Construction and validation of a prognostic model for colon cancer based on anoikis-related genes
10.3760/cma.j.cn115355-20240719-00362
- VernacularTitle:基于失巢凋亡相关基因的结肠癌预后模型构建及验证
- Author:
Tao ZHANG
1
;
Ziyao LI
1
;
Yingying SUN
1
;
Boyang LI
1
;
Zhao WANG
1
;
Zhifu YANG
1
Author Information
1. 中国中医科学院西苑医院外一科，北京 100091
- Publication Type:Journal Article
- Keywords:
Colonic neoplasms;
Anoikis;
Prognosis;
Tumor microenvironment;
Immunity
- From:
Cancer Research and Clinic
2025;37(1):55-63
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To construct and validate a prognostic model of colon cancer based on differentially expressed anoikis-related genes, and to preliminarily investigate the relationship between anoikis-related genes and the tumor immune microenvironment of colon cancer.Methods:A total of 472 cancer tissues samples of patients with colon cancer, RNA sequencing data and clinical data of 41 normal tissues samples were downloaded from the Cancer Genome Atlas (TCGA) database between the establishment time and July in 2024. A total of 919 genes related to anoikis were screened out from GeneCards database, and the common genes were selected from the RNA sequencing gene datasets of colon cancer and normal colon tissues in the TCGA database, among which the differentially expressed anoikis-related genes of colon cancer and normal colon tissues were screened out based on P < 0.05. Furthermore, genes related to the prognosis of 446 colon cancer patients with prognostic data in the TCGA database were screened by using univariate Cox proportional risk model; the genes with P < 0.05 were further screened out and a colon cancer prognosis model was constructed by using LASSO-Cox proportional risk model. The risk score of the above 446 colon cancer patients in the TCGA database was calculated according to the prognostic model, and the patients were divided into high-risk (≥ median value) group and low-risk (< median value) group according to the median risk score, and the overall survival of the 2 groups was analyzed by using the Kaplan-Meier method. The risk score based on R software-based time ROC program package was used to predict 1-year, 2-year, 3-year overall survival therapeutic efficacy of colon cancer patients in the TCGA database. According to the median risk score of colon cancer patients in the TCGA database, the patients in the International Cancer Genome Consortium (ICGC) database were divided into high-risk group and low-risk group. Kaplan-Meier method and receiver operating characteristic (ROC) curve were used to verify the predictive effect of the prognostic model. The differentially expressed genes between low-risk group and high-risk group stratified by prognostic model risk score in the TCGA database were used to perform single sample gene set enrichment analysis (ssGESA) of immune cells and immune function by using R software related programs. The differences in risk scores of patients with different immunophenotypes (including inflammator response type, wound healing type, interferon gamma dominant type and lymphocyte depletion type) were compared; and correlation analysis of infiltration and risk scores between immune cells and stromal cells in tumor microenvironment was made. Based on the tumor immune function and rejection (TIDE) database, the relationship between the prognostic model risk score and programmed death receptor ligand 1 (PD-L1) gene expression level was analyzed. Results:Based on anoikis-related genes in the GeneCards database, 236 differentially expressed anoikis-related genes between colon cancer tissues and normal tissues were obtained from the TCGA database. LASSO Cox regression was applied to establish a prognostic model constructed by 7 differentially expressed anoikis-related genes in cancer tissues and normal colon tissues related to the prognosis of colon cancer. Risk score = 0.366×TIMP1-0.404×NAT1+0.207×LTB4R2+0.075×INHBB+0.140×CD36-0.109×MMP3+2.994×OFCC1. The median risk score of 446 colon cancer patients in the TCGA database was 1.754 719 545. Survival analysis showed that the overall survival of colon cancer patients in high-risk group of the TCGA database was worse than that in low-risk group ( P < 0.001); ROC curve analysis showed that the area under the curve for predicting 1-year, 2-year and 3-year overall survival of patients in the TCGA database based on the prognostic model risk score was 0.705, 0.731 and 0.723, respectively. Kaplan-Meier method analysis showed that in the ICGC database, the overall survival of colon cancer patients in high-risk group was worse than that in low-risk group ( P = 0.041); ROC curve analysis showed that the area under the curve of prognostic model risk score for predicting 1-year and 2-year overall survival of colon cancer patients in the ICGC database was 0.663 and 0.966, respectively. ssGESA analysis showed that macrophage level in high-risk group was higher than that in low-risk group, helper T (Th) 1 cell and Th2 cell levels in high-risk group were lower than those in low-risk group (all P < 0.01). In terms of immune function, the cell killing activity and histocompatibility complex Ⅰ level in high-risk group were lower than those in low-risk group, and type Ⅱ interferon response score in high-risk group was higher than that in low-risk group (all P < 0.05). The analysis of immunophenotype showed that the risk score of inflammatory response type was higher than that of wound healing type ( P < 0.05), and there was no statistically significant difference in risk score between the other 2 types (all P > 0.05). Risk score was positively correlated with stromal cell infiltration score ( R = 0.340, P < 0.001) and immune cell infiltration score ( R = 0.148, P < 0.05); the expression level of PD-L1 in high-risk group was higher than that in low-risk group in the TCGA database ( P = 0.048), and the expression level of PD-L1 was positively correlated with risk score ( R = 0.130, P = 0.009). Conclusions:A prognostic model of colon cancer constructed by anoikis-related genes can better predict the prognosis of colon cancer patients, and anoikis-related genes may play an important role in tumor immunity of colon cancer.
