Developing efficient purification processes for non-mAb protein therapeutics can be a major challenge. These biomolecules are often not amenable to affinity capture and developing a robust train of chromatographic operations is time-consuming due to the large resin and buffer design space. To address this issue, we have developed a platformable approach to purify these challenging biomolecules by leveraging high throughput screening in combination with an
in silico process design tool to quickly identify purification sequences. The success of this approach relies on identifying and quantifying orthogonal chromatography sequences to select resins which remove non-overlapping sets of impurities to provide efficiency and robustness. Impurities such as HCPs and DNA found in null-producing cell culture fluid were chromatographically characterized on a set of multimodal, salt-tolerant, and ion exchange resins for a wide range of conditions. Fractions collected from this one-time screen were analyzed using UP-RPLC to generate a database of impurity fingerprints. An
in silico tool was then developed that leveraged product retention behavior to propose complete sets of integrated purification sequences containing up to 3 chromatography operations. This tool then quantified the extent to which sets of resins and conditions are orthogonally selective to each other and scored processes based on their predicted ability to remove impurities and recover a product.
To test this approach, several cytokine and hormone therapeutics containing impurities and challenging product variants were selected for purification from cell culture fluid. This workflow was applied and top-scoring purification sequences were identified for each product and experimentally tested and refined. For each product, complete purification development was achieved in approximately 3 weeks and impurities and product variants were both reduced to acceptable levels. This success not only demonstrates this as a promising tool for rapidly purifying newly emerging modalities, but also showcases the importance of quantitatively considering orthogonality in process development. Further, the orthogonality framework developed here may expand the implementation of multimodal resins in purification processes, which often exhibit unintuitive selectivies making them challenging to incorporate during process design.
