Bile acid-receptor TGR5 deficiency worsens liver injury in alcohol-fed mice by inducing intestinal microbiota dysbiosis

Abstract

Background & aims: Bile-acid metabolism and the intestinal microbiota are impaired in alcohol-related liver disease. Activation of the bile-acid receptor TGR5 (or GPBAR1) controls both biliary homeostasis and inflammatory processes. We examined the role of TGR5 in alcohol-induced liver injury in mice.

Methods: We used TGR5-deficient (TGR5-KO) and wild-type (WT) female mice, fed alcohol or not, to study the involvement of liver macrophages, the intestinal microbiota (16S sequencing), and bile-acid profiles (high-performance liquid chromatography coupled to tandem mass spectrometry). Hepatic triglyceride accumulation and inflammatory processes were assessed in parallel.

Results: TGR5 deficiency worsened liver injury, as shown by greater steatosis and inflammation than in WT mice. Isolation of liver macrophages from WT and TGR5-KO alcohol-fed mice showed that TGR5 deficiency did not increase the pro-inflammatory phenotype of liver macrophages but increased their recruitment to the liver. TGR5 deficiency induced dysbiosis, independently of alcohol intake, and transplantation of the TGR5-KO intestinal microbiota to WT mice was sufficient to worsen alcohol-induced liver inflammation. Secondary bile-acid levels were markedly lower in alcohol-fed TGR5-KO than normally fed WT and TGR5-KO mice. Consistent with these results, predictive analysis showed the abundance of bacterial genes involved in bile-acid transformation to be lower in alcohol-fed TGR5-KO than WT mice. This altered bile-acid profile may explain, in particular, why bile-acid synthesis was not repressed and inflammatory processes were exacerbated.

Conclusions: A lack of TGR5 was associated with worsening of alcohol-induced liver injury, a phenotype mainly related to intestinal microbiota dysbiosis and an altered bile-acid profile, following the consumption of alcohol.

Lay summary: Excessive chronic alcohol intake can induce liver disease. Bile acids are molecules produced by the liver and can modulate disease severity. We addressed the specific role of TGR5, a bile-acid receptor. We found that TGR5 deficiency worsened alcohol-induced liver injury and induced both intestinal microbiota dysbiosis and bile-acid pool remodelling. Our data suggest that both the intestinal microbiota and TGR5 may be targeted in the context of human alcohol-induced liver injury.

Keywords: ALD, alcohol-related liver diseases; ALT, alanine aminotransferase; Alc, alcohol; Alcoholic liver disease; BA, bile acids; BHI, brain heart infusion; Bile acid; C57, conventional mice; C57C57, conventional mice transplanted with their own IM; CA, cholic acid; CCL, CC motif chemokine ligands; CDCA, chenodeoxycholic acid; Col1a1, collagen type-I alpha-1 chain; DCA, deoxycholic acid; Dysbiosis; FDR, false-discovery rate; FXR, farnesoid X receptor; Gut-liver axis; IM, intestinal microbiota; Inflammation; KC, Kupffer cells; KO, knockout; Kupffer cells; LCA, lithocholic acid; LDA, linear discriminative analysis; LEfsE, LDA effect size; MCA, muricholic acid; MO, monocytes/macrophages; Microbiome; NFkB, nuclear factor-kappa B; OTU, operational taxonomic unit; PCA, principal component analysis; PCoA, principal coordinate analysis; PICRUSt, phylogenetic investigation of communities by reconstruction of unobserved states; RIN, RNA integrity number; TBA, total bile acids; TG, triglycerides; TGF, transforming growth factor; TIMP1, tissue inhibitor of metalloproteinase 1; TNF, tumour necrosis factor; UDCA, ursodeoxycholic acid; WT, wild-type; WTKO, WT mice transplanted with the IM of TGR5-KO mice; alpha-SMA, alpha-smooth muscle actin; mMMP9, matrix metallopeptidase 9.

Graphical abstract

Fig. 1

Liver injury in WT and TGR5-deficient mice after chronic alcohol consumption. Wild-type (WT) and Takeda-G-protein-receptor-5-deficient (TGR5-KO) mice were fed alcohol (Alc) or isocaloric maltodextrin (Ctrl). (A) Representative histological images of H&E and Oil Red O staining of the liver of alcohol-fed mice (scale bar: H&E = 100 μm and Oil Red O = 400 μm). (B) Quantification of Oil Red O staining. (C) Plasma ALT levels. (D–E) Liver mRNA levels of (D) pro-inflammatory cytokines and chemokines and (E) genes related to the macrophage phenotype. Kruskall-Wallis test and Dunn’s multiple comparison post-hoc test, ∗p <0.05, ∗∗p <0.01, ∗∗∗p <0.001. WT Ctrl n = 4, WT Alc n = 6, TGR5-KO Ctrl n = 4, TGR5-KO Alc n = 6. ALT, alanine aminotransferase.

Fig. 2

Intestinal microbiota of control and alcohol-fed WT and TGR5-deficient mice. WT and TGR5-KO mice were fed alcohol (Alc) or isocaloric maltodextrin (Ctrl). (A) Bray Curtis PCoA Unifrac distances showing a difference in the composition of the faecal microbiota between groups (p = 0.001). (B) Box plots showing alpha diversity based on the Shannon Index (Kruskall-Wallis test and Dunn’s multiple comparison post-hoc test). (C) Cladogram showing taxa with the largest differences (LDA >2) in abundance by LEfSe analysis. (D) LEfSe for the predicted metagenome metabolic pathways (KEGG modules). Only taxa with an LDA >2 and p <0.05 are shown (Wilcoxon test). WT Ctrl n = 3, WT Alc n = 5, TGR5-KO Ctrl n = 4, TGR5-KO Alc n = 6. LDA, linear discriminative analysis; LEfSe, linear discriminative analysis effect size; PCoA, principal coordinate analysis; WT, wild-type.

Fig. 3

Faecal microbiota transplantation from TGR5-KO to WT mice worsens alcohol-induced liver injury. (A) Experimental design. (B–D) Wild-type (WT), Takeda-G-protein-receptor-5-deficient (TGR5-KO), and WT mice transplanted with the IM of TGR5-KO mice (WT^KO) were fed alcohol. (B) Liver mRNA levels of pro-inflammatory cytokines and chemokines. (C) Plasma ALT and quantification of Oil Red O staining. (D) Representative histological images of the liver (upper, scale bar = 100 μm), Oil Red O staining (middle, scale bar = 400 μm), and F4/80 labelling (lower, scale bar = 100 μm). (E) Liver mRNA expression of F4/80 and CD68. Kruskall-Wallis test and Dunn’s multiple comparison post-hoc test, ∗p <0.05, ∗∗p <0.01, ∗∗∗p <0.001. WT n = 5, WT^KO n = 5, KO n = 9. IM, intestinal microbiota.

Fig. 4

Liver injury in alcohol-fed mice transplanted with their own IM and analysis of the IM. C57BL/6J (C57) mice were fed alcohol. A group was transplanted with their own IM (C57^C57). (A) Plasma ALT and liver TG levels. (B) Liver mRNA levels of pro-inflammatory cytokines and chemokines. (C) Bray Curtis PCoA Unifrac distances showing no difference in the composition of the faecal microbiota between groups. Mann-Whitney test. C57 n = 9, C57^C57 n = 9. ALT, alanine aminotransferase; IM, intestinal microbiota; PCoA, principal coordinate analysis; TG, triglycerides.

Fig. 5

Intestinal microbiota of alcohol-fed mice transplanted with the IM of TGR5-deficient mice. WT, KO, and WT^KO mice were fed alcohol. (A) Unweighted PCoA Unifrac distances showing the composition of the IM (ANOSIM test, p = 0.001). (B) Histogram showing the relative abundance between phyla (FDR test, p <0.05) (upper panel) and Deferribacteres (FDR test, p <0.05) (lower panel). (C–E) Comparisons between WT and WT^KO mice. (C) Cladogram showing taxa with the largest differences in abundance (LDA >2). (D) Plot showing differences in the relative abundance of taxa. (E) Venn diagram showing common taxa. (F) Predictive expression of choloylglycine hydrolase and BA metabolism genes. Kruskall-Wallis test and Dunn’s multiple comparison post-hoc test, ∗p <0.05. WT n = 5, WT^KO n = 5, KO n = 9. BA, bile acid; IM, intestinal microbiota; KO, knockout; LDA, linear discriminative analysis; PCoA, principal coordinate analysis; WT, wild-type.

Fig. 6

Bile acids in alcohol-fed WT and TGR5-deficient mice. WT and TGR5-KO mice were fed alcohol (Alc) or isocaloric maltodextrin (Ctrl). Bile-acid composition in the plasma (upper), liver (middle), and caecum (lower). (A) Total bile acids. (B) Ratio of secondary/primary bile acids. (C) Hydrophobic index. (D) Unconjugated bile acids. (E) Percentage of each bile acid. Kruskall-Wallis test and Dunn’s multiple comparison post-hoc test, ∗p <0.05, ∗∗p <0.01, ∗∗∗p <0.001. WT Ctrl n = 4, WT Alc n = 6, TGR5-KO Ctrl n = 4, TGR5-KO Alc n = 6. CA, cholic acid; CDCA, chenodeoxycholic acid; DCA, deoxycholic acid; HCA, hyocholic acid; LCA, lithocholic acid; MCA, muricholic acid; WT, wild-type.

Fig. 7

Quantification of mRNA levels for genes involved in bile-acid synthesis. WT and KO mice were fed alcohol (Alc) or isocaloric maltodextrin (Ctrl). (A–C) mRNA level of genes quantified by qPCR related to (A) hepatic bile-acid synthesis enzymes (Cyp7a1, Cyp8b1, Cyp27a1) and the FGF15 pathway in the (B) ileum (FGF15, SHP) and (C) liver FXR, SHP and SREBP1. Kruskall-Wallis test and Dunn’s multiple comparison post-hoc test, ∗p <0.05, ∗∗p <0.01. WT Ctrl n = 4, WT Alc n = 6, TGR5-KO Ctrl n = 4, TGR5-KO Alc n = 6. FGF15, fibroblast growth factor-15; KO, knockout; WT, wild-type.

Fig. 8

Graphical summary of TGR5 and IM in alcohol-fed mice. Total TGR5 deficiency in alcohol-fed mice induces dysbiosis of the intestinal microbiota, which is associated with a decrease in secondary BA levels in the plasma and liver. This decrease is associated with lower ileal FXR signaling and thus a decrease in FGF15 synthesis. Decreased FGF15 levels result in increased hepatic BA synthesis. The changes in the BA pool in the liver may result in liver FXR activation. These changes all result in an increase in liver injury. The solid lines are based on the results, whereas the dotted lines represent hypothetical relationships. BA, bile acids.