Presented at the ASMS 2018, San Diego, CA, USA
Chinese hamster ovary (CHO) cell lines are the workhorse for protein production in the biopharmaceutical industry. Although the protein yields from CHO cells have increased due to the understanding of cell environmental perturbation, the lack of knowledge on intracellular metabolic activities caused limited exploitation of major cellular processes for protein productivity. To understand CHO cell metabolic activities, mass spectrometry (MS) were used to study intracellular biomolecules, like aqueous metabolites (metabolomics) and lipids (lipidomics). However, the integration of both metabolomics and lipidomics data present challenges in raw file formats, parameter customization, automation, and standardization, which makes high-throughput analysis unfeasible. In this study, the use of a single software platform to facilitate the processing of metabolomics and lipidomics data will be presented
Raw LC-MS data from liquid-liquid extracted samples were directly loaded into the software platform (Genedata AG, Basel, Switzerland). The data comprises of samples from biologically replicated CHO cell bioreactors across ten days, lipid and aqueous metabolite extracted fractions, and data acquired in both positive and negative ionization mode on the MS. Automated data processing workflow was executed with a one-click operation. All the data were integrated and compared across different CHO cell bioreactors and at different culture days. The trends of metabolic changes in the CHO cells were analyzed with statistics within the software platform. The workflows were optimized and applied to routine bioprocess monitoring of CHO cell bioreactors.
In this study, the implementation of a routine workflow for multi-omics bioprocess monitoring with a single software platform for the data processing, analysis, and management of MS data will be presented. In this approach, dedicated workflows were customized to the application before integration for further analysis. Optimized data processing was applied to large data sets and execution times scaled linearly with the number of samples. Browsing and downstream data analyses, including statistical tests, visual verification of the results, biomarker identification and generation of customized reports, were performed. This approach can be fully automated and employed as part of a bioprocess control strategy. For this study, a proof-of-concept was shown as an example to describe the possibility of using the optimized workflows to routinely monitor bioreactors at a larger scale. A compliance module including GxP functionalities such as audit trails, electronic signatures and data security allows the deployment of this software platform in regulated environments.