(82a) Implementation of a Small-Scale Capillary Shear Model to Facilitate Process Development of Disc Stack Centrifugation during the Harvest of Biopharmaceuticals

Harvest is generally the first step in the downstream processing of protein therapeutics and serves to clarify the cell broth by removing cells and debris. Disc stack centrifuges (DSCs) are often used during harvest, as they are capable of continuous operation and alleviate load on subsequent harvest filtration steps. It is important to understand how the DSC process parameters affect cell debris clearance and the degree of cell shearing, which can impact product quality attributes such as host cell protein and DNA concentration. However, as these DSCs require a relatively large amount of cell broth to operate, process development of the harvest step becomes impractical at the lab scale. Therefore, a small-scale model of the DSC is desirable but difficult to implement as DSCs impart shear forces on cells that differ from those generated by laboratory-scale centrifuges. Still, some small-scale DSC models have been proposed in literature. One such model uses a capillary shear device (CSD) to shear cell broth, which is then centrifuged with a lab-scale fixed-angle centrifuge to create DSC-like centrate. We have explored the use of this small-scale DSC model to process cell broth obtained from commercial-scale manufacturing batches of monoclonal antibodies. Specifically, we have evaluated how the model’s parameters (flow rate of cell broth through the CSD, centrifugation time, and centrifugation speed) affect the characteristics of the model-generated centrate (turbidity, particle size distribution, and LDH activity – a measure of cell lysis). With a better understanding of how model parameters affect the model-generated centrate, the model can be used to replicate manufacturing-scale DSC centrate at the small scale such that harvest filtration studies can be conducted more efficiently. Further study of how the model’s parameters relate to DSC parameters (bowl speed and flow rate) will ultimately allow us to evaluate if the model can be used to conduct process development of the entire harvest step (centrifugation and filtration) at the lab scale.