Predictive scale-up of ion exchange chromatography using high throughput miniature columns

Abstract

These findings presented in this thesis will help to support the industry trend towards increased usage of miniaturized chromatography columns as a scale-down model system combined with column simulations for downstream process development.

The biopharmaceutical industry is consistently evolving to create new platforms and methods to implement in purification processes. Miniaturized chromatography columns have the potential to innovate the development and characterization of these processes. Their lower material requirements, parallel operation, and implementation into robotic liquid handling systems for autonomous operation make them suitable for increasing the number of process parameters that can be evaluated and reducing experimental time in development. In addition to the growing interest in these advantages, there is a desire for the miniature columns to be implemented as scale-down models. There is, however, a lack of understanding with respect to this implementation due to differences associated with the use of miniature columns and robotic liquid handling systems as compared to traditional scale-down systems. In this thesis, the utility of these columns and systems are investigated using both experimental and theoretical methods.

A detailed comparison was carried out using both experimental data and column simulations for miniature and traditional benchtop columns. Appropriate translations to account for extracolumn system volume differences were developed and shown to result in comparability. A parameter estimation protocol was then implemented to obtain linear steric mass-action (SMA) isotherm and lumped mass transport parameters from gradient experiments in both systems. The parameters estimated from the miniature columns were successfully employed to simulate the benchtop experimental data.

Column simulations were then employed to help understand the effects of system differences between operational formats. Step elution experiments were then designed in silico using the simulator to evaluate its potential for process development. The identified conditions were then carried out experimentally and simulation comparisons were employed to demonstrate that increased dispersion of the induced salt step changes in the benchtop system could help to explain the observed differences. These results showed that not only can the simulations be used to provide insight regarding the data differences observed between the two experimental formats, but they can also successfully account for the subtle differences in operation. These findings further support the use of miniature chromatography columns as a scale-down model system combined with column simulations for downstream process development.

A comparison was also carried out using both experimental data and column simulations for small scale RoboColumns®, using a robotic liquid handling system, and lab scale chromatography columns, using the traditional benchtop system. Importantly, the parameters obtained from the RoboColumn® experiments were shown to successfully predict the scaled-up behavior of the lab scale system.