Traceless delivery of prime editor ribonucleoprotein (RNP) complexes in engineered nanoparticles

Genome editing has enormous potential to transform the future of genetic medicines, but current delivery strategies suffer from several shortcomings. By incorporating ribonucleoprotein (RNP) complexes instead of vector genomes into lentivirus-derived nanoparticles (LVNPs), we address key limitations of viral and non-viral delivery, including: (i) overcoming limited packaging capacity, (ii) preventing unwanted genomic integration, and (iii) reducing exposure to the therapeutic payload. The practical utility of this LVNP platform has been demonstrated, directing robust levels of genome editing in both primary human T cells ex vivo and the murine retina in vivo (PMID: 37678882). We also demonstrated LVNP-mediated delivery of larger base/prime editors, however, the substitution of Cas9 with larger RNP complexes caused a decline in both potency and viral yield.

With strategic emphasis on clinical applications, successful vectors should converge in terms of potency, safety, and manufacturing. To improve potency of LVNPs loaded with prime editors, we added effector domains to favour membrane anchoring during manufacturing and removed redundant viral proteins to improve LVNP packaging capacity. In addition, we incorporated protein/RNA aptamers to increase (e)pegRNA loading and recruit cognate effector proteins. Collectively, these alterations substantially improved prime editing efficacy (~8-fold to ~34% across targeted loci). We also tested prime editor configurations that lack the RNAseH domain but observed slightly reduced efficacy. Finally, we have established protocols for large-scale and cGMP-compliant manufacturing using VSV-G or Sendai F/HN pseudotypes. Collectively, these improvements in our LVNP platform will enable preclinical testing in patient-derived cells and in vivo models to assess potency and safety.