Every PCSK9 allele was edited; the bystander sites stayed mosaic

Editing at every allele

Delivering ABE8e-V106W as a protein at fertilisation achieved editing at all PCSK9 alleles in human embryos. Development continued to the blastocyst stage, and homozygous edited stem cell lines were derived. Two things are being claimed at once: that the edit reached every copy of the target, and that the embryo carrying it kept growing.^[1]

The mechanism the authors point to is repair. Cas9-induced double-strand breaks in human embryos are genotoxic and cause frequent aneuploidy and large deletions, while the nicks and mismatches a base editor leaves behind are efficiently repaired. That claim is about which lesion the early embryo can handle; it carries no verdict on which tool is better in general. Another reading is that the difference sits in the delivery and timing used here rather than in the type of lesion, since the editor arrived as protein at fertilisation instead of at a later stage.^[1]

The parts that stayed messy

No insertions or deletions were detected. Rare on-target chromosome breakage and chromosomal abnormalities still occurred, and that second finding limits how far the first one travels. A clean insertion and deletion count describes what the editor did to the sequence; chromosome breakage describes what happened to the chromosome carrying it.^[1]

Editing at bystander and off-target sites was mosaic. Mosaicism means the cells of one embryo do not carry the same genome, so a result read from a few cells does not describe the whole. In a research setting that is a limitation to measure and report; in a setting where the embryo is meant to become a person, it is the difference between knowing what was edited and hoping.^[1]

The distance to a clinic

The delivery format is part of the result. When the editor was introduced as mRNA, guide-independent deaminase activity caused frequent embryo arrest. The same molecule, handed to the embryo in a different form, stopped development instead of surviving it, which makes protein delivery at fertilisation a condition the finding depends on.^[1]

The authors' own conclusion is the sentence to keep. Undesirable consequences for the genome and for development can occur, and those consequences currently preclude clinical use in reproduction. Read as medicine, the useful sentence in this paper is the negative one: a base editor can reach every copy of PCSK9 in a human embryo, and that capability alone does not make the embryo safe to transfer.^[1]