Sebastian, Martina Fineas; (2023) Robust design of integrated upstream and recovery operations for antibody manufacture using scale-down mimics and predictive modelling. Doctoral thesis (Eng.D), UCL (University College London).

Preview

Text Martina_SebastianThesis_EngD_after_VIVA_clean.pdf - Accepted Version Download (5MB) | Preview

Abstract

Disk-stack centrifugation followed by depth filtration is the cornerstone in the clarification of mammalian cell cultures used for biopharmaceutical manufacture. The drive for higher cell culture productivities has led to cell culture harvests with increased solids concentration and cell debris that can be difficult to clarify with the existing recovery platforms. Due to the limitations in the available scale-down models for the upstream-recovery sequence, these challenges are often first faced upon scale-up leading to facility fit issues and an increase in the costs of manufacture. The aim of this thesis is to address this challenge through building a novel integrated framework for the upstream-recovery sequence. The framework leverages scale-down mimics with modelling techniques using multivariate data analysis (MVDA) and bioprocess economics for a holistic view of the upstream and recovery operations. To map the upstream-recovery design space, an integrated small scale platform was set up comprising a high throughput microbioreactor system, a centrifugation mimic based on a capillary shear device (CSD) and cholesterol concentration as a filterability predictor. The thesis focused on using this novel set up with Design of Experiments (DoE) and MVDA to derive cause-and effect correlations for upstream (e.g. product titre) and recovery (e.g. solids concentration and filterability) metrics as a function of operating conditions. The results revealed trade-offs between increasing titre, the solids load onto the centrifuge and the centrate’s filterability. The work also identified the impact of cell culture viability and centrifugation shear on intact antibody monomer content. The predictive models and cholesterol-based correlation for filterability were integrated into a bioprocess economics model, which was built to determine the cost of goods (COG) for the recovery operation. This enabled a better prediction of the impact of shear on facility fit, throughput and COG [under uncertainty]. Windows of Operation for the integrated upstream-recovery sequence were generated to determine the sweet spot in terms of cell culture operating conditions (harvest day, seeding density) and recovery shear levels that meet multiple conflicting output targets, namely titre, product loss, filter area, COG and throughput. A major challenge in using high throughput experimentation such as the framework proposed in this thesis is the associated burden of high volume of analytics. To address this challenge, novel rapid approaches were developed that accurately predicted solids concentration and cell damage using the particle size distribution (PSD) data from a routinely used automated cell counter. By leveraging existing PSD data, this has the potential to reduce analytical resource and time. The work in this thesis showed the benefits of using a holistic approach to the optimisation of the upstream-recovery sequence of operations. The developed framework can be used to design more efficient processes earlier in the development cycle accelerating development timelines and reducing costs.

Download activity - last month

Export as XMLCSVJSON

Download activity - last 12 months

Export as JSONXMLCSV

Downloads by country - last 12 months

Export as XMLCSVJSON

Archive Staff Only

View Item