An age stratified analysis of the biomarkers in patients with colorectal cancer

Abstract

Colorectal cancer (CRC), a common malignant tumor of the digestive tract, has a high incidence and mortality rate. Several recent studies have found that aging is associated with the increasing risk of cancer. Nevertheless, the expression status and function of age-related genes in CRC is still not well understood. In the study, we comprehensively analyzed the gene expression data of CRC patients from The Cancer Genome Atlas (TCGA) database. Age-related differential expression genes (age-related DEGs) in tumor tissues compared with normal tissues of CRC were further identified. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses of age-related DEGs were performed by clusterProfiler of R. Afterwards, we used the STRING database to map the protein-protein interaction network of DEGs. We constructed prognostic model through univariate and multivariate COX regression analyses, and further evaluated their predictive power. The prognostic gene signature-related functional pathways were explored by gene set enrichment analysis (GSEA). The weighted gene co-expression network analysis (WGCNA) was used to identify key module associated with two prognostic gene signatures. Finally, we used the Metascape to perform functional enrichment analysis of genes in the key module. A total of 279 age-related DEGs were identified from the TCGA database. GO and KEGG enrichment analysis showed that the age-related DEGs were enriched in the Modulation of chemical synaptic transmission and Neuroactive ligand-receptor interaction. Moreover, we established a novel age-related gene signature (DLX2 and PCOLCE2) for overall survival in CRC, which was further predicted in both the training and validation sets. The results of GSEA demonstrated that numerous disease-related pathways were enriched in the high-risk group. We identified 43 genes related to the DLX2 and PCOLCE2 by the WGCNA co-expression network. We also found that these 43 genes were enriched in the cancer-related pathways. To sum up, the study identified an age-related gene signature for predicting the prognosis of CRC patients, which is conducive to the identification of novel prognostic molecular markers.

Figure 1

Differentially expressed genes in the TCGA database. (A) The DEGs between CRC and normal samples. (B) The DEGs between young and old groups. (C) The Venn diagram describes 279 common differentially expressed genes. (D) PPI network of 279 common genes. (software: The STRING database, version number: 11.0, URL link: https://version-11-0.string-db.org/).

Figure 2

Functional enrichment analysis of DEGs. (A) The top 10 GO terms. (B) The top 10 KEGG pathways.

Figure 3

Combined DLX2 and PCOLCE2 genes in CRC overall survival prediction. (A) Kaplan–Meier survival curve of CRC patients between the low-risk group and the high-risk group. (B) 3- and 5-year ROC curve. (C, D) Risk score and survival status of the high-risk and low-risk groups. (E) The heat map of expression profile of the 2 prognostic genes. (software: pheatmap package in R, version number: 0.7.7, RUL link: https://cran.r-project.org/src/contrib/Archive/pheatmap/).

Figure 4

Validation of age related genes in CRC patients derived from TCGA dataset. (A) Kaplan–Meier survival curve of CRC patients between the low-risk group and the high-risk group. (B) 3- and 5-year ROC curve. (C, D) Risk score and survival status of the high-risk and low-risk groups. (E) The heat map of expression profile of the 2 prognostic genes. (software: pheatmap package in R, version number: 0.7.7, RUL link: https://cran.r-project.org/src/contrib/Archive/pheatmap/).

Figure 5

The relationship between clinicopathological factors and prognosis. (A) Univariate COX regression analysis forest map. (B) Multivariate COX regression analysis forest map. (C) A nomogram for predicting the probability of survival at 1, 3 and 5 years. (D) Nomogram calibration curve.

Figure 6

Gene set enrichment analysis with the 2 gene signature.

Figure 7

Weighted gene co-expression network analysis. (A) Soft-thresholding filtering. (B) Module screening, 20 modules have been identified. (C) The relationship between the 20 modules and the clinical traits. The numbers in rectangles indicate the correlation coefficient and the numbers in brackets indicate the P value. (D) Bar graph of enriched pathway of 43 genes in MElightcyan1 module.