Immunogenic cell death-related gene landscape predicts the overall survival and immune infiltration status of ovarian cancer.

Abstract

Background: Ovarian cancer (OC) is the most troubling malignant tumor of the female reproductive system. It has a low early diagnosis rate and a high tumor recurrence rate after treatment. Immunogenic cell death (ICD) is a unique form of regulated cell death that can activate the adaptive immune system through the release of DAMPs and cytokines in immunocompromised hosts and establish long-term immunologic memory. Therefore, this study aims to explore the prognostic value and underlying mechanisms of ICD-related genes in OC on the basis of characteristics. Methods: The gene expression profiles and related clinical information of OC were downloaded from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) database. ICD-related genes were collected from the Genecards database. ICD-related prognostic genes were obtained by intersecting ICD-related genes with the OC prognostic-related genes that were analyzed in the TCGA database. Functional enrichment, genetic mutation, and immune infiltration correlation analyses were further performed to identify underlying mechanisms. Subsequently, we developed a TCGA cohort-based prognostic risk model that included a nine-gene signature through univariate and multivariate Cox regression and LASSO regression analyses. Meanwhile, external validation was performed on two sets of GEO cohorts and the TCGA training cohort for three other common tumors in women. In addition, a nomogram was established by integrating clinicopathological features and ICD-related gene signature to predict survival probability. Finally, functional enrichment and immune infiltration analyses were performed on the two risk subgroups. Results: By utilizing nine genes (ERBB2, RB1, CCR7, CD38, IFNB1, ANXA2, CXCL9, SLC9A1, and SLAMF7), we constructed an ICD-related prognostic signature. Subsequently, patients were subdivided into high- and low-risk subgroups in accordance with the median value of the risk score. In multivariate Cox regression analyses, risk score was an independent prognostic factor (hazard ratio = 2.783; p < 0.01). In the TCGA training cohort and the two GEO validation cohorts, patients with high-risk scores had worse prognosis than those with low-risk scores (p < 0.05). The time-dependent receiver operating characteristic curve further validated the prognostic power of the gene signature. Finally, gene set enrichment analysis indicated that multiple oncological pathways were significantly enriched in the high-risk subgroup. By contrast, the low-risk subgroup was strongly related to the immune-related signaling pathways. Immune infiltration analysis further illustrated that most immune cells showed higher levels of infiltration in the low-risk subgroup than in the high-risk subgroup. Conclusion: We constructed a novel ICD-related gene model for forecasting the prognosis and immune infiltration status of patients with OC. In the future, new ICD-related genes may provide novel potential targets for the therapeutic intervention of OC.