Faraaz Noor Khan Yusufi

Combined In Silico Modeling and Metabolomics Analysis to Characterize Fed-Batch CHO Cell Culture

The increasing demand for recombinant therapeutic proteins highlights the need to constantly improve the efficiency and yield of these biopharmaceutical products from mammalian cells, which is fully achievable only through proper understanding of cellular functioning. Towards this end, the current study exploited a combined metabolomics and in silico modeling approach to gain a deeper insight into the cellular mechanisms of Chinese hamster ovary (CHO) fed‐batch cultures. Initially, extracellular and intracellular metabolite profiling analysis shortlisted key metabolites associated with cell growth limitation within the energy, glutathione, and glycerophospholipid pathways that have distinct changes at the exponential‐stationary transition phase of the cultures. In addition, biomass compositional analysis newly revealed different amino acid content in the CHO cells from other mammalian cells, indicating the significance of accurate protein composition data in metabolite balancing across required nutrient assimilation, metabolic utilization, and cell growth. Subsequent in silico modeling of CHO cells characterized internal metabolic behaviors attaining physiological changes during growth and non‐growth phases, thereby allowing us to explore relevant pathways to growth limitation and identify major growth‐limiting factors including the oxidative stress and depletion of lipid metabolites. Such key information on growth‐related mechanisms derived from the current approach can potentially guide the development of new strategies to enhance CHO culture performance. Biotechnol. Bioeng. 2012; 109:1415–1429.

Metabolomics-driven approach for the improvement of Chinese hamster ovary cell growth: Overexpression of malate dehydrogenase II

We have established a liquid chromatography-mass spectrometry based metabolomics platform to identify extracellular metabolites in the medium of recombinant Chinese hamster ovary (CHO) fed-batch reactor cultures. Amongst the extracellular metabolites identified, malate accumulation was the most significant. The contributing factors to malate efflux were found to be the supply of aspartate from the medium, and an enzymatic bottleneck at malate dehydrogenase II (MDH II) in the tricarboxylic acid cycle. Subsequent metabolic engineering to overexpress MDH II in CHO resulted in increases in intracellular ATP and NADH, and up to 1.9-fold improvement in integral viable cell number.

Metabolomics profiling of extracellular metabolites in recombinant Chinese Hamster Ovary fed-batch culture.

A metabolomics-based approach was used to time profile extracellular metabolites in duplicate fed-batch bioreactor cultures of recombinant Chinese Hamster Ovary (CHO) cells producing monoclonal IgG antibody. Culture medium was collected and analysed using a high-performance liquid chromatography (HPLC) system in tandem with an LTQ-Orbitrap mass spectrometer. An in-house software was developed to pre-process the LC/MS data in terms of filtering and peak detection. This was followed by principal component analysis (PCA) to assess variance amongst the samples, and hierarchical clustering to categorize mass peaks by their time profiles. Finally, LC/MS2 experiments using the LTQ-Orbitrap (where standard was available) and SYNAPT HDMS (where standard was unavailable) were performed to confirm the identities of the metabolites. Two groups of identified metabolites were of particular interest; the first consisted of metabolites that began to accumulate when the culture entered stationary phase. The majority of them were amino acid derivatives and they were likely to be derived from the amino acids in the feed media. Examples included acetylphenylalanine and dimethylarginine which are known to be detrimental to cell growth. The second group of metabolites showed a downward trend as the culture progressed. Two of them were medium components--tryptophan and choline, and these became depleted midway into the culture despite the addition of feed media. The findings demonstrated the potential of utilizing metabolomics to guide medium design for fed-batch culture to potentially improve cell growth and product titer.

Developing genomic platforms for Chinese hamster ovary cells

Chinese hamster ovary (CHO) cells are widely used in recombinant protein production, yet despite their importance in bioprocessing, few genomic resources have been developed for this cell line. Over the past several years, we have made considerable progress in the development of genomic tools for CHO. Using Sanger-based sequencing technology, we have accrued a sequence repertoire of more than 68,000 expressed sequence tags (ESTs), representing more than 28,000 unique CHO transcripts. Using closely related species, we have functionally annotated this sequence set and have currently achieved significant representation in a number of functional classes, including some closely tied to recombinant protein production. This sequence repository has been used to design custom CHO Affymetrix arrays for transcriptome analysis. Illumina Solexa deep sequencing technology was also applied to study the CHO cell transcriptome and survey the identity and expression of small RNAs. These applications demonstrate the utility of genomic tools, and illustrate the applicability of emerging next-generation sequencing technologies.

Reaching the depth of the Chinese hamster ovary cell transcriptome

The high-throughput DNA sequencing Illumina Solexa GAII platform was employed to characterize the transcriptome of an antibody-producing Chinese hamster ovary (CHO) cell line. More than 55 million sequencing reads were generated and mapped to an existing set of CHO unigenes derived from expressed sequence tags (ESTs), as well as several public sequence databases. A very significant fraction of sequencing reads has not been previously seen. The frequency with which fragments of a unigene were sequenced was taken as an estimate of the abundance level of the corresponding transcripts. A wide dynamic range of transcript abundance levels was observed, spanning six orders of magnitude. However, the distribution of coverage across transcript lengths was found to vary, from relatively uniform to highly variable. This observation suggests that more challenges are yet to be resolved before direct sequencing can be used as a true quantitative measure of transcript level and for differential gene expression analysis. With the depth that high-throughput sequencing methods can reach, one can expect that the entire transcriptome of this industrially important organism will be decoded in the near future.

Metabolomics-based identification of apoptosis-inducing metabolites in recombinant fed-batch CHO culture media.

A liquid chromatography-mass spectrometry (LC-MS) based metabolomics platform was previously established to identify and profile extracellular metabolites in culture media of mammalian cells. This presented an opportunity to isolate novel apoptosis-inducing metabolites accumulating in the media of antibody-producing Chinese hamster ovary (CHO mAb) fed-batch bioreactor cultures. Media from triplicate cultures were collected daily for the metabolomics analysis. Concurrently, cell pellets were obtained for determination of intracellular caspase activity. Metabolite profiles from the LC-MS data were subsequently examined for their degree of correlation with the caspase activity. A panel of extracellular metabolites, the majority of which were nucleotides/nucleosides and amino acid derivatives, exhibited good (R² > 0.8) and reproducible correlation. Some of these metabolites, such as oxidized glutathione, AMP and GMP, were later shown to induce apoptosis when introduced to fresh CHO mAb cultures. Finally, metabolic engineering targets were proposed to potentially counter the harmful effects of these metabolites.

Global Insights Into the Chinese Hamster and CHO Cell Transcriptomes

Transcriptomics is increasingly being used on Chinese hamster ovary (CHO) cells to unveil physiological insights related to their performance during production processes. The rich transcriptome data can be exploited to provide impetus for systems investigation such as modeling the central carbon metabolism or glycosylation pathways, or even building genome‐scale models. To harness the power of transcriptome assays, we assembled and annotated a set of RNA‐Seq data from multiple CHO cell lines and Chinese hamster tissues, and constructed a DNA microarray. The identity of genes involved in major functional pathways and their transcript levels generated in this study will serve as a reference for future studies employing kinetic models. In particular, the data on glycolysis and glycosylation pathways indicate that the variability of gene expression level among different cell lines and tissues may contribute to their differences in metabolism and glycosylation patterns. Thereby, these insights can potentially lead to opportunities for cell engineering. This repertoire of transcriptome data also enables the identification of potential sequence variants in cell lines and allows tracing of cell lineages. Overall the study is an illustration of the potential benefit of RNA‐Seq that is yet to be exploited. Biotechnol. Bioeng. 2015;112: 965–976.

Enhanced expression of codon optimized interferon gamma in CHO cells

The human interferon-gamma (IFN-γ) is a potential drug candidate for treating various diseases due to its immunomodulatory properties. The efficient production of this protein can be achieved through a popular industrial host, Chinese hamster ovary (CHO) cells. However, recombinant expression of foreign proteins is typically suboptimal possibly due to the usage of non-native codon patterns within the coding sequence. Therefore, we demonstrated the application of a recently developed codon optimization approach to design synthetic IFN-γ coding sequences for enhanced heterologous expression in CHO cells. For codon optimization, earlier studies suggested to establish the target usage distribution pattern in terms of selected design parameters such as individual codon usage (ICU) and codon context (CC), mainly based on the hosts highly expressed genes. However, our RNA-Seq based transcriptome profiling indicated that the ICU and CC distribution patterns of different gene expression classes in CHO cell are relatively similar, unlike other microbial expression hosts, Escherichia coli and Saccharomyces cerevisiae. This finding was further corroborated through the in vivo expression of various ICU and CC optimized IFN-γ in CHO cells. Interestingly, the CC-optimized genes exhibited at least 13-fold increase in expression level compared to the wild-type IFN-γ while a maximum of 10-fold increase was observed for the ICU-optimized genes. Although design criteria based on individual codons, such as ICU, have been widely used for gene optimization, our experimental results suggested that codon context is relatively more effective parameter for improving recombinant IFN-γ expression in CHO cells.

Web-based applications for building, managing and analysing kinetic models of biological systems

Mathematical modelling and computational analysis play an essential role in improving our capability to elucidate the functions and characteristics of complex biological systems such as metabolic, regulatory and cell signalling pathways. The modelling and concomitant simulation render it possible to predict the cellular behaviour of systems under various genetically and/or environmentally perturbed conditions. This motivates systems biologists/bioengineers/bioinformaticians to develop new tools and applications, allowing non-experts to easily conduct such modelling and analysis. However, among a multitude of systems biology tools developed to date, only a handful of projects have adopted a web-based approach to kinetic modelling. In this report, we evaluate the capabilities and characteristics of current web-based tools in systems biology and identify desirable features, limitations and bottlenecks for further improvements in terms of usability and functionality. A short discussion on software architecture issues involved in web-based applications and the approaches taken by existing tools is included for those interested in developing their own simulation applications.

In silico genome-scale modeling and analysis for identifying anti-tubercular drug targets

Mycobacterium tuberculosis is the deadly pathogen responsible for causing tuberculosis in humans, continuing to infect and kill millions of people globally. Despite the availability of a number of anti‐tuberculosis drugs and advances in high‐throughput drug discovery technology there is an urgent need for designing novel anti‐tubercular treatments due to growing parasite resistance and compromised immune systems in some patients. Therefore, it is highly necessary to develop systematic approaches that can facilitate the drug discovery by identification of drug targets in effective and efficient ways. In this sense, with the availability of whole genome sequence, application of genome‐scale modeling is becoming increasingly important for deriving rational drug target identification. This approach is indeed powerful in unraveling the metabolic behavior of pathogens and helps in identifying most relevant metabolites/genes as drug targets, which are experimentally testable. Herein, we present a review on the application of genome‐scale modeling and analysis in the context of identification of anti‐tubercular drug targets. Drug Dev Res 72: 121–129, 2011.  © 2010 Wiley‐Liss, Inc.