(601f) Computational Analysis of the Compartmentalization of Phosphatase-Mediated Regulation of the Epidermal Growth Factor Receptor

Cell signaling via the epidermal growth factor receptor (EGFR) is broadly involved in development, normal physiology, and pathology. Phosphorylated EGFR (pEGFR) levels, which dictate receptor signaling strength and persistence, are the net result of two opposing processes, auto-phosphorylation of cytoplasmic tyrosines by the receptor kinase and dephosphorylation of those tyrosines by protein tyrosine phosphatases. Several studies have concluded that pEGFR dynamics are regulated by phosphatases mainly in the cell interior, subsequent to receptor endocytosis. Here, we challenge this view through the development of a novel mechanistic model of EGFR phosphorylation. Our model is based on a set of coupled ordinary differential equations that describe receptor-level processes including ligand binding, competitive binding between ATP and kinase inhibitors, receptor dimerization, tyrosine phosphorylation, receptor dephosphorylation at the cell membrane and in the cell interior, and receptor trafficking. The list of model reactions was chosen to minimize the number of model parameters and species, while still incorporating key molecular rate processes. We found that model results were in general unchanged by expanding the model scope to include molecular species that might be predicted to form but which were excluded from the model. The few model parameters which were not taken from the literature were estimated by performing a multi-parameter fit of the model to several different types of experimental data taken from the literature using a simulated annealing algorithm. Whereas virtually all previous models of pEGFR dynamics involve fits to measurements of relative changes in pEGFR levels in response to ligand alone, we simultaneously fit our model to data of pEGFR dynamics in response to ligand, phosphatase inhibitors, or EGFR kinase inhibitors. In addition to these dynamic data, the model was also fit to measurements of basal pEGFR levels and actual percentages of total EGFR phosphorylated for particular conditions. Finally, the fit also involved an approach to recapitulate measurements of EGFR endocytosis rate constants which is more consistent with the definition of the experimentally measured rate constant than has been implemented in previous models. Use of this broad collection of different types of data for pEGFR levels represents a significant departure from previous modeling approaches. Comparison of our fitted parameters to those from previous studies revealed several noteworthy results. First, the difference between our fitted constants for receptor dimerization kinetics in the presence or absence of ligand was significantly greater than that reported in previous modeling studies, a difference which results mainly from our incorporation of kinetic data for receptor dephosphorylation in the fit. A more surprising result was that the fitted rate constants for phosphatase kinetics suggested that receptor dephosphorylation may actually occur more rapidly at the cell surface than in the cell interior. Based on these fitted parameters, the model further predicted that such high phosphatase activity at the cell surface may have a profound effect on EGFR endocytosis kinetics. Specifically, by reducing the concentration of EGFR phosphotyrosines at the cell surface, which are a primary driving force for receptor internalization, phosphatase activity at the cell surface was predicted to slow the net rate of EGFR endocytosis. This prediction is qualitatively consistent with a recent report of phosphatase-mediated regulation of EGFR degradation in cancer cells. Not surprisingly, only phosphatase activity at the cell surface was predicted to affect EGFR endocytosis rates. Thus, whereas the activities of phosphatases at the cell surface and in the cell interior may both regulate pEGFR levels directly through receptor dephosphorylation, phosphatase activity at the cell surface may also determine the cellular distribution of EGFR. Overall, our results suggest that the dephosphorylation of EGFR occurs on a time scale which is much shorter than is generally appreciated and that phosphatases may regulate higher level receptor processes such as trafficking in ways which have not been fully explored. Further investigation of these phenomena may reveal new avenues for the rational modulation of EGFR signaling dynamics for therapeutic purposes.