Bioinformatics Analysis of RNA-seq Data Reveals Genes Related to Cancer Stem Cells in Colorectal Cancerogenesis

Abstract

Cancer stem cells (CSC) play one of the crucial roles in the pathogenesis of various cancers, including colorectal cancer (CRC). Although great efforts have been made regarding our understanding of the cancerogenesis of CRC, CSC involvement in CRC development is still poorly understood. Using bioinformatics and RNA-seq data of normal mucosa, colorectal adenoma, and carcinoma (n = 106) from GEO and TCGA, we identified candidate CSC genes and analyzed pathway enrichment analysis (PEI) and protein-protein interaction analysis (PPI). Identified CSC-related genes were validated using qPCR and tissue samples from 47 patients with adenoma, adenoma with early carcinoma, and carcinoma without and with lymph node metastasis and were compared to normal mucosa. Six CSC-related genes were identified: ANLN, CDK1, ECT2, PDGFD, TNC, and TNXB. ANLN, CDK1, ECT2, and TNC were differentially expressed between adenoma and adenoma with early carcinoma. TNC was differentially expressed in CRC without lymph node metastases whereas ANLN, CDK1, and PDGFD were differentially expressed in CRC with lymph node metastases compared to normal mucosa. ANLN and PDGFD were differentially expressed between carcinoma without and with lymph node metastasis. Our study identified and validated CSC-related genes that might be involved in early stages of CRC development (ANLN, CDK1, ECT2, TNC) and in development of metastasis (ANLN, PDGFD).

Keywords: RNA-seq; adenoma-carcinoma progression; bioinformatics analysis; cancer stem cells; colorectal cancer; differentially expressed genes; metastases; qPCR.

Figure 1

Venn diagrams of overlapping differentially expressed genes: (a) overlap between projects GSE50760 (blue color), GSE104836 (green color), and GSE100243 (red color); (b) overlap between project GSE50760 (green color) and GSE104836 (violet color). Legend: GSE: gene expression omnibus series.

Figure 2

Heatmap of top 100 DEGs between normal tissue and carcinomas. Legend: GEO, gene expression omnibus; GSE, gene expression omnibus series; group N, normal tissue; group C, carcinoma tissue; gene names are shown on the right side of the heatmap; downregulation is associated with blue color, no difference in expression is associated with white color, and upregulation is associated with red color.

Figure 3

Heatmap of top 100 DEGs between adenomas and carcinomas. Legend: GEO, gene expression omnibus; GSE, gene expression omnibus series; group A, adenoma; group C, carcinoma tissue; gene names are shown on the right side of the heatmap; downregulation is associated with blue color, no difference in expression is associated with white color, and upregulation is associated with red color.

Figure 4

Protein interaction network: (a) full protein–protein interaction network; (b) physical protein–protein interaction network. Legend: nodes are associated with proteins; edges are associated with protein–protein interactions.

Figure 5

Top 100 DEGs heatmap of expression between the primary tumor and normal samples. Legend: group Primary Tumor, carcinoma; group Solid Tissue Normal, normal tissue; gene names, where available, otherwise Ensembl Gene identificators are shown on the right side of the heatmap; downregulation is associated with blue color, no difference in expression is associated with white color, and upregulation is associated with red color.

Figure 6

Workflow of identification of differentially expressed genes in the bioinformatics analysis. Legend: N, normal colon tissue; A, adenomas; CRC, colorectal cancer; DEGs, differentially expressed genes; TCGA, the cancer genome atlas; R, R programming language; R studio, R studio programme; DAVID, The Database for Annotation, Visualization and Integrated Discovery; STRING; STRING database for protein–protein interaction analysis.

Figure 7

Expression of six genes (TNC, TNXB, PDGFD, CDK1, ANLN, ECT2) in colorectal cancerogenesis: (a) ∆Cq of six genes in normal mucosa, adenoma, and adenoma with early carcinoma; (b) ∆Cq of six genes in colorectal cancer without lymph node metastases (CRC N0) compared to corresponding normal mucosa; (c) ∆Cq of six genes in colorectal cancer with lymph node metastases (CRC N+) compared to corresponding normal mucosa; (d) ∆∆Cq of six genes in colorectal cancer with (CRC N+) compared to those without lymph node metastases (CRC N0). Legend: x, mean; ◦, outlier; *, p ≤ 0.05; **, p ≤ 0.01; ***, p ≤ 0.001.