Intracellular Profiling for Biopharmaceutical Cultivation Processes

Abstract

It is only about 30 years ago when the foundation was laid for the revolution towards modern biopharmaceutical production of monoclonal antibodies (mAb). Since then a fast-paced development has taken place. The demand for high yields, product quality, and process efficiency has increased over the years, as mAb have evolved to a profitable line of pharma business with a yearly revenue of over 50 billion US dollars. Nowadays the industrial production of humanized mAb is done by fed-batch cultivation of mammalian cell lines, particularly Chinese hamster ovary (CHO) ones, in bioreactors up to a volume of 15,000 liters. Another production mode gaining popularity is based on the perfused operation of bioreactors, enabling continuous long-term cultivation with high cell densities and constant conditions. This steady-state (SS) is beneficial to increase the product quality and homogeneity while ensuring high yields. The development of the corresponding production processes is complex, involving modern micro scale-down models, cell line development and cell media optimization. Current process development relies strongly on process parameters and extracellular determined factors. Rising ‘omics technologies enable the analysis of the transcriptome, proteome and metabolome. The gained information contributes to enlighten the cell line’s phenotypes and this knowledge can help to improve process development and –optimization. In the context of this doctoral thesis the application and development of intracellular analyses for the characterization of topics relevant for biopharmaceutical process development and optimization are shown. This includes the validation of micro-scale bioreactors as scale-down models, the investigation of the SS in continuous perfusion processes and the development of an amino acid (AA) analysis method, which can be applied for cell line and cell media development. Each study was based on a CHO mAb producing platform process. For the intracellular scale-down model validation of modern micro-bioreactors, the proteome of a 10 mL ambrTM high throughput system and a 300 L pilot scale fed-batch bioreactor were compared. A hierarchical clustering of the proteome time evolution delivered day-wise clusters independent of the scale, which proved their similarity on a microscopic base. To capture small differences between the datasets each individual relative protein abundance, the so-called fold change (FC), was compared applying a dataset based FC threshold. Only a negligible number varied between the two scales, showing no common biological function. The characterization study was finalized with the analysis of proteins related to N-glycosylation, one of the most important mAb quality attributes. The study proves micro-bioreactors to be sophisticated scale-down models. The SS of a continuous cultivation was characterized performing transcriptomics and proteomics analysis to capture two levels of gene expression. For both datasets, the FC distribution over time was analyzed and the proteome was compared to the previously studied fed-batch distribution, showing consistency for the perfusion and large variation in time for the fed-batch process. Transient behavior to SS within seven days, based on the characteristic time of fluid dynamics and intracellular processes superimposing each other, was observed. To investigate the transition on the ‘omics level, the transcripts and proteins were grouped into three groups, according to their transition to SS. Only 0.4 % of all the entities never reached SS conditions. Functional annotation was performed to understand the biological function of the entities. Galactose processing and cell productivity, two parameters important for successful long-term cultivations of mAb, were investigated in detail. The observed decrease in specific productivity over time was most likely from epigenetic origin and not due to previously reported loss in copy numbers. In general, the perfusion process showed very stable intracellular conditions over time. As metabolites differ in their chemical structure, whole metabolome analysis is still a long way off. In order to expand the metabolite fingerprint of our CHO process in a fast and sensitive manner we developed a quantitative AA analysis method using MALDI-TOF-MS. Initially the matrix showing highest AA detection yields and suitable spotting for a microarray for mass spectrometry (MAMS) target, was selected. Automatic data processing established with MATLAB® was used to analyze spectral data. A customized internal standard was developed, where four isotopically labeled AA were carefully selected based on three parameters, the AA`s chemical structure, the linear fit of the calibration and the comparison to a UPLC standard method. The method was applied for AA profiling of two independent perfusion experiments. One aiming to reflect stable conditions of a SS and the second one capturing the variation caused by different asparagine and glucose concentrations in the cell culture media. The developed method showed consistency for most quantified AAs with the UPLC method and the process conditions were reflected in their time profiles. These studies further emphasize that intracellular profiling delivers valuable insights for the characterization of cultivation processes used in the development and optimization of biopharmaceutical production.