INCREASED REFLUX SECONDARY BILE ACIDS ARE ASSOCIATED WITH SHIFTS TO THE MICROBIOME AND TRANSCRIPTOME IN BARRETT’S ESOPHAGUS

Background: Bile acids comprise a major component of refluxate in GERD, the major risk factor for Barrett’s esophagus (BE) and esophageal adenocarcinoma (EAC). Bile acids are metabolized by bacteria, and bacterial composition is also impacted by exposure to bile acids. To understand how these factors may inter-relate in EAC development, we performed a comprehensive assessment of the microbiome, refluxate bile acids, and tissue gene expression in patients without and with BE, dysplasia, and EAC.
Methods: Clinical data were recorded and biospecimens were collected from subjects undergoing endoscopy as part of a prospective multi-center cross-sectional study. LC-MS was performed to quantify bile acids from gastric aspirates, 16S rRNA gene sequencing was performed to characterize the tissue microbiome, and RNA sequencing was performed on BE or control gastric cardia tissue.
Results: 152 subjects were enrolled and analyzed (51 controls; 101 BE: 50 no dysplasia, 10 indef. dysplasia, 17 low grade dysplasia, 17 high grade dysplasia, 7 EAC). BE gene expression demonstrated increases in bile acid secretion and decreases in nitrogen metabolism pathways compared to controls. There were two distinct BE gene expression clusters independent of associated histology and not associated with bile acid or microbiome composition (Fig 1A). No genes in cluster 1 were associated with dysplasia/EAC; in cluster 2 progression was associated with upregulation of FOSB, ITGA4, and CSF1. Almost all bile acids were conjugated, suggesting little bacterial bile salt hydrolase activity in the stomach. BE subjects had significantly higher levels of secondary bile acids (esp. conjugated forms of DCA; Fig 1B). There was no association between bile acid levels and dysplasia/EAC. The bile acid GDCA was associated with upregulation of CCKBR, HDC, and CXCL5 and keratinocyte differentiation pathway downregulation. There were trends towards inverse correlations between bile acid levels and Gram-negative bacteria (Fig 1C). BE subjects had tissue microbiome alterations, including increased Streptococcus. Dysplasia/EAC were associated with increased Lactobacillus and decreased Actinomyces and several other genera. Among bacteria, Actinomyces had the greatest number of significant correlations with tissue gene expression, including shifts in Notch signaling and oxidative phosphorylation pathways.
Conclusions: Bile acids appear to shape the BE microbiome, and both in turn associate with gene expression changes potentially contributing to the development of EAC. Two observed BE gene expression clusters were independent of bile acids, bacteria, and associated histology, suggesting that there may be distinct neoplastic pathways. Future studies are needed to determine whether bile acid or microbiome modification can modify EAC risk and to assess the biological significance of the BE gene clusters.