Linking biomarkersof environmental chemical exposure and endometriosis: integrating the exposome and metabolome in the ENDOXOMICS-b Study

Background/Aim Endometriosis is a gynaecological disease impacting a staggering 5-15% of women with dramatic socio-economic impacts, aggravated by long diagnostic delays of 8-10 years between symptom onset and surgical confirmation. Previous studies suggest a relationship between exposure to some persistent organic pollutants (POPs) and endometriosis, but its aetiology remains uncertain. In this study, we developed a comprehensive analytical framework based in mass spectrometry (MS) and advanced computational approaches to characterise exposure and metabolic profiles to identify potential biomarkers of exposure and effect related to endometriosis. Methods The ENDOXOMICS-β study comprises women seeking obstetric intervention for endometriosis, infertility, and other gynaecologic issues at University Hospital of Nantes. Endometriosis phenotype was determined histologically. 140 serum samples (92 cases, 48 controls) were analysed for POPs using ultra-trace methods based on liquid and gas chromatography coupled to high-resolution MS (LC- and GC-HRMS). Lipid and metabolite profiling was performed using the Biocrates MxP® Quant 500 Kit and targeted flow injection analysis (FIA-MS/MS) and LC-MS/MS. Data analyses included a battery of algorithms for single-block (elastic-net, neural networks, support vector machine) and multiblock (partial-least-square) variable selection. Multivariate logistic regression (MLR) was used to validate the associations, adjusting for covariates. Results Preliminary targeted metabolomics analyses quantified over 600 metabolites, including a number of functional/structural families (acylcarnitines, ceramides, glycerophospolipids, phosphatidylcholines, sphingolipids). 80 lipids were found associated with endometriosis (p<0.05) with elevated adjusted odds-ratios. Several organochlorine pesticides, polychlorobiphenyls and perfluoroalkyl substances were also identified. Multiblock models appear a powerful approach to integrate highly correlated, multidimensional data to identify potential exposure and effect biomarkers. Conclusion To our knowledge, this is the first application of quantitative MS-based metabolomics and exposomics in serum to identify potential biomarkers associated with endometriosis. Integrative multiblock methods offer promising potential for the discovery of biomarkers of exposure and effect with large applications in etiological and clinical research.