Analysis of Fecal, Salivary, and Tissue Microbiome in Barrett's Esophagus, Dysplasia, and Esophageal Adenocarcinoma.

Background and Aims: Esophageal adenocarcinoma (EAC) incidence has risen dramatically in the Western countries over the past decades. The underlying reasons are incompletely understood, and shifts in the esophageal microbiome have been postulated to increase predisposition to disease development. Multiple factors including medications, lifestyle, and diet could influence microbiome composition and disease progression. The aim of this study was (1) to identify a feasible method to characterize the tissue-associated microbiome, and (2) to investigate differences in the microbiome of saliva, esophageal tissue, and fecal samples by disease state and validate with 2 external cohorts. Methods: Forty-eight patients (15 Barrett's esophagus [BE], 4 dysplasia, 15 EAC, and 14 healthy) were enrolled in this cross-sectional study (Munich cohort). Demographics, epidemiologic and clinical data, medications, smoking, and alcohol consumption were assessed. 16S rRNA Gene sequencing was performed on saliva, tissue biopsy and fecal samples. PAXgene fixation was used as a novel methodology. Microbial community alpha- and beta-diversity, as well as microbial composition at phylum and genus level, were characterized for this cohort and compared with 2 external cohorts: New York cohort and Cooperative Health Research in the Augsburg Region cohort. Results: We first established PAXgene fixation is a feasible method for microbiome analysis and utilized it to identify a distinct microbial shift in tissue biopsies from patients with EAC, whereas overall microbial diversity in salivary and fecal samples did not differ significantly between disease states. Our findings were similar in a reanalysis to those from a US cohort that used a standardized fresh frozen biopsy collection protocol (New York cohort, N = 75 biopsies). Nevertheless, we could not distinguish German Munich cohort patients from a German population-based cohort (Cooperative Health Research in the Augsburg Region cohort, N = 2140 individuals) when fecal bacterial profiles were compared between both cohorts. In addition, we used data integration of diagnosis and risk factors of patients and found associations with microbiome alterations. Conclusion: Sample collection and microbiome analysis are indeed feasible and can be implemented into clinical routine by an easy-to-use biopsy protocol. The presence of BE and EAC together with epidemiologic factors can be associated with alterations of the salivary, tissue, and fecal microbial community in an easy-to-use data integration concept. Given a possible role of the microbiome in BE and EAC, it will be important in future studies to take tissue-specific microbial communities and individual taxa into account in larger prospective studies.