309. Optimization of Production of AAVrh.10 Viral Vectors

Abstract

With the increasing interest in the use of adeno-associated virus (AAV) vectors to treat human disorders, there is increasing focus on optimizing AAV production to improve yield, purity and potency. The production of AAV vectors can be conceptualized in two stages: upstream cell culture and downstream purification and formulation. As an example of how AAV production can be optimized, we assessed upstream (cell density at transfection, plasmid and transfection reagent concentration, cell culture nutrients and harvest timing) and downstream (virus recovery, purification methods, potency) steps in the production of AAVrh. 10, a nonhuman primate vector, using adherent 293T cells and a 2 plasmid transfection system using polyethylenimine (PEI). Upstream, vector yield, as measured by quantitative PCR in the cell harvest from transfections at each of several different cell densities peaked at 8 × 104 cell/cm2. Total vector yield as a function of time after transfection peaked at day 3 post-transfection with diminishing returns and greater partitioning into the cell supernatant at longer times. Higher vector yield correlated with a greater total quantity of plasmid and PEI, but changing the PEI:plasmid ratio from 1:1 diminished productivity. Fetal bovine serum concentration had minimal effect. Downstream, freeze-thaw methods were not reproducible. Physically lysing the cells was not as effective for AAV recovery as the detergents, triton or tween, which were equally effective, with the lowest effective detergent concentration important to minimize the residual detergent in the final product. While iodixanol gradients are the traditional AAV purification systems, depth filtration followed by tangential flow filtration worked well for small scale production (>90%), but for large scale was less efficient (>70%). Affinity chromatography using AVB (GE Healthcare Life Sciences), provided an effective purification with maximum vector loads of up to 3 × 1012 vector genomes per ml of packed column. Finally, concentration of the final bulk AAV product was critically dependent on a formulation that stabilizes the solubility of the AAV monomer, with 0.01% pluronic acid providing AAV recovery of nearly 80%. Together, the optimized methods for the best yield, purity and potency relied on transfection with equal amounts of plasmid and PEI and a cell density of 8 × 104 cells/cm2 combined with the use of detergent to disrupt and recover AAV and sequential purification by depth filtration, tangential flow filtration and affinity chromatography. This process improved the yield from harvest to final product from 9% to 40% and achieved final purified AAV at 7 × 109 vector genomes/cm2 of adherent cell culture. These systematic refinements to the individual steps of AAV production demonstrate improved yield and facilitates transfer of a method for GMP production for clinical use.