Quantifying the partition of demoresources between cellular and recombinant protein glycosylation in GS-CHO cells

Conference Dates

May 6-11, 2018


All commercially-available therapeutic monoclonal antibodies (mAbs) contain a consensus N-linked glycosylation site on the constant fragment (Fc) of their heavy chains. The composition of the carbohydrates (glycans) bound to these products determines their safety and therapeutic efficacy [1]. While production cell lines synthesize, glycosylate, and secrete the mAb, they also glycosylate their own components. Therefore, there is a partition of glycan biosynthetic precursors – nucleotide sugars (NSs) – that directly couples mAb glycosylation with cellular growth and metabolism. With this work, we present a novel metabolic flux analysis (MFA) model that establishes a mechanistic and quantitative representation of the partition of metabolic resources between cellular and mAb glycosylation.

Experimentally, IgG-producing GS-CHO cells were cultured with three amino acid feeding strategies. Data for cell density, nutrient availability, metabolite accumulation, product titer, intracellular NS concentration [2], and mAb glycoprofiles [3] were collected. Computationally, a metabolic flux analysis (MFA) model that represents 101 metabolites connected by 143 reactions has been developed. In addition to central carbon, amino acid, nucleic acid and lipid metabolism, the MFA also includes the aspartate-malate shuttle, the urea cycle and detailed balances for ATP as well as the NAD(P)+/NAD(P)H redox pair. Demand of cellular resources towards cellular glycosylation has been represented by including stoichiometric coefficients for O-GalNAc, N-linked, glycosphingolipid and GPI anchor glycans in the biomass equation [4]. Product glycosylation has been included in the equation describing mAb composition. The underdetermined MFA (42 degrees of freedom) was constrained to account for reaction reversibility and was solved through multi-objective optimization, where the squared error between experimentally-determined and MFA-calculated fluxes is minimized and ATP synthesis per flux unit was simultaneously maximized [5].

The proposed MFA framework allows us to analyze how metabolic resources are partitioned between cellular and mAb glycosylation. Our results indicate that during exponential growth, cellular glycosylation consumes considerably higher amounts of NS biosynthetic precursors (ATP, glucose, and glutamine). As growth ceases, a larger fraction of metabolic resources is allocated to mAb glycosylation, but total NS consumption decreases. This suggests that cellular glycosylation is the larger metabolic ‘sink’ within our cell line, a result that is consistent with the intracellular accumulation of NSs observed towards latter stages of culture. With further refinement – in particular, with data for dynamic variations in cellular glycosylation – our MFA framework can serve as a computational tool to design optimal NS precursor feeding strategies that control mAb glycosylation and minimize negative impacts on cell growth and productivity.


1. Jefferis R, 2009. Nat Rev Drug Discov, 8(3):226.

2. del Val IJ, et al., 2013. Anal Biochem, 443(2):172.

3. Stockmann H, et al., 2013. Anal Chem, 85(18):8841.

4. del Val IJ, et al., 2016. Sci Rep, 6:28547.

5. Schuetz R, et al., 2007. Mol Syst Biol, 3:119.

This document is currently not available here.