A fragment of an ancient virus switches on a gene that early human embryo models need

In Short. A copy of an ancient virus in human DNA turns on calbindin, a calcium-handling protein, in early human embryos. When scientists lowered calbindin in stem cells, the cells grew poorly and failed to form models of early embryos.

ABOUT 13M years ago, after the ancestors of humans, chimpanzees and gorillas split from those of the orangutan, a virus left a copy of its genes in the DNA of an ape. That copy is still there, on chromosome 8 of every living person, and according to a peer-reviewed paper published on October 9th in Science Advances, it now helps switch on a gene that early human embryos make. The work, by Pape, Kassiotis and colleagues at the Francis Crick Institute in London, at the University of Cambridge and Imperial College London, (doi:10.1126/sciadv.aee4523) ties one specific viral insert to the growth of stem-cell models of the earliest stages of human life.

A DNA copy of a virus that infected an ancestor and is now inherited in every cell is called an endogenous retrovirus. Some of these copies contain stretches of DNA that turn nearby genes on or off. One family of them, known as HERVH, is especially active in human embryos and stem cells, and a 2022 analysis in eLife by Carter and colleagues, who are not part of the Crick team, found that different subtypes of HERVH are active at different stages of early development.

The gene in question makes calbindin, a protein that binds calcium inside cells. In adults it is made mainly in the brain and kidney, where, the authors write, it protects cells from too much calcium and stops them entering senescence, a state in which cells stay alive but stop dividing. The new paper finds that calbindin is also made in human embryos in the days before they attach to the womb and while the embryo is still a hollow ball of cells. It is not made in marmoset or macaque embryos at the same stages, which lack the viral insert. In human embryos most of the calbindin comes from a version of the gene that starts inside the viral DNA.

Researchers cannot easily experiment on human embryos, so the team used structures grown from human embryonic stem cells. Blastoids copy the hollow ball of cells before it attaches to the womb. Amnioids copy the amnion, the membrane that will later surround the fetus. The team used CRISPR, a gene-editing tool, to damage the calbindin gene in stem cells. None of the edited cell lines lost calbindin entirely. Each kept between 13% and 53% of the normal amount, through a shortened version of the protein that skips the damaged section. Edits that weaken a gene rather than remove it are called hypomorphic, and the authors suggest the cells may be under pressure to keep a minimum level of calbindin.

Even with some calbindin left, the edited stem cells grew more slowly, showed signs of senescence and began turning into other cell types on their own. They formed fewer colonies of the kind that resemble the earliest embryo. Their blastoids formed poorly, and the few that did form lacked GATA4, a marker of one of the embryo’s first specialized layers. They failed completely to form amnioids.

The same viral switch has a role in disease. In 2023 the same Crick group reported in the Journal of Clinical Investigation (Attig and colleagues) that lung-cancer cells use the HERVH-driven form of calbindin to keep growing and avoid senescence. The new paper suggests that the cancer cells are switching on a program that early human embryos already rely on. Because both papers come from one laboratory, the link between them has yet to be tested by another group.

The results come from stem cells and embryo models, not from intact human embryos, and the edits lowered calbindin rather than removing it. The authors also say they have not worked out exactly how calbindin’s handling of calcium keeps the cells growing. Taken together, the work suggests that a viral insert acquired within roughly the last 13m years became part of how early human embryo models develop.