Scale-Down Perfusion Process for Recombinant Protein Expression

Abstract

The complexity and expense associated with large-scale perfusion ­culture systems make them impractical for medium optimization studies involving DOE approaches. We have employed scale-down processes involving shake flask cultures seeded at high density, or at several different densities with daily replacement of the culture media (maintaining a constant cell density). These model systems were used to test effects of media components on growth of recombinant CHO cells expressing a therapeutic protein with the ultimate goal being to apply the media modifications to a large-scale perfusion system. Basal media employed were protein-free, chemically defined formulations that were supplemented with various combinations of additives to test effects on cell growth and recombinant protein expression. Additives tested included amino acids, vitamins, lipids, growth factors, nucleic acid precursors, iron chelating substances, and trace elements. In addition, the effects of osmolality and ammonium concentration were examined. Results suggest that perfusion systems are difficult to model in shake flasks. Several media components showed positive effects in standard batch shake flask culture, but had no effect in simulated perfusion shake flask cultures. Several media components showed positive effects when tested in simulated perfusion shake flask cultures, but had no effect in the larger perfusion system. One possible explanation for the discrepancy in results between the standard batch shake flask cultures, simulated perfusion shake flask cultures and large-scale perfusion system is the accumulation of metabolites and resulting pH shift downward in the shake flask cultures. Batch shake flask cultures have no pH regulation; simulated perfusion shake flask cultures have only limited pH control (when the medium is replaced once each day) while large-scale perfusion cultures have real-time pH control.