High-Throughput Screening For High-Performance Cell Culture Media Development

By Shan Gao, Matthew Tenorio, Hamid Soleymani, Xiaofei Wang, Yosuke Kurokawa, Ken Naruse, Omid Taghavian, David T. Ho, and Chandana Sharma, department of research and development, FUJIFILM Biosciences

Developing an optimized, chemically defined cell culture medium is crucial yet challenging, often involving over 40 components to achieve high yield and quality for biologics like Adeno-associated virus serotype 2 (AAV2). A high-throughput screening method offers a streamlined, cost-effective solution with a fast turnaround.

This process involves establishing a library of diverse media formulations, such as the 279 formulations generated for HEK293 cells. By screening these formulations for AAV2 production using transient transfection in 96-deep well plates assisted by automated liquid handlers, a rich dataset is generated. This data is used to build Design of Experiments (DoE) and Machine Learning (ML) models for cell growth and viral genome (VG) titer prediction, which are subsequently validated at different scales.

This approach successfully demonstrated a greater than 3-fold improvement in VG titer compared to the control medium. Furthermore, the resulting dataset can be leveraged by ML for further component optimization. Explore the full poster to see the detailed workflow and validation data.