Claudin-3 and Claudin-4 Glycovariants in Ovarian Cancer Diagnostics

Abstract

Ovarian cancer has one of the highest mortality rates among gynecological cancers. A large contributing factor is that the disease is often only detected in the later stages, where 5-year survival rates are much worse, 29 % for patients diagnosed in stage 3 or 4, compared to 92 % in patients diagnosed in stage 1 or 2. There is a clear need for new biomarkers to supplement or replace CA125 and HE4 currently used gold standard biomarkers. Four potential glycovariant biomarkers were selected for this study. Claudin-3, claudin-4, CD24, and CD63 were utilized as capture antibody. Different glycovariants of these biomarkers were measured in serum samples from ovarian cancer patients in a well-plate based sandwich assay. The assay utilized ultra-high sensitivity europium-chelate doped nanoparticles conjugated to lectins to detect specific cancer-related glycans on the captured four biomarker proteins. An initial round of screening was conducted to find antibody-lectin combinations that showed high expression in ovarian cancer serum samples. Selected combinations were tested with a larger sample cohort, based on those results the study was continued with two antibody-lectin combinations: claudin-3 capture with macrophage galactose lectin (MGL) and claudin-4 with MGL tracer. The combinations were tested with an 81-patient cohort with healthy (n = 14), benign (n = 23) and ovarian cancer (n = 44) samples. Both combinations were overexpressed in samples across all FIGO stages and all measured histological classifications. Claudin-3/MGL performed better of the two combinations. Ovarian cancer sample signals were on average 2.9 times higher than healthy samples. High elevation was observed in early stage cases as well, enabling early detection of ovarian cancer. However, signal from benign samples was even more elevated, on average 4.8 times higher than healthy samples. Despite promising results in cancer vs. healthy samples, real world diagnostic utility is limited by overexpression in benign cases. The promising differentiation between healthy and cancer samples calls for further research into claudins as ovarian cancer biomarkers