Modeling of protein monomer/aggregate purification by hydrophobic interaction chromatography: Application to column design and process optimization

Despite the use of Hydrophobic Interaction Chromatography (HIC) in the purification protocol of a wide variety of protein therapeutics, the mechanism of separation of protein monomer/aggregate species within the HIC column is complex and still not fully understood. Several attempts have been made to develop a comprehensive model to explain the separation process including the solvophobic, preferential interaction and homogenous diffusion/competitive Langmuir binary isotherm. A mechanistic model was developed based on competitive binary Langmuir kinetics as the rate-limiting step and a mass balance in the liquid phase was used to predict protein/aggregate distribution in the column under binding and elution conditions. A Partial Least Square (PLS) regression model was used to estimate the concentration and composition of the etuent stream. The PLS predictions were compared to mechanistic model predictions. The mechanistic model developed successfully predicted the binding and elution curves for both monomer and aggregates. Optimum amount of resin necessary for a specification, flow rate, and ideal binding and elution points can be estimated using the mechanistic model. This could result in significant savings within the biopharmaceutical industry by reducing the process development time and optimizing resin use in chromatography.