Base editing hit every PCSK9 allele in human embryos

A team led from Columbia University, with Stepan Jerabek as lead author, published results in Nature on 9 September 2026 from editing the PCSK9 gene in human embryos. The editor was ABE8e-V106W, an adenine base editor, delivered as protein at fertilisation rather than as mRNA. It achieved editing at all PCSK9 alleles, no insertions or deletions were detected, embryos developed to the blastocyst stage, and homozygous edited stem cell lines were derived from them.

The failure modes are reported with equal precision. Rare on-target chromosome breakage and chromosomal abnormalities occurred. Editing at bystander and off-target sites was mosaic, meaning some cells carried changes others did not. Delivering the editor as mRNA instead caused frequent embryo arrest, which the authors attribute to guide-independent deaminase activity — the enzyme acting without being aimed.

The conclusion is the part that will be quoted out of context. Base editor lesions do repair more efficiently than the double-strand breaks Cas9 leaves. The authors nonetheless state that undesirable consequences for the genome and for development can occur, and that this currently precludes clinical use in reproduction. PCSK9 was chosen because knocking it out lowers cholesterol for life; the paper is about whether the method is controllable, and its answer is not yet.