Antibiotic Disruption of the Gut Microbiota Enhances the Murine Hepatic Dysfunction Associated With a High-Salt Diet

Abstract

Epidemiological and experimental evidence indicates that antibiotic exposure is related to metabolic malfunctions, such as obesity and non-alcoholic fatty liver disease (NAFLD). Liver impairment and hypertrophy of adipose cells are related to high salt consumption. This research aims to investigated the physiological mechanism of a high salt diet (HSD) enhanced antibiotic-induced hepatic injury and mitochondrial abnormalities in mice. The mice were fed a HSD with or without penicillin G (PEN) for 8 weeks and the gut metabolome, untargeted faecal metabolomics, and intestinal function were evaluated. The results revealed that HSD, PEN and their combination (HSPEN) significantly changed the gut microbial community. HSPEN mice exhibited more opportunistic pathogens (such as Klebsiella and Morganella) and reduced probiotic species (including Bifidobacterium and Lactobacillus). The main variations in the faecal metabolites of the HSPEN group were identified, including those connected with entero-hepatic circulation (including bile acids), tryptophan metabolism (i.e., indole derivatives) and lipid metabolism (e.g., erucic acid). Furthermore, increased intestinal permeability and immunologic response caused greater hepatic damage in the HSPEN group compared to the other groups. These findings may have important implications for public health.

Keywords: antibiotic exposure; gut microbiome; hepatic steatosis; high-salt diet; mitochondrial function.

FIGURE 1

Experimental process.

FIGURE 2

High-salt diet, Penicillin G, and their combination caused liver injury in mice. The daily diet intake (A), fluid consumption (B), and body weights (C) of the different groups throughout the experiment. (D) Liver/body mass ratio. Values are presented as the mean ± SD; n = 8. (E) Representative mouse liver sections with H&E, ORO and Masson staining. Black arrows indicate ballooning degeneration in liver tissues. Green arrows indicate loss of cellular boundaries around the central vein of hepatocytes. Yellow arrows indicate obvious fibrosis in liver tissues.

FIGURE 3

HSD and antibiotic exposure induce mitochondrial dysfunction in the liver. (A) Swollen mitochondria and mitochondrial membrane rupture (red arrows) in the livers of mice as shown in typical TEM photomicrographs (×3000 magnification). (B) MitoTracker red was used to stain hepatocytes for 30 min and liver cells were analysed using confocal microscopy (scale bar: 10 μm). (C) The mRNA levels of the mitochondrial fusion marker Mfn1 and fission markers Drp1 and Fis1 in the liver tissue, normalized to levels in the liver tissue of the ND group. (D) Mitochondrial release of H₂O₂ in mouse liver. Data are shown as the mean ± SD; n = 8. Different letters were significantly different (p < 0.05).

FIGURE 4

HSD and antibiotic exposure induce dysregulation of the gut microbiome. (A) Heatmap showing normalised values of 18 phyla in the faeces of ND-, HSD-, PEN- and HSPEN-treated mice. The normalised abundance values are indicated from red to green, where red indicates that the abundance is the highest and green indicates that it is the lowest. (B) The chord diagram reveals the top 20 abundant genera (abundance > 0.2%) in the four groups. (C) Histogram scores from linear discriminant analysis (LDA) comparing inter-group variance level by LDA effect size (LEfSe) analysis of the relative abundance. (D) Microbial richness estimates (Chao 1 index) and diversity indices (Shannon−Wiener) in the different groups at the 8-weeks point. (E) Separation of the faecal bacterial structure expressed using PCoA in the four groups. Data are shown as the mean ± SD; n = 8. *p < 0.05, **p < 0.01 and ***p < 0.001.

FIGURE 5

Faecal metabolic analysis in the ND, HSD, PEN, and HSPEN groups. (A) Scatterplot of PCoA scores in the various groups. (B) The differential variables between the ND and HSPEN groups are shown via a volcano plot. Each metabolite is indicated by a dot, with down-regulation represented by red dots, up-regulation represented by blue dots, and no statistical difference represented by green dots. (C) Heatmap of 90 metabolites that were differentially (p < 0.01) abundant at standardised levels between the ND and HSPEN groups. The distances of the metabolites are expressed by the dendrogram according to their relative abundances. The normalised abundance values are described intuitively from red to blue, expressing the maximum and minimum abundances, respectively. (D) Enriched KEGG pathways in the ND group in contrast to the HSPEN group. The statistical significance values (p < 0.05) are described intuitively from red to green, showing the most and least differences, respectively. The size of the dot on the vertical axis indicates the metabolite count in the metabolic pathway.

FIGURE 6

The regulatory networks between 32 gut microbiota (p < 0.01) and 33 faecal metabolites (p < 0.001) in the ND group in comparison with the HSPEN group. The Pearson’s rank correlations (R-value > 0.40) are expressed as red and green lines, representing positive and negative correlations, respectively.

FIGURE 7

Effects of HSD and antibiotic exposure on colonic function and morphology in mice. Levels of DAO (A) and D-LA (B) in serum, which signify the intestinal permeability of the ND-, HSD-, PEN- and HSPEN-treated mice. (C) Photomicrographs (×200 magnification) of the colon in the various groups of mice. Inflammatory cell osmosis and oedema in the submucosal layer are expressed respectively by black and red arrows. (D) Microscopic observation of CD4⁺, CD8⁺, and IgA lymphocytes on the colon epithelium (×200 magnification). Image Pro-Plus 6.0 software was employed to analyse the average numbers of CD4⁺ (E), CD8⁺ (F), and IgA (G) lymphocytes in the colon tissues and the (H) pH values of the colonic constituents of the groups. Data are expressed as the mean ± SD; n = 8. Different letters were significantly different (p < 0.05).