Genetic, Functional, and Epidemiologic Analyses of Genomic Variants in Uterine Leiomyomata

Abstract

Uterine leiomyomata (UL), commonly known as fibroids, are one of the primary causes of morbidity in reproductive-aged women. Epidemiologic analyses, including twin studies, familial aggregation, and racial biases in disease prevalence and morbidity, suggest heritable factors influence disease risk. Previous genetic association analyses in Japanese women reported three independent loci at 10q24.33, 11p15.5, and 22q13.1; genome-wide linkage and association analyses in European women identified a genomic region encompassing FASN at 17q25.3; and, a genome-wide association study (GWAS) in African American women associated a separate region at 22q13.1 with UL. To replicate and identify additional risk variants, we assembled FibroGENE – a consortium of conventional, population-based and direct-to-consumer cohorts. Discovery-phase meta-analysis of 244,324 European women replicated the three previously associated loci in Japanese women and identified 21 novel genomic regions. Of the 24 loci in total, 17 were replicated (p < 2.08e-3) in an independent, direct-to-consumer cohort of 58,655 European women obtained through 23andMe’s Research Portal. Meta-analysis across all five cohorts of 302,979 European women revealed an additional six discovery loci. Estrogen receptor alpha (ESR1) and forkhead box protein O1 (FOXO1) were located in two candidate regions associated with UL in our analyses at 6q25.2 and 13q14.11, respectively. FOXO1 is a well-characterized tumor suppressor and hormone-regulated transcription factor that interacts with ESR1 in the estrogen-signaling pathway. We observed significant, 2.18-fold greater nuclear FOXO1 levels in UL patient samples compared to myometrial control tissue in an immunohistochemistry-based assay. Interestingly, four regions that were identified and replicated in our FibroGENE analysis were also previously associated with endometriosis – another estrogen-dependent disease affecting uterine tissue. Epidemiologic meta-analysis of multivariate-adjusted risk estimates from the Nurses’ Health Study II and Women’s Health Study revealed women with a history of endometriosis had greater odds of UL when compared to women with no previous diagnosis of endometriosis. Here, we report the identification of a previously undocumented genetic overlap between endometriosis and UL and further underscore the relationship by formal epidemiologic quantification of the comorbidity of these two highly common gynecologic diseases.