leung & Zernicka-Goetz 2013). Over time, this molecular interplay results in the mutually exclusive expression of CDX2 in outer cells and SOX2 in inner cells prior to blastocyst formation (Frum et al. 2018). The Hippo pathway is also involved in resolving embryonic cell fate conflicts in the mouse embryo. Around the 32-cell stage, outer cells with partial apolar phenotypes activate Hippo (Frum et al. 2018). The activation of the Hippo pathway in those apolar outer cells promotes either apoptosis or internalization and transition to the ICM fate (Frum et al. 2018). This internalization coincides with the loss of TE plasticity and with the formation of mature tight junctions among outside cells (Frum et al. 2018), suggesting that TE is committed by this time as has been recently demonstrated (Posfai et al. 2017).
Mechanisms of human ICM/TE specification
In the human preimplantation embryo, studies investigating the mechanisms and timing of ICM-TE specification are less abundant than those conducted in the mouse. However, the gradually growing body of human-focused work has provided evidence that there are discrepancies in the temporality of gene expression between the species, which are potentially indicative of underlying mechanistic differences. For example, at both the protein and mRNA level, CDX2, which marks the development of the mouse TE at the early morula stage, is not appreciably expressed in human embryos prior to blastocyst formation (Niakan & Eggan 2013, Petropoulos et al. 2016, Molè et al. 2020). In addition, while in the mouse, ICM and TE markers become mutually exclusive in the cells of the morula (Chazaud & Yamanaka 2016), in the human morula (E4), TE genes such as GATA2 and 3 remain co-expressed with pluripotency markers SOX2 and KLF17 at both the mRNA and protein level (Blakeley et al. 2015, Petropoulos et al. 2016, Kilens et al. 2018, Gerri et al. 2020). Moreover, TCFAP2C/AP2γ, a key transcriptional regulator of TE in the mouse, is localized to both the TE and EPI in the human blastocyst (Blakeley et al. 2015, Gerri et al. 2020). Further, while in both species TE-associated genes such as CDX2, GATA3 and KRT18 are highly expressed in the late blastocyst TE, transcripts for ELF5 and EOMES are absent in the human, suggesting they are not involved in TE specification (Blakeley et al. 2015, Petropoulos et al. 2016). Whether EOMES is expressed in other cell types in the human embryo is unclear, with one study reporting expression in the PE (Blakeley et al. 2015), while another found no expression of EOMES from the morula to the blastocyst stage (Petropoulos et al. 2016). An important player for human TE specification may instead be GATA3, as its protein expression is initiated at the morula stage (E4) in all cells (Gerri et al. 2020, Zhu et al. 2021) and is then quickly restricted to outer cells (Gerri et al. 2020). Generally, studies are in agreement that TE differentiation is regulated by cell polarity in both humans and mice (Gerri et al. 2020, Zhu et al. 2021), though whether the initiation of GATA3 expression is dependent on cell polarity remains unclear (Zhu et al. 2021). While suggested TE specifiers TEAD1, YAP, and GATA3 mostly co-localize in polarized outer cells of the human 16-cell morula, they are also found in some cells of compacting embryos before polarity is established, supporting the notion that TE differentiation may be initiated independently of polarity (Regin et al. 2022). Further work is warranted to identify the functional impact of the identified discrepancies in gene expression between the mouse and human TE markers and to tease out the precise mechanism(s) driving TE fate in humans.
As discussed earlier, the Hippo pathway is thought to be the main driver of ICM-TE specification in the mouse, and studies investigating this pathway’s role in the human have shown that there are some similarities between the TE programs. Similar expression patterns of aPKC and AMOT at the apical membrane of outer cells have been found in mouse and human morula stage embryos (Gerri et al. 2020), suggesting that the Hippo pathway is inactive in the outer cells of both species. Moreover, inhibition of aPKC leads to downregulation of nuclear YAP in the outer cells, suggesting that the pattern of expression of the Hippo pathway is similar between human and mouse TE; however, the functional role of YAP remains to be directly examined in the human embryo. While this paper provides evidence in support of the involvement of the Hippo pathway in the establishment of the TE, it suggests that its role in the establishment of the ICM is unclear and differs from the mouse (Gerri et al. 2020). For example, SOX2 is quickly restricted to inner cells via regulation by members of the Hippo pathway as early as the 8- to 16-cell morula stage in the mouse, but remains ubiquitously expressed in the human even in the early blastocyst (Gerri et al. 2020). In addition, in the mouse, knockout of TEAD4 (interacts with YAP/TAZ to activate TE genes) inhibits blastocyst formation and results in the downregulation of CDX2 and GATA3 expression (Nishioka et al. 2008, Ralston et al. 2010). In contrast, in humans, knockout of TEAD4 did not impact the rate of blastocyst formation or GATA3 expression, suggesting GATA3 does not act downstream of TEAD4 (Stamatiadis et al. 2022). Instead, it has been proposed that YAP may interact with TEAD1 to initiate TE differentiation, as its expression precedes TEAD4 and co-localizes with YAP and GATA3 in precursor TE cells at the morula stage (Regin et al. 2022). Therefore, although the Hippo pathway may be involved in reinforcing the TE program in humans, its mechanisms of action may differ from the mouse. Further, there is the possibility that additional signaling pathways such as the MAPK pathway and WNT pathway play an important role (Blakeley et al. 2015, Krivega et al. 2015) or perhaps there is cross-talk amongst the aforementioned pathways. In human embryos, degradation of the WNT signaling molecule beta-catenin leads to inhibition of TE fate, with decreased CDX2 protein expression in outer cells (Krivega et al. 2015). Accordingly, while in control E6 embryos, there is no expression of EOMES (Blakeley et al. 2015, Krivega et al. 2015, Petropoulos et al. 2016), addition of the WNT signaling ligand WNT3 to culture medium leads to promotion of EOMES expression (Krivega et al. 2015), suggesting that WNT signaling may promote TE development in the human. Further studies are required to dissect the specific roles of these signaling pathways and any crosstalk which may occur in driving TE segregation in humans.
In humans, as in mice, the transcription factors POU5F1, SOX2, and NANOG comprise the core of the pluripotency network (Boroviak et al. 2018). In contrast to the mouse embryo where POU5F1 is gradually restricted to the ICM at E3.5 by a reciprocal downregulation with CDX2 (Chazaud & Yamanaka 2016), POU5F1 is expressed equally in all cells of early human blastocysts (Niakan & Eggan 2013), potentially due to the lack of CDX2 expression in early human blastomeres. Nonetheless, POU5F1 has been shown to be essential for human blastocyst formation as microinjection of targeting guide RNAs resulted in impaired blastocyst formation and downregulation of pluripotency genes such as NANOG (Fogarty et al. 2017). Also in contrast to the mouse, SOX2 is not the first pluripotency factor restricted to the inner cells at the morula stage but instead occurs after the restriction of NANOG in the blastocyst (Gerri et al. 2020). This suggests that the temporal expression of pluripotency genes is different between the two species and raises the possibility that there may also be differences in the mechanisms controlling the expression dynamics. To this end, while PKC inhibition promotes ectopic expression of SOX2 in mouse embryos, in humans, it does not have a discernible effect on SOX2 expression (Gerri et al. 2020). Taken together, these studies suggest that there are differences in the programs of the core pluripotency factors in humans when compared to the mouse. Understanding the molecular state of the early human blastocyst may provide further insight into the state of human naïve pluripotent cells and vice versa. Further functional studies are required to better understand when and how these pluripotency factors act in the human, and what signaling pathways are involved in driving and maintaining lineage specification in the human.
The second cell fate decision (EPI/PE specification)
The second cell fate decision of mammalian preimplantation development is conventionally considered the differentiation of the ICM into two distinct cell types: the pluripotent EPI and the PE. The EPI exists as a group of cells situated between the PE and the TE, while the PE appears as a monolayer of cells along the surface of the EPI facing the fluid-filled blastocyst cavity. In the murine model, it is well-established that EPI/PE specification begins after the first lineage decision is completed at the 32-cell stage (early blastocyst, E3.5), and those defined layers are clearly distinct and committed by the 64-cell stage (late blastocyst, E4.5) (Guo et al. 2010, Soszynska et al. 2019). Human blastocyst development lags in timing compared to the mouse, and therefore, the period from early to late blastocyst corresponds to approximately E5.0 to E7.0 in human development (Molè et al. 2020). These discrepancies in embryo development dynamics could be related to differences in the timing of EGA (Zhu et al. 2020). As discussed in detail earlier, there may be temporal discrepancies in EPI/PE emergence between mouse and human preimplantation development which may be indicative of important distinctions on a molecular level. In the context of the human, our understanding of EPI/PE specification is limited. Later we will highlight the studied signaling pathways that may be involved and spatio-temporal expression of EPI/PE factors.
Mechanisms of mouse EPI/PE specification
NANOG (an EPI marker) and GATA6 (a PE marker) are expressed in the mouse ICM (E3.5) in the commonly referenced ‘salt and pepper’ pattern (Chazaud et al. 2006), with mutually exclusive cellular expression occurring by the 64-cell stage (E3.75–E4.0) (Plusa et al. 2008, Guo et al. 2010, Posfai et al. 2017, Soszynska et al. 2019) (Fig. 1). A series of loss of function studies analyzing phenotypes of Grb2 (Cheng et al. 1998, Chazaud et al. 2006), Fgf4 and Fgfr2 (Wilder et al. 1997) mutants and inhibition of the FGF pathway (Frankenberg et al. 2011) have explained the acquisition of this pattern via the FGF/MAPK-dependent model for differentiation of the EPI/PE, which has been reviewed in great detail (Soszynska et al. 2019). Cells of the ICM are initially bipotential (Plusa et al. 2008), but differences in FGF signaling cause them to become either NANOG or GATA6 positive. In one model, this is established through sequential waves of internalization, whereby cells internalized early express higher levels of FGF4 and cells internalized later exhibit higher levels of FGF receptors (Soszynska et al. 2019). Another model however reports a bias in the expression of Fgf4 and Fgfr2 in inner cells, with Fgf4 being the first gene expressed between the waves of internalization and suggests that EPI/PE fate is not solely dependent on the timing of asymmetric divisions but also on asynchronicity in the cell cycle (Krupa et al. 2014, Saiz et al. 2016). Nonetheless, at a molecular level, FGF signaling is believed to regulate the final fates. FGF4 is secreted by NANOG-positive cells that are to become the EPI (Frankenberg et al. 2011), which neighboring PE-fated cells are able to respond to via receptor tyrosine kinase activation and GRB2-RAS-MAP signaling (Soszynska et al. 2019). The balance between NANOG and GATA6 results in the initial specification of the ICM derivatives, subsequently leading to stabilization of the pluripotency network in EPI cells and initiation of the endodermal program in PE cells (Soszynska et al. 2019). FGF4 also controls the proportion of EPI and PE cells within the fully developed ICM by coupling regulation of cell number to fate decisions through changes in local growth factor concentration (Saiz et al. 2020).
FGF4 acts on PE-fated cells by binding to FGF receptor 1 (FGFR1) and to a lesser extent FGFR2 on the cell surface (Soszynska et al. 2019). After this signal is relayed, GATA6 initiates a step-wise pluripotency factor disengagement, initially repressing Nanog and Esrrb, then Sox2, followed by Pou5f1, while simultaneously activating extraembryonic endoderm genes Sox17, Gata4 and Sox7 in a similar step-wise manner (Wamaitha et al. 2015). Importantly, while GATA6 is essential for the activation of PE maturation genes it cannot function alone, as in Nanog mutant (FGF4-deficient) embryos, Gata6 expression is unaffected while Sox17 and Gata4 are severely downregulated (Messerschmidt & Kemler 2010, Frankenberg et al. 2011). SOX17 functions to downregulate ESC-associated gene expression and directly activate genes functioning in endodermal differentiation (Niakan et al. 2010). Its induction in mESCs interferes with the binding of NANOG at shared sites, suggesting that SOX17 may compete with pluripotency maintenance proteins for the same DNA-binding sites (Niakan et al. 2010). GATA4 contains an enhancer sequence that is bound and activated by the PE marker FOXA2 following GATA6 expression (Rojas et al. 2010). The expression of FOXA2 and GATA4 likely reinforces PE identity, potentially preventing reversion to an EPI identity and later serving as pioneer factors for the induction of definitive endoderm gene expression. Murine EPI formation depends on the stabilization of the same pluripotency network as in ICM specification; POU5F1, SOX2, and NANOG (Wicklow et al. 2014, Chazaud & Yamanaka 2016). POU5F1 partners with SOX2 to bind to a ‘canonical’ response element to activate EPI-specific genes (Aksoy et al. 2013). Interestingly, both SOX2 and POU5F1 have also been implicated in the proper segregation of the PE (Frum et al. 2013, Wicklow et al. 2014), highlighting the continued importance of cell–cell communication between the two groups not only prior to commitment but also during EPI/PE maturation. Further, it is possible that TEAD1 activity resulting from the nuclear accumulation of YAP is involved in the acquisition of this pluripotency, as Tead1−/− cells showed significantly weaker expression of SOX2 and NANOG compared to wild-type cells at the late mouse blastocyst stage (Hashimoto & Sasaki 2019).
While the PE is established via FGF signals in the embryo environment, the EPI must have a way to maintain its pluripotency (and thus not differentiate into PE). Initially, it was suggested that shielding of presumptive EPI cells may be achieved by the downregulation of FGF receptor 2 (FGFR2) expression, as it was not expressed in the EPI but is in both the TE and PE (Guo et al. 2010). However, more recent studies suggest that it is FGFR1 and not FGFR2 that is critical for the establishment of PE identity (Kang et al. 2017, Molotkov et al. 2017). Since FGFR1 is expressed in all cells of the ICM (Kang et al. 2017), differences in FGF receptor expression could no longer explain the shielding of the presumptive EPI. A model has instead been suggested whereby differential signaling activity in lineage-biased ICM cells is elicited via divergent intracellular responses from a common signal, FGF4 (Kang et al. 2017). Downstream FGF target Etvs4/5 has been found to cluster close to EPI-related genes using single-cell transcriptomics (Kang et al. 2017, Azami et al. 2019), and in murine embryoid bodies knocking down Etv5 increases EPI specification by the upregulation of Fgf5 and downregulation of Gata6, Gata4, and Pdgfrα (Zhang et al. 2018), suggesting it may be a downstream EPI-specific response. ETV5 levels depend on the presence of NANOG, likely through direct activation as NANOG binds to the Etv5 locus in mESCs (Azami et al. 2019). Therefore, NANOG may activate the expression of FGF4 and ETV5 simultaneously, inducing paracrine PE cell differentiation and protecting EPI cells against autocrine/paracrine FGF4 activity (Azami et al. 2019). In PE cells, differential response to FGF4 could be achieved via an FGF-ERK target/regulator called DUSP4. A close association exists between PE-related genes and DUSP4 (Kang et al. 2017, Azami et al. 2019), which enables the direct inactivation of ERK by dephosphorylation. Co-labeling experiments of pERK and DUSP4 identified DUSP4 accumulation tightly following ERK phosphorylation (Azami et al. 2019), suggesting a negative feedback loop exists that potentially regulates the MAPK-ERK pathway (Kang et al. 2017).
Mechanisms of human EPI/PE specification
Patterns similar to the classic ‘salt and pepper’ formation have been detected in the late blastocyst of human embryos, with some inner cells exhibiting high levels of NANOG and low levels of GATA6, with others containing the expression of both at approximately the same level (Kuijk et al. 2012, Roode et al. 2012, Regin et al. 2022). However, a dramatic difference in the process of human EPI/PE specification compared to the mouse is that it does not seem dependent on FGF signaling (Kuijk et al. 2012, Roode et al. 2012). When human embryos are treated with a MEK inhibitor at the morula stage, the number of NANOG-positive cells does not significantly increase relative to controls (Kuijk et al. 2012, Roode et al. 2012) as observed in murine embryos (Frankenberg et al. 2011, Kang et al. 2013, Krawchuk et al. 2013). In accordance with these findings, the vast majority of human blastocysts assessed lack expression of FGF receptors 1–4 as determined by RT-PCR and immunostaining (Kunath et al. 2014). Further, scRNA-seq studies have identified a number of factors found in the mouse pluripotency network that are absent from the human ‘ICM’, including KLF2, Nr0B1, and FBXO15 (Blakeley et al. 2015, Stirparo et al. 2018). Similarly, genes highly enriched in the human EPI (LEFTY1, NODAL, ACVRL1/ALK1, and KLF17 (mRNA and protein)) are not expressed in the mouse ICM (Blakeley et al. 2015, Kilens et al. 2018). Many of these genes identified remain to be validated by immunostaining in both species. Together, these findings suggest a different mechanism controlling EPI/PE specification in human preimplantation development.
scRNA-seq studies have been able to shed some light on alternative cell signaling pathways, finding that several agonists of TGFβ/Activin/Nodal and WNT signaling are upregulated in EPI (Blakeley et al. 2015, Petropoulos et al. 2016, Boroviak et al. 2018, Stirparo et al. 2018). The implication of these pathways is plausible, as both are known for their roles in maintaining pluripotency in hESCs and their modulation can commit cells to one of the three germ-layer cell types (Murry & Keller 2008). To date, studies directly implicating TGFβ/Activin/Nodal and WNT signaling factors functionally in human embryos remain sparse. While one study found that inhibiting the TGFβ/Activin/Nodal signaling pathway from E3 to E5 increases the mean number of NANOG-positive EPI cells in the human embryo (Van der Jeught et al. 2014), a subsequent study found that this same inhibitor downregulated NANOG and POU5F1 expression, which was also confirmed in hESCs (Blakeley et al. 2015). The discrepancy in results may be explained by the four-fold lower concentration of SB-431542, a TGFβ receptor inhibitor, used in the earlier experiment. In terms of WNT signaling, it has been shown that WNT3 is upregulated in human and marmoset EPIs (Boroviak et al. 2018) and that WNT inhibition interferes with NANOG and GATA6 segregation in the marmoset embryo (Boroviak et al. 2015), possibly supporting a functional requirement for WNT in primate PE specification. However, when human embryos were treated with Cardamonin (WNT/β-catenin inhibitor) or WNT3, there was no effect on the mRNA levels of pluripotency markers NANOG, POU5F1, and SOX2 or any effect on the number of NANOG-positive cells compared to controls (Krivega et al. 2015). Further studies are required to more firmly implicate the TGFβ/Activin/Nodal and WNT pathways in EPI-PE specification, and moreover, the complete mechanism by which EPI-PE specification is initiated in the human remains to be elucidated.
Figure 1 outlines the temporal dynamics of different genes involved in EPI/PE specification in mice and shows the expression of those same genes during human preimplantation development. Studies have identified differences between the expression of key mouse lineage markers in humans, making human EPI/PE lineage specification difficult to study and compare (Niakan & Eggan 2013, Blakeley et al. 2015, Boroviak et al. 2018). For example, while in mice, GATA6 begins to restrict to inner cells as early as the 32-cell stage and is restricted to PE-fated cells by the 64-cell stage (Chazaud & Yamanaka 2016), in humans, GATA6 has been found to be constitutively expressed until the mid-blastocyst stage (E6) (Kimber et al. 2008, Kuijk et al. 2012, Roode et al. 2012, Boroviak et al. 2018). As such, GATA6 and NANOG do not become expressed in a mutually exclusive manner until after blastocyst hatching (Kuijk et al. 2012, Roode et al. 2012). In contrast, ‘ICM’ cells collected from early human blastocysts have been shown to express either NANOG or GATA4 exclusively, and GATA6 is expressed at very low levels compared with GATA4 in both early and late human blastocysts (Durruthy-Durruthy et al. 2016), suggesting a more important role for GATA4 (Meistermann et al. 2021). These findings have been used in support of the argument that in the human embryo specification of the EPI occurs first, followed by differentiation of the PE from the EPI (Meistermann et al. 2021). Validating the functional relevance of these genes in the human EPI/PE and further establishing novel human-specific EPI/PE markers is of great importance for our understanding of human lineage specification. This will be critical for functional studies seeking to identify the molecular pathways involved in establishing the lineages, where having accurate markers as read-outs of the manipulation applied is key.
Loss of lineage plasticity
The major difference between EPI/PE specification and commitment is the capability of the cellular program to be reversed. In mice, evidence leads to the conclusion that the ICM cell lineages (EPI and PE) are determined and cannot be reversed as of E4.0 coinciding with the exclusion of the pluripotency marker POU5F1 from the PE (Grabarek et al. 2012, Posfai et al. 2017). At this time, PE cells turn on mature markers such as SOX17, GATA4, DAB2, and PDGFRα, and any experimental condition or manipulation that decreases the cell count of one lineage is not compensated for by an increase in the other (Chazaud & Yamanaka 2016). Further supporting the notion of plasticity in the early embryo, when donor cells from 8-cell murine embryos were aggregated to a wild-type host morula, they were able to contribute to both the ICM and TE lineages in the resulting chimera (Posfai et al. 2017). However, this ability of a single cell to give rise to both lineages begins to decrease beginning at the 16- to 32-cell transition, whereby CDX2 low cells at the early 32-cell stage exclusively contribute to the ICM (Posfai et al. 2017). The specific mechanisms by which plasticity is lost remain largely unknown.
In human embryos, the emergence of the TE, EPI, and PE (Petropoulos et al. 2016, Meistermann et al. 2021) may reflect a longer window of plasticity compared to the mouse. Indeed, isolated E5.0 TE cells which are repositioned in the center of the human embryo lack the ability to sort back to their original position and instead integrate within the ‘ICM’ and start to express NANOG (De Paepe et al. 2013). Recently, the capacity of EPI cells from late human blastocysts (E6 and E7; PE is already specified) to differentiate into other cellular lineages was determined. ICMs were isolated via immunosurgery and further cultured for 5 days, at which point cells with TE morphology and marker expression (GATA3, CK7, and hCGB+) outgrew from the still undifferentiated cell masses (NANOG+) (Guo et al. 2021). When explants were cultured in media supplemented with PD0325901 (MEK inhibitor) and A83-01 (Activin/NODAL/TGF-β pathway inhibitor), they differentiated almost entirely into TE-like cells with GATA3 expressed throughout by culture day 5 (Guo et al. 2021). This suggests that pluripotent cells in the human embryo are capable of retaining extraembryonic lineage plasticity until implantation. Although these seminal studies have provided insight into the plasticity of the human preimplantation embryo, the precise mechanisms and timing of loss of the plasticity and commitment of cells within the lineages remain to be determined.
Stem cell-based 3D models of the human blastocyst
The first stem cell-based models resembling the human preimplantation blastocyst (’blastoids’) were recently developed (Fan et al. 2021, Liu et al. 2021, Sozen et al. 2021, Yanagida et al. 2021, Yu et al. 2021, Kagawa et al. 2022). These groundbreaking models have provided the scientific community with a platform to better understand previously inaccessible aspects of early human development by circumventing some of the ethical constraints and limitations associated with human blastocyst research. Further, the predictive value of blastoids in studies involving drug screening and therapeutic interventions was recently demonstrated by exposing the Kawaga et al. model to exogenous glucocorticoid (GC) and comparing it to the human blastocyst (Kagawa et al. 2022, Zhao et al. 2022). The changes observed in protein expression of genes following GC treatment in the human blastocyst were similarly impacted in blastoids, suggesting that the effects of interventions on gene expression or exposures are modeled well (Zhao et al. 2022). Amongst the published studies, the production of human blastoids has mainly involved culturing human naïve or reprogrammed pluripotent stem cells with tailored culture media in systems that permit 3D cellular interactions and differentiation. Although these studies (Table 1) have made a promising stride toward the goal of a model for human preimplantation development, they are not without limitations. One of these is the low efficiency of blastoid formation in some models after a lengthy derivation process, which ranges from 5.8 to 18% (Liu et al. 2021, Yu et al. 2021). However, in more recent models, this has largely been overcome providing an efficiency up to greater than 80% depending on the stem cell source have been reported (Yanagida et al. 2021, Kagawa et al. 2022). This improvement in efficiency and higher transcriptional similarity was achieved by culturing PSCs in the presence of ERK, ROCK, and NODAL inhibitors (Yanagida et al. 2021), or in the presence of Hippo, TGF-β and ERK inhibitors (with Hippo inhibition noted as being essential for efficiency) (Kagawa et al. 2022). Validation of these models included functional readouts (hCG detection), analysis of lineage markers via immunofluorescence and scRNA-seq, allowing for comparison of transcriptomic signatures to those of previously published human blastocyst datasets. However, the incorporation of comprehensive human-derived reference datasets, that include both pre- and post-implanted lineages were lacking. The Incorporation of proper reference data is critical for the accurate classification of cell types and adequate benchmarking of these models using multiple levels of assessment is critical (Zhao et al. 2021). Further, some of these studies have noted differences in EPI (Fan et al. 2021), PE (Sozen et al. 2021, Yanagida et al. 2021), or TE (Yanagida et al. 2021, Yu et al. 2021) lineage proportions relative to that of human embryos and/or have generated cell populations not representative of preimplantation lineages (Liu et al. 2021, Yu et al. 2021). In addition, the status of XCI and the epigenome remains to be characterized in these blastoids, which are additional molecular layers of importance to consider. As such, further optimization of culturing conditions and benchmarking would achieve a 3D model even more representative of the human blastoid.
Human 3D stem cell-based blastocyst models.
| Publication | Stem cell source | Acquisition of naïve pluripotency | Media | 3D culture system | Efficiency |
|---|---|---|---|---|---|
| Liu et al. 2021 | iPSC from primary human adult dermal fibroblasts (HDFa) | Reprogrammed into naïve stem cells with POU5F1, KLF4, SOX2, and c-MYC. | Human iBlastoid medium (2:1:1 mixture of IVC1 medium, iBlastoid basal medium 1 (50:50 mixture of DMEM/F-12) and neurobasal medium + supplements) | AggreWell | 5.8–18% |
| Yu et al. 2021 | hESCs (WIBR3) | Reprogrammed into naïve stem cells via kinase inhibitors that induce and maintain the activity of a naïve pluripotency reporter. | Sequential hypoblast differentiation and trophoblast stem cell medium treatments | AggreWell | ~10% |
| Yanagida et al. 2021 | Naïve: HNES1, HNES2, HNES3 Primed: Shef6 iPSC from adult primary dermal fibroblasts (HDF75) | Primed cells were reprogrammed into naïve stem cells with PXGL as previously described. | PXGL → PD+A83+Y → A83 → N2B27 | Non-adherent 96-well U-bottom plates | 30 to >80% depending on the source of stem cells |
| Sozen et al. 2021 | hPSC lines; RUES2-GLR and ESI07 | Human expanded potential stem cells were grown using ‘human expanded potential’ (hEP) medium | 50% IVF media (CSCM-NXC), 25% hEP media, 25% 1q trophoblast stem cell media, supplemented with CHIR99021, Y27632, BMP4, FGF2, and A83-01. A83-01 omitted after 48h | AggreWell | 7.2% |
| Fan et al. 2021 | iPSC from human skin fibroblasts (isolated from chest of a female aborted fetus) | iPSCs generated by electroporation of fibroblasts with episomal vectors (pCXLE-hOCT3/4-shp53-F, pCXLE-hSK, and pCXLE-hUL). iPSCs converted to expanded potential stem cells. | BMP4 differentiation medium → EPS medium and IVC1 medium in a ratio of 1.5:1 | AggreWell | 1.9% |
| Kagawa et al. 2022 | Human ES cell lines: H9, iPSC lines: cR-NCRM2 and niPSC 16.2. b | The H9 lines reset to naive state were provided by the lab of Y. Takashima. Other naive human ES cells and iPSCs were provided by the lab of A. Smith. | N2B27 + Y-27632 → PALLY medium (N2B27 + PD0325901, A83-01, LPA, hLIF and Y-27632) → N2B27 + LPA + Y-27632 | Non-adherent hydrogel microwells | >70% |
The generation of new 3D models has sparked important discussions around ethical oversight and as such the International Society for Stem Cell Research (ISSCR) has revisited its guidelines. Importantly, these models are not human blastocysts but are instead considered integrated models with the ability to develop within the ‘14 day’ rule. However, to date, development past this point has not been demonstrated in either the mouse or human. There will likely be a continued dialogue around legal and ethical legislation surrounding the use of these models as further advancements are made.
Conclusion
Despite over 44 years following the first birth from ART, mechanisms underlying the key developmental events during the first 7 days of human development are still poorly understood. The limited availability of human embryos, ethical concerns, and legislation still present an obstacle toward our advancement of knowledge. Live birth is largely governed by proper blastocyst formation, embryo competence, implantation, and continued healthy development during pregnancy. An in-depth understanding of how the human preimplantation embryo develops will support the optimization of culturing conditions and the use of additives, enabling an increased success rate in ART. While great efforts have been made to understand these mechanisms in other species, including the mouse, bovine, rat, guinea pig and non-human primate, with increased access to human embryos and the advancement of technology, it is increasingly apparent that species-related differences exist and the knowledge acquired must be translated with care. This highlights the need for more human-focused studies and the importance of cross-species analysis. A better understanding of key similarities and differences amongst the preimplantation embryos of various species and stem cell-based models is necessary to effectively translate the knowledge acquired and aid in choosing the most appropriate model. Overall, this review has highlighted some of the gaps in our knowledge related to blastocyst formation and lineage differentiation pertaining to human preimplantation development. Studies aimed at filling these knowledge gaps will collectively result in new insights and potential recommendations that can be applied to human health, derivation of stem cells, and fertility.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of this review.
Funding
This work was supported by grants from the Swedish Research Council (2016-01919), Swedish Society for Medical Research (Dnr4-236-2107), The Canadian Institutes of Health Research (PJT-178082). Fonds de Recherche Santé Québec (K V), Natural Sciences and Engineering Research Council (K V, S P (DGECR-2019-00347)). S P holds the Canada Research Chair in Functional Genomics of Reproduction and Development (950-233204).
Author contribution statement
S P conceived the topic of the review. S B, J C, K V, C Z, and S P contributed to the writing and editing of the review.
References
Aksoy I, Jauch R, Eras V, Chng WB, Chen J, Divakar U, Ng CKL, Kolatkar PR & Stanton LW 2013 Sox transcription factors require selective interactions with Oct4 and specific transactivation functions to mediate reprogramming. Stem Cells 31 2632–2646. (https://doi.org/10.1002/stem.1522)
Alpha Scientists in Reproductive Medicine and ESHRE Special Interest Group of Embryology 2011 The Istanbul consensus workshop on embryo assessment: proceedings of an expert meeting. Human Reproduction 26 1270–1283. (https://doi.org/10.1093/humrep/der037)
Azami T, Bassalert C, Allègre N, Estrella LV, Pouchin P, Ema M & Chazaud C 2019 Regulation of the ERK signalling pathway in the developing mouse blastocyst. Development 146 dev177139. (https://doi.org/10.1242/dev.177139)
Blakeley P, Fogarty NME, Valle del I, Wamaitha SE, Hu TX, Elder K, Snell P, Christie L, Robson P & & Niakan KK 2015 Defining the three cell lineages of the human blastocyst by single-cell RNA-seq. Development 142 3151–3165. (https://doi.org/10.1242/dev.123547
Boroviak T, Loos R, Lombard P, Okahara J, Behr R, Sasaki E, Nichols J, Smith A & Bertone P 2015 Lineage-specific profiling delineates the emergence and progression of naive pluripotency in mammalian embryogenesis. Developmental Cell 35 366–382. (https://doi.org/10.1016/j.devcel.2015.10.011)
Boroviak T, Stirparo GG, Dietmann S, Hernando-Herraez I, Mohammed H, Reik W, Smith A, Sasaki E, Nichols J & Bertone P 2018 Single cell transcriptome analysis of human, marmoset and mouse embryos reveals common and divergent features of preimplantation development. Development 145 dev167833. (https://doi.org/10.1242/dev.167833)
Chazaud C & Yamanaka Y 2016 Lineage specification in the mouse preimplantation embryo. Development 143 1063–1074. (https://doi.org/10.1242/dev.128314)
Chazaud C, Yamanaka Y, Pawson T & Rossant J 2006 Early lineage segregation between epiblast and primitive endoderm in mouse blastocysts through the Grb2-MAPK pathway. Developmental Cell 10 615–624. (https://doi.org/10.1016/j.devcel.2006.02.020)
Cheng AM, Saxton TM, Sakai R, Kulkarni S, Mbamalu G, Vogel W, Tortorice CG, Cardiff RD, Cross JC & Muller WJ et al.1998 Mammalian Grb2 regulates multiple steps in embryonic development and malignant transformation. Cell 95 793–803. (https://doi.org/10.1016/s0092-8674(0081702-x)
Cockburn K & Rossant J 2010 Making the blastocyst: lessons from the mouse. Journal of Clinical Investigation 120 995–1003. (https://doi.org/10.1172/JCI41229)
De Paepe C, Cauffman G, Verloes A, Sterckx J, Devroey P, Tournaye H, Liebaers I & Van de Velde H 2013 Human trophectoderm cells are not yet committed. Human Reproduction 28 740–749. (https://doi.org/10.1093/humrep/des432)
Durruthy-Durruthy J, Wossidlo M, Pai S, Takahashi Y, Kang G, Omberg L, Chen B, Nakauchi H, Reijo Pera R & Sebastiano V 2016 Spatiotemporal reconstruction of the human blastocyst by single-cell gene-expression analysis informs induction of naive pluripotency. Developmental Cell 38 100–115. (https://doi.org/10.1016/j.devcel.2016.06.014)
Fan Y, Min Z, Alsolami S, Ma Z, Zhang E, Chen W, Zhong K, Pei W, Kang X & Zhang P et al.2021 Generation of human blastocyst-like structures from pluripotent stem cells. Cell Discovery 7 81. (https://doi.org/10.1038/s41421-021-00316-8)
Fogarty NME, McCarthy A, Snijders KE, Powell BE, Kubikova N, Blakeley P, Lea R, Elder K, Wamaitha SE & Kim D et al.2017 Genome editing reveals a role for OCT4 in human embryogenesis. Nature 550 67–73. (https://doi.org/10.1038/nature24033)
Foty RA & Steinberg MS 2005 The differential adhesion hypothesis: a direct evaluation. Developmental Biology 278 255–263. (https://doi.org/10.1016/j.ydbio.2004.11.012)
Frankenberg S, Gerbe F, Bessonnard S, Belville C, Pouchin P, Bardot O & Chazaud C 2011 Primitive endoderm differentiates via a three-step mechanism involving Nanog and RTK signaling. Developmental Cell 21 1005–1013. (https://doi.org/10.1016/j.devcel.2011.10.019)
Frum T, Halbisen MA, Wang C, Amiri H, Robson P & Ralston A 2013 Oct4 cell-autonomously promotes primitive endoderm development in the mouse blastocyst. Developmental Cell 25 610–622. (https://doi.org/10.1016/j.devcel.2013.05.004)
Frum T, Murphy TM & Ralston A 2018 HIPPO signaling resolves embryonic cell fate conflicts during establishment of pluripotency in vivo. eLife 7 e42298. (https://doi.org/10.7554/eLife.42298)
Gerri C, McCarthy A, Alanis-Lobato G, Demtschenko A, Bruneau A, Loubersac S, Fogarty NME, Hampshire D, Elder K & Snell P et al.2020 Initiation of a conserved trophectoderm program in human, cow and mouse embryos. Nature 587 443–447. (https://doi.org/10.1038/s41586-020-2759-x)
Grabarek JB, Żyżyńska K, Saiz N, Piliszek A, Frankenberg S, Nichols J, Hadjantonakis AK & Plusa B 2012 Differential plasticity of epiblast and primitive endoderm precursors within the ICM of the early mouse embryo. Development 139 129–139. (https://doi.org/10.1242/dev.067702)
Guo G, Huss M, Tong GQ, Wang C, Li Sun L, Clarke ND & Robson P 2010 Resolution of cell fate decisions revealed by single-cell gene expression analysis from zygote to blastocyst. Developmental Cell 18 675–685. (https://doi.org/10.1016/j.devcel.2010.02.012)
Guo G, Stirparo GG, Strawbridge SE, Spindlow D, Yang J, Clarke J, Dattani A, Yanagida A, Li MA & Myers S et al.2021 Human naive epiblast cells possess unrestricted lineage potential. Cell Stem Cell 28 1040–1056.e6. (https://doi.org/10.1016/j.stem.2021.02.025)
Hashimoto M & Sasaki H 2019 Epiblast formation by TEAD-YAP-dependent expression of pluripotency factors and competitive elimination of unspecified cells. Developmental Cell 50 139 .e5–154.e5. (https://doi.org/10.1016/j.devcel.2019.05.024)
Hirate Y & Sasaki H 2014 The role of angiomotin phosphorylation in the Hippo pathway during preimplantation mouse development. Tissue Barriers 2 e28127. (https://doi.org/10.4161/tisb.28127)
Kagawa H, Javali A, Khoei HH, Sommer TM, Sestini G, Novatchkova M, Scholte op Reimer Y, Castel G, Bruneau A & Maenhoudt N et al.2022 Human blastoids model blastocyst development and implantation. Nature 601 600–605. (https://doi.org/10.1038/s41586-021-04267-8)
Kang M, Piliszek A, Artus J & Hadjantonakis AK 2013 FGF4 is required for lineage restriction and salt-and-pepper distribution of primitive endoderm factors but not their initial expression in the mouse. Development 140 267–279. (https://doi.org/10.1242/dev.084996)
Kang M, Garg V & Hadjantonakis AK 2017 Lineage establishment and progression within the inner cell mass of the mouse blastocyst requires FGFR1 and FGFR2. Developmental Cell 41 496.e5–510.e5. (https://doi.org/10.1016/j.devcel.2017.05.003)
Kilens S, Meistermann D, Moreno D, Chariau C, Gaignerie A, Reignier A, Lelièvre Y, Casanova M, Vallot C & Nedellec S et al.2018 Parallel derivation of isogenic human primed and naive induced pluripotent stem cells. Nature Communications 9 360. (https://doi.org/10.1038/s41467-017-02107-w)
Kimber SJ, Sneddon SF, Bloor DJ, El-Bareg AM, Hawkhead JA, Metcalfe AD, Houghton FD, Leese HJ, Rutherford A & Lieberman BA et al.2008 Expression of genes involved in early cell fate decisions in human embryos and their regulation by growth factors. Reproduction 135 635–647. (https://doi.org/10.1530/REP-07-0359)
Krawchuk D, Honma-Yamanaka N, Anani S & Yamanaka Y 2013 FGF4 is a limiting factor controlling the proportions of primitive endoderm and epiblast in the ICM of the mouse blastocyst. Developmental Biology 384 65–71. (https://doi.org/10.1016/j.ydbio.2013.09.023)
Krivega M, Essahib W & Van de Velde H 2015 WNT3 and membrane-associated β-catenin regulate trophectoderm lineage differentiation in human blastocysts. Molecular Human Reproduction 21 711–722. (https://doi.org/10.1093/molehr/gav036)
Krupa M, Mazur E, Szczepańska K, Filimonow K, Maleszewski M & Suwińska A 2014 Allocation of inner cells to epiblast vs primitive endoderm in the mouse embryo is biased but not determined by the round of asymmetric divisions (8→16- and 16→32-cells) Developmental Biology 385 136–148. (https://doi.org/10.1016/j.ydbio.2013.09.008)
Kuijk EW, van Tol LTA, Van de Velde H, Wubbolts R, Welling M, Geijsen N & Roelen BAJ 2012 The roles of FGF and MAP kinase signaling in the segregation of the epiblast and hypoblast cell lineages in bovine and human embryos. Development 139 871–882. (https://doi.org/10.1242/dev.071688)
Kunath T, Yamanaka Y, Detmar J, MacPhee D, Caniggia I, Rossant J & Jurisicova A 2014 Developmental differences in the expression of FGF receptors between human and mouse embryos. Placenta 35 1079–1088. (https://doi.org/10.1016/j.placenta.2014.09.008)
Leung CY & Zernicka-Goetz M 2013 Angiomotin prevents pluripotent lineage differentiation in mouse embryos via Hippo pathway-dependent and -independent mechanisms. Nature Communications 4 2251. (https://doi.org/10.1038/ncomms3251)
Liu X, Tan JP, Schröder J, Aberkane A, Ouyang JF, Mohenska M, Lim SM, Sun YBY, Chen J & Sun G et al.2021 Modelling human blastocysts by reprogramming fibroblasts into iBlastoids. Nature 591 627–632. (https://doi.org/10.1038/s41586-021-03372-y)
Maître JL, Niwayama R, Turlier H, Nédélec F & Hiiragi T 2015 Pulsatile cell-autonomous contractility drives compaction in the mouse embryo. Nature Cell Biology 17 849–855. (https://doi.org/10.1038/ncb3185)
Marikawa Y & Alarcón VB 2009 Establishment of trophectoderm and inner cell mass lineages in the mouse embryo. Molecular Reproduction and Development 76 1019–1032. (https://doi.org/10.1002/mrd.21057)
Meistermann D, Bruneau A, Loubersac S, Reignier A, Firmin J, François-Campion V, Kilens S, Lelièvre Y, Lammers J & Feyeux M et al.2021 Integrated pseudotime analysis of human pre-implantation embryo single-cell transcriptomes reveals the dynamics of lineage specification. Cell Stem Cell 28 1625 .e6–1640.e6. (https://doi.org/10.1016/j.stem.2021.04.027)
Messerschmidt DM & Kemler R 2010 Nanog is required for primitive endoderm formation through a non-cell autonomous mechanism. Developmental Biology 344 129–137. (https://doi.org/10.1016/j.ydbio.2010.04.020)
Molè MA, Weberling A & Zernicka-Goetz M 2020 Chapter Four – Comparative analysis of human and mouse development: from zygote to pre-gastrulation. In Current Topics in Developmental Biology, pp. 113–138. Ed Solnica-Krezel L Academic Press. (https://doi.org/10.1016/bs.ctdb.2019.10.002)
Molotkov A, Mazot P, Brewer JR, Cinalli RM & Soriano P 2017 Distinct requirements for FGFR1 and FGFR2 in primitive endoderm development and exit from pluripotency. Developmental Cell 41 511 .e4–526.e4. (https://doi.org/10.1016/j.devcel.2017.05.004)
Murry CE & Keller G 2008 Differentiation of embryonic stem cells to clinically relevant populations: lessons from embryonic development. Cell 132 661–680. (https://doi.org/10.1016/j.cell.2008.02.008)
Niakan KK & Eggan K 2013 Analysis of human embryos from zygote to blastocyst reveals distinct gene expression patterns relative to the mouse. Developmental Biology 375 54–64. (https://doi.org/10.1016/j.ydbio.2012.12.008)
Niakan KK, Ji H, Maehr R, Vokes SA, Rodolfa KT, Sherwood RI, Yamaki M, Dimos JT, Chen AE & Melton DA et al.2010 Sox17 promotes differentiation in mouse embryonic stem cells by directly regulating extraembryonic gene expression and indirectly antagonizing self-renewal. Genes and Development 24 312–326. (https://doi.org/10.1101/gad.1833510)
Nishioka N, Yamamoto S, Kiyonari H, Sato H, Sawada A, Ota M, Nakao K & Sasaki H 2008 Tead4 is required for specification of trophectoderm in pre-implantation mouse embryos. Mechanisms of Development 125 270–283. (https://doi.org/10.1016/j.mod.2007.11.002)
Petropoulos S, Edsgärd D, Reinius B, Deng Q, Panula SP, Codeluppi S, Plaza Reyes A, Linnarsson S, Sandberg R & Lanner F 2016 Single-cell RNA-seq reveals lineage and X chromosome dynamics in human preimplantation embryos. Cell 165 1012–1026. (https://doi.org/10.1016/j.cell.2016.03.023)
Plusa B, Piliszek A, Frankenberg S, Artus J & Hadjantonakis AK 2008 Distinct sequential cell behaviours direct primitive endoderm formation in the mouse blastocyst. Development 135 3081–3091. (https://doi.org/10.1242/dev.021519)
Posfai E, Petropoulos S, de Barros FRO, Schell JP, Jurisica I, Sandberg R, Lanner F & Rossant J 2017 Position- and Hippo signaling-dependent plasticity during lineage segregation in the early mouse embryo. eLife 6 e22906. (https://doi.org/10.7554/eLife.22906)
Radley A, Corujo-Simon E, Nichols J, Smith A & Dunn SJ 2022 Entropy sorting of single-cell RNA sequencing data reveals the inner cell mass in the human pre-implantation embryo. Stem Cell Reports 18 47–63. (https://doi.org/10.1016/j.stemcr.2022.09.007)
Ralston A, Cox BJ, Nishioka N, Sasaki H, Chea E, Rugg-Gunn P, Guo G, Robson P, Draper JS & Rossant J 2010 Gata3 regulates trophoblast development downstream of Tead4 and in parallel to Cdx2. Development 137 395–403. (https://doi.org/10.1242/dev.038828)
Regin M, Essahib W, Demtschenko A, Dewandre D, David L, Gerri C, Niakan K, Verheyen G, Tournaye H & Sterckx J et al.2022 Lineage segregation in human pre-implantation embryos is specified by YAP1 and TEAD1. bioRxiv (https://doi.org/10.1101/2022.09.29.509946)
Rojas A, Schachterle W, Xu SM, Martín F & Black BL 2010 Direct transcriptional regulation of Gata4 during early endoderm specification is controlled by FoxA2 binding to an intronic enhancer. Developmental Biology 346 346–355. (https://doi.org/10.1016/j.ydbio.2010.07.032)
Roode M, Blair K, Snell P, Elder K, Marchant S, Smith A & Nichols J 2012 Human hypoblast formation is not dependent on FGF signalling. Developmental Biology 361 358–363. (https://doi.org/10.1016/j.ydbio.2011.10.030)
Saiz N, Williams KM, Seshan VE & Hadjantonakis AK 2016 Asynchronous fate decisions by single cells collectively ensure consistent lineage composition in the mouse blastocyst. Nature Communications 7 13463. (https://doi.org/10.1038/ncomms13463)
Saiz N, Mora-Bitria L, Rahman S, George H, Herder JP, Garcia-Ojalvo J & Hadjantonakis AK 2020 Growth-factor-mediated coupling between lineage size and cell fate choice underlies robustness of mammalian development. eLife 9 e56079. (https://doi.org/10.7554/eLife.56079)
Soszynska A, Klimczewska K & Suwinska A 2019 FGF/ERK signaling pathway: how it operates in mammalian preimplantation embryos and embryo-derived stem cells. International Journal of Developmental Biology 63 171–186. (https://doi.org/10.1387/ijdb.180408as)
Sozen B, Jorgensen V, Weatherbee BAT, Chen S, Zhu M & Zernicka-Goetz M 2021 Reconstructing aspects of human embryogenesis with pluripotent stem cells. Nature Communications 12 5550. (https://doi.org/10.1038/s41467-021-25853-4)
Stamatiadis P, Cosemans G, Boel A, Menten B, De Sutter P, Stoop D, Chuva de Sousa Lopes SM, Lluis F, Coucke P & Heindryckx B 2022 TEAD4 regulates trophectoderm differentiation upstream of CDX2 in a GATA3-independent manner in the human preimplantation embryo. Human Reproduction 37 1760–1773. (https://doi.org/10.1093/humrep/deac138)
Stirparo GG, Boroviak T, Guo G, Nichols J, Smith A & Bertone P 2018 Integrated analysis of single-cell embryo data yields a unified transcriptome signature for the human pre-implantation epiblast. Development 145 dev158501. (https://doi.org/10.1242/dev.158501)
Van der Jeught M, Heindryckx B, O’Leary T, Duggal G, Ghimire S, Lierman S, Van Roy N, Chuva de Sousa Lopes SM, Deroo T & Deforce D et al.2014 Treatment of human embryos with the TGFβ inhibitor SB431542 increases epiblast proliferation and permits successful human embryonic stem cell derivation. Human Reproduction 29 41–48. (https://doi.org/10.1093/humrep/det400)
Wamaitha SE, Valle del I, Cho LTY, Wei Y, Fogarty NME, Blakeley P, Sherwood RI, Ji H & Niakan KK 2015 Gata6 potently initiates reprograming of pluripotent and differentiated cells to extraembryonic endoderm stem cells. Genes and Development 29 1239–1255. (https://doi.org/10.1101/gad.257071.114)
Wicklow E, Blij S, Frum T, Hirate Y, Lang RA, Sasaki H & Ralston A 2014 HIPPO pathway members restrict SOX2 to the inner cell mass where it promotes ICM fates in the mouse blastocyst. PLoS Genetics 10 e1004618. (https://doi.org/10.1371/journal.pgen.1004618)
Wilder PJ, Kelly D, Brigman K, Peterson CL, Nowling T, Gao QS, McComb RD, Capecchi MR & Rizzino A 1997 Inactivation of the FGF-4 gene in embryonic stem cells alters the growth and/or the survival of their early differentiated progeny. Developmental Biology 192 614–629. (https://doi.org/10.1006/dbio.1997.8777)
Yan L, Yang M, Guo H, Yang L, Wu J, Li R, Liu P, Lian Y, Zheng X & Yan J et al.2013 Single-cell RNA-Seq profiling of human preimplantation embryos and embryonic stem cells. Nature Structural and Molecular Biology 20 1131–1139. (https://doi.org/10.1038/nsmb.2660)
Yanagida A, Spindlow D, Nichols J, Dattani A, Smith A & Guo G 2021 Naive stem cell blastocyst model captures human embryo lineage segregation. Cell Stem Cell 28 1016 .e4–1022.e4. (https://doi.org/10.1016/j.stem.2021.04.031)
Yu L, Wei Y, Duan J, Schmitz DA, Sakurai M, Wang L, Wang K, Zhao S, Hon GC & Wu J 2021 Blastocyst-like structures generated from human pluripotent stem cells. Nature 591 620–626. (https://doi.org/10.1038/s41586-021-03356-y)
Zhang J, Cao H, Xie J, Fan C, Xie Y, He X, Liao M, Zhang S & & Wang H 2018 The oncogene Etv5 promotes MET in somatic reprogramming and orchestrates epiblast/primitive endoderm specification during mESCs differentiation. Cell Death and Disease 9 1. (https://doi.org/10.1038/s41419-018-0335-1
Zhao C, Reyes AP, Schell JP, Weltner J, Ortega NM, Zheng Y, Björklund ÅK, Rossant J, Fu J & Petropoulos S et al.2021 Reprogrammed blastoids contain amnion-like cells but not trophectoderm. bioRxiv. (https://doi.org/10.1101/2021.05.07.442980
Zhao C, Biondic S, Vandal K, Björklund ÅK, Hagemann-Jensen M, Sommer TM, Canizo J, Clark S, Raymond P & Zenklusen DR et al.2022 Single-cell multi-omics of human preimplantation embryos shows susceptibility to glucocorticoids. Genome Research 32 1627–1641. (https://doi.org/10.1101/gr.276665.122)
Zhu M, Cornwall-Scoones J, Wang P, Handford CE, Na J, Thomson M & Zernicka-Goetz M 2020 Developmental clock and mechanism of de novo polarization of the mouse embryo. Science 370 eabd2703. (https://doi.org/10.1126/science.abd2703)
Zhu M, Shahbazi M, Martin A, Zhang C, Sozen B, Borsos M, Mandelbaum RS, Paulson RJ, Mole MA & Esbert M et al.2021 Human embryo polarization requires PLC signaling to mediate trophectoderm specification. eLife 10 e65068. (https://doi.org/10.7554/eLife.65068)
