Identification of Malignant Cell Populations Associated with Poor Prognosis in High-Grade Serous Ovarian Cancer Using Single-Cell RNA Sequencing

SIMPLE SUMMARY: Ovarian cancer has a high recurrence rate (~75%), and tumor heterogeneity is associated with such tumor recurrence. However, it is still poorly understood in ovarian cancer. To reveal tumor heterogeneity, we performed single-cell RNA sequencing (RNA-seq) of serous ovarian cancer cells from four different patients: two with primary carcinoma, one with recurrent carcinoma, and one with carcinoma treated with interval debulking surgery. As a result, we found two malignant tumor cell subtypes associated with poor prognosis. One malignant population included the earliest cancer cells and cancer stem-like cells. SLC3A1 and PEG10 were identified as the marker genes of cancer-initiating cells. The other malignant population expressing CA125 (MUC16), the most common biomarker for ovarian cancer, is associated with a decrease in the number of tumor-infiltrating cytotoxic T lymphocytes (CTLs). Our findings will offer new markers for diagnosis and choosing treatments targeting the malignant populations in ovarian cancer. ABSTRACT: To reveal tumor heterogeneity in ovarian cancer, we performed single-cell RNA sequencing (RNA-seq). We obtained The Cancer Genome Atlas (TCGA) survival data and TCGA gene expression data for a Kaplan–Meier plot showing the association of each tumor population with poor prognosis. As a result, we found two malignant tumor cell subtypes associated with poor prognosis. Next, we performed trajectory analysis using scVelo and Monocle3 and cell–cell interaction analysis using CellphoneDB. We found that one malignant population included the earliest cancer cells and cancer stem-like cells. Furthermore, we identified SLC3A1 and PEG10 as the marker genes of cancer-initiating cells. The other malignant population expressing CA125 (MUC16) is associated with a decrease in the number of tumor-infiltrating cytotoxic T lymphocytes (CTLs). Moreover, cell–cell interaction analysis implied that interactions mediated by LGALS9 and GAS6, expressed by this malignant population, caused the CTL suppression. The results of this study suggest that two tumor cell populations, including a cancer-initiating cell population and a population expressing CA125, survive the initial treatment and suppress antitumor immunity, respectively, and are associated with poor prognosis. Our findings offer a new understanding of ovarian cancer heterogeneity and will aid in the development of diagnostic tools and treatments.