Six days old is the age at which a human embryo makes one of its most fateful decisions — how to stitch its DNA back together after a cut. And a new study has just revealed something surprising about that repair: embryos do a great job fixing small nicks, but they stumble badly when the damage is bigger. That finding, published in the journal Nature, could one day make gene therapy far safer.
Here is why it matters. Scientists dream of correcting inherited diseases at the very start of life, before an illness has a chance to develop or be passed on. But editing an embryo's genes means deliberately wounding its DNA, and the whole approach only works if the embryo repairs that wound faithfully. For years, no one really knew how well early embryos handled the job. The team behind this study, led by first author Štěpán Jeřábek of Columbia University and the Czech academy's IOCB Prague, set out to find out.
They used two different editing tools on two well-studied genes in living embryos. The famous CRISPR–Cas9 scissors cut clean through both strands of the DNA — the bigger kind of wound. Base editing, by contrast, caused only a tiny single-strand nick. Then the researchers watched how the embryos' own repair machinery responded to each.
The results were stark. "Early human embryos are able to effectively repair single-strand DNA damage, whereas the repair of double-strand breaks is significantly less reliable at this stage of development," Jeřábek said. In plain terms: small cuts heal well; big cuts are risky business.
The team also pulled off a technical feat. They managed to isolate stem cells from the edited six-day-old embryos, providing enough genetic material for deep analysis and a way to study how the changes ripple into later generations of cells.
Two scientists from IOCB Prague, Michal Doležal and Iva Pichová, played a key supporting role. They purified the protein-based editors used to deliver those precise single-strand changes. And they made a practical breakthrough: delivering the editors directly into cells as proteins, rather than as mRNA, turned out to be compatible with normal embryonic development. Their collaboration began back in 2021.
Jeřábek is careful to frame the study as basic science, not a recipe for designer babies. "Our goal was not to develop a method for genetically modifying human embryos but rather to better understand, through basic research, how DNA repair mechanisms function in the early stages of human development." He adds that while base editing is a real improvement in precision, many safety questions must still be answered before these tools could ever be used in human embryos in a clinic. The research ran in Dieter Egli's laboratory under the oversight of Columbia University's ethics committee — and even before publication, it was already attracting international attention. The next step is clear: figure out how to make double-strand repair as reliable as the small-nick version.