For non-platformable separation challenges, often times non-affinity ligands such as mixed-mode, IEX, or HIC are tasked with the burden for HCP removal. In such systems, various approaches have been utilized to characterize relevant HCP properties using techniques such as 2D gels or octanol-water partition coefficients. Although these approaches often suggest the type of chromatography likely to remove group-specific HCPs, they do not exploit the subtle differences between different ligands within a given resin family. Here we present a novel platformable approach for the rapid in silico identification of downstream processes that provide highly orthogonal HCP selectivity between individual chromatographic steps. Null-producing cell culture fluid supernatant from Pichia pastoris was loaded onto 12 commercially available mixed-mode resins and subjected to fractionation by salt and pH gradients. Fractions were analyzed by UP-RPLC and used to generate a database for the chromatographic behavior of group-specific HCPs. This one-time HCP screening approach was combined with rapid pure-component product screens in a plug-and-play manner using 2 relevant targets, hGH and G-CSF. A novel in silico process design tool was developed to search for processes that provide the highest degree of HCP orthogonal selectivity while also satisfying constraints to allow for integrated operation. The process outputs for these 2 molecules were then experimentally verified and found to reduce HCP levels by more than 5 logs to below 10 ppm. The work presented here will provide a rapid generic approach for efficient HCP removal in non-platformable systems and we have also shown that this approach can be extended to identify processes for the efficient removal of both process and product-related variants.

In silico process design tool for rapid identification of orthogonally selective multimodal chromatographic materials for HCP removal

253rd National Meeting (2017)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.