Two teams of scientists have announced that they have grown embryo-like structures, made entirely from human stem cells, that are more advanced than any previous efforts. The synthetic embryos developed to a stage equivalent to that of natural embryos about 14 days after fertilization.

Such experiments could provide opportunities to study human embryonic development at later stages than ever before. But they also raise ethical and legal questions about the status of such ‘embryo models’ and how they should be regulated.

The work is described in two preprint studies1,2, posted to the bioRxiv server on 15 June by teams led by developmental biologist Magdalena Zernicka-Goetz at the University of Cambridge, UK, and stem-cell biologist Jacob Hanna at the Weizmann Institute of Science in Rehovot, Israel. Both groups had previously presented their findings at scientific meetings, with the work making headlines after Zernicka-Goetz spoke about her results at the annual meeting of the International Society for Stem Cell Research in Boston, Massachusetts, on 14 June.

Nature spoke to scientists about what these developments could mean for research on human embryos.

What did the researchers do and how does it differ from previous work?

Both teams allowed their embryo-like structures to self-assemble from human embryonic stem cells, some of which had been converted into cell types resembling the stem cells that form a placenta and the cells that form the yolk sac outside a naturally developing embryo.

The researchers say that the resulting embryo models show structures and gene transcription profiles found in human embryos between 6 and 14 days after fertilization — up to the onset of the stage called gastrulation, when the cells that will form the embryo become organized into a layer between the amniotic cavity and the yolk sac.

Researchers have made similar entities before from the stem cells of humans and other animals. Last year, both Zernicka-Goetz’s and Hanna’s teams used similar techniques to create embryo models from mouse cells that developed all the way up to the stage at which organs such as the heart and brain begin to form3,4. Human embryo models haven’t got that far, but in a preprint posted on bioRxiv on 17 May, stem-cell biologist Ali Brivanlou at the Rockefeller University in New York City and his co-workers reported the development of human embryo models that show signatures of gastrulation equivalent to those seen at around 12 days after fertilization5. The latest studies1,2 say that they have made the most advanced human embryo models so far.

What is the significance of the embryos lasting for 14 days?

Research on natural human embryos tends to observe a widely adopted guideline — enforced by law in many countries — that human embryos should not be cultured in the laboratory beyond 14 days. This means that researchers have to use animal models to study later stages of embryonic development. These do not necessarily reflect the corresponding processes in humans.

But because in most countries embryo models do not meet the formal definition of an embryo, they are not subject to such restrictions. “We sought to develop a tool to ask specific questions about the second week of human embryo development, since using actual human embryos in research is ethically and technically challenging,” says Zernicka-Goetz.

Models that are older than 14 days could therefore offer important insight into human embryonic development that cannot currently be obtained. They could be used to study developmental defects, for example, or pregnancy loss.

Why is the research scientifically controversial?

Growing embryo models to ever-later stages of development has become a highly competitive race, provoking many arguments about the merits of claims made.

It remains to be seen whether the claims made by the latest studies, neither of which has yet been peer-reviewed, will pass muster. Alfonso Martinez Arias, a developmental biologist at Pompeu Fabra University in Barcelona, Spain, says there is “nothing” in the results described by Zernicka-Goetz and her colleagues that can be considered analogous to real 14-day embryos. “What we can see is masses of cells separated into compartments, but no embryo-like organization,” he says. He thinks that the over-expression of some genes needed to produce the extra-embryonic cell types “confuses what cells do”, and argues that the results do not show anything that goes beyond earlier work.

Zernicka-Goetz acknowledges the limitations of embryo models for studying development. “These structures do not recapitulate all aspects of the embryo,” she says, “but rather serve as a complementary tool for us to study the differentiation of specific tissues during key stages of development.”

What about ethical concerns?

The results have sparked a discussion about the status of human embryo models in general, and whether they should continue to fall outside legislation on human embryos. Although they are not subject to the 14-day rule, the embryo-like structures reported by Zernicka-Goetz’s and Hanna’s groups do need to respect guidelines and regulations on the use of the human embryonic stem cells from which they are made. But other groups have made embryo models using ‘induced’ stem cells derived from adult tissues6, which are not governed by such rules. Those embryo models “are not regulated at all”, says Robin Lovell-Badge, a cell biologist at the Francis Crick Institute in London.

So far, no one has made embryo models that have the capacity to develop into human beings, but a recent study on monkey embryo models showed that such models could induce pregnancy (which terminated spontaneously soon after) if placed in the uterus7.

Some researchers think that a revised definition of an embryo is needed to clarify the issues. For others, the whole purpose of embryo models is to circumvent the current constraints on embryo research. “These models do challenge the need to stick to the 14-day rule,” says Lovell-Badge, who was part of a committee that, in 2021, recommended relaxing the guideline.

In any case, there are significant challenges to making human embryo models that live much longer, says Martinez Arias. Creating structures that develop up to 21 days “will not be easy”, he says. “I will be surprised if [human embryo models] can go beyond it.”