Evaluation of information flows in the RAS-MAPK system using transfer entropy measurements

Abstract

The RAS-MAPK system serves as a prominent signal processing network within animal cells, controlling various cell fates such as proliferation, differentiation, and cell death. Dysregulation of this system has been identified in genetic diseases and cancer affecting diverse tissues. To better understand the dynamics of this system, we employed information flow analysis based on transfer entropy (TE) between the activation dynamics of two key elements: SOS, a guanine nucleotide exchanger for the small GTPase RAS, and RAF, a RAS effector serine/threonine kinase. TE analysis enables a model-free assessment of the direction and strength of the control capability between components within the network. Here, we analyzed the information flows HeLa cells stimulated with EGF. Notably, we observed a switching of information flows during the progression of signal transduction. In the early phase, the dominant regulator was the forward information flow from SOS to RAF, contingent on the EGF dose. However, in the late phase, the EGF-dependency diminished, and mutual regulation between SOS and RAF took precedence. TE analyses on cells treated with a MEK inhibitor or expressing a SOS mutant associated with Noonan syndrome, a human genetic disease revealed anomalies in the information flow, leading to decreased followability of RAF inputs. Our findings demonstrate that reaction network analysis based on TE holds significant promise for applications in molecular pharmacology and pathology.SignificanceIntracellular signal processing relies on intricate regulatory networks composed of various molecular reactions. Ordinary differentiation equation (ODE) modeling, which proposes the reaction formula and parameter values, is a conventional approach to analyzing intracellular reaction networks. In contrast, statistical analyses focus on correlations between the reactions, complementing ODE modeling. Among these, TE analysis is advantageous in dealing with the reaction dynamics. The dynamics of molecular activation estimated using the ODE models do not directly mean the control capability between the molecules, a factor that TE excels at evaluating. In this study, we analyzed the dynamics of control capabilities between SOS and RAF. The analysis has revealed context-specific information flows within the cell.