Comparative aspects of early lineage specification events in mammalian embryos – insights from reverse genetics studies

ABSTRACT

Within the mammalian class, formation of the blastocyst is morphologically highly conserved among different species. The molecular and cellular events during preimplantation embryo development have been studied extensively in the mouse as model organism, because multiple genetically defined strains and a plethora of reverse genetics tools are available to dissect specific gene functions and regulatory networks. However, major differences in preimplantation developmental kinetics, implantation, and placentation exist among mammalians, and recent studies in species other than mouse showed, that even regulatory mechanisms of the first lineage differentiation events and maintenance of pluripotency are not always conserved. Here, we focus on the first and the second lineage segregation in mouse and bovine embryos, when the first differentiated cell types emerge. We outline their common features and differences in the regulation of these essential events during embryonic development with a glance at further species. In addition, we show how new reverse genetics strategies aid the study of regulatory circuits in embryos of domestic species, enhancing our overall understanding of mammalian preimplantation development.

Lessons from the mouse model

After fertilization, the zygote is generally regarded as totipotent and cells from the embryo retain that totipotency until the 2-cell stage with the ability to generate a complete organism and until the 8-cell stage being competent to contribute to all three cell lineages of a chimeric embryo [1]. The three lineages in a late blastocyst emerge from two consecutive lineage differentiations. First the trophectoderm (TE) separates from the inner cell mass (ICM) and second the primitive endoderm (PE) differentiates within the ICM, leaving the epiblast (EPI) as the only pluripotent lineage, from which the embryo proper will develop. In the specification of ICM and TE, the HIPPO/YAP signaling pathway plays a central role. Segregation of the two lineages begins, after the first 1–2 cells are internalized at the 16-cell stage; cells at the outside of the embryo experience less cell-to-cell contact and are polarized, which negatively regulates HIPPO signaling and leads to nuclear translocation of YAP and the expression of TE-specific genes, i.e. Cdx2 and Gata3, through activation of Tead4 during formation of the blastocyst (reviewed in [2]). The formerly ubiquitously expressed pluripotency regulating transcription factor OCT4/POU5F1 is then restricted to the ICM and extinguished in TE cells by CDX2 [3]. After the first lineage differentiation of the TE, cells of the ICM form a homogeneous population co-expressing OCT4, NANOG, GATA6 and SOX2. At the 32-cell stage, ICM cells begin to express either NANOG or GATA6 in a mutually exclusive salt and pepper expression pattern, leading to an EPI or PE cell fate, respectively (reviewed in [1,2,4,5]). This second lineage differentiation is mainly regulated by FGF4/MAPK signaling. In wild-type embryos, expression of PE genes can be induced with exogenous FGF4, resulting in GATA6 expression in all cells of the ICM [6]. In contrast, embryos lacking FGF4 have an ICM entirely made up of NANOG positive cells [7] and the same is also true for embryos treated with inhibitors of FGF/MAPK signaling [8]. New insights into the roles of FGF receptors during PE differentiation and maturation of the EPI cell lineage revealed, that both FGFR1 and FGFR2 are indispensable for correct lineage specification in the ICM. FGFR1 is expressed in all cells of the ICM and plays a dominant role, while FGFR2 – in contrast to previous models – is only expressed in PE cells and not required for initial GATA6 expression [9,10]. After establishing the mutually exclusive expression of PE and EPI specific transcription factors, PE cells begin to migrate to their definitive position by E4.0 and implantation of the late blastocyst begins shortly after at E4.5 [4].

New insights through comparative embryology

It remains largely unclear, how exactly the first lineage differentiation of the TE is initiated and regulated in species other than mouse. In bovine, polarization of outside cells and expression of TEAD4 is first observed at the 16-cell stage [11,12]. YAP1 is expressed in the nuclei of most cells from day 5 onwards but not co-expressed with NANOG in later stages [13]. Knockdown of YAP1 or inhibition of the interaction between TEAD4 and YAP1 using verteporfin both result in a decreased number of CDX2 positive cells, indicating a role in TE differentiation [13]. But while in mouse knockout of Tead4 results in the embryo’s inability to form a blastocyst and is lethal before implantation [14], knockdown of TEAD4 transcripts in bovine had no effect on the development of zygotes until blastocyst stage, while effects on the expression levels of TE specific genes remain controversial [12,15]. Akizawa and coworkers [15] proposed a reciprocal regulation of CTGF (CCN2) and TEAD4 in bovine TE development, as knockdown of TEAD4 led to reduced CTGF transcripts and vice versa and normal expression of both genes was required for the stable expression of the TE specific genes GATA2 and CDX2, but not GATA3. Angiomotin (AMOT) is a junction-associated component of the HIPPO/YAP signaling pathway and therefore present in inside cells of the mouse morula, where it phosphorylates YAP1 and thus inhibits expression of Tead4 [16,17]. AMOT was also associated with differentiation events in bovine embryos, where exposure of embryos to mRNA of the WNT-signaling antagonist Dickkopf-1 (DKK1) resulted in an increased number of TE and hypoblast (HB, equivalent to the murine PE) cells and simultaneously reduced the transcript abundance of AMOT [18,19]. Evidence suggests a role of HIPPO/YAP signaling in the differentiation of the TE also in bovine embryos, but functional studies of the key components are needed to dissect the mechanisms involved in the first lineage differentiation.

The second lineage differentiation in bovine embryos starts after formation of the early blastocyst 6–7 days after fertilization [20], while implantation begins at day 18 (Figure 1 [21]). At day 8, the same salt and pepper distribution of GATA6 and NANOG positive cells is present in the bovine ICM when compared with the corresponding stage of mouse embryogenesis. Nevertheless, the role of FGF4 is not analog to the situation in the mouse: inhibition of MAPK signaling only partially blocks GATA6 expression and increases the number of NANOG positive cells; inhibition of the FGF receptor has no effect on the composition of the ICM in the bovine embryo, while addition of FGF4 leads to GATA6 expression in all ICM cells. Therefore, FGF signaling is not essential for expression of GATA6 and the effect of exogenous FGF4 on GATA6 expression is indirect in bovine embryos [13,22]. Interestingly, in rabbit embryos exogenous FGF4 also induces PE fate in all cells of the ICM and inhibition of FGF/MAPK signaling prevents cells from acquiring a PE fate. But in contrast to mouse and bovine embryos, this inhibition of the PE cell fate does not increase the number of EPI precursor cells, resulting in an increased number of cells which lack expression of any marker of either EPI or PE [23]. Interference with FGF/MAPK signaling in human preimplantation embryos has no effect on the expression of the HB and EPI markers GATA4/6 and NANOG, respectively [22,24], while the effect of exogenous FGF4 on the second lineage differentiation in human remains unknown.

Figure 1.

Developmental kinetics of preimplantation development and timing of embryonic genome activation (EGA) in mouse [51–53], rabbit [54], pig [55], cattle [21,44,56] and human embryos (reviewed in [57]). Time points of 2^nd and 3^rd cleavage divisions indicate the first cleavage after reaching 2-cell and 4-cell stage, respectively.

The role of OCT4 during the first differentiation events

In mouse preimplantation development, it is now established that Oct4, together with Nanog and Sox2, is at the top of the pluripotency regulatory hierarchy [25], although it is neither necessary for the first lineage segregation into ICM and TE nor for the initiation of toti- or pluripotency [26–28]. After formation of the TE, OCT4 expression is restricted to cells in the ICM and extinguished in TE cells by CDX2 [3]. A variety of different species though co-express OCT4 and CDX2 in the TE shortly after the first lineage segregation, because CDX2 does not actively extinguish OCT4. This is the case in human, rabbit, pig, bovine and even rat blastocysts [29–33], indicating that suppression of Oct4 expression in the TE by CDX2 is a mechanism developed uniquely in mouse, probably to enable fast implantation [3].

Embryos with a maternal and zygotic knockout (KO) of Oct4 show normal development until the blastocyst stage including the first lineage differentiation [27,28,34]. At E3.5, Oct4 KO embryos exhibit normal cell numbers in the TE and ICM, and expression of GATA3 and CDX2 – factors that induce differentiation of the TE – is repressed in the ICM [26,28,35]. OCT4-deficient embryos also exhibit a mutually exclusive expression of NANOG and GATA6 in the ICM at E3.75. With ongoing development, GATA6 positive cells disappear from the ICM and the proportion of cells neither expressing GATA6 nor NANOG (unlabeled cells) increases until E4.25, when almost no GATA6 positive cells are found. Activation of PE specific gene expression, i.e. Gata6, Sox17 and Sox7, fails and thus there is no development of PE. Treatment of Oct4 KO embryos with exogenous FGF4 cannot activate PE gene expression or repress NANOG expression. In chimeras from an Oct4 KO embryo and wild-type ESCs, the ESCs are not able to rescue PE development, so OCT4 is required cell-autonomously for PE differentiation. In summary, the current working model states, that in EPI cells OCT4 regulates the expression of FGF4, which induces PE cell fate and that additionally, OCT4 in PE cells activates the expression of PE genes [27,34].

To address the question, whether the roles of OCT4 uncovered in the mouse model are conserved across species or unique to mouse preimplantation development, we recently set out to study the effects of loss of OCT4 in bovine embryogenesis [36]. In a non-homologous end-joining repair approach, we induced a CRISPR/Cas9 mediated loss of function mutation by deletion of a single nucleotide in exon 2 of OCT4 in adult fibroblasts. By reconstructing embryos via somatic cell nuclear transfer (SCNT [37]), we were able to study the development of OCT4 KO embryos in vitro with RNA sequencing and immunofluorescence analyses. Similar to OCT4-deficient murine [27] and human [38] embryos, transcriptome analysis revealed that there is no conversion of bovine OCT4 KO blastocysts toward one particular lineage, but rather an overall decrease in relevant gene expression in all three lineages, which eventually leads to developmental failure. To examine effects of an OCT4 KO during the first lineage differentiation, we analyzed the expression patterns of OCT4 and CDX2 at day 5 morula and day 7 blastocyst stage. By day 5, OCT4 protein was still detectable in OCT4 KO embryos, albeit at a lower proportion as in control embryos, showing that maternal OCT4 mRNA is sufficient to maintain OCT4 expression until day 5. By day 7, OCT4 was completely absent and as observed in mouse OCT4 deficient embryos, the total number of cells was unchanged and CDX2 was repressed in ICM cells [26,28,35]. This indicates that embryonic OCT4 has no effect on the quantitative allocation of cells to either ICM or TE during the first lineage differentiation and that OCT4 is not required to repress CDX2 in the ICM at formation of the blastocyst. Immunostaining of GATA6 and NANOG at day 7 revealed, that in bovine OCT4 KO blastocysts, NANOG is lost and that the embryos fail to establish a mutually exclusive expression pattern of EPI and HB precursor cells, although GATA6 is still present. Zygotic activation of NANOG is not dependent on embryonic OCT4, as NANOG was still present at day 5. Our results rather indicate that maintenance of EPI cells in the absence of OCT4 fails at the beginning of the second lineage differentiation. This is similar to the phenotype of human OCT4 KO blastocysts, where the absence of NANOG protein was observed [38], but in sharp contrast to Oct4-null mouse blastocysts, where NANOG persists while GATA6 expression is lost [27,34]. Interestingly, the different OCT4 KO phenotypes show parallels to the effects of inhibition of FGF4/MAPK signaling. In mouse, inhibition of this pathway completely ablates GATA6 expression and all cells of the ICM express NANOG (reviewed in [8]), while in bovine the proportion of NANOG positive cells increases but GATA6 is still present [22]. In human embryos, loss of MAPK signaling has no effect on the EPI and HB precursor cells [22,24]. In mouse EPI and PE cells, OCT4 regulates the expression of FGF4 and the activation of PE genes, respectively, and is therefore essential for the maintenance of GATA6 expression. On the other hand, bovine and human embryos apparently possess an additional activation pathway for GATA6, as FGF4-independent expression of GATA6 is observed, and one may speculate that this is the reason why in contrast to mouse, GATA6 expression is maintained in human and bovine OCT4 KO embryos (Figure 2). How in mouse the expression of NANOG is maintained in the absence of OCT4, while it is lost in human and bovine, remains unclear.

Figure 2.

Regulation of the second lineage differentiation in mouse and bovine embryos. In mouse, Oct4, Sox2 and Nanog promote expression of FGF4. FGF4 signal transduction by FGFR1 maintains epiblast identity, while FGFR2, which is only expressed in primitive endoderm (PE) precursor cells, induces expression of PE specific genes downstream of Gata6 together with FGF4 [4,9,10]. In bovine, NANOG expression in epiblast cells is OCT4 dependent. NANOG and GATA6 expression is mutually exclusive in a subset of ICM cells by day 7 and in OCT4 deficient embryos that have lost all NANOG expression, there are no GATA6 negative cells. Exogenous FGF4 induces GATA6 expression in all cells of the ICM, but inhibition of MAPK signaling does not repress GATA6 expression in all cells and inhibition of FGFR has no effect. If in the bovine embryo epiblast cells express FGF4 to induce differentiation in hypoblast precursor cells, remains unclear. FGF/MAPK signaling induces GATA6 expression but it is unknown if, like in mouse, this is also OCT4 dependent. OCT4 deficient embryos still express GATA6, but this might also be induced by an unknown factor, which also maintains GATA6 expression when FGF/MAPK signaling is inhibited [22,36].

Domestic species as model organisms for preimplantation development

Our strategy to induce mutations in somatic cells and reconstruct embryos through SCNT provides a general approach for studying the regulation of preimplantation development in domestic species. The great advantage of SCNT is the guarantee, that all embryos produced will uniformly carry the mutations previously identified in somatic cells and that thorough screening of the used cell line, e.g. for off-target effects of CRISPR/Cas9, is feasible. This allows the implementation of proper controls, which should also include wild-type in vitro fertilized embryos (IVF) to exclude effects of the SCNT procedure. Importantly, in our study on the function of OCT4 in bovine preimplantation embryos [36], we did not record any major effects of SCNT on the parameters investigated. Furthermore, RNA sequencing data revealed, that control embryos from SCNT and IVF clustered closely together against OCT4 KO embryos in a principal component analysis. An alternative to the above-described approach for inducing mutations in preimplantation embryos is the injection of target specific nucleases into the zygote. This can induce mutations at a relatively high rate, but the type of mutation is unknown during development of the embryo and it also often results in mosaic mutations [39–42], hampering analysis and repeatability of experiments. In addition, the limited genomic material per specimen impedes in-depth investigations, especially when combined with imaging procedures or transcriptome analyses. Recently, OCT4 has been knocked out by zygote injection and the authors report, that embryos deficient of OCT4 fail before blastocyst stage and lose OCT4 expression already at morula stage [43]. This is in contrast to our findings that OCT4 was maintained until day 5, probably by maternal mRNAs, and that OCT4 KO embryos reach blastocyst stage without major alterations during the first lineage differentiation. Apparently, differences between embryos from SCNT and IVF can influence the embryo’s response to loss of OCT4. In this particular case evidence suggests, that in IVF embryos maternal OCT4 mRNA is degraded faster, leading to a completely OCT4 negative embryo already by day 5, whereas in SCNT embryos, OCT4 was still present and may have enabled further development to the blastocyst stage. Considering the different phenotypes of OCT4-deficient embryos from SCNT and zygote injection, the validation of obtained results from either approach with an alternative strategy seems imperative to ultimately decipher regulatory mechanisms using reverse genetics (Figure 3).

Figure 3.

A general strategy for studying the role of specific genes during preimplantation development in domestic species. Embryos from somatic cell nuclear transfer provide a uniform platform with a defined genotype and enable thorough analysis together with in vitro fertilized control embryos. Validation of observed phenotypes by genome editing in zygotes or mRNA knockdown provides ultimate proof.

The technological developments in single-cell RNA sequencing enable researchers to study the heterogeneity in cell populations, permitting the analysis of the particular lineages of the preimplantation embryo [44–46]. In combination with reverse genetics experiments, single-cell RNA sequencing will accelerate the dissection of the specific genes’ role also within an embryonic cell lineage, allowing a more detailed insight into regulatory mechanisms [38].

Differences in key developmental features across mammalian species stress the importance and value of comparative embryology. With the recent upcoming of highly efficient genome editing tools, we are for the first time able to investigate the dynamics and mechanisms during preimplantation development in alternative model organisms in a resource and time efficient manner. Domestic species such as pig and cattle have the great advantage, that artificial reproductive technologies are highly advanced. This includes the in vitro maturation and fertilization of oocytes as well as efficient cloning protocols and the subsequent culture of embryos until blastocyst stage, but also the possibility to transfer embryos to recipients for the analysis of later developmental stages [47,48]. In cattle, non-surgical, transcervical collection of embryos is even possible at all stages until implantation [49]. Recently, the establishment of bovine embryonic stem cells (bESCs) with stable pluripotency marker gene expression and the ability to form teratomas has been reported [50]. Although the study does not include a chimeric complementation assay including possible germline transmission, the results are very promising, as bESCs would immensely broaden the available toolkit for genome editing, functional in vitro studies, etc., and advantageously, bESCs were shown to be fit for SCNT.

Groundbreaking studies in the mouse model have investigated how the first differentiation events in early embryos are induced and regulated and how pluripotency is maintained. Research in other species than mouse uncovered features that are species specific and therefore suggests, that alternative model organisms might serve as complementary research tools to study human development, as they share more similarities. The bovine embryo, being similar to the human regarding FGF/MAPK signaling and in his OCT4 KO phenotype, is especially interesting, because state of the art in bovine embryology already provides the necessary foundation for innovative research that will enhance our overall understanding of preimplantation development.

Funding Statement

This work was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under grant 405453332.

Disclosure statement

No potential conflict of interest was reported by the authors.