Mining novel cell glycolysis related gene markers that can predict the survival of colon adenocarcinoma patients

Abstract

Colon adenocarcinoma (COAD) is a malignant gastrointestinal tumor, often occurring in the left colon, which is regulated by glycolysis-related processes. In past studies, multiple genes that influence the prognosis for survival have been discovered through bioinformatics analysis. However, the prediction of disease prognosis using a single gene is not an accurate method. In the present study, a mechanistic model was established to achieve better prediction for the prognosis of COAD. COAD-related data downloaded from The Cancer Genome Atlas (TCGA) were correlated with the glycolysis process using gene set enrichment analysis (GSEA) to determine the glycolysis-related genes that regulate COAD. Using COX regression analysis, glycolysis-related genes associated with the prognosis of COAD were identified, and the genes screened to establish a predictive model. The risk scores of this model were correlated with relevant clinical data to obtain a connection diagram between the model and survival rate, tumor characteristic data, etc. Finally, genes in the model were correlated with cells in the tumor microenvironment, finding that they affected specific immune cells in the model. Seven genes related to glycolysis were identified (PPARGC1A, DLAT, 6PC2, P4HA1, STC2, ANKZF1, and GPC1), which affect the prognosis of patients with COAD and constitute the model for prediction of survival of COAD patients.

Keywords: Colon adenocarcinoma; Glycolysis; Prognostic; Survival; mRNA; tumor microenvironmentsts.

Figure 1. Screen for glycolysis-related genes

(A) Enrichment plots of three gene sets that were significantly different (P<0.05) between normal and COAD tissues by performing GSEA. (B) Volcano map of 253 glycolytic genes expressed differentially in tumor and normal tissues (P<0.05 and logFc≠0).

Figure 2. The build of risk scores correlation model

(A) Analysis of survival prognosis of the constructed glycolysis-related models. Red represents the high-risk group and blue represents the low-risk group. (B) Receiver Operating Characteristic (ROC) curve of the glycolysis-related model. (C) Differences in risk scores for the two types of colon adenocarcinoma. (D) Relationship between the expression of genes of the glycolysis model, red representing a positive correlation, and blue representing a negative correlation. The values represent the degree of correlation.

Figure 3. Analysis of risk factors

(A) Univariate Cox regression analysis of the relationship between glycolysis and related clinical data. (B) Multivariate Cox regression analysis of the relationship between glycolysis and related clinical data. (C) Gene mutations that constitute the glycolysis model.

Figure 4. Analysis of mRNAs expression level

Expression of 7 mRNAs in colon adenocarcinoma and normal tissues. (A) The expression of ANKZF1 in tumors and normal tissue. (B) The expression of DLAT in tumors and normal tissue. (C) The expression of G6PC2 in tumors and normal tissue. (D) The expression of GPC1 in tumors and normal tissue. (E) The expression of P4HA1 in tumors and normal tissue. (F) The expression of PPARGC1A in tumors and normal tissue. (G) The expression of STC2 in tumors and normal tissue (*P<0.05, **P<0.01, ***P<0.001).

Figure 5. Analysis of biological function

Interaction network diagram of biological function analysis of the seven genes that constitute the glycolysis model. The modules represented by specific colors are displayed in the pie chart (*P<0.05, **P<0.01).

Figure 6. The seven‐mRNA signature associated with risk scores predicts overall survival in patients with colon adenocarcinoma

(A) mRNA risk score distribution in each patient. (B) Survival in days of colon adenocarcinoma patients in ascending order of risk scores. (C) Heatmap of the expression profile of the seven genes.

Figure 7. Kaplan–Meier survival analysis for COAD patients in TCGA data set

Kaplan–Meier survival analysis of clinical features and survival rate. Clinical features included (A) (age), (B) (grade), (C) (M), (D) (N), (E) (stage), and (F) (T).

Figure 8. K-M survival analysis for COAD patients in TCGA data set

K-M curves for prognosis of risk scores for the patients categorized by clinical feature. (A) (age), (B) (gender), (C) (M), (D) (N), (E) (stage), (F) (T).

Figure 9. Trend between risk score of the glycolysis model and the number of cells that comprise the microenvironment

(A)(Macrophage M0_CIBERSORT), (B) (Macrophage M0_CIBERSORT-ABS), (C) (Macrophage M1_CIBERSORT-ABS), (D) (Myeloid dendritic cell_TIMER), (E) (T-cell CD4+ memory resting_CIBERSORT).

Figure 10. Selection of the 10 groups (left) of samples with the highest risk scores and 10 groups (right) of samples with the lowest risk scores in correlation analysis

(A) The difference in cell numbers in the 20 samples. (B) Heat map of immune cell expression of the 20 sets of samples. (C) Immune cell correlation matrix, positive correlations displayed in purple, and negative correlations in gold.