OPTIMIZATION STUDY ON PERIODIC COUNTER-CURRENT CHROMATOGRAPHY (PCC) INTEGRATED IN A MAB DOWNSTREAM PROCESS

摘要

In the biopharmaceutical market, there is an increasing pressure to reduce prices. Nowadays, a big part of the manufacturing cost is due to the product purification. A big effort has been made in this matter to reduce costs, and continuous downstream processes like periodic counter-current chromatography (PCC) have gained interest in the last years due to a higher productivity and resin utilization compared to a batch process. Herein, a PCC process is integrated with a virus inactivation and two ion-exchange chromatography steps (polishing steps) for the purification of mAb at lab scale, and multi-objective optimization is used to compare several process configurations using two resins with different particle sizes. The optimization was based on a mechanistic model of the breakthrough curve in the protein A capture step. The breakthrough curve and the length of the product recovery determined the scheduling of the PCC process, since the loading of the capture step had to be at least so long as the time that takes to purify the product. Residence time during the loading of the capture column and cycle time were selected as decision variables, and productivity and resin utilization were the objective functions, while yield during the loading of the capture column was set as a constraint. A set of optimal solutions were obtained (Figure 1). In the solutions with high productivity, the residence time was lower, and as a result, the breakthrough curve was flatter. Therefore, the PCC cycle ended earlier to avoid product loss, which led to a decrease in resin utilization. The opposite happened for the solutions with higher resin utilization. The process configuration affected the optimization by setting a lower bound on the cycle length. Therefore, several process configurations were compared: only the capture step in one system, the complete purification in series in one system, and the same process in parallel in two systems. In Figure 1, the effect of the process configuration on the Pareto front is shown. Experimentally, the process was integrated and run simultaneously in two systems, which were centrally controlled by an external software called Orbit. In general, the resin with smaller particle size had a higher dynamic binding capacity at all residence times, which allowed to obtain a better compromise between productivity and resin utilization.