Efficient transgenic mouse modeling through CRISPR-RNP gene editing delivered via virus-like particles

초록

The production of diverse/accurate transgenic models for therapeutic or basic research into various diseases is crucial. While CRISPR technology has enabled precise gene editing, many researchers have delivered them to zygotes as plasmids, mRNA or RNPs using traditional methods (e.g. microinjection, electroporation, GONAD) to create mutant mice. However, these technologies require specialized skills and expensive equipment, limiting access to them for many labs. Moreover, they can cause physical damage to embryos and interfere with normal development. We present a novel, non-viral approach using virus-like particles (VLPs) to address these issues and enable more research facilities to create the exact animal models they need. VLP can carry larger gene-editing cargo, do not contain viral DNA, and pose no risk of viral genome integration. Additionally, packaging CRISPR as RNPs in VLPs minimizes off-target problems. We produced VLPs containing CRISPR-RNPs and then cocultured them with fertilized zygotes or during IVF to introduce mutations into embryos without causing physical damage. VLP-Cas9 showed 44.7% editing efficiency and had a heterozygous genotype for frameshift-induced knockout by deletion of 29-nucleotide in Plin1. The mutation was transmitted to the next generation and mice had smaller adipocytes. Editing by VLP-BEs were also successful and we could observe increases in VLP productivity and up to 6.5-fold higher editing, especially after codon-optimization of a part of gag. We were also demonstrated multi-target editing with different CRISPR systems simultaneously. For knock-in, we co-cultured AAVs containing donor DNA and VLPs with zygotes, and succeeded in replacing exon5 of mouse Kcnq4 and part of intron with human sequence. VLP-treatment during the IVF procedure, we obtained hetero Tyr-mutant mice with H420R (A-to-G) without off-target effects. This method further simplifies and accelerates transgenic model generation without requiring special techniques or equipment and can be widely applied for customized mouse models in diverse research fields.