Sixty specific-pathogen-free (SPF) male Sprague-Dawley rats (batch number:110324211103957112), 8-week-old and weighing (220 ± 20) g, were obtained from SiPeiFu Biotechnology Co., Ltd. (Beijing, China). The rats had ad libitum access to food and water, and were housed in the barrier environmental facility at Beijing University of Chinese Medicine, which was maintained at (23 ± 2) °C temperature, 55%–65% relative humidity, and a 12 h/12 h light/dark cycle. The experiment was approved by the Medical and Laboratory Animal Ethics Committee of Beijing University of Chinese Medicine and the Animal Ethics Committee of University of Macau, and the approval numbers were BUCM-4-2021071207-3178 and UMARE-008-2021.

Rats were acclimated for 1 week, then were randomly assigned 10 to the Normal group and 50 to the high-fat, high-fructose diet (Model) group based on body weight and fasting blood glucose (FBG) levels. During the 12-week modeling period, the rats in the Normal group were fed a standard chow diet (SWS9102), while those in the Model group were fed a high-fructose, high-fat diet (XT301-1). The diets were supplied by Xietong Pharmaceutical Bio-engineering Co., Ltd. (Jiangsu, China). And the specific compositions of these diets are shown in Supplementary Table S1. The evaluation of the MetS model was conducted according to a translational guide developed by Inês Preguiça (Preguiça et al., 2020). The evaluation criteria for MetS model in rats was determined by the presence of at least three out of the following five manifestations: 1) obesity; 2) reduced high-density lipoprotein; 3) elevated triglycerides; 4) elevated fasting blood glucose or impaired glucose tolerance; 5) elevated systolic blood pressure. Successful modeling is defined as having mean values in the Model group that are 10% higher than those in the Normal group.

The successful model rats were randomly divided into five groups (n = 8 per group) according to body weight, blood glucose and lipid levels: Model group, Rosiglitazone (Rosi) group, XSJ high dose (XSJH) group, XSJ medium dose (XSJM) group, and XSJ low dose (XSJL) group. The Normal and Model groups received equal amounts of distilled water daily.

The raw extract Xiasangju (XSJ) (Batch No. 20210812) were obtained from of the Chinese patent medicine Baiyunshan Xingqun XSJ granules (Chinese drug approval number Z44022217) provided by Baiyunshan Xingqun Pharmaceutical Co., Ltd. (Guangzhou, China). According to the conversion ratio between rat and humans, the daily extract dose (g/kg) of XSJ for rats can be calculated as the following formula: the conversion ratio of surface area between rat and human (6) × a daily dose of an adult (4.8 g extract)/the average weight of an adult (60 kg). That is: 6 × 4.8 (g)/60 (kg) = 0.48 g/kg as the rat equivalent dose. Therefore, in the study, 0.24 g/kg is regarded as low dose and 0.96 g/kg as high dose. And the rats of XSJ groups were given XSJ orally once daily. The rats of Rosi group were given rosiglitazone (No. H20030569, Hengrui, China) 0.4 mg/kg by gavage once a day as the daily human dose of rosiglitazone (4 mg/d).

During administration, the Normal group continued to be fed a standard chow diet and the other five groups were fed a high-fructose, high-fat diet. Body weight, abdominal circumference, nasoanal length and blood glucose were measured after a 12-h fast at a fixed time each week. And body mass index (BMI) was calculated as weight (kg)/[height (m)] ^ 2. Blood pressure was measured using the BP-2010A tail-cuff blood pressure measurement system (Softron Biotechnology, China) which records systolic (SBP), diastolic (DBP), and mean (MBP) arterial pressures with 3 repetitions at 15-min intervals.

An oral glucose tolerance test (OGTT) was performed at the 12th week of high-fat, high-fructose diet modeling and the end of the 16th week of XSJ treatment. Before administering oral glucose loading (2.0 g/kg) for the OGTT, rats were fasted for 12 h. Blood glucose was measured before (0) and 15, 30, 60, 90, and 120 min after glucose administration with a glucometer (ACCU-CHEK^® Performa, Roche, U.S.A). Glucose levels were plotted over time and the area under the curve (AUC) of OGTT was calculated. HOMA-IR was calculated according to the following formula: fasting glucose levels (mmol/L) × fasting serum insulin (mU/L)/22.5.

Rats were fasted for 10 h before each assay. After a 12-week modeling period, blood was collected from the retro-orbital venous plexus of rats after 4% isoflurane (RWD, China) inhalation anesthesia for model evaluation. After 16 weeks of XSJ administration, rats were anesthetized with 25% urethane (Yuanye, China) and blood was collected from the abdominal aorta. Serum was collected after centrifugation at 4°C, 3,000 rpm for 10 min. Samples of the liver, interscapular brown adipose tissue (BAT), epididymal white adipose tissue (eWAT), inguinal subcutaneous white adipose tissue (iWAT) and colon were collected. The organ coefficients for the liver, BAT, eWAT, and iWAT were calculated as follows: organ coefficient = organ weight (g)/rat body weight (kg) × 1000. Samples were either tested immediately after collection or frozen in liquid nitrogen and stored at −80°C.2.5.

The pathology of liver, BAT, iWAT, and colon tissues was assessed using hematoxylin and eosin (H&E) staining. All tissue samples were fixed in 10% neutral formalin, dehydrated in ethanol, and embedded in paraffin wax. 5 μm-thick tissue sections were stained with H&E dye before pathology slice scanner observation (Roche, Switzerland). Hepatic fat accumulation was assessed using Oil Red O staining. Colon tissues were embedded in glutaraldehyde and dehydrated in an ethanol-acetone gradient, and colonic epithelial tight junctions were assessed by TEM.

The biochemical assay kits were used to determine serum lipid profile (triglycerides (TG), total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C) and low-density lipoprotein cholesterol (LDL-C)) in rats for model evaluation. The serum lipid indices, as well as the concentrations of TG and TC in both liver and BAT, were measured following a 16-week treatment period. All of the above-mentioned kits were acquired from Nanjing Jiancheng Bioengineering Institute, China. Lipopolysaccharide (LPS), tumor necrosis factor-alpha (TNF-α), monocyte chemotactic protein-1 (MCP-1), interleukin-6 (IL-6), glucagon-like peptide-1 (GLP-1) and insulin (INS) of rat serum were quantified by enzyme-linked immunosorbent assay (ELISA) kits (Kete, China).

100 mg of cecal contents were homogenized and resuspended in pre-chilled 80% methanol. After centrifuging at 15,000 g, 4°C for 20 min, then the supernatant was diluted to a final concentration of 53% methanol, transferred to a new tube and centrifuged again. Finally, the supernatant was injected into the UHPLC-MS/MS system (Thermo Fisher, Germany) using a 17 min linear gradient at a flow rate of 0.2 mL/min. The eluents for the positive polarity mode were eluent A (0.1% formic acid in Water) and eluent B (Methanol). The eluents for the negative polarity mode were eluent A (5 mM ammonium acetate, pH 9.0) and eluent B (Methanol). The solvent gradient was set as follows: 2% B, 1.5 min; 2%–85% B, 3 min; 85%–100% B, 10 min; 100%–2% B, 10.1 min; 2% B, 12 min. Mass spectrometer was operated in positive/negative polarity mode with spray voltage of 3.5 kV, capillary temperature of 320°C, sheath gas flow rate of 35 psi and aux gas flow rate of 10 L/min, S-lens RF level of 60, Aux gas heater temperature of 350°C. The raw data files generated by UHPLC-MS/MS were processed using Compound Discoverer 3.1 (CD3.1, Thermo Fisher, United States) to quantitation for each metabolite. These metabolites were annotated using the KEGG Database, HMDB Database and LIPIDMaps Database. Principal components analysis (PCA) and Partial least squares discriminant analysis (PLS-DA) were performed. We applied univariate analysis (t-test) to calculate the statistical significance (p-value). The metabolites with VIP >1, p < 0.05 and |log₂(FoldChange)| ≥ 1.5 were considered to be differential. The metabolic pathways were enriched by differential metabolites using the KEGG Database.

Total genomic DNA was extracted from 200 mg of cecal contents samples using the CTAB method and amplified with 341F—806R barcoded primers for the 16S rDNA V3—V4 regions. PCR products were mixed in equidensity ratios and purified with Qiagen Gel Extraction Kit (Qiagen, Germany). Then purified amplification products were sequenced to construct a sequencing library using TruSeq^® DNA PCR-Free Sample Preparation Kit (Illumina, United States). The sample library was sequenced on an Illumina NovaSeq platform. Sequences with ≥97% similarity were assigned to the same operational taxonomic units (OTUs) using the Silva database (http://www.arb-silva.de/) to annotate taxonomic information based on the Mothur algorithm.

According to the taxonomic classification and OTU abundance, the abundance of each sample at each taxonomic level of phylum, class, order, family and genus was calculated and visualized by phyloseq (McMurdie and Holmes, 2013). Macrogenomic function predictions were based on Tax4Fun2 (Wemheuer et al., 2020).

Total genomic DNA of BAT and iWAT was isolated by DNeasy Blood & Tissue Kit (Qiagen, United States) according to the manufacturer’s instructions. Quantitative PCR was carried out using a Step One Real-Time PCR System (Applied Biosystems, United States) and PCR products were detected by using the ChamQ Universal SYBR qPCR Master Mix (Vazyme, China). Mitochondrial DNA (mtDNA) copy number was estimated by quantifying the genomic DNA levels of mtDNA-transcribed Cox2, and normalized to that of the nuclear-specific gene β-actin (Kowluru et al., 2014). Primers used are listed in Supplementary Table S2.

The BAT and colon tissue were homogenized and the protein concentrations were measured using a BCA protein detection kit (NCMbiotech, China). Total protein (10 μg) was separated by SDS-PAGE, and transferred to PVDF membranes (Millipore, United States). Membranes were incubated overnight at 4°C with primary antibodies. BAT primary antibodies sources were as follows: Acyl-CoA oxidase 1 (ACOX1) (1:1000, ab184032, Abcam), Adipose triglyceride lipase (ATGL) (1:1000, ab207799, Abcam), Carnitine palmitoyl transferases 1 α (CPT1α) (1:1000, ab234111, Abcam), β3 Adrenoceptor (ADRB3) (1:1000, YT0363, Immunoway), Dopamine β-hydroxylase (DBH) (1:1000, YT1296, Immunoway), Protein kinase A (PKA) (1:10000, ab134901, abcam), Uncoupling Protein 1 (UCP1) (1:1000, 23673-1-AP, Proteintech), Peroxisome proliferator-activated receptor α (PPARα) (1:1000, ab178860, Abcam), PPAR γ coactivator 1α (PGC-1α) (1:1000, ab188102, Abcam), PRD1-BF-1-RIZ1 homologous domain containing protein 16 (PRDM16) (1:1000, NBP1-71992, Novus), NADH dehydrogenase [ubiquinone] 1 beta subcomplex subunit 8 (NDUFB8) (1:5000, 14794-1-AP, Proteintech), Succinate dehydrogenase complex iron sulfur subunit B (SDHB) (1:5000, 10620-1-AP, Proteintech), Ubiquinol-cytochrome c reductase core protein 2 (UQCRC2) (1:2000, 14742-1-AP, Proteintech), Mitochondrially encoded cytochrome c oxidase II (MTCO2) (1:2000, 55070-1-AP, Proteintech), ATP synthase F1 subunit α (ATP5A1) (1:5000, 14676-1-AP, Proteintech) and β-Tubulin (1:1000, 10094-1-AP, Proteintech). Colon tissue primary antibodies sources were as follows: ZO-1 (1:1000, 21773-1-AP, Proteintech), Occludin (1:1000, 27260-1-AP, Proteintech), Claudin 1 (1:1000, 13050-1-AP, Proteintech) and β-Tubulin (1:1000, 10094-1-AP, Proteintech). Membranes were incubated for 60 min with Goat anti-Rabbit IgG-HRP (1:1000, ab205718, Abcam) at room temperature. Protein band signals were detected using the ChemiDoc MP Imaging System (Bio-Rad, United States). ImageJ software (NIH, United States) was used for quantitative analysis of protein bands. All target proteins were normalized by β-Tubulin.

GraphPad Prism 9.0 and R 4.3.1 were used to perform the statistical analyses and visualize the data. All data were expressed as the means ± standard deviation (SD). Statistical significance was tested with one-way ANOVA (and nonparametric or mixed). LSD test was used to assess intergroup differences for normal data. And Mann-Whitney U test was used to assess intergroup differences for non-normal data. p < 0.05 was considered statistically significant.

After 12 weeks of high-fructose, high-fat diet, the body weight, abdominal circumference and body mass index (BMI) in the Model group was higher than that in the Normal group (Figures 1A–C, p < 0.01). In addition, the adipose tissue and liver index of the Model group rats was higher than that of the Normal group (Supplementary Table S3). In terms of lipid indices, Model rats exhibited decreased HDL-C and increased LDL-C, whereas no statistical differences were seen between TG and TC and Normal group rats (Figure 1D, p < 0.01). FBG level in the Model group was higher than that in the Normal group (Figure 1E, p < 0.05). Moreover, the Model group exhibited impaired glucose tolerance in the oral glucose tolerance test compared with the Normal group (Figures 1F,G, p < 0.05). Taken together, the Model group rats showed increased body weight, abdominal circumference and BMI, higher adipose tissue and liver index, decreased HDL-C, and increased FBG along with impaired glucose tolerance (Supplementary Table S4), so the Model group rats were assessed to be consistent with the MetS model.

Model rats lost weight after 16 weeks of XSJ administration (Figure 2A, p < 0.05). The Model group rats’ abdominal circumference was significantly larger than the Normal group, whereas XSJ was able to reduce the abdominal circumference (Figure 2B, p < 0.01). There was no significant difference in BMI between the Model group rats and the Normal group rats (Figure 2C, p > 0.05). XSJ also reduced the calorie intake compared with the Model group rats (Figures 2D,E, p < 0.01). And rats in the Model group showed a decrease in BAT index and an increase in eWAT, iWAT and liver index, whereas XSJ was able to increase BAT index and decrease eWAT, iWAT and liver index (Figures 2F–I, p < 0.01). There was a significant increase in BAT TG and liver TG, TC levels in the Model rats compared with the Normal group. Conversely, XSJ was able to improve BAT TG and liver TG, TC levels ((Figure 2J–Q, p < 0.01). In addition, adipocytes in the BAT of the Model group showed the morphology of large lipid droplets in a single compartment in H&E staining, and adipocytes were reduced after XSJ administration. Furthermore, based on H&E and Oil Red O staining analysis of the liver, XSJ treatment reduced the number of lipid droplets in the liver and the degree of liver tissue lesions compared to the Model group (Figures 2R,S, p < 0.05). These results indicate that XSJ can effectively reduce fat accumulation in Model rats.

The rats in the Model group showed elevated fasting blood glucose (FBG) and impaired glucose tolerance, while XSJ was able to improve FBG and glucose tolerance in the rats in the Model group (Figures 3A–E, p < 0.01). The results of the insulin resistance index showed that the HOMA-IR in the Model group was significantly higher than that in the Normal group, whereas XSJ could significantly reduce the HOMA-IR (Figure 3E, p < 0.01). Moreover, XSJ could significantly increase the serum GLP-1 concentration (Figure 3F, p < 0.05). Our findings show that XSJ benefits insulin resistance in Model rats.

Based on the attenuation of fat accumulation in BAT and liver and the improvement of insulin resistance, we investigated the mtDNA copy number and the expression of lipid β-oxidation proteins in BAT, including PPARα, ACOX1, CPT1, and lipolysis proteins such as ATGL in BAT. The findings revealed that the mtDNA copy number was enhanced in BAT after XSJ treatment (Figure 3G, p < 0.01). Additionally, XSJ increased mtDNA copy number in iWAT compared to the Model group (Supplementary Figure S1, p < 0.01). Both PPARα, ACOX1, CPT1, and ATGL were downregulated in the Model rats, whereas XSJ upregulated their expression (Figures 3H,I, p < 0.05). These results indicate that XSJ can promote lipolysis and β-oxidation in BAT to ameliorate metabolism disorders by regulating the expression of PPARα, ACOX1, CPT1 and ATGL.

XSJ exhibited a beneficial influence on the stimulation of BAT, prompting us to delve deeper into the underlying mechanism by which XSJ fosters metabolism through metabolite alterations. The metabolomics data indicated clear separation between the Normal and Model groups in both PCA and PLS-DA plots. The XSJH group exhibited a gradual separation from the Model group, indicating its ability to regulate the metabolites of the rats in the Model group, ultimately leading to an improvement in MetS. (Figures 4A,B).

Differential metabolite screening was performed according to |log2FC|≥1.5, VIP≥1, p < 0.05, as shown in Figure 4C where there was a total of 504 altered metabolites between the Normal group and the Model group, and 186 altered metabolites between the XSJH group and the Model group, with a total of 52 shared altered metabolites (Supplementary Table S5). The altered metabolites were analyzed separately for KEGG pathway enrichment using the MBRole 2.0 database. There are 6 enrichment pathways between the Model group and the Normal group, and 4 differential metabolite enrichment pathways between the XSJH group and the Model group, and there are 3 overlapping pathways between them, which were metabolic pathway, phenylalanine, tyrosine and tryptophan biosynthesis pathway, and tyrosine metabolism pathway (Figure 4D). The details of the KEGG pathway enrichment results are listed in Supplementary Tables S6, S7. Among them, tyrosine metabolism pathway is part of the phenylalanine, tyrosine and tryptophan biosynthesis pathway, suggesting that the tyrosine metabolism pathway may be the key pathway for the treatment of MetS by XSJ.

Based on the altered metabolites in tyrosine metabolism, the concentration of noradrenaline (NA) in the colonic contents of rats in the Model group was significantly decreased compared with that of rats in the Normal group, whereas XSJH was able to restore the concentration of NA. So, we examined the concentration of NA in rat serum and found that serum NA concentration in Model rats was significantly decreased compared with rats in the Normal group, whereas XSJ was able to increase serum NA concentration after administration of medium and high doses of XSJ (Figure 4E, p < 0.05). The results showed that the expression of NA synthesis and effector proteins (DBH, ADRB3, PKA), thermogenic proteins (UCP1, PGC1α, PRDM16), OXPHOS proteins (NDUFB8, SDHB, UQCRC2, MTCO2, and ATP5A1) in BAT in the Model group was decreased, and XSJ could upregulated the expression of the above-mentioned proteins (Figures 4F,G, p < 0.05). Although the concentration of NA levels rose, the blood pressure in XSJ group rats was lower than that in Model group rats (Supplementary Figure S2, p < 0.05). Combined with previous results that XSJ promotes BAT lipolysis and β-oxidation, our study suggests that XSJ accelerates NA synthesis, and therefore the consumption of excess glucose and lipids in the form of thermogenesis through OXPHOS to reduce lipid accumulation.

The rarefaction curve can reflect the diversity of microbial communities, the rank abundance curve can explain the species richness and community evenness, and PCoA (principal co-ordinates analysis) can evaluate the separate clustering of the gut microbiota from the different treatment groups. The rarefaction curves flattened, and the rank abundance curve had a larger horizontal axis range and a smoother curve indicating that the sequencing results of this experiment could truly reflect the bacterial abundance in each sample (Figures 5A,B). Based on the weighted unifrac method, the PCoA result showed that the Normal group was significantly separated from the Model group. However, the cluster distance of the XSJH group converged to the Normal group (Figure 5C).

The dominant phyla were mainly classified into Firmicutes, Proteobacteria, Bacteroidetes, unidentified Bacteria and Actinobacteria. XSJH reduced the abundance of Actinobacteria (Figures 5D,E). And the dominant genera were mainly categorized into Lactobacillus, Allobaculum, Romboutsia, Blautia and unidentified Ruminococcaceae. XSJH upregulated the abundance of Oscillibacter and downregulated the abundance of Bilophila, Candidatus Stoquefichus, Holdemania, Parasutterella and Rothia (Figures 5F,G). The above results suggested that XSJ improved the composition of the gut microbiota in MetS. However, the result of LEfSe (LDA effect size) analysis of the Tax4fun2 pathway prediction also showed that the gut microbiota after XSJ administration was associated with upregulated expression of amino acid metabolism pathway, phenylalanine, tyrosine and tryptophan biosynthesis pathway, and PPAR signaling pathway (Supplementary Figure S3). In terms of functional prediction, as shown in Figure 5H, the microbiota gene of Model rats involved osmoprotectant transport system substrate-binding protein (opuC), osmoprotectant transport system permease protein (opuBD) and glycine betaine transporter (betL) were upregulated (p < 0.05), suggesting an altered intestinal osmolality in the Model rats. XSJ restored the expression of osmoprotectant transport gene.

Intestinal barrier leakage may occur with an increase in the abundance of colitis-associated microbiota. As shown in Figures 5I–L, the Model group rats had significantly higher LPS, TNF-α, MCP-1 and IL-6, compared to the rats in the Normal group (p < 0.05). Through H&E staining analysis, the mucosal intrinsic glands in the colon villi in the Model group were loosely arranged with atrophy, while the intrinsic glands in the colon tissue of the XSJ group were intact or with mild atrophy (Figure 5M). Colon intrinsic glands atrophy indicated that intestinal barrier leakage happened in Model rats. So, we used transmission electron microscopy (TEM) to observe the colonic epithelial tight junctions (TJs). Compared to the Normal group, the TJs of the colonic epithelial cells in the Model group were disrupted and the cellular gaps were enlarged, whereas XSJH was able to restore cellular TJs (Figure 5N). The expression of the TJ protein ZO-1, Occludin and Claudin1 was significantly decreased in the Model group, whereas the expression of TJ proteins was elevated after XSJH administration (Figures 5O, P, p < 0.05). The above results demonstrate that XSJ can protect the intestinal barrier to reduce low-grade inflammation due to LPS leakage.

Additionally, Bilophila, Candidatus Stoquefichus, Holdemania, Parasutterella and Rothia showed a negative correlation with most metabolites, including NA (Figure 5Q). In contrast, Oscillibacter had a positive correlation with NA. LPS, TNF-α, MCP-1 and IL-6 were positively correlated with Candidatus Stoquefichus and Holdemania. So, through our analysis, it is suggested that high-fructose, high-fat diet may lead to a disruption of the intestinal environment and an increase in LPS levels, resulting in intestinal barrier leakage and low-grade inflammation.