Qualitative and Quantitative DNA- and RNA-Based Analysis of the Bacterial Stomach Microbiota in Humans, Mice, and Gerbils

Abstract

Clinical interventions in the stomach have been linked to fecal microbiota alterations, suggesting a function of the stomach in gastrointestinal (GI) homeostasis. We sought to determine the taxonomic bacterial biogeography of the upper GI tract, including different sites within the human stomach (cardia, corpus, and antrum), adjacent upstream (esophagus) and downstream (duodenum) locations, and luminal contents (aspirate), as well as whole-stomach samples from mice and gerbils. Qualitative and quantitative DNA- and RNA-based taxonomic microbiota analyses were combined to study the relationship of relative and absolute bacterial abundances and transcriptionally active bacterial microbiota components in the stomach of humans and mice. Stomach microbiota compositions resembled those of esophagus and duodenum. However, along the descending GI tract, the relative abundances of specific oropharyngeal commensals decreased (Streptococcus) or increased (Rothia mucilaginosa, Porphyromonas, and Lachnospiraceae). Furthermore, the compositional similarity (weighted UniFrac) between stomach aspirates and esophageal biopsy samples increased with gastric Streptococcus relative abundance. In both human aspirate and mouse stomach samples, Firmicutes were more abundant among transcriptionally active bacteria than Bacteroidetes. The relative abundance of Firmicutes in the stomach was negatively correlated and that of Bacteroidetes was positively correlated with absolute bacterial abundance, suggesting a disproportionate increase of Bacteroidetes over Firmicutes at higher bacterial densities. Human, mouse, and gerbil stomach samples showed similarities at higher taxonomic levels but differences at lower taxonomic levels. Our findings suggest selective enrichment and depletion of specific bacterial taxa in the stomach and Firmicutes being transcriptionally more active than Bacteroidetes that increase in relative abundance with total bacterial load. IMPORTANCE Clinical stomach interventions, such as acid inhibition or bypass surgery, have been linked to fecal microbiota alterations. We demonstrate that the stomach microbiota largely overlaps those of adjacent gastrointestinal locations and identify gradual decreases and increases in the relative abundances of specific bacteria within the stomach, suggesting selective enrichment and depletion. Moreover, similarities between stomach and esophagus samples are proportional to the concentrations of Streptococcus (Firmicutes) in the stomach. The relative abundance of Firmicutes in the stomach, compared to that of Bacteroidetes, is increased in RNA relative to DNA, indicating higher transcriptional activity. Moreover, increased absolute bacterial loads are associated with decreased relative abundance of Firmicutes and higher relative abundance of Bacteroidetes. Our findings characterize the stomach microbiota as influenced by Bacteroidetes influx against a background of transcriptionally more active Firmicutes. Human, mouse, and gerbil stomach microbiotas differ at lower taxonomic levels, which might affect the utility of these model organisms.

Keywords: 16S rRNA; absolute abundance; quantitative microbiota analysis; stomach microbiota; transcriptional activity.

FIG 1

Histology and locations of the sampled human upper GI tract sections. (A) Gastroesophageal junction, showing the stratified multilayered squamous epithelium of the esophagus (Es) and the single layered columnar epithelium of the gastric cardia region (Ca). (B) Stomach corpus (Co) mucosa with oxyntic gastric glands. (C) Stomach antrum (An) with mucous glands in the mucosa. (D) Small-bowel architecture with villi and crypts of the duodenum (Du). (E) Schematic overview, including gastric aspirate (As). Abbreviations: s, surface; m, mucosa. Magnifications, ×40 (A and B) and×100 (C and D).

FIG 2

Mucosal microbiota analysis of the human stomach in relation to esophagus and duodenum. (A and B) Principal-coordinate analysis (PCoA) plots of microbiota comparisons of biopsy samples based on taxonomic distance (unweighted UniFrac), showing significant differences between patients (A) but not between sampled locations (B). Significance was determined with a nonparametric analysis of similarity (ANOSIM) test, using 999 permutations, as implemented in QIIME (64). (C) Mean relative microbiota compositions of all upper GI samples at the taxonomic order and phylum levels, including all OTUs with ≥1% relative abundance in at least one sample, excluding samples from H. pylori-positive patients. (D) Cladogram showing bacterial taxa with differential relative abundances, from kingdom (innermost ring) to species (outermost ring), between esophagus and stomach biopsy samples from H. pylori-negative patients, using a linear discriminant analysis (LDA) effect size of >2.0 as determined with LEfSe (65). Circle diameters used to represent taxa are proportional to their relative abundances. Circle colors indicate a significant increase in esophagus samples (red) or in stomach samples (green) or no difference between the two sample types. Shaded circle fractions in red and green mark the lower taxonomic groups that comprise each taxon with significantly different relative abundances in the two sample groups.

FIG 3

Changes in relative abundance of specific bacterial taxa along the descending axis of the human upper GI tract. Relative abundances of bacterial taxa with a mean relative abundance of >0.1% and an LDA effect size of >2.0 for the LEfSe comparison of stomach and esophagus biopsy samples from H. pylori-negative patients (Fig. 2D) were analyzed. Correlations were determined separately from esophagus to antrum (black lines) and from esophagus to duodenum (red lines), using the two-tailed, nonparametric Spearman’s rank correlation coefficient test.

FIG 4

Firmicutes and Bacteroidetes in human gastric aspirate samples. (A) The relative dominance of Firmicutes and Bacteroidetes over the gastric microbiota was more pronounced in gastric aspirate samples than in biopsy samples, as demonstrated by the results of comparisons of combined fractions performed using the nonparametric Mann-Whitney U test (*, P < 0.05). (B) Variations in the relative abundances of Firmicutes and Bacteroidetes among gastric aspirate samples are apparently not linked to patient disease status. Symbols represent samples from individuals with different disease backgrounds from two separate studies, i.e., data from this study (outlined symbols) and a previously published data set (44), including H. pylori (HP)-positive and-negative patients.

FIG 5

Relationship of human stomach and esophagus microbiota. Analyses of correlations of taxonomic distances (weighted UniFrac) between gastric aspirate and esophageal biopsy samples with the relative abundances of the phylum Firmicutes (A) or the genus Streptococcus (B) in aspirate samples were performed on the basis of Spearman’s (two-tailed) rank correlation coefficient analysis. Each bar shows the relative abundance of Firmicutes (A) or Streptococcus (B) of an individual aspirate sample, as well as the range of taxonomic distances found between the microbiota of this sample and all esophagus biopsy samples, including samples from H. pylori-positive and -negative patients.

FIG 6

Microbiota comparison of H. pylori-negative human, mouse, and gerbil stomach samples. (A) Taxonomic microbiota compositions differ between humans (aspirate plus biopsy samples), mice, and gerbils, as determined with nonparametric ANOSIM tests, using 999 permutations, based on taxonomic distance (unweighted UniFrac). (B) Microbial diversity comparison using Shannon and Simpson indices, with significance determined based on a nonparametric two-sample t test, using 999 Monte Carlo permutations and Bonferroni correction for multiple comparisons (**, P < 0.01). (C) Mean relative microbiota compositions at the genus and phylum levels, including all OTUs with ≥1% relative abundance in at least one group of samples (human, mouse, or gerbil). (D) Variations in the relative abundances of Firmicutes (Firm.) and Bacteroidetes (Bact.) among human, mouse, and gerbil samples. Actino., Actinobacteria; HA, human aspirate; HB, human biopsy; G, gerbil; M, mouse; uncl., unclassified.

FIG 7

DNA- and RNA-based relative and quantitative taxonomic microbiota compositions of human gastric aspirate (n = 7) and mouse stomach (n = 6) samples. (A) Comparison of phylum relative abundances in DNA (blue) and RNA (orange) extracts of all samples based on 16S rRNA gene PCR and transcript RT-PCR amplicon sequencing (***, P < 0.001; **, P < 0.01 [Mann-Whitney U test]). (B) Comparison of bacterial phyla based on the ratio of 16S rRNA transcripts to 16S rRNA genes per sample (*, P < 0.05 [Kruskal-Wallis test]).

FIG 8

Association of phylum relative abundance and absolute 16S rRNA gene and transcript copy number. Correlations are shown for the five most abundant phyla, including Firmicutes (red), Bacteroidetes (orange), Actinobacteria (gray), Proteobacteria (light blue), and Fusobacteria (dark blue). (A and B) Correlation of phylum relative abundance in DNA, based on 16S rRNA gene amplicon sequencing with bacterial 16S rRNA gene copy number in human gastric aspirate (A) and mouse whole-stomach (B) samples. (C and D) Correlation of phylum relative abundance in RNA, based on 16S rRNA transcript amplicon sequencing and bacterial 16S rRNA transcript number in human gastric aspirate (C) and mouse whole-stomach (D) samples. Correlation coefficients (R) and significance (P) were calculated using Spearman’s rank correlation.