Guanxinning Tablet Attenuates Coronary Atherosclerosis via Regulating the Gut Microbiota and Their Metabolites in Tibetan Minipigs Induced by a High-Fat Diet

Coronary atherosclerosis (CA) is a chronic and evolving inflammatory disease characterized by the build-up of atherosclerotic plaque in the wall of coronary arteries. Guanxinning tablet (GXNT) is a novel Chinese medicine formula, which has been clinically used to treat coronary heart disease for many years. However, the potential mechanism for treating CA remains unclear. Thus, the study was aimed at investigating the therapeutic effect of GXNT on CA and further explore the underlying mechanisms from the perspective of gut microbiota. Following the establishment of a CA model in Tibetan minipigs, GXNT was orally administrated. We simultaneously detected blood lipid levels, observed ventricular function using ultrasound examination, measured platelet aggregation, and checked changes in inflammatory factors, oxidative stress factors, and vascular endothelial injury-related indexes applying ELISA assays. Histopathological changes of coronary artery tissue were subsequently evaluated using Sudan IV staining, HE staining, Oil red “O” staining, and immunohistochemistry assays. Finally, alterations of the gut microbiota and microbial metabolites were detected using metagenomic sequencing and targeted metabolomics, respectively. The results have suggested that GXNT could regulate dyslipidemia, improve heart function, and inhibit the levels of ox-LDL, CRP, TNF-α, IL-1β, SOD, MDA, vWF, and ET-1, as well as platelet aggregation. Additionally, histopathological findings revealed that GXNT could reduce lipid deposition, alleviate AS lesions, and restrain the expressions of NF-κB, TNF-α, and MMP-9. Furthermore, the composition of the gut microbiota was altered. Specifically, GXNT could upregulate the relative abundance of Prevotellaceae and Prevotella and downregulate the abundance of Proteobacteria, Enterobacteriaceae, and Escherichia. As for microbial metabolites, GXNT could increase fecal propionic acid, butyric acid, and LCA-3S and decrease fecal TMA-related metabolites, CDCA, and serum TMAO. In sum, the results showed that GXNT had a satisfactory anti-CA effect, and the mechanism was closely associated with modulating gut microbiota and related metabolites.


Introduction
Coronary atherosclerosis (CA) continues to be a leading cause of illness and death worldwide, which refers to a chronic inflammatory pathological change in the wall of coronary arteries [1]. It causes vascular lumen stenosis or thrombosis, resulting in angina pectoris, myocardial infarction, and even sudden death. Atherosclerosis (AS), as the pathological basis of CA, is characterized by progressive lipid deposition, inflammatory cell infiltration, and fibrosis [2]. The occurrence and progression of CA are attributable to genetic and environmental factors, such as dyslipidemia, metabolic syndrome, and air pollution, whereas the mechanisms are still noncompletely understood.
Recently, emerging studies reveal that gut microbes are likely to be one of the important pathogenic factors for CA. Gut microbiota participates in host nutrient absorption and metabolism. It can not only affect glucose and lipid metabolism by directly regulating energy intake in food but also affect the circulating endotoxin level and inflammatory response by indirectly changing intestinal permeability, thereby facilitating CA development [3]. Liu et al. have found that the homeostasis of the gut microbiome in CA patients was disrupted, accompanied by significant changes in the composition of microbes and metabolites [4]. Once the integrity of the gut barrier is broken, the gut microbiota and metabolites are more likely to enter the blood circulation, followed by chronic systemic inflammatory responses and AS formation. Accumulating evidence suggests that the microbiota-dependent metabolite trimethylamine-Noxide (TMAO) promotes AS formation, and diet-induced inflammation will be alleviated if a TMAO inhibitor is applied [5]. In addition, it is reported that the proportion of Bacteroidetes and Firmicutes changed in CA patients, as well as gene functions, including lipopolysaccharide biosynthesis, fermentative capacity, and propanoate metabolism [6]. These findings hint that differences in the structure and function of the gut microbiota may affect CA.
Guanxinning tablet (GXNT) is a new preparation of Chinese medicine formula, developed from GuanXinNing injection by changing the route of medication. GXNT has a unique therapeutic effect on cardiovascular diseases, especially coronary heart disease (CHD), which is reported to be a safe and effective treatment for stable angina pectoris patients [7]. GXNT consists of two traditional Chinese medicines, Salvia miltiorrhiza Bunge. and Ligusticum chuanxiong Hort. in a 1 : 1 ratio, possessing the effects of activating blood circulation, removing stasis, dredging arteries, and nourishing the heart [8]. Our previous studies have found that GXNT contains a variety of phenolic acid active components, and salvianolic acid B in Salvia miltiorrhiza Bunge. and ferulic acid in Ligusticum chuanxiong Hort. are used as the quality control specification. Despite these active ingredients having low oral bioavailability [9,10], they can still exert efficacy, so we infer that the mechanism may be related to the gut microbiota. The Tibetan minipig CA model induced by a high-fat (HF) diet is an ideal model for simulating the formation of human AS plaques. We preliminarily found that GXNT could improve HF diet-induced chronic low-grade inflammation and inhibit AS formation in minipigs, suggesting that GXNT has an anti-AS effect [11]. From this, our study observed the efficacy of GXNT on long-term HF diet-induced CA in Tibetan minipigs and explored the potential mechanism by which GXNT treats CA based on the gut microbiota.  Figure 2: Effects of GXNT on blood lipids of the CA model. Changes of (a) TG, (b) TC, (c) LDL-C, and (d) HDL-C levels and (e) AI during 12 weeks of GXNT administration. n = 6 in each group. Data are presented as the means ± SE. # P < 0:05 or ## P < 0:01 vs. the control group; * P < 0:05 or * * P < 0:01 vs. the model group.
University (Hangzhou, China) under the conditions of temperature: 22 ± 1°C, relative humidity: 40%~65%, and 12 h/ 12 h light and dark cycle. They were fed two meals a day and had free access to water. All animal experiment procedures in this study were authorized by the Committee on the Management and Use of Laboratory Animals (IACUC     artery tissue with a thickness of 8 μm were fixed in neutral formaldehyde for 15 min, stained with newly prepared and filtered oil red "O" solvent (Sigma, USA) for 15 min, differentiated with 60% isopropanol for seconds, washed with running water, immersed with hematoxylin for 5 min, washed again with water, and finally mounted. Then, pathological sections were scanned with a 2.0 RS Nana Zoomer digital slide scanner (Hamamatsu, Japan), and the lipid deposition was analyzed using Image-Pro Plus 6.0 software.  (Proteintech, USA) at 4°C overnight, respectively. After washing, it was incubated with the secondary antibody for 1 h at room temperature, colored with DAB (ASGB-BIO, China), counterstained with hematoxylin, dehydrated, and mounted. The primary antibody was replaced with PBS as a negative control. Brownish-yellow or light yellow represented the positive substance. The pathological sections were scanned with a 2.0 RS Nana Zoomer digital slide scanner (Hamamatsu, Japan), and the positive expression area and total area in the field of vision were analyzed with Image-Pro Plus 6.0 software. The positive expression rate was calculated with the formula of positive expression area/total area × 100%.

Fecal Metagenomic Analysis for Microbiome and Function
2.8.1. Fecal DNA Extraction and Sequencing. DNA was extracted from fecal samples by the CTAB method as previously described [12]. Briefly, each 200 mg feces were added to the CTAB lysate, incubated at 65°C, reversed to fully dissociate the sample, and centrifuged. Afterward, the supernatant was added with mixtures of phenol (PH = 8:0), chloroform, and isoamyl alcohol (v/v/v, 25 : 24 : 1), inverted, and centrifuged (12000 rpm, 10 min). The supernatant was added with a mixture of chloroform and isoamyl alcohol (v/v, 24 : 1), inverted thoroughly, and centrifuged (12000 rpm, 10 min). Then, the supernatant was again added with isopropanol, inverted several times, precipitated at -20°C, and centrifuged (12000 rpm, 10 min). After discarding the supernatant, the DNA pellets were washed with 1 mL 75% ethanol twice, air-dried on the benchtop or at room temperature, dissolved with ddH 2 O, and incubated (55-60°C, 10 min) if necessary. In addition, RNA could be removed from the DNA samples using 1 μL RNase A (37°C, 15 min). DNA integrity and purity were separately verified by 1% agarose gel electrophoresis and NanoDrop spectrophotometry (NanoDrop, Germany), and the concentrations were quantified using Qubit Fluorometric Quantification (Thermo Fisher, USA). According to the Illumina protocol of NEBNext Ultra DNA Library Prep Kit, a DNA library was constructed and high-throughput sequencing was conducted on the Illumina NovaSeq platform, following the manufacturer's guidelines.

Sequencing Data Analysis.
KneadData software was used for raw data quality control (based on Trimmomatic) and dehosted (based on Bowtie2) prior to acquiring valid sequences (clean data). And FastQC was used to test the rationality and effectiveness of quality control. Microbial databases were downloaded from Kraken2 (http://ccb.jhu .edu/software/kraken2/) to identify the species in the sample, and then, Bracken (http://ccb.jhu.edu/software/bracken/) was used to predict the relative abundance of species. Based on Bracken's absolute abundance and annotation information, β diversity analysis and construction of abundance profile were performed on the Tutools platform (https://www .cloudtutu.com). Using HUMAnN2 software (based on DIAMOND), the reads of each sample were mapped to the KEGG database (https://www.kegg.jp/). At last, the annotation information and relative abundance table of each functional database were obtained according to the corresponding relationship between the UniRef90 ID and the database. 7 Journal of Immunology Research 60 s, sonicated 10 min at room temperature, incubated 12 h at 4°C, and then centrifuged at 12,000 rpm for 20 min at 4°C. Subsequently, 20 μL fecal supernatant or serum was accurately pipetted, mixed with 10 μL internal standard and 750 μL 1% formic acid-acetonitrile solvent, vortexed for 30 s, and centrifuged at 12,000 rpm for 5 min at 4°C. The supernatant was filtered with a 0.22 μm filter and transferred to analysis.
2.9.2. Detection of Fecal SCFA Based on GC-MS. Each 50 mg sample was prepared, mixed with 50 μL 15% phosphoric acid, 100 μL 125 μg/mL internal standard solution, and 400 μL ether, homogenized for 1 min, and centrifuged at 12000 rpm for 10 min at 4°C.  Journal of Immunology Research GC-MS technology detects fecal SCFA profiles based on an Agilent HP-INNOWAX capillary column (30 m * 0.25 mm ID * 0.25 μm) with a split inlet, 1 μL of injection volume, and 10 : 1 split ratio. The inlet temperature was 250°C, the ion source temperature was 230°C, the transfer line temperature was 250°C, and the quadrupole temperature was 150°C. The programmed temperature was initially set to 90°C, heated to 120°C at 10°C/min, then heated to 150°C at 5°C/min, and finally heated to 250°C at 25°C/min for 2 min. The carrier gas was helium at a flow rate of 1.0 mL/ min. MS conditions are as follows: electron impact ionization (EI) source, SIM scan mode, and 70 eV electron energy.

Detection of Fecal BAs
Based on GC-MS. Each 50 mg sample was accurately weighed, extracted by 1000 μL methanol solvent (-20°C) by vortexing 60 s, vibrated using glass beads (50 Hz, 60 s), repeated the above steps at least 2 times, sonicated for 30 min at room temperature, centrifuged at 4°C (12000 rpm, 10 min), and finally filtered with a 0.22 μm filter.
2.10. Statistical Analysis. Data were processed using SPSS 24.0 software (SPSS Inc., Chicago, IL, USA). All measurement data were tested for normal distribution, expressed as mean ± standard error ( x ± SEM). T-test was used for data comparison between two groups, and one-way ANOVA was used to compare data between multiple groups. The statistical graphs were drawn with GraphPad Prism 8.0 software (GraphPad Software Inc., San Diego, CA, USA). All statistics were two-tailed tests, and a P value < 0.05 was considered statistically significant.

Effects of GXNT on General
Signs of the CA Model. During the experiment period, there were no obvious abnormalities of general signs, such as spirits, activity, food intake, feces, and urine of minipigs in the control group. However, after HF diet induction, pigs in the model group exhibited greasy coats, hair loss, obesity, laziness, and lethargy, which were notably improved after GXNT administration for 12 weeks (Figure 1(a)). In addition, habitus index ( Figure 1 Figure 2(e)) in the model group were much higher than those in the control group (P < 0:01). After GXNT administration, the LDL-C level significantly decreased at the 12 th week (P < 0:05), as well as the AI index at the 8 th and 12 th weeks (P < 0:05, P < 0:01). Meanwhile, GXNT could significantly increase HDL-C after 8 weeks of GXNT administration (P < 0:01).
3.4. Effects of GXNT on Serum SOD and MDA of the CA Model. SOD and MDA are important markers of oxidative stress. As shown in Figure 4, we found that the serum SOD activity (Figure 4(a)) in the model group significantly decreased compared to the control group (P < 0:01), while the MDA content (Figure 4(b)) significantly increased   13 Journal of Immunology Research (P < 0:01). Fortunately, GXNT could remarkably ameliorate these changes (P < 0:01).

Effects of GXNT on Serum ET-1 and vWF of the CA
Model. Both ET-1 and vWF can reflect endothelial injury. They were measured for observing the protective effect of GXNT on the endothelium. As shown in Figure 5, the serum levels of ET ( Figure 5(a)) and vWF ( Figure 5(b)) in the model group were much higher than those in the control group (P < 0:01). After GXNT treatment, the ET and vWF levels significantly decreased (P < 0:05, P < 0:01).

Effect of GXNT on Left Ventricular Structure and
Function of the CA Model. Patients with CA are often associated with reduced cardiac function; thus, the impact of GXNT on heart function was also evaluated. Compared with the control group, LVd mass ( Figure 6(a)), IVSd ( Figure 6(b)), and LVPWd ( Figure 6(c)) in the model group significantly increased (P < 0:05, P < 0:01), while EF ( Figure 6(d)) and FS ( Figure 6(e)) significantly decreased (P < 0:01). GXNT treatment could distinctly decrease LVd mass, IVSd, and LVPWd (P < 0:05) and increase EF and FS compared with the model group (P < 0:05).

Effects of GXNT on Lipid Deposition in Aortic
Vessels of the CA Model. The results of Sudan IV staining showed the red-stained lipid in the thoracic aorta and abdominal aorta in the model group (indicated by white arrows) was much more than that in the control group, which was greatly improved in the GXNT group (Figure 7(a)). Subsequently, quantitative analysis (Figure 7(b)) further showed that the lipid deposition in the thoracic aorta, abdominal aorta, and whole aorta in the model group markedly increased compared to the control group (P < 0:01). And GXNT administration could significantly reduce the lipid deposition (P < 0:01).

Effects of GXNT on the Pathological Morphology of
Coronary Artery Tissue of the CA Model. In the control group, the intima and media structures of coronary artery tissue were intact and clear, without obvious abnormalities. In the model group, the intima of the coronary artery had obvious hyperplasia, infiltration of inflammatory cells such as macrophages, increased foam cells, severe atherosclerotic plaques, fibrous cap formation, and serious vascular stenosis, which were attenuated by GXNT as shown in Figure 8(a). Similarly, quantitative analysis showed that IMT (Figure 8(b)), NIA (Figure 8(c)), NIA/MA (Figure 8(d)), and NIA/IELA (Figure 8(e)) in the model group significantly increased as compared to the control group (P < 0:01), which could be clearly decreased by GXNT treatment (P < 0:05, P < 0:01).

Effects of GXNT on Lipid Deposition of Coronary Artery
Tissue of the CA Model. The results of oil red "O" staining showed that only a small amount of lipid was deposited in coronary artery tissue in the control group, while extensive lipid deposition appeared in the model group, and the lipid was visually reduced in the GXNT group, as shown in Figure 9(a). Compared with the control group, the lipid deposition rate of the coronary artery in the model group significantly increased using quantitative assessment (P < 0:01) (Figure 9(b)). On the contrary, the lipid deposition rate was significantly decreased after GXNT administration (P < 0:05).
3.10. Effects of GXNT on the Protein Expressions of NF-κB, TNF-ɑ, and MMP-9 in Coronary Artery Tissue of the CA Model. The immunohistochemical staining results of NF-κB, TNF-α, and MMP-9 proteins in coronary vessels were shown in Figures 10(a)-10(c), separately. Compared with the control group, the apparent positive expressions of NF-κB, TNF-ɑ, and MMP-9 proteins in the intima and middle layers in the model group could be observed, which had a visual reduction in GXNT group. Furthermore, quantitative analysis showed that the positive expression rates of NF-κB ( Figure 10(d)), TNF-ɑ (Figure 10(e)), and MMP-9 ( Figure 10(f)) in the model group significantly increased as compared to the control group (P < 0:01). And the positive expression rates in the GXNT group were much less than those in the model group (P < 0:05, P < 0:01).

Effects of GXNT on the Composition of Gut Microbiota in the CA Model.
To observe the β diversity changes, we carried out methods of principal component analysis (PCA) based on the similarity coefficient matrix and principal coordinates analysis (PCoA) based on the Bray-Curtis distance matrix. Significant differences existed in the composition and structure between the control group and model group, which could be improved by GXNT, as shown in Figure 11(a).
The relative abundance of the gut microbiota was analyzed, and it was found that there were significant differences at the phylum, family, and genus levels. The top 10 species in relative abundance at the phylum level were selected to construct the abundance profile as shown in Figure 11(b). Proteobacteria, Bacteroidetes, and Firmicutes were the three dominant phyla in the control group, model group, and GXNT group. The relative abundance of Proteobacteria in the model group significantly increased as compared with the control group (P < 0:05) (Figure 11(c)). And GXNT administration could decrease its relative abundance (P < 0:05).
As to the levels of family and genus, the top 20 species were selected to construct abundance profiles (Figures 11(d) and 11(g), respectively). Enterobacteriaceae was the dominant bacteria shared by 3 groups. Compared with the control group, the relative abundance of Enterobacteriaceae in the model group significantly increased (P < 0:05) (Figure 11(e)), while the abundance of Prevotellaceae obviously decreased (P < 0:05) (Figure 11(f)). GXNT treatment could distinctly regulate their relative abundance (P < 0:05). At the level of the genus, Escherichia, belonging to the Enterobacteriaceae family, was the common dominant bacteria in the 3 groups. The relative abundance of Escherichia in the model group significantly increased as compared with the control group (P < 0:05) (Figure 11(h)), and the relative abundance of Prevotella decreased (P < 0:05) (Figure 11(i)). After GXNT administration, the relative 14 Journal of Immunology Research abundance of Escherichia apparently reduced (P < 0:05), and the relative abundance of Prevotella significantly increased (P < 0:05).

The Effects of GXNT on Gut Microbiota Function
Based on KEGG Analysis. The KEGG orthology database is the basic database for studying gene function. Using the KO database, genes that have the same function can be aggregated. Then, differences in metabolic pathways (Pathways) and functional modules (Modules) between groups can be analyzed based on the KO. As for KO, the gut microbiota genes associated with the functions of bile acid metabolism (K03455), choline metabolism (K00108), carnitine metabolism (K22443 and K22444), acetate kinase production (K00925), and lipopolysaccharide biosynthesis (K00748) in the model group significantly rose as compared to the control group (P < 0:05, P < 0:01), while those related to short-chain fatty acid transport (K02106), butyrate kinase production (K00929), and propionate kinase production (K00932) clearly decreased (P < 0:05, P < 0:01). All of these genes involved in the functions above were significantly regulated by GXNT (P < 0:05), as shown in Table 1.

Effects of GXNT on Metabolites of CA-Related Gut Microbiota
3.13.1. Effects of GXNT on Metabolite SCFAs in the CA Model. The contents of fecal SCFAs, including acetic acid, propionic acid, butyric acid, isobutyric acid, valeric acid, isovaleric acid, and caproic acid, were detected by the GC-MS method. The results in Figure 13 showed that acetic acid, propionic acid, and butyric acid were the main SCFAs, accounting for about 90%-95% of the total content. The propionic acid and butyric acid contents in the model group significantly decreased as compared with the control group (P < 0:05), which could be significantly increased by GXNT (P < 0:01).

Discussion
GXNT is clinically used for treating CHD and ischemic cardiomyopathy with significant curative effects [13]. Our previous pharmacological experiments confirmed that GXNT has the functions of reducing whole blood viscosity, improving blood rheology abnormalities, antimyocardial ischemia, antithrombosis, and vasodilation [14][15][16][17][18][19]. In this study, we established a CA model induced by a HF diet in Tibetan minipigs and simultaneously interfered with GXNT. It was found that GXNT could regulate blood lipids, inhibit platelet aggregation, improve cardiac dysfunction, reduce vascular lipid deposition, inhibit vascular inflammation, protect vascular endothelium, and reduce AS lesions. Using   Figure 13: Effects of GXNT on fecal SCFAs in the CA model. n = 6 in each group. # P < 0:05 vs. the control group; * * P < 0:01 vs. the model group. Journal of Immunology Research metagenomic and targeted metabolomics analysis, we revealed that GXNT could regulate the composition of gut microbiota and related metabolites in the CA model, thereby preventing AS formation. Dyslipidemia is an independent risk factor for AS [20]. The relationship between diet, lipid level, and AS is well established, and atherosclerotic lesions in animals and humans appear to be associated with elevated cholesterol and excess fat intake [21]. Dyslipidemia can damage vascular endothelial cells and consequently stimulate vascular intimal hyperplasia. Endometrial injury is an initiating factor for AS, with injury markers ET-1 and vWF increasing [22]. vWF can activate platelets to release coagulation substances for participating in platelet aggregation, thrombosis, and AS plaque formation, thereby resulting in coronary stenosis, insufficient blood supply, and decreased cardiac function. As an important factor, oxidative stress can also damage vascular endothelial cells. The level of MDA is one of the indicators reflecting the degree of oxidative stress, and the activity of SOD reflects the ability to scavenge reactive oxygen species. Our results showed that GXNT can regulate dyslipidemia, reduce visceral and arterial lipid deposition, reduce AS plaques, resist oxidation and endothelial damage, inhibit platelet aggregation, and improve cardiac function, suggesting that GXNT can resist AS plaque formation and protect endothelial function and cardiac function, accordingly.
AS is a chronic low-grade inflammatory process [23,24]. In the early stage of AS, LDL infiltrates and is oxidized into ox-LDL after endothelial damage, followed by inflammatory factor expression, the adherence of monocytes to endothelial cells, foam cell forming by phagocytosis of ox-LDL. With the progression of AS, inflammatory factors increase, such as TNF-ɑ, IL-1β, and IL-6, which aggravate the development of AS plaques. As an independent risk factor for AS, CRP directly participates in the formation and aggregation of AS. The study showed that GXNT could inhibit MDA content, enhance SOD activity, and reduce the levels of proinflammatory factors, thus contributing to restraining the development of AS.
NF-κB is an important transcription factor in the process of an inflammatory response, which can regulate the expression of various inflammatory genes related to the occurrence of AS [25]. It is reported that TNF-ɑ can activate NF-κB,

Control
Model GXNT is considered a biochemical marker of plaque instability, and it can degrade almost all extracellular substances, leading to increased plaque vulnerability. The results of our study showed that GXNT can significantly reduce the expressions of NF-κB, TNF-ɑ, and MMP-9 in AS plaques, indicating that GXNT can reduce the inflammatory response, playing an important role in plaque stability.
Accumulating evidence suggests that gut microbiota play a critical role in cardiovascular disease. There were significant differences in the composition and diversity of gut microbiota in patients with different types of CHD, and Proteobacteria increased in CA patients [26], which is in line with our results. Another research reported that the relative abundances of Escherichia-Shigella and Enterococcus in patients with coronary artery disease significantly increased, while the relative abundances of Faecalibacterium and Roseburia significantly decreased, which may be related to functions such as lipopolysaccharide synthesis and propionic acid metabolism enhancement [27]. The current study showed decreased gut microbiota diversity in the CA model induced by a HF diet in minipigs, which is consistent with our previous findings [11], and GXNT can improve the bacterial composition to a certain extent. GXNT could decrease the relative abundances of Proteobacteria, Enterobacteriaceae, and Escherichia and increase the relative abundances of Prevotellaceae and Prevotella. Proteobacteria are known as microbial hallmarks of gut dysbiosis and are associated with immune inflammation and metabolic dysfunction. After GXNT intervention, the relative abundance of Proteobacteria was significantly reduced, speculating that proteobacteria were the characteristic bacteria of GXNT in treating AS. Jie et al. found that the abundance of Enterobacteriaceae family and Streptococcus species increased in patients with coronary artery disease, while the abundance of Roseburia Intestinalis and Faecalibacterium prausnitzii decreased, which was associated with reduced fermentative capacity and enhanced proinflammatory properties [28]. Studies have shown that Prevotella are the most abundant bacteria in the human gut and play a crucial role in nutrition and metabolism, such as degrading dietary fiber, fermenting polysaccharides, lowering cholesterol [29]. Thus, GXNT may ameliorate the formation of a HF diet-induced CA by modulating gut microbiota composition.
In addition, functional changes in gut microbiota were predicted using metagenomics, and targeted metabolites were detected based on the predicted results. The results showed that GXNT could affect the relative abundances of bacteria related to bile acid metabolism, choline biosynthesis, SCFA production in a HF diet-induced CA model, and TMAO, SCFAs, and BAs are closely related to AS [30]. TMAO has been recognized as a biomarker for predicting AS [31]. TMA is produced by TMA-forming microbial communities (e.g., Escherichia coli) from a HF diet mainly containing choline and carnitine. Subsequently, it is further oxidized to TMAO by hepatic flavin monooxygenase, resulting in AS occurrence through multiple mechanisms [32]. TMAO can induce the activation of MAPK and NF-κB in endothelial cells and smooth muscle cells and promote the expression of downstream inflammatory factors such as IL-8 and IL-1β. High levels of TMAO can also affect lipid metabolism, increasing the risk of cardiovascular disease by reducing reverse cholesterol transport and altering bile acid transport, composition, pool size, etc. [33]. In this experiment, GXNT can reduce the generation of TMA metabolites and lower the circulating TMAO level, thereby preventing the AS formation, which may be related to the regulation of Escherichia. SCFAs are the main products of intestinal dietary fiber fermentation. Acetic acid, propionic acid, and butyric acid are the most abundant SCFAs, accounting for about 95% of the total SCFAs. SCFAs can regulate energy metabolism, and blood pressure, fight inflammation and improve insulin resistance by binding to corresponding receptors [34]. Particularly, propionate and butyrate can inhibit the NF-κB signaling pathway by regulating receptor signaling, reduce the production of downstream inflammatory factors, and prevent AS progression. Butyrate can enhance intestinal barrier function by regulating the expression of claudin, which may be achieved by activating AMPK 20 Journal of Immunology Research or downregulating the expression of claudin. In this study, GXNT may increase intestinal acetate and propionate content, possibly affecting energy metabolism, intestinal permeability, and inflammatory factor production. Additionally, BAs are the main metabolites of cholesterol. They are now considered signaling molecules that interact with cell membranes and nuclear receptors and have regulatory effects on physiological processes such as energy balance and glycolipid metabolism. In the colorectal, primary bile acids, such as CA and CDCA, are metabolized to secondary bile acids, including DCA, LCA, and UDCA [35]. Conjugated or unconjugated bile acids are reabsorbed through the enterohepatic circulation, while LCA and UDCA are mostly excreted in the feces. Amazingly, gut microbes can regulate the ratio of bile acids. When the body is in an unhealthy state, gut microbes can cause a decrease in secondary bile acids, an increase in primary bile acids, activation of FXR, and downregulation of bile acid production, leading to cholesterol increase and AS occurrence [36]. Our results showed that GXNT can clearly reduce CDCA and increase LCA-3S, hinting that GXNT can promote primary bile acid conversion and increase bile acid excretion, thereby reducing cholesterol level. Overall, GXNT may also exert anti-AS function by influencing the metabolites of the gut microbiota.
The therapeutic effect of GXNT on AS may be associated with its phenolic acids. In previous studies, we have identified 14 active components using LC-MS [37], and 7 phenolic acids of which were quantified, including tanshinol, protocatechualdehyde, chlorogenic acid, caffeic acid, ferulic acid, rosmarinic acid, and salvianolic acid B. The phenolic acids have been widely reported for AS treatment. Tanshinol has a pronounced anti-inflammatory effect involving TLR2 and macrophages through the NF-κB signaling pathway, which supports the novel application of DSS in the treatment of relevant diseases including AS and ischemic or ischemic/ perfusion injury of the heart and brain [38]. Chlorogenic acid mitigates ox-LDL-induced endothelial oxidative stress and mitochondrial dysfunction by activating SIRT1 and modulating the AMPK/PGC-1 signaling pathway [39]. The mix of chlorogenic acid and caffeic acid exhibits a high antioxidant effect, which can reduce lipid storage in macrophages by a reduction in the expression of transcription factors C/EBPβ and PPAR-γ1 [40]. Gu et al. reported that FA could significantly ameliorate atherosclerotic injury, which may be partly by modulating gut microbiota and lipid metabolism via the AMPKα/SREBP1/ACC1 pathway [41]. Rosmarinic acid inhibits endothelial dysfunction which is shown in diabetic atherosclerosis through downregulating the p38-FOXO1-TXNIP pathway and inhibiting inflammasome activation [42]. Salvianolic acid B can attenuate the development of AS, the anti-AS effect of which is related to regulating the YAP/TAZ/JNK signaling pathway [43]. Moreover, these phenolic acids are linked to gut health. Substances containing phenolic acids, such as protocatechuic acid, chlorogenic acid, caffeic acid, ferulic acid, and rosmarinic acid, have the function of suppressing intestinal inflammation and modulating gut microbial populations [44,45]. Thus, we speculate that the phenolic acids contained in GXNT can regulate the composition and structure of the gut microbiota in the AS model, as well as the production of metabolites, exerting anti-inflammatory and anti-AS effects. Meanwhile, the phenolic acids can also be metabolized by the gut microbiota to generate beneficial components, which are absorbed by the body and contribute to AS treatment. In general, the therapeutic effect of GXNT on AS may be the result of the interaction between phenolic acids and gut microbiota, which needs further study.

Conclusions
GXNT has the functions of regulating blood lipids, antioxidative stress, antivascular endothelial injury, inhibiting platelet aggregation, anti-inflammatory, and reducing AS plaque area, thereby exerting an anti-CA effect. The mechanism may be closely related to regulating the composition of the gut microbiota and bacterial metabolites such as TMAO, SCFAs, and BAs. We found that GXNT could modulate the relative abundance of Proteobacteria, Enterobacteriaceae, Escherichia, Prevotellaceae, and Prevotella, possibly involved in AS development, which may be the target bacteria for GXNT in AS therapy. The experiment provides an experimental basis for the clinical application of GXNT in the treatment of CA and new inspiration for researching the underlying mechanism. However, the composition of GXNT is complex, and its specific mechanism of action needs to be further studied.

Data Availability
All related data have presented in the manuscript.

Conflicts of Interest
All authors declare that there is no conflict of interest.

Authors' Contributions
Chen Minli is responsible for project administration and funding acquisition. Yang Qinqin is responsible for conducting experiments, analyzing data, and writing the article. Xu Yanyun and Shen Liye are responsible for performing some experiments and data collection. Pan Yongming is responsible for supervision. Ma Quanxin and Chen Yu are responsible for providing experiment guidance for the research. Yu Chen is responsible for performing the pathological examination and statistical analysis. Qinqin Yang and Yanyun Xu are co-first authors.