On-line control of an immobilized hybridoma culture with multi-channel flow injection analysis.

An immobilized hybridoma cell line was cultivated at controlled glucose and glutamine concentrations. On-line analysis of the substrates was carried out with a multi-channel flow injection analysis system. The analysis system also determined on-line the lactate and ammonium concentration. The substrate concentrations were controlled using an adaptive-control strategy. This strategy consisted of the estimation of the real-time concentrations and volumetric substrate consumption rates by an Extended Kalman Filter, and a minimum variance controller, which used the estimated parameters to set the feed rates of the substrates. The closed-loop control was used to start-up two cultures with either glucose or glutamine as control-substrate for the medium feed rate. The controller kept the concentration of the control-substrate constant by enhancing the medium feed rate simultaneously to the increasing volumetric consumption rate of the substrate. When glutamine was used as control-substrate, the glucose concentration remained relatively constant, whereas the glutamine concentration decreased during the start-up at a constant glucose concentration. This indicates that glutamine is consumed faster than glucose and will be a better control-substrate to avoid limitation during the start-up of a culture with the applied hybridoma cell line. During the colonization of the microcarriers, the yield of ammonium on glutamine decreased from 0.80 to 0.55 (mol mol-1), indicating a change in the glutamine metabolism. The yield of lactate on glucose stayed constant for both experiments. During long-term culture of more than 800 h, the controller kept both the glucose and glutamine concentrations constant at perfusion rates between 0.50 h-1 and 0.15 h-1. The medium, glucose and glutamine feed rate were independently controlled. Both the specific glutamine and glucose consumption rates remained constant for all perfusion rates, which was probably as a result of the constant concentrations. The specific monoclonal antibody production rate decreased with the perfusion rate decreasing from 0.40 h-1 to 0.20 h-1. The immobilized-cell concentration decreased only at the lowest perfusion rate. Both effects could not be explained directly by the increasing ammonium and lactate concentrations nor by the decreasing amino-acid concentrations.