The early stages of human development are normally hidden within the womb, but improved techniques for culturing embryos from the blastocyst stage promise to make these steps easier to investigate. See Letter p.251
Studying early development in human embryos is a challenge. Few embryos are available, and research is subject to considerable ethical and legal constraints. But understanding early-stage development is vital for improving reproductive technologies, enhancing stem-cell cultures for regenerative medicine and examining early pregnancy losses. In two papers, Deglincerti et al.1 (page 251 of this issue) and Shahbazi et al.2 (in Nature Cell Biology) report that human embryos derived from in vitro fertilization (IVF) can self-organize in a Petri dish, forming the founding cell lineages of the fetus and its supporting tissues. This is a first step towards a clearer view of the beginnings of human life.
In mammals, including humans, a fertilized egg undergoes a series of cell divisions over the first days of development, leading to the formation of a structure called the blastocyst (Fig. 1). The first cell-lineage decisions are made at this stage, with a lineage called the epiblast, which goes on to form the entire fetus, becoming separated from two lineages that will produce non-embryonic tissues — first the trophectoderm and then the primitive endoderm. The trophectoderm gives rise to cells that form most of the placenta, whereas the primitive endoderm forms some layers of the yolk sac, which is required for early fetal blood supply.
The mechanisms that underlie blastocyst-stage lineage specification are well understood in mice, and it had been assumed that these pathways are evolutionarily conserved. However, that assumption has been challenged3. Although many of the genes that direct lineage decisions in mouse embryos are expressed in the same lineages in humans, the timing of onset and the upstream pathways that regulate their expression differ between the species4,5.
The blastocyst becomes implanted in the lining of the uterus just five days after fertilization in mice and seven days after in humans. This is a vital period, in which trophectoderm-derived cells begin to interact with the uterus, and the embryo progresses towards perhaps the most crucial step in development — gastrulation, in which an epiblast-derived cell mass called the primitive streak gives rise to the three basic cell layers from which every bodily structure is derived.
In mice, signals from the primitive endoderm and trophectoderm initiate formation of the primitive streak6. But in humans, this period of development has been a complete black box. The only available information has come from rare cross-sections cut through human embryos and from non-human primates, such as rhesus monkeys. Those studies7,8 show that there are major differences between primate and mouse development as the embryo implants in the uterus. Most notably, the mouse epiblast forms a cup-like structure, on one side of which form the primitive streak and amniotic folds (which will later form the fluid-filled amniotic membranes). By contrast, the primate epiblast first forms a central amniotic cavity and then flattens out to form a disc, from which the primitive streak arises at one end (Fig. 1). The spatial relationships between lineages therefore differ between species.
The development of a strategy for culturing human embryos in vitro over the early post-implantation period could improve our understanding of the significance of these differences. Using a system developed for culturing mouse embryos9, Deglincerti et al. and Shahbazi et al. did just that, culturing human embryos derived from IVF up to a stage equivalent to 13 days post-fertilization in vivo. An improved culture medium and a better substrate for embryo attachment seem to have been the key to these advances.
The groups report that blastocysts attach to the dish, the trophectoderm spreads out and shows signs of differentiation into specialized placental cell types, and the primitive endoderm segregates from the epiblast. Shahbazi and colleagues found that a small central cavity develops within the epiblast, apparently because the lineage reorganizes into a radially polarized structure. This is reminiscent of how the amniotic cavity is thought to form in both human and rhesus-monkey embryos7,8, although the absence of good molecular markers of amniotic tissue precludes a conclusive identification of this structure.
Both groups also observed a second cavity in the spreading primitive endoderm, which they equate to the yolk-sac cavity. Deglincerti et al. report that the cells lining this cavity express genes characteristic of trophectoderm-derived cells. The authors suggest that this is a previously unidentified cell type, which they name yolk-sac trophectoderm. However, comparison with anatomical descriptions7 suggests that these cells are more likely to be derived from the primitive endoderm — perhaps with a gene-expression profile that differs from that in mice. Extension of the cultures beyond 12 days led to cavity collapse and disorganized development, although trophectoderm differentiation continued.
Although these studies represent steps towards a firmer understanding of human development over the implantation period, there are many limitations still to overcome. The cultured embryos are largely flattened and two-dimensional, and so are clearly imperfect models of normal 3D embryonic development. In addition, unequivocally identifying cell types, cavities and structures in the cultures is challenging. Genome-wide expression analysis of these features might help to refine the system.
This culture method could enable researchers to probe the role of signalling molecules from the extraembryonic tissues in patterning the epiblast. By comparing these results with the signalling molecules detected in embryonic-stem-cell cultures that mimic events of gastrulation10, we might better understand how to induce human stem cells to differentiate into cell types that have therapeutic potential. The development of a 3D blastocyst culture system, akin to the 'organoid' systems used to model more-mature tissues, could also be informative. If the topological relationships between the different cell types were more normal in such 3D cultures, this might enable gastrulation to occur in vitro.
Currently, human embryo cultures are restricted, by international agreement, to 14 days of growth or the beginning of primitive-streak formation, whichever comes first. If gastrulation were achievable in vitro, what would be the impact on this 14-day rule? Improved and longer cultures could provide important information for basic human biology, improving IVF success rates and the understanding of stem-cell differentiation. However, the development of such culture systems would again raise the question of where to place the ethical limits on human embryo development in vitro. See also Comment page 169
About this article
See also Comment p.169
Science China Life Sciences (2018)