(724b) Applying Topological Metabolic Analysis to Industrially Relevant CHO Cell Cultures

Despite extensive efforts to model mammalian cell metabolism in academic research labs, dating back over 30 years and similar levels of bioprocess analysis in industrial research laboratories, much about mammalian cell metabolism in industrial biprocesses remains poorly understood. Recent advances in metabolomics in conjunction with the sequencing of the CHO genome have generated the ability to create and analyze genome-scale models of CHO cell metabolism. However, this level of analysis is impractical for analysis of day-to-day bioprocesses or even for selection of clones that are compatible with platform technologies.

To address this modeling and analysis "disconnect", in this academic-industrial collaboration, we applied an existing, large-scale metabolic model^1,2 to analyze bioprocess data with the following objectives:

• To identify  the minimum amount of cellular metabolite data that needs to be collected to sufficiently model cellular metabolism (e.g. amino acids, TCA cycle intermediates, etc.).

•  To determine whether metabolite data from supernatant samples is sufficient for this purpose or if cell pellet samples would need to be analyzed as well.

• To develop and implement a model that gives a reasonable overview of the cellular state, in terms of specific metabolic signatures, which can be used to identify clones that do or do not fit in with the platform process and/or to determine when a cellular state will adversely affect product quality attributes.

• To determine whether these cellular states can be used to identify changes in cell culture performance due to changes in inputs, e.g., raw materials, which may affect protein quality attributes.

In this presentation, we discuss our success in this effort and provide guidance for the use of large-scale models for analysis of industrial bioprocesses. In addition, we compare our approach with other multivariate analysis approaches to identify optimal approaches for applying metabolic analysis to industrial bioprocesses.

(1) A.C. Baughman, S.T. Sharfstein, and L.L. Martin, A Flexible State-space Approach for the Modeling of Metabolic Networks I: Development of Mathematical Methods, Metabolic Engineering, 13: 125-137 (2011). doi:10.1016/j.ymben.2010.12.002

(2) A.C. Baughman, S.T. Sharfstein, and L.L. Martin, A Flexible State-space Approach for the Modeling of Metabolic Networks II: Advanced Interrogation of Hybridoma Metabolism, Metabolic Engineering, 13: 138-149 (2011). doi:10.1016/j.ymben.2010.12.003