Functional studies of signaling pathways in peri-implantation development of the mouse embryo by RNAi

Background

Studies of gene function in the mouse have relied mainly on gene targeting via homologous recombination. However, this approach is difficult to apply in specific windows of time, and to simultaneously knock-down multiple genes. Here we report an efficient method for dsRNA-mediated gene silencing in late cleavage-stage mouse embryos that permits examination of phenotypes at post-implantation stages.

Results

We show that introduction of Bmp4 dsRNA into intact blastocysts by electroporation recapitulates the genetic Bmp4 null phenotype at gastrulation. It also reveals a novel role for Bmp4 in the regulation the anterior visceral endoderm specific gene expression and its positioning. We also show that RNAi can be used to simultaneously target several genes. When applied to the three murine isoforms of Dishevelled, it leads to earlier defects than previously observed in double knock-outs. These include severe delays in post-implantation development and defects in the anterior midline and neural folds at headfold stages.

Conclusion

Our results indicate that the BMP4 signalling pathway contributes to the development of the anterior visceral endoderm, and reveal an early functional redundancy between the products of the murine Dishevelled genes. The proposed approach constitutes a powerful tool to screen the functions of genes that govern the development of the mouse embryo.

Cleavage pattern and emerging asymmetry of the mouse embryo

Early mammalian development is regulative - it is flexible and responsive to experimental intervention. This flexibility could be explained if embryogenesis were originally completely unbiased and disordered; order and determination of cells only arising later. Alternatively, regulative behaviour could be consistent with the embryo having some order or bias from the very beginning, with inflexibility and cell determination increasing steadily over time. Recent evidence supports the second view and indicates that the sequence and the orientations of cell divisions help to build the first asymmetries.

Spatial arrangement of individual 4-cell stage blastomeres and the order in which they are generated correlate with blastocyst pattern in the mouse embryo

In the unperturbed development of the mouse embryo one of the 2-cell blastomeres tends to contribute its progeny predominantly to the embryonic and the other to the abembryonic part of the blastocyst. However, a significant minority of embryos (20-30%) do not show this correlation. In this study, we have used non-invasive lineage tracing to determine whether development of blastocyst pattern shows any correlation with the orientation and order of the second cleavage divisions that result in specific positioning of blastomeres at the 4-cell stage. Although the orientation and order of the second cleavages are not predetermined, in the great majority (80%) of embryos the spatial arrangement of 4-cell blastomeres is consistent with one of the second cleavages occurring meridionally and the other equatorially or obliquely with respect to the polar body. In such cleaving embryos, one of the 2-cell stage blastomeres tends to contribute to embryonic while the other contributes predominantly to abembryonic part of the blastocyst. Thus, in these embryos the outcome of the first cleavage tends to correlate with the orientation of the blastocyst embryonic-abembryonic axis. However, the order of blastomere divisions predicts a specific polarity for this axis only when the earlier 2-cell blastomere to divide does so meridionally. In contrast to the above two groups, in those embryos in which both second cleavage divisions occur in a similar orientation, either meridionally or equatorially, we do not observe any tendency for the 2-cell blastomeres to contribute to specific blastocyst parts. We find that all these groups of embryos develop to term with similar success, with the exception of those in which both second cleavage divisions occur equatorially whose development can be compromised. We conclude that the orientations and order of the second cleavages are not predetermined; they correlate with the development of blastocyst patterning; and that the majority, but not all, of these cleavage patterns allow equally successful development.

First cell fate decisions and spatial patterning in the early mouse embryo

A growing body of evidence indicates that although the early mouse embryo retains flexibility in responding to perturbations, its patterning is initiated at the earliest developmental stages. There are a few spatial cues that are able to influence the pattern of cleavage divisions: one of these lies in the vicinity of the previous meiotic division, the second is associated with the sperm entry and, related to this, the third is the cell shape. Furthermore, the first cleavage separates the zygote into two cells that tend to follow distinguishable fates: one contributes mainly to the embryonic part of the blastocyst, and the other to the abembryonic. The cumulative effect of the early asymmetries generated through cleavage might lead to asymmetric interactions between the first lineages of cells. This could influence development of patterning after implantation. These early polarity cues serve to bias patterning and not as definitive determinants.

The first cleavage of the mouse zygote predicts the blastocyst axis

One of the unanswered questions in mammalian development is how the embryonic-abembryonic axis of the blastocyst is first established. It is possible that the first cleavage division contributes to this process, because in most mouse embryos the progeny of one two-cell blastomere primarily populate the embryonic part of the blastocyst and the progeny of its sister populate the abembryonic part. However, it is not known whether the embryonic-abembryonic axis is set up by the first cleavage itself, by polarity in the oocyte that then sets the first cleavage plane with respect to the animal pole, or indeed whether it can be divorced entirely from the first cleavage and established in relation to the animal pole. Here we test the importance of the orientation of the first cleavage by imposing an elongated shape on the zygote so that the division no longer passes close to the animal pole, marked by the second polar body. Non-invasive lineage tracing shows that even when the first cleavage occurs along the short axis imposed by this experimental treatment, the progeny of the resulting two-cell blastomeres tend to populate the respective embryonic and abembryonic parts of the blastocyst. Thus, the first cleavage contributes to breaking the symmetry of the embryo, generating blastomeres with different developmental characteristics.

Four-cell stage mouse blastomeres have different developmental properties

Blastomeres of the early mouse embryo are thought to be equivalent in their developmental properties at least until the eight-cell stage. However, the experiments that have led to this conclusion could not have taken into account either the spatial origin of individual blastomeres or the spatial allocation and fate of their progeny. We have therefore readdressed this issue having defined cell lineages in mouse embryos undergoing different patterns of cleavage in their second division cycle. This has enabled us to identify a major group of embryos in which we can predict not only the spatial origin of each given four-cell blastomeres, but also which region of the blastocyst is most likely to be occupied by its progeny. We show that a pattern of second cleavage divisions in which a meridional division is followed by one that is equatorial or oblique allows us to identify blastomeres that differ in their fate and in their developmental properties both from each other and from their cousins. We find that one of these four-cell stage blastomeres that inherits some vegetal membrane marked in the previous cleavage cycle tends to contribute to mural trophectoderm. The progeny of its sister tend to donate cells to part of the ICM lining the blastocyst cavity and its associated trophectoderm. Chimaeras made entirely of these equatorially or obliquely derived blastomeres show developmental abnormalities in both late preimplantation and early postimplantation development. By contrast, chimaeras made from four-cell stage blastomeres from early meridional divisions develop normally. The developmental defects of chimaeras made from the most vegetal blastomeres that result from later second cleavages are the most severe and following transplantation into foster mothers they fail to develop to term. However, when such individual four-cell blastomeres are surrounded by blastomeres from random positions, they are able to contribute to all embryonic lineages. In conclusion, this study shows that while all four-cell blastomeres can have full developmental potential, they differ in their individual developmental properties according to their origin in the embryo from as early as the four-cell stage.

Downregulation of Par3 and aPKC function directs cells towards the ICM in the preimplantation mouse embryo

Generation of inside cells that develop into inner cell mass (ICM) and outside cells that develop into trophectoderm is central to the development of the early mouse embryo. Critical to this decision is the development of cell polarity and the associated asymmetric (differentiative) divisions of the 8-cell-stage blastomeres. The underlying molecular mechanisms for these events are not understood. As the Par3/aPKC complex has a role in establishing cellular polarity and division orientation in other systems, we explored its potential function in the developing mouse embryo. We show that both Par3 and aPKC adopt a polarized localization from the 8-cell stage onwards and that manipulating their function re-directs cell positioning and consequently influences cell fate. Injection of dsRNA against Par3 or mRNA for a dominant negative form of aPKC into a random blastomere at the 4-cell stage directs progeny of the injected cell into the inside part of the embryo. This appears to result from both an increased frequency by which such cells undertake differentiative divisions and their decreased probability of retaining outside positions. Thus, the natural spatial allocation of blastomere progeny can be over-ridden by downregulation of Par3 or aPKC, leading to a deceased tendency for them to remain outside and so develop into trophectoderm. In addition, this experimental approach illustrates a powerful means of manipulating gene expression in a specific clonal population of cells in the preimplantation embryo.

Directing pluripotent cell differentiation using ?diced RNA? in transient transfection

Embryonic stem (ES) and embryonic carcinoma (EC) cells are pluripotent and have the capacity to differentiate into many cell types. The ability to direct their differentiation should have considerable practical applications. Here, we first report the use of diced short interfering RNAi against Oct4 in a transient approach, to direct differentiation of ES towards the trophectoderm lineage. We then apply this approach to downregulate Smad4 in mouse P19 EC cells. We have found that this leads to an increase in the levels of Pax6 (a neuroectoderm marker), reduction in the levels of Brachyury (a mesoderm marker), and a 3-fold increase in the number of betaIII tubulin-positive colonies when these cells were allowed to differentiate. This indicates a redirection of cell fate towards the neuroectoderm lineage. Thus, transient RNAi could provide a valuable tool to direct pluripotent cells along specific pathways of differentiation while circumventing permanent genetic changes.

A genome-wide study of gene activity reveals developmental signaling pathways in the preimplantation mouse embryo

The preimplantation development of the mammalian embryo encompasses a series of critical events: the transition from oocyte to embryo, the first cell divisions, the establishment of cellular contacts, the first lineage differentiation-all the first subtle steps toward a future body plan. Here, we use microarrays to explore gene activity during preimplantation development. We reveal robust and dynamic patterns of stage-specific gene activity that fall into two major phases, one up to the 2-cell stage (oocyte-to-embryo transition) and one after the 4-cell stage (cellular differentiation). The mouse oocyte and early embryo express components of multiple signaling pathways including those downstream of Wnt, BMP, and Notch, indicating that conserved regulators of cell fate and pattern formation are likely to function at the earliest embryonic stages. Overall, these data provide a detailed temporal profile of gene expression that reveals the richness of signaling processes in early mammalian development.

First Cleavage of the Mouse Embryo Responds to Change in Egg Shape at Fertilization

Although mouse development is regulative, the cleavage pattern of the embryo is not random. The first cleavage tends to relate to the site of the previous meiosis. Sperm entry might provide a second cue, but evidence for and against this is indirect and has been debated. To resolve whether sperm entry position relates to the first cleavage, we have followed development from fertilization by time-lapse imaging. This directly showed cytokinesis passes close to the site of the previous meiosis and to both the sperm entry site and trajectory of the male pronucleus in a significant majority of eggs. We detected asymmetric distribution of Par6 protein in relation to the site of meiosis, but not sperm entry. Unexpectedly, we found the egg becomes flattened upon fertilization in an actin-mediated process. The sperm entry position tends to lie at one end of the short axis along which cleavage will pass. When we manipulated eggs to change their shape, this repositioned the cleavage plane such that eggs divided along their experimentally imposed short axis. Such manipulated eggs were able to develop to term, emphasizing the regulative nature of their development.

The Anterior-Posterior Axis Emerges Respecting the Morphology of the Mouse Embryo that Changes and Aligns with the Uterus before Gastrulation

BACKGROUND

When the anterior-posterior axis of the mouse embryo becomes explicit at gastrulation, it is almost perpendicular to the long uterine axis. This led to the belief that the uterus could play a key role in positioning this future body axis.

RESULTS

Here, we demonstrate that when the anterior-posterior axis first emerges it does not respect the axes of the uterus but, rather, the morphology of the embryo. Unexpectedly, the emerging anterior-posterior axis is initially aligned not with the long, but the short axis of the embryo. Then whether the embryo develops in vitro or in utero, the anterior-posterior axis becomes aligned with the long axis of embryo just prior to gastrulation. Of three mechanisms that could account for this apparent shift in anterior-posterior axis orientation-cell migration, spatial change of gene expression, or change in embryo shape-lineage tracing studies favor a shape change accompanied by restriction of the expression domain of anterior markers. This property of the embryo must be modulated by interactions with the uterus as ultimately the anterior-posterior and long axes of the embryo align with the left-right uterine axis.

CONCLUSIONS

The emerging anterior-posterior axis relates to embryo morphology rather than that of the uterus. The apparent shift in its orientation to align with the long embryonic axis and with the uterus is associated with a change in embryo shape and a refinement of anterior gene expression pattern. This suggests an interdependence between anterior-posterior gene expression, the shape of the embryo, and the uterus.

RNA interference by production of short hairpin dsRNA in ES cells, their differentiated derivatives, and in somatic cell lines

dsRNA of several hundred nucleotides in length is effective at interfering with gene expression in mouse oocytes, pre-implantation embryos, and embryonic stem (ES) cells but is not as efficient in differentiated cell lines. Here we describe a method to achieve RNA interference in totipotent and differentiated ES cells together with a wide range of other mammalian cell types that is both simple and efficient. It utilizes a linearized plasmid that directs the expression of a hairpin RNA with a 22-nucleotide-paired region. This molecule has a 13-nucleotide 5' overhang that would be subject to capping on its 5' phosphoryl group and thus differs from the ideal structure suggested for effective small interfering RNAs. Thus, it appears either that the structure of small inhibitory RNA molecules may not need to be as precise as previously thought or that such a transcript is efficiently processed to a form that is effective in interfering with gene expression.

Developmental fate of embryonic germ cells (EGCs), in vivo and in vitro

Embryonic germ cells (EGCs) derived from mouse primordial germ cells (PGCs) are known both to colonize all cell lineages of the fetus and to make tumors in vivo. When aggregated with eight-cell embryos, EGCs from a new EGC line expressing green fluorescent protein (GFP) were found to contribute preferentially to the epiblast but unexpectedly were also capable of colonizing primary endoderm. When injected under the kidney capsule, EGCs derived from 12.5 days post coitum (dpc) PGCs formed differentiated tumors. The ability of EGCs to differentiate in an organ culture system depends upon their partners in cell culture. When EGCs, marked with a LacZ transgene, were mixed with disaggregated and reaggregated mouse fetal lung in an organ culture system, they remained undifferentiated. In urogenital ridge reaggregates on the other hand, some EGCs were capable of differentiating to form small epithelial cysts.

Early patterning of the mouse embryo--contributions of sperm and egg

The first cleavage of the fertilised mouse egg divides the zygote into two cells that have a tendency to follow distinguishable fates. One divides first and contributes its progeny predominantly to the embryonic part of the blastocyst, while the other, later dividing cell, contributes mainly to the abembryonic part. We have previously observed that both the plane of this first cleavage and the subsequent order of blastomere division tend to correlate with the position of the fertilisation cone that forms after sperm entry. But does sperm entry contribute to assigning the distinguishable fates to the first two blastomeres or is their fate an intrinsic property of the egg itself? To answer this question we examined the distribution of the progeny of early blastomeres in embryos never penetrated by sperm - parthenogenetic embryos. In contrast to fertilised eggs, we found there is no tendency for the first two parthenogenetic blastomeres to follow different fates. This outcome is independent of whether parthenogenetic eggs are haploid or diploid. Also unlike fertilised eggs, the first 2-cell blastomere to divide in parthenogenetic embryo does not necessarily contribute more cells to the blastocyst. However, even when descendants of the first dividing blastomere do predominate, they show no strong predisposition to occupy the embryonic part. Thus blastomere fate does not appear to be decided by differential cell division alone. Finally, when the cortical cytoplasm at the site of sperm entry is removed, the first cleavage plane no longer tends to divide the embryo into embryonic and abembryonic parts. Together these results indicate that in normal development fertilisation contributes to setting up embryonic patterning, alongside the role of the egg.

Site of the previous meiotic division defines cleavage orientation in the mouse embryo

The conservation of early cleavage patterns in organisms as diverse as echinoderms and mammals suggests that even in highly regulative embryos such as the mouse, division patterns might be important for development. Indeed, the first cleavage divides the fertilized mouse egg into two cells: one cell that contributes predominantly to the embryonic part of the blastocyst, and one that contributes to the abembryonic part. Here we show, by removing, transplanting or duplicating the animal or vegetal poles of the mouse egg, that a spatial cue at the animal pole orients the plane of this initial division. Embryos with duplicated animal, but not vegetal, poles show abnormalities in chromosome segregation that compromise their development. Our results show that localized factors in the mammalian egg orient the spindle and so define the initial cleavage plane. In increased dosage, however, these factors are detrimental to the correct execution of division.

Efficient delivery of dsRNA into zona-enclosed mouse oocytes and preimplantation embryos by electroporation

Conditions for the electroporation of mouse oocytes and preimplantation embryos have been optimised by following the incorporation of rhodamine labeled dextran. This procedure includes a step to weaken but not remove the zona pellucida that helps achieve good survival. This approach has been applied to introduce double-stranded RNA for c-mos into oocytes and green fluorescent protein (GFP) into transgenic GFP-expressing embryos at the 1- and 4-cell stages. In both cases we were able to observe sequence-specific interference with the expression of the target gene--a failure of oocytes to arrest at metaphase II and a loss in the green fluorescence of embryos by the morula or blastocyst stages. These effects could be observed in multiple oocytes or embryos allowed to develop together following electroporation.

Sperm entry position provides a surface marker for the first cleavage plane of the mouse zygote

The sperm entry position (SEP) of the mouse egg, labelled by placing a bead at the fertilisation cone, tends to be associated with the first cleavage plane (Piotrowska and Zernicka-Goetz: Nature 409:517-521, 2001). Nevertheless, in up to one-fourth of embryos the cleavage furrow did not pass close to the bead, and following the division the bead marked the cleavage plane in only 60% of cases. This raised the question of whether such variability arose from the labelling itself or had a biological basis. The zona pellucida was not responsible for this effect because similar results were obtained in its presence or absence. However, this variability could be attributable to the large size of the fertilisation cone relative to the SEP. Therefore, we have developed a means of fluorescently labelling sperm that can record the exact site of its penetration when the label transfers to the egg surface. This approach indicates that the SEP marks the first cleavage in the great majority (88%) of embryos. In conclusion, direct sperm labelling shows precisely the correlation between the SEP and the first cleavage, although there is natural variability in this process.

Patterning of the embryo: the first spatial decisions in the life of a mouse

Although in most species the polarity of the embryo takes its roots from the spatial patterning of the egg, mammals were viewed as an exception. This was because the anteroposterior polarity of the mouse embryo could not be seen until gastrulation, and no developmental cues were known that could define polarity at earlier stages. Why should we now re-consider this view? While mechanisms of axis formation in mammals could, in principle, be unique, the evolutionary conservation of numerous other developmental processes raises the question of why mammals would have evolved a different way or timing of organising their embryonic polarity. Indeed, recent evidence shows that well before the onset of gastrulation, the mouse embryo initiates asymmetric patterns of gene expression in its visceral endoderm. Although this extra-embryonic tissue does not contribute to the body itself, it is involved in axis formation. Other recent work has revealed that spatial distribution of cells in the visceral endoderm can be traced back to polarity present at the blastocyst stage. These insights have raised the possibility that embryonic polarity might also originate early during development of mammalian embryos. Indeed it now appears that there are at least two spatial cues that operate in the mouse egg to shape polarity of the blastocyst. One of these is at the animal pole, which is defined by the site of female meiosis, and another is associated with the position of sperm entry. In this review I discuss these recent findings, which have led to the recognition that mouse embryos initiate development of their polarity at the earliest stages of their life. This novel perspective raises questions about the nature of cellular and molecular mechanisms that could convert developmental cues in the zygote to axes of the blastocyst, and hence into polarity of the post-implantation embryo. It also brings to light the need to understand how such mechanisms could enable early mouse development to be so regulative.

Blastomeres arising from the first cleavage division have distinguishable fates in normal mouse development

Two independent studies have recently suggested similar models in which the embryonic and abembryonic parts of the mouse blastocyst become separated already by the first cleavage division. However, no lineage tracing studies carried out so far on early embryos provide the support for such a hypothesis. Thus, to re-examine the fate of blastomeres of the two-cell mouse embryo, we have undertaken lineage tracing studies using a non-perturbing method. We show that two-cell stage blastomeres have a strong tendency to develop into cells that comprise either the embryonic or the abembryonic parts of the blastocyst. Moreover, the two-cell stage blastomere that is first to divide will preferentially contribute its progeny to the embryonic part. Nevertheless, we find that the blastocyst embryonic-abembryonic axis is not perfectly orthogonal to the first cleavage plane, but often shows some angular displacement from it. Consequently, there is a boundary zone adjacent to the interior margin of the blastocoel that is populated by cells derived from both earlier and later dividing blastomeres. The majority of cells that inhabit this boundary region are, however, derived from the later dividing two-cell stage blastomere that contributes predominantly to the abembryonic part of the blastocyst. Thus, at the two-cell stage it is already possible to predict which cell will contribute a greater proportion of its progeny to the abembryonic part of the blastocyst (region including the blastocyst cavity) and which to the embryonic part (region containing the inner cell mass) that will give rise to the embryo proper.

Use of Green Fluorescent Protein in mouse embryos

Green Fluorescent Protein (GFP) has rapidly been established as a versatile and powerful cell marker in many organisms. Initial problems in using it in mammalian cells were solved by introducing mutations to increase its solubility at higher temperatures, such that GFP has now been used as a reporter in both gene expression and cell lineage studies, and to localize proteins within mammalian cells. GFP has two unique advantages: (i) the protein becomes fluorescent in an autocatalytic reaction, so that it can be introduced into any cell type simply as a cDNA or mRNA, or as protein; (ii) it is "bright" enough to be visualized in living cells under conditions that do not cause photodamage to the cells. In this article we outline the ways in which we have used GFP mRNA and cDNA in our studies of mouse cell lineages, and to characterize the behavior of proteins within the embryos.

Role for sperm in spatial patterning of the early mouse embryo

Despite an apparent lack of determinants that specify cell fate, spatial patterning of the mouse embryo is evident early in development. The axis of the post-implantation egg cylinder can be traced back to organization of the pre-implantation blastocyst. This in turn reflects the organization of the cleavage-stage embryo and the animal-vegetal axis of the zygote. These findings suggest that the cleavage pattern of normal development may be involved in specifying the future embryonic axis; however, how and when this pattern becomes established is unclear. In many animal eggs, the sperm entry position provides a cue for embryonic patterning, but until now no such role has been found in mammals. Here we show that the sperm entry position predicts the plane of initial cleavage of the mouse egg and can define embryonic and abembryonic halves of the future blastocyst. In addition, the cell inheriting the sperm entry position acquires a division advantage and tends to cleave ahead of its sister. As cell identity reflects the timing of the early cleavages, these events together shape the blastocyst whose organization will become translated into axial patterning after implantation. We present a model for axial development that accommodates these findings with the regulative nature of mouse embryos.

Animal and vegetal poles of the mouse egg predict the polarity of the embryonic axis, yet are nonessential for development

Recent studies suggest early (preimplantation) events might be important in the development of polarity in mammalian embryos. We report here lineage tracing experiments with green fluorescent protein showing that cells located either near to or opposite the polar body at the 8-cell stage of the mouse embryo retain their same relative positions in the blastocyst. Thus they come to lie on either end of an axis of symmetry of the blastocyst that has recently been shown to correlate with the anterior-posterior axis of the postimplantation embryo (see R. J. Weber, R. A. Pedersen, F. Wianny, M. J. Evans and M. Zernicka-Goetz (1999). Development 126, 5591–5598). The embryonic axes of the mouse can therefore be related to the position of the polar body at the 8-cell stage, and by implication, to the animal-vegetal axis of the zygote. However, we also show that chimeric embryos constructed from 2-cell stage blastomeres from which the animal or the vegetal poles have been removed can develop into normal blastocysts and become fertile adult mice. This is also true of chimeras composed of animal or vegetal pole cells derived through normal cleavage to the 8-cell stage. We discuss that although polarity of the postimplantation embryo can be traced back to the 8-cell stage and in turn to the organisation of the egg, it is not absolutely fixed by this time.