Kinetic Modeling of Antibody Secretion in IgG1-Producing CHO Cells Using Intracellular Antibody and Intermediate Dynamics During Translation Inhibition

Chinese hamster ovary (CHO) cells are widely used host cells for therapeutic antibody production. Identifying and alleviating bottlenecks in the antibody secretion process in CHO cells is important for improving antibody productivity. However, secretion bottlenecks are often antibody- or cell line-specific, and their identification and validation based solely on experimental approaches can be time-consuming and costly. Computational approaches combined with experimental data could facilitate the identification of secretion bottlenecks. In this study, a kinetic model of the antibody secretion process was developed, and its parameters were optimized using Particle Swarm Optimization (PSO) to describe intracellular and extracellular concentration changes of antibodies and intermediates quantified during translation inhibition in CHO cell lines producing trastuzumab, rituximab, or bevacizumab. The proposed model successfully reproduced the concentration changes of antibodies and assembly intermediates in all three cell lines. Parameter optimization suggested the presence of previously unobserved intracellular molecular species, such as HC2 and HC2LC. The estimated initial flux of fully assembled IgG secretion into the extracellular space was 1.68–1.99 times higher in CHO cells producing trastuzumab and rituximab than in those producing bevacizumab, based on the interquartile spread of parameter estimates across accepted PSO runs. Furthermore, the reaction rate constants for this secretion step were highest in CHO cells producing rituximab, with values 1.80–3.95 times higher than those in cells producing trastuzumab and bevacizumab. These results suggest a relatively high extracellular secretion capacity in CHO cells producing rituximab, consistent with the experimental results. The optimized parameters enabled the evaluation of specific secretion reactions and suggested potential secretion bottlenecks. Although further improvement in parameter identifiability and model refinement is required, the results of this study improve the systematic understanding of the antibody secretion process and may facilitate the rational quantitative identification of secretion bottlenecks and the development of CHO cell lines with improved antibody productivity.