Chetan T. Goudar

A framework for real‐time glycosylation monitoring (RT‐GM) in mammalian cell culture

Glycosylation is a critical characteristic of biotherapeutics because of its central role in in vivo efficacy. Multiple factors including medium composition and process conditions impact protein glycosylation and characterizing cellular response to these changes is essential to understand the underlying relationships. Current practice typically involves glycosylation characterization at the end of a fed‐batch culture, which in addition to being an aggregate of the process, reflects a bias towards the end of the culture where a majority of the product is made. In an attempt to rigorously characterize the entire time‐course of a fed‐batch culture, a real‐time glycosylation monitoring (RT‐GM) framework was developed. It involves using the micro sequential injection (μSI) system as a sample preparation platform coupled with an ultra‐performance liquid chromatography (UPLC) system for real‐time monitoring of the antibody glycan profile. Automated sampling and sample preparations were performed using the μSI system and this framework was used to study manganese (Mn)‐induced glycosylation changes over the course of a fed‐batch culture. As expected, Mn‐supplemented cultures exhibited higher galactosylation levels compared to control while the fucosylation and mannosylation were consistent for both supplemented and control cultures. Overall, the approach presented in the study allows real time monitoring of glycosylation changes and this information can be rapidly translated into process control and/or process optimization decisions to accelerate process development. Biotechnol. Bioeng. 2015;112: 1146–1154.

Transcriptomics as a Tool for Assessing the Scalability of Mammalian Cell Perfusion Systems

DNA microarray-based transcriptomics have been used to determine the time course of laboratory and manufacturing-scale perfusion bioreactors in an attempt to characterize cell physiological state at these two bioreactor scales. Given the limited availability of genomic data for baby hamster kidney (BHK) cells, a Chinese hamster ovary (CHO)-based microarray was used following a feasibility assessment of cross-species hybridization. A heat shock experiment was performed using both BHK and CHO cells and resulting DNA microarray data were analyzed using a filtering criteria of perfect match (PM)/single base mismatch (MM) > 1.5 and PM-MM > 50 to exclude probes with low specificity or sensitivity for cross-species hybridizations. For BHK cells, 8910 probe sets (39 %) passed the cutoff criteria, whereas 12,961 probe sets (56 %) passed the cutoff criteria for CHO cells. Yet, the data from BHK cells allowed distinct clustering of heat shock and control samples as well as identification of biologically relevant genes as being differentially expressed, indicating the utility of cross-species hybridization. Subsequently, DNA microarray analysis was performed on time course samples from laboratory- and manufacturing-scale perfusion bioreactors that were operated under the same conditions. A majority of the variability (37 %) was associated with the first principal component (PC-1). Although PC-1 changed monotonically with culture duration, the trends were very similar in both the laboratory and manufacturing-scale bioreactors. Therefore, despite time-related changes to the cell physiological state, transcriptomic fingerprints were similar across the two bioreactor scales at any given instance in culture. Multiple genes were identified with time-course expression profiles that were very highly correlated (> 0.9) with bioprocess variables of interest. Although the current incomplete annotation limits the biological interpretation of these observations, their full potential may be realized in due course when richer genomic data become available. By taking a pragmatic approach of transcriptome fingerprinting, we have demonstrated the utility of systems biology to support the comparability of laboratory and manufacturing-scale perfusion systems. Scale-down model qualification is the first step in process characterization and hence is an integral component of robust regulatory filings. Augmenting the current paradigm, which relies primarily on cell culture and product quality information, with gene expression data can help make a substantially stronger case for similarity. With continued advances in systems biology approaches, we expect them to be seamlessly integrated into bioprocess development, which can translate into more robust and high yielding processes that can ultimately reduce cost of care for patients.

Elucidating the role of copper in CHO cell energy metabolism using 13C metabolic flux analysis

13C‐metabolic flux analysis was used to understand copper deficiency‐related restructuring of energy metabolism, which leads to excessive lactate production in recombinant protein‐producing CHO cells. Stationary‐phase labeling experiments with U‐13C glucose were conducted on CHO cells grown under high and limiting copper in 3 L fed‐batch bioreactors. The resultant labeling patterns of soluble metabolites were measured by GC‐MS and used to estimate metabolic fluxes in the central carbon metabolism pathways using OpenFlux. Fluxes were evaluated 300 times from stoichiometrically feasible random guess values and their confidence intervals calculated by Monte Carlo simulations. Results from metabolic flux analysis exhibited significant carbon redistribution throughout the metabolic network in cells under Cu deficiency. Specifically, glycolytic fluxes increased (25%–79% relative to glucose uptake) whereas fluxes through the TCA and pentose phosphate pathway (PPP) were lower (15%–23% and 74%, respectively) compared with the Cu‐containing condition. Furthermore, under Cu deficiency, 33% of the flux entering TCA via the pyruvate node was redirected to lactate and malate production. Based on these results, we hypothesize that Cu deficiency disrupts the electron transport chain causing ATP deficiency, redox imbalance, and oxidative stress, which in turn drive copper‐deficient CHO cells to produce energy via aerobic glycolysis, which is associated with excessive lactate production, rather than the more efficient route of oxidative phosphorylation.

Evaluating the impact of high Pluronic® F68 concentrations on antibody producing CHO cell lines

Pluronic® F68 (P-F68) is an important component of chemically-defined cell culture medium because it protects cells from hydrodynamic and bubble-induced shear in the bioreactor. While P-F68 is typically used in cell culture medium at a concentration of 1 g/L (0.1%), higher concentrations can offer additional shear protection and have also been shown to be beneficial during cryopreservation. Recent industry experience with variability in P-F68-associated shear-protection has opened up the possibility of elevated P-F68 concentrations in cell culture media, a topic that has not been previously explored in the context of industrial cell culture processes. Recognizing this gap, we first evaluated the effect of 1-5 g/L P-F68 concentrations in shake flask cultures over ten 3-day passages for cell lines A and B. Increase in terminal cell density and cell size was seen over time at higher P-F68 concentrations but protein productivity was not impacted. Results from this preliminary screening study suggested no adverse impact of high P-F68 concentrations. Subsequently fed-batch bioreactor experiments were conducted at 1 and 5 g/L P-F68 concentrations with both cell lines where cell growth, viability, metabolism, and product quality were examined under process conditions reflective of a commercial process. Results from these bioreactor experiments confirmed findings from the preliminary screen and also indicated no impact of elevated P-F68 concentration on product quality. If additional shear protection is desired, either due to raw material variability, cell line sensitivity, or a high-shear cell culture process, our results suggest this can be accomplished by elevating the P-F68 concentration in the cell culture medium without impacting cell culture performance and product quality.

An evaluation of public genomic references for mapping RNA-Seq data from Chinese hamster ovary cells.

While RNA‐Seq is increasingly used as the method of choice for transcriptome analysis of mammalian cell culture processes, no universal genomic reference for mapping RNA‐Seq reads from CHO cells has been reported. In previous publications, de novo transcriptomes assembled using these RNA‐Seq reads were subsequently used for mapping. Potential caveats with this approach include the incomplete coverage and the non‐universal nature of the de novo assemblies, leading to challenges in comparing results across studies. In order to facilitate future RNA‐Seq studies in CHO cells, we performed a comprehensive evaluation of four public genomic references for CHO cells hosted by the NCBI Reference Sequence Database (RefSeq), including two annotated genomes released in 2012 and 2014 and their accompanying transcriptomes. Each genome showed significantly higher mapped rates compared to its accompanying transcriptome. Furthermore, higher mapped rates in deep intra‐genic regions, especially within exons, were observed for the more recent genome release (2014) compared to the older one (2012), indicating that the 2014 genome was the preeminent reference among the four. Sequential addition of human and mouse genomes increased the total mapped rate to 87.3 and 89.7%, respectively, from 73.5% using the 2014 Chinese hamster genome alone. Thus, the sequential combination of the 2014 RefSeq Chinese hamster genome, the Ensembl human genome (h38), and the Ensembl mouse genome (m38) was suggested as the most effective strategy for mapping RNA‐Seq data from CHO cells. Biotechnol. Bioeng. 2015;112: 2412–2416.

Accelerating patient access to novel biologics using stable pool-derived product for non-clinical studies and single clone-derived product for clinical studies

Cell cloning and subsequent process development activities are on the critical path directly impacting the timeline for advancement of next generation therapies to patients with unmet medical needs. The use of stable cell pools for early stage material generation and process development activities is an enabling technology to reduce timelines. To successfully use stable pools during development, it is important that bioprocess performance and requisite product quality attributes be comparable to those observed from clonally derived cell lines. To better understand the relationship between pool and clone derived cell lines, we compared data across recent first in human (FIH) programs at Amgen including both mAb and Fc‐fusion modalities. We compared expression and phenotypic stability, bioprocess performance, and product quality attributes between material derived from stable pools and clonally derived cells. Overall, our results indicated the feasibility of matching bioprocess performance and product quality attributes between stable pools and subsequently derived clones. These findings support the use of stable pools to accelerate the advancement of novel biologics to the clinic.

An automated RNA-Seq analysis pipeline to identify and visualize differentially expressed genes and pathways in CHO cells.

Recent advances in RNA‐Seq based comparative transcriptomics have opened up a unique opportunity to understand the mechanisms of different phenotypes in bioprocessing‐related cell lines including Chinese hamster ovary (CHO) cells. However, simple and powerful tools are needed to translate large data sets into biologically relevant information that can be leveraged for genetic engineering and cell culture medium and process development. While tools exist to perform specific tasks associated with transcriptomics analysis, integrated end to end solutions that span the entire spectrum of raw data processing to visualization of gene expression changes on canonical pathways are rare. Additionally, these are not automated and require substantial user intervention. To address this gap, we have developed an automated RNA‐Seq analysis pipeline in R which leverages the latest public domain statistical advances in transcriptomics data analysis. This pipeline reads RNA‐Seq gene count data, identifies differentially expressed genes and differentially expressed pathways, and provides multiple intuitive visualizations as outputs. By using two publicly available CHO RNA‐Seq datasets, we have demonstrated the utility of this pipeline. Subsequently, this pipeline was used to demonstrate transcriptomic similarity between laboratory‐ and pilot‐scale bioreactors, helping make a case for the suitability of the lab‐scale bioreactor as a scaled‐down model. Automated end to end RNA‐Seq data analysis approaches such as the one presented in this study will shorten the time required from acquiring sequencing data to biological interpretation of the results and can help accelerate the adoption of RNA‐Seq analysis and thus mechanism‐driven approaches for cell line and bioprocess optimization.

A Comparative Transcriptomics Workflow for Analyzing Microarray Data From CHO Cell Cultures

Microarray-based comparative transcriptomics analysis is a powerful tool to understand therapeutic protein producing mammalian cell lines at the gene expression level. However, an integrated analysis workflow specifically designed for end-to-end analysis of microarray data for CHO cells, the most prevalent host for commercial recombinant protein production, is lacking. To address this gap, an automated data analysis workflow in R that leverages public domain analysis modules is developed to analyze microarray based gene expression data. In addition to testing the global transcriptome differences of CHO cells at different conditions, the workflow identifies differentially expressed genes and pathways with intuitive visualizations as the outputs. The utility of this automated workflow is demonstrated by comparing the transcriptomic profiles of recombinant protein expressing CHO cells with and without a temperature shift. Statistically significant differential expression at the gene, pathway, and global transcriptome levels are identified and visualized. An automated workflow like the one developed in this study will enable rapid translation of CHO culture microarray data into biologically relevant information for mechanism-driven cell line optimization and bioprocess development.

Characterization of intrinsic variability in time‐series metabolomic data of cultured mammalian cells

In an attempt to rigorously characterize the intrinsic variability associated with Chinese Hamster Ovary (CHO) cell metabolomics studies, supernatant and intracellular samples taken at 5 time points from duplicate lab‐scale bioreactors were analyzed using a combination of gas chromatography (GC)‐ and liquid chromatography–mass spectrometry (LC–MS) based metabolomics. The intrinsic variability between them was quantified using the relative standard deviation (RSD), and the median RSD was 9.4% and 12.4% for supernatant and intracellular samples, respectively. When exploring metabolic changes between lab‐ and pilot‐scale bioreactors, a high number of metabolites (65–105) were significantly different when no corrections were made for this intrinsic variability. This distinction also extended to principal component and metabolic pathway analysis. However, when intrinsic variability was taken into account, the number of metabolite with significant changes reduced substantially (20–25) as did the separation in principal component and metabolic pathway analysis, suggesting a much smaller change in physiology across bioreactor scale. Our results also suggested the contribution of biological variability to the total variability across replicates (∼0.4%) was significantly lower than that from technical variability (∼9–12%). Our study highlights the need for understanding and accounting for intrinsic variability in CHO cell metabolomics studies. Failure to do so can result in incorrect biological interpretation of the observations which could ultimately lead to the identification of a suboptimal set of targets for genetic engineering or process development considerations. Biotechnol. Bioeng. 2015;112: 2276–2283.