The SNS decoction was prepared as described in our previous report (Jiang et al., 2023). SNS consists of Bupleurum falcatum L. [Apiaceae], Citrus × aurantium f. aurantium [Rutaceae], Paeonia lactiflora Pall. [Paeoniaceae], and Glycyrrhiza uralensis Fisch. [Leguminosae, Glycyrrhiza uralensis Fisch.ex DC] in a ratio of 1:1:1:1. SNS was decocted with 10 times the volume of distilled water for 1.5 h. The residue was then added to eight times the volume of distilled water for a second decoction. Finally, the residue was added to six times the volume of distilled water and decocted for 0.5 h. The solution of three decoctions were combined and filtered through a sterile gauze, the residue was removed, and the liquid was subsequently concentrated to 0.31 g/mL. SNS decoctions was freshly stored in a −20°C freezer for up to 2 days before use. The content determination of the SNS decoction has been published in our previous study (Jiang et al., 2023).

Olive oil (Cat. S30503) was purchased from Yuanye Bio-Technology Co., Ltd. (Shanghai, China). Carbon tetrachloride (CCl₄) (Cat. C805329) was purchased from Macklin (Shanghai China). Vancomycin (Cat. V10549), neomycin sulfate (Cat. N109017), ampicillin (Cat. A105483), and metronidazole (Cat. M109874) were purchased from Shanghai Aladdin Biochemical Technology Co., Ltd. (Shanghai, China). Aspartate aminotransferase (AST, Cat. C010-2-1), alanine aminotransferase (ALT, Cat. C009-2-1), and alkaline phosphatase (ALP, Cat. A059-2-2) kits were purchased from Nanjing Jiancheng Bioengineering Institute (Nanjing China). Recombinant human TNF-α (Cat. 031825) was purchased from PeproTech (Shanghai, China). Actinomycin D (Cat. HY-17559) was purchased from MedChemExpress (Shanghai, China). Cell Counting kit-8 (Cat.RK001099) was purchased from Beijing Medical Technology Co., Ltd. (Anhui, China). Anti-caspase-3 (Cat. 19677-1-AP), anti-FFAR2 (Cat. 19952-1-AP), and anti-FFAR3 (Cat. 66811-1-Ig) antibodies were purchased from Proteintech (Wuhan, China). TRIzol reagent (Cat. 15596026) was obtained from Invitrogen (Carlsbad, CA, United States). SYBR Green Premix (Cat. AG11701) and Evo M-MLV RT Premix (Cat. AG11706) were purchased from Accurate Biology (Hunan, China). The annexin V-FITC/PI double staining apoptosis detection kit (BB-4101) was purchased from Bibo Biotech (Shanghai, China). Sodium butyrate (Cat. IS0190) was purchased from Solarbio (Beijing, China). Sodium propionate (Cat. Abs42086166) and isovaleric acid (Cat. Abs47050746) were purchased from Shanghai Universal Biotech Co., Ltd. (Shanghai, China).

Male C57BL/6J mice, 6–8 weeks old and weighing between 16 and 22 g, were purchased from Hangzhou Zhiyuan Laboratory Animal Technology Co. (SCXK (Zhe) 2019-0004) and maintained in the Anhui Academy of Medical Sciences. The mice were bred in an environmentally controlled room (temperature, 22–24°C; relative humidity, 50%–60%) under a 12-h light/dark cycle with unlimited access to food and water. After adapting for 1 week, the age and body weight-matched mice were divided into three groups (n = 6): the control group (NC group), CCl₄ group, and CCl₄ group administered SNS (CCl₄+SNS group). The mice were injected intraperitoneally with CCl₄ (0.5 mL/kg, dissolved in olive oil) three times a week for 6 weeks to establish a liver fibrosis model, and NC group mice were injected with equal amounts of lysates. After 2 weeks of CCl₄ injection, the SNS group was administered with 6.2 g/kg of SNS (20 mL/kg) once daily, based on the optimal dose determined in our previous experiments. After 6 weeks, mice were anesthetized using intraperitoneal injections of tribromoethanol (300 mg/kg) and executed by cervical dislocation after orbital blood extraction (once; 0.2 mL) when they were unresponsive to mild stimuli.

Male C57BL/6J mice, 6–8 weeks old and weighing between 16 and 22 g, were used to construct pseudo-sterile mice and housed under the same conditions as described above. After a 1-week adaptive feeding, the age- and body weight-matched mice were divided into six groups (n = 6): NC group, CCl₄ group, CCl₄ + SNS group, pseudo-sterile group (PG), PG + CCl₄ group, and PG + CCl₄ + SNS group. Pseudo-sterile mice were prepared using a previously described method: antibiotics (0.5 g/L vancomycin, 1 g/L neomycin sulfate, 1 g/L ampicillin, and 1 g/L metronidazole) were added to the drinking water, which was changed daily for 6 weeks (Chambers et al., 2022). The dosing method for CCl₄ and SNS was the same as described above. All experiments were conducted under the approval of the Ethics Committee of Anhui Medical University (LLSC20210242).

Mice serum was centrifuged at room temperature at 3,000 rpm for 15 min, and the supernatant was collected and stored at −80°C. Levels of ALT, AST, and ALP were measured using automated chemistry kits.

Liver tissue was fixed with 10% formalin solution and embedded in paraffin, cut at 5 μm, and triple-stained with hematoxylin–eosin (H&E) and Masson according to standard procedures. Three independent fields of view were randomly selected from the specimen at ×20 magnification. Images were obtained through Pannoramic MIDI (3DHISTECH, Hungary).

After dewaxing, liver tissue sections were incubated sequentially using 20 μg/mL of DNase-free protease and a TUNEL assay solution. The reaction was subsequently terminated by a labeled reaction stop solution, and streptavidin–HRP working solution was added. After washing, the DAB chromogenic solution was added, and hematoxylin staining was performed. Finally, sections were graded using a 1% ethanol solution of hydrochloric acid and stained blue using a lithium carbonate solution. Images were obtained using Pannoramic MIDI (3DHISTECH, Hungary).

Immunofluorescence staining was used to determine the expression of α-smooth muscle actin (α-SMA), FFAR2, and FFAR3. Briefly, after dewaxing and antigen retrieval, tissue sections were incubated with α-SMA, FFAR2, and FFAR3 primary antibodies overnight and then stained with appropriate secondary antibodies for 2 h. Finally, sections were incubated with 4′,6-diamidino-2-phenylindole (DAPI) for 10 min. Images were obtained by Pannoramic MIDI (3DHISTECH, Hungary).

The mouse feces were collected the day before euthanasia and put into a new sterile tube. The fecal samples were submitted to Oyi Biotech for 16s intestinal flora sequencing, fecal DNA was extracted using the FastDNA SPIN Kit (MP Biomedicals, Santa Ana, CA, United States), and the V16 region of the 4S rRNA gene was amplified using dual index primers. Reads were processed and quality-filtered using Quantitative Insights into Microbial Ecology (QIIME) (v1.9.1) software, and chimera-free sequences were compared using the SILVA database (http://www.arb-silva.de) at 97% identity sex threshold. The amplicons were then sequenced using Illumina MiSeq and MiSeq Reagent Kit V3 (Illumina, San Diego, CA, United States). Paired-end sequencing was performed using the Illumina MiSeq platform. Each deduplicated feature sequence (ASV) was generated by the DADA2 method after denoising, splicing, and other quality control operations; α-diversity is determined using diversity indices such as Shannon and Simpson; PCA, LEfSe, LDA, heatmap, etc. are performed using R software analysis to visualize β-diversity.

The serum was submitted to Luming Biotechnology Company, and SCFAs were quantified via liquid chromatography–tandem mass spectrometry (LC–MS/MS). Briefly, 80 μL of the sample was taken, 80 μL of 50% acetonitrile–water solution (v/v) was added; then ultrasonic extraction in an ice–water bath was performed for 10 min, and samples were centrifuged. The clear supernatant liquid is transferred to the injection vial, derivatized, separated by UPLC, and detected via MS/MS.

Proteins were extracted from liver tissues using RIPA buffer containing protease inhibitors and phenylmethylsulfonyl fluoride (PMSF). Proteins were separated using sodium dodecyl sulfate polyacrylamide (SDS-PAGE) gels and transferred to polyvinylidene fluoride (PVDF) membranes. Proteins were blocked using milk for 2 h at room temperature and incubated with α-SMA, caspase-3, FFAR2, FFAR3, collagen-I (Col-I), apoptosis regulator Bax (Bax), and apoptosis regulator Bcl-2 (Bcl-2) primary antibodies at a ratio of 1:1000 at 4°C overnight. After incubation with secondary antibodies for 1 h, luminescence imaging was performed.

Total RNA was first extracted from liver tissue using TRIzol reagent, and reverse transcription was performed using the PrimeScript™ RT kit. Real-time fluorescence quantitative PCR was performed using the SYBR green real-time fluorescence quantitative PCR kit. β-actin was used as a normalization control. Relative expression levels were assessed using the 2^−ΔΔCT method. The primer sequences used in this study are included in Supplementary Table S1.

The human liver cancer cell HepG2 purchased from Procell Life Sciences (Wuhan, China) was maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum and 1% penicillin–streptomycin in an incubator at 37°C and 5% CO₂. Cells were incubated with actinomycin D (0.2 μM) for 30 min, after which the supernatant was discarded. The fresh medium containing TNF-α (20 ng/mL) was then added to induce apoptosis over a 24-h period. To assess the effects of propionic acid, butyric acid, and isovaleric acid, these compounds were added simultaneously with TNF-α.

The data are presented as the mean ± standard error of the mean. Group differences were analyzed using one-way analysis of variance (ANOVA), followed by Tukey’s test for normally distributed data, and performed using GraphPad Prism 8.0 software. For non-normally distributed data, the Kruskal–Wallis test was conducted using SPSS 17. The chi-square test was employed to compare categorical data between groups. Each experiment was repeated three times, and P < 0.05 was considered statistically significant. Principal component analysis (PCA) and heatmap analyses were performed using the MetaboAnalyst 5.0 online platform (www.metaboanalyst.ca).

In this experiment, we first examined the effect of SNS on the liver function of CCl₄-treated mice, and the results showed that SNS significantly reduced the CCl₄-induced elevation of ALT, AST, and ALP (Supplementary Figure S1A). H&E staining demonstrated that SNS ameliorated hepatocyte cell death and the infiltration of inflammatory cells in CCl₄-treated mice. Masson’s staining showed that SNS reduced pseudo-leaflet and fibrotic septum formation (Supplementary Figure S1B), while this anti-fibrosis effect of SNS was further confirmed by the expression of α-SMA and collagen-1 (Supplementary Figures S2A–C). These results demonstrate that SNS protects against CCl₄-induced fibrosis.

Apoptosis in the liver tissue was further measured using the TUNEL assay. The results showed that a large amount of positive cells were observed in the CCl₄-induced model mice, which was significantly reduced by SNS treatment (Supplementary Figure S3A), suggesting that SNS could suppress CCl₄-induced hepatocyte apoptosis. Western blotting was used to measure the protein expression of the apoptosis-related markers Bcl-2, Bax, and caspase-3. In CCl₄-treated mice, SNS reduced the expression of caspase-3 and Bax/Bcl-2 (Supplementary Figure S3B), which was consistent with the results of TUNEL assay. These data collectively support the protective effects of SNS against CCl4-induced liver fibrosis and hepatocyte apoptosis.

To determine whether SNS treatment altered the microbiome, we performed a 16S rRNA high-throughput gene sequencing analysis of fecal bacterial DNA isolated from the mice in the NC, CCl₄, and CCl₄ + SNS groups. The alpha diversity of intestinal microbiota was first analyzed using a generalized linear model (Figure 1A), and SNS was found to improve the abnormal increase in alpha diversity in CCl₄-treated mice. We then performed beta diversity analysis to generate a principal component analysis plot (PCA plot) (Figure 1B). The PCA plot showed clear cluster separation among the operational taxonomic units (OTUs), revealing different community structures among the three groups. These data suggest that the community structure of the intestinal flora is similar within the same group of mice but varies considerably among different groups. To further examine the potential differences in the composition of intestinal flora among different groups, a TOP15 phylum level species stacking diagram was created (Figure 1C). Phylum-level bacterial community composition was analyzed. All samples exhibited similar taxonomic communities. CCl₄-induced liver fibrosis mice showed an increased abundance of Bacteroidota and a decreased abundance of Firmicutes. Treatment with SNS, however, mitigated these alterations. To further determine which bacteria were affected by the SNS treatments, high-dimensional class comparisons were performed using effect size linear discriminant analysis to evaluate the differences in bacterial community dominance among the three groups (Figures 1D, E). The results indicated an imbalance in major bacterial types in the CCl₄ group, including Bacteroides, Parasutterella, Muribaculum, and Bifidobacterium. In contrast, these bacteria were not similarly enriched in the NC and SNS groups. Thus, SNS attenuated the abnormal enrichment of these bacterial types observed in the CCl₄-induced liver fibrosis mice. In addition, we organized the top 15 abundances at the genus level in the intestinal microbiota of the three groups (Figure 1F). Similarly, SNS treatment decreased the abundance of genera Bacteroides and Muribaculum in CCl₄-treated mice, which was consistent with the linear discriminant analysis (LDA) effect size (LEfSe) (Figures 1D, E). Overall, SNS treatment significantly changed the diversity and composition of the intestinal microbiota and restored CCl₄-induced intestinal microbiota disturbances.

Based on the linear discriminant analysis (LDA) effect size (LEfSe), notable differences were observed in the genus-level bacterial flora among the three groups. Therefore, a heatmap of the genus-level composition of the intestinal flora was generated for the three groups, while the results consistently demonstrated that SNS treatment significantly decreased the levels of Bacteroides, Parasutterella, Muribaculum, and Bifidobacterium in CCl₄-treated mice (Figure 2A). The SCFA content was next quantitated using LC–MS/MS, and decreased levels of acetic and hexanoic acids were found in the CCl₄ group, while this decrease in acetic and hexanoic acids was more pronounced after SNS administration (Figure 2B). Isobutyric acid levels were increased in the CCl₄ group, while SNS treatment tends to rescue the change in isobutyric acid with no statistical significance (Figure 2B). Propionic, butyric, and isovaleric acids were increased in the CCl₄ group, which was significantly alleviated by SNS administration (Figure 2B). To study whether changes in bacterial communities have an impact on microbial metabolic output, a correlation analysis was conducted on three groups of bacterial communities with differences in the genus level and SCFAs (Figure 2C). We observed that at the genus level, bacterial groups positively correlated with propionic acid include Faecalibaculum, Enterorhabdus, Bifidobacterium, and Parasutterella, whereas those negatively correlated include Incertae sedis. Bacterial groups positively correlated with butyric acid comprise Faecalibaculum, Enterorhabdus, Monoglobus, Bifidobacterium, Parasutterella, and Muribaculum, with negatively correlated groups being Incertae sedis and Prevotellaceae-UCG-001. Finally, bacterial groups positively correlated with isovaleric acid are Faecalibaculum, Bifidobacterium, and Parasutterella. Next, we constructed a histogram of the intestinal flora related to propionic, butyric, and isovaleric acids. The results showed that SNS reduced the abnormal increase in Bifidobacterium, Enterorhabdus, Parasutterella, Monoglobus, and Muribaculum bacteria and alleviated the decrease in Incertae sedis (Figure 2D) Typically, SNS treatment regulated the structure and composition of the intestinal bacterial community related to the contents of propionic, butyric, and isovaleric acids, which suggests that SNS improves liver fibrosis possibly by regulating the levels of propionic, butyric, and isovaleric acids metabolized by intestinal flora.

Furthermore, we examined the expression of SCFA receptors FFAR2 and FFAR3 in liver tissues. Western blot results showed that SNS administration significantly reduced CCl₄-induced hepatic expression of both FFAR2 and FFAR3 (Figure 3A), which was further consistently supported by the analyses of RT-qPCR and liver immunofluorescence (Figures 3B–D). These data suggest that SCFAs may play an important role in the progression of liver fibrosis.

Given the close association between SCFAs and hepatocyte injury, we further investigated the effects of various SCFAs on hepatocyte apoptosis (Singh et al., 2015; Kobayashi et al., 2018; Shao et al., 2024). HepG2 cells were first treated with increasing doses of propionic, butyric or isovaleric acids alone for 24 h, and cell viability was assessed. Cell viability was not found to be significantly decreased after treatment with 500 μM propionic acid, 200 μM butyric acid, and 20 μM isovaleric acid, suggesting that the tested SCFAs alone did not cause hepatocyte cell death (Figure 4A). Next, the effect of SCFAs on potentiating the TNF-α-induced apoptosis was further explored. FFAR2/3 expression was increased, and apoptosis was aggravated in the TNF-α-treated group compared with the control group. After the administration of propionic, butyric, and isovaleric acids, the expression of FFAR2/3 was further increased compared to that in the TNF-α group, while apoptosis was found to be significantly aggravated by CCK-8 assay (Figure 4B), which was further confirmed by the fluorescence results of apoptosis (Figure 5A).

To investigate the potential influence of intestinal microbes on the effectiveness of SNS against liver fibrosis, intestinal microbiota were eradicated by administering a cocktail of antibiotics in drinking water. The analyses of biochemical parameters showed that the SNS treatment failed to improve CCl₄-induced liver injury and fibrosis in the presence of antibiotics in drinking water (Figure 6A). Consistently, H&E staining showed that in the absence of intestinal flora, SNS failed to improve hepatocyte cell death and inflammatory cell infiltration in CCl₄-treated mice. Additionally, in terms of fibrosis, Masson staining confirmed that the ameliorative effect of SNS on the formation of pseudo-leaflet and fibrotic septa was also dependent on the presence of intestinal flora (Figure 6B), which was further confirmed by the fluorescence results of α-SMA (Figure 6C) and the Western blot results (Figure 6D).

We then asked whether SNS treatment reduces hepatocyte apoptosis by regulating the intestinal microbiota-derived SCFA metabolites and propionic, butyric, and isovaleric acid content. The results showed that mice fed antibiotics had lower levels of propionic acid, butyric acid, and isovaleric acid compared to CCl₄-treated wild-type mice, though the difference in butyric acid was not statistically. Additionally, antibiotics mitigated the effects of SNS on propionic acid, butyric acid, and isovaleric acid (Figure 7A). Next, we detected the expression of FFAR2/3 using immunofluorescence, and the results showed that the expression of FFAR2/3 was significantly reduced in antibiotic-treated mice (Supplementary Figure S4). To further investigate the impact of antibiotic cocktail on liver cell apoptosis, we performed TUNEL staining and Western blot analyses. The results showed that apoptosis of liver cells in the PG + CCl₄ group was significantly lower than that in the wild-type CCl₄-induced mice (Figures 7B, C). All these data demonstrate that in the presence of antibiotics, the anti-apoptotic effect of SNS on hepatocytes may be related to the modulation of intestinal flora, particularly SCFA-associated species.