Conference Dates

May 6-11, 2018


Many cGMP cell culture processes are now based on a continuous high cell density perfusion bioreactor system. Control of the feed or addition rates to maintain pseudo-steady-state conditions in these bioreactors can be especially challenging due to high and fluctuating cell concentrations that can rapidly change environmental conditions. With infrequent manual daily sampling based on offline cell counting, the control system can have too little information on which to base an appropriate decision to manipulate the process. Tight control of the perfusion or concentrate addition rate allows the bioreactor to be operated under the optimum conditions for maximum recombinant protein production. The perfusion is typically started on day 2-3 after inoculation when the cells are still in an exponential batch growth phase and before nutrient limitation occurs. A robust automatic perfusion rate control system based on the Alternating Tangential Flow or ATF system (Repligen,USA)) combined with an on-line live biomass monitor probe using Radio Frequency Impedance ( Aber Instruments, UK) is now being used in cell culture manufacturing processes. The system operates in a completely closed loop i.e. no samples need to be taken to obtain process information. In the control algorithm, a cell specific perfusion rate (CSPR) is specified and the signal from the biomass probe is converted into a perfusion flow rate through calculation and implementation with a variable speed controlled pump. For a continuous increase in perfusion, a live biomass probe is interfaced with the harvest pump, such that the perfusion rate is increased as a linear function of the cell density determined by the biomass probe, based on a desired CSPR. In this poster we show three examples of combining the ATF and live biomass probes. In the first example the bioreactor perfusion was controlled at a constant CSPR by the biomass probe to automatically increase the feed rate as cell density increased. The probe accurately estimated the viable cell density throughout the run with cell densities up to 110 million viable cells /mL. In the second example the probe was shown to immediately spot a sudden increase in live cell density caused by too much media being pumped out in error. For the final study, the capacitance was scanned within a wide range of frequency values (100–19,490 kHz). For the measured spectroscopic data, partial least squares regression (PLS), Cole–Cole, and linear modeling were applied and compared in order to predict VCD. The Cole–Cole model and the PLS model provided reliable prediction over the entire cultivation.

Included in

Engineering Commons