Abnormal bile acid metabolism is an important feature of gut microbiota and fecal metabolites in patients with slow transit constipation

Abstract

Destructions in the intestinal ecosystem are implicated with changes in slow transit constipation (STC), which is a kind of intractable constipation characterized by colonic motility disorder. In order to deepen the understanding of the structure of the STC gut microbiota and the relationship between the gut microbiota and fecal metabolites, we first used 16S rRNA amplicon sequencing to evaluate the gut microbiota in 30 STC patients and 30 healthy subjects. The α-diversity of the STC group was changed to a certain degree, and the β-diversity was significantly different, which indicated that the composition of the gut microbiota of STC patients was inconsistent with healthy subjects. Among them, Bacteroides, Parabacteroides, Desulfovibrionaceae, and Ruminiclostridium were significantly upregulated, while Subdoligranulum was significantly downregulated. The metabolomics showed that different metabolites between the STC and the control group were involved in the process of bile acids and lipid metabolism, including taurocholate, taurochenodeoxycholate, taurine, deoxycholic acid, cyclohexylsulfamate, cholic acid, chenodeoxycholate, arachidonic acid, and 4-pyridoxic acid. We found that the colon histomorphology of STC patients was significantly disrupted, and TGR5 and FXR were significantly downregulated. The differences in metabolites were related to changes in the abundance of specific bacteria and patients' intestinal dysfunction. Analysis of the fecal genomics and metabolomics enabled separation of the STC from controls based on random forest model prediction [STC vs. control (14 gut microbiota and metabolite biomarkers)-Sensitivity: 1, Specificity: 0.877]. This study provided a perspective for the diagnosis and intervention of STC related with abnormal bile acid metabolism.

Keywords: 16S rRNA amplicon sequencing; bile acid metabolism; diagnosis and intervention; metabolomics; slow transit constipation.

Figure 1

The shift of gut microbiome in slow transit constipation (STC) and control C subjects accroding to the 16S rRNA data. (A) Venn diagram of the observed OTUs in STC and C. (B) The relative abundance of the two groups at the genus level (Top 10). *p-value < 0.05 and ** p-value < 0.01, t-test. (C, D) Difference analysis of α-diversity index between the STC and C groups. Boxplot of difference between groups of Shannon index (C) with p = 0.094, t-test, and Simpson (D) with p = 0.021, Wilcoxon rank-sum test. (E) Principal coordinate analysis (PCoA) of the microbiota based on the unweighted UniFrac distance metrics for the STC and C groups. ANOSIM, R = 0.134, p = 0.001. (F) The differences between the STC and C groups were observed based on Non-Metric Multi-Dimensional Scaling (NMDS). (G, H) Cladograms generated by LEfSe indicating differences in the bacterial taxa between the STC and C groups. Red bars indicate taxa with enrichment in the C group, and green bars indicate taxa with enrichment in the STC group. (I) The LEfSe analysis of KEGG pathway (Welch’s t-test test).

Figure 2

Random forest model prediction of potential diagnostic biomarkers of gut microbiota between STC patients and healthy subjects. (A) Classification performance of a random forest model using 16S rRNA abundance of 14 different bacterial genera or species. The cross-validation prediction performance of models with increasing number of predictors in order, and sorted by importance. (B) ROC curve displaying the classification for STC and C employing five potential diagnostic gut microbiota biomarkers (AUC = 0.785). (C) The abundance of 5 potential diagnostic gut microbiota biomarkers in each sample including g:Bacteroides, g:Parabacteroides, f:Desulfovibrionaceae, g:Ruminiclostridium 5, and g:Subdoligranulum. (D) Co-occurrence network of five potential diagnostic gut microbiota biomarkers in both the STC group and the control group based on the Spearman correlation algorithms. Each node presents a bacterial genus or species. The node size indicates the relative importance of each genus or species, and the density of the edges represents the Spearman coefficient. Red links stand for positive interactions between nodes, and green links stand for negative interactions.

Figure 3

The fecal metabolites of STC patients and healthy subjects are significantly different. (A) The different metabolites were visualized in the form of volcano plot. The abscissa is the logarithm value of log2 of the fold change, and the ordinate is the logarithm value of −log10 of significance p-value (STC vs. C). Metabolites of difference with FC > 1.5, p-value < 0.05 are represented by rose red, while those with FC < 0.67 and p-value < 0.05 are shown in blue. Metabolites that are not significantly different are shown in black. (B) The score plot of orthogonal partial least-squares discriminant analysis (OPLS-DA), where t[1] represents principal component 1, t[2] represents principal component 2, and the ellipse represents the 95% confidence interval. The distribution of points reflects the degree of difference between groups and within group. The model evaluation parameter Q ² obtained by sevenfold cross-validation is 0.311. (C) Permutation test of OPLS-DA. The x-coordinate represents the degree of permutation, and the y-coordinate represents the values of R ² and Q ². The green dot represents R ², the blue dot denotes Q ², and the two dashed lines represent the regression lines of R ² and Q ², respectively. (D) Top 20 enriched KEGG pathways of 21 screened differential metabolites (VIP value ≥ 1 and p-value < 0.05). (E) Classification performance of a random forest model using abundance of 21 differential metabolites based on multivariate ROC curve exploratory analysis. The cross-validation prediction performance of models with inreasing number of predictors in order, and sorted by importance. (F) The receiver operating characteristic curve (ROC) of 12 potential diagnostic biomarker metabolites [each area under curve (AUC) ≥ 0.7]. (G) The abundance of 9 different metabolites involved in the BA synthesis, metabolism, secretion, and lipid metabolism, which was believed to have diagnostic efficacy. Significance compared with the control group, **P < 0.01 or *P < 0.05 vs. the control group. (H) ROC curve displaying the classification for STC and C employing 9 diagnostic biomarker metabolites (AUC = 0.81).

Figure 4

The morphology and BA-related receptor expression in colon tissues of STC patients and healthy subjects. (A) HE staining of colon tissues. Red arrows indicated inflammatory cell infiltration, and black arrows indicated mucosa layer structure of colon tissues. (B) Immunohistochemical staining results of TGR5 and FXR. The number of positive cells was quantitatively analyzed in six random selected fields of the same size using NIH ImageJ software. Significance compared with the control group, **p < 0.01 or *p < 0.05 vs. the control group.

Figure 5

The gene and protein expression levels of TGR5 and FXR in colon tissues of STC patients and healthy subjects. (A) qRT-PCR results of gene expression of TGR5 and FXR. (B) Gene primer sequence. The relative gene expression levels were calculated by the 2^-△△CT method with GAPDH as the internal reference (n = 6). (C) Western blotting results of protein expression of TGR5 and FXR. NIH ImageJ software was used to quantify the relative optical density of protein bands. Values are means ± SD (n = 6). The protein expression was detected in the same gel in which GAPDH was used as the internal control. Significance compared with the control group, **p < 0.01 or *p < 0.05 vs. the control group.

Figure 6

The summary of gut microbiota composition and metabolism analysis between STC and healthy subjects. Compared to the healthy subjects, the relative abundance of Bacteroides, Parabacteroides, Desulfovibrionaceae, and Ruminiclostridium 5 were increased while Subdoligranulum was decreased significantly in feces of STC patients. STC patients displayed alternation of gut microbiota composition. STC patients displayed decreased level in metabolites that were associated with BA synthesis, metabolism, secretion, and lipid metabolism. Experimental studies on human colon histology showed decreased expression of TGR5 and FXR in STC patients. Wexner constipation score was negatively correlated with these diagnostic metabolites, while GIQLI and BSFS scores were positively correlated. There was obvious positive correlation among metabolites. There was a certain negative correlation between the differential microbiota and the differential metabolites. Comprehensive use of 14 diagnostic gut microbiota and metabolite biomarkers may accurately distinguish STC and healthy population with AUC = 0.877.