In our previous work (Xie and Wang, 1994a), a simplified stoichiometric model on energy metabolism for animal cell cultivation was developed. Fed-batch experiments were performed in T-flasks using this model in supplemental medium design (Xie and Wang, 1994b). In this work, the major pathways of glucose and glutamine metabolism were incorporated into the stoichiometric model. Fed-batch culture was conducted in a 2-liter bioreactor with appropriate process control strategies. Nutrient concentrations, especially glucose and glutamine, were maintained at constant but low levels through the automated feeding of a supplemental medium formulated using the improved stoichiometric model. The formation of toxic byproducts, such as ammonia and lactate (Hassell et al., 1991), was greatly reduced. The specific lactate production rate was decreased by 62-fold compared with batch culture in bioreactor and by 8-fold compared to fed-batch culture in T-flask using the previous stoichiometric model. Ammonia formation was also decreased compared with both the batch and fed-batch cultures. Most importantly, the monoclonal antibody concentration reached 900 mg l-1, an increase of 17- and 1.6-fold compared with the batch and fed-batch cultures respectively.

A metabolic reaction network is developed for the estimation of the stoichiometric production of adenosine triphosphate (ATP) in animal cell culture. By using the material balance data from fed‐batch and batch cultures of hybridoma cells, the stoichiometric ATP productions are determined with estimated effective P/O ratios of 2 for NADH and 1.2 for FADH2. A significant percentage of the ATP requirement (16–41%) in hybridoma cells is generated directly from free energy release without the participation of oxygen. The oxidative phosphorylation of NADH accounts for about 60% of the total ATP production in the fed‐batch cultures and about 47% in the batch culture. The oxidative phosphorylation of FADH2 accounts for less then 20% of the total ATP production in all cases. A fractional model is devised to analyze the contribution of each nutrient to the ATP production. Results show that a majority of the ATP is produced from glucose metabolism (60–76%). Less than 30% of the ATP is derived from glutamine, and less than 11% is derived from other essential amino acids. The analysis also shows that the glycolytic pathway generates more ATP in the batch (41%) than in the fed‐batch (<27%) cultures. The TCA cycle provides 51–68% of the total ATP production. The calculated stoichiometric oxygen consumption differs among the batch and fed‐batch cultures, depending on the glucose concentration. This result suggests that the relationship between the oxygen uptake rate (OUR) and cell growth may change with the culture conditions. However, the calculated respiratory quotient (RQ) is relatively constant in all cases. A linear relationship is obtained between the specific ATP production rate and the specific cell growth rate. The maximum ATP yield and the maintenance ATP requirement are determined based on this linear relationship. The biosynthetic ATP demand estimated from the dry cell weight and cell composition is significantly lower than that calculated from the maximum ATP yield, indicating that the non‐growth‐associated ATP demand may contain other factors than what is considered in the estimation of the biosynthetic ATP demand. © 1996 John Wiley & Sons, Inc.