3. Information Flow in a Mammalian Signal Transduction Pathway

The mammalian signal transduction network relays detailed information about the presence and concentration of ligands on the outside of the cell to the nucleus, and alters cellular behaviour by changing gene expression. Since signal transduction pathways exhibit striking similarities to typical communication systems, the framework of information theory can be directly applied to better understand cellular signalling. During the current funding period of the priority program InKoMBio, we determined the information transmission capacities of the prototypic MAPK pathway using a combination of single cell experimentation and information theoretical calculations. Surprisingly, our results indicate that the signalling network transmits less than one bit of information. Rather than faithfully reporting extracellular concentrations of the ligand EGF, it responds in a binary manner. In addition, molecular noise interferes with a robust encoding of the presence of the input signal, limiting the information content even further. We observed similarly limited channel capacities for two other signalling pathways, the TGF\(\beta \)/SMAD and p53 networks. As many studies in different biological model systems suggest that cells can gain more information than 1-bit about their environment using signalling pathways, we aim to investigate what is limiting the information transmission capabilities at the single cell level and how cells maximise the amount of information gained from external and internal sources to ensure a proper physiological response. We hypothesise that the pathways integrate information from the cellular context, which could explain the apparently low-channel capacity. We therefore propose to use information theory, single cell experimentation and mathematical modelling to study the influence of contextual information, by addressing the following specific questions: (i) how does the state of a cell influence the response to an external signal, (ii) how does the context of previous stimuli influence the response and (iii) what are common principles of context-dependent signalling across different pathways? We will use live-cell imaging and immunofluorescence assays to measure signalling and context, and calculate the contribution of contextual information using conditional mutual information, context trees and parsimonious Bayesian networks. To gain a predictive understanding of the underlying molecular mechanisms, we will expand existing mathematical models of the pathways to include the interacting regulatory processes that provide context and analyse their information theoretical properties. Using network perturbations, we will experimentally validate model predictions.