Proteomic signatures of biofluid-derived extracellular vesicles for benign and malignant ovarian tumor differentiation, prognosis, and chemotherapy response prediction.

e17562 Background: Small extracellular vesicles (sEVs) are nano-sized particles that reflect the pathological state of their origin cells and serve as promising biomarkers for various tumor. In high-grade serous ovarian carcinoma (HGSC), the most lethal ovarian cancer subtype, sEV proteomics offers significant potential for uncovering diagnostic markers, prognostic indicators, and therapeutic targets. Methods: Proteomic analyses of sEVs from plasma, urine, and ascites were conducted using 4D-label free mass spectrometry. sEV proteins specific to malignant tumors or linked to varying prognoses and chemotherapy responses were identified and analyzed through functional pathway studies, combined with single-cell sequencing data, to uncover their characteristics in biofluids. The diagnostic and prognostic potential of these altered sEV proteins in HGSC was also evaluated. Results: We identified 2,172 proteins in plasma-derived sEVs and 4,596 proteins in urine-derived sEVs across all patient samples. Principal component analysis of abundant proteins in plasma-derived sEVs clearly distinguished HGSC from benign diseases. Under integrated analysis of vesicle proteomics and single-cell transcriptomics, sEV proteins in plasma and urine samples from the malignant tumor group were predominantly secreted by tumor, endothelial, and stromal cells within the tumor microenvironment. In plasma-derived sEVs, proteins elevated in HGSC were enriched in pathways regulating the citrate cycle, complement and coagulation cascades, and ECM-receptor interaction. CDH5 and TGFB1, associated with tumor progression, showed higher levels in HGSC than borderline tumors and further increased in advanced compared to early HGSC. Notably, UBA6 was identified as a plasma-specific sEV protein unique to HGSC compared to borderline tumors and benign diseases. Additionally, MUC1 and CDH5 in plasma-derived sEVs effectively distinguished HGSC from benign cases, with ROC-AUC values of 0.96 and 0.91, respectively. Urine sEV proteomics revealed PTCD3 and F10 were higher in HGSC than benign disease, with AUC values of 0.95 and 0.89, distinguishing malignant tumors. Furthermore, GP2, CCNY, and VWF in plasma sEVs predicted chemotherapy resistance with ROC values of 0.93, 0.92, and 0.84, respectively, and were associated with shorter PFS. Similarly, VCP in ascites (ROC 0.94) and NOB1 in urine (ROC 0.96) strongly predicted chemotherapy resistance and were also linked to shorter PFS. Conclusions: SEVs from plasma, urine, and ascites, predominantly secreted by tumor and microenvironmental cells, serve as tumor-specific or -enriched biomarkers for HGSC. They enable non-invasive detection, prognosis, and therapy guidance by distinguishing HGSC and predicting chemotherapy resistance.