A Novel Network Profiling Analysis Reveals System Changes in Epithelial-Mesenchymal Transition

Human Genome Center, Institute of Medical Science, University of Tokyo, Minato-ku, Tokyo, Japan.
PLoS ONE (Impact Factor: 3.23). 06/2011; 6(6):e20804. DOI: 10.1371/journal.pone.0020804
Source: PubMed


Patient-specific analysis of molecular networks is a promising strategy for making individual risk predictions and treatment decisions in cancer therapy. Although systems biology allows the gene network of a cell to be reconstructed from clinical gene expression data, traditional methods, such as bayesian networks, only provide an averaged network for all samples. Therefore, these methods cannot reveal patient-specific differences in molecular networks during cancer progression. In this study, we developed a novel statistical method called NetworkProfiler, which infers patient-specific gene regulatory networks for a specific clinical characteristic, such as cancer progression, from gene expression data of cancer patients. We applied NetworkProfiler to microarray gene expression data from 762 cancer cell lines and extracted the system changes that were related to the epithelial-mesenchymal transition (EMT). Out of 1732 possible regulators of E-cadherin, a cell adhesion molecule that modulates the EMT, NetworkProfiler, identified 25 candidate regulators, of which about half have been experimentally verified in the literature. In addition, we used NetworkProfiler to predict EMT-dependent master regulators that enhanced cell adhesion, migration, invasion, and metastasis. In order to further evaluate the performance of NetworkProfiler, we selected Krueppel-like factor 5 (KLF5) from a list of the remaining candidate regulators of E-cadherin and conducted in vitro validation experiments. As a result, we found that knockdown of KLF5 by siRNA significantly decreased E-cadherin expression and induced morphological changes characteristic of EMT. In addition, in vitro experiments of a novel candidate EMT-related microRNA, miR-100, confirmed the involvement of miR-100 in several EMT-related aspects, which was consistent with the predictions obtained by NetworkProfiler.

10 Reads
  • Source
    • "Mining of gene functions associated with the core invasiveness genes suggests a tight link between CIG expression and EMT, a process in which tumour cells lose their epithelial phenotype to acquire a more mesenchymal phenotype. THBS1, FN1, CTGF and E-Cadherin are bona fide markers of EMT [52]. LYN is reported to be a top-ranked EMT signature gene and RNAi-mediated knockdown of LYN inhibited cell migration and invasion [53]. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Metastases remain the primary cause of cancer-related death. The acquisition of invasive tumour cell behaviour is thought to be a cornerstone of the metastatic cascade. Therefore, gene signatures related to invasiveness could aid in stratifying patients according to their prognostic profile. In the present study we aimed at identifying an invasiveness gene signature and investigated its biological relevance in breast cancer. We collected a set of published gene signatures related to cell motility and invasion. Using this collection, we identified 16 genes that were represented at a higher frequency than observed by coincidence, hereafter named the core invasiveness gene signature. Principal component analysis showed that these overrepresented genes were able to segregate invasive and non-invasive breast cancer cell lines, outperforming sets of 16 randomly selected genes (all P<0.001). When applied onto additional data sets, the expression of the core invasiveness gene signature was significantly elevated in cell lines forced to undergo epithelial-mesenchymal transition. The link between core invasiveness gene expression and epithelial-mesenchymal transition was also confirmed in a dataset consisting of 2420 human breast cancer samples. Univariate and multivariate Cox regression analysis demonstrated that CIG expression is not associated with a shorter distant metastasis free survival interval (HR = 0.956, 95%C.I. = 0.896-1.019, P = 0.186). These data demonstrate that we have identified a set of core invasiveness genes, the expression of which is associated with epithelial-mesenchymal transition in breast cancer cell lines and in human tissue samples. Despite the connection between epithelial-mesenchymal transition and invasive tumour cell behaviour, we were unable to demonstrate a link between the core invasiveness gene signature and enhanced metastatic potential.
    PLoS ONE 02/2014; 9(2):e89262. DOI:10.1371/journal.pone.0089262 · 3.23 Impact Factor
  • Source
    • "KLF5 activates CDH1 expression and downregulation of VIM in non-small cell lung cancer cells [40]. Both KLF4 and KLF5 show competition for promoter binding sites with antagonizing effects regarding proliferation [41]. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Krüppel-like factors (KLFs) comprise a highly conserved family of zinc finger transcription factors, that are involved in a plethora of cellular processes, ranging from proliferation and apoptosis to differentiation, migration and pluripotency. During the last few years, evidence on their role and deregulation in different human cancers has been emerging. This review will discuss current knowledge on Krüppel-like transcription in the epithelial-mesenchymal transition (EMT), invasion and metastasis, with a focus on epithelial cancer biology and the extensive interface with pluripotency. Furthermore, as KLFs are able to mediate different outcomes, important influences of the cellular and microenvironmental context will be highlighted. Finally, we attempt to integrate diverse findings on KLF functions in EMT and stem cell biology to fit in the current model of cellular plasticity as a tool for successful metastatic dissemination.
    Oncotarget 11/2013; 5(1). DOI:10.18632/oncotarget.1456 · 6.36 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: A very important topic in systems biology is developing statistical methods that automatically find causal relations in gene regulatory networks with no prior knowledge of causal connectivity. Many methods have been developed for time series data. However, discovery methods based on steady-state data are often necessary and preferable since obtaining time series data can be more expensive and/or infeasible for many biological systems. A conventional approach is causal Bayesian networks. However, estimation of Bayesian networks is ill-posed. In many cases it cannot uniquely identify the underlying causal network and only gives a large class of equivalent causal networks that cannot be distinguished between based on the data distribution. We propose a new discovery algorithm for uniquely identifying the underlying causal network of genes. To the best of our knowledge, the proposed method is the first algorithm for learning gene networks based on a fully identifiable causal model called LiNGAM. We here compare our algorithm with competing algorithms using artificially-generated data, although it is definitely better to test it based on real microarray gene expression data.
Show more

Preview (2 Sources)

10 Reads
Available from