How exactly do embryo models differ from natural human embryos?

A newly devised benchmark provides a measure of how closely different human stem-cell-based embryo models (SCBEMs) resemble early natural human embryos.

Researchers in Australia have built a detailed reference map of how cells behave in human embryos around the time of implantation, and used this to evaluate blastoids – a type of SCBEM that imitates the blastocyst stage of embryo development. The researchers compared four published methods for generating blastoids, finding that the resulting SCBEMs differ widely from one another and that none of them fully matches natural human blastocysts.

'Human embryo models have enormous potential for studying the earliest days of an embryo's development, but there has been no consistent way to assess how accurately these reflect real human development', said Dr Pengyi Yang from the University of Sydney, lead author of the study published in Cell Systems.

To determine whether a blastoid that resembles an embryo also behaves like one, the team first pooled previously published data on more than 14,000 individual cells from human embryos, to map how cells normally change and organise themselves before, during and after implantation.

Each of the four methods of generating blastoids was then assessed against this map, with consideration given to the identity of blastoid cells, the timing of cell development and how the different cell types relate to one another.

The team found substantial differences between the four protocols. Some methods produced all three of the main cell types present in natural blastocysts relatively well, while others represented certain cell types poorly, or generated many cells that could not be confidently matched to a known equivalent in a natural embryo. No method perfectly recapitulated a natural blastocyst.

SCBEMs such as blastoids have potential uses in the study of fertility, pregnancy success and the first days of human development (see BioNews 1335 and 1354). However, Dr Yang cautioned that current models are 'not biologically equivalent to real human embryos,' and urged care over the conclusions drawn from them.

The authors noted that their approach only measures which genes are active in cells, and cannot show definitively whether a SCBEM is capable of implanting in a womb lining (or in a model of a womb lining) or whether it is capable of further development. They noted that according to earlier studies, blastoids do not develop beyond a few days, and improved culture systems will be needed in order to make further progress (see BioNews 1321a and 1321b).

The researchers hope that their computational framework can be used as a benchmark to help refine SCBEMs, and make comparisons between them more rigorous. They have made their reference data and testing tools publicly available, so that other laboratories can test their own models against the same standards.

'Our work provides a roadmap for improving embryo models and ensuring scientific claims remain grounded in what the models can actually support,' said Dr Yang.