(660g) Modeling the Effect of Bioreactor pH on Chinese Hamster Ovary Cell Metabolism and Site-Specific N-Linked Glycosylation of VRC01 Mab

Process optimization of bioprocessing requires the use of time-consuming and expensive experiments. The development of mathematical models that can provide accurate predictions of experimental conditions could aid with reducing the number of experiments performed [1]. CHO cell fed-batch processes require the addition of feed media at regular intervals to sustain the nutrient requirements of the cells. Large number of experiments are needed to optimize the concentrations of various metabolites in the media and the feeding schedule [2]. Development of a mathematical model capable of predicting dynamic concentration profiles of various metabolites can reduce the number of experiments required. In order to predict dynamic metabolic profiles of a large number of metabolites a dynamic metabolic flux analysis (DMFA) model has been developed in the literature [3]. However, this model only predicts the dynamic metabolic profiles of 13 metabolites and does not include the effect of bioreactor pH. It has been shown in the literature that bioreactor pH can significantly affect metabolism and N-linked glycosylation of CHO cells and optimization of pH shift can lead to improve process performance [4]. Mechanistic mathematical models of N-linked glycosylation have been used in the literature to study the relationship between bioreactor temperature and the N-linked glycosylation in the Fc region of a mAb [5] but similar studies have not been performed to study the effect of pH. About 15 to 20 percent of human IgG molecules have N-linked glycosylation sites on the Fc and Fab region. Glycosylation in the Fab region has also been shown to significantly affect mAb half-life, aggregation and antigen binding [6] but the effect of cell culture process parameters on the Fab region glycosylation has not been studied in the literature. The work performed in this study uses mathematical modeling to understand the effect of bioreactor pH on cellular metabolism and the site-specific N-linked glycosylation of the VRC01 mAb.

Experimental data to develop a database for model development were performed by growing VRC01 producing CHO cells in bioflo 120 bioreactor at pH values 6.75, 7 and 7.25. The uptake/production rates of nutrients were modeled as a function of concentration of the metabolites and bioreactor pH via kinetic expressions. These rates were fed to a stoichiometric model to perform metabolic flux analysis. The solution from the metabolic flux analysis was assumed to be steady over 0.25 days. The extracellular metabolite concentrations after 0.25 days was calculated by using the metabolic flux analysis solution. The dynamic metabolic profiles were obtained by iterating process until cell viability dropped below 80%. This led to development of a model to predict the effect of pH on dynamic metabolic profiles of 23 metabolites (glucose, lactate, essential amino acids, non-essential amino acids, ammonia, titer and cell density). A mechanistic model for N-linked glycosylation that approximates the Golgi as a series of stirred tank reactors was modified to include the presence of multiple glycosylation sites [7]. This model was used to study the differences between the rates of glycosylation reactions at the two different glycosylation sites, revealing increased rates of galactosylation and sialylation in the Fab region while compared to the Fc region. The model indicated that the increased activity of galactosyltransferase and sialyltransferase resulted in the experimentally observed increase in galactosylated and sialylated fractions in the Fab region while compared to the Fc region. The model also indicated that decreased activity of fucosyltransferase at higher pH values resulted in the experimentally observed decrease in fucosylation in the Fab region but the decrease was not large enough to cause a change in Fc region fucosylation and Fc region was fully fucosylated in all pH conditions. This study led to the development of a mathematical model for metabolism of CHO cells that can predict the dynamic metabolic profiles of many nutrients and the development of a mathematical model to study the effect of pH on site specific N-linked glycosylation of the VRC01 mAb.

1) Sommeregger, W., et al., Quality by control: Towards model predictive control of mammalian cell culture bioprocesses. Biotechnol J, 2017. 12(7).

2) Ritacco, F.V., Y. Wu, and A. Khetan, Cell culture media for recombinant protein expression in Chinese hamster ovary (CHO) cells: History, key components, and optimization strategies. Biotechnology Progress, 2018. 34(6): p. 1407-1426.

3) Nolan, R.P. and K. Lee, Dynamic model of CHO cell metabolism. Metabolic Engineering, 2011. 13(1): p. 108-124.

4) Hogiri, T., et al., Optimization of a pH-shift control strategy for producing monoclonal antibodies in Chinese hamster ovary cell cultures using a pH-dependent dynamic model. J Biosci Bioeng, 2018. 125(2): p. 245-250.

5) Sou, S.N., et al., Model-based investigation of intracellular processes determining antibody Fc-glycosylation under mild hypothermia. Biotechnol Bioeng, 2017. 114(7): p. 1570-1582.

6) Van De Bovenkamp, F.S., et al., The Emerging Importance of IgG Fab Glycosylation in Immunity. The Journal of Immunology, 2016. 196(4): p. 1435-1441.

7) Krambeck, F.J., et al., Model-based analysis of N-glycosylation in Chinese hamster ovary cells. PLOS ONE, 2017. 12(5): p. e0175376.