Gut Microbiota-Tryptophan Metabolism-GLP-1 Axis Participates in β-Cell Regeneration Induced by Dapagliflozin

Abstract

Sodium-glucose cotransporter 2 inhibitors, efficacious antidiabetic agents that have cardiovascular and renal benefits, can promote pancreatic β-cell regeneration in type 2 diabetic mice. However, the underlying mechanism remains unclear. In this study, we aimed to use multiomics to identify the mediators involved in β-cell regeneration induced by dapagliflozin. We showed that dapagliflozin lowered blood glucose level, upregulated plasma insulin level, and increased islet area in db/db mice. Dapagliflozin reshaped gut microbiota and modulated microbiotic and plasmatic metabolites related to tryptophan metabolism, especially l-tryptophan, in the diabetic mice. Notably, l-tryptophan upregulated the mRNA level of glucagon-like peptide 1 (GLP-1) production-related gene (Gcg and Pcsk1) expression and promoted GLP-1 secretion in cultured mouse intestinal L cells, and it increased the supernatant insulin level in primary human islets, which was eliminated by GPR142 antagonist. Transplant of fecal microbiota from dapagliflozin-treated mice, supplementation of l-tryptophan, or treatment with dapagliflozin upregulated l-tryptophan, GLP-1, and insulin or C-peptide levels and promoted β-cell regeneration in db/db mice. Addition of exendin 9-39, a GLP-1 receptor (GLP-1R) antagonist, or pancreatic Glp1r knockout diminished these beneficial effects. In summary, treatment with dapagliflozin in type 2 diabetic mice promotes β-cell regeneration by upregulating GLP-1 production, which is mediated via gut microbiota and tryptophan metabolism.

Graphical abstract

Figure 1

Alteration of the plasma metabolomic profile induced by dapagliflozin (Dapa) in db/db mice. A: Principal component (PC) analysis plots of plasma metabolomic profiles. B: Differential plasmatic metabolites in the top 20 of VIP scores. C: Metabolic pathway analysis of differential plasmatic metabolites. The circles that are closer to the top-right diagonal region are significantly changed, and they are more likely to have significant impacts on the pathways. D: Relative levels of l-tryptophan, indoleacetate, and indolepyruvate in plasma. n = 5 − 6 mice/group. Data are mean ± SEM. Statistical analysis was performed by one-way ANOVA followed by Dunnett T3 multiple comparisons test in D. §P < 0.05 (vs. db/m); *P < 0.05 (vs. control [Ctrl]). 2,5-pyrrolidinedione, 3,3-dichloro-4-(dichloromethylene)-2,5-pyrrolidinedione; 3-methyl-N-butanamide, 3-methyl-N-(2-naphthyl)butanamide; N-γ-glutamylhistamine, N(α)-γ-l-glutamylhistamine; sig, significant; S-mercaptopyruvate, S-(4-bromophenyl)-mercaptopyruvate.

Figure 2

The composition of gut microbial communities after treatment with dapagliflozin (Dapa) in db/db mice. A and B: Abundance-based coverage estimator (ACE) and Simpson index to describe the α-diversity of gut microbial communities. C: Nonmetric multidimensional scaling (MDS) to describe the β-diversity of gut microbial communities. D: The top 14 maximum abundances of gut microbiota at the genus level. E: The top 30 differential gut microbiota at the genus level. F: Cladogram representation of differentially abundant taxonomic clades with linear discriminant analysis score >4. Significantly distinct taxonomic nodes are colored, and the branch areas are shaded according to the effect size of the taxa. n = 6 mice/group. Data are the mean ± SEM or median (interquartile range). Statistical analysis was performed by one-way ANOVA followed by Tukey multiple comparisons test in A or by Kruskal-Wallis test followed by Dunn multiple comparisons test in B. Ctrl, control.

Figure 3

The relationship among the gut microbiota, microbiotic metabolites, and GLP-1 production. A: Principal component (PC) analysis plots of microbial metabolomic profiles. B: Spearman correlation analysis between the top 30 differential gut microbiota and differential microbiotic metabolites related to l-tryptophan (Trp) metabolism. C: The number of overlapping metabolites between microbiotic and plasmatic metabolites. D and E: Relative Gcg and Pcsk1 mRNA levels in the intestinal L-cell line STC1 cells incubated with different concentrations of Trp. F: GLP-1 secretion in STC1 cells cultured with different concentrations of Trp. G–I: Relative Gcg and Pcsk1 mRNA levels and GLP-1 secretion in STC1 cells incubated with Trp (10 mmol/L) or vehicle with or without CLP-3094 (CLP) (10 μmol/L). n = 4. Data are mean ± SEM. Statistical analysis was performed by one-way ANOVA followed by Dunnett multiple comparisons test in D and E, one-way ANOVA followed by Tukey multiple comparisons test in G and H, or one-way ANOVA followed by Dunnett T3 multiple comparisons test in F and I. *P < 0.05 (vs. control [Ctrl]); ‡P < 0.05 (vs. Trp).

Figure 4

Histological analysis in the pancreatic tissues of db/db mice receiving FMT from dapagliflozin (Dapa)- or vehicle-treated mice with or without exendin 9-39 (Ex9) for 5 weeks. A: Representative photographs of the entire pancreata immunostained for glucagon and insulin. Scale bar = 2,000 μm. B: Quantification of the islet area per pancreatic area. C: Quantification of the islet number per section. D: Quantification of the small islet number per section. An islet with a cell number less than eight is defined as a small islet. E: Representative photographs of an islet immunostained for glucagon and insulin. Scale bar = 20 μm. F: Quantification of the β-cell area per pancreatic area. G: Quantification of the α-cell area per pancreatic area. H: The ratio of β-cell area to α-cell area. n = 3 sections/mouse and 3 mice/group. Data are mean ± SEM or median (interquartile range). Statistical analysis was performed by one-way ANOVA followed by Dunnett T3 multiple comparisons test in B and C, by Kruskal-Wallis test followed by Dunn multiple comparisons test in F, or by one-way ANOVA or two-way ANOVA followed by Tukey multiple comparisons test as appropriate in D, G, and H. *P < 0.05 (vs. control [Ctrl]-FMT + vehicle).

Figure 5

Histological analysis in the pancreatic tissues of db/db mice treated with vehicle, l-tryptophan (Trp), or Trp combined with exendin 9-39 (Ex9) for 5 weeks. Age-matched db/m mice treated with vehicle were included as normal control (Ctrl). A: Representative photographs of the entire pancreata immunostained for glucagon and insulin. Scale bar = 2,000 μm. B: Quantification of the islet area per pancreatic area. C: Quantification of the islet number per section. D: Quantification of the small islet number per section. An islet with a cell number less than eight is defined as a small islet. E: Representative photographs of an islet immunostained for glucagon and insulin. Scale bar = 20 μm. F: Quantification of the β-cell area per pancreatic area. G: Quantification of the α-cell area per pancreatic area. H: Ratio of β-cell area to α-cell area. n = 3 sections/mouse and 3 mice/group. Data are mean ± SEM. Statistical analysis was performed by one-way ANOVA followed by Dunnett T3 multiple comparisons test in B–D and F, or by one-way ANOVA or two-way ANOVA followed by Tukey multiple comparisons test as appropriate in G and H. §P < 0.05 (vs. db/m); *P < 0.05 (vs. Ctrl); ‡P < 0.05 (vs. Trp).

Figure 6

Histological analysis in the pancreatic tissues of db/db mice treated with vehicle, dapagliflozin (Dapa), or Dapa with exendin 9-39 (Ex9) for 5 weeks. Age-matched db/m mice treated with vehicle served as normal control (Ctrl). A: Representative photographs of the entire pancreata immunostained for glucagon and insulin. Scale bar = 2,000 μm. B: Quantification of the islet area per pancreatic area. C: Quantification of the islet number per section. D: Quantification of the small islet number per section. An islet with a cell number less than eight is defined as a small islet. E: Representative photographs of an islet immunostained for glucagon and insulin. Scale bar = 20 μm. F: Quantification of the β-cell area per pancreatic area. G: Quantification of the α-cell area per pancreatic area. H: Ratio of β-cell area to α-cell area. n = 3–4 sections/mouse and 4 mice/group. Data are mean ± SEM or median (interquartile range). Statistical analysis was performed by one-way ANOVA followed by Dunnett T3 multiple comparisons test in C, D, and G, two-way ANOVA followed by Tukey multiple comparisons test in H, or Kruskal-Wallis test followed by Dunn multiple comparisons test in B and F. §P < 0.05 (vs. db/m); *P < 0.05 (vs. Ctrl); ‡P < 0.05 (vs. Dapa).

Figure 7

Histological analysis in the pancreatic tissues of diabetic Glp1r^pan−/− (KO) mice and Flox/Cre littermates treated with dapagliflozin (Dapa) or vehicle control (Ctrl) for 6 weeks. Diabetes was induced by high-fat diet and streptozotocin before the treatment. A: Representative photographs of the entire pancreata immunostained for glucagon and insulin. Scale bar = 2,000 μm. B: Quantification of the islet area per pancreatic area. C: Quantification of the islet number per section. D: Quantification of the small islet number per section. An islet with a cell number less than eight is defined as a small islet. E: Representative photographs of an islet immunostained for glucagon and insulin. Scale bar = 20 μm. F: Quantification of the β-cell area per pancreatic area. G: Quantification of the α-cell area per pancreatic area. H: The ratio of β-cell area to α-cell area. n = 3 sections/mouse and 3 mice/group. Data are mean ± SEM or median (interquartile range). Statistical analysis was performed by one-way ANOVA or two-way ANOVA followed by Tukey multiple comparisons test as appropriate in C, D, and H, one-way ANOVA followed by Dunnett T3 multiple comparisons test in B and F, or Kruskal-Wallis test followed by Dunn multiple comparisons test in G. *P < 0.05 (vs. Ctrl in the same genotype).