The exploration of mitochondrial-related features helps to reveal the prognosis and immunotherapy methods of colorectal cancer

Abstract

Background: Cancer cell survival, proliferation, and metabolism are all intertwined with mitochondria. However, a complete description of how the features of mitochondria relate to the tumor microenvironment (TME) and immunological landscape of colorectal cancer (CRC) has yet to be made. We performed subgroup analysis on CRC patient data obtained from the databases using non-negative matrix factorization (NMF) clustering. Construct a prognostic model using the mitochondrial-related gene (MRG) risk score, and then compare it to other models for accuracy. Comprehensive analyses of the risk score, in conjunction with the TME and immune landscape, were performed, and the relationship between the model and different types of cell death, radiation and chemotherapy, and drug resistance was investigated. Results from immunohistochemistry and single-cell sequencing were utilized to verify the model genes, and a drug sensitivity analysis was conducted to evaluate possible therapeutic medicines. The pan-cancer analysis is utilized to further investigate the role of genes in a wider range of malignancies.

Methods and results: We found that CRC patients based on MRG were divided into two groups with significant differences in survival outcomes and TME between groups. The predictive power of the risk score was further shown by building a prognostic model and testing it extensively in both internal and external cohorts. Multiple immune therapeutic responses and the expression of immunological checkpoints demonstrate that the risk score is connected to immunotherapy success. The correlation analysis of the risk score provide more ideas and guidance for prognostic models in clinical treatment.

Conclusion: The TME, immune cell infiltration, and responsiveness to immunotherapy in CRC were all thoroughly evaluated on the basis of MRG features. The comparative validation of multiple queues and models combined with clinical data ensures the effectiveness and clinical practicality of MRG features. Our studies help clinicians create individualized treatment programs for individuals with cancer.

Keywords: colorectal cancer; immune landscape; mitochondrial; prognosis.

FIGURE 1

Differentially expressed genes (DEGs) extraction and identification of mitochondrial‐related genes (MRGs). (A) Volcano plot of the DEGs of MRGs; (B) Gene Ontology enrichment analysis; (C) Kyoto Encyclopedia of Genes and Genomes analysis of the related pathways; (D) Heatmap of the NMF clustering (k = 2); (E) Heatmap of differences in clinicopathological factors between the two different Clusters; (F, G) The Kaplan–Meier curve of overall survival and progression‐free survival rates among the Clusters. **p < .01; *p < .05.

FIGURE 2

Immune‐microenvironment analysis between the cluster subtypes. (A) The ImmuneScore results between the cluster subtypes; (B) The StromalScore results between the cluster subtypes; (C) The ESTIMATEScore results between the cluster subtypes; (D) The TumorPurity results between the cluster subtypes; (E–G) Analysis of the T cells, Endothelial cells, Fibroblasts infiltration among the cluster subtypes; (H) Sankey plot of the correlation between the cluster subtypes and immune subtypes.

FIGURE 3

Construction and validation of the mitochondrial‐related gene risk scores. (A, C, D, G, I) Risk distribution, survival status, related gene expression, overall survival (OS) rates, receiver operating characteristic (ROC) curves, and progression‐free survival rates between the two groups in TCGA cohort; (B, E, F, H) Risk distribution, survival status, related gene expression, OS rates, ROC curves of patients between the two groups in Gene Expression Omnibus cohort.

FIGURE 4

Prognosis stratified by mitochondrial‐related gene risk scores with clinicopathologic factors in colorectal cancer. (A—L) Subgroup analysis of the clinical data.

FIGURE 5

Identification of the clinical application value. (A) Clinical stratification comparison between the two groups; (B) Specific differences in the two groups in the N stage; (C–E) Risk scores predicted 1‐, 3‐, 5‐years for patients stratified by clinicopathologic factors. *p < .05.

FIGURE 6

Verification of the risk scores' predictive ability. (A, B) Validation of the prognostic effect of the risk scores; (C) Survival nomograms for the TCGA cohort, estimating OS rates; (D) Predictions of the nomogram's impact using its associated ROC curves; (E) The Calibration curve of the nomogram; (F)The DCA curves used to test the prediction accuracy of the nomogram. ***p < 0.001; **p < .01.

FIGURE 7

The prognostic properties of mitochondrial‐related gene (MRG) risk scores compared to other prognostic models. (A–C) Kaplan–Meier curves established by Qi, Du, and MRG; (D–F) The receiver operating characteristic curves established by Qi, Du, and MRG; (G) The Restricted Survival Time curves compared the prognostic effects of the three prognostic models; (H) The C‐index comparison of the three prognostic models.

FIGURE 8

Biological enrichment results based on prognosis model genes for mitochondrial‐related gene risk factors. (A, B) Gene Ontology enrichment results for the two groups; (C, D) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment results; (E) GSVA supplementation of KEGG pathway enrichment results.

FIGURE 9

Tumor microenvironment (TME) and immune cell infiltration predicted by the prognostic model. (A–D) StromalScore, immuneScore, ESTIMATEScore, and TumorPurity results for the TME; (E) Infiltration of immune cells and their relationship to the mitochondrial‐related gene risk score; (F) Differences in immune cell levels; (G) With the use of several algorithms, the heatmap displays the variation in immune cell expression. ***p < .001; **p < .01; *p < .05.

FIGURE 10

Forecasting the immune response using the mitochondrial‐related gene (MRG) risk score. (A) Analysis of immune‐related functional differences; (B) Immune checkpoint genes were differentially expressed; (C) Analysis of the connection between the MRG risk score and the immune checkpoint genes; (D– F) Key immunological checkpoint genes: correlation analysis results; (G) The TCIA analysis of the immunotherapy response results in the two groups; (H–K) The IPS analysis of the response. ***p < .001; **p < .01; *p < .05.

FIGURE 11

Association between mitochondrial‐related gene (MRG) risk score and microsatellite instability (MSI), tumor mutational burden, and cancer stem cells in colorectal cancer. (A, B) Connecting the MRG risk score to the MSI; (C) Correlation of MRG risk score and cancer stem cells; (D, E) Somatic mutation analysis via the MRG risk score on oncoplots.

FIGURE 12

Several genes are expressed differently. (A–E) Differential expression analysis of TIS‐, multidrug resistance‐, chemoradiotherapy‐, cuproptosis‐, and disulfidptosis‐related genes. ***p < .001; **p < .01; *p < .05; ns, non significant.