Base Editing in Human Embryos Achieves Precision but Produces Mosaicism in Majority of Cases

Scientists injected base editors into zygotes and tracked editing outcomes through blastocyst stage, focusing on genes essential for early lineage specification. The approach avoided double-strand breaks that previously triggered large deletions and p53 activation. However, editing efficiency varied across cells within individual embryos, leaving 60-80% of samples as mosaics containing both edited and wild-type lineages. This pattern mirrors earlier CRISPR reports but occurs despite the single-base mechanism.

Mosaicism raises distinct regulatory stakes beyond somatic editing. Germline interventions could transmit mixed genotypes across generations, complicating consent frameworks and long-term surveillance requirements already debated after the 2018 He Jiankui case. Existing US and EU moratoria on heritable editing rest partly on safety thresholds; persistent mosaicism suggests those thresholds remain unmet even with refined tools.

Related observational data from UK and Chinese IVF cohorts show natural mosaicism rates of 20-30% in preimplantation embryos, indicating base editing may amplify rather than introduce the problem. Policy discussions at the 2025 ISSCR and NASEM meetings have begun linking technical mosaicism metrics directly to possible conditional approval pathways, a connection rarely foregrounded in coverage focused only on technical feasibility.

Next steps require systematic comparison of base, prime, and newer epigenome editors in larger embryo cohorts with single-cell resolution, followed by multi-generational animal studies before any regulatory body considers lifting current prohibitions.