Circulating anti-Müllerian hormone levels in pre-menopausal women: novel genetic insights from a GWAS meta-analysis

Abstract

Study questionCan a genome-wide association study (GWAS) meta-analysis, including a large sample of young premenopausal women from a founder population from Northern Finland, identify novel genetic variants for circulating anti-Mullerian hormone (AMH) levels and provide insights into biological pathways and tissues involved in AMH regulation? Summary answerWe identified six loci associated with AMH levels at P < 5 x 10-8, including the previously reported MCM8, AMH and TEX41 loci, and three novel signals in or near CHEK2, BMP4 and EIF4EBP1. Gene set enrichment analysis highlighted significant enrichment in renal system vasculature morphogenesis and tissue enrichment analysis ranks the pituitary gland as a top associated tissue. What is known alreadyAMH is expressed by preantral and small antral stage ovarian follicles in women, and variation in age-specific circulating AMH levels has been associated with several health conditions. However, the biological mechanisms underlying the association between health conditions and AMH levels are not yet fully understood. Previous GWAS have identified loci associated with AMH levels in pre-menopausal women, but they were limited by small sample sizes or focused mostly on older pre-menopausal women. Study design, size, durationWe performed a GWAS meta-analysis for AMH level measurements in 9,668 pre-menopausal women. Participants/materials, setting, methodsWe performed a GWAS meta-analysis in which we combined 2,619 AMH measurements (at age 31 years old) from a prospective founder population cohort (Northern Finland Birth Cohort 1966, NFBC1966) with a previous GWAS meta-analysis that included 7,049 pre-menopausal women (spanning age range 15-48). NFBC1966 AMH measurements were quantified using an automated assay (Elecsys(R) AMH Plus (Roche)). We annotated the genetic variants, combined different data layers to prioritise potential candidate genes, described significant pathways and tissues enriched by the GWAS signals, identified plausible regulatory roles using colocalization analysis and leveraged publicly available summary statistics to assess genetic and phenotypic correlations with multiple traits. Main results and the role of chanceThree novel genome-wide significant loci were identified. One of these is in complete linkage disequilibrium with c.1100delC in CHEK2, which is found to be 4-fold enriched in the Finnish population compared to other European populations. We propose a plausible regulatory effect of some of the GWAS variants linked to AMH, as they colocalise with GWAS signals associated with gene expression levels of BMP4, TEX41 and EIFBP41. Gene set analysis highlighted significant enrichment in renal system vasculature morphogenesis and tissue enrichment analysis ranked the pituitary gland as the top association. Large scale dataThe GWAS meta-analysis summary statistics will be available for download from the GWAS Catalog. Accession numbers will be provided upon publication. Limitations, reasons for cautionThis study only included women of European ancestry and the unavailability of sufficiently sized relevant tissue data in gene expression datasets hinders the assessment of potential regulatory effects in reproductive tissues. Wider implications of the findingsOur results highlight the increased power of founder populations and larger sample sizes to boost the discovery of novel trait-associated variants underlying variation in AMH levels, which aided to characterise novel biological pathways and plausible genetic regulatory effects linked with AMH levels variation for the first time. Study funding / competing interest(s)This work has received funding from the European Unions Horizon 2020 research and innovation programme under the MATER Marie Sklodowska-Curie grant agreement No. 813707 and Oulu university scholarship foundation (N.P.-G.), Academy of Finland, Sigrid Juselius Foundation, Novo Nordisk, University of Oulu, Roche Diagnostics (T.T.P). This work was supported by the Estonian Research Council grant 1911 (R.M.). J.R. was supported by the European Unions Horizon 2020 research and innovation program under grant agreements No. 874739 (LongITools), 824989 (EUCAN- Connect), 848158 (EarlyCause) and 733206 (LifeCycle). U.V. was supported by the Estonian Research Council grant PRG (PRG1291). The NFBC1966 received financial support from University of Oulu Grant no. 24000692, Oulu University Hospital Grant no. 24301140, ERDF European Regional Development Fund Grant no. 539/2010 A31592.