769d Monitoring and Control of Cell Culture Bioreactors Using on-Line Scanning Dielectric Spectroscopy

Cary Opel and Ashraf Amanullah. Oceanside Process R&D, Genentech, Inc., 1 Antibody Way, Oceanside, CA 92056

Achieving reliable on-line cell density measurements is one of the most valuable, but difficult goals in process monitoring. Capacitance probes offer one solution to this problem by measuring the enclosed viable biovolume of a culture. One limitation to this method is that the relationship between biovolume and cell density changes during batch and fed-batch cultures, making it difficult to predict cell density from capacitance readings. Recent advances in scanning dielectric spectroscopy show promise alleviating this limitation by providing information on the size and size distribution of a culture. With this knowledge, one can theoretically calculate the number of cells in a given volume. Capacitance probes improve the online characterization of a cell culture process, contributing to the Process Analytical Technology Initiative

Batch and fed-batch cultures using CHO cells were performed and analyzed using a Fogale Biomass System. Offline samples were analyzed using a Nova Biomedical Flex, a Beckman-Coulter Z2 Cell Analyzer and a Beckman-Coulter ViCell. Packed cell volumes were also measured. Several methods of analyzing the set of capacitance data were used to derive useful culture parameters such as viable cell density, viable packed cell volume, integrated viable cell density, growth rates, average cell diameter and cell size distribution. First, a simple model was created using a capacitance-to-parameter ratio that was corrected each offline sample. Second, Partial Least Squares (PLS) regression was used to create a prediction model based on a finite training set of batch and fed-batch cultures. Finally, theoretical calculations based on the Beta-Dispersion curve were used to predict the parameters of interest. Growth rates from these methods were calculated then used to test the accuracy of viable cell density predictions.

Finally, some practical applications of capacitance probes were demonstrated through the integration of the signal with a DeltaV bioreactor control system. This includes an automated batch re-feed for the maintenance of inoculum cells, monitoring of base addition via conductivity (an additional measurement provided by the capacitance probe), exact target densities for inoculum trains, online indication of glucose and lactate exhaustion, and the implementation of a capacitance controlled continuous bioreactor.