Construction and analysis of a prognostic risk scoring model for gastric cancer anoikis-related genes based on LASSO regression

Abstract

ObjectiveTo construct a prognostic risk scoring model for gastric cancer (GC) anoikis-related genes (ARGs) based on least absolute shrinkage and selection operator (LASSO) regression, and analyze the relationship between anoikis and prognosis of gastric cancer as well as the significance of the model in immunotherapy and chemotherapy of gastric cancer patients. MethodsDifferentially expressed prognostic anoikis-related genes (ARGs) in gastric cancer and adjacent tissues were screened through The Cancer Genome Atlas (TCGA) database and the Molecular Signatures Database (MSigDB); key anoikis genes were selected based on LASSO regression analysis to construct a prognostic risk scoring model, and patients were divided into high-risk and low-risk groups with the median risk score as the cutoff point. Gene expression levels in gastric cancer clinical samples and cells were detected by real-time quantitative PCR (RT-qPCR); Kaplan-Meier (KM) survival curves, univariate and multivariate Cox regression analyses were used to verify the predictive efficiency of the prognostic risk scoring model for the prognosis of gastric cancer patients; CIBERSORT and ESTIMATE algorithms were used to analyze the immune cell infiltration levels in patients with different risk groups; the correlation between risk scores and immune checkpoint expression levels in gastric cancer patients was analyzed using the R package "ggplot2" and "ggExtra", and the correlation between tumor mutation burden (TMB) and risk scores was assessed; chemotherapy drug sensitivity analysis was used to evaluate the value of the constructed prognostic risk scoring model in gastric cancer chemotherapy. ResultsSix key ARGs (VCAN, FEN1, BRIP1, CNTN1, P3H2, DUSP1) were screened out based on LASSO regression analysis, and a prognostic risk scoring model was constructed. RT-qPCR detection showed that VCAN, FEN1, and BRIP1 genes were highly expressed in gastric cancer tissues and cells (P < 0.05), while CNTN1, P3H2, and DUSP1 genes were lowly expressed (P < 0.05); Kaplan-Meier survival curve analysis found that the survival rate of patients in the high-risk group was significantly lower than that in the low-risk group (P < 0.001), and there were statistically significant differences in survival rate and TMB levels between high- and low-risk group patients (P < 0.05); univariate and multivariate analyses showed that tumor stage, age, and risk score were significantly associated with survival in gastric cancer patients (P < 0.05); immune cell infiltration analysis showed that the tumor immune matrix score was significantly higher in the high-risk group than in the low-risk group (P < 0.05), and the constructed prognostic risk scoring model can be used as an indicator of the tumor immune microenvironment (TIME) status; analysis using the R package "ggplot2" and "ggExtra" showed that the risk score of the constructed model was positively correlated with the upregulated expression of immune checkpoints TIM3, VISTA, TIGIT, BTLA, and B7-H3 (r = 0.26, 0.40, 0.16, 0.26, 0.21, P < 0.05), and TMB level was negatively correlated with risk score (R = – 0.4, P < 0.05); the constructed prognostic risk scoring model can be used to guide chemotherapy for gastric cancer patients. ConclusionThe prognostic risk scoring model constructed based on anoikis-related genes can be used to predict the prognosis of gastric cancer patients; the risk genes in the model can serve as potential targets for gastric cancer treatment, providing a reference for individualized treatment of gastric cancer.

Keywords

• gastric cancer
• anoikis
• the cancer genome atlas
• risk scoring model
• prognosis