Regression Models of Viability and Apoptosis Can be Generated Leveraging Cluster Analysis From Ovizio iLineF PRO Image Data.

Chinese hamster ovary (CHO) cells serve as the backbone of modern large-scale manufacturing of monoclonal antibodies. Central to this process are fed-batch cultures, where cells are grown from low to high cell densities and go through a mAb production phase. Despite CHO cells' widespread usage and vital role in the production of biologics, the cellular states during fed-batch coinciding with high specific productivity and apoptosis are poorly understood. Crucial to this understanding is a clear depiction of the various subpopulations of cells that exist in a fed-batch culture over time. In this work, an Ovizio iLine F PRO was used to image the cells in benchtop bioreactors and gather morphological and optical information for several CHO cell lines. A cluster analysis was applied to the Ovizio iLine F PRO raw data, revealing a diverse set of cellular sub-populations within each culture. Raw data from the Ovizio were then transformed into tank-level data and applied to offline measurements for viability and apoptosis in regression analysis. We used the results from the cluster analysis and feature averaging to set up regressions for offline measurements. Our regression analysis illustrates the power of in-line imaging of CHO cells as a rich process analytical tool.