(815g) Phenotypic Heterogeneity in the Epithelial to Mesenchymal Transition (EMT) Is Controlled By the NFAT and Phosphorylated-Sp1 Transcription Factors

The epithelial to mesenchymal transition (EMT) is important in development and pathological processes such as fibrosis and cancer. Several extracellular signals trigger EMT, central amongst these are soluble transforming growth factor β (TGFβ) family members. TGFβ isoforms are potent inducers of EMT in both embryonic and pathological conditions. Receptor mediated signaling in response to TGFβ triggers a program that ultimately represses the expression of epithelial genes such as E-cadherin, while simultaneously activating the expression of mesenchymal genes such as Vimentin. The EMT signaling program is complex, containing structural features such as redundancy, feedback and crosstalk. While these architectural features promote robustness, they complicate the understanding of signal flow and rational signal reprogramming. In this study, we developed a dynamic network model of TGFβ induced EMT signaling. Our model used mass action kinetics within an ordinary differential equation (ODE) framework to describe the EMT program initiated by TGFβ isoforms. The interaction network contained 995 protein or mRNA components interconnected through 1700 interactions. A family of model parameters (1700 kinetic constants and 56 non-zero initial conditions) was estimated using 41 protein signaling data data sets generated in DLD1 colon carcinoma, MDCKII and A375  melanoma cells using the Pareto optimal ensemble technique (POETs) multiobjective optimization algorithm. POETs identified more than 15,000 likely TGFβ/EMT models, from which we selected a low-correlation population of approximately 1000 models for analysis. Analysis of the model population suggested that both MCF10A and DLD1 cells could exhibit phenotypic heterogeneity if treated simultaneously with TGFβ2 and VEGFA. Furthermore, the model population predicted that the NFAT and phosphorylated Sp1 transcription factors were key elements driving this heterogeneity. This heterogeneity was characterized by a significant fraction of the population being in a “mixed state” having both high E-cadherin and high Vimentin expression. We tested these predictions experimentally using qRT-PCR and flow-cytometry studies in the presence and absence of NFAT inhibitors. These studies verified that upwards of 45% of the cellular population could be put into the mixed state in the presence of both TGFβ2 and VEGF. Moreover, this response depended upon both the MAPK activation of Sp1 and NFAT action (consistent with the molecular signaling predicted by the model).  Lastly, we found that the heterogenous population had significantly different functional behavior at the cellular level. For example, the extent of ductal branch formation significantly increased with MCF10A cells in the mixed phenotype, compared to cells treated with VEGFA alone. Together, these results establish a predictive mechanistic model of EMT susceptibility, and reveal a novel signaling axis between phosphorylated Sp1/NFAT which could possibly regulate carcinoma progression through an EMT versus tubulogenesis response.