Based on RNA modification-related programmed cell death signature predicts prognosis and immunotherapy in patients with ovarian cancer

Ovarian cancer (OC) is the most common cause of cancer-related mortality in gynecology. It is crucial to develop novel therapeutic strategies to achieve effective long-term disease stabilization for this malignant tumor. In this study, we evaluated prognostic genes associated with RNA modification (RM) andProgrammed cell death (PCD) in OC and explored the underlying mechanisms of these genes through enrichment analysis and drug sensitivity analysis. Additionally, based on single-cell RNA sequencing data of OC, we further investigated the association between these genes and the efficacy of immunotherapy for OC. The OC-related datasets, TCGA-OC, GSE63885, and E-MTAB-8107, were obtained from the public databases. Firstly, the Weighted gene co-expression network analysis (WGCNA) was performed to identify the module genes related to RM, which was intersected with DEGs and PCDRGs. Subsequently, the optimal combination was selected from the 101 machine learning algorithms in TCGA-OC and verified in GSE63885. Key genes were subsequently identified from this optimal combination, and their risk coefficients were calculated to construct a risk model using the risk score formula.Then, genes with significant overall survival differences based on Kaplan-Meier (KM) curve were selected as prognostic genes. Finally, a single cell-RNA sequencing dataset, E-MTAB-8107, was used to explore the expression levels in different cell types and identify key cells. The cell-cell communication and pseudotime analysis were used to study the heterogeneity of key cells. Additionally, the clinic samples were collected to verify the expression of prognostic genes by real-time qPCR and Immunohistochemistry. In TCGA-OC, a sum of 122 candidate genes were screened, and then 12 feature genes related to prognosis were screened in the univariate Cox analysis. Afterwards, survival SVM (C-index = 0.607) was determined as the optimal combination generating a risk model. Specially, the high-risk subgroups had worse prognosis. Then, HERC1, TPCN2, EPHA2, CASP2, and BLOC1S1 were identified as prognostic genes. In E-MTAB-8107, a total of 12 different cell types were annotated. The fibroblasts had strong interaction with monocytes. In real-time qPCR and IHC assay, the expression level of CASP2 was significantly up-regulated in the OC group. Overall, HERC1, TPCN2, EPHA2, CASP2, and BLOC1S1 were mined as potential therapeutic targets, and fibroblasts were determined as key cells in OC, which provides new perspectives for the treatment of OC.