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

Authors: Lillian Kay Petersen*, Garyk Brixi*, Jun Li, Jie Hu, Zicheng Wang, Xikun Han, Anat Yaskolka Meir, Jaakko Tyrmi, Shruthi Mahalingaiah, Terhi Piltonen, Liming Liang

Abstract

Polycystic ovarian syndrome (PCOS) is a common hormonal disorder with high metabolic and psychological morbidity and is a leading cause of anovulatory infertility. PCOS is thought to be associated with increased systemic low grade inflammation, but the genetic and causal basis of this relationship is not yet understood. Here, we study the relationship between PCOS and obesity, testosterone, sex hormone binding globulin (SHBG), and chronic inflammation. First, we create the largest-to-date meta-analysis of PCOS (cases = 7,747 controls = 498,227), identifying 4 novel PCOS genetic loci. By further analyzing this alongside GWAS for obesity (n=681,275), SHBG (n=190,366), testosterone (n=176,687), and a meta-analysis of 138 inflammation biomarkers (average n=30,000), we replicate potential causal relationships from obesity and SHBG to PCOS. We further find extensive genetic correlations and causality (via Mendelian Randomization) between these traits and inflammation biomarkers. We identify genetic correlation between PCOS and 11 inflammatory biomarkers, suggesting a shared genetic architecture, including new relationships such as a strong correlation between PCOS and death receptor 5 (TRAILr2, LDSC rg=0.54, FDR=0.043). Additionally, Mendelian Randomization suggests 28 inflammation biomarkers with a causal effects on SHBG and four biomarkers with significant causal effects on testosterone, with TWEAK and MMP9 increasing testosterone while IL2Rb and IP10 are found to decrease testosterone levels in females. We further find evidence that testosterone increases CRP levels. Although we do find shared genetic architecture between PCOS and inflamation, we did not find evidence of causality between inflammation and PCOS. While our results support the hypothesis that chronic inflammation can influence androgen and SHBG levels in females, and that these traits can cause PCOS, further study and other forms of evidence is needed to verify the results from Mendelian randomization. Finally, to create a better genetic prediction of PCOS we combine the genetics of PCOS-related traits showing that including polygenic risk scores of BMI, WHR, Testosterone, and SHBG improves accuracy compared to using only PCOS risk scores in the UK Biobank (0.72 AUC from 0.59 AUC) and MGB Biobank (0.61 AUC from 0.59 AUC).

Stay tuned for full pdf, coming soon!