(714a) Optimal Experimental Design for Isotopic Tracing Experiments of Tumour Metabolism

13C tracing is becoming more and more popular as a practically unique method that gives access to intracellular metabolic flux profiles under various conditions of cell incubation in situ [1]. Metabolic flux profiling is used for various biomedical applications. For instance, it enhances the understanding of differences between normal pathologically transformed cells; it gives indications of metabolic consequences of application of drugs that increase stress resistance of normal cells or kill cancer cells. In this way 13C- based metabolic profiling has a perspective to be used for the development of novel strategies in therapy. The efficient analysis of 13C labelling experiments requires detailed simulations to compute metabolic fluxes, which underlie a specific measured distribution of 13C isotopic isomers (isotopomers) in the products and intermediates of cellular metabolism. For this purpose, the computation of isotopomer dynamics is essential in order to understand the underlying metabolic mechanisms, which results in the construction of large-scale dynamic systems modeled with a home-made powerful computational tool [2].

Nevertheless, measurement of isotopomers distributions possibly requires a combination of different methodologies resulting in very expensive and time-consuming tasks. Furthermore, the efficiency of the subsequently constructed models and, obviously their predictive capabilities largely depend on the substrates labeled, the initial labeling and the metabolites, intermediates, products and isotopic isomers measured. For this purpose classical sensitivity analysis and experimental design-oriented techniques could be of immense help. In this work we exploit a recently developed methodology [3] based on the Proper Orthogonal Decomposition technique [4] which is used directly as a novel sensitivity analysis tool, essentially coupling model reduction and optimal experimental design to identify metabolites as well as families of isotopomers exhibiting different levels of sensitivity for different experimental conditions/different cells. This information is used as a tool to effectively guide experimental isotopomer measurements. The relevance of this methodology is investigated for isotopic labeling kinetic models spanning a range of initial labeling conditions as well as different labeled substrates and experimental guidelines are extracted. Furthermore, the applicability of this method to compute optimal time points at which (after starting the incubation with labelled substrates) the samples should be taken is discussed.

References

[1] Vizán P, Sánchez-Tena S, Alcarraz-Vizán G, Soler M, Messeguer R, Pujol MD, Lee WN, Cascante M. (2009). Carcinogenesis 30:946-52.

[2] Selivanov VA, Marin S, Lee PW, Cascante M. (2006) Bioinformatics 22:2806-12. Epub 2006 Sep 25. PubMed PMID: 17000750.

[3] Alana J., Theodoropoulos C. (2010) Comput. Chem. Eng. http://dx.doi.org/10.1016/j.compchemeng.2010.04.014