No eggs, no sperm, no uterus: extending the boundaries of mammalian development in vitro

Synthetic embryos developed in vitro until the eighth embryonic day (E8.5) from the report of Tarazi et al. (1)

Published 28 September 2022

Two recently published papers have described experiments in which mouse embryo models were developed from pluripotent stem cells. Mina Popovic and Susana Chuva de Sousa Lopes from ESHRE’s SIG Stem Cells report.

Two research groups have recently achieved the unthinkable, demonstrating that mouse embryo models derived from stem cells have the potential to develop from pre-gastrulation until early organogenesis in vitro. (1,2). The mouse embryo-like structures in these experiments were grown until the equivalent of embryonic day (E)8.5 (a third of a mouse pregnancy). Although many showed clear morphological abnormalities, some structures contained a beating heart, a brain rudiment with fore- and midbrain, patterned neural and gut tubes, migrating primordial germ cell-like cells and progenitors of other organs. Remarkably, they also developed extra-embryonic structures, such as an umbilical cord, amnion and yolk sac that formed blood islands ─ all without the need for maternal tissues.

Over the past years, a flurry of studies have demonstrated the remarkable ability of (mouse and human) pluripotent stem cells to self-assemble into organised embryo-like structures in vitro.(3,4) While traditional developmental biology has been limited by the availability of natural (fertilised) embryos for research, this enhanced stem cell toolkit has enabled several aspects of mammalian peri-implantation development to be captured in vitro. Accordingly, ‘blastoids’ have recapitulated the blastocyst, ‘gastruloids’ model features of axis development and gastrulation, while various other embryoids mimic aspects of epiblast, trophoblast and (pro)amniotic cavity formation. Yet, up until now, none of these models have been able to demonstrate the full developmental potential of natural embryos.

Now, these two independent studies, published in Cell and Nature, have applied similar technologies to generate and culture stem cell-based embryo-like structures, demonstrating their self-organising capacity and unprecedented developmental potential. To generate these structures, the group of Magdalena Zernicka Goetz combined mouse embryonic stem cells (mESCs) with trophoblast stem cells, and extraembryonic endoderm-like cells, using methods previously pioneered by their own group.(5,6) The second group, led by Jacob Hanna, started solely from mESCs, yet some were made to transiently overexpress master regulatory transcription factors to induce both the trophoblast and extra-embryonic endoderm lineage. The aggregated stem cells first assembled into egg-cylinders and then further progressed into complete mouse embryo-like structures.

To maintain the embryo-like structures in culture, both groups used a platform for extended ex-utero culture of natural embryos from E5 to E11, previously optimised by Hanna’s team.(7) In this system, mouse embryos are cultured in glass vials rotating on a drum in the presence of rat (or human) blood serum, with an electronic ventilation system regulating gas and pressure.(8) Following extended culture on the rolling platform, the mouse embryo-like structures showed notable similarities to their natural E8.5 counterparts grown either in utero or ex-utero. Remarkably, they increased in complexity over time towards the formation of differentiated organ primordia. However, unlike natural embryos that can be cultured up to E11 using this system, the embryo-like structures could only reach the morphology of an E8.5 embryo. It remains unknown whether these differences are a result of the stem cell aggregation protocols or varying culture requirements.

Certainly, further optimisation of the technology will be necessary. A large proportion of the embryo-like structures developed abnormally, exhibiting a variety of abnormalities during ex-utero culture, including the complete lack of body segments. Of the normal egg-cylinder-shaped embryo-like structures at E5, only around 2% developed to E8.5, yielding an effective 0.1%–0.5% efficiency from the total initial aggregates generated. These results were similar across both studies.

Heterogeneity during the formation of embryo-like structures also remains a challenge. Efficiency varied substantially between mESC lines, with some lines not able to generate embryo-like structures beyond E6.

Although further work is necessary to improve efficiency and reproducibility, mouse embryo-like models do hold some advantages over natural embryos. Primarily, they are more amenable to genetic modifications and may provide a powerful in vitro system for elucidating the diverse roles of genes during early organogenesis. This may ultimately reduce the need for experimental animals and natural embryos for research. Evaluating developmental pathways in greater detail than ever before could also enhance the efficiency and control of stem cell differentiation protocols for regenerative medicine.

To demonstrate the functionality of their model, the team of Zernicka-Goetz knocked-out Pax6 (a key gene involved in neural tube patterning, brain and eye development) in the embryo-like structures. Markedly, neural tube development was compromised in the structures lacking Pax6, which is consistent with natural embryos missing this gene.

Accordingly, the future development of similar embryo-like models in human may provide insights into longevity, (in)fertility and developmental diseases. Beyond basic research, Jacob Hanna is hopeful that this method may provide a source of new organs and tissues for human transplantation biotechnology. Yet, their use for reproductive purposes is not and should not be considered, especially since the embryo-like structures are in fact genetic clones of the donor stem cells used for their formation.

Nevertheless, translating this system from mouse to human will not be straightforward. Reaching these same stages of organogenesis in the human would correspond to a first-trimester fetus, a path undoubtedly fraught with technical as well as ethical concerns. In practice, capturing the length of human gestation, sheer size of human organ primordia and complexity of these developmental milestones, will certainly be an immense challenge. At present, the possibility of culturing human embryo-like structures beyond gastrulation, particularly in the absence of key maternal cellular constituents and proper implantation assays, remains unknown.

Concurrent to scientific innovation, continued ethical reflection and societal debate remain imperative. Given their benefits for research, it is reasonable to assume that the quality and developmental potential of human embryo-like structures will gradually improve. Moving forward, considering the extent to which the use of these models raises moral concerns characteristic of human embryo research will be essential. At present, it is unclear whether human embryo-like structures which mimic the intact human embryo show developmental potency beyond gastrulation – because of a lack of adequate culture platforms and the 14 day rule, which prohibits in vitro culture of human embryos beyond 14 days. Last year, the International Society for Stem Cell Research recommended relaxing this standard.(9) However, any proposal for the culture of natural embryos or stem cell-based embryo models that mimic the intact human embryo beyond the current 14 day limit must gain broad public support and would require changes in national legislation.

Nonetheless, with the field of developmental biology brimming with continued efforts to refine embryo models, the stem cell toolbox is becoming increasingly valuable. We do anticipate that stem cell-based embryo-like structures will enhance the roadmap for studying early development, offering novel opportunities for exploring the early days of development in real-time and in unprecedented detail.

1. Tarazi S, Aguilera-Castrejon A, Joubran C, et al. Post-gastrulation synthetic embryos generated ex utero from mouse naive ESCs. Cell 2022; 185: 3290-3306.e25.
2. Amadei G, Handford CE, Qiu C, et al. Synthetic embryos complete gastrulation to neurulation and organogenesis. Nature 2022:
3. Rossant J, Tam PPL. Opportunities and challenges with stem cell-based embryo models. Stem cell reports 2021.
4. Veenvliet JV, Lenne PF, Turner DA, et al. Sculpting with stem cells: how models of embryo development take shape. Development 2021; 148: dev192914.
5. Sozen B, Amadei G, Cox A, et al. Self-assembly of embryonic and two extra-embryonic stem cell types into gastrulating embryo-like structures. Nature Cell Biology 2018; 20: 979-989.
6. Amadei G, Lau KYC, De Jonghe J, et al. Inducible stem-cell-derived embryos capture mouse morphogenetic events in vitro. Dev Cell 2021; 56: 366-382.e9.
7. Aguilera-Castrejon A, Oldak B, Shani T, et al. Ex utero mouse embryogenesis from pre-gastrulation to late organogenesis. Nature. 2021;593(7857):119-124.
8. Tam PP, Snow MH. The in vitro culture of primitive-streak-stage mouse embryos. J Embryol Exp Morphol. 1980;59:131-143.
9. See

Get notified of new articles with our ESHRE newsletter.

Sign up and never miss an update