High-fat diet-disturbed gut microbiota-colonocyte interactions contribute to dysregulating peripheral tryptophan-kynurenine metabolism

Abstract

Background: Aberrant tryptophan (Trp)-kynurenine (Kyn) metabolism has been implicated in the pathogenesis of human disease. In particular, populations with long-term western-style diets are characterized by an excess of Kyn in the plasma. Host-gut microbiota interactions are dominated by diet and are essential for maintaining host metabolic homeostasis. However, the role of western diet-disturbed gut microbiota-colonocyte interactions in Trp metabolism remains to be elucidated.

Results: Here, 4-week-old mice were fed with a high-fat diet (HFD), representing a typical western diet, for 4 weeks, and multi-omics approaches were adopted to determine the mechanism by which HFD disrupted gut microbiota-colonocyte interplay causing serum Trp-Kyn metabolism dysfunction. Our results showed that colonocyte-microbiota interactions dominated the peripheral Kyn pathway in HFD mice. Mechanistically, persistent HFD-impaired mitochondrial bioenergetics increased colonic epithelial oxygenation and caused metabolic reprogramming in colonites to support the expansion of Proteobacteria in the colon lumen. Phylum Proteobacteria-derived lipopolysaccharide (LPS) stimulated colonic immune responses to upregulate the indoleamine 2,3-dioxygenase 1 (IDO1)-mediated Kyn pathway, leading to Trp depletion and Kyn accumulation in the circulation, which was further confirmed by transplantation of Escherichia coli (E.coli) indicator strains and colonic IDO1 depletion. Butyrate supplementation promoted mitochondrial functions in colonocytes to remodel the gut microbiota in HFD mice, consequently ameliorating serum Kyn accumulation.

Conclusions: Our results highlighted that HFD disrupted the peripheral Kyn pathway in a gut microbiota-dependent manner and that the continuous homeostasis of gut bacteria-colonocytes interplay played a central role in the regulation of host peripheral Trp metabolism. Meanwhile, this study provided new insights into therapies against western diet-related metabolic disorders. Video Abstract.

Keywords: Gut microbiota; High-fat diet; Kynurenine; Mitochondrial dysfunction; Tryptophan metabolism.

Fig. 1

High-fat diet-disturbed tryptophan metabolism. a Experimental workflow started with animal dietary interventions and sample collection (n = 12 for each group). Multi-omics analysis was conducted to elucidate the mechanisms by which high-fat diet (HFD) affects metabolites. b Principal component analysis (PCA) score plot to assess serum metabolomic data comparing HFD-fed mice with standard diet-fed mice (Chow) (n = 6 for each group). Permutational multivariate analysis of variance (PERMANOVA) by Adonis was used to determine statistical significance. c Chemical similarity enrichment analysis (ChemRICH) clustering of 614 statistically HFD-altered serum metabolites by chemical similarity with x-axis of mediation logarithmic additive octanol–water partition coefficients (XlogP) and y-axis for sets statistical significance (Kolmogorov–Smirnov test, p-value < 0.05); the node size depicted total compound numbers for each cluster set and the node color scale the proportion of Chow-enriched vs. HFD-enriched metabolites. d Quantitative metabolite set enrichment analysis (qMSEA) based on 99 metabolite sets associated with human metabolic pathways identified the top 25 serum metabolic pathways significantly perturbed by HFD (p-value < 0.05). e HFD significantly upregulated the kynurenine (Kyn) metabolic pathway. Data are represented as mean ± SD and normalized to the Chow. In e, p-values were determined by independent samples t-test. NS not significant, * p-value ≤ 0.05, ** p-value ≤ 0.01, *** p-value ≤ 0.001. LC–MS liquid chromatography-mass spectrometry; Dim dimension

Fig. 2

HFD-induced gut dysbiosis was highly correlated with serum metabolic profiles. a Gut bacterial alpha diversity was estimated by Shannon index (n = 8 for each group). b Principal coordinate analysis (PCoA) plot showing microbial compositional differences quantified by Bray–Curtis distance (PERMANOVA by Adonis). c Bar graph of bacterial abundance at the phylum level. d HFD enhanced the Firmicutes/Bacteroidetes ratio in the fecal microbiota. e Cladogram generated from linear discriminant analysis effect size (LEfSe) showing the most differentially enriched bacterial taxa in the colonic contents of Chow (green) or HFD (blue) mice (LDA value = 2.0; p-value < 0.05). f Heatmap of bacterial taxa abundance determined by LEfSe analysis. g Pairwise comparisons of HFD-altered bacterial taxa are shown, with a color gradient indicating Pearson’s correlation coefficients. HFD-altered serum metabolites (blue: HFD-enriched serum metabolites; green: serum metabolites downregulated in HFD-fed mice) were related to each bacterial taxon by Mantel tests. Edge width corresponds to Mantel’s r statistic for the corresponding distance correlations, and edge color indicates the statistical significance based on 999 permutations. h The top 10 bacterial biomarkers were identified by random forests regression of relative abundances of LEfSe-determined bacterial taxon against serum Kyn concentration. Statistical significance of selected bacterial biomarkers was assessed by permutation test (999 times). Data are represented as mean ± SD. In a and d, p-values were determined by independent samples t-test. NS not significant, * p-value ≤ 0.05, ** p-value ≤ 0.01, *** p-value ≤ 0.001. IncMSE increase in mean squared error

Fig. 3

The causal role of gut microbiota in mediating HFD-induced dysregulation of peripheral tryptophan-kynurenine metabolism. a Fecal pellets were collected from HFD mice at 8 weeks and then used to perform fecal microbiota transplantation (FMT). Standard diet-fed mice were treated with an antibiotic cocktail for 3 days, followed by FMT lasting 4 weeks (C-FMT). HFD-fed mice were concurrently treated with an antibiotic cocktail for 4 weeks (H-Abx) to further verify the critical role of gut microbiota in diet-mediated disruption of tryptophan (Trp)-kynurenine (Kyn) metabolism. b,c PCoA based on Bray–Curtis distance showing the similarity of microbiota composition in mice after FMT to that of the donor (n = 8 for each group). d Bar graph of bacterial abundance at the phylum level. e–f C-FMT mice exhibited a taxonomic profile of gut bacteria more comparable to that of HFD mice. g PCA was performed to assess the serum metabolic profile in standard diet-fed mice after FMT donated by HFD mice (n = 6 for each group; PERMANOVA by Adonis). h FMT donated by HFD mice enhanced the peripheral Trp-Kyn pathway in standard diet-fed mice. i Clearing gut bacteria by antibiotic cocktail abolished HFD-induced dysregulation of peripheral Trp-Kyn metabolism. Data are represented as mean ± SD. In h, data are normalized to Chow after FMT. In i, data are normalized to HFD after antibiotic cocktail treatment. In c, e, and f, p-values were determined by one-way analysis of variance (ANOVA) with Dunnett’s multiple comparison test. In h,i, statistical significance was assessed by independent samples t-test. NS not significant, * p-value ≤ 0.05, ** p-value ≤ 0.01, *** p-value ≤ 0.001

Fig. 4

Hyperactivated indoleamine 2,3-dioxygenase 1 in colonocytes mediated the gut dysbiosis-induced dysregulation of peripheral Trp-Kyn metabolism. a Volcano plot showing the HFD-altered transcriptomes in colonocytes (n = 3 for each group). b Gene set enrichment analysis (GSEA) showing enrichment of the Trp metabolism gene sets (left) and heatmap of involved genes (right) (n = 3 for each group). c Fluorescent immunostaining of indoleamine 2,3-dioxygenase 1 (IDO1) (green) in mouse colon sections. Nuclei were counterstained with DAPI (blue) (n = 9 slices from 3 mice). d Enriched gene sets of lipopolysaccharide (LPS)-induced inflammation (left) and the heatmap of involved genes (right) (n = 3 for each group). e HFD enhanced the LPS concentration in the colonic contents (n = 8 for each group). f Escherichia coli (E. coli) indicator strains were isolated from the feces of HFD mice and enriched in vitro for bacterial transplantation to confirm the causal role of Proteobacteria expansion in HFD-induced dysregulation of Kyn pathway (C-E.coli). To determine the central role of colonic IDO1 in E.coli-mediated upregulation of the Kyn pathway, mice were administrated with palmatine to eliminate colonic IDO1 before receiving E.coli (C-Pal). g E. coli transplantation enhanced LPS concentration in the colonic contents (n = 8 for each group). h IDO1 activity in the colon (n = 9 slices from 3 mice). i E. coli transplantation upregulated the serum Kyn pathway, which was reversed by palmatine pre-treatment (n = 6 for each group). Data are represented as mean ± SD. In e, g, and i, data are normalized to Chow. In c and e, p-values were determined by independent samples t-test. In g–i, statistical significance was assessed by ANOVA with Dunnett’s multiple comparison test. NS not significant, * p-value ≤ 0.05, ** p-value ≤ 0.01, *** p-value ≤ 0.001

Fig. 5

HFD disrupted mitochondrial function and caused metabolic reorientation in colonocytes. a HFD reduced butyrate concentrations in colonic contents (n = 6 for each group). HFD inhibited mitochondrial gene expression (b) and mitochondrial oxidative phosphorylation (c). d Binding of pimonidazole (red) was used to qualify the oxygen gradient in the colon. Nuclei were counterstained with DAPI (blue). e Pimonidazole staining was quantified by scoring blinded sections of the colon (n = 9 slices from 3 mice). f Adenosine triphosphate (ATP) concentrations in colonic tissue (n = 8 for each group). g HFD downregulated HIF-α-mediated epithelial hypoxia in colonic tissue. h HFD induced a shift in the metabolism of short-chain fatty acids (SCFAs) towards glycolysis in colonic tissue. i PCA score plot for assessing colonic metabolomic data comparing HFD mice with Chow mice (n = 6 for each group; PERMANOVA by Adonis). j qMSEA method was adopted to identify the colonic metabolic pathways perturbed by HFD, and the top 30 were shown. HFD enhanced the lactate in colonic contents (k, n = 6) and nitrate in colonic tissue (l, n = 8). Data are represented as mean ± SD. In a, f, k, and l, data are normalized to Chow, and statistical significance was assessed by independent samples t-test. In e, the p-value was determined by the Wilcoxon test. NS not significant, * p-value ≤ 0.05, ** p-value ≤ 0.01, *** p-value ≤ 0.001. NES normalized enrichment scores

Fig. 6

Colonic mitochondrial dysfunction-induced gut dysbiosis disturbed systemic Trp-Kyn metabolism. a Standard diet-fed mice were administrated with GW9662 to examine the relationship between colonic mitochondrial dysfunction and Trp metabolism (C-GW). To confirm the beneficial effect of remodeling colonic mitochondrial bioenergetics on systemic Trp-Kyn metabolism, butyrate was provided to HFD-fed mice in drinking water (H-Buy). b Colonic epithelial oxygenation was determined by pimonidazole (red). Nuclei were counterstained with DAPI (blue) (n = 9 slices from 3 mice). Concentrations of ATP (c) and nitrate (d) in colonic tissue (n = 8 for each group). e PCoA showed the similarity of gut microbiota determined by 16S profiling (n = 8 for each group; PERMANOVA by Adonis). f Relative abundance of the bacterial phylum. g Relative abundance of phylum Proteobacteria, family Enterobacteriaceae, and genus Desulfovibrio and the ratio of Firmicutes/Bacteroidetes. h Fluorescent immunostaining of IDO1 (green) in mouse colon sections. Nuclei were counterstained with DAPI (blue) (n = 9 slices from 3 mice). i PCA score plot showing serum metabolic profiles (n = 6 for each group; PERMANOVA by Adonis). j Serum metabolites involved in the Kyn pathway (n = 6 for each group). Data are represented as mean ± SD. In c, d, and j, data are normalized to Chow or HFD respectively. In b, the p-value was determined by the Wilcoxon test. Other statistical procedures were performed by independent samples t-test. NS not significant, * p-value ≤ 0.05, ** p-value ≤ 0.01, *** p-value ≤ 0.001

Fig. 7

Schematic diagram of the mechanism underlying HFD-dysregulated peripheral Trp-Kyn metabolism. a In the healthy gut, the gut microbiota is dominated by obligate anaerobic bacteria that convert fiber into fermentation products, such as butyrate, to maintain colonocyte metabolism. b Sustained HFD reduced colonic butyrate concentrations and enhanced β-oxidation of long- and very long-chain fatty acids in colonic cells, which impaired mitochondrial bioenergetics and caused a metabolic reprogramming of colonic cells, allowing Proteobacteria to thrive. Bacteria-derived LPS stimulated colonic immune responses to upregulate the IDO1-mediated Kyn pathway