Phocaeicola vulgatus alleviates diet-induced metabolic dysfunction-associated steatotic liver disease progression by downregulating histone acetylation level via 3-HPAA

ABSTRACT

Diet-induced metabolic dysfunction-associated steatotic liver disease (MASLD) is a prevalent metabolic disorder with limited effective interventions available. A novel approach to address this issue is through gut microbiota-based therapy. In our study, we utilized multi-omics analysis to identify Phocaeicola vulgatus (P. vulgatus) as a potential probiotic for the treatment of MASLD. Our findings from murine models clearly illustrate that the supplementation of P. vulgatus mitigates the development of MASLD. This beneficial effect is partly attributed to the metabolite 3-Hydroxyphenylacetic acid (3-HPAA) produced by P. vulgatus, which reduces the acetylation levels of H3K27 and downregulates the transcription of Squalene Epoxidase (SQLE), a rate-limiting enzyme in steroid biosynthesis that promotes lipid accumulation in liver cells. This study underscores the significant role of P. vulgatus in the development of MASLD and the critical importance of its metabolite 3-HPAA in regulating lipid homeostasis. These findings offer a promising avenue for early intervention therapy in the context of MASLD.

KEYWORDS:

• Intestinal microbiology
• lipid metabolism
• fatty liver
• probiotics
• molecular mechanisms

Introduction

Diet-induced metabolic dysfunction-associated steatotic liver disease (MASLD) is a chronic metabolic disease characterized by hepatic steatosis without a history of significant alcohol consumption. It is associated with hepatocellular fat deposition, inflammation, and varying degrees of fibrosis, which can progress to cirrhosis and hepatocellular carcinoma.¹ The global prevalence of MASLD is estimated to be 24%, and even higher in some regions, making it the most common liver disease worldwide.² Despite the high prevalence and increasing impact on global health, there are few approved treatments for MASLD currently.³ Lifestyle changes, including dietary modification and exercise, remain the cornerstone of MASLD management.^4,5

In recent years, an increasing number of studies have shown a close relationship between gut microbiota and the development of MASLD.^6,7 Gut microbiota interacts with liver through the gut-liver axis, involving specific metabolites such as bile acids (BAs)⁸ and short-chain fatty acids (SCFAs).⁹ Animal model-based studies have shown that probiotics, prebiotics, and synbiotics can regulate gut microbiota homeostasis and further effectively improve MASLD, including liver fat content reduction, liver inflammation, and fibrosis alleviation.^10–12 In recently published meta-analyses of randomized clinical trials (RCTs) involving probiotic or synbiotic treatments for individuals with MASLD/NASH, microbial therapies have shown a significant association with improved levels of alanine amino transferase (ALT) and aspartate transaminase (AST).^13–15

Probiotics may affect the progression of metabolic diseases in multiple ways. Among them, they can even directly affect gene transcription in liver cells through metabolites via the gut-liver axis. For example, supplementation of prebiotics and synbiotics regulates the expression of genes involved in beta-oxidation and fat synthesis, including peroxisome proliferator-activated receptor alpha (PPAR-α), sterol regulatory element-binding protein 1c (SREBP-1c) and malic enzyme (ME).¹⁶ However, due to the complexity of the gut microbiome and the interaction among different bacterial groups, there is a relative scarcity of research related to the interplay among gut microbiota, metabolites, and gene transcription in the liver.

To deeper understand the relationship between gut microbiota and MASLD, we established murine models and carried out the longitudinal 16S rRNA gene sequencing at different phases as well as liver transcriptome. Multi-omics analysis showed Phocaeicola vulgatus (P. vulgatus, recently reclassified from Bacteroides vulgatus ¹⁷) significantly correlated with the progression of MASLD, which might be a potential probiotic for MASLD. Then, the protective role of P. vulgatus for MASLD was determined through murine model. We revealed that P. vulgatus attenuated the development of MASLD through the production of 3-hydroxyphenylacetic acid (3-HPAA), which is the metabolites of P. vulgatus. This metabolite was further discovered to reduce the acetylation status of histones, consequently resulting in decreased expression of squalene epoxidase (SQLE), which effectively inhibits fat accumulation. In this study, we focused on different stages of MASLD and demonstrated that P. vulgatus and its metabolite 3-HPAA could serve as early interventions for MASLD. This finding provided a novel approach to mitigate the development of the disease.

Result

Dynamic alterations of general indicators and gut microbiota during MASLD development

We established a murine model of MASLD using a high-fat, high-carbohydrate (HFHC) diet.^18,19 Mice were sacrificed at the fourth, eighth, and sixteenth weeks respectively.²⁰ Their serum, liver tissues, intestinal tissues, and feces samples were collected to investigate the dynamic alterations of general indicators and gut microbiota (Figure 1a). The general indicators included weight gain, hematoxylin-eosin staining (H&E score), as well as the content of ALT, AST, and triglyceride (TG) in the serum. At the fourth week of modeling, the mice in the HFHC group were already heavier than those in the Chow group (Figure 1b), as well as the TG content in the serum (Figure 1c). Moreover, the HFHC group exhibited increased levels of ALT and AST, which are the indicators of liver damage, as compared to the Chow group (Figure 1d–e). These trends became more pronounced over time. To understand the conditions in liver, H&E and Oil Red O staining assays showed that steatosis had already appeared at the fourth week. As the disease progresses, liver cells showed more ballooning and were even filled with lipid droplets at the sixteenth week (Figure 1f–g). Sirius red staining reveal that liver tissue also exhibited fibrosis at this time (Supplementary Fig S1). The quantified scores of H&E staining and Oil Red O staining are shown in Supplementary Fig S2. The results indicated that both pathological and functional changes occurred during the early stages of MASLD, highlighting the importance of early intervention in managing this disease. Subsequently, we collected feces from mice at these three end points and performed 16S rRNA gene sequencing. Microbial richness was initially decreased in the fourth week but showed a slight rebound trend afterward based on ɑ-diversity indexes (Chao1, Shannon, and Simpson) (Supplementary Fig S3). Gut microbiota compositional discrimination was observed between mice in the HFHC and the Chow group through principal coordinates analysis (PCoA), utilizing both unweighted and weighted UniFrac (Bray-curtis) distances. It could be found that the gut microbiota structure of these two groups was already distinct at the fourth week, and the differences always present during modeling time (Figure 1h, Supplementary Fig S4; PERMANOVA, p < .05). At the phylum level, the top eleven most abundant gut taxa (including “others”) are listed at three time points. Among all the time points, the abundance of Firmicutes was significantly higher than that of Bacteroidota in the HFHC group, while in the Chow group, the abundance of both phyla was similar. In detail, at the fourth week, Campilobacterota and Proteobacteria were enriched in the HFHC group compared with Chow group, while Actinobacteriota was depleted (Figure 1i). At the eighth week, the tendencies of Proteobacteria, Campilobacterota, and Actinobacteriota were similar to that at the fourth week. Whereas at the sixteenth week, Proteobacteria was depleted in HFHC group compared with Chow group, which was opposite to the previous trend. This suggested that in the late stage of MASLD, the structure of gut microbiota might not fully represent the overall microbial composition during the progression of MASLD. To better show the trend of microbial changes, heatmap and cluster analysis of the 16S rRNA gene sequencing data at three endpoints were performed. Heatmap showed that one cluster of the microbiota was clearly separated from others, which had the characteristic of a significant decrease in content as the disease progressed (Figure 1j, Cluster 1). These gut microbiotas were screened out for subsequent analysis (Supplementary Table S1). Interestingly, we found that there was a sharp change in the abundance of gut microbiota at the eighth week, which suggested interventions targeting the gut microbiota may have a specific temporal window.

Figure 1. Dynamic alterations of liver pathological change and gut microbiota during MASLD development. (a) Schematic diagram of the HFHC diet-induced MASLD model. (b-e) body weight gain, serum TG levels, ALT levels, and AST levels of two groups of mice at three time points. (f, g) H&E staining images and oil red O staining images of liver sections. The arrow represents the typical ballooning hepatocytes in the 16w HFHC group. Scale bar 100 μm. (h-i) PCoA clustering plot of 16S rRNA gene sequencing data and relative abundance of microbial phyla at three time points. (j) OTU clustering heatmap based on the progression of MASLD modeling. The ‘Ward.D’ method was used for clustering, and the ‘Euclidean’ method was used for distance calculation. Cluster 1, marked by blue frame, were selected further analysis. Data are presented as mean±SEM. *p < .05, **p < .01.

Multi-omics analysis provides evidence for the involvement of phocaeicola vulgatus in the development of MASLD

In addition to 16S rRNA gene sequencing above, a series of transcriptome sequencing of liver tissues was also conducted. Through PCoA analysis, it was found that the differences in liver cell transcriptome appeared slightly later than the differences in 16s rRNA gene sequencing (Figure 2a–c). Similar to alterations in the microbiota, transcriptome clustering analysis revealed two series of genes that showed gradual changes over time as the disease progressed (Figure 2d). Gene lists in these two clusters were shown in the Supplementary Table S2 , S3. Interestingly, a sharp change was also observed from the eighth week in transcriptome data. Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis showed the decreased genes were mainly concentrated in the cGMP-PKG pathway (Figure 2e, Supplementary Fig S5). In contrast, upregulated genes were mainly concentrated in metabolic pathways (Figure 2f, Supplementary Fig S6), including heat shock protein family D (Hsp60) member 1 (HSPD1), nicotinamide phosphoribosyltransferase (NAMPT), apolipoprotein A4 (APOA4), and other genes which played crucial roles in MASLD development.^21–23 To further explore the relationship between gene transcription and gut microbiota structure, we specifically focused on the down-regulated microbiota (Cluster 1) and the up-regulated genes (Cluster B). Employing a Spearman correlation analysis and utilizing the ‘Ward.d2’ clustering method, we identified strongly correlated microbiota-gene pairs, comprising 31 bacterial taxa at the genus level and 238 genes (Figure 2g). Subsequently, through stringent criteria (r < −0.6 and correlation test p < .05), we refined our selection to 27 bacterial taxa at the genus level. Detailed information on their correlation with gene expression, along with the correlation test results, is available in Supplementary Table S4. Additionally, general indicators (including weight gain, ALT and TG content, H&E staining score) were linked to the microbiota content using a Redundancy (RDA) analysis (Figure 2h). It can be found that the blue-labeled microbiota, including Parabacteroides, Helicobacter, Akkermansia, Bacteroides, Bifidobacterium and Dubosiella, were negatively correlated with general indicators. Two genera of microbiota: Parabacteroides and Bacteroides were identified by overlapping the results of Spearman analysis and the RDA analysis (Supplementary Fig S7a), and their strongly correlated genes were marked in red in Supplementary Table S4. A quantitative reverse transcription PCR (RT-qPCR) analysis was conducted at the species level to evaluate the colonization of these two genera. Seven bacterial species belonging to these genera were analyzed, including P. vulgatus, which was historically classified under the genus Bacteroides but has recently been reclassified as a member of the Phocaeicola genus (not yet reflected in the 16S rRNA analysis annotations). In these seven bacterial species, P. vulgatus emerged as the most abundant bacterium (Supplementary Fig S7b). Interestingly, when examined the changes in the abundance of different species within Parabacteroides and Bacteroides at three time points, we observed that only P. vulgatus and B. acidifaciens showed consistently lower abundances compared to the Chow group across all three time points during HFHC modeling, beginning at the fourth week. Furthermore, the abundance of P. vulgatus in the Chow group remained relatively stable over time (Supplementary Fig S8). We speculated that P. vulgatus might confer an important protective role against MASLD in the early stage.

Figure 2. Dynamic alterations of liver transcriptome and multiomics analysis. (a-c) PCA plot of RNA sequencing data at three time points. (d) Heatmap of RNA clustering analysis based on the progression of disease. Clustering was performed using the ‘Median’ method and the ‘Manhattan’ distance calculation method. The ‘cluster B’, highligted by the red frame, was used for further analysis. (e) KEGG pathway analysis of the down-regulated genes. (f) KEGG pathway analysis of the up-regulated genes. (g) Spearman correlation heatmap of the up-regulated genes’ cluster (cluster B) and the down-regulated bacteria’s cluster (cluster 1). Clustering was performed using the ‘Ward.D2’ method and the ‘Manhattan’ distance calculation method. (h) RDA plot showing the correlation between the abundance of bacteria and general indicators. Permutation test showed that the RDA model was statistically significant (p < .05).

P. vulgatus confers a protective role against MASLD

To verify the function of P. vulgatus on MASLD, an eight-week study on antibiotic-depleted ApoE^−/− mice was conducted initially. ApoE^−/− mice are genetically engineered mice that lack a functional copy of the apolipoprotein E (ApoE) and are commonly used in research related to metabolic disorders such as MASLD.^24,25 The choice of ApoE^−/− mice in our study is further supported by their characteristic lipid profile, where the absence of functional ApoE leads to alterations in cholesterol distribution, mimicking aspects of dyslipidemia observed in MASLD patients.²⁶ This unique physiological trait makes ApoE^−/− mice a suitable model for investigating the interplay between P. vulgatus and MASLD. Mice were given HFHC diet for six weeks after two weeks of antibiotic clearance (Supplementary Fig S9). Meanwhile, the reinforced clostridial agar (RCM) group was given the RCM medium as negative control. The group that received P. vulgatus (1 × 10⁹ cfu/100 μL) was referred to as the P. vul group (Figure 3a, Supplementary Fig S10). Mice in P. vul group showed lower TG levels compared to the RCM group at the end of the modeling time (Figure 3b–c, Supplementary Fig S11). Hepatic ballooning degree were reduced in the mice fed with P. vulgatus according to H&E staining (Figure 3d, f). Oil Red O staining also showed a significant reduction in hepatic fat deposition in P. vul group (Figure 3e, g). In our previous study, we used this model to investigate the therapeutic effect of Akkermansia muciniphila on MASLD, and we found that in this model, P. vulgatus had more stable performance (Supplementary Fig S12).

Figure 3. P. vulgatus alleviated disease progression in MASLD. (a) Schematic diagram of HFHC-induced APOE^−/− mice orally gavaged with P. vulgatus. (b-e) body weight, TG levels, ALT levels, and AST levels of two groups of mice at the end of the modeling time. (f, g) H&E staining and oil red O staining of liver sections at the end of the modeling period. Scale bar 100 μm, 200 μm. (h) Schematic diagram of HFHC-induced C57BL/6 mice orally gavaged with P. vulgatus. (i-l) body weight, TG levels, ALT levels, and AST levels of four groups of mice at the end of the modeling time. (m) H&E staining of liver sections from four groups of mice at the end of the modeling period. (n) Nile red staining, oil red O staining and Sirius red staining of liver sections. Scale bar 100 μm. Data are presented as mean±SEM. *p < .05, **p < .01, ns, no significance.

For further investigation of the specific therapeutic effect of P. vulgatus on MASLD, a sixteen-week HFHC modeling study on C57BL/6 mice were conducted. Given the significant alteration of the transcriptome and microbiota at the eighth week, we divided the mice into four groups: the TP group received P. vulgatus throughout the modeling period, the F8w group received for the first eight weeks, the B8w group received for the next eight weeks, and the RCM group served as a negative control (Figure 3h). It was found that the weight gain in the TP, F8w and B8w groups was significantly lower, especially in the F8w group (Figure 3i). Similar results were also found in the serum TG (Figure 3j), ALT and AST levels (Figure 3k, l). The pathological staining including H&E and oil red O demonstrated that the F8w group had the least fat deposition and collagen fiber deposition among these four groups (Figure 3m–p and Supplementary Fig S13). It could be concluded that treatment with P. vulgatus confers a protective role against MASLD, and early treatment may improve some aspects of whole-body metabolism.

P. vulgatus relieves MASLD via three-hydroxyphenylacetic acid

It is well known that the gut microbiota generally influence the development of disease through its metabolites.²⁷ To further investigate the functional metabolite of gut microbiota, the feces from mice in Chow and HFHC group were subjected to metabolome analysis. It can be observed that in the HFHC group, the content of short chain fatty acids (SCFAs) was significantly reduced, while the content of amino acids was significantly increased (Figure 4a). The volcano plot shows ten substances with decreased levels in the HFHC group (Figure 4b, Supplementary Fig S14). Among these metabolites, 3-HPAA is the intestinal metabolite of quercetin, which has been widely reported to be inversely associated with the progression of MASLD.^28,29 Its content was found to be reduced in the HFHC group (Figure 4c). However, it was unclear whether the metabolites of P. vulgatus play a role in reducing fat accumulation. Therefore, fecal samples were collected from mice after oral gavage of P. vulgatus, then dissolved with PBS to form fecal supernatant. The fecal supernatant was applied to the human normal liver cell line MIHA and the mouse normal liver cell line AML12 in vitro (Figure 4d). The fat deposition in liver cells was reduced after treatment with the supernatant of P. vul group, despite being subjected to high-fat modeling (Figure 4e, Supplementary Fig S15-S16). In order to find out whether it was 3-HPAA in the metabolite that played a role, in vitro experiments were conducted and showed that compared to other substances such as quercetin, γ-aminobutyric acid (GABA) and pentadecanoic acid (C15:0), 3-HPAA could inhibit lipid deposition in liver cells more significantly (Figure 4f, Supplementary Fig S17). At the same time, in in vitro experiments, after adding 3-HPAA, the transcription of key genes that contributed to fatty acid metabolism in the liver cell lines MIHA and AML12 was also reduced (Supplementary Fig S18). To confirm that 3-HPAA is indeed a metabolite of P. vulgatus, we introduced quercetin into the culture medium of P. vulgatus and analyzed its metabolites with high performance liquid chromatography (HPLC) (Figure 4g). After 24 hours of cultivation, we observed a decrease in the quercetin content and a significant increase in its metabolite, 3-HPAA (Figure 4h–i). In in vivo experiments, after oral gavage of P. vulgatus, the content of 3-HPAA in feces of mice also increased significantly (Figure 4j). In conclusion, we demonstrated that P. vulgatus confers a protective role against MASLD through the metabolite 3-HPAA.

Figure 4. P. vulgatus confers a protective role against MASLD through the metabolite 3-HPAA. (a) Distribution of metabolite content of different types in feces between the high-fat diet group and the normal diet group. (b) Volcano plot of substances with significant differences between the two groups. (c) Bar graph of the content of 3-HPAA. (d) Schematic diagram of treatment with fecal supernatant on liver cell lines. (e) Oil red O and nile red staining of AML12 treated with fecal supernatant. PAOA: high-fat modeling (palmitic acid and oleic acid); RCM: the fecal supernatant of mice in the RCM group based on high-fat modelling; P. vul: the fecal supernatant of mice in the P. vul group based on high-fat modeling. Scale bar 50 μm. (f) Oil red O and nile red staining of cell lines treated with different candidate metabolites. C15:0: pentadecanoic acid. Scale bar 50 μm. (g) Schematic diagram of HPLC detection of P. vulgatus metabolites and the standard curve of 3-HPAA. HPLC peaks and statistics of quercetin(h) and 3-HPAA(i) in these two groups of bacterial culture medium. (j) HPLC peaks and statistics of 3-HPAA in the feces of two groups of mice. The red box represents the characteristic peak. Data are presented as mean±SEM. *p < .05, **p < .01.

3-HPAA alleviated the progression of MASLD

To further explore the function of 3-HPAA in alleviating the progression of MASLD, 3-HPAA was added into the culture medium of liver cell lines at various concentrations (Figure 5a, Supplementary Fig S19). In the condition of high fat modeling, addition of 3-HPAA significantly reduced the degree of hepatocytic fat deposition in a dose-dependent manner up to 200 μM (Figure 5b, c, Supplementary Fig S20), whereas this effect was not observed with the treatment of quercetin (Figure 4f). At the same time, the transcription levels of the lipid metabolism-related genes were significantly decreased (Figure 5d, e), which indicated 3-HPAA, rather than quercetin, could attenuate the progression of the fat deposition. Mice which were fed with the HFHC diet were administered 3-HPAA for sixteen weeks to conduct in vivo experiments (Figure 5f). The intake of 3-HPAA during the modeling process did not affect the food intake of mice (Supplementary Fig S21). After administering 3-HPAA to mice, we found that the 3-HPAA content in the mouse serum increased (Supplementary Fig S22). After sixteen weeks of oral administration, the body weight gain and serum TG levels in mice fed with 3-HPAA were significantly lower than those fed with PBS (Figure 5g, h). In addition, hepatic injury in 3-HPAA group was also obviously reduced (Figure 5i, j). H&E, Oil red O and Sirius red staining showed that 3-HPAA treatment significantly reduced the ballooning of liver cell, declined lipid deposition and reduced liver fibrosis (Figure 5k–m, Supplementary Fig S23). Results above confirmed that 3-HPAA inhibited the progression of MASLD and regulated the transcription of lipid metabolism-related genes.

Figure 5. 3-HPAA alleviated the progression of MASLD. (a) Schematic diagram of the effect of 3-HPAA on liver cells in vitro. Oil red O staining is shown in (b-c). Scale bar 50 μm. (d-e) transcriptional effects of different concentrations of 3-HPAA on genes related to lipid metabolism in two cell lines. (f) Schematic diagram of 3-HPAA gavage in mice. (g-j) weight gain, TG levels, ALT levels and AST levels of two groups of mice after gavage. (k) H&E staining images of two groups of mice after gavage. Oil red O staining images(l) of four groups of mice after modeling and Nile red staining(m). Scale bar 200 μm. Data are presented as mean±SEM. *p < .05, **p < .01.

3-HPAA promotes transcription of SQLE through affecting histone acetylation

Three-hydroxyphenylacetic acid is a type of phenolic acid that can easily penetrated the lipid bilayer, then directly affect the epigenetic modifications of genes, thereby affecting their transcription.^30,31 Therefore, we speculated that 3-HPAA might have an epigenetic regulatory effect on genes related to fat metabolism, and one important form of such epigenetic modification is histone acetylation.³² To validate this hypothesis, the acetylation levels of vital histones were measured in hepatic cell lines after treated with 3-HPAA in vitro. It was found that the acetylation levels of H3K27 and H3K9 were obviously decreased after the treatment of 3-HPAA, which could be rescued by the histone acetylation activator CTPB (Figure 6a, Supplementary Fig S24). Furthermore, Oil Red O staining demonstrated that CTPB could reverse the inhibition of lipid accumulation caused by 3-HPAA (Figure 6b, Supplementary Fig S25–27). In mice fed with HFHC diet, we found that oral gavage of P. vulgatus and 3-HPAA both significantly reduced H3K27 acetylation (Figure 6c). To identify the critical downstream target genes of 3-HPAA, we screened out genes which highly correlated with P. vulgatus (|r| > 0.4，p<.05, Supplementary Table S5). Analysis of KEGG implied that apart from the overall metabolic pathway, the steroid biosynthesis pathway was particularly significant. Among genes in this pathway, the transcription of the squalene epoxidase (SQLE) and NAD(P) dependent steroid dehydrogenase-like (NSDHL) both decreased with the addition of 3-HPAA, and these changes were both rescued by CTPB (Figure 6e). Previous study has demonstrated that SQLE acts as a connection between steroid synthesis and lipid metabolism, thus accelerates the progression of MASLD.³³ In light of this, our study sought to investigate the impact of 3-HPAA on SQLE expression levels. Encouragingly, results showed that treatment with 3-HPAA significantly decreased the expression of SQLE in cells, and this effect could be reversed by CTPB (Figure 6f and Supplementary Fig S28). In Oil Red O staining, knocking down SQLE led to a reduction in hepatic fat deposition, while its overexpression promoted its deposition, providing compelling evidence for the promoting effect of SQLE on MASLD (Figure 6g and Supplementary Fig S29). Furthermore, in in vivo experiment, gavage of P. vulgatus and 3-HPAA both markedly reduced the expression of SQLE in the liver (Figure 6h). The regulation of histone acetylation is achieved through the combined action of acetyltransferases (HATs) and deacetylases (HDACs).³⁴ To confirm that 3-HPAA is indeed capable of reducing HAT activity to modulate histone acetylation, we measured HAT enzyme activity. Results showed that HAT activity in cells significantly decreased after treatment with 3-HPAA and increased with CTPB (Figure 6i). To explore the direct binding between 3-HPAA and HAT, AutoDock Vina was employed to perform docking simulations and calculate the interaction between 3-HPAA and E1A binding protein P300 (EP300), which is the key protein of HAT.^35,36 There were multiple groups of residues used to form interactions between the receptor protein EP300 and ligand 3-HPAA, such as the hydrogen bond formed by TRP1466 of EP300 and 3-HPAA. With these interaction forces, the binding energy of protein-ligand complex was −5.9 kcal/mol, which shows a good performance (Figure 6j and Supplementary Table S6). Overall, these results suggested that 3-HPAA, the metabolism of P. vulgatus, affected the transcription levels of downstream genes such as SQLE by inhibiting the activity of HAT and reducing the level of H3K27 acetylation (Figure 7).

Figure 6. 3-HPAA affects the transcription of downstream genes by histone acetylase. (a) Western blot analysis of histone acetylation in two liver cell lines treated with 3-HPAA and the HAT inhibitor CTPB. (b) Oil red O staining of two liver cell lines treated with 3-HPAA and CTPB under the high-fat molding by PA and OA. Scale bar 50 μm. (c) Immunohistochemical staining of H3K27ac in liver tissue sections from four groups of mice with time-divided gavage of P. vulgatus. scale bar 200 μm. (d) The KEGG of genes strongly related with P. vulgatus. (e) The transcriptional changes of genes related to cholesterol biosynthesis pathway were investigated in AML12 cell line after treatment with 3-HPAA and CTPB. (f) Western blot of the expression of SQLE in two liver cell lines treated with 3-HPAA and the HAT inhibitor CTPB. (g) Oil red O staining on two liver cell lines treated with siRNA and overexpression of SQLE under high-fat conditions. The scale bar is 50 μm. (h) Immunofluorescence staining of SQLE in liver tissue sections from four groups of mice with time-divided gavage of P. vulgatus. Scale bar 50 μm. (i) Activity assay of HAT inhibition by 3-HPAA. (j) The protein-ligand interaction figure was generated by PyMOL. The EP300 protein is represented as a slate cartoon model, ligand is shown as a cyan stick, and their binding sites are shown as magentas stick structures. Nonpolar hydrogen atoms are omitted. The hydrogen bond, ionic interactions, and hydrophobic interactions are depicted as yellow, magentas and green dashed lines, respectively. PA: palmitic acid; OA: oleic acid. Data are presented as mean±SEM. *p < .05, **p < .01.ns, no significance.

Figure 7. The mechanism diagram of P. vulgatus alleviating MASLD. P. vulgatus converts quercetin into 3-HPAA in the intestine. 3-HPAA inhibits HAT in liver cells, resulting a decreased the acetylation level of H3K27, and further leading the inhibition of SQLE transcription, ultimately slowing down intracellular lipid deposition.

Discussion

Metabolic dysfunction-associated steatotic liver disease is a highly prevalent metabolic disorder and the incidence is likely to increase with further social development. Previous study reported the gut microbiota represent an environmental factor associated with the development of MASLD. However, how the specific microbiota is involved in the progression of MASLD has not been studied thoroughly. Here, we have identified the crucial role of the microbiota P. vulgatus in the progression of MASLD, especially in early stage. Furthermore, we have demonstrated that this bacterium reduced fat accumulation by down regulating the expression of SQLE through its metabolite 3-HPAA, which exerted its effect by inhibiting HAT enzyme activity and reducing acetylation levels of H3K27. These results revealed the interplay among gut microbiota, metabolites, and gene transcription in the liver.

In this study, a multi-omics approach was used to explore the functional microbiota. The integrational analysis of transcriptomics and 16S rRNA gene sequencing data allowed for a more comprehensive biological interpretation of the microbiota. Through correlation analysis, we identified several bacteria associated with MASLD development, including Parabacteroides, Dubosiella, Akkermansia, Phocaeicola. Among which, the Akkermansia muciniphila has received considerable attention in recent research, particularly for its potential application in MASLD.^37,38 Yet, the role of other selected bacteria remains unclear and deserve further investigation. P. vulgatus is a genus of gram-negative bacteria, which can break down the complex carbohydrates in human, such as glycans, to generate energy and other essential nutrients.³⁹ Some studies have shown that Phocaeicola can alleviate atherosclerosis⁴⁰ and colitis,⁴¹ but its underlying mechanism remains obscure. As one of the potential next-generation probiotics, P. vulgatus seems has both positive and negative effects. In most studies, it is beneficial for maintaining intestinal barrier, reducing DSS-induced colitis,^41–43 and preventing atherosclerosis.⁴⁰ However, some studies have shown that P. vulgatus may positively associated with metabolic disease, such as type 2 diabetes.^44–46 These disparities may arise from variations in study populations, methodologies, and sequencing techniques. Gut microbiota comprising numerous species with distinct interactions further complicates the picture. In addition, as found in our study, changes in microbiota are not monotonic in disease, and different sampling time points during the development of the disease may also lead to different research results. Given these discrepancies, further research is imperative. Future investigations should delve into the interaction among microbiota and consider broader gut microbiota factors. This nuanced approach promises a better understanding of the role of P. vulgatus in metabolic health and potential therapeutic applications.

As for the exploration of the specific mechanisms of P. vulgatus, we further demonstrated that it act through its metabolites 3-HPAA. There is already a large body of literature on the therapeutic effects of bacteria metabolites on MASLD, such as butyrate⁴⁷ and bile acids.⁴⁸ 3-HPAA was one of the metabolites of quercetin, which abundant in human dietary.²⁸ One study has found that the Parabacteroides was positively correlated with 3-HPAA, and that the microbial metabolite 3-HPAA can protect cardiac dysfunction by activating NRF2.⁴⁹ In addition, studies have shown that 3-HPAA can protect the liver from oxidative damage.^50–52 However, there is little literature to verify the exact relationship between bacterial strains and 3-HPAA metabolism, and there is no study on the alleviation effect of 3-HPAA in MASLD. In our study, we identified 3-HPAA as a metabolite of P. vulgatus through HPLC and found that it can prevent the progression of MASLD. Our data suggested that 3-HPAA might serve as a candidate drug for subsequent intervention in MASLD, with great application potential.

Histone acetylation is an epigenetic modification that regulates gene transcription and is closely related to the development of MASLD.⁵³ Acetylation can relax the structure of histone proteins, making DNA more accessible for transcription.⁵⁴ Previous studies have reported acetylation of histone promoted the expression of genes involved in fatty acid synthesis in hepatocytes, such as acyl-CoA synthase 5 (ACSL5)⁵⁵ and the fatty acid translocase CD36.⁵⁶ Overexpression of these genes leads to the accumulation of fatty acids in hepatocytes. The acetylation of H3K27 is one of the common types of histone acetylation.⁵⁷ Studies have shown that the gene expression influenced by the acetylation of H3K27 is closely related to the progression of MASLD.⁵⁸ Our study has uncovered a novel function of 3-HPAA in regulating histone acetylation, filling a gap in our understanding of the biological effects of 3-HPAA. As the downstream of 3-HPAA, Sqle is a key enzyme in the cholesterol biosynthesis pathway, involved in the oxidation of squalene.⁵⁹ In this study, we showed SQLE could also be regulated by histone acetylation.

Several limitations exist in the present study. Firstly, the use of antibiotics to deplete the gut microbiota may not entirely replicate a germ-free environment in the mice. Subsequent experiments with germ-free mice would provide a more accurate model. Secondly, our analysis relied on time-series heatmaps to depict changes in the gut microbiota of Chow group mice over time. While insightful, this approach may not fully capture nuanced temporal trends in microbial composition. Future research could explore more sophisticated methods to elucidate dynamic shifts in gut microbiota over time. In addition, to study the specific metabolites of the target bacterium, a relatively single bacterial specie was involved in the intervention. We also focused on a specific bacterium, P. vulgatus, to produce 3-HPAA, without considering the potential contribution of other microbial groups to 3-HPAA metabolism. It is essential to recognize that the gut microbiota is a complex ecosystem with multiple species capable of metabolizing various compounds. Regarding metabolite analysis, this study did not comprehensively encompass all gut metabolites, including those associated with quercetin, such as 3,4-dihydroxyphenylacetic acid. In terms of application of our findings, the starting point of this experiment was the modeling in mice, the experimental results have not been validated in the human. The situation in the human body is more complex, and more follow-up experiments are needed.

In summary, this study provided novel evidence that P. vulgatus confers a protective role against MASLD through the production of its metabolite 3-HPAA, which, in turn, reduces the acetylation of histones in liver cells. This reduction in histone acetylation affects the transcription of genes involved in lipid metabolism, such as SQLE. These findings highlight the relationship between gut microbiota and gene transcription in the liver.

Materials and methods

Animals and diets

All animals were kept in specific pathogen-free (SPF) conditions with a constant temperature of 20–22°C and a 12-hour light-dark cycle. Six mice were housed per cage. Autoclaved corn cob bedding was used as the substrate. Bedding was replaced and cages were cleaned twice a week.

Male C57BL/6J mice (six weeks old) were purchased from Charles River Laboratories and bred in the Laboratory Animal Center of Zhejiang University (Zhejiang, China). To establish the murine MASLD model, mice were subjected to a two-week course of antibiotics at 6-weeks old (ampicillin 1 g/L, gentamicin 1 g/L, metronidazole 1 g/L, and vancomycin 0.5 g/L)^60,61. After antibiotic cleaning, mice were fed chow diet and high-fat, high-calorie fructose (HFHC) diet which was reported to induce murine MASLD.⁶² HFHC diet (TrophicDiet, Nantong, China, Ltd, China) contains 5.56 kcal/g, the Chow diet (TrophicDiet, Nantong, China, Ltd, China) contains 4.11 kcal/g. The nutrient composition of the ingredients and their proportions in the diet are shown in Table 1. Drinking water was supplemented with fructose (55%) 23.1 g/L and sucrose (45%) 18.9 g/L. Mice were sacrificed and sampled at weeks 0, 4, 8, and 16, respectively.²⁰

Table 1. Ingredients table of HFHC diet.

Calories	Proportion	Source	Proportion
HFHC die
Fat	58%	fully hydrogenated coconut oil	93%
		soybean oil	7%
Carbohydrates	26%	Mainly from sucrose and maltodextrin
Protein	16%
Chow diet
Fat	13%	fully hydrogenated coconut oil	93%
		soybean oil	7%
carbohydrates	62%	Mainly from sucrose and maltodextrin
protein	25%

In intervention experiments, mice were fed with P. vulgatus (for varying durations) or 3-HPAA after antibiotic clearance and then fed with HFHC diet until week 16. Mice were maintained under a 12-hour light/12-hour dark cycle. Body weight and food intake of each group of mice were measured weekly. At the end of the experiment, fecal samples were collected from each group of mice for bacterial 16S rRNA gene sequencing. Mice were fasted for 12 hours, and blood was collected from the retroorbital sinus. Body and visceral weights were recorded. Liver was rapidly excised, weighed, and divided into three parts, one part was fixed in 4% paraformaldehyde for histological staining, another part was snap-frozen in liquid nitrogen for frozen sectioning, and the last part was stored at −80°C for further experiments. Mice were euthanized by cervical dislocation.

Eight-week-old male ApoE^−/− mice were purchased from the Nanjing GemPharmatech Company (Nanjing, China) and subjected to the same antibiotic cocktail described above, along with a HFHC diet for six weeks.^24,63 In intervention experiments, mice were fed with P. vulgatus for six weeks. All subsequent operations were the same as for C57BL/6J mice.

All animal protocols were approved by Zhejiang University Experimental Animal Ethics Committee and used the ARRIVE1reporting guidelines.⁶⁴

Additional methods are provided in supplementary file.

Authors’ contributions

SX Jin and P Chen were involved in study design, performed experiments and drafted the manuscript; J Yang performed original draft and review & editing; DG Li, QM Xia, XL Liu and YF Tong performed animal experiments; WH Yu performed histological evaluation; YY Zhang performed original draft and review & editing; YL Li, YX Li, GQ Chen performed review & editing; XX Fan, H Lin designed, supervised the study and revised the manuscript.

Data availability material

The datasets analyzed in our study are openly available in figshare at 10.6084/m9.figshare.23622366.

Acknowledgments

We thank Xiaoli Hong and Chao Bi from the Core Facilities, Zhejiang University School of Medicine for their technical support. We also thank Wei Liu and Jiaqing Wang from Zhejiang Academy of Agricultural Sciences for their technical support.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/19490976.2024.2309683

Additional information

Funding

This work was supported by the National Natural Science Foundation of China [81874059 and 82102105]; Natural Science Foundation of Zhejiang Province [LQ22H160017, LQ21H030012]; Medical and Health Research Project of Zhejiang Province [2019325792].