Mechanistic modelling of microscale chromatography

Abstract

Microscale chromatography as an experimental tool has shown much utility in process development due to reduced material consumption and ease of parallelisation which are of major benefit when compared to conventional lab-scale studies. Microscale columns are commonly used in early process development where the most impactful decisions, such as choice of unit operation, purification strategy, resin, and the choice of candidate are made with limited resources and knowledge. Understanding the behaviour of microscale chromatography and better applying the knowledge gained from microscale studies to large scale chromatography may allow faster, more efficient and more robust early process development, and therefore more effective processes once a bioprocess is fully developed and products commercialised. It is the overall aim of the project to develop a model to determine large scale mass transfer parameters describing a lab-scale chromatographic process from microscale data, and allow one to simulate and optimise large scale separations whilst enjoying the benefits of reduced resource consumption of the microscale domain. From the outset, characterisation ofthe differences between lab-scale columns operated on a conventional Fast Protein Liquid Chromatography (FPLC) system and microscale columns on a robotic Liquid Handling System (LHS) was performed. Determining the common metrics of column performance, HETP, asymmetry and experimentto-experiment or column-to-column variation between columns and experiments provides an understanding of some of the key differences between lab-scale and microscale column formats with regards to system, scale and data quality, as well as providing an opportunity to optimise the experimental design of microscale experiments. This was performed through evaluating methods of improving resolution, including fashioning rigs to use microscale columns on a conventional system, evaluating various tracer substances and evaluating a novel strategy of pre-filling collection plates. Investigations into ascertaining the dynamic binding capacity (DBC) of IgG to Protein A resin using microscale data has been performed with 3 microscale column volumes at several residence times using the high throughput system, and repeated at lab scale, with further work into understanding the effect of intermittent flow on resin: target interaction by mimicking the microscale operation on a larger system. This effort has led towards data used to calibrate a mechanistic model of chromatography at both lab scale and microscale with the intention of predicting lab scale behaviour. By correcting for scale, operational and flow effects, one may predict large scale performance through calibrating a model with microscale data, enabling better process understanding with reduced material consumption.