(674f) Mechanistic Modeling and Control of Column-Free Continuous Chromatography Process

Chromatography is a critical process used to purify biological molecules. The dominant configuration employs resin particles in columns typically operated in batch mode. The resin either captures the biological molecule of interest or allows the biological molecule to flow through while removing impurities (aka “polishing”). As bioreactors have become more productive, the operating costs have increasingly been shifted to downstream purification, with the cost of goods reported as high as 75% for the chromatography columns for some biotherapeutics such as monoclonal antibodies. Continuous configurations that oscillate between multiple columns have become commercially available, reducing the operating costs of downstream purification, which are referred to as “continuous chromatography.”¹

This work involves an alternative continuous chromatography design that does not employ columns and does not have an oscillatory operation. This Continuous Countercurrent Tangential Chromatography (CCTC) technology more efficiently uses resin (by far the most expensive part of the process for some biotherapeutics, such as monoclonal antibodies) while operating in quasi-steady conditions so that the material leaving the process has greater consistency.² This technology is also able to capture and release biological molecules in less than ten minutes, which is essential for those molecules that have poor stability when adsorbed on or within the resin particles.

This work derives the first distributed parameter model for the CCTC process. The model is validated experimentally for the most expensive step in the purification of monoclonal antibodies (mAbs), which is the capture of mAbs from clarified bioreactor material. The separation is achieved due to the adsorption of mAb from the feed into the surface and pores of resin particles suspended in an aqueous buffer solution.^3-4 The mechanistic model developed for this process relies on the kinetic parameters, including the binding rate constant, pore diffusion coefficient, and film transfer coefficient. A full Bayesian inference is performed using the Markov-chain Monte-Carlo (MCMC) method that incorporates prior knowledge and quantifies parameter uncertainties. The model accurately predicts the concentration of purified mAb as a function of residence time. Furthermore, we derive a model-based control system for the automated operation of the system. The concentration profile of mAb shows stable output and quickly responds to variations in the feed composition. The robustness, scalability, and high yield for large biomolecules with very short residence times enable the capture and elution of poorly stable biotherapeutic molecules.

References:

1. Behere, K.; Yoon, S., Chromatography bioseparation technologies and in-silico modelings for continuous production of biotherapeutics. J Chromatogr A 2020, 1627, 461376.
2. Shinkazh, O.; Kanani, D.; Barth, M.; Long, M.; Hussain, D.; Zydney, A. L., Countercurrent tangential chromatography for large-scale protein purification. Biotechnol Bioeng 2011, 108 (3), 582-91.
3. Dutta, A. K.; Tran, T.; Napadensky, B.; Teella, A.; Brookhart, G.; Ropp, P. A.; Zhang, A. W.; Tustian, A. D.; Zydney, A. L.; Shinkazh, O., Purification of monoclonal antibodies from clarified cell culture fluid using Protein A capture continuous countercurrent tangential chromatography. J Biotechnol 2015, 213, 54-64.
4. Dutta, A. K.; Fedorenko, D.; Tan, J.; Costanzo, J. A.; Kahn, D. S.; Zydney, A. L.; Shinkazh, O., Continuous countercurrent tangential chromatography for mixed mode post-capture operations in monoclonal antibody purification. J Chromatogr A 2017, 1511, 37-44.