Proteomic discovery of tissue- and serum-based prognostic biomarkers for high-grade serous ovarian carcinoma

연구배경: This study aimed to investigate tissue- and serum-based prognostic biomarkers of high-grade serous ovarian carcinoma (HGSOC) through proteomic analysis. 대상 및 방법: We conducted label-free liquid chromatography-mass spectrometry using chemotherapy-naive, fresh-frozen primary HGSOC specimens (n=12) and compared the results between a favorable prognosis group (progression-free survival (PFS) ≥18 months) and a poor prognosis group (PFS <18 months). In parallel, we also conducted proteomic profiling of serum samples obtained from chemotherapy-naive, HGSOC patients (n=20). Differentially expressed serum proteins between the favorable and poor prognosis groups were compared with those identified from cancer tissues. Candidate protein biomarkers were validated via immunohistochemistry (IHC) and Enzyme-linked Immunosorbent Assay (ELISA) using an independent set of chemotherapy-naive primary HGSOC specimens (n=107) and serum samples (n=100), respectively. 결과: Among 658 differentially expressed proteins, 288 proteins were upregulated in the favorable prognosis group and 370 proteins were upregulated in the poor prognosis group. Using hierarchical clustering, we selected AAT, NFKB, PMVK, VAP1, FABP4, PF4, APOA1, and AGP for further validation via IHC staining. Survival analyses revealed that high expression of AAT, NFKB, and PMVK were independent biomarkers for favorable PFS. Conversely, high expression of VAP1, FABP4, and PF4 were identified as independent biomarkers for poor PFS. Furthermore, we constructed models predicting the 18-month PFS by combining clinical variables and IHC results. The optimal model was based on initial serum CA-125, germline BRCA1/2 mutations, residual tumors after surgery, International Federation of Gynecology and Obstetrics (FIGO) stage, and expression levels of the six proteins. Among the six biomarkers, AAT and FABP4 were quantified in the serum samples by ELISA. Dose-dependent prognostic significance of each protein''s serum concentration for PFS was observed. A model consisting of serum AAT and FABP4 concentrations and clinical variables showed god performance in predicting the poor prognosis groups. 결론: The present results elucidate the proteomic landscape of HGSOC and tissue- and serum-based protein biomarkers to predict the prognosis of HGSOC.