Fecal microbial determinants of fecal and systemic estrogens and estrogen metabolites: a cross-sectional study

Background: High systemic estrogen levels contribute to breast cancer risk for postmenopausal women, whereas low levels contribute to osteoporosis risk. Except for obesity, determinants of non-ovarian systemic estrogen levels are undefined. We sought to identify members and functions of the intestinal microbial community associated with estrogen levels via enterohepatic recirculation. Methods: Fifty-one epidemiologists at the National Institutes of Health, including 25 men, 7 postmenopausal women, and 19 premenopausal women, provided urine and aliquots of feces, using methods proven to yield accurate and reproducible results. Estradiol, estrone, 13 estrogen metabolites (EM), and their sum (total estrogens) were quantified in urine and feces by liquid chromatography/tandem mass spectrometry. In feces, beta-glucuronidase and beta-glucosidase activities were determined by realtime kinetics, and microbiome diversity and taxonomy were estimated by pyrosequencing 16S rRNA amplicons. Pearson correlations were computed for each log(e) estrogen level, loge enzymatic activity level, and microbiome alpha diversity estimate. For the 55 taxa with mean relative abundance of at least 0.1%, ordinal levels were created [zero, low (below median of detected sequences), high] and compared to loge estrogens, beta-glucuronidase and beta-glucosidase enzymatic activity levels by linear regression. Significance was based on two-sided tests with alpha=0.05. Results: In men and postmenopausal women, levels of total urinary estrogens (as well as most individual EM) were very strongly and directly associated with all measures of fecal microbiome richness and alpha diversity (R >= 0.50, P <= 0.003). These non-ovarian systemic estrogens also were strongly and significantly associated with fecal Clostridia taxa, including non-Clostridiales and three genera in the Ruminococcaceae family (R=0.57-0.70, P=0.03-0.002). Estrone, but not other EM, in urine correlated significantly with functional activity of fecal beta-glucuronidase (R=0.36, P=0.04). In contrast, fecal beta-glucuronidase correlated inversely with fecal total estrogens, both conjugated and deconjugated (R <=-0.47, P <= 0.01). Premenopausal female estrogen levels, which were collected across menstrual cycles and thus highly variable, were completely unrelated to fecal microbiome and enzyme parameters (P >= 0.6). Conclusions: Intestinal microbial richness and functions, including but not limited to beta-glucuronidase, influence levels of non-ovarian estrogens via enterohepatic circulation. Thus, the gut microbial community likely affects the risk for estrogen-related conditions in older adults. Understanding how Clostridia taxa relate to systemic estrogens may identify targets for interventions.