Predicting the Gemcitabine efficacy by a stochastic language based model
P. Lecca (1). Kahramanoğulları (1); ), D. Morpurgo (1), C. Priami (1, 2)
(1) The Microsoft Research – University of Trento Centre for Computational and Systems Biology, Trento, Italy; (2) DISI – University of Trento, Italy
Objectives: This study is aimed at developing a discrete algorithmic model of biochemical action mechanisms of gemcitabine. The model has two main functions: (i) to predict the gemcitabine efficacy in terms of kinetic parameters governing the dynamics of drug phosphorilation and incorporation into growing DNA, and (ii) to provide a molecular interpretation of the observed tumor shrinkage curves.
Methods: BlenX [1] is the stochastic language that we used to specify the model. BlenX enables the specification of parallel and concurrent interactions; it is compositional as it allows the modeler to extend a model incrementally when new experimental data are available. Our model describes the metabolic reactions that gemcitabine undergoes, the competition between gemcitabine and the deoxycytidine triphosphate for the incorporation into the DNA, and the death and survival of the cell as a function of the amount of accumulated damage on the DNA. We calibrated the model on experimental time series data of the concentration of the main gemcitabine metabolites [2,3] with KInfer [4]. KInfer infers the kinetics by maximizing a Gaussian probability model of the observed time changes of the gemcitabine metabolites and substrates concentration.
Results: The simulations fit the observed dynamics of the metabolites and the tumor shrinkage curves, as reported in Lecca et al. [5]. The gemcitabine efficacy has been found to be directly proportional to the kinetic rate constants of the following reactions: phosphorilation of the gemcitabine, interaction of the ribonuclease reductase with gemcitabine diphosphate, incorporation of the gemcitabine triphosphate into the DNA. Conversely, the drug efficacy proved to be inversely proportional to the kinetic rates of the dephosphorilation of the gemcitabine monophosphate and of the gemcitabine deamination.
Conclusions: The BlenX model reproduces the experiments in [2,3] and predicts the drug efficacy value reported in [6]. Our approach to the pharmacodynamic modeling proposes an unconventional way to estimate the efficacy of the therapy. We built a modular multilevel model of the mechanism of drug action and identified the reactions giving the highest contribution to the toxicity of gemcitabine. In this context, the drug efficacy has an interpretation at molecular level in terms of the contribution of these reactions, whereas usually the drug efficacy is estimated as derivative of the tumor shrinkage with respect to the drug dose.
References:
[1] L. Dematté, C. Priami, and A. Romanel, "The BetaWorkbench: a computational tool to study the dynamics of biological systems," Briefings in Bioinformatics, vol. 9, no. 5, pp. 437-448, 2008.
[2] S. A. Veltkamp et al., "New insights into the pharmacology and cytotoxicity of gemcitabine and 2',2'-difluorodeoxyuridine," Mol Canc Ther, vol. 7, no. 8, pp. 2415-2425, 2008.
[3] V. Heineman et al., "Cellular elimination of 2',2'-difluorodeoxycytidine 5'-triphosphate: A mechanism of selfpotentiation," cancer research, vol. 52, pp. 533-539, 1992.
[4] P. Lecca, P. Palmisano, A. E. C: Ihekwaba, and C. Priami, "Calibration of dynamic models of biological systems with KInfer," European Journal of Biophysics, vol. 39, no. 6, p. 1019, 2010.
[5] P. Lecca, O. Kahramanogulları, D. Morpurgo, C. Priami, and R. A. Soo, "Modelling the tumor shrinkage pharmacodynamics with BlenX," in ICCABS 2011, Orlando, Florida, 2011.
[6] L.-S. Tham et al., "A Pharmacodynamic Model for the Time Course of Tumor Shrinkage by Gemcitabine + Carboplatin in Non-Small Cell Lung Cancer Patients," Clin Cancer Re, vol. 14, no. 13, pp. 4213-4218, 2008.