(485r) A High-Throughput, Scale-Down Model for Process Optimization Using Product Quality as a Performance Metric

Quality by Design (QbD) initiatives are driving the need for more comprehensive understanding of cell culture processes. While maximizing product titer remains important in clone selection and process optimization, there is a significant effort in identifying cell lines and processes that result in better product quality. The development of such upstream processes may not only result in better downstream performance, but can also result in better safety and efficacy profiles for the final drug product. Therefore, the development of scale-down models which can accurately predict product quality profiles is of great interest to the biotechnology industry.

Currently, clone selection and process optimization are conducted across a range of vessels including well plates, shaken tubes and flasks and bench-scale bioreactors. Each vessel differs not only in its volume, but its shear profile as well as process measurement and control capabilities. As a result, culture conditions are typically varied across the various platforms. For example, simple batch processes are performed in well plates, uncontrolled batch or fed-batch processes are conducted in tubes and/or flasks, and fully controlled fed-batch processes are finally conducted in bench-scale bioreactors. With these differences in process, product titer and quality may vary significantly between the various platforms and the final manufacturing scale process.

Until now, product quality assessments were typically made during late-stage process development in bench-scale bioreactors. However, Process Analytical Technology and QbD require additional information and larger experimental designs than can be conveniently and economically conducted at the bench-scale. Therefore, a high-throughput scale-down model with the measurement and control capabilities of a bioreactor is required to produce the data sets and identify the design space required for these initiatives. We have developed a microbioreactor scale-down model with a working volume of less than 1mL. These microbioreactors are capable of performing complex fed-batch processes with pH, DO and glucose feed-back control and can accurately predict the performance of culture conditions at scales more than 1000x larger than the model.

A case study of process optimization using a CHO cell line producing a monoclonal antibody was performed in the microbioreactor platform. Optimal conditions were determined based on product titer and product quality. Select conditions from the Design of Experiments were also run in bench-scale bioreactors and the results for the two platforms were compared. A strong correlation was found between the different scales demonstrating the use of this scale-down model for process optimization and QbD applications.