Multivariate Process Optimization for Fixed-Bed Bioreactor-Based AAV Production Improves Total Batch Yield

Abstract

Adeno-associated viral vectors (AAVs) are promising tools for gene therapy. However, scaling up the production of AAVs to produce high-quality vectors at high yields for clinical purposes has proven to be challenging. In the present study, we optimized the production process of AAV in a fixed-bed bioreactor using transient transfection in adherent HEK-293T cells. We systematically optimized the key process parameters, namely cell seeding density, cell density at transfection, and DNA-to-cell ratio, based on the yield obtained, starting from a prototype batch, followed by ten batch runs. Here, we packaged a reporter gene (enhanced green fluorescent protein) and a therapeutic gene (lysyl oxidase) into AAV9 capsids as part of our process development program to be applied for future current Good Manufacturing Practices production and clinical trial application. Throughout the experiments, media conditions, transfection processes, and mechanical parameters were kept identical, while monitoring pH, dissolved oxygen, and media glucose concentration during a production process of approximately 10 days. We demonstrate that by optimizing these parameters, the fixed-bed bioreactor was able to support as many as 1.6–2.8 × 10⁶ cells/carrier strip, up to 3 × 10⁹ cells/m² bioreactor. Through this multivariate optimization process, we increased viral yield by about 7.6-fold (range of 5.7–10.4-fold for the optimized process runs) over the prototype batch. The total AAV vector yield average was 2.3 × 10¹⁴ vg (range 1.1 × 10¹⁴ vg to 4.95 × 10¹⁴ vg), corresponding to an average per cell yield of 1.4 × 10⁵ vg/cell (range 0.85 × 10⁵–2.46 × 10⁵vg/cell). In conclusion, our findings highlight that optimizing process parameters in a fixed-bed bioreactor presents a promising strategy for scalable and cost-effective AAV vector production.