Minimizing batch‐to‐batch variability of a live virus vaccine by process analytical technologies

For bioprocesses producing live virus, such as enterovirus Coxsackievirus A21, viral titer (infectivity basis) decay rates can exceed 30% within a day. Consequently, harvest timing is paramount. To optimize titer at harvest, a continuous viral product titer model was generated to elucidate kinetics. The model leveraged experimentally determined viable cell density, cell-specific viral productivity, and viral specific decay rates. Next, three separate online process analytical technology (PAT) harvest triggers were developed to predict maximal viral titer. Finally, the PAT harvest triggers were tested alongside traditional time-based harvests. The harvest triggers utilized common bioprocessing tools - dissolved oxygen (DO) and capacitance probes - to track DO and viable cell volume (VCV) and derived a third parameter, cell-specific oxygen uptake rate. Harvesting with PAT triggers allowed for significantly improved batch-to-batch consistency. The standard deviation of harvest yield was reduced by 41% (DO), 56% (OUR) and 71% (capacitance) as compared to the industry standard time-based harvest. Even when a process deviation in inoculated cell density occurred, causing a significant shift in viral titer kinetics, the PAT harvest triggers yielded greater than 87% of peak titer. By comparison, the time-based harvest yielded 16%.