Hydrophobic interaction chromatography (HIC) is one of the common separation modes for the purification of biomolecules. It is especially powerful as polishing step for monoclonal antibodies to remove aggregates and impurities such as host cell proteins. The separation principle is based on the reversible interaction between hydrophobic patches on the protein surface and the hydrophobic ligands of the stationary phase.
Advances in fundamental understanding of HIC have been achieved by mechanistic studies and modeling work based on the Mollerup’s thermodynamic framework (2008). Mechanistic modeling provided a deeper insight into adsorption mechanism and offers time saving in process development. Strategies for model-based process development of salt- and pH-dependent IEX and mixed mode chromatography were reported by Nfor et al. (2010), and Lee et al. (2015).
pH extension and mixed-mode binding
Modeling of pH-dependence is a necessity for practical model-based development of industrial HIC processes, as these commonly involve an alteration of the binding strength by changing the pH. In this work, we present how we extended our previously developed HIC model (Wang et al., 2016) to account for changes in the pH. Bind/elute experiments were performed by applying linear salt gradients at constant pH. The adsorption behavior of three pure proteins (lysozyme, hemoglobin, IgG) on TOYOSCREEN PPG-600M® (Tosoh), was investigated for pH values near the pI and could be modeled successfully by introducing just a single additional model parameter. Model validation was conducted using combined linear pH and salt gradients.
Furthermore, we combined our HIC model with the Steric Mass Action (SMA) isotherm to construct a mixed-mode model. A case study with the model protein glucose oxidase investigates identifiability of adsorption isotherm parameters on Capto adhere® (GE Healthcare) and compared the models’ ability to match the observed peak shapes. Compared to the model by Nfor et al. (2010) which required a complex column and pore model, the newly constructed model provided equally good results with a much simpler column model and, thus, less parameters.
Successful application in contract modeling projects
In summary, we could successfully extend our previously developed HIC model to account for pH-changes and mixed-mode binding. The model proved to be superior to existing isotherm models and has already been used for industrial process development in service projects at GoSilico.
Please fill out this form to
download your files