Mucosa-Associated Microbial Profile Is Altered in Small Intestinal Bacterial Overgrowth

Abstract

The overall gut microbial profile of patients with small intestinal bacterial overgrowth (SIBO) has not been thoroughly investigated. We investigated the microbial communities of mucosal specimens from the duodenum, ileum, sigmoid colon, and feces of patients with and without SIBO, as diagnosed by lactulose breath testing. The bacteria present in the mucosal and fecal samples were identified using 16S rRNA gene sequencing. Further analysis was performed using the linear discriminant analysis (LDA) effect size method, random forest analysis, and receiver operating characteristic analysis. The microbial diversities of the fecal samples were significantly lower than those of the mucosal samples from the duodenum, ileum, and sigmoid colon (P < 0.001, P < 0.001, and P < 0.001, respectively), while the bacterial compositions of the ileac mucosal samples and sigmoid mucosal samples were similar. The bacterial composition of either the fecal or duodenal mucosal samples were significantly different from those of the other three groups (ANOSIM R = 0.305, P = 0.001). The bacterial compositions of the mucosal samples of the duodenum, ileum, and sigmoid colon in the SIBO + subjects were significantly different from those of the SIBO- subjects (ANOSIM P = 0.039, 0.002, and 0.007, respectively). The relative abundances of 7, 18, and 8 genera were significantly different (LDA score > 3) in the mucosal samples of the duodenum, ileum, and sigmoid colon between the SIBO + and SIBO- groups. Four genera (Lactobacillus, Prevotella_1, Dialister, and norank_f__Ruminococcaceae) showed similar changes among the mucosal samples of the duodenum, ileum, and sigmoid colon in the SIBO + subjects. A signature consisting of four genera in the duodenal mucosa, three genera in the ileac mucosa, or six genera in the mucosa of the sigmoid colon exhibited predictive power for SIBO (area under the curve = 0.9, 0.93, and 0.87, respectively). This study provides a comprehensive profile of the gut microbiota in patients with SIBO. Dysbiosis was observed in the mucosa-associated gut microbiome but not in the fecal microbiome of patients with SIBO. Furthermore, we identified mucosa-associated taxa that may be potential biomarkers or therapeutic targets of SIBO. Further investigation is needed on their mechanisms and roles in SIBO.

Keywords: 16S rRNA gene sequencing; SIBO; biomarker; duodenum; ileum; microbiota; mucosa.

FIGURE 1

Microbial richness and diversity of the mucosal and fecal microbiota of all subjects. (A,B) Ace richness and Shannon diversity indices of mucosal and fecal samples: the richness and diversity were significantly lower in the fecal samples compared to the mucosal samples of the duodenum, ileum, and sigmoid colon. Data are presented as mean ± standard deviation of the mean for each sample type. The differences between groups were calculated using the Wilcoxon rank-sum test. ***P < 0.001.

FIGURE 2

Distinct bacterial composition at the phylum level in the mucosal and fecal samples of all subjects. (A) The predominant phylum in the 4 types of samples. (B) PCoA analysis based on the bray_curtis distance showed that the bacterial composition of the fecal and duodenal mucosal samples were significantly different from the other two groups (ANOSIM R = 0.305, P = 0.001). (C) Firmicutes and Bacteroidetes were significantly enriched (P = 1.2 × 10^–3 and P = 5 × 10^–3, respectively), while Proteobacteria, unclassified_k_norank_d_Bacteria, and Actinobacteria were significantly lower in the fecal samples compared to the mucosal samples (P = 5.2 × 10^–5, P = 1.5 × 10^–12, and P = 1.7 × 10^–4, respectively). (D) Firmicutes and Bacteroidetes were significantly lower in the duodenal mucosal samples than in the ileac mucosal samples (P = 0.047 and P = 0.026, relatively). (E) Fusobacteria, Epsilonbacteraeota, Patescibacteria, and Spirochetes were significantly enriched in the mucosal samples of the duodenum than in the mucosal samples of the sigmoid colon (P = 1.83 × 10⁻⁴, P = 1.9 × 10⁻⁴, P = 1.39 × 10⁻⁴, and P = 5.33 × 10⁻³, respectively). The phyla with < 0.1% abundance were aggregated as “others.” The Kruskal–Wallis H test was used to test differences among multiple groups with fdr controlling the false discovery rates. Wilcoxon rank-sum test was used to test differences between two groups. *P < 0.05, **P < 0.01, and ***P < 0.001.

FIGURE 3

Distinct bacterial composition at the genus level in the mucosal and fecal samples of all subjects. (A) The predominant genera in the 4 types of samples. (B) Bacteroides (P < 0.001), Agathobacter (P < 0.001), and Roseburia (P < 0.01) were more abundant in the fecal samples compared to the mucosal samples. On the other hand, the relative numbers of unclassified_k_norank_d_bacteria (P < 0.001), lactobacillus (P < 0.001), Ralstonia (P < 0.001), and Streptococcus (P < 0.001) were more enriched in the mucosa compared to those in the feces. (C,D) Bacteroides (P < 0.05), Faecalibacterium (P < 0.01), Blautia (P < 0.001), Subdoligranulum (P < 0.01), Bifidobacterium (P < 0.01), and Agathobacter (P < 0.05) were relatively less abundant, whereas Streptococcus (P < 0.01), Neisseria (P < 0.001), Veillonella (P < 0.001), and Prevotella_7 (P < 0.001) were more abundant in the mucosa of the duodenum than in the ileum and sigmoid colon. The genera with < 0.1% abundance were aggregated as “others.” Kruskal–Wallis H test was used to test differences among multiple groups with fdr controlling the false discovery rates. Wilcoxon rank-sum test was used to test differences between two groups. *P < 0.05, **P < 0.01, and ***P < 0.001.

FIGURE 4

Microbial richness and diversity of the mucosa-associated microbial community in the SIBO + and SIBO− subjects. (A,B) The Sobs richness and Shannon diversity indices of the ileac mucosal samples were significantly lower in the SIBO + subjects compared to the SIBO− subjects. (C) The Sobs richness index of the mucosal samples of the sigmoid colon was significantly lower in the SIBO + subjects than in the SIBO− subjects. *P < 0.05 and ***P < 0.001.

FIGURE 5

Distinct bacterial composition in the mucosal samples of the SIBO + and SIBO− groups. (A–C) PCoA analysis. The bacterial community compositions were clustered significantly separately between the SIBO + and SIBO− groups in the duodenal mucosa based on abund_jaccard distance (ANOSIM R = 0.102, P = 0.039), in the ileac mucosa based on binary_jaccard (ANOSIM R = 0.216, P = 0.002), and in the mucosa of sigmoid colon based on binary_jaccard (ANOSIM R = 0.155, P = 0.007). (D–F) Linear discriminant analysis (LDA) identified distinct bacterial genera that were enriched in the SIBO + and SIBO− groups. Genera with LDA score > 3 were considered significant. The abundances of 7, 18, and 8 genera differed significantly in the mucosal samples of duodenum, ileum, and sigmoid colon between the SIBO + and SIBO− groups.

FIGURE 6

Identification of microbial signatures of SIBO + subjects. Random forest analysis was used to identify taxa in the mucosal samples of the duodenum, ileum, and sigmoid colon to distinguish the SIBO + and SIBO− subjects. The genera were ranked in descending order according to their importance to the accuracy of the model. The insert shows the AUC verification results. X represents the number of species in the importance ranking and Y represents the AUC value. According to the solid point with the largest Y value, the corresponding X value was considered the ultimate number of top species to build a prediction model. The prediction performance was assessed by ROC analysis. (A,B) The random forest analysis showed that use of the top 4 important features (Ruminococcus_1, [Eubacterium]_ventriosum_group, norank_f__NS9_marine_group, and unclassified_c__Parcubacteria) of the duodenal mucosal samples for prediction of SIBO was the most optimal model, and ROC Analysis assessed its prediction performance with an AUC value of 0.9 (95% CI: 0.76–1). (C,D) In the ileum, random forest analysis found that use of the top 3 important features (Rhodococcus, norank_f__norank_o__NB1-j, and Candidatus_Solibacter) to construct the prediction model yielded an AUC value of 0.93 (95% CI: 0.81–1) based on ROC analysis. (E,F) In the sigmoid colon, the random forest analysis found that use of the top 6 important features (Gallionella, Aggregatibacter, Gordonibacter, Selenomonas_1, Pediococcus, and Oribacterium) to build the prediction model yielded an AUC value of 0.87 (95% CI: 0.71–1) based on ROC analysis.