Understanding and Predicting Polycystic Ovary Syndrome through Shared Genetics with Testosterone, SHBG, and Chronic Inflammation

Abstract

AbstractPolycystic ovary syndrome (PCOS) is a common hormonal disorder that affects one out of eight women and has high metabolic and psychological comorbidities. PCOS is thought to be associated with increased systemic low-grade inflammation, but the underlying mechanisms remain unclear. Here we study the genetic relationship between PCOS and obesity, testosterone, sex hormone binding globulin (SHBG), and chronic inflammation. First we create a large meta-analysis of PCOS (7,747 PCOS cases and 498,227 controls) and identify four novel genetic loci associated with PCOS. These novel loci have been previously associated with changes in gene expression in multiple tissues including the thyroid and ovary. We then analyze the PCOS meta analysis alongside GWASs for obesity (n=681,275), SHBG (n=190,366), testosterone (n=176,687), and meta-analyses of 130 inflammatory biomarkers (average n=30,000). We replicate potential causal relationships (via Mendelian randomization) from obesity and SHBG to PCOS and find extensive genetic correlations and causality between these traits and inflammatory biomarkers. We identify significant genetic correlations between PCOS and eight inflammatory biomarkers, including new relationships such as a strong correlation to death receptor 5 (LDSC rg = 0.54, FDR = 0.043). Although these results point to a shared genetic architecture between PCOS and inflammation, we did not find statistically significant causality between them; however, we did find 27 inflammatory biomarkers with significant causal effects on SHBG and four biomarkers with significant causal effects on testosterone. These results support the hypothesis that chronic inflammation can influence androgen and SHBG levels, though more research is needed to verify these results. Finally, we show that combining the polygenic risk scores of PCOS-related traits improves genetic prediction of PCOS cases in the UK Biobank and MGB Biobank compared to using only PCOS risk scores (0.72 AUC from 0.59 AUC and 0.61 AUC from 0.59 AUC in each biobank respectively).