DO powder was purchased from Yunnan Tianbao Betula Biological Resources Development Co., Ltd. (Yunnan, China) and Atorvastatin calcium (AT) tablets were purchased from Beijing Jialin Pharmaceutical Co., Ltd. (Beijing, China). The normal diet was purchased from Jiangsu Medisen Biological Medicine Co., Ltd. (Bei Jing, China). Kits used to measure alanine transaminase (ALT), aspartate transaminase (AST), triglyceride (TG), total cholesterol (TC), low-density lipoprotein cholesterol (LDL-c), high-density lipoprotein cholesterol (HDL-c), and gamma-glutamyl transpeptidase (GGT) and the total protein assay were purchased from the Nanjing Jiancheng Bioengineering Institute (Nanjing, China). Kits used to measure interleukin-6 (IL-6), interleukin 1-β (IL-1β), LPS, tumor necrosis factor-α (TNF-α), and diamine oxidase (DAO) were purchased from the Jiangsu Meimian Industrial Co., Ltd. (Jiangsu, China). D-lactate (D-LA) kit was purchased from Jiangsu Addison Biotechnology Co., Ltd. (Jiangsu, China). Anti-Occludin rabbit pAb, anti-Claudin-1 rabbit pAb, anti- Zona occludens-1 (ZO-1) rabbit pAb, and anti-NF-kB p65 (p 65) rabbit pAb were purchased from Servicebio (Wuhan, China). Ultrapure RNA Kit, cDNA Synthesis Kit, and UltraSYBR Mixture were purchased from CoWin Biosciences (Jiangsu, China). TLR4 rabbit pAb was purchased from Proteintech Group, Inc (Rosemont, USA). Phospho-NF-kB p65 (p-p65) antibody was purchased from Affinity Biosciences Ltd. (OH, USA). Anti-beta actin (β-actin) antibody was obtained from Abcam Inc. (Cambridge, UK).

All experimental procedures followed the guidelines of the Animal Ethics Committee of the Yunnan University of Chinese Medicine (Approval lot: R-062021024).

Healthy male Sprague-Dawley (SD) rats (180–200 g; SPF) were provided by Hunan Slike Jingda Laboratory Animal Co., Ltd. (Hunan, China). The rats were maintained in a specific pathogen-free standard environment on the public platform for animal experiments in the Science and Technology Department of the Yunnan University of Chinese Medicine. The rearing temperature was 20–25°C and the relative humidity was 50% ± 10%, with 12 hours of alternating light. All experimental rats had free access to distilled water and were fed a normal or a high-fat diet (HFD, 82.5% normal diet, 10% lard, 2% cholesterol, 0.5% sodium cholate, and 5% egg yolk powder).

After 1 week of adaptive feeding, the rats were randomized into the following six groups (n = 8 rats per group): 1) Control group, fed with a normal diet and gavaged with distilled water; 2) HFD group, fed with a HFD and gavaged with distilled water; 3) AT group, fed with a HFD and gavaged with 20 mg/(kg·d) AT; 4) High-dose DO group (DOH), fed with a HFD and gavaged with 1000 mg/(kg·d) DO powder; 5) Middle-dose DO group (DOM), fed with a HFD and gavaged with 500 mg/(kg·d) DO powder; 6) Low-dose DO group (DOL), fed with a HFD and gavaged with 250 mg/(kg·d) DO powder. AT and DO powder were prepared separately using distilled water. Rats were gavaged with the corresponding drug (or distilled water) once a day for 10 weeks.

Body weights were recorded weekly during the experiment. After the last administration at the end of 10 weeks, rat feces were collected from the anus of each rat using sterile EP tubes and immediately preserved in liquid nitrogen. Rats were euthanized after fasting for 12 hours, and liver tissue, small intestine tissue, and serum samples were collected. Serum samples and a portion of both the liver and small intestine tissues were stored at -80°C.

Serum AST, ALT, GGT, TG, TC, LDL-c, HDL-c, and D-LA levels and liver TG, TC, LDL-c, and HDL-c levels were measured using a commercial kit.

LPS levels in the liver, serum, and ileum, IL-6, IL-1β, TNF-α levels in the liver and ileum, and DAO levels in the ileum were detected using ELISA kits.

Liver and ileum tissues fixed in 4% paraformaldehyde solution were dehydrated with different concentrations of ethanol, made transparent with xylene, embedded in liquid paraffin, stained with H&E, and sealed with neutral gum. The fixed liver tissue was dehydrated in different concentrations of sucrose solution, embedded in an optimal cutting temperature compound, sliced using a cryostat, stained with oil red O, and sealed with glycerol gelatin. A slide scanning image analysis system (Shenzhen Shengqiang Technology, China) was used to observe the staining of the pathological sections at 400x, and the oil red O-positive area was analyzed by ImageJ software (NIH, Bethesda, MA, USA).

TLR4, p-p65 and p65 expression in liver tissues were determined by Western blot. Liver tissue (50 mg) and 0.5 ml RIPA lysate were added to an EP tube, ground for 60 s, and centrifuged at 4°C at 10,000×g for 10 min. BCA protein quantification was used to measure the protein concentration. The protein solution was added to a 5x reduced protein loading buffer at a ratio of 4:1 and denatured in a boiling water bath for 15 min. Electrophoresis was conducted at 80V for 20 min and then at 120V until the bromophenol blue ran to a position 1 cm from the lower end of the glass plate. TLR4 (1:8000), p-p65 (1:1000), p65 (1:1000), and β-Actin (1:2000) were incubated for 60 min and washed with TBST until no skimmed milk powder was present. The universal secondary antibodies (1:5000) were incubated for 60 min at room temperature and washed three times with TBST for 5 min each. Immunoreactive protein bands were visualized with a chemiluminescence HRP substrate using a ChemiDoc XRS image detector (Jena Analytical Instruments AG, Jena, Germany). The blots were analyzed using ImageJ software.

Sections of paraffin-embedded ileum tissue were deparaffinized and rehydrated. Antigen was repaired using citric acid antigen repair buffer and 3% hydrogen peroxide was used to block any endogenous peroxidase. BSA was added dropwise for serum blocking followed by the addition of ZO-1 (1:1000), occludin (1:1000), or claudin-1 (1:800). After incubating the samples overnight, a secondary antibody was added dropwise. The colour was developed with DBA, the cell nuclei were re-stained, and the samples were dehydrated to seal the slides. A slide scanning image analysis system (Shenzhen Shengqiang Technology, China) was used at 400x to observe the samples. Ultimately, the Image-Pro Plus software (U.S. MEDIA CYBERNETICS) was used to count the mean density and analyze the integrated optical density (IOD) of positive staining.

Total RNA was extracted from liver tissue using Ultrapure RNA Kit and was reverse transcribed into cDNA using cDNA Synthesis Kit. The PCR cycle system was set as follows: at 95°C for 10 min, at 95°C for 15 s, and at 60°C for 60 s, for a total of 40 cycles. A total of 1μL cDNA template was used for PCR amplification using the following primers: TLR4, forward: 5′-CCGCTCTGGCATCATCTTCA-3′, reverse: 5′-TGGGTTTTAGGCGCAGAGTT-3′; GAPDH, forward: 5′-GCCCAGCAAGGATACTGAGA-3′, reverse: 5′-GGTATTCGAGAGAAGGGAGGGC′.

Statistical analysis was conducted using SPSS software (Version 25, SPSS Inc., Chicago, USA) and graphs were created using GraphPad Prism software (Version 9.4.0, GraphPad Software Inc., CA, USA). Data are shown as the mean ± SD. Normally distributed data were tested by one-way ANOVA followed by the LSD or Dunnett’s test. Other types of data were tested using the non-parametric Mann-Whitney test (Kruskal-Wallis test for multiple groups). Differences were considered statistically significant at a p-value <0.05.

HFD rats had a higher food intake than those in the other groups at 4 weeks. While food intake gradually decreased in the last 6 weeks, possibly due to an anorexic reaction caused by overeating the HFD, there was no significant difference in intake among the groups ( Figure 1A ). The HFD rats gained weight faster than rats in the other groups ( Figure 1B ). HFD rats had significantly higher body weight (p-value < 0.001), body mass index (p-value <0.05), liver weight (p-value <0.001), and liver index (p-value <0.001) than Control rats. The AT and DO interventions significantly reduced the body weight, body mass index, liver weight, and liver index (p-value <0.05, <0.01, and <0.001, respectively) ( Figures 1C–F ). In addition, the livers of HFD rats were significantly larger and more yellow than those of the Control, AT, and DO rats, suggesting that HFD rat livers may have accumulated more lipids ( Figure 1G ). These results indicate that DO effectively inhibited HFD-induced weight gain, lipid deposition, and liver enlargement.

After 10 weeks of HFD feeding with or without DO treatment, liver cells from Control rat livers were neatly arranged, the liver cords were clear, and no obvious lipid deposition was observed. In contrast, liver cells from HFD rat livers were disordered, with more extensive and robust steatosis accompanied by intralobular inflammatory foci and balloon-like changes. Inflammation and steatosis were lower in the livers of rats in the DO and AT groups than those in the HFD group ( Figure 2A ). The NAFLD activity score (NAS), semi-quantitative data used to assess NAFLD progression (NAS >4), showed that NASH was occurring in the HFD group, indicating that a HFD successfully induced NASH. NAS scores were significantly lower following DO and AT treatment (p-value <0.001) ( Figure 2C ). Liver oil red O staining showed no obvious lipid deposition in Control rats and a large amount of deposition in HFD rats (p-value <0.001). The area of lipid deposition was decreased following DO and AT treatment (p-value <0.01 and <0.001, respectively) ( Figures 2B, D ). These results supported the success of NASH modeling in this study and indicated that DO treatment reduces lipid accumulation and inflammation in rat livers.

TC, TG, HDL-c, and LDL-c are clinical indicators used to reflect blood lipid levels and lipid metabolism. HFD significantly increased serum and liver TC, TG, and LDL-c and reduced HDL-c levels (p-value <0.001). These blood and liver lipid indexes were significantly improved following DO and AT treatment (p-value <0.05, <0.01, and <0.001, respectively) ( Figures 2E, F ). ALT, AST, and GGT are the most used clinical indicators of liver function. When the liver is damaged, hepatocytes produce these proteins, resulting in an increase in serum ALT, AST, and GGT levels and indicating the occurrence of liver disease and inflammation. HFD significantly increased serum ALT, AST, and GGT levels (p-value <0.001) and these indexes were significantly decreased following DO and AT treatment (p-value <0.05, <0.01, and <0.001, respectively) ( Figure 2G ).

Structural changes in the gut microbiota of rats that received DO treatment were assessed using 16S rRNA sequencing analysis. A total of 1,548,295 sequences were obtained from 30 samples, and 1,033 OTUs were collected at a 97% similarity level. The species accumulation and rank-abundance curves showed that most of the diversity and species were included and the amount of sequencing data was adequate ( Figures 3A, B ).

Alpha diversity was determined using Chao, Qstat, and Smithwilson to calculate the complexity of species diversity in samples. The Chao, Qstat, and Smithwilson indexes describe the richness, diversity, and evenness of the gut microbiota, respectively. In the current study, changes in these indexes are shown in Figures 3C–E . The Chao (p-value <0.01) and Qstat (p-value <0.001) were lower and the Smithwilson (p-value <0.001) was higher in the HFD group than in the Control group, suggesting that a HFD can significantly reduce microbiome richness, diversity, and evenness. The Chao and Qstat increased and the Smithwilson decreased following DO treatment (p-value <0.05 and <0.01, respectively) suggesting that DO may effectively improve the diversity, richness, and evenness of gut microbiota in NASH rats.

Beta diversity principal component analysis (PCA) was performed to clarify the effects of HFD and DO intervention on the composition and structure of the gut microbiota. As expected, PCA revealed a clear separation between the Control and HFD groups, and the composition of the gut microbiota exhibited a clear response to DO intervention ( Figure 3F ). The unique and common OTUs in the Venn diagram more directly indicated the unique species information of each group. Different numbers of OTUs were detected in each group, including 884 in the Control group, 627 in the HFD group, 711 in the DOH group, 702 in the DOM group, and 754 in the DOL group. A total of 462 OTUs were shared by all groups and each group had unique OTUs. These results showed that DO could ameliorate the gut microbiota disorder induced by HFD in NASH rats ( Figure 3G ).

To evaluate the effect of DO on the gut microbiota of NASH rats, the microbial abundance at the phylum, genus, and species levels was determined using taxonomic analysis. At the phylum level, 13 phyla were found in five groups, of which Firmicutes and Bacteroidetes accounted for the largest proportion. The abundance of Firmicutes, Bacteroidetes, and Proteobacteria related to NASH was changed ( Figure 4A ). Proteobacteria were positively correlated with the HFD group ( Figure 4B ). While the relative abundances of Firmicutes and Proteobacteria were higher in the HFD group than in the Control group (p <0.05), Bacteroidetes were lower in the HFD group (p <0.05). After DO intervention, the relative abundance of Proteobacteria was significantly decreased in the DOH group (p <0.05) ( Figure 4C ). These results indicated that compared with the Control group, the composition of gut microbiota in the HFD group changed significantly at the phylum level, especially for Proteobacteria.

The abundance of gut microbiota at the genus and species levels also differed by group ( Figure 5A , 6A ). To identify specific bacterial taxa that arose after DO supplementation, LEfSe analysis (all-against-all) with a 3.0 threshold for discriminative features on the logarithmic LDA scale was performed. In LEfSe, different colours represent different groups. The potentially harmful bacteria, Romboutsia, Turicibacter, Lachnoclostridium, Blautia, Ruminococcus_torques_group, Sutterella, and Escherichia-Shigella were enriched in the HFD group at the genus level ( Figure 5B ). DO treatment reduced the abundance of Romboutsia, Turicibacter, Lachnoclostridium, Blautia, Ruminococcus_torques_group, Sutterella, and Escherichia-Shigella than the HFD group (p-value <0.05, <0.01, and <0.001, respectively) ( Figure 5C ). Meanwhile, the abundance of the potentially beneficial bacteria, Prevotella and Alistipes, was significantly lower in the HFD group than in the Control group (p-value <0.05 and <0.001, respectively), and DO intervention increased the levels of these organisms (p-value <0.05) ( Figure 5C ). At the species level, the probiotic Lactobacillus_acidophilus was significantly enriched in the DOH group ( Figure 6B ). The HFD group had a significantly lower abundance of Lactobacillus_acidophilus than the Control group (p-value <0.05), and the level of this bacteria increased significantly in the DOH and DOL groups (p-value <0.01) ( Figure 6C ).

The gut microbiota plays an important role in maintaining the integrity of the intestinal mucosal barrier, and intestinal tight junction protein is the principal determinant of intestinal permeability. Each layer of the ileum tissue in the Control and the AT groups was clearly structured, the mucosal epithelium was intact, cell morphology was normal, the intestinal villi were evenly distributed, the intestinal glands were abundant and tightly arranged, and no obvious abnormalities were found. In contrast, there was necrotic cellular debris in the intestinal lumen of the HFD group and the apical epithelium of the intestinal villi was separated from the lamina propria. This separation was much rarer in rats receiving different doses of DO ( Figure 7A ). Expression of the tight junction proteins, ZO-1, claudin-1, and occludin was significantly lower in the HFD group than in the Control group (p-value <0.001), and all three proteins were significantly increased following DO intervention (p-value<0.05, <0.01, and <0.001, respectively) ( Figure 7B ). Ileum DAO and serum D-LA levels were significantly higher in the HFD group (p-value <0.001), and DAO and D-LA levels were significantly lower after DO and AT treatment (p-value <0.05, <0.01, and <0.001, respectively) ( Figure 7C ). Ileum IL-6, TNF-α, and IL-1β levels were all significantly higher in the HFD group than in the Control group (p-value <0.001), and significantly decreased following DO and AT treatment (p-value <0.05, <0.01, and <0.001, respectively) ( Figure 7D ). The protective effect of DO on intestinal permeability was manifested by increased tight junction protein expression and lower DAO activity and D-LA and inflammatory cytokine production.

At the phylum level, Proteobacteria were most dramatically changed after 10 weeks of HFD, suggesting that there was a concomitant rise in LPS. Disruption of the gut microbiota and increased intestinal permeability allow LPS to enter the liver. TLR4 and NF-κB are important proteins related to liver inflammation during NASH, and IL-6, IL-1β, and TNF-α were the major inflammatory cytokines induced by TLR4 and NF-κB.

LPS levels in the ileum, serum and liver were significantly higher in rats in the HFD group than in the Control group (p-value <0.001) ( Figure 8A ), and relative TLR4 mRNA expression in the liver was significantly increased (p-value <0.001) ( Figure 8B ). DO and AT treatment resulted in significantly lower LPS levels in the ileum, serum, and liver (p-value <0.05, <0.01 and <0.001, respectively) and reduced relative TLR4 mRNA expression (p-value <0.001). Protein expression of TLR4 and the p-p65/p65 ratio were significantly higher in the HFD group than in the Control group (p-value <0.001) and both were decreased following DO and AT treatment (p-value <0.05, <0.01, and <0.001, respectively) ( Figure 8C ). IL-6, IL-1β, and TNF-α levels were significantly higher in the HFD group than in the Control group (p-value <0.001) and were significantly decreased following DO and AT treatment (p-value <0.01 and <0.001, respectively) ( Figure 8D ). These results suggested that DO was able to attenuate liver inflammation.

Spearman correlation analysis was performed to assess the potential correlation between gut microbiota and the levels of biochemical factors and LPS. The abundances of Firmicutes, Proteobacteria, Romboutsia, Turicibacter, Lachnoclostridium, Blautia, Ruminococcus_torques_group, Sutterella, and Escherichia-Shigella were positively correlated with TG, TC, LDL-c, AST, ALT, GGT, and LPS levels. Meanwhile, the abundances of Bacteroides, Lactobacillus_acidophilus, Prevotella, and Alistipes were positively correlated with HDL-c levels ( Figure 9A ).