African genetic ancestry interacts with body mass index to modify risk for uterine fibroids

Race, specifically African ancestry, and obesity are important risk factors for uterine fibroids, and likely interact to provide the right conditions for fibroid growth. However, existing studies largely focus on the main-effects rather than their interaction. Here, we firstly provide evidence for interaction between categories of body mass index (BMI) and reported-race in relation to uterine fibroids. We then investigate whether the association between inferred local European ancestry and fibroid risk is modified by BMI in African American (AA) women in the Vanderbilt University Medical Center bio-repository (BioVU) (539 cases and 794 controls) and the Coronary Artery Risk Development in Young Adults study (CARDIA, 264 cases and 173 controls). We used multiple logistic regression to evaluate interactions between local European ancestry and BMI in relation to fibroid risk, then performed fixed effects meta-analysis. Statistical significance threshold for local-ancestry and BMI interactions was empirically estimated with 10,000 permutations (p-value = 1.18x10-4). Admixture mapping detected an association between European ancestry and fibroid risk which was modified by BMI (continuous-interaction p-value = 3.75x10-5) around ADTRP (chromosome 6p24); the strongest association was found in the obese category (ancestry odds ratio (AOR) = 0.51, p-value = 2.23x10-5). Evaluation of interaction between genotyped/imputed variants and BMI in this targeted region suggested race-specific interaction, present in AAs only; strongest evidence was found for insertion/deletion variant (6:11946435), again in the obese category (OR = 1.66, p-value = 1.72x10-6). We found nominal evidence for interaction between local ancestry and BMI at a previously reported region in chromosome 2q31-32, which includes COL5A2, and TFPI, an immediate downstream target of ADTRP. Interactions between BMI and SNPs (single nucleotide polymorphisms) found in this region in AA women were also detected in an independent European American population of 1,195 cases and 1,164 controls. Findings from our study provide an example of how modifiable and non-modifiable factors may interact to influence fibroid risk and suggest a biological role for BMI in fibroid etiology.Author summary: Although it is postulated that obesity and non-modifiable risk factors such as race or genetic ancestry may interact to jointly influence uterine fibroid growth, most existing studies have not evaluated their interaction. In this study we exhibit evidence for interaction across several tiers of investigation. We first show that the association between reported/third-party identified-race (African American (AA) and European American (EA)) and fibroid risk is modified by body mass index (BMI) categories. We then reveal evidence for interaction between two genetically-inferred local European ancestry regions (top two regions: chromosome 6p24, ADTRP; chromosome 2q31-32, TFPI, COL5A2) along the genome and BMI in relation to fibroids in two independent AA populations. Of intrigue, genes in these top two regions are mechanistically related, where ADTRP gene-product is an immediate upstream regulator for TFPI. Then at the genotype level, we show that interaction between genotyped/imputed variants and BMI is race-specific for chromosome 6p24 region, present in AAs but not in EAs, whereas, trans-ethnic, common across AAs and EAs for the 2q31-32 region, as replicated in an independent population of EA women. Our multi-tiered investigation supports evidence for interaction between reported race, genetic ancestry and BMI in relation to fibroid risk.