5 signature genes revealed by single-cell profiling identified unique immune subtypes affecting the prognosis of ovarian cancer.

Abstract

Ovarian cancer (OC) ranks among the most prevalent gynecological malignancies, with surgery, chemotherapy, and immunotherapy constituting primary treatment modalities. However, despite advancements, immunotherapy, particularly immune checkpoint inhibitors, has yielded suboptimal outcomes. The pressing need to identify biomarkers predictive of clinical prognosis underscores our objective. We aim to discern gene signatures and establish prognostic subgroups, specifically in the context of immunotherapy and chemotherapy, guiding clinical decision-making. We used the Tumor Immunotherapy Gene Expression Resource (TIGER) and The Cancer Genome Atlas (TCGA) databases to extract signature genes of prognostic significance. Unsupervised consensus clustering was employed to classify patients based on these signature genes. The Tumor Immune Estimation Resource (TIMER) database, along with the R packages "maftools" and "ESTIMATE" facilitated immune infiltration estimation. Gene set variation analysis (GSVA) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis were implemented to probe immune-related cell signaling pathways among distinct subtypes. The Tumor Immune Dysfunction and Exclusion (TIDE) database was used to assess immunotherapy effects, while the R package "OncoPredict" evaluated drug sensitivity differences among subtypes. We identified five prognostically influential genes in ovarian cancer: IGFBP7, JCHAIN, CCDC80, VSIG4, and MS4A1. Utilizing these signature genes, we categorized TCGA-OV patients into five clusters, each associated with varying clinical prognoses. Notably, 2 clusters exhibited superior prognoses, accompanied by enhanced immune cell infiltration. KEGG enrichment analysis revealed their heightened enrichment in cellular immunity and immune cell interaction pathways. Given the elevated expression levels of multiple immune checkpoint molecules, these clusters may substantially benefit from immune checkpoint inhibitor therapy. Additionally, chemotherapy sensitivity analysis indicated their favorable responses to first or second-line chemotherapy regimens. We subclustered ovarian cancer patients by 5 signature genes obtained from the Single-cell RNA sequencing (scRNA-seq) dataset, which demonstrated a good typing effect. Patients in the two molecular subtypes showed better survival, higher immune cell infiltration, and higher drug sensitivity. This meticulous typing may help clinicians to quickly assess the prognosis of patients and the response to immunotherapy and chemotherapy.