CHO cell line specific prediction and control of recombinant monoclonal antibody N‐glycosylation

ABSTRACT Here we demonstrate that it is possible to predict and control N ‐glycan processing of a secreted recombinant monoclonal antibody during manufacturing process development using a combination of statistical modelling and comparative measurement of cell surface glycans using fluorescent lectins. Using design of experiments—response surface modelling (DoE‐RSM) methodology to adjust the relative media concentrations of known metabolic effectors of galactosylation (manganese, galactose, and uridine) we have shown that β1,4‐galactosylation of the same recombinant IgG 4 monoclonal antibody produced by different CHO cell lines can be precisely controlled in a cell line specific manner. For two cell lines, monoclonal antibody galactosylation could be increased by over 100% compared to control, non‐supplemented cultures without a reduction in product titre and with minimal effect on cell growth. Analysis of galactosylation effector interactions by DoE‐RSM indicated that Mn 2+ alone was necessary but not sufficient to improve galactosylation, and that synergistic combinations of Gal and Urd were necessary to maximize galactosylation, whilst minimizing the deleterious effect of Urd on cell growth. To facilitate rapid cell culture process development we also tested the hypothesis that substrate‐level control of cellular galactosylation would similarly affect both cell surface and secreted monoclonal antibody glycans, enabling facile indirect prediction of product glycan processing. To support this hypothesis, comparative quantitation of CHO cell surface β1,4‐galactosylation by flow cytometry using fluorescent derivatives of RCA and ConA lectins revealed that substrate‐controlled variation in monoclonal antibody galactosylation and cell surface galactosylation were significantly correlated. Taken together, these data show that precision control of a complex, dynamic cellular process essential for the definition of protein product molecular heterogeneity and bioactivity is possible. Moreover, real‐time, or near real‐time control can be enabled by facile, rapid measurement of cell surface biomarkers of cellular biosynthetic capability. Biotechnol. Bioeng. 2013;110: 2970–2983. © 2013 Wiley Periodicals, Inc.