Analysis of the Molecular Network Regulating Epithelial-Mesenchymal Transition

Epithelial to mesenchymal transition (EMT) has been associated with diseased conditions in adult tissues such as  fibrosis [1] and cancer [2].  This phenomenon occurs when epithelial cells differentiate towards mesenchymal lineage and invade the underlying matrix.  The accumulation of these cells creates an environment producing aberrant extracellular matrix and other components at the site of chronic inflammation.  Further progression leads to infiltration of the surrounding tissue and ultimately metastasizes to distant sites.  The fate processes unique to these cells are dependent upon their morphological appearance and phenotype.  Typical epithelium is a sheet of cells connected by cell-cell tight junctions and inhibits the movement of individual cells away from the monolayer.  This layer is polarized so that apical and basal surfaces are dependent on their function such as adhering to the basal lamina or securing the vessel of blood flow.  Mesenchymal cells, however, are phenotypically different and exhibit neither an oriented structure nor tight intracellular adhesion.  Movement within these cells is quite different since they can migrate or invade into the surrounding tissue while leaving the trailing region behind [3].  The mechanisms underlying this phenomenon reside in the molecular machinery and regulate these specific phenotypes.  Commonly used molecular markers for EMT include the expression of E-cadherin, VE-cadherin, Alpha Smooth Muscle Actin (α-SMA), Vimentin, and nuclear localization of B-catenin.  Transcription factors such as Snail, Slug, and Twist are also present and inhibit E-cadherin expression [4].  

Interestingly, EMT is a natural phenomenon in embryonic development and necessary for the formation tissues such as atrioventricular cushions (precursor to valves) and palate-fusion.  Atrioventricular cushion formation is a complex process involving the coordinated response endocardial cells [5].  At the beginning of HH14, myocardial cells lining these regions secrete factors such as BMP2, TGFB2, and TGFB3 that activate the overlaying endocardium to undergo EMT [6].   During this complex phenomenon, the endocardial cells shift from an epithelial to mesenchymal phenotype by reducing cell-cell contacts, increasing cell-matrix adhesion proteins, and invade into the underlying matrix.  Specifically, BMP2 and TGFB2 regulate E-cadherin expression by activating tight junction repressors Snail and Slug via MAPK signaling.  This repression releases B-catenin from the cell membrane and forms complexes with TCF-4 directly up-regulating the expression of TGFB3 [7].  TGFB3 further promotes EMT through SMAD signaling by regulating the levels of LEF-1 (which also complexes with B-catenin) [8, 9].  This complex directly effects the expression of mesenchymal markers such as α-SMA and Vimentin.  TGFB2 and TGFB3 have also been shown to further stabilize the EMT transition by signaling through PI3K.   PI3K signaling occurs through AKT, which can phosphorylate and inactivate GSK-3B.  This inhibition is critical as GSK-3B is a protein that targets both Snail and B-catenin for degradation through the ubiquitin proteosome pathway [10].  Upon completion, cushions consisting of mesenchymal cells and an outer layer of endocardial cells are present.       

Although EMT has been observed in both normal development and diseased tissue, fundamental differences have been noticed.   These events have been described as subtle and controlled movements in embryogenesis while aggressive and uncontrolled movements in tumorigenesis [11].  The mechanisms distinguishing these states have yet been answered although canonical signaling between the two appears conserved.  Detailed molecular studies of EMT may provide additional markers and allow further investigation into the role of metastability.   Metastability indicates the existence of cells with features of both epithelial and mesenchymal features.  This may be important for identifying critical expression levels in both embryonic and diseased regulation.  However, understanding the complex multi-scale and temporal regulation of EMT is a challenging task.  Multiple cytokines, extracellular matrix receptors, and mechanics have all been shown important and synergizing the cross talk within these signaling pathways is improbable [12, 13].  Therefore, using a systems biology approach could potentially elucidate some of the key underlying features that regulate EMT in a normal developmental environment.  Our approach was therefore to model the transcriptional regulation of EMT to capture the dynamic effects of multiple cytokines.  Sensitivity analysis and simulation protocols will hopefully reveal cross-talk mechanisms between pathways to expose synergistic or inhibitory effects and present targets for therapeutic applications.    

In this study, we constructed and analyzed a mathematical model integrating multiple signaling pathways regulating the epithelial to mesenchymal phenomenon.  Of particular interest was the role of Transforming Growth Factor Beta and Notch signaling within atrioventricular cushion EMT.  The objective of this study was to investigate which signaling systems were critical to EMT and conduct simulations to evaluate synergistic or inhibitory effects between the signaling pathways.  The model was formulated as a set of ordinary differential equations which described 992 species and 1770 interactions, including both the mRNA and protein levels for key species.  The system was also compartmentalized to incorporate nuclear transport proteins within the nucleus and cytoplasm.  Model parameters to estimate the magnitude and duration of EMT were taken from the literature. A population approach was chosen to constrain model parameters and to estimate the impact of parametric uncertainty on model predictions.  These populations were constrained by Pareto optimization and cross-validation methods so that minimal error with greatest predictability was achieved.  Within the model, we were able to statistically capture the role of TGFB isoforms against literature values.  TGFB2 was shown to repress E-cadherin expression through Snail and Slug expression but, was also dependent upon TGFb3 activation of the Smad2-LEF1-Smad4 complex.  TGFB3 further inhibited GSK-3B through phosphorylation allowing Snail and Slug effectively stay active.  Target genes necessary for mesenchymal transition, were assessed upon treatment of TGFB3 and many were directly regulated by the Bcatenin-LEF1 complex.  Synergistic effects of TGFB2 and Notch revealed increased Snail expression while Notch alone only produced Slug expression.  A sensitivity analysis of network components using a coupling approach revealed the importance of both MAPK and SMAD signaling.  Our model predictions were confirmed using RT-PCR for measuring gene expression and IHC to localize proteins of interest. 

1.             Kalluri, R. and E.G. Neilson, Epithelial-mesenchymal transition and its implications for fibrosis. J Clin Invest, 2003. 112(12): p. 1776-84.

2.             Thiery, J.P., Epithelial-mesenchymal transitions in tumour progression. Nat Rev Cancer, 2002. 2(6): p. 442-54.

3.             Hay, E.D., EMT concept and examples from the vertebrate embryo, in Rise and Fall of Epithelial Phenotype: Concepts of Epithelial-Mesenchymal Transition, P. Savagner, Editor. 2005, Springer: Berlin.

4.             Thiery, J.P. and J.P. Sleeman, Complex networks orchestrate epithelial-mesenchymal transitions. Nat Rev Mol Cell Biol, 2006. 7(2): p. 131-42.

5.             Person, A.D., S.E. Klewer, and R.B. Runyan, Cell biology of cardiac cushion development. Int Rev Cytol, 2005. 243: p. 287-335.

6.             Boyer, A.S., et al., TGFbeta2 and TGFbeta3 have separate and sequential activities during epithelial-mesenchymal cell transformation in the embryonic heart. Dev Biol, 1999. 208(2): p. 530-45.

7.             Liebner, S., et al., Beta-catenin is required for endothelial-mesenchymal transformation during heart cushion development in the mouse. J Cell Biol, 2004. 166(3): p. 359-67.

8.             Nawshad, A. and E.D. Hay, TGFbeta3 signaling activates transcription of the LEF1 gene to induce epithelial mesenchymal transformation during mouse palate development. J Cell Biol, 2003. 163(6): p. 1291-301.

9.             Nawshad, A., et al., TGFbeta3 inhibits E-cadherin gene expression in palate medial-edge epithelial cells through a Smad2-Smad4-LEF1 transcription complex. J Cell Sci, 2007. 120(Pt 9): p. 1646-53.

10.           Medici, D., E.D. Hay, and B.R. Olsen, Snail and Slug promote epithelial-mesenchymal transition through beta-catenin-T-cell factor-4-dependent expression of transforming growth factor-beta3. Mol Biol Cell, 2008. 19(11): p. 4875-87.

11.           Lee, J.M., et al., The epithelial-mesenchymal transition: new insights in signaling, development, and disease. J Cell Biol, 2006. 172(7): p. 973-81.

12.           Shook, D. and R. Keller, Mechanisms, mechanics and function of epithelial-mesenchymal transitions in early development. Mech Dev, 2003. 120(11): p. 1351-83.

13.           Quaranta, V., et al., Mathematical modeling of cancer: the future of prognosis and treatment. Clin Chim Acta, 2005. 357(2): p. 173-9.