Insulin signaling network: mathematical modeling and parameter estimation from experimental data

The insulin signaling network (ISN) is an important metabolic network that, upon the insulin binding to its receptor at the cell surface, triggers the glucose uptake into the cell. The study of this mechanism within muscle cells, hepatocytes and cells of the adipose tissue is of major interest since it is crucial for understanding more clearly the factors that may induce the insulin resistance. However, the structure and the behaviour of the insulin signaling network are only partially known and the current research on this topic is fragmented into various lines of investigation. Because of the high degree of complexity of the ISN, it is diffcult to understand, without a theoretical framework, how the network responses evidenced from the experimental data determine the cell behaviour. In the present thesis, we proposed a detailed mathematical model of the ISN in order to investigate the factors that affect the basal concentrations and the dose-response curves (i.e., the steady state concentrations at given insulin levels) of the main components of the whole network. Our model concentrated particularly on single and double phosphorylation of Akt protein, and hypothesized the existence of a putative factor released by the small intestine that induces insulin resistance by activating the mammalian target of rapamycin complex 2 (mTORC2) in an insulin-independent manner and possibly operating through the IGF-1 receptor. Such hypothesis is based on clinical and experimental observations. The parameters of the ISN model were estimated from the experimental data of two skeletal muscle cell lines using a least squares approach. As the available data consisted in the equilibrium concentrations of many of the known signaling components at given values of the insulin, we derived the concentrations of the chemicals at the steady-state from the kinetic equations and then we implemented an algorithm that minimizes the distance between the model outputs and the data. For the numerical solution, we used a local optimization routine based on a derivative-free algorithm for bound constrained optimization. The ISN model was able to adequately fit the available experimental data. The model could thus become a useful tool to generate and test hypotheses, leading to a deeper understanding of the molecular mechanisms underlying insulin resistance and, in future perspective, to find drugs able to counterbalance the effects of this disease. Finally, as it is now widely recognized that Akt and mTOR complexes have a major role also in the regulation of cell proliferation, and then in cancer development, we combined the ISN model with a mathematical model that described the evolution of a AML (acute myeloid leukemia) cell population in order to investigate the effects of mTOR inhibitors with antitumor activity on the ISN and on the cell population response. Based on literature data of AML cell response to mTOR inhibitors with antitumor activity (the dual ATP-competitive mTOR inhibitor AZD8055), the two models provided simple relationships between the concentrations of proteins of the ISN and parameters representative of cell cycle progression and cell death.