Testing models of insulin binding in rat adipocytes using network thermodynamic computer simulations

Abstract

Many different models have been proposed to explain the complex binding behavior of insulin to its receptor, but a systematic comparison of models with experimental results is lacking. We have used network thermodynamic computer simulations to compare models of insulin binding against the results of several experimental tests designed to differentiate between the models. Six models of insulin binding were tested (simple, diffusion-reaction, conversion, dissociation, heterogeneous site, and two-step intramembrane) against results reported in the literature for isolated rat adipocytes. Although still a matter of experimental controversy, the criteria selected for modeling were curvilinear Scatchard plots, bi-or multi-exponential dissociation, insulin-accelerated dissociation, lack of dependence of the overall dissociation constant on receptor number, and receptor reserve. Using a given set of parameter values most appropriate for each model, none was able to account for all of the observed experimental results. This indicates both the complexity of the binding reaction and the need for further model development. The approach of using computer simulations to systematically test models against experimental results affords not only insight into the critical features of a model enabling it to pass a test, but also indicates potential experiments which might differentiate between models.