ABSTRACT

   The gut microbiota drives individual sensitivity to excess
   acetaminophen (APAP)-mediated hepatotoxicity. It has been reported that
   the bacterium Akkermansia muciniphila protects hosts against liver
   disease via the liver-gut axis, but its therapeutic potential for
   drug-induced liver injury remains unclear. In this study, we aimed to
   investigate the effect of A. muciniphila on APAP-induced liver injury
   and the underlying mechanism. Administration of A. muciniphila
   efficiently alleviated APAP-induced hepatotoxicity and reduced the
   levels of serum alanine aminotransferase (ALT) and aspartate
   transaminase (AST). A. muciniphila significantly attenuated
   APAP-induced oxidative stress and the inflammatory response, as
   evidenced by restoration of the reduced glutathione/oxidized
   glutathione (GSH/GSSG) balance, enhanced superoxide dismutase (SOD)
   activity, reduced proinflammatory cytokine production, and alleviation
   of macrophage and neutrophil infiltration. Moreover, A. muciniphila
   maintained gut barrier function, reshaped the perturbed microbial
   community and promoted short-chain fatty acid (SCFA) secretion. The
   beneficial effects of A. muciniphila were accompanied by alterations in
   hepatic gene expression at the transcriptional level and activation of
   the phosphatidylinositol 3-kinase (PI3K)/Akt signaling pathway. Our
   results suggested that A. muciniphila could be a potential pretreatment
   for APAP-induced liver injury.

   IMPORTANCE Our work revealed that A. muciniphila attenuated
   APAP-induced liver injury by alleviating oxidative stress and
   inflammation in the liver, and its hepatoprotective effect was
   accompanied by activation of the PI3K/Akt pathway and mediated by
   regulation of the composition and metabolic function of the intestinal
   microbiota. This finding suggested that the microbial community is a
   non-negligible impact on drug metabolism and probiotic administration
   could be a potential therapy for drug-induced liver injury.

   KEYWORDS: probiotic, Akkermansia, liver injury, microbiota,
   transcriptome, short-chain fatty acids

INTRODUCTION

   Acetaminophen (APAP) is a frequently used nonsteroidal antipyretic
   analgesic drug in clinical treatment. At therapeutic doses, APAP is
   mainly metabolized by phase II conjugating enzymes such as
   UDP-glucuronosyltransferases (UGTs) and sulfotransferases (SULTs) and
   converted to nontoxic soluble compounds ([54]1). However, overdose of
   APAP has potent hepatotoxicity and is the major cause of acute liver
   injury worldwide. After the nontoxic metabolic pathway reaches
   saturation, the cytotoxic compound N-acetyl-p-benzoquinone imine
   (NAPQI) is generated from the excess APAP via catalysis by cytochrome
   P450 (CYP) family enzymes and depletes glutathione (GSH) in the liver
   ([55]2). Hepatocytes cannot detoxify NAPQI when GSH is exhausted, and
   the remaining NAPQI attempts to combine with mitochondrial proteins,
   leading to oxidative stress and mitochondrial dysfunction and
   eventually to hepatocyte necrosis ([56]3, [57]4).

   Trillions of mutualistic bacteria belonging to more than 1,000 species
   colonize the human gastrointestinal tract and are defined as the gut
   microbiota ([58]5). The gut microbiota maintains a symbiotic
   relationship with its host. Changes in the pathological condition in
   the host are often accompanied by dynamic variability in the microbial
   profile, while the microbiota, in turn, participates in disease
   development. An increasing number of studies have reported the
   important association between the gut microbiota and liver disease.
   Through contributing to bile acid circulation, endotoxin translocation,
   modulation of Kupffer cells via Toll-like receptors, gut-derived T cell
   migration and other effects, microbiota dysbiosis is strongly related
   to liver diseases such as nonalcoholic fatty liver disease (NAFLD),
   autoimmune liver disease, and cirrhosis ([59]6[60]–[61]8). Modulation
   of microbial structure has become an important treatment method to
   improve the outcomes of liver diseases. Although it has been reported
   that the microbiota mediates APAP hepatotoxicity and that selected
   species protect against drug-induced liver damage, the impact of the
   microbiota on APAP-induced liver injury and its mechanism are
   insufficiently studied ([62]9). Only a few probiotics have been
   confirmed to be hepatoprotective against APAP-induced liver injury in
   animal studies.

   Akkermansia muciniphila is a strictly anaerobic Gram-negative species
   belonging to Verrucomicrobia, which constitutes of 0.5%∼5% of the human
   intestinal microflora ([63]10). A. muciniphila is considered a highly
   promising probiotic due to its representative benefits in promoting
   immunomodulatory and metabolic functions and maintaining the host’s
   homeostatic condition. A. muciniphila adheres to the mucus layer and
   consumes mucus glycan oligosaccharides as a nutrient ([64]11). The
   degradation metabolites of mucin, such as short-chain fatty acids
   (SCFAs), can regulate the expression of genes related to lipid
   metabolism and the immune response. A large number of studies have
   demonstrated that A. muciniphila is an important factor in pathological
   progression. Reduced abundance of A. muciniphila leads to a thinner
   mucous layer and impaired intestinal barrier integrity ([65]12). A.
   muciniphila abundance is negatively correlated with bowel symptoms in
   inflammatory bowel disease (IBD) patients ([66]13). A. muciniphila
   administration significantly ameliorated immune-mediated liver injury
   in a concanavalin A (ConA) model ([67]14).

   This study was designed to explore the effect of A. muciniphila on
   APAP-induced liver injury and investigate the underlying mechanisms
   through multiomic integration, incorporating 16S sequencing,
   transcriptomics, and metabolomics.

RESULTS

A. muciniphila ameliorated APAP-induced liver injury.

   A mouse model of acute liver injury model was established at 24 h after
   300 mg/kg APAP dosing. As typical biological markers of liver function,
   serum alanine aminotransferase (ALT), and aspartate transaminase (AST)
   levels were determined to evaluate the severity of liver injury. The
   ALT and AST levels were significantly elevated after intraperitoneal
   injection of APAP. Moreover, administration of A. muciniphila (AkAP)
   markedly decreased ALT and AST levels compared with those in the group
   treated with normal saline + APAP (NsAP) ([68]Fig. 1B). Hematoxylin and
   eosin (H&E) staining analysis revealed signs of extensive necrosis and
   massive inflammatory cell infiltration in the NsAP group. Only sporadic
   small necrotic foci and relatively low inflammatory cell infiltration
   observed in the group treated with A. muciniphila + APAP (AkAP)
   ([69]Fig. 1C). No obvious pathological findings were detected in the
   group treated with A. muciniphila + saline control (AkC) and that
   treated with normal saline + saline control (NsC). Consistent with the
   changes in ALT and AST levels, the modified histological activity index
   (HAI) presented a similar tendency: the HAI score was significantly
   increased in NsAP compared with NsC and was partially restored after A.
   muciniphila administration ([70]Fig. 1C). Taken together, the above
   results indicated that A. muciniphila effectively attenuated
   APAP-induced liver injury.

FIG 1.

   [71]FIG 1
   [72]Open in a new tab

   Administration of A. muciniphila ameliorated APAP-induced liver injury
   and oxidative stress. (A) Schematic diagram of the design of the animal
   experiment. (B) Serum levels of ALT and AST. (C) Representative liver
   H&E staining (left panel, original magnification × 20, scale bar =
   200 μm) and liver pathological HAI (right panel). The data are
   presented as the mean ± SEM. (D) Reduced GSH level, GSSG level, and
   GSH/GSSG ratio. (E) SOD activity. *, P < 0.05; **, P < 0.01; and ***,
   P < 0.001 for the comparison. AkAP, A. muciniphila+APAP; AkC, A.
   muciniphila+saline control; NsAP, normal saline+APAP; NsC, normal
   saline+saline control; n = 10 each group.

A. muciniphila alleviated APAP-induced oxidative stress and the inflammatory
response.

   The increased mitochondrial permeability and oxidative stress caused by
   intracellular GSH depletion are considered the initial but critical
   triggers of APAP-induced degeneration ([73]15). We evaluated GSH,
   oxidized glutathione (GSSG), and superoxide dismutase (SOD) levels to
   assess vulnerability to oxidative stress. As shown in [74]Fig. 1D and
   [75]E, 24 h after APAP treatment, the GSH, GSH/GSSG, and SOD levels
   were substantially reduced. However, AkAP exhibited evident restoration
   of those indicators compared with NsAP. In addition, the GSSG levels
   were not significantly different among the four groups.

   The inflammatory response mediated by activated Kupffer cells and
   infiltrated immune cells plays a crucial role in APAP toxicity cascade
   development ([76]1). To evaluate hepatic inflammation, we detected the
   alterations in the expression of representative inflammatory markers
   (IL-1β, IL-2, IL-6, IL-10, TNF-α, and IFN-γ) at the transcriptional
   level. Mice in NsAP showed evidently lower expression of IL-1β, IL-2,
   IL-6, and TNF-α than those in NsC, demonstrating that APAP resulted in
   an inflammatory response in the liver. However, A. muciniphila
   attenuated inflammation, as indicated by the downregulation of these
   cytokines in AkAP compared with NsAP ([77]Fig. 2C). IL-10 and IFN-γ
   showed no noticeable difference among the four groups (Fig. S1).
   Moreover, immunohistochemical analysis with anti-F4/80 or anti-Ly6G was
   performed to estimate the hepatic infiltration of macrophages and
   neutrophils. As expected, the staining results indicated that the
   recruitment of macrophages and neutrophils was markedly promoted by
   APAP, while A. muciniphila evidently reduced macrophage and neutrophil
   infiltration ([78]Fig. 2A and [79]B). Flow cytometric analysis also
   confirmed increased accumulation of macrophages and neutrophils in AkAP
   compared with NsC, while AkAP showed lower numbers of macrophages and
   neutrophils compared with NsAP ([80]Fig. 2D). Besides, we observed that
   other immune cell subsets (CD4^+ T cell, CD8^+ T cell, B cell, NK cell,
   NKT cell) had no significant variation (Fig. S2). These results
   revealed that A. muciniphila improved APAP hepatotoxicity-induced
   oxidative stress and the inflammatory response in the liver.

FIG 2.

   [81]FIG 2
   [82]Open in a new tab

   A. muciniphila improved APAP-induced oxidative stress and the
   inflammatory response in the liver. (A) Representative images of
   immunohistochemical staining of F4/80 and Ly6G (original magnification
   × 40, scale bar = 100 μm). (B) H-score of immunohistochemical staining.
   The H-score was assessed by ImageJ, demonstrating the staining
   intensity. (C) Relative hepatic mRNA expression of proinflammatory
   cytokines. (D) Representative flow cytometric plots of macrophages and
   neutrophils and quantitative analysis of cell percentage (n = 4 each
   group). The data are presented as the mean ± SEM. *, P < 0.05; **,
   P < 0.01; and ***, P < 0.001 for the comparison. AkAP, A.
   muciniphila+APAP; AkC, A. muciniphila+saline control; NsAP, normal
   saline+APAP; NsC, normal saline+saline control; n = 6 each group for
   immunohistochemical analysis; n = 10 each group for RT-qPCR; n = 4 each
   group for flow cytometry.

A. muciniphila modulated the alteration of hepatic transcriptional expression
during APAP challenge.

   To further investigate how A. muciniphila contributed to liver
   protection from APAP-induced injury, we performed transcriptomic
   analysis of livers harvested from AkAP, NsAP, and NsC (n = 4 per
   group). In the comparison between NsAP and NsC, 2810 genes were
   identified as differentially expressed genes (DEGs) by using the
   criteria P < 0.05 and |log[2] (fold change)| > 1. Among these DEGs,
   1,558 and 1,252 were significantly upregulated and downregulated,
   respectively, in NsAP compared with NsC. In the comparison between AkAP
   and NsAP, 838 and 846 DEGs were significantly upregulated and
   downregulated, respectively, in NsAP. [83]Fig. 3A shows an intuitive
   visualization of the hierarchical clustering of DEGs in the two
   comparisons. NsAP and NsC had apparently distinct gene expression
   patterns, whereas the clustering patterns of AkAP and NsC were
   relatively similar and homogeneous, indicating the improvement of the
   pathological process. We identified 777 DEGs that were promoted by APAP
   injection but suppressed by A. muciniphila administration and 711 DEGs
   that were suppressed by APAP injection but promoted by A. muciniphila
   administration through overlap analysis of all DEGs from the two
   transcriptomic comparisons, suggesting that A. muciniphila markedly
   modulated the transcriptional alteration caused by APAP. Kyoto
   Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of the
   DEGs from the comparison of NsAP versus NsC revealed 73 pathways based
   on the criterion adj-P <0.05. The affected pathways were mainly
   assigned to drug metabolism, involving CYPs, UGTs, SULTs, and the
   glutathione S-transferase (GST) family ([84]Fig. 3B). KEGG enrichment
   analysis of the DEGs from the comparison of AkAP versus NsAP identified
   50 pathways, such as steroid hormone biosynthesis, metabolism of
   xenobiotics by cytochrome P450 and cytokine-cytokine receptor
   interaction. A protein-protein interaction (PPI) network was
   constructed by mapping genes from these 50 significant pathways to
   classify genes participating in hepatoprotection. Four predominant
   clusters were extracted from the PPI network, representing drug
   metabolism, inflammation, amino acid biosynthesis, and signaling
   conduction (the top 10 hub genes of each cluster are shown in
   [85]Fig. 3C, cluster 1–3 are shown in Fig. S3, cluster 4 is shown in
   [86]Fig. 3D, and all hub genes are shown in the Table S3). Notably, the
   phosphatidylinositol 3-kinase (PI3K)/Akt pathway was identified as the
   most enriched pathway in cluster 4. These results suggested that the
   protection by A. muciniphila could be attributed to modulation of
   transcriptional expression related to metabolism, inflammation, and
   signal transduction.

FIG 3.

   [87]FIG 3
   [88]Open in a new tab

   A. muciniphila intervention affected hepatic transcriptomic alteration.
   (A) Heatmap of DEGs in AkAP versus NsAP or NsAP versus NsC. (B) KEGG
   enrichment analyses of DEGs between the comparisons (upper, NsAP versus
   NsC; lower, AkAP versus NsAP). (C) Heatmap of the top 10 hub genes from
   four clusters of the PPI network. (D) Visualization of the cluster 4
   network. AkAP, A. muciniphila+APAP; NsAP, normal saline+APAP; NsC,
   normal saline+ saline control; n = 3 each group.

A. muciniphila activated the PI3K/Akt pathway and inhibited apoptosis in the
liver.

   The transcriptomic results showed that the PI3K/Akt pathway might be
   responsible for the effects of A. muciniphila under APAP challenge. The
   activated PI3K/Akt pathway inhibits cellular apoptosis via modulation
   of Bcl-2 and Bax ([89]16). APAP treatment had no influence on Akt and
   PI3K phosphorylation, but A. muciniphila promoted the phosphorylation
   of Akt and PI3K after APAP challenge ([90]Fig. 4A and [91]C). The
   protein expression of Bax in NsAP was strikingly higher than that in
   AkAP and NsC, and Bcl-2 expression in AkAP was significantly higher
   than that in NsAP, with no noticeable alteration compared with NsC. As
   an important indicator of apoptosis, the Bcl-2/Bax ratio was decreased
   in NsAP, and A. muciniphila tended to reverse this change to the level
   observed in NsC. Activation of c-Jun N-terminal kinase (JNK) and
   downstream proteins acts as a critical mediator amplifying
   mitochondrial dysfunction and triggering apoptosis. We also found that
   A. muciniphila prevented the APAP-elicited enhancement in the
   phosphorylation levels of ERK1/2 and JNK without a prominent effect on
   the phosphorylation level of NF-κB. Consistent with these findings, the
   terminal deoxynucleotidyltransferase-mediated dUTP nick end labeling
   (TUNEL) assay demonstrated a reduced percentage of TUNEL-positive
   hepatocytes in AkAP compared with NsAP, indicating that A. muciniphila
   reduced nuclear DNA fragmentation caused by APAP hepatotoxicity
   ([92]Fig. 4B). Therefore, A. muciniphila promoted the activation of the
   PI3K/Akt pathway, corrected the Bcl-2/Bax ratio, and further suppressed
   hepatic apoptosis.

FIG 4.

   [93]FIG 4
   [94]Open in a new tab

   A. muciniphila intervention inhibited APAP-induced apoptosis and
   activated the PI3K/Akt signaling pathway. (A) Representative Western
   blot bands. (B) Representative liver DAPI/TUNEL immunofluorescence
   staining (upper panel, original magnification × 20, scale bar = 200 μm)
   and percentage of TUNEL-positive cells in each group (lower panel). (C)
   Quantitative analysis of protein expression. The data are presented as
   the mean ± SEM. *, P < 0.05; **, P < 0.01; and ***, P < 0.001 for the
   comparison. AkAP, A. muciniphila+APAP; AkC, A. muciniphila+saline
   control; NsAP, normal saline+APAP; NsC, normal saline+saline control;
   n = 6 each group.

A. muciniphila enhanced gut barrier function and inhibited LPS stimulation.

   Disruption of gut barrier integrity is an extrahepatic characteristic
   of APAP intoxication ([95]17). Our results demonstrated that serum
   lipopolysaccharide (LPS) levels were elevated in NsAP, and the
   elevation in LPS levels induced by APAP toxicity was significantly
   alleviated in AkAP ([96]Fig. 5A). LPS is recognized by TLR4 in the
   liver and triggers MyD88-dependent inflammation. The mRNA expression of
   TLR4 and MyD88 in the liver was dramatically activated by APAP and
   effectively suppressed to normal levels by A. muciniphila, indicating
   restriction of APAP-mediated LPS leakage and bacterial translocation
   ([97]Fig. 5B). Additionally, we assessed several typical gut barrier
   factors (Occludin, Claudin, MUC2, ZO-1, cannabinoid receptor 1, and
   cannabinoid receptor 2) in mRNA and protein expression levels
   ([98]Fig. 5B and [99]C and [100]D). The mRNA expression of Occludin and
   MUC2 was downregulated by 45.8% and 50.3%, respectively, in NsAP
   compared with NsC, and A. muciniphila reinforced Occludin and MUC2 mRNA
   expression to a normal level. Significant difference of Occludin
   protein expression was only observed between NsAP and AkC. MUC2 protein
   expression followed a similar pattern to mRNA expression. The
   comparison between NsAP and NsC revealed that APAP significantly
   reduced Claudin protein level. And AkAP showed a recovered Claudin
   protein level compared with NsAP. APAP did not induce a reduction in
   Claudin mRNA expression, but A. muciniphila still exerted a stimulative
   effect. There was no significant difference in cannabinoid receptor 1,
   cannabinoid receptor 2, and ZO-1 levels among these groups (data not
   shown).

FIG 5.

   [101]FIG 5
   [102]Open in a new tab

   A. muciniphila enhanced gut barrier function. (A) Serum LPS
   concentration. (B) Relative hepatic and gut mRNA expression (liver,
   TLR4 and MyD88; gut, Occludin, Claudin, and MUC2). (C) Representative
   Western blot bands and quantitative analysis of protein expression. (D)
   Representative images of MUC2 immunohistochemical staining (original
   magnification × 40, scale bar = 100 μm) and H-score. The data are
   presented as the mean ± SEM. *, P < 0.05; **, P < 0.01; and ***,
   P < 0.001 for the comparison. AkAP, A. muciniphila+APAP; AkC, A.
   muciniphila+saline control; NsAP, normal saline+APAP; NsC:=, normal
   saline+saline control; n = 6 each group for LPS detection; n = 10 each
   group for RT-qPCR; n = 6 each group for Western blot; n = 3 each group
   for immunohistochemical staining.

A. muciniphila regulated APAP-induced intestinal dysbiosis.

   To further reveal the extrahepatic influence of APAP overdose on
   intestinal homeostasis, we performed 16S rRNA sequencing analysis to
   explore the alteration in gut microbiota composition during APAP
   challenge and the influence of A. muciniphila administration on
   microflora structure. The Chao1 and Shannon index showed increased
   α-diversity after APAP treatment (Fig. S4).The fecal samples were
   generally clustered according to the similarity of the microflora and
   are illustrated by an unweighted UniFrac principal coordinates analysis
   (PCoA) plot ([103]Fig. 6A). The microbiome profiles of AkAP and NsAP
   were distinctly separated from that of NsC. The taxonomic abundance of
   the microbiome was dramatically altered. The 10 most abundant phyla and
   20 most abundant genera are presented in the Fig. S5 and [104]Fig. 6B.
   At the phylum level, Verrucomicrobioa accumulated in AkAP and AkC upon
   gavage administration of A. muciniphila. Following APAP injection, the
   relative abundances of Deferribacteres, Cyanobacteria, and
   Desulfobacterota were considerably increased. The enrichment of
   Actinobacteria was significantly reduced after APAP treatment with or
   without A. muciniphila administration. The relative abundance of
   Desulfobacterota declined in AkAP compared with NsAP (Fig. S6). At the
   genus level, high proportions of Bacteroides, Oscillibacter,
   Mucispirillum, and Colidextribacter and decreased proportions of
   Dubosiella, Lactobacillus, Bifidobacterium, Prevotellaceae_UCG-001 and
   Candidatus_Saccharimonas were observed in NsAP compared with NsC
   ([105]Fig. 6C). Lactobacillus and Candidatus_Saccharimonas were
   markedly enriched, and Oscillibacter was markedly depleted, in AkAP
   compared with NsAP. Mice in AkAP showed a markedly higher abundance of
   Lactobacillus than mice in AkC. The abundance of
   Candidatus_Saccharimonas and Oscillibacter demonstrated no significant
   difference in AkAP compared with AkC or NsC. Next, we conducted linear
   discriminant analysis (LDA) effect size (LEfSe) analysis among AkAP,
   NsAP and NsC to confirm the alteration in the microbiota structure and
   identify the signature taxonomic markers (Fig. S7). According to the
   comparison of NsAP and NsC, Oscillibacter, Mucispirillum,
   Colidextribacter, and Bacteroides were enriched, and Dubosiella,
   Bifidobacterium, Allobaculum, Prevotellaceae_UCG_001, and Lactobacillus
   were depleted, by APAP treatment. Comparison between AkAP and NsAP
   showed that Lactobacillus, Candidatus_Saccharimonas, and Akkermansia
   were enriched in AkAP, while Oscillibacter, Colidextribacter,
   Pseudaminobacter, Ruminiclostridium, and Idiomarina were enriched in
   NsAP.

FIG 6.

   [106]FIG 6
   [107]Open in a new tab

   Modulatory effect of A. muciniphila on the alterations in microbial
   taxonomic abundance after APAP treatment. (A) Unweighted UniFrac PCoA
   plot. (B) Relative abundance of the 20 most abundant taxa at the genus
   level. (C) Bar chart of the relative abundance of genera with
   significant changes (Kruskal-Wallis test). The data are presented as
   the mean ± SEM. *, P < 0.05; **, P < 0.01; and ***, P < 0.001 for the
   comparison. AkAP, A. muciniphila+APAP; AkC, A. muciniphila+saline
   control; NsAP, normal saline+APAP; NsC, normal saline+saline control;
   n = 10 each group.

The A. muciniphila-reshaped microbiota restored SCFA production.

   SCFAs, which are only produced by microbial fermentation in hosts, are
   regarded as indispensable signals in the liver-gut axis. Therefore, we
   conducted a targeted metabolomic analysis to determine the
   concentrations of specific SCFAs (acetic acid, propionic acid,
   2-methylpropionic acid, butyric acid, 2-methylbutyric acid, and valeric
   acid) in feces ([108]Fig. 7A). AkC presented consistent enhancement in
   the levels of acetic acid, butyric acid and 2-methylbutyric acid
   compared with NsC, indicating that A. muciniphila elicits SCFA
   production. The concentrations of acetic acid, propionic acid and
   butyric acid were strikingly decreased by APAP injection with or
   without A. muciniphila administration, demonstrating that APAP suddenly
   reduced SCFA production. A. muciniphila increased the levels of these
   SCFAs, although the restored levels remained lower than those in NsC
   and AkC. APAP-treated mice showed elevation in 2-methylbutyric acid and
   valeric acid elevations, but the difference was not significant. Even
   so, A. muciniphila still increased the concentrations of these SCFAs
   after APAP injection to conventional levels. No intergroup variation
   was observed for 2-methylpropionic acid.

FIG 7.

   [109]FIG 7
   [110]Open in a new tab

   A. muciniphila improved SCFA production after APAP challenge. (A)
   Levels of SCFAs in fecal samples. (B) Correlation analysis of
   differentially abundant microbial genera and SCFAs. Correlation
   analysis was conducted by the Spearman rank correlation test, and red
   (positive correlation) and blue (negative correlation) colors represent
   different correlation coefficients. The data are presented as the mean
   ± SEM. *, P < 0.05; **, P < 0.01; and ***, P < 0.001 for the
   comparison. AkAP, A. muciniphila+APAP; AkC, A. muciniphila+saline
   control; NsAP, normal saline+APAP; NsC, normal saline+saline control;
   n = 10 for each group.

Correlations of differentially abundant microbial genera and representative
biomarkers.

   Furthermore, we performed correlation analysis involving differential
   genera, SCFAs and injury indicators to determine the putative
   association between altered intestinal genera and biomarkers. The
   correlation analysis of differentially abundant microbial genera and
   SCFAs showed that Lactobacillus and Candidatus_Saccharimonas were
   conserved under Akkermansia administration, as their relative
   abundances were positively correlated with that of Akkermansia.
   Corresponding positive correlations were identified among
   Oscillibacter, Mucispirillum, and Colidextribacter, as these genera
   were regarded as unfavorable genera and were enriched in NsAP. Acetic
   acid was positively correlated with Dubosiella and Bifidobacterium and
   negatively correlated with Bacteroides, Mucispirillum, and
   Colidextribacter. Propionic acid was positively correlated with
   Candidatus_Saccharimonas and negatively correlated with Oscillibacter
   and Colidextribacter. Butyric acid and 2-methylbutyric acid were
   positively correlated with Akkermansia, Lactobacillus, and
   Candidatus_Saccharimonas ([111]Fig. 7B).

   Dysbiotic bacteria might aggravate the severity of APAP-induced liver
   injury. Oscillibacter, Mucispirillum, and Colidextribacter, which were
   abundant in NsAP, had positive associations with the liver injury
   indices (ALT, AST, HAI, and TUNEL positivity rate) and inflammatory
   factors (cytokines and inflammatory cell infiltration). In contrast,
   Lactobacillus, Candidatus_Saccharimonas, and SCFAs were generally
   negatively associated with the liver injury indices, inflammatory
   factors, serum LPS level, and JNK phosphorylation. In addition, the
   antioxidant capacity (Bcl-2/Bax ratio and SOD activity) was positively
   correlated with Lactobacillus, Candidatus_Saccharimonas,
   Bifidobacterium and SCFAs. The levels of Akt phosphorylation were
   positively correlated with Akkermansia, Lactobacillus, and Bacteroides
   ([112]Fig. 8A). Based on the integrated analysis, we revealed the
   essential interaction among the A. muciniphila-reshaped microbial
   community, microbial metabolic function and hepatoprotective effect of
   A. muciniphila.

FIG 8.

   [113]FIG 8
   [114]Open in a new tab

   The interactions among the gut microbiota, microbial metabolism, and
   APAP-induced pathological changes. (A) Correlation heatmap of
   differentially abundant microbial genera, SCFAs, and representative
   biomarkers. Red (positive correlation) and blue (negative correlation)
   colors represent different correlation coefficients. *, P < 0.05; **,
   P < 0.01; and ***, P < 0.001. (B) Schematic diagram of A. muciniphila
   intervention in APAP-induced liver injury.

DISCUSSION

   Bidirectional interaction between the gut microbiota and liver, which
   is termed the gut–liver axis, is accomplished in multiple ways, such as
   bile acid circulation, LPS leakage, and transmission of inflammatory
   mediators ([115]18). Individualized microbial composition is considered
   a potential factor underlying the variable specificity of drug
   metabolism in the liver. It has been reported that the intestinal
   microflora contributes to susceptibility to APAP-induced hepatoxicity
   ([116]9). This may be due to competitive inhibition against APAP by
   gut-derived metabolites or LPS-induced augmentation of hepatic
   inflammation, but the exact mechanism remains unclear ([117]17).
   Probiotics have been regarded as a therapeutic strategy for liver
   diseases and the concomitant gut dysbiosis. The beneficial activities
   of A. muciniphila in improving immune and metabolic disorders have been
   well recognized in liver disease, IBD, and type 2 diabetes ([118]14,
   [119]19). However, recent studies suggest that A. muciniphila
   aggravates acute graft-versus-host disease (aGVHD) and contributes to
   enhance pathogen susceptibility of Citrobacter rodentium and
   Clostridium difficile ([120]20, [121]21). Therefore, it’s necessary to
   investigate the effects of A. muciniphila in specific conditions. Our
   study revealed that A. muciniphila administration notably attenuated
   APAP-induced hepatocellular death by restoring the GSH/GSSG
   equilibrium, inhibiting mitogen-activated protein kinase (MAPK)
   activation, and alleviating oxidative stress and inflammation in the
   liver. Its beneficial effect was accompanied by activation of the
   PI3K/Akt pathway and mediated by regulation of the composition and
   metabolic function of the intestinal microbiota ([122]Fig. 8B).

   Following the undesired formation of NAPQI, mitochondrial oxidative
   stress and dysfunction are considered the primary pathogenic mechanisms
   of APAP hepatotoxicity. GSH/GSSG imbalance and diminished SOD activity
   indicate severe disruption of the antioxidant system against superoxide
   radicals ([123]2). The accumulation of reactive species that are
   generated by the electron transport chain and NADPH oxidases and not
   scavenged immediately leads to lipid peroxidation and DNA oxidative
   damage ([124]22). A number of studies have found that APAP treatment
   promotes the generation of myeloperoxidase (MPO) and malondialdehyde
   (MDA), depletes GSH and reduces SOD activity, consequently compromising
   the buffering capacity against oxidative stress events in the liver
   ([125]23, [126]24). Consistent with these studies, our results showed a
   decreased GSH/GSSG ratio and SOD activity under APAP intoxication, and
   A. muciniphila administration restored the antioxidant levels,
   suggesting that A. muciniphila corrected the redox imbalance. Because
   A. muciniphila mitigates oxidative stress in hepatocytes, the
   activation of the downstream MAPK pathway was further suppressed. This
   was supported by our Western blot analysis, which showed that A.
   muciniphila inhibited JNK and ERK1/2 phosphorylation. JNK and ERK MAPKs
   are vital intracellular signaling pathways that are responsible for
   apoptosis, inflammation, and proliferation ([127]25). Activation of
   MAPKs is a necessary condition for APAP-induced hepatocyte death due to
   its amplification of mitochondrial oxidative stress and downstream
   transduction of the MAPK signaling pathway ([128]26, [129]27). Previous
   studies have indicated that JNK inhibitors or ERK1/2 inhibitors protect
   against APAP hepatotoxicity ([130]28, [131]29).

   In addition to oxidative stress, inflammation is recognized as a
   necessary aggravated cofactor during the progression of APAP
   hepatotoxicity ([132]30). The levels of proinflammatory cytokines such
   as IL-1β and TNF-α generally increase in APAP-induced liver injury
   ([133]31, [134]32). Kupffer cells are continuously stimulated by
   gut-derived bacterial components such as LPS, inducing low-grade
   inflammation. The TLR4 receptor on the Kupffer cell membrane binds to
   LPS and triggers inflammation via MyD88-dependent signaling ([135]33).
   Increased intestinal permeability allows vast translocation of
   bacterial components into the portal vein, leading to portal
   endotoxemia and accelerating liver disease progression ([136]15,
   [137]34). As an agent causing extrahepatic damage, overdose of APAP
   destroys the gut barrier and enhances bacterial translocation,
   exacerbating hepatic inflammation ([138]17, [139]35). Inhibition of
   LPS-binding protein (LBP) effectively reduces intrahepatic cytokine
   production and susceptibility to APAP-induced liver injury ([140]36).
   In this study, A. muciniphila suppressed hepatic inflammation, as
   evidenced by the reduced expression of IL-1β and TNF-α and reduced
   infiltration of macrophages and neutrophils. Furthermore, we found that
   A. muciniphila reduced serum LPS levels, downregulated hepatic TLR4 and
   MyD88 expression, and upregulated intestinal Occludin and MUC2
   expression as well as SCFA production, indicating that its
   anti-inflammatory property was closely associated with reinforcement of
   APAP-impaired gut barrier function.

   Hepatocyte death caused by APAP hepatotoxicity is a complex process
   mediated by the coordination of multiple intracellular signal
   transduction pathways ([141]1, [142]27). A variety of apoptosis-related
   pathways, such as the AMPK-dependent pathway and SIRT1 pathway, are
   known to participate in this process. Our transcriptomic results showed
   that modulation of the PI3K/Akt pathway is a considerable
   hepatoprotective mechanism of A. muciniphila. Moreover, we examined Akt
   phosphorylation and found that it was improved after A. muciniphila
   administration. The PI3K/Akt pathway plays an important role in liver
   injury and regeneration. Recent studies have revealed that the PI3K/Akt
   pathway alleviates liver injury by phosphorylating target proteins such
   as Bcl-2, NF-κB and Nrf2 ([143]37). Bcl-2 family proteins are the
   classic downstream regulators of activated Akt and further regulate DNA
   repair and apoptosis ([144]38). The Bcl-2 family of proteins contains
   both pro- and antiapoptotic members and governs apoptotic mitochondrial
   events. The proportion of proapoptotic proteins and antiapoptotic
   proteins determines cell persistence to the apoptotic response
   ([145]39). Activated Akt can enhance the antiapoptotic protein Bcl-2
   and inhibit the proapoptotic protein Bax, thus alleviating apoptosis
   and promoting cell survival ([146]16, [147]40). Consistent with these
   findings, our experiments showed that A. muciniphila increased the
   Bcl-2/Bax ratio, accompanied by reduced TUNEL positivity, suggesting
   that A. muciniphila alleviated apoptosis by regulating Bcl-2 family
   proteins via the PI3K/Akt pathway ([148]41).

   Recent studies have increasingly focused on the interaction mechanism
   of the liver-gut axis and the application of probiotics in liver
   disease ([149]42, [150]43). Intestinal dysbiosis has been demonstrated
   as a latent factor involved in APAP-induced liver injury, as germfree
   animals present a lower susceptibility to APAP hepatotoxicity
   ([151]44). Modulation of other probiotic strains, such as Lactobacillus
   acidophilus LA14 and Lactobacillus rhamnosus GG, protects against
   APAP-induced liver injury ([152]44, [153]45). In our study, we found
   that opportunistic pathogens, such as Oscillibacter, Colidextribacter,
   and Mucispirillum, were highly enriched in NsAP, and the abundances of
   these genera were positively correlated with the liver injury indices,
   barrier disruption, and hepatic inflammation. Oscillibacter shows a
   close association with depression and obesity. Previous studies found
   enrichment of the genus Oscillibacter in patients with chronic
   metabolic disorders such as obesity and NAFLD ([154]46, [155]47). APAP
   increased the level of Mucispirillum, which is linked to IBD and
   regarded as an aggravated factor associated with intestinal
   inflammation ([156]48, [157]49). Potential probiotic genera, such as
   Dubosiella, Lactobacillus, Bifidobacterium, and
   Candidatus_Saccharimonas were depleted by APAP treatment, whereas the
   abundances of Lactobacillus and Candidatus_Saccharimonas were
   observably reversed by A. muciniphila administration ([158]50,
   [159]51). Many studies have demonstrated that Lactobacillus can
   stimulate Treg cell proliferation, promote SCFA production and prevent
   epithelial barrier dysfunction, exerting immunomodulatory effects
   ([160]52, [161]53). Importantly, the abundance of Lactobacillus is
   decreased in a variety of liver diseases, and supplementation with
   Lactobacillus shows a protective effect against liver damage ([162]44,
   [163]45, [164]54, [165]55). Previous studies have reported the
   protective effect of Lactobacillus administration against APAP-induced
   liver injury via modulation of BclII/Bax, activation of the Nrf-2
   antioxidant pathway and induction of autophagy ([166]44). In our study,
   we found that the abundance of Lactobacillus was positively correlated
   with Akt phosphorylation, identifying that Lactobacillus played a
   pivotal role in the PI3K/Akt-mediated antioxidant response. Activation
   of the PI3K/Akt pathway by Lactobacillus strains has been realized in
   other studies, although the underlying mechanism still needs further
   investigation.

   As the APAP-treated mice showed the disruption of microbial
   homeostasis, we hypothesized that microbiota-derived metabolites might
   be key mediators of the response to APAP-induced pathological
   conditions. SCFAs are metabolic products generated by the cecal and
   colonic microbiota by fermenting nondigestible dietary fibers
   ([167]56). The results of targeted metabolomics revealed that fecal
   SCFA concentrations were considerably reduced by the impact of APAP due
   to the depletion of SCFA-producing commensals, such as Lactobacillus
   and Bifidobacterium. A. muciniphila is an SCFA-producing bacterial
   strain itself ([168]57). In addition, the A. muciniphila-reshaped
   microbial community was dominated by beneficial bacteria in favor of
   SCFA synthesis. In this context, A. muciniphila restored acetate,
   propionate and butyrate to appropriate levels, as expected. Propionate
   and butyrate were negatively correlated with gut barrier dysfunction,
   suggesting that SCFAs reinforced the mucosal barrier under the
   extrahepatic impact of APAP. Butyrate is absorbed by colonic epithelial
   cells as an energy source, thereby supporting epithelial cell
   proliferation and maintaining gut barrier integrity ([169]58, [170]59).
   Increased butyrate levels suppress intestinal inflammation and
   chemotaxis by stimulating the differentiation of Treg cells via GPR
   signaling activation or histone deacetylase (HDAC) inhibition
   ([171]60[172]–[173]62). SCFAs can also be transported into the liver by
   the portal vein and play an important role in hepatic pathological
   progression ([174]63, [175]64). Butyrate-producing strains alleviate
   steatohepatitis and autoimmune hepatitis through enterohepatic
   immunologic modulation of T cell differentiation ([176]65). In
   addition, as an HDAC inhibitor, SCFAs probably suppress oxidative
   stress during APAP-induced mitochondrial dysfunction ([177]66).

   Taken together, the results of our study demonstrated that A.
   muciniphila protects against APAP-induced hepatotoxicity by
   ameliorating oxidative stress and inflammation. With latent risk, the
   disrupted gut barrier is partly recovered. The hepatoprotective effect
   of A. muciniphila was closely correlated with the reshaped gut
   microbiota and reinforced SCFA secretion.

MATERIALS AND METHODS

Preparation of A. muciniphila strains.

   A. muciniphila Muc^T was purchased from ATCC (ATCC Number: BAA-835) and
   cultured in 0.5% (wt/vol) mucin-supplemented brain heart infusion (BHI)
   (Difco, MI, USA) medium anaerobically at 37°C. After 48 h of
   incubation, the cultures were centrifuged at 7,000 g for 15 min, washed
   twice in sterile saline, and resuspended at a bacterial concentration
   of 1.5 × 10^10 CFU/mL by measuring absorbance at 630 nm for further
   intragastric administration.

Animals and treatment.

   Specific pathogen-free (SPF) C57BL/6 mice (male, 6 weeks) were
   purchased from Shanghai SLAC Laboratory Animal Co., Ltd., and
   maintained in an SPF environment. After 2 weeks of adaptive feeding,
   the mice were randomly assigned into four groups (n = 10 each group)
   ([178]Fig. 1A). The mice in the groups AkAP (A. muciniphila+APAP) and
   AkC (A. muciniphila+saline control) were orally gavaged with 3 × 10^9
   CFU A. muciniphila suspended in 0.2 mL of sterile saline per day, while
   those in the groups NsAP (normal saline+APAP) and NsC (normal
   saline+saline control) were treated with an equivalent dose of sterile
   saline. The treatments were administered for 14 days. After 14 days of
   gavage treatment, a model of APAP-induced liver injury was established
   at day 15 as previously described ([179]9). Briefly, following 12 h of
   fasting, mice in the groups AkAP and NsAP were injected
   intraperitoneally with APAP (300 mg/kg in normal saline), while those
   in AkC and NsC were administered an equivalent volume of normal saline
   as a control. Fresh feces were collected 24 h after intraperitoneal
   injection, followed by sacrifice. Blood, liver tissues, colon tissues,
   and colon contents were harvested for further experiments. All mice
   were housed at ambient temperature under humidity-controlled conditions
   with an automatic 12 h light-dark cycle and provided with chow diet and
   water ad libitum. All experiments were approved by the Animal Care and
   Use Committee of the First Affiliated Hospital, School of Medicine,
   Zhejiang University.

Serum and liver biochemical parameter analysis.

   Serum ALT and AST levels were evaluated by a dry chemistry analyzer
   (FUJI DRI‐CHEM 7000V, FUJIFILM, Tokyo, Japan).

   Liver GSH and GSSG levels were measured by a green fluorescence assay
   kit (ab205811, Abcam, Cambridge, UK), Liver SOD activity was measured
   by a colorimetric assay kit (Sigma–Aldrich, MO, USA), and the serum LPS
   level was measured by a limulus amebocyte lysate (LAL) chromogenic
   endpoint assay kit (Hycult Biotech, Uden, the Netherlands) according to
   the manufacturer’s instructions.

Histology and immunohistochemistry.

   The left central liver lobe and proximal colon was freshly isolated,
   fixed with 10% formalin for 24 h, embedded in paraffin. Liver sections
   were stained with H&E for pathological assessment. Three independent
   visual fields were randomly picked at ×20 magnification in each
   specimen. Liver necrosis and inflammation were evaluated according to
   the modified Knodell HAI by a pathologist who was blinded to this study
   ([180]67).

   Immunohistochemical analysis was performed using standard procedures
   ([181]68). Paraffin-embedded sections were deparaffinized and
   rehydrated. Subsequently, liver sections were incubated with anti-F4/80
   (Cell Signaling Technology, MA, USA) or anti-Ly6G (Abcam) antibodies to
   estimate the infiltration of macrophages and neutrophils, and colon
   sections were incubated with anti-Mucin 2 (MUC2, Abcam) or anti-zonula
   occludens-1 (ZO-1, Proteintech) to estimate gut barrier proteins. Five
   independent visual fields were randomly picked at ×40 magnification in
   each specimen. The relative quantification of staining intensity was
   performed by calculating the H-score by using ImageJ software. The
   formula for the H-score was as follows: H-score = low positive staining
   percentage × 1 + positive staining percentage × 2 + high positive
   staining percentage × 3 ([182]69).

Flow cytometry.

   Fresh livers were isolated and dissociated using gentleMACS Dissociator
   (Miltenyi Biotec, Bergisch Gladbach, Germany) and Mouse Liver
   Dissociation Kit (Miltenyi) according to the manufacturer’s protocols.
   The cell suspension was filtered through a 70 μm cell strainer,
   followed by Percoll density gradient centrifugation to enrich
   leukocytes. The antibodies used in cytometry are listed in Table S2.
   Flow cytometry analysis was performed on CytoFLEX LX (Beckman Coulter,
   CA, USA).

TUNEL assay.

   Apoptosis of hepatocytes was evaluated by employing the TUNEL Apoptosis
   Detection Kit (Vazyme, Nanjing, China) according to the manufacturer’s
   instructions. Briefly, paraffin-embedded liver sections were
   deparaffinized, rehydrated, and blocked, followed by incubation with
   the TUNEL labeling solution for 2 h at room temperature. Thereafter,
   the sections were immediately observed and imaged using a fluorescence
   microscope. Five independent visual fields were randomly picked at ×20
   magnification in each specimen. TUNEL-positive cells were counted by
   ImageJ software.

16S rRNA gene sequencing.

   The feces collected before sacrifice were used for 16S rRNA sequencing.
   Fecal genomic DNA was extracted with the Qiagen Power PowerSoil Kit
   (Qiagen, Hilden, Germany) according to the manufacturer’s protocols.
   The V3-V4 variable region of 16S rRNA were amplified by universal
   primers (Table S1). The amplicon was used to construct sequencing
   libraries after purification, followed by sequencing performed on the
   Illumina NovaSeq platform. Paired‐end reads were generated and filtered
   through Trimmomatic software. Qualified reads were merged, denoised,
   and clustered to generate operational taxonomic units (OTUs) by using
   QIIME software and Vsearch software. Subsequently, the Silva database
   was employed to annotate OTUs. α-Diversity was determined with the
   Shannon and Chao1 diversity indices. PCoA and LEfSe analysis were
   conducted to visualize β-diversity by using R software.

Targeted fecal metabolomics.

   Targeted metabolomics was carried out using gas chromatography-mass
   spectrometry (GC-MS) to quantify the levels of representative SCFAs.
   Feces were suspended in ultrapure water, homogenized, centrifuged, and
   filtered. The filtrate was blended with an equal volume of ethyl
   acetate and incubated at 4°C for 30 min, followed by centrifugation to
   separate and collect the ethyl acetate layer. Different concentrations
   of SCFA standards were treated with the same procedures. SCFAs in the
   samples were characterized through the comparison of their retention
   times with those of standards. Calibration curves were established
   based on the peak areas of the standards at different concentrations to
   quantify SCFA levels.

RNA extraction.

   Total RNA from stored liver segments was extracted using the Qiagen
   RNeasy Plus mini kit (Qiagen) according to the instructions provided.
   The quality and quantity of RNA were measured by a NanoDrop 2000
   spectrophotometer (Thermo Fisher, MA, USA), and the RNA integrity
   number (RIN) was determined by an Agilent 2100 Bioanalyzer (Agilent
   Technologies, CA, USA). Only RNAs satisfying the criteria (RNA
   concentration ≥ 2,500 ng/μL and RIN ≥ 7) were subjected to RNA
   sequencing or quantitative real-time PCR (RT-qPCR).

Transcriptome sequencing.

   Four micrograms of the total RNA collected above were purified with
   Agencourt AMPure XP (Beckman Coulter, CA, USA) magnetic beads to enrich
   the mRNA. Then, the mRNA was randomly fragmented into small pieces and
   subjected to template construction of the mRNA library with the TruSeq
   Stranded mRNA LT Sample Prep Kit (Illumina, CA, USA). Then, 150-bp
   paired-end reads were generated through sequencing conducted on the
   Illumina HiSeq 2500 platform (Illumina). After filtration using
   Trimmomatic, trimmed reads were mapped to the reference genome using
   Hisat2. DEGs were identified by the DESeq2 package in R, and these DEGs
   were aligned against the KEGG database for pathway enrichment analysis.

RT-qPCR.

   To validate the transcriptomic results and detect target gene
   expression precisely, RT-qPCR was applied to quantify mRNA levels.
   Briefly, 500 ng of RNA was subjected to reverse transcription using
   PrimeScript RT Master Mix (TaKaRa Biomedicals, Kusatsu, Japan), and
   subsequently, the synthesized cDNA was used for RT-qPCR with TB Green
   Premix Ex Taq II (TaKaRa) and paired primers on a ViiA real-time PCR
   system (Applied Biosystems, CA, USA). The reaction of each sample was
   performed in duplicate.

   The relative expression of target genes was standardized to the GAPDH
   mRNA levels, and fold changes between groups were calculated based on
   the 2^-△△Ct method. All paired primer sequences are listed in Table S1.

Western blotting.

   Liver samples (30 mg) were added to 1 mL of radioimmunoprecipitation
   (RIPA) lysis buffer containing 1% protease inhibitor and 1% phosphatase
   inhibitor, fully homogenized for 30 min and centrifuged at 10,000 × g
   for 15 min. The supernatants were diluted to the same protein
   concentration and mixed with NuPage loading buffer (Invitrogen, CA,
   USA) prior to being denatured by boiling for 5 min. Proteins were
   electrophoresed, transferred onto polyvinylidene fluoride (PVDF)
   membranes, and incubated with primary antibodies and secondary
   antibodies successively. Bands were visualized with ECL detection
   reagents (Beyotime, Beijing, China) and captured. Protein expression
   was quantified in an optical densitometric manner using ImageJ
   software. All antibodies used are listed in Table S2.

Statistical analysis.

   Most data were analyzed by one-way analysis of variance (ANOVA) with
   Fisher’s LSD post hoc test conducted using SPSS 20.0 (SPSS, Inc., IL,
   USA) and are presented as the means ± standard deviations (SDs). A P
   value < 0.05 indicated statistical significance. Graphs were plotted
   using GraphPad Prism 7.0 (GraphPad Software Inc., CA, USA) and R
   software.

Data availability.

   The 16S rRNA gene sequencing data are available at NCBI’s Sequence Read
   Archive (SRA) database under the BioProject accession code PRJNA750743.
   The raw data of RNA sequencing are available at SRA database under the
   BioProject accession code PRJNA751139.

ACKNOWLEDGMENTS