Susceptibility Factor TNF-α Synergizes with Polygonum multiflorum to Drive Idiosyncratic Liver Injury in Mice by Disrupting Gut Microbiota Composition and Hepatic Metabolite Homeostasis

Abstract

Background: Polygonum multiflorum (PM), known as a traditional Chinese herb renowned for its tonic properties, has been used medicinally for millennia. However, it has drawn attention significantly due to the potential to induce idiosyncratic drug-induced liver injury (IDILI) in recent years. Previous studies identified the TNF-α, the pro-inflammatory cytokine, as a key factor contributing to susceptibility to PM induced-IDILI (PM-IDILI). However, the effects by which TNF-α mediates PM-IDILI remain poorly understood.

Methods: This study sought to elucidate the role of TNF-α in PM-IDILI using a TNF-α-sensitized C57BL/6J mouse model, integrating analyses of the gut microbiota and metabolomics We employed biochemical analysis, inflammatory markers, inflammatory liver histopathological, sequencing of 16S rRNA gene, as well as untargeted metabolomics based on LC-MS to systematically evaluate the extent of liver injury and characterize alterations in gut microbiota and liver metabolites following PM administration in TNF-α pre-treated mice.

Results: The results demonstrated that PM treatment in TNF-α-sensitized mice significantly elevated levels of indicators as AST (3.6-fold compared to the control group, P < 0.05) and ALT (3.9-fold compared to the control group, P < 0.01), increased plasma levels of IL-6 and IL-1β (P < 0.05 or P < 0.01), induced infiltration of inflammatory cell substantially in the liver. TNF-α-mediated PM disrupted the intestinal microbiota structure, characterized by reduced abundance of Akkermansia and increased abundance of Lachnospiraceae_NK4A136_group, Bacteroides, Alloprevotella, and Blautia. Furthermore, hepatic metabolomics analysis revealed that significant perturbations in TNF-α + PM treated mice, particularly affecting glutathione metabolism, purine metabolism, and arachidonic acid metabolism pathways.

Conclusion: These findings suggest that TNF-α sensitization predisposes mice to PM-IDILI, potentially by disrupting gut microbial homeostasis and altering host hepatic metabolism. This research provides critical theoretical and experimental evidence relevant to the safe and effective clinical application of PM.

Keywords: Polygonum multiflorum; TNF-α; gut microbiota; idiosyncratic drug-induced liver injury; metabolomics; susceptibility.

Graphical abstract

Figure 1

Chromatograms of PM decoction (A) and five mixed reference standards (B). 1: TSG; 2: E8OG; 3: P8OG; 4: emodin; 5: physcion.

Figure 2

Effects of co-administration of TNF-α and PM on mouse liver. (A) Experimental design of animal studies. (B) Plasma level of AST. (C) Plasma level of ALT. (D) Plasma level of IL-6. (E) Plasma level of IL-1β. Data are means ± SEM. Significant differences indicated as: *P < 0.05, **P < 0.01 vs N group. (F) Pathological changes in liver tissue. Scale bar = 100 μm. (A-F, n = 8).

Figure 3

Co-administration of TNF-α and PM water decoction modulates the diversity of the gut microbiome in mice. (A) The Rank Abundance plot in each group. (B) PcoA analysis on intestinal microbiota. (C) NMDS analysis on intestinal microbiota. Alpha diversity indicates microbiota in the faecal contents of mice. (D) Chao1 index; (E) Shannon index; (F) Simpson index; Data are means ± SD. (A-F, n = 8).

Figure 4

Proportion of intestinal microbes in N, T, PH and TPH groups. (a) Phylum level relative abundance. (b) Class level relative abundance. (c) Order level relative abundance. (d) Family level relative abundance. (e) Genus level relative abundance. (f) Species level relative abundance. Data are means ± SD. (x±s, n=8).

Figure 5

Linear discriminant analysis effect size (LEfSe) analysis was conducted on intestinal flora of mice in each group. (A) LEfSe analysis highlighted the dominant biomarker taxa in the four groups (LDA score > 4). (B) Taxonomic cladogram derived from LEfSe analysis (LDA score>4, n=8 per group).

Figure 6

Metabolic profile and multivariate analysis among the four groups. (A) PCA score plot (positive ion mode). (B) PCA score plot (negative ion mode). (C) OPLS-DA score plot (N vs PH) (D) OPLS-DA score plot (N vs T). (E) OPLS-DA score plot (N vs TPH). (F) Permutation tested for OPLS-DA model (N vs PH). (G) Permutation test for OPLS-DA model (N vs T). (H) Permutation tested for OPLS-DA model (N vs TPH).

Figure 7

Screening for differential metabolites and pathways enrichment analysis associated with liver injury. (A) Venn diagram showing differential metabolites in comparisons of the N group with the PH, T, and TPH groups. (B) Analysis of metabolic pathways of biomarkers.

Figure 8

Spearman correlation heatmap of liver function markers, metabolomic profiles, and fecal microbiome dysbiosis. (A) The relationship between gut microbiota and liver enzymes (ALT/AST). (B) Correlation network of hepatic metabolic signatures, and differential gut Microbiota. Color intensity indicates the correlation coefficient (Red: positive, Blue: negative). (*P < 0.05; **P < 0.01).