Abstract
BACKGROUND: Rifaximin is a poorly absorbed broad-spectrum antibiotic used for hepatic encephalopathy. Although increased Lactobacillaceae and decreased Bacteroidetes abundance are characteristic of hepatic encephalopathy, rifaximin does not dramatically alter the stool microbiota. As the antimicrobial effect of rifaximin increases by micellization with bile acids, we hypothesized that rifaximin alters the microbiota in the duodenum and jejunum, where the levels of bile acids are abundant. METHODS AND RESULTS: Eight-week-old BALB/c mice were injected with carbon tetrachloride (CCl(4)) intraperitoneally for 12 weeks to induce liver fibrosis. The mice were grouped into the control (n = 9), CCl(4) (n = 13), and rifaximin group in which mice were treated with rifaximin for two weeks after CCl(4) administration (n = 13). We analyzed the microbiota of the duodenum, jejunum, ileum, cecum, and stool using 16S ribosomal RNA gene analysis. The content of Lactobacillaceae, the most abundant bacterial family in the duodenum and small intestine, increased in the CCl(4) group, especially in the jejunum (median 67.0% vs 87.8%, p = 0.03). Rifaximin significantly decreased Lactobacillaceae content in the duodenum (median 79.4% vs 19.0%, p = 0.006) and jejunum (median 87.8% vs 61.3%, p = 0.03), but not in the ileum, cecum, and stool. Bacteroidetes abundance tended to decrease on CCl(4) administration and increased following rifaximin treatment in the duodenum and jejunum. S24_7, the most abundant family in Bacteroidetes, demonstrated a significant inverse correlation with Lactobacillaceae (duodenum, r = - 0.61, p < 0.001; jejunum, r = - 0.72, p < 0.001). In the ileum, cecum, and stool, the effect of rifaximin on the microbiota was minimal, with changes within the same phylum. The percentage of bacterial families, such as Lactobacillaceae and S24_7 in the duodenum and small intestine, did not correlate with that in the stool. CONCLUSIONS: The abundance of Lactobacillaceae increased in the jejunum of mice with CCl(4)-induced liver fibrosis, while rifaximin significantly reduced it in the duodenum and jejunum. Thus, rifaximin possibly exerts its effect by altering the duodenal and jejunal microbiota. Furthermore, changes in the duodenal and small intestinal microbiota were not associated with that of stool, suggesting that the analysis of stool microbiota is insufficient to evaluate upper intestinal microbiota.