Vanderbilt Institute of Chemical Biology



Discovery at the VICB







Signature of Aggressive Colorectal Cancer


By: Carol A. Rouzer, VICB Communications
Published: November 12, 2014



Bioinformatics analysis reveals role for NFATc1-regulated genes in colorectal cancer metastasis.


Colorectal cancer (CRC) is a major cause of cancer-related deaths in the United States. Despite significant advances in treating CRC, correct staging of the disease, the identification of patients most likely to suffer metastasis or recurrence, remains a challenge. Attempts to use gene expression profiling to identify signatures of aggressive disease have provided some clues, but no unified functional insights have so far emerged from this approach. These ongoing challenges led VICB member Bing Zhang, his Vanderbilt collaborator Daniel Beauchamp, and their laboratories to take a comprehensive approach to discover a functionally relevant signature of metastatic potential in CRC. Their research identifies a transcription factor that plays an important role as a modulator of CRC metastasis, a finding with important with prognostic significance [M.K. Tripathi, et al. (2014) Cancer Res., published online October 15, DOI: 10.1158/0008-5472.].




Figure 1. Identification of genes associated with metastatic potential in colorectal cancer (CRC). Human CRC gene expression datasets were assembled and validated. Then correlation analysis identified gene pairs that were likely to be co-expressed. GO semantic similarities were used to assign functional relevance to the pairs. An integrated co-expression network was then derived from the functional relevance data. Modules of functionally related genes were identified and filtered to focus on genes that are up-regulated in highly metastatic murine CRC cells. The resulting metastasis-related networks were then analyzed to identify transcription factors common to the up-regulated genes. The results identified the NFAT family, and further experiments verified the importance of NFATc1 for CRC invasiveness and patient survival.


The central hypothesis driving the Zhang and Beauchamp labs' work was that deregulation of co-regulated gene networks in cancer cells has functional consequences that affect patient prognosis and that this same network deregulation can be observed in mouse models of metastatic disease. Their goal was to combine patient data with data from appropriate CRC mouse models to identify signatures of the most important deregulated networks. They started their work by accumulating 11 human microarray datasets comprising data from 1,295 CRCs. They used these datasets to first identify pairs of genes that are highly likely to be co-expressed in CRC and then employed gene ontology (GO) biological process annotation to determine the functional similarities between members of each pair. They combined the functional similarity and co-expression data to calculate a log likelihood ratio (LLR), which was a measure of functional relevance for each gene pair. In general, LLR increased as the probability that a gene pair is co-expressed increased.


The investigators next created a co-expression network containing all gene pairs with an LLR greater than 1. They visualized the network by representing each gene as a circle, called a node, attached to every co-expressed gene by a line, called an edge. The network contained 2,285 genes and 13,083 edges. Within the network, they identified 441 modules containing functionally related genes. The five largest modules contained 187, 107, 105, 85, and 72 genes and were related to developmental processes, cell cycle regulation, translational elongation, immune system processes, and unknown functions, respectively.


Having established a large co-expression network for human CRC, the investigators now turned to a mouse model of the disease. They used parental MC-38 colon carcinoma cells (MC-38Par) with low metastatic potential and a highly metastatic cell line (MC-38Met) derived from MC-38Par cells. They evaluated each gene in their human CRC co-expression network to determine if that gene was up- or down-regulated in MC-38Met versus MC-38Par. The results indicated that the 187 gene module was enriched for up-regulated genes while the 107 gene module was enriched with down-regulated genes in MC-38Met cells as compared to the parental cell line. These findings were consistent with reduced proliferation and increased cell motility in the MC-38Met cells, characteristics that are typical of epithelial-to-mesenchymal transition, a process strongly associated with cancer metastasis.


Focusing on the 187 gene module, the investigators identified 63 core genes that were enriched in the highly metastatic cells. Of these, 21 were targets of the NFAT (nuclear factor of activated T cell) family of transcription factors and were involved in processes such as skeletal and muscular system development, connective tissue development, and regulation of cell movement, growth, and proliferation. These findings led them to search for expression of NFAT family members in 22 freshly frozen stage II and III primary CRCs. They found that the mRNAs for 19 out of the 21 NFAT genes were upregulated in these tumors. More careful examination revealed that in both the MC-38Met cells and the human tumors, NFATc1 was the most differentially expressed member of the NFAT family.


Both a broad NFAT family small molecule inhibitor and specific knockdown of NFATc1 using RNAi reduced invasiveness and expression of metastasis-associated target genes in MC-38Met cells. Consistently, RNAi-mediated knockdown of NFATc1 reduced invasiveness of the highly metastatic HCT116 human CRC cell line, while overexpression of NFATc1 increased invasiveness of the poorly metastatic HT29 human CRC cell line. These results confirmed that the role of NFATc1 in metastasis applied to human as well as mouse CRC. To test the role of NFATc1 in CRC invasiveness in vivo, the investigators injected MC-38Par cells transfected with the gene for NFATc1 or empty vector into the spleens of mice and monitored degree of invasiveness on the basis of spread of tumor to the liver. MC-38Par cells expressing NFATc1 were much more invasive than those transfected with empty vector. Similarly, MC-38Met cells transfected with an shRNA directed against NFATc1 were much less invasive in this model than MC-38Met cells transfected with a scrambled shRNA. Together the results strongly support a role for NFATc1-mediated transcription in CRC cell invasiveness and metastatic potential.


Increased activity of NFAT family members in CRC may result from over-expression of the transcription factor or from over-activation of normal levels of the transcription factor. Thus, expression levels of NFATc1 alone cannot reliably be used to screen for up-regulation of its target genes. The investigators therefore looked for gene expression patterns that could serve as a signature for NFATc1 activation. In MC-38Met cells, they identified 8 out of the 21 NFAT-regulated genes that were consistently up-regulated. These genes coded for Angiopoeitin (ANGPTL2), Asporin (ASPN), Collagen 3A1 (COL3A1), Matrix Gla protein (MGP), Mannose receptor (MRC2), Fibroblast activating protein (FAP), Polymerase I and transfer release factor (PTRF), and Twist-related protein 1 (TWIST1). The investigators evaluated the expression of these genes in two clinically annotated cohorts of stage II and stage III CRCs. They found that tumors that expressed high levels of the 8 signature genes were associated with reduced overall survival, reduced disease-specific survival, and reduced disease-free survival when compared to tumors that expressed low levels of the genes. These results confirm the clinical relevance of the NFATc1 pathway to CRC prognosis.


This new discovery of the role for NFATc1-mediated transcription regulation in CRC metastasis provides a valuable tool that can be used by clinicians to predict the prognosis and guide therapy of CRC patients. It also provides critical new insight in to CRC pathogenesis and malignant behavior.








The Vanderbilt Institute of Chemical Biology, 896 Preston Building, Nashville, TN 37232-6304, phone 866.303 VICB (8422), fax 615 936 3884
Vanderbilt University is committed to principles of equal opportunity and affirmative action. Copyright © 2014 by Vanderbilt University Medical Center