By Pooja Toshniwal PahariaReviewed by Lauren HardakerSep 4 2025In a recent study published in the British Journal of Cancer, researchers conducted proteome-level Mendelian randomization (MR) to identify blood-based proteins for early detection and treatment of epithelial ovarian cancer (OC).
They found a strong association between plasma follitropin subunit beta (FSHB) and, more strikingly, follicle-stimulating hormone (FSH) with endometrioid OC. The findings position FSH as a potential serum marker for earlier diagnosis, a critical need given OC’s frequent late-stage presentation and poor prognosis. However, further validation is still required.
Image credit: Nemes Laszlo/Shutterstock
Crucially, the FSH receptor (FSHR) is already druggable as an agonist in infertility treatment, and an investigational antagonist (suramin) exerts antineoplastic activity. Identifying additional protein associations despite pleiotropic effects at known risk loci further highlights the value of protein-level insights for risk stratification. These results may support biomarker-driven prognosis, precision screening, and repurposing of existing drugs as innovative therapies for OC.
Introduction
Ovarian cancer is the most lethal gynecological malignancy, accounting for nearly half of all deaths from female genital tract cancers. Its high mortality stems largely from late-stage diagnosis, with five-year survival rates below 30% for stage III–IV disease, compared to over 90% for stage I. Current screening strategies, primarily cancer antigen 125 (CA125) testing and transvaginal ultrasound, lack sensitivity for early-stage tumors, highlighting the urgent need for improved detection methods.
About The Study
The present study used a proteome-wide MR framework to identify circulating protein markers and potential drug targets for OC and its histological subtypes.
The team obtained summary-level genomic data on ovarian cancers from the Ovarian Cancer Association Consortium (OCAC). The dataset comprised 25,509 European individuals with (case) and 40,941 individuals without (control) OC. Subtype-specific analyses included low-grade and high-grade cases of serous type (n=14,049), endometrioid type (n=2,810), clear cell type (n=1,366), and mucinous type (n=1,417).
The researchers sourced proteomic data on 2,923 plasma proteins from the United Kingdom Biobank Pharma Proteomics Project (UKB-PPP). Analysis focused on 34,557 European individuals, identifying genetic associations with 2,414 proteins. After excluding proteins with palindromic instruments, the MR analyses retained 2,337 proteins, including cis- and trans-protein quantitative trait loci (pQTLs) to instrument circulating protein concentrations.
The team expressed effect estimates as odds ratios (ORs). Depending on instrument availability, they derived causal estimates using the Wald ratio and inverse-variance weighting (IVW). Sensitivity analyses included MR-Egger regressions, weighted mode, and weighted median to account for pleiotropy. The leave-one-out analysis and MR-Steiger directionality tests evaluated the robustness of the findings.
To validate the findings, the researchers replicated the analyses using independent proteomic datasets, including the KARMA, INTERVAL, KORA, Fenland, and deCODE cohorts. They evaluated the proteins with evidence of causal association for therapeutic relevance using the DrugBank and the Drug-Gene Interaction Database (DGIdb).
Results
Of 2,337 plasma proteins examined, 12 showed associations with ovarian cancer (OC) or its subtypes. The strongest signal was between follitropin subunit beta (FSHB) and endometrioid OC, with a two-fold increased risk (OR 2.4). This association was consistent across weighted median (OR 2.61) and weighted mode (OR 2.64) analyses, with no evidence of pleiotropy or bias in sensitivity tests.
The result aligns with the gonadotropin hypothesis, suggesting excess exposure to FSH increases OC risk. Importantly, FSH and its receptor are druggable targets, with existing fertility treatments and the investigational antagonist suramin having demonstrated antineoplastic activity. Colocalization analysis was suggestive but not definitive (PPH4 ≈ 0.65), and replication was mixed: the FSHB signal was observed in Olink/KARMA but not in SomaScan cohorts, whereas instruments for FSH (the protein complex) did replicate across SomaScan, supporting the main signal.
Beyond FSHB, 11 additional proteins showed suggestive evidence of causal associations with OC. These included vesicle amine transport protein-1 (VAT1), keratin, type I cytoskeletal 18 (KRT18), cathepsin L (CTSL), lacritin (LACRT), carcinoembryonic antigen-related cell adhesion molecule 5 (CEACAM5), prosaposin-like 1 (PSAPL1), corneodesmosin (CDSN), langerin (CD207), lactoperoxidase (LPO), GDNF family receptor alpha-3 (GFRA3), and F3 tissue factor.
Directionality and sensitivity analyses supported these findings, though pleiotropy from ABO and Microtubule-associated protein tau-antisense RNA 1 (MAPT-AS1) loci explained some associations. Functionally, several proteins showed subtype-specific effects; for example, CTSL was inversely associated with serous OC, and F3 with lower mucinous OC risk, indicating that results should be interpreted cautiously.
The researchers identified several proteins as actionable drug targets. FSHR antagonism with suramin, KRT18 inhibition, CTSL-targeting agents, and CEACAM5-directed therapies (e.g., labetuzumab) represent promising therapeutic avenues. F3, a tissue factor central to coagulation and tumor biology, is already targeted by approved therapies such as tisotumab vedotin-TFTV, which is approved for cervical cancer and under investigation for ovarian cancer.
Similarly, LPO and GFRA3 were linked to multiple investigational drugs, while agents annotated as targeting CTSL (e.g., trastuzumab deruxtecan) and labetuzumab (targeting CEACAM5) illustrate repurposing opportunities.
Download your PDF copy now!
Conclusions
The study identifies FSHB and its receptor as key biomarkers and druggable targets for endometrioid ovarian cancer, alongside 11 additional plasma proteins with potential relevance across OC subtypes. These findings highlight proteins as potential biomarkers for early detection, prognosis, and therapeutic innovation, complementing existing screening methods.
Approved or investigational drugs target many proteins, which supports drug repurposing. The authors emphasize the need for future studies with larger case numbers, tissue-specific proteomics, and non-European populations to validate these associations and assess clinical utility.
Journal Reference
Mulugeta, A., Stacey, D., Lumsden, A.L., et al. (2025). Protein markers of ovarian cancer and its subtypes: insights from proteome-wide Mendelian randomization analysis. Br J Cancer. DOI: 10.1038/s41416-025-03143-w https://www.nature.com/articles/s41416-025-03143-w
UCLA Researchers Reprogram Stem Cells to Produce Continuous Cancer-Fighting T Cells