Genes involved in TGFβ1-driven epithelial-mesenchymal transition of renal epithelial cells are topologically related in the human interactome map

Abstract

Background: Understanding how mesenchymal cells arise from epithelial cells could have a strong impact in unveiling mechanisms of epithelial cell plasticity underlying kidney regeneration and repair. In primary human tubular epithelial cells (HUTEC) under different TGFβ1 concentrations we had observed epithelial-to-mesenchymal transition (EMT) but not epithelial-myofibroblast transdifferentiation. We hypothesized that the process triggered by TGFβ1 could be a dedifferentiation event. The purpose of this study is to comprehensively delineate genetic programs associated with TGFβ1-driven EMT in our in vitro model using gene expression profile on large-scale oligonucleotide microarrays. Results: In HUTEC under TGFβ1 stimulus, 977 genes were found differentially expressed. Thirty genes were identified whose expression depended directly on TGFβ1 concentration. By mapping the differentially expressed genes in the Human Interactome Map using Cytoscape software, we identified a single scale-free network consisting of 2630 interacting proteins and containing 449 differentially expressed proteins. We identified 27 hub proteins in the interactome with more than 29 edges incident on them and encoded by differentially expressed genes. The Gene Ontology analysis showed an excess of up-regulated proteins involved in biological processes, such as "morphogenesisquot;, "cell fate determination" and "regulation of development", and the most up-regulated genes belonged to these categories. In addition, 267 genes were mapped to the KEGG pathways and 14 pathways with more than nine differentially expressed genes were identified. In our model, Smad signaling was not the TGFβ1 action effector; instead, the engagement of RAS/MAPK signaling pathway seems mainly to regulate genes involved in the cell cycle and proliferation/apoptosis. Conclusion: Our present findings support the hypothesis that context-dependent EMT generated in our model by TGFβ1 might be the outcome of a dedifferentiation. In fact: 1) the principal biological categories involved in the process concern morphogenesis and development; 2) the most up-regulated genes belong to these categories; and, finally, 3) some intracellular pathways are involved, whose engagement during kidney development and nephrogenesis is well known. These long-term effects of TGFβ1 in HUTEC involve genes that are highly interconnected, thereby generating a scale-free network that we named the "TGFβ1 interactome", whose hubs represent proteins that may have a crucial role for HUTEC in response to TGFβ1. © 2007 Campanaro et al; licensee BioMed Central Ltd.