Genetic, proteomic, and metabolomic analysis of endometriosis-associated pain and its comorbidity with bladder pain syndrome

Endometriosis affects up to 10% of women in reproductive age and is frequently co-diagnosed with bladder pain syndrome (BPS) among chronic pelvic pain patients. The diagnostic pathways are long and there are no validated blood-based biomarkers. It has been hypothesised that the two conditions share aetiological features in pain generating and sustaining mechanisms, but no comprehensive evaluation of protein and metabolite profiles was conducted to date. In this thesis I explore genotypes, inflammation and neurology-related plasmatic proteins, and global metabolomic profile of 786 women recruited into the Translational Research in Pelvic Pain study encompassing patients with endometriosis, BPS, or both, along participants with only chronic pelvic pain or without any pain symptoms. In the first chapter I investigate whether polygenic risk scores can predict endometriosis and suggest avenues for biomarker exploration. My results highlight lack of portability of a risk score previously validated in European population, and highlight proteins, metabolites and phenotypes correlating with specific risk scores. I also define a previously unexplored polygenic risk score for endometriosis, which outperforms the previously validated score; however, the best-performing score comes from an earlier, UK Biobank-based study. In the second chapter I explore the proteomic and metabolomic correlates of endometriosis-associated pain and BPS, with the former yielding two putative biomarkers of limited discrimination power, but with supporting evidence in previous literature. This chapter also provides baseline biomarker panels characterising the promising potential of plasmatic proteins and metabolites to predict endometriosis, chronic pelvic pain and (to a lesser extend) BPS status. The final chapter explores the molecular associations of pain phenotypes, symptoms, and surgical endometriosis features. Comprehensively integrating widespread pain, quantitative sensory testing, fatigue, conditioned pain modulation and surgical phenotypes this chapter discerns protein and metabolite associations representing disease-specific from pain-related processes. Molecular associations of prevalence and severity of patient-reported pain outcomes are analysed, highlighting new candidate pain biomarkers. Finally, a feasibility of a biomarker panel predicting dysmenorrhea severity from blood proteome and metabolome is demonstrated, however yielding a modest predictive performance, and presenting an interpretation challenge due to the impact of xenobiotics and medications.