To investigate the therapeutic potential of base editing, the authors delivered the base editor and targeting sgRNA using a lentivirus followed by puromycin selection. Treatment resulted in 87% and 91% correction of the disease-associated mutation in patient-derived primary fibroblast cells 20 days later. Progerin mRNA and protein levels were reduced by 4.4-8.1-fold and 6.1-15-fold, respectively, and nuclear morphology improved. No off-target DNA or RNA editing could be detected.
The authors then used two trans-splicing adeno-associated viral vectors (AAV9) to deliver the base editor and LMNA c.1824-targeting sgRNA to homozygous mice containing the human LMNA c.1824 C>T allele. Injection of 5 × 10^11 vg each at P14, achieved 10-60% genomic DNA correction in a variety of tissues six weeks later and these levels only increased 6 months after treatment. Treated mice showed normal vascular pathology and life spans, the latter a 137% increase over control mice.
These results compare extremely favourably with the 25% increase in median survival seen with CRISPR–Cas9 gene editing and with recently published results using peptide-conjugated phosphorodiamidate morpholino oligomers (61.6% increase) or phosphorothioate/MOE-modified splice switching oligonucleotides (33% increase).
However, all three tested AAV-injected mice showed evidence of rare AAV integrations in known AAV integration sites. Five of nine deceased treated mice had liver tumours which are likely associated with the use of AAV or with the two copies of human RAB25 contained in the transgene allele.