An integrative modeling framework reveals plasticity of TGF-β signaling

Background: The TGF-beta transforming growth factor is the most pleiotropic cytokine controlling a broad range of cellular responses that include proliferation, differentiation and apoptosis. The context-dependent multifunctional nature of TGF-beta is associated with complex signaling pathways. Differential models describe the dynamics of the TGF-beta canonical pathway, but modeling the non-canonical networks constitutes a major challenge. Here, we propose a qualitative approach to explore all TGF-beta-dependent signaling pathways. Results: Using a new formalism, CADBIOM, which is based on guarded transitions and includes temporal parameters, we have built the first discrete model of TGF-beta signaling networks by automatically integrating the 137 human signaling maps from the Pathway Interaction Database into a single unified dynamic model. Temporal property-checking analyses of 15934 trajectories that regulate 145 TGF-beta target genes reveal the association of specific pathways with distinct biological processes. We identify 31 different combinations of TGF-beta with other extracellular stimuli involved in non-canonical TGF-beta pathways that regulate specific gene networks. Extensive analysis of gene expression data further demonstrates that genes sharing CADBIOM trajectories tend to be co-regulated. Conclusions: As applied here to TGF-beta signaling, CADBIOM allows, for the first time, a full integration of highly complex signaling pathways into dynamic models that permit to explore cell responses to complex microenvironment stimuli.