Identification of Prognostic Biomarker Signatures and Candidate Drugs in Colorectal Cancer: Insights from Systems Biology Analysis

Abstract

Colorectal cancer (CRC) is the second most common cause of cancer-related death in the world, but early diagnosis ameliorates the survival of CRC. This report aimed to identify molecular biomarker signatures in CRC. We analyzed two microarray datasets (GSE35279 and GSE21815) from the Gene Expression Omnibus (GEO) to identify mutual differentially expressed genes (DEGs). We integrated DEGs with protein⁻protein interaction and transcriptional/post-transcriptional regulatory networks to identify reporter signaling and regulatory molecules; utilized functional overrepresentation and pathway enrichment analyses to elucidate their roles in biological processes and molecular pathways; performed survival analyses to evaluate their prognostic performance; and applied drug repositioning analyses through Connectivity Map (CMap) and geneXpharma tools to hypothesize possible drug candidates targeting reporter molecules. A total of 727 upregulated and 99 downregulated DEGs were detected. The PI3K/Akt signaling, Wnt signaling, extracellular matrix (ECM) interaction, and cell cycle were identified as significantly enriched pathways. Ten hub proteins (ADNP, CCND1, CD44, CDK4, CEBPB, CENPA, CENPH, CENPN, MYC, and RFC2), 10 transcription factors (ETS1, ESR1, GATA1, GATA2, GATA3, AR, YBX1, FOXP3, E2F4, and PRDM14) and two microRNAs (miRNAs) (miR-193b-3p and miR-615-3p) were detected as reporter molecules. The survival analyses through Kaplan⁻Meier curves indicated remarkable performance of reporter molecules in the estimation of survival probability in CRC patients. In addition, several drug candidates including anti-neoplastic and immunomodulating agents were repositioned. This study presents biomarker signatures at protein and RNA levels with prognostic capability in CRC. We think that the molecular signatures and candidate drugs presented in this study might be useful in future studies indenting the development of accurate diagnostic and/or prognostic biomarker screens and efficient therapeutic strategies in CRC.

Keywords: biomarkers; candidate drugs; colorectal cancer; differentially expressed genes; drug repositioning; protein–protein interaction; reporter biomolecules; systems biology.

Figure 1

The integrative analytical pipeline employed in the present study. (A) The colorectal cancer (CRC) datasets were analyzed under the Bioconductor platform in R. We used linear models for microarray data (LIMMA) to detect the differentially expressed genes (DEGs) in CRC compared to normal samples. (B) Gene ontology (GO) terms and molecular pathways were identified by DEGs enrichment via the Database for Annotation, Visualization and Integrated Discovery (DAVID). (C) The hub proteins were identified by protein–protein interaction (PPI) analysis. (D) The reporter feature algorithm was used to identify reporter biomolecules as transcriptional regulatory elements. (E) The survival analysis of the hub biomolecules was done through The Cancer Genome Atlas (TCGA) CRC datasets via SurvExpress and oncomiR. (F) The candidate drug molecules were identified by Connectivity Map (cMap) and geneXpharma.

Figure 2

Identification of differentially expressed genes (DEGs) in colorectal cancer (CRC) from microarray CRC datasets: (A) the upregulated genes in the CRC expression profiling datasets; (B) the downregulated genes in the CRC expression profiling datasets.

Figure 3

The significant pathways altered in colorectal cancer: (A) upregulated pathways in colorectal cancer; (B) downregulated pathways in colorectal cancer.

Figure 4

The survival analysis of the hub genes in the prognosis of colorectal cancer. (A) The box-plot represents the differential expression of the 10 hub genes in two risks groups. (B) The Kaplan–Meier plot represents the prognostic ability of the hub gene signatures in CRC.

Figure 5

The survival assessment of the reporter transcription factor (TF) signatures in the prognosis of colorectal cancer. (A) The box-plot represents the differential expression of the 10 TFs between both risk groups. (B) The Kaplan–Meier plot represents the prognostic power of the TF signatures in colorectal cancer.

Figure 6

The survival analysis of the reporter microRNA (miRNA) signatures in colorectal cancer. The Kaplan–Meier plot represents the prognostic ability of miRNA signatures (miR-193b-3p and miR-615-3p) in colorectal cancer.

Figure 7

Drug repositioning results in colorectal cancer. (A) Classification of repurposed drugs according to drug development stages. (B) Distribution of approved drugs into anatomical therapeutic chemical drug classes.