Ramulus Mori (Sangzhi) alkaloids reduce obesity caused by high-fat diet in rats by changing gut bacteria and bile acid processing

SZ-A powder (J202302021) is characterized by a total polyhydroxy alkaloid content of 54.50%, with specific constituents including 42.66% 1-deoxynojirimycin (DNJ), 7.40% fagomine (FAG), and 4.44% 1,4-dideoxy-1,4-imino-D-arabinitol (DAB). The remaining components are mostly polysaccharides and amino acids (17). This product was graciously provided by Beijing Wuhe Boao Pharmaceutical Co., Ltd. (Beijing, China). Orlistat capsules (Hangzhou Zhongmei East China Pharmaceutical Co., Ltd.) were used. Carboxymethyl cellulose sodium (CMC-Na) (A05925). High-fat feed (MD12032, comprising 45% fat, 20% protein, and 35% carbohydrates) was obtained from Jiangsu Medison Biopharmaceutical Co., Ltd.

The Rat Insulin Competitive ELISA Kit (EK3220-96), Rat Leptin ELISA Kit (EK397-96), Rat IL-1β ELISA Kit (EK301B/3-96), Rat IL-6 ELISA Kit (EK306/3-96), and Rat TNF-A ELISA Kit (EK382/3-96) were obtained from Lianke Biotechnology Co., Ltd. The Rat Glucagon-Like Peptide 1 (GLP-1) ELISA Kit (E-EL-R3007) was obtained from Elabscience Biotechnology Co., Ltd. FXR (25055-1-AP) antibody was purchased from Wuhan Sanying Biotechnology Co., Ltd.; FGF15 (GB114817) antibody; ZO-1 (GB111402) antibody; and occludin (GB111401) antibody. The antibody against claudin-1 (GB112543) was obtained from Wuhan Servicebio Technology Co., Ltd., and the CYP7A1 (A10615) antibody and the TGR5 (A20778) antibody were obtained from ABclonal Biotechnology Co., Ltd.

Specific pathogen-free (SPF)-grade male Sprague-Dawley (SD) rats, aged 6 weeks and weighing 190–220 g, were used. The research protocol was sanctioned by the Animal Welfare Committee at the Henan Academy of Traditional Chinese Medicine (HNTCMDW-202303011). Housing conditions were maintained at a temperature of 22 ± 2°C, humidity of 45 ± 5%, and light/dark cycle of 12 h/12 h, with unrestricted availability of food and water. Following a 7-d adaptation period, the rats were assigned at random to 2 distinct groups: the normal control group (NCD) and the HFD control group (HFD). Control group rats received a standard diet, whereas HFD control group rats were given HFD for a period of 12 weeks (considered successful modelling if the BW exceeded the normal rat weight by more than 20%). After successful model establishment (rats that did not successfully model were excluded), the successfully modelled rats were divided into 4 groups; HFD control group (Model), SZ-A high-dose group (SZ-A400), SZ-A low-dose group (SZ-A300), and orlistat capsule group (orlistat).

Subsequently, the rats in the control and model groups were gavaged with physiological saline, while those in the SZ-A high-dose group received SZ-A at a dosage of 400 mg/kg/day. Those in the SZ-A low-dose group received SZ-A at a dosage of 300 mg/kg/day. The orlistat group was gavaged with orlistat dissolved in 0.5% CMC-Na at a dosage of 30 mg/kg/day once daily for a continuous period of 6 weeks. During the experimental period, the BWs of the rats were measured weekly. At the conclusion of the study, fecal samples from the rats were collected under aseptic conditions and preserved at a temperature of -80°C.

After collection, serum was extracted from the blood samples via centrifugation and was then utilized for the quantitative determination of total cholesterol (TC), triglyceride (TG), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), aspartate aminotransferase (AST), and alanine aminotransferase (ALT) levels using an automated biochemical analyzer.

The rat livers were fixed in 10% formalin solution for a period of 48 h, followed by embedding in paraffin. Subsequently, 5-μm-thick sections were prepared. These sections were treated with hematoxylin and eosin (H&E) staining and observed under a light microscope to evaluate the tissue structure. Additionally, rat liver tissue sections were rapidly frozen in liquid nitrogen. Oil Red O staining was conducted on these frozen sections to visualize lipid deposition within hepatocytes, and the stained slides were scrutinized under a microscope to evaluate any pathological alterations indicative of lipid accumulation.

The levels of LPS, INS, and GLP-1 in rat serum, as well as the expression levels of the IL-1β, IL-6, and TNF-α in rat small intestine tissue, were determined via ELISA.

The tissue was ground thoroughly and centrifuged to obtain the supernatant; 100 μl of alternative chloroform was added; and the mixture was mixed well. After centrifugation, the supernatant was mixed with isopropanol and allowed to stand; the liquid was removed; ethanol was added; and the mixture was inverted for washing. This process was repeated once, and the liquid was completely removed. The mixture was placed on a super clean bench and blown for 3–5 min, and an RNA dissolution solution was added to dissolve the RNA. Complete reverse transcription was performed on a regular PCR machine using the following amplification protocol. The RT-qPCR results were analyzed using a Bio-Rad CFX Connect (Bio-Rad, USA) for fluorescence quantification. The mRNA levels were standardized against GAPDH levels and determined using the ΔΔCT technique (shown in Table 1).

After the tissue was washed with PBS, it was homogenized and lysed. The supernatant was collected, and electrophoresis and membrane transfer were subsequently performed. The membrane was subjected to overnight incubation with the primary antibody, followed by incubation with the secondary antibody at room temperature. After the unbound antibodies were removed, the PVDF membrane was placed in a chemiluminescence detector. After exposure, the original image was saved in TIFF format.

The paraffin sections were deparaffinized and cleaned, immersed in retrieval solution, and heated for antigen retrieval. After blocking with 3% BSA for 30 min, the sections underwent overnight incubation with the primary antibody. After rinsing, the sections were treated with the secondary antibody in darkness for 50 min. Subsequent to washing with PBS buffer, the sections were stained with DAPI, followed by a 10 min incubation period in the dark. API was used to visualize the cell nuclei. Images were obtained using a Zeiss LSM 780 confocal microscope.

The BA standard solution was prepared, and the samples were extracted after low-temperature grinding; then the supernatant was collected. The presence and amount of the target compounds in the samples were determined using LC-ESI-MS/MS (UHPLC-Qtrap). The default parameters in the AB SCIEX quantitative software OS were used for automatic recognition and integration of the ion fragments, with manual checks as an auxiliary. Standard curves for linear regression were constructed by plotting the mass spectrometry peak area of the analyte on the y-axis and the analyte concentration on the x-axis. Sample concentration determination: The peak area of the analyte from mass spectrometry was input into the linear regression equation to determine the concentration.

The sample was mixed with 200 μL of precooled 75% methanol, ground for 3 min with a steel ball, vortexed, and extracted on ice via ultrasonication for 3 min. Then, 30 μL of methanol and 600 μL of methyl tert-butyl ether were added, and the mixture was shaken for 30 min. Afterwards, 160 μL of water was added, the mixture was mixed well and allowed to stand for 30 min, and the mixture was centrifuged. After centrifugation, the upper layer was transferred to dryness with nitrogen, after which the sample was reconstituted with 500 μL of a mixture of acetonitrile-isopropanol (CH3CN-IPA, 1:1, V/V) and centrifuged, and the supernatant was collected (the AB SCIEX corporation's state-of-the-art ultrahigh-performance liquid chromatography hyphenated to a triple quadrupole time-of-flight mass spectrometer, known as the UPLC-TripleTOF system).

After DNA extraction from the samples, the genomic DNA extracted was assessed. Fragmentation was performed to an approximate size of 300 bp using a Covaris M220. A paired-end (PE) library was constructed, followed by bridge PCR amplification. Sequencing was then completed on an Illumina HiSeq platform. The data analysis began with the raw sequences obtained from the sequencer. Initially, the raw sequences were subjected to optimization processes, including splitting, quality trimming, and contamination removal. Subsequently, the optimized sequences were used for contig assembly and gene prediction.

All the data are presented as the means ± standard deviations (x ± s) and were analyzed using GraphPad Prism 9.5 (San Diego, CA, USA). For multiple group comparisons, one-way analysis of variance and Fisher's least significant difference post hoc test were employed. For comparisons between two groups, an unpaired two-tailed Student's t-test was used. Statistical significance was determined at p < 0.05 and p < 0.01.

After 1 week of acclimatized feeding, the rats were divided into two groups and fed HFD or a normal diet for 12 weeks, after which SZ-A was administered by gavage (Figure 1A). Starting from the second week, there was a notable rise in the BW of the rats on a high-fat diet as opposed to those on a regular diet (P < 0.01). After 12 weeks, the BW of the rats fed by HFD was 692.28 ± 37.5 g, while that of the rats fed a normal diet was 537.38 ± 68.6 g, with the HFD group being more than 20% heavier on average than the normal diet group, indicating a successful modelling of obesity in the rats (Figure 1B). Intervention began in the 12th week, and after 6 weeks of intervention, there was a significant reduction in BW among rats treated with either SZ-A or orlistat capsules in the 18th week (P < 0.05). This indicates that SZ-A can effectively suppress weight gain in obese rats (Figure 1C).

Consistent with other studies, we found that SZ-A can effectively ameliorate lipid metabolism disorders and liver function abnormalities. After treatment with SZ-A, the serum levels of TC, LDL, ALT, and AST were notably reduced (P < 0.05) (Figures 1D–I). This indicates that SZ-A contributes to a decrease in blood lipid levels and the amelioration of hepatic injury in obese rats.

In the control group, the rat livers exhibited a characteristic reddish-brown hue without surface macules. In contrast, livers from the model group were moderately enlarged, had a soft texture, and presented a yellowish tint with distinct mottling and a greasy texture. The liver appearance of rats treated with SZ-A or orlistat capsules was closer to normal, with minimal yellowing and maculation. H&E staining revealed distinct differences in liver morphology. Livers from the control group displayed well-defined lobular structures with hepatocytes free from fatty degeneration, the absence of discernible fat globules, and an orderly arrangement of cellular nuclei. Conversely, the model group exhibited pronounced hepatocellular ballooning degeneration, intracellular fat vacuoles of varying dimensions, perturbed nuclear organization, and inflammatory necrosis.

Following treatment with SZ-A at both low and high doses or with orlistat capsules, there was a marked reduction in hepatocellular steatosis and ballooning degeneration, as well as a decrease in inflammatory necrotic areas, compared to those in the model group. Staining with Oil Red O exposed a notably increased presence of red-colored lipid droplets in the hepatic tissues of the model group rats, signifying severe fatty liver changes. Treatment with SZ-A or orlistat capsules noticeably ameliorated lipid droplet accumulation. Collectively, these findings suggest that SZ-A effectively mitigates pathological hepatic lipid deposition, inflammation, and fibrotic changes in obese rats (Figure 1J).

Subsequently, we evaluated the protective effects of SZ-A on fat accumulation. HFD feeding caused a notable enlargement of adipocytes in the epididymal white adipose tissue (epiWAT) compared to the normal group. Treatment with SZ-A at both low and high doses, as well as orlistat capsules, significantly reduced the adipocyte size in epiWAT compared to the model group. These findings further substantiate that the addition of SZ-A markedly improves fat accumulation in rats subjected to an HFD (Figure 1K).

Intestinal bacteria have a significant association with obesity and imbalances in lipid metabolism; therefore, this study conducted metagenomic sequencing of rat feces to compare the composition of the microbial community in the feces of rats in the control group, model group, and SZ-A. We utilized the abundance of gut bacteria at the operational taxonomic unit level for principal coordinate analysis (PCoA), and the findings highlighted considerable variations in the gut microbial communities across the three rat groups (P = 0.001) (Figure 2A). Nonmetric multidimensional scaling analysis indicated that among the three comparisons, the stress value was 0.076 < 0.1, suggesting good ordination, with samples within the same group being closer together, indicating good repeatability within groups and a higher degree of dispersion between different groups. The R value of 0.96, nearing 1, signified greater disparities between groups compared with within-group variations, and P < 0.05 (Figure 2B).

Based on the NR species annotation results, the dominant species in the rat communities at the phylum level were analyzed for different groups. The community bar chart shows the changes in the composition of the dominant species (Figure 2C), with the dominant bacterial groups in rat feces being Firmicutes, Bacteroidota, and Actinobacteriota. In contrast to the control group, the model group exhibited a markedly reduced presence of Proteobacteria in their fecal bacteria (P < 0.01), whereas the SZ-A group showed a significant increase in both Actinobacteria and Bacteroidota (P < 0.01, P < 0.001), and the relative abundance of Firmicutes was obviously lower (P < 0.001). As opposed to the model group, the SZ-A group's fecal microbiota showed a notably higher relative abundance of Actinobacteria and Proteobacteria (P < 0.05). Additionally, there was a notable rise in the relative abundance of Bacteroidota (P < 0.001) and a corresponding significant decrease in Firmicutes (P < 0.001). At the phylum level, compared with those in the control group, the ratio of Firmicutes to Bacteroidota in the model group was severely disrupted, with a notable rise in the abundance of Firmicutes and a corresponding decline in the abundance of Bacteroidota. SZ-A is capable of modulating the proportion of these bacterial groups to resemble the control group's community composition. The Firmicutes to Bacteroidota ratio serves as an indicator of the gut microbiota's equilibrium, with a reduced Firmicutes/Bacteroidota ratio being directly associated with weight reduction (Figures 2D–G). Our subsequent analyses at the genus and species levels revealed significant changes in the gut microbiota among the groups. At the genus level, in contrast to the control group, the relative abundance of Blautia increased, while that of Ligilactobacillus significantly decreased in the model group. The SZ-A group exhibited a notably higher relative abundance of Bacteroides and a considerably lower relative abundance of Blautia, in contrast to the model group. At the genus level, the abundances of Bifidobacterium animalis, Ligilactobacillus murinus, and Muribaculaceae bacterium increased in the SZ-A group, while the abundances of Firmicutes bacterium and Lachnospiraceae bacterium decreased. These results indicate that SZ-A directly influences the modulation of gut microbiota composition in rats with HFD (Figures 2H–I).

Using the LEfSe (LDA>4.0) method, statistical analysis of the relative abundance of gut microbiota of rats in each group was performed to identify group-specific and representative differential species, resulting in an evolutionary branch diagram (Figure 3A) and a statistically significant biomarker abundance diagram (Figure 3B). There were 14 trees in the control group, 15 in the model group, and 19 in the SZ-A group, and these branches had different abundances and were considered characteristic bacterial phenotypes. The result indicated a notable change in microbial abundance under the intervention of HFD and SZ-A. Among them, Oscillospiraceae (Firmicutes) was a characteristic genus in the control group. In the model group, Blautia (genus) and Firmicutes (phylum) were characteristic genera. In the SZ-A group, Bacteroides (genus), Bacteroidaceae (Firmicutes), Bifidobacteriaceae (Firmicutes), Bifidobacterium (genus), and Bifidobacteriales (order) were characteristic genera.

Spearman correlation analysis assessed the relationship among the predominant 20 bacterial genera and the rats' blood lipid levels and hepatic function indices. Blautia and Dorea showed positive associations with blood lipids and liver health biomarkers, whereas Ligilactobacillus exhibited an inverse relationship with these parameters. The abundance of Bifidobacterium and Parabacteroides was positively associated with TG levels. The abundance of Bacteroides was negatively associated with AST levels, and that of Ruminococcus had negative correlations with TG levels, TC levels, and LDL levels. Spearman correlation analysis indicated that these bacteria were closely related to parameters associated with hyperlipidemia (Figure 3C).

We also conducted functional analysis using the KEGG database to predict the functional pathways associated with gut microbiota. The control group's gut microbiota displayed enhanced capabilities in genetic information processing, cellular activities, human diseases, and organismal systems, while that of the model group showed an increase in metabolism and environmental information processing. The SZ-A group experienced a notably reduced capacity for environmental information processing relative to the model group (Figures 3D, E). To screen for differential metabolic functional pathways, we performed LEfSe analysis of the KEGG metabolic pathways. The results indicated that functional name, mismatch repair, ribosome, flagellar assembly, purine metabolism, aminoacyl-tRNA biosynthesis, homologous recombination, teichoic acid biosynthesis (phospholipid biosynthesis), and quorum sensing were significantly enriched in the control group. Peptidoglycan biosynthesis, porphyrin metabolism, two-component system, metabolic pathways, ABC transporters, biosynthesis of amino acids, 2-oxocarboxylic acid metabolism, and histidine metabolism were notably enriched in the model group. Sphingolipid metabolism, lysosome, amino sugar and nucleotide sugar metabolism, biosynthesis of nucleotide sugars, galactose metabolism, other glycan degradation, O-antigen nucleotide sugar biosynthesis, and biosynthesis of secondary metabolites were significantly enriched in the SZ-A group (Figure 3F).

Based on the KEGG annotation results, we conducted a differential significance analysis of enzymes involved in the secondary BA pathway. The KEGG gene enrichment analysis indicated that, compared with those in the model group, 7α-hydroxysteroid dehydrogenase (7α-HSDH) and choloylglycine hydrolase were significantly enriched in both the control group and the SZ-A group. 7α-HSDH plays an important role in bile acid (BA) metabolism, efficiently converting taurochenodeoxycholic acid (TCDCA) into tauroursodeoxycholic acid (TUDCA). Choloylglycine hydrolase, also known as bile salt hydrolase or conjugated BA hydrolase, is an enzyme that plays an important role in BA metabolism. Conversely, the concentration of 3α-hydroxycholanate dehydrogenase(3α-HSDH) was markedly reduced in both the control and SZ-A groups as compared with the model group. This finding suggested that HFD-induced obesity can affect the composition of gut microbiota through the regulation of BAs (Figures 4A, B).

For a comprehensive grasp of SZ-A's impact on lipid and metabolic processes, we used lipidomics methods to detect differences and changes in lipids in rat adipose tissue. We identified 607 lipid metabolic products from rat adipose tissue and performed PLS-DA on the samples. Under the positive and negative modes, there was a clear distinction in lipid grouping among the control group, model group, and SZ-A, indicating notable differences in fat lipid metabolism among the groups (Figures 5A, B).

Student's t-test was applied to analyze lipid metabolites, and differentially abundant metabolites were screened with the criteria of a fold change (FC) ≥1 and P < 0.05. Overall, the comparison between the control group and the model group revealed a total of 387 distinct metabolic compounds, with 286 experiencing an increase in abundance and 101 showing a decrease. The main changes were the upregulation of triglycerides (TGs) and diglycerides (DGs), as well as the downregulation of phosphatidylcholine (PC) and phosphatidylethanolamine (PE). Between the model group and SZ-A, 258 differentially abundant metabolites were detected, with 72 upregulated and 186 downregulated. This indicates that SZ-A is involved in modulating lipid metabolism in rats with obesity (Figures 5C, D).

Next, we utilized OPLS-DA model analysis to obtain the top 30 VIP scores, among which TG (18:2e/6:0/9:0), TG (18:2/11:3/18:2), TG (4:0/15:0/16:0), TG (4:0/18:2/22:6), and TG (18:1/11:4/20:4) were the top 5 lipid components contributing to the separation between the model group and SZ-A. These are lipid subclasses within the glycerolipid (GL) category and play a key role in improving lipid metabolism disorders in obese rats (Figures 5E, F).

For KEGG enrichment analysis of differentially enriched adipose tissue between the model group and the SZ-A group, we selected the top 13 metabolic function categories with significant relevance, which were mainly enriched in the following 13 pathways: choline metabolism in cancer, insulin resistance, neurotrophin signaling mechanism, Kaposi's sarcoma-associated herpesvirus infection, glycerophospholipid metabolism, retrograde endocannabinoid signaling, adipocytokine signaling pathway, AGE-RAGE signaling mechanism in diabetic complications, fat digestion and absorption, regulation of lipolysis in adipocytes, sphingolipid signaling mechanism, thermogenesis, and diabetic cardiomyopathy (Figure 5G).

We utilized BA-targeted metabolomics to measure the BA content in rat feces. The results showed that the BA metabolic profiles of the model group and the SZ-A group were significantly different, with a greater degree of separation between the two groups, indicating a notable classification effect and significant differences (Figures 6A, B). Student's t-test was applied to analyze BAs in rat feces, and differentially abundant metabolites were screened using the criteria of FC ≥ 1.5, P < 0.05, and a variable importance (VIP) > 1.0 in the OPLS-DA model between the two groups (Figures 6C, D). The levels of hyodeoxycholic acid (HDCA), deoxycholic acid (DCA), 12-ketolithocholic acid (12-KLCA), lithocholic acid (LCA), and murideoxycholic acid (MDCA) significantly decreased in the SZ-A group (Figures 6E, F).

Spearman's correlation analysis was employed to assess the relationship between the 20 predominant bacterial genera and the 20 key BA metabolite genera identified in rat fecal samples. The study revealed that murideoxycholic acid, deoxycholic acid, ursodeoxycholic acid, lithocholic acid, hyodeoxycholic acid, and isolithocholic acid were negatively correlated with BAs, such as Bacteroides. Ursodeoxycholic acid (3β-ursodeoxycholic acid), 7-ketodeoxycholic acid, 3-dehydrocholic acid, beta-muricholic acid, alpha-muricholic acid, and ursocholic acid were positively correlated with Bifidobacterium and Blautia (Figure 6G).

In contrast to the control group, the model group exhibited a marked increase in the levels of FXR and FGF15 proteins within the rat small intestine (P < 0.01), while the CYP7A1 content was notably decreased (P < 0.01). This indicates that obesity triggers the activation of the FXR/FGF15 signaling pathway in the rat small intestine, which coincides with a pronounced decrease in hepatic CYP7A1 levels (P < 0.01). Relative to the model group, the small intestinal FXR and FGF15 levels in the rats in both the low- and high-dose SZ-A groups, as well as those in the orlistat group, were notably lower (P < 0.05, P < 0.01). Additionally, the protein expression of CYP7A1 was substantially upregulated (P < 0.01). These results indicate that SZ-A can effectively counteract obesity-induced activation of the FXR/FGF15 signaling pathway in rats and suppress CYP7A1 expression.

Relative to the control group, the TGR5 protein content in the small intestine of rats in the model group was notably reduced (P < 0.01), suggesting that SZ-A can notably suppress the activation of TGR5 in the rat small intestine caused by obesity and inhibit the expression of TGR5 (Figures 7A–F).

Relative to the control group, the LEP levels in the model group were greater (P < 0.001). Relative to the model group, both the high-dose SZ-A group and the orlistat group showed a reduction in serum LEP levels, reflecting an increase in leptin sensitivity (P < 0.001). We further measured serum insulin levels and GLP-1 activity levels as indicators of insulin and GLP-1 secretion, respectively. Compared to those in the control group, the model group rats had increased insulin levels (P < 0.001). Relative to the model group, both the high- and low-dose SZ-A groups and the orlistat group showed a decrease in serum insulin levels, reflecting an increase in insulin sensitivity (P < 0.001). Relative to the control group, GLP-1 levels were lower in the model group (P < 0.05). Relative to the model group, both the high- and low-dose SZ-A groups and the orlistat group had increased serum GLP-1 levels (P < 0.01) (Figures 7G–I).

To further investigate the impact of SZ-A on intestinal inflammation in obese rats, we measured the levels of inflammatory factors in the colonic tissues. Relative to the model group, the levels of IL-1β, IL-6, and TNF-α in the colonic tissues of rats in the low- and high-dose SZ-A groups and the orlistat group were notably reduced (P < 0.01); therefore, SZ-A can alleviate intestinal inflammation in obese rats (Figures 8A–C).

Under normal physiological conditions, an intact intestinal mucosal barrier can prevent the permeation of bacteria and their mediators into the systemic circulation. Epithelial tight junctions (TJs) are composed of ZO-1, occludin and claudin-1 among others. These tight junction complexes effectively seal intercellular spaces, preventing the invasion of pathogens, endotoxins, and bacteria. Relative to the model group, the protein and mRNA expression of ZO-1 in the SZ-A group increased (Figures 8D–F). Our research results indicate that compared to the model group, SZ-A can regulate the expression of ZO-1, occludin, and claudin-1 staining in TJs, thereby maintaining the integrity of the intestinal barrier (Figure 8G).