Scientists create primate embryo model to unlock gastrulation's mysteries

Alfonso Martínez-Arias believes gastrulation should be a household word. Most people know they began as a single fertilized cell that divided and multiplied. By day fourteen, they were still an undifferentiated sphere. But seven days later—by day twenty-one—they had acquired a body plan and the rudimentary seeds of every organ, all contained in a structure the size of a pencil tip. That transformation is gastrulation, and until this week, no one had ever watched it happen in a primate.

Observing gastrulation in humans has been technically impossible and ethically forbidden. The legal and moral boundary sits at fourteen days—the point beyond which an embryo cannot split into twins, the moment an individual is said to exist. No researcher has cultivated a human embryo past that threshold, leaving gastrulation as perhaps the deepest mystery in developmental biology. This Thursday, Martínez-Arias and a team of Chinese colleagues published a breakthrough in Nature: the first primate embryo model to reach day twenty-five, three days past the completion of gastrulation. "We have crossed the frontier," Martínez-Arias said.

The stakes are substantial. Of every three fertilized human eggs, only one becomes a baby. In many cases, no one knows why the other two fail. Many congenital malformations that affect people throughout their lives originate in gastrulation errors that are never identified. Understanding what goes wrong during those critical days could reshape how we approach miscarriage and birth defects. Martínez-Arias, an Icrea researcher and emeritus professor at Pompeu Fabra University in Barcelona, collaborated with Zhen Liu's team at the Institute of Neuroscience of the Chinese Academy of Sciences in Shanghai.

Two years ago, Liu's group had attempted to open this black box by implanting primate embryo models into three macaque females. Those models reached day seventeen inside the womb—the first glimpse of the cellular choreography that builds a new individual. But the pregnancies failed shortly after. This time, the team started with macaque embryonic stem cells, capable of becoming any tissue in the body. They guided these cells to the blastocyst stage—a sphere of a few hundred cells—then applied a new three-dimensional liquid culture technique to allow them to grow continuously to day twenty-five. When the researchers compared these engineered embryoids to naturally developing macaque embryos, they found the models faithfully reproduced the essential features: the seeds of muscle, digestive tract, blood, brain, and even the germ cells from which eggs and sperm arise.

Until now, only mouse embryos had been cultured past gastrulation, but mice differ from humans in crucial ways, Martínez-Arias notes. These new models open the door to studying the process in primates—the group to which humans belong. It is the closest anyone has come to studying human gastrulation. The models also sidestep the ethical complications of using and destroying natural primate embryos.

Martínez-Arias cautions that in developmental biology, almost nothing begins when we think it does. Mutations produce visible consequences when a fetus is seven or eight weeks old, but their origin often traces back to gastrulation. "If we can search for that beginning and find it, we are taking the first step toward a possible treatment," he said. He predicts these models will prove especially valuable in teratology—the study of embryonic malformations—and in testing the toxicity of environmental contaminants.

Nicolas Rivron, a biologist whose team was among the first to create human embryo models before gastrulation, called the work important. He noted it is the first demonstration that such models can grow past the critical moment when the body plan emerges. But he cautioned that these structures—whether from mice, primates, or humans—remain imperfect, and it is unclear whether they truly reflect natural development. Xiangyu Kong and Thorold Theunissen, biologists at Washington University in St. Louis, highlighted that the work provides "for the first time a model that opens a window to observe primate development continuously." They wondered whether the same innovations could create human embryo models beyond fourteen days. The team is trying, but the embryoid models resist maturing beyond their current limits.