Exploration of Reduced Mitochondrial Content-Associated Gene Signature and Immunocyte Infiltration in Colon Adenocarcinoma by an Integrated Bioinformatic Analysis

Abstract

Purpose: Mitochondrial dysfunction refers to cancer immune evasion. A novel 7-gene prognostic signature related to the mitochondrial DNA copy number was utilized to evaluate the immunocyte infiltration in colon cancer according to the risk scores and to predict the survival for colon cancer. Experimental design: We performed an integrated bioinformatic analysis to analyze transcriptome profiling of the EB-treated mitochondrial DNA-defected NCM460 cell line with differentially expressed genes between tumor and normal tissues of COAD in TCGA. The LASSO analysis was utilized to establish a prognostic signature. ESTIMATE and CIBERSORT validated the differences of immunocyte infiltration between colon cancer patients with high- and low-risk scores. Results: Our study identified a 7-gene prognostic signature (LRRN2, ANKLE1, GPRASP1, PRAME, TCF7L1, RAB6B, and CALB2). Patients with colon cancer were split into the high- and low-risk group by the risk scores in TCGA (training cohort: HR = 2.50 p < 0.0001) and GSE39582 (validation cohort: HR = 1.43 p < 0.05). ESTIMATE and CIBERSORT revealed diverseness of immune infiltration in the two groups, especially downregulated T-cell infiltration in the patients with high-risk scores. Finally, we validated the colon patients with a low expression of the mitochondrial number biomarker TFAM had less CD3⁺ and CD8⁺ T-cell infiltration in clinical specimens. Conclusion: An mtDNA copy number-related 7-gene prognostic signature was investigated and evaluated, which may help to predict the prognosis of colon cancer patients and to guide clinical immunotherapy via immunocyte infiltration evaluation.

Keywords: Bioinformatic analysis; Colon Cancer; Gene signature; Immunocyte infiltration; mitochondiral DNA.

FIGURE 1

DEGs identified in the EB-treated NCM460 cell line. (A) Heatmap of the top 200 DEGs by |logFC|. The color from blue to red indicates low to high expression level. (B) Volcano map between EB-treated and control groups of the NCM460 cell line. (C,D) Enriched terms in GO analysis of DEGs in EB-treated NCM460, respectively. (E,F) Enriched terms in KEGG analysis of DEGs in EB-treated NCM460 DEGs, respectively; DEGS means differentially expressed genes; logFC means log fold change.

FIGURE 2

Intersection of the mtDNA-related gene and prognostic gene in TCGA (A). Heatmap of the top 200 DEGs identified between tumor and compared normal tissues from COAD in TCGA based on the value of |logFC| (B). Intersection of 475 common DEGs with the prognostic value in TCGA and EB-treated NCM460 using the Venn diagram. Top three hub modules were classified in MODE. Red circles represented upregulated genes, while blue circles represented downregulated genes. (C,D) Heatmap of GO enriched terms and biological process (BP) GO enriched terms. (E) PPI network of the common 475 genes calculated by STRING software from the Metascape website (F) Top eight modules of the PPI network; GO, gene ontology; PPI, protein–protein interaction [Color figure can be viewed at wileyonlinelibrary.com].

FIGURE 3

Construction and validation of the mtDNA-related prognostic signature. Assessment of LASSO regression analysis of 475 DGEs is shown here. (A) Risk score, (C) survival status, and (E) profiles of mtDNA-associated gene expression. The value of the mtDNA-related prognostic signature was validated in GSE39582 by the (B) risk score, (D) survival status, and (F) mtDNA-related gene expression profiles, and GSE39582 was displayed in Figure 4.

FIGURE 4

Survival analysis and ROC analysis of the 7-gene prognostic signature. Kaplan–Meier survival curves of OS (A,B), OS predictive nomogram (C,D), univariate (E,F) and multivariate Cox proportional hazards regression (G,H) and ROC curves (I,J) in patients with colon carcinoma in TCGA and GSE39582.

FIGURE 5

Construction and certification of mtDNA-related patients groups in colon carcinoma samples. (A) Binary risk group constructed by LASSO regression analysis of 475 DGEs and further validated using the ESTIMATE algorithm. (B–E) Differences of scores calculated by the ESTIMATE between the high-risk and low-risk group. (F) Boxplot shows the expression of HLA family genes. (G–H) Expressions of PD-L1 (CD274) and CTLA4. (*:p < 0.05 **:p < 0.01 ***: p < 0.001).

FIGURE 6

Immune infiltration conditions between the two mtDNA-related groups in colon carcinoma samples. Immune infiltration of high- (A) and low-(B) risk group calculated using the CIBERSORT algorithm was shown in the two barplot, respectively. (C–I) Differences of tumor-infiltrating immune cells during the two groups. Seven subtypes of tumor-infiltrating immune cells between the high-risk and low-risk group showed significant differences (p <0.05). (J) Radar plot shows the whole survey of CIBERSORT analysis (*:p < 0.05 **:p < 0.01 ***: and p < 0.001).

FIGURE 7

Verification of the correlation between T-cell infiltration and mitochondrial copy number. Results of immunohistochemistry of 39 colon patients are shown here. (A,B,D,E) were paired IHC tissues. As shown in results, tissues with a higher TFAM expression [left of (A,B,D,E)] have a stronger CD3 [right of (A,B)] and CD8 [right of (D,E)] expression. (C,F) shows positive correlation relationship of CD3 and CD8 with the TFAM expression (p = 0.0031 and 0.0237, respectively). (G) Results of the mRNA expression of seven signature genes in TFAM-knockdown SW480 cells by qPCR.