Unreviewed Preprint Reports Efficient Base Editing of Human Embryos Without Chromosomal Damage

A preprint posted to bioRxiv on June 1 by a team led by Dieter Egli, a developmental cell biologist at Columbia University, describes what the authors call efficient base editing of human embryos at two genomic targets - and claims to do so without the chromosomal damage that has long plagued standard CRISPR-Cas9 approaches in early embryonic cells.

The distinction matters mechanistically. According to the bioRxiv preprint, DNA double-strand breaks induced by CRISPR/Cas9 result in frequent aneuploidy and large deletions in early human embryos, a repair deficiency that has limited clinical translation of that technology. Base editing sidesteps that problem by swapping individual nucleotide letters rather than cutting both DNA strands. The team used adenine base editors delivered as ribonucleoprotein complexes to introduce targeted A-to-G edits at two loci: PCSK9, a gene involved in cholesterol regulation, and HBG1/HBG2, fetal hemoglobin genes relevant to sickle cell disease and beta-thalassemia. The preprint reports that editing was efficient and, unlike Cas9-induced double-strand breaks, did not result in chromosomal abnormalities or large deletions.

That is the headline result. But the methods section tells a more complicated story. As reported in Scientific American, most of the edited embryos showed mosaicism - meaning the genome varied from cell to cell within the same embryo, a finding that would complicate any eventual therapeutic use. The embryos also survived only when the editing machinery was delivered as a protein; when researchers used an RNA delivery vehicle instead, embryo development halted. The authors themselves acknowledge in the preprint that "translation to a clinical context remains premature."

The study has not yet been peer reviewed, a fact that reporting in Nature News noted prominently. That caveat is not cosmetic. Embryo editing research has a documented history of results that do not replicate or that reveal confounds missed in preprint form - the most infamous example being He Jiankui's 2018 claims, which collapsed under scrutiny. The current work is from a credentialed academic group, but the peer-review pipeline exists precisely to pressure-test efficiency numbers, off-target rates, and the statistical basis for claiming the absence of large deletions.

The funding chain adds another variable worth naming. According to reporting by NewsNation, the research is partially funded by Nucleus Genomics, a company that screens IVF embryos for genetic disorders and also analyzes genes linked to traits such as height and intelligence. A co-author listed on the preprint is now employed by Nucleus Genomics, per a clarification added to Nature's coverage on June 8. The genes targeted in this study were chosen because they were well-studied, not for potential therapeutic purposes - the investigators said so in the preprint - but the funder's commercial profile has drawn scrutiny from bioethicists.

The broader regulatory backdrop sharpens that scrutiny. As Chemical and Engineering News reported, the American Society of Gene and Cell Therapy, the International Society of Cell and Gene Therapy, and the Alliance for Regenerative Medicine jointly issued a moratorium in May 2025 proposing a 10-year ban on heritable human genome editing, which includes edits made in embryos. UC Berkeley geneticist Fyodor Urnov, quoted in media coverage, questioned the premise of embryo editing when standard preimplantation genetic testing via IVF has already been used safely for decades.

None of those critiques resolve the underlying biology question, which this preprint advances incrementally but does not close. The key outstanding issues are mosaicism rate, long-term developmental outcomes beyond the blastocyst stage, and off-target editing frequency as a function of guide RNA design - all areas where peer review and independent replication are the only reliable arbiters. The authors derived edited stem cell lines from some blastocysts, which could enable downstream mechanistic work, but that work has not yet appeared.

The preprint is citable as a data contribution. It is not yet evidence that base editing of human embryos is safe or effective in any clinically meaningful sense.

Sources cited:
- bioRxiv preprint (Jerabek et al., DOI: 10.64898/2026.05.30.728989) (https://www.biorxiv.org/content/10.64898/2026.05.30.728989v1)
- Chemical and Engineering News (https://cen.acs.org/biological-chemistry/gene-editing/base-editing-human-embryos/104/web/2026/06)
- Scientific American (https://www.scientificamerican.com/article/report-of-gene-edited-human-embryos-sparks-worries-about-the-technologys-future-uses/)
- Nature News (https://www.nature.com/articles/d41586-026-01827-8)
- NewsNation (https://www.newsnationnow.com/science/scientists-human-embryo-dna-gene-editing/)