Mammalian cell cultures constitute a major source for producing high-value bio-therapeutics such as monoclonal antibodies, recombinant proteins, viral vaccines, etc. The increasing demand for these commercially valuable compounds can be satisfied by developing high-yielding mammalian cell culture processes (1). The development of such cultures requires a good understanding of the metabolism and cellular response of the production host. However, the host's structural, functional, and dynamic complexities make it difficult to elucidate the behavior of the mammalian cellular system. In order to alleviate this difficulty, statistical techniques have been recently adopted for investigating correlations among the available experimental profiles (2). These techniques do not require the development of structured models and they use only a minimal set of experimental data, which makes them applicable to cell cultures with different operating modes. However, they still have limitations in explaining the effect of culture conditions on the internal cell metabolism. In this sense, in silico analyses using genome-scale models of organisms are invaluable in characterizing the cellular physiology and internal metabolism. A combined approach involving statistical and genome-scale analyses can be crucial in elucidating the effect of environmental conditions and its impact on internal cell metabolism. However, such techniques have received limited attention in the literature. They offer promise for identifying optimal feeding strategies and potential cell engineering targets as well as facilitating media optimization for the enhanced production of bio-therapeutics in mammalian cells. Furthermore, using such a systematic approach may also reduce the number of wet experiments.

In this work, we present a combined statistical and in silico genome-scale analysis for mouse cell lines producing mAb to elucidate their physiological and metabolic states under different environmental conditions and genetic manipulations. First, we applied multivariate statistical data analysis techniques on experimental fermentation data to identify the relationships among nutrient uptakes and desired products. The analysis identified significant correlations between mAb production and certain amino acids. Subsequently, we expanded and curated an existing genome-scale metabolic model (3) of mouse cells using updated biochemical and genomic data. The internal metabolic activities were then explored to elucidate the effect of positive correlation obtained from statistical analysis. The effect of culture media on intracellular metabolism was also explored further (4). The in silico simulation results revealed interesting results on ATP utilization, metabolic shifting and nutrient utilization. Thus we see that a combined approach involving statistical analysis and in silico genome-scale analysis can be very effective for analyzing the cellular metabolism.

References:

1)Wurm FM. 2004. Production of recombinant protein therapeutics in cultivated mammalian cells. Nat Biotechnol 32:1393-1398.

2)Alwis DMD, Dutton RL, Scharer J, Moo-Young M. 2007. Statistical methods in media optimization for batch and fed-batch animal cell culture. Bioprocess Biosyst Engg, 30:107-113.

3)Sheikh K, Forster J, Nielson LK. 2005. Modeling hybridoma cell metabolism using a generic genome-scale metabolic model of Mus musculus. Biotech. Prog. 21:112-121.

4)Selvarasu. S., Lee, D-Y., Wong, V. V. T., Karimi, I. A. Elucidation of metabolism in hybridoma cells grown in fed-batch culture by genome-scale modeling. Submitted