ChBE Seminar Series: Jeff Varner
Tuesday, May 8, 2012
11:00 a.m.-12:00 p.m.
Room 2110 Chemical and Nuclear Engineering Bldg.
Professor Panagiotis Dimitrakopoulos
Modeling and Analysis of the Core Architecture Regulating TGFß Induced Epithelial to Mesenchymal Transition (EMT)
School of Chemical and Biomolecular Engineering
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 E-cadherin, while simultaneously activating the expression of mesenchymal genes such as Vimentin. EMT signaling pathways are complex, containing structural features such as redundancy, feedback and crosstalk. While these architectural features ensure robustness, they complicate the understanding of signal flow and rational signal reprogramming. Systems biology is an essential tool for understanding the operation and regulation of intracellular networks, such as those responsible for initiating EMT. In this study, we developed a first-generation 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 signaling and gene expression 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 experimental 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. Signal flow, sensitivity, and robustness analysis, suggested three important levels of regulation controlling TGFβ induced EMT. First, the AP1 and SP1 transcription factors played differential roles during EMT induction. Deletion of AP1 enhanced E-cadherin expression, while decreasing Vimentin expression. On the other hand, SP1 deletion enhanced Vimentin expression with little or no effect on E-cadherin. Second, overexpression of ERK-specific phosphatases increased the level of phosphorylated Smad2, which enhanced the transition to a mesenchymal phenotype (instead of the expected epithelial phenotype). Lastly, we explored factors regulating the availability of LEF1 and its ability to regulate the shift to a mesenchymal phenotype. Taken together, these results provided insight into the core molecular architecture of TGFβ induced EMT, and revealed possible operational paradigms of phenotypic conversion.