| Literature DB >> 31825847 |
Kate Lawrenson1, Marcos A S Fonseca2, Annie Y Liu2, Felipe Segato Dezem2, Janet M Lee3, Xianzhi Lin2, Rosario I Corona4, Forough Abbasi4, Kevin C Vavra4, Huy Q Dinh5, Navjot Kaur Gill6, Ji-Heui Seo7, Simon Coetzee3, Yvonne G Lin8, Tanja Pejovic9, Paulette Mhawech-Fauceglia10, Amy C Rowat6, Ronny Drapkin11, Beth Y Karlan2, Dennis J Hazelett3, Matthew L Freedman7, Simon A Gayther12, Houtan Noushmehr13.
Abstract
Fallopian tube secretory epithelial cells (FTSECs) are likely the main precursor cell type of high-grade serous ovarian cancers (HGSOCs), but these tumors may also arise from ovarian surface epithelial cells (OSECs). We profiled global landscapes of gene expression and active chromatin to characterize molecular similarities between OSECs (n = 114), FTSECs (n = 74), and HGSOCs (n = 394). A one-class machine learning algorithm predicts that most HGSOCs derive from FTSECs, with particularly high FTSEC scores in mesenchymal-type HGSOCs (padj < 8 × 10-4). However, a subset of HGSOCs likely derive from OSECs, particularly HGSOCs of the proliferative type (padj < 2 × 10-4), suggesting a dualistic model for HGSOC origins. Super-enhancer (SE) landscapes were also more similar between FTSECs and HGSOCs than between OSECs and HGSOCs (p < 2.2 × 10-16). The SOX18 transcription factor (TF) coincided with a HGSOC-specific SE, and ectopic overexpression of SOX18 in FTSECs caused epithelial-to-mesenchymal transition, indicating that SOX18 plays a role in establishing the mesenchymal signature of fallopian-derived HGSOCs.Entities:
Keywords: RNA-seq; SOX18; dual origins; fallopian tube secretory epithelial cell; high-grade serous ovarian cancer; machine learning; one-class logistic regression models; ovarian surface epithelial cell; single-cell RNA-seq; super enhancers; transcription factors
Year: 2019 PMID： 31825847 DOI： 10.1016/j.celrep.2019.10.122
Source DB: PubMed Journal: Cell Rep Impact factor: 9.423