Biologists trying to create laboratory models of the early growth of a human embryo have taken a major step forward. In preprints posted online on Thursday and Friday, four research teams reported using various kinds of human stem cells, some genetically modified, to create ersatz embryos that closely resemble real embryos that are up to 14 days old, replicating a period in human development that is very difficult to study.

The rush of papers was triggered when the leader of one group, developmental biologist Magdalena Zernicka-Goetz, briefly described her team’s results at the International Society for Stem Cell Research (ISSCR) meeting in Boston on Wednesday morning. Other scientists are still assessing the four groups’ claims, but some are already impressed. “The similarity to the natural embryo is remarkable, almost uncanny,” developmental biologist Jesse Veenvliet of the Max Planck Institute of Molecular Cell Biology and Genetics says about the embryo mimics from a lab led by stem cell biologist Jacob Hanna of the Weizmann Institute of Science.

Besides clarifying early human development, the new embryo models could help researchers better understand birth defects and probe the safety of drugs used during pregnancy. But they pose fraught issues. U.K. law, for example, prohibits research on donated in vitro fertilization (IVF) embryos that are beyond 14 days old. But because these new models are formed from human stem cells, not eggs and sperm, that law does not apply to them. The ethical and moral status of creations like these also remains unresolved. Embryo models are “a matter of significant moral discussion and of significant moral concern,” says ethicist J. Benjamin Hurlbut of Arizona State University.

Scientists have previously spurred human stem cells to form structures that closely resemble the blastocyst, the embryonic stage that begins around 5 days after fertilization and implants in the womb. These imitation blastocysts, known as blastoids, help researchers probe events around implantation, which may be key to some infertility. But the blastoids stagnate developmentally, making it hard to investigate changes that occur slightly later. The period of human development immediately after implantation is “truly a black box,” says Hanna.

One way in which researchers have tried to delve into events after the blastocyst stage is by producing ersatz embryos from mouse stem cells. Last year, for example, Hanna and colleagues grew such cells into embryo mimics that sported a beating heart, the rudiments of a brain and spinal cord, and incipient muscles.

To create similar models of human development, Hanna and his team started with cell lines previously cultured from early human embryos and with stem cells converted from adult cells. A good mimic needs to contain not just the cells of the embryo proper, but also the cells that produce the so-called extraembryonic tissues that help nurture the growing offspring, such as the placenta and umbilical cord. Using cell culture media developed by other scientists and mixtures they crafted themselves, the researchers nudged their stem cells to specialize into extraembryonic cell lineages found in a genuine embryo. They then allowed these lineages to mingle with stem cells. The resulting cell clusters showed hallmarks of post-blastocyst embryos, the team reports in its paper on the Biorxiv preprint server.

Zernicka-Goetz, who divides her research between the University of Cambridge and the California Institute of Technology, and her colleagues took a different approach. They too had previously grown mouse stem cells into cellular structures that recapture post-implantation events. In the new work, led by Cambridge PhD student Bailey Weatherbee, they generated extraembryonic cell lineages from two kinds of human embryonic stem cells that had been genetically modified to differentiate into the tissue types when exposed to the antibiotic doxycycline. The scientists then mixed the results with unmodified human embryonic stem cells, and the three types of cells coalesced to produce embryo-like clusters. Several important features of real embryos are present in the mimics, the team says, including the predecessors of sperm and eggs and the beginnings of the amnion, the membrane-like structure that encloses the embryo in a watery environment.

But some researchers in the field argue that this group’s synthetic embryos, which are described in a preprint posted on bioRxiv on Thursday, do not duplicate natural embryos as closely as the mimics produced by Hanna and colleagues because they are missing some cell lineages and their organization is not well-defined. “These structures do not recapitulate all aspects of the embryo, but rather serve as a complementary tool for us to study … key stages of development,” Weatherbee says.

Two other groups also released pre-prints this week describing the generation of post-implantation-like embryo models. One, from a University of Pittsburgh team led by Mo Ebrahimkhani, used genetically modified induced pluripotent stem cells—the embryo-like cells created from adult human cells—to form embryonic and extraembryonic tissues in in cell clusters they call iDiscoids. The second group, a Chinese team led by Tianqing Li of Kunming University of Science and Technology, created its “E-assembloids” by mixing freshly derived human embryonic stem cells and extraembryonic cells.

Researchers now want to improve their procedures to produce even more accurate embryo stand-ins, says Hanna, so that they can obtain a clearer picture of how early human development unfolds. His team and the Cambridge team stopped their experiments when the ersatz embryos were the equivalent of natural 14-day-old human embryos. That cutoff was long the agreed-upon limit for growing IVF-derived human embryos. However, Zernicka-Goetz and other researchers have argued that studying IVF embryos for longer would be valuable. ISSCR, the organization that sets permissible research practices in the field, dropped its strict 14-day limit on culture of IVF embryos in a 2021 update to its guidelines. The guidelines do not set a specific time limit for embryo models like the ones generated from stem cells.

These new embryo mimics add a fresh element to discussions about the status of such models of human development. “The big question is how the boundary between a tissue culture and a human organism is going to be drawn and on what criteria,” says Hurlbut.