Recombinant vectors based on adeno-associated virus (rAAV) are showing immense therapeutic promise in human trials, resulting in greater demand for clinical-grade rAAV vector production capacity.
rAAVs are traditionally produced via a transient transfection of three plasmids encoding:
1) a transgene flanked by inverted terminal repeats,
2) adenoviral helper proteins, and
3) two AAV genes, rep and cap,
into HEK293 cells. We hypothesised that the equimolar ratio of rep-to-cap genes, while evolutionarily optimised for packaging of the wild type virus, may be suboptimal for packaging rAAV vector. By separating rep/cap onto individual plasmids we were able to test the impact of different rep-to-cap ratios on the yield of natural and bioengineered AAV variants.
As the current AAV packaging model indicates that Rep proteins are involved in loading replicated AAV genomes into preassembled AAV particles, we hypothesised that increasing the amount of rep gene would enhance rAAV packaging. Additionally, since current vector production methods produce more empty than full capsids when quantified prior to purification, we hypothesised that the cap gene was provided in excess. In contrast, the data indicates that increasing the amount of cap was associated with improved packaging whilst increasing rep actually resulted in a reduction in packaging efficiency. The effect was more significant for variants that are known to be difficult to package, such as AAV2 and LK03, for which up to 3-fold improvements were achieved. Interestingly, vectors that were efficient at packaging using the canonical rep/cap plasmid such as AAV8 were found to be already optimal at a 1:1 rep-to-cap ratio.
The observed improvements in vector yields using our novel packaging system have the potential to revolutionise not only the lab-scale preclinical AAV manufacturing, but also large-scale clinical manufacturing, enhancing preclinical research and making AAV-based therapies more affordable for patients suffering from genetic disorders.