Results and Discussion

This study successfully demonstrated a next-generation continuous purification process, running for 14-days at a rate of 50L of cell culture harvest material per day. Bioburden control was maintained for the duration of the run and product quality results were within acceptable limits. Advanced automation enabled tight control of the process to meet specifications. Automation control strategies also enabled break vessel volumes between 150 ml and 10L (20-fold or greater reduction compared to batch). Facility footprint of the process was reduced by 75% compared to the batch process, primarily due to the reduction in vessel volumes and ability to perform all three process steps in a single system. Setup time was to 1.5 days and was followed by 14 days of hands-free operation with the exception of daily sampling (90% reduction in operator hours). These outcomes demonstrate the robustness of this commercially available automation system for continuous purification processes.

Bioburden Control

Most downstream operations for monoclonal antibody are not considered aseptic and operate as low-bioburden. The guidance commonly used sets a limit at 10 colony-forming units (CFU) per mL.
Bioburden control was demonstrated for the duration of the 14-day run. Samples were taken daily from the harvest break vessel, protein A product vessel, VI neutralized product vessel and AEX final product. None of the sampling points tested positive for bioburden except, on day 7, when 1CFU/mL was observed for the protein A product vessel (below the acceptable criteria of <10CFU/mL). This hit was attributed to intentionally opening the protein A product vessel to fix an incorrectly installed tubing assembly on the vessel. The bioburden objectives for this process were clearly met. The 1CFU/mL sample from the protein A elution product vessel was cleared by implementing a bioburden mitigation strategy. This strategy consisted of six important steps: (1) pausing the protein A step, (2) removing the contaminated vessel from the system, (3) autoclaving a new vessel, (4) filtering material into the new vessel under aseptic conditions, (5) integrating the new vessel into the process and (6) flushing lines with sanitization buffer before restarting the protein A step. No bioburden was detected after the implementation of this strategy. Successful implementation of this mitigation strategy demonstrates the ability to clear bioburden from an ongoing process without significant process disruption, a necessity to save a batch when operating a continuous operation.

Capture

Seventy protein A capture chromatography cycles were successfully run during the 14-day process. The protein A capture step began automatically once the harvest vessel reached a setpoint volume of 2L. The capture step operated for the duration of the run.
The capture step was operated under a constant mass load principle, targeting the resin capacity of 60g of mAb per L of resin: the load volume onto the column changed as the titer varied throughout the run to ensure that the grams of mAb loaded was the same for each cycle, as illustrated in Figure 4. Load volume decreases as titers increase until the middle of the run. Load volume then increases again as the titer drops towards the end of the run. This strategy ensures a constant concentration downstream of the protein A capture step, which greatly simplifies the downstream process dynamics. The standard deviation for the Protein A concentration over the duration of the run was 1.4 g/L. With constant downstream concentration, the virus inactivation step does not need to account for large mAb concentration changes that can affect the titration profile. Additionally, polishing chromatography step throughput can be calculated from volumetric throughput and the known concentration. The capture step was triggered to stop once the harvest vessel dropped to a low level due to lack of flow from the cell culture feed.