Enhancing The Upstream Performance Of Adeno-Associated Virus (AAV) Vector Manufacturing Via Multivariate Data Analysis

By Apezteguia A., Keune J., and Iglesias M.

Viralgen, a leading AAV Contract Development and Manufacturing Organization (CDMO), has created an extensive proprietary dataset from over 1,000 batches produced using its Pro10™ production platform. This initiative aims to optimize the AAV upstream process by identifying critical parameters that significantly affect titer. Given the complexity of biological processes and the interplay of numerous variables, traditional univariate techniques are insufficient for capturing multi-variable relationships. To address these challenges, Viralgen applied dimensionality reduction methods, building a multivariate data model based on 89 historical batches. Analysis revealed that higher viable cell density (VCD) correlates with lower pH and glucose levels before transfection, negatively impacting productivity. Subsequent experiments confirmed the pivotal role of VCD, determining its optimal levels for maximizing titer.

To operationalize these insights, Viralgen developed an automated application that identifies key factors influencing titer in historical batches. This tool supports Manufacturing Sciences & Technology and R&D teams in continuously monitoring performance and pinpointing areas for process improvement. In the context of cell and gene therapy (CGT)—a field with transformative potential for treating genetic disorders and cancers—such data-driven approaches are critical. Viralgen leverages proprietary technologies licensed from AskBio, including the Pro10™ cell line, a high-yield, universal system capable of producing all serotypes and chimeric forms of recombinant AAV (rAAV). The diverse range of CGT products, the small quantities required for treatment, and the associated high production costs highlight the necessity of platform data for understanding and managing process variability.

Key challenges in the AAV industry include the high variability stemming from manual manufacturing steps, raw materials, product diversity, and analytical methods, as well as the complexity of interrelated process parameters such as VCD, metabolite concentration, and pH. Notably, the productivity at the transfection pool (TP) stage is critical for achieving high drug product (DP) yields, underscoring the importance of optimizing upstream processes for better outcomes.