The Hypolipidemic Effect of Hawthorn Leaf Flavonoids through Modulating Lipid Metabolism and Gut Microbiota in Hyperlipidemic Rats

Objective The purpose of this study was to explore the potential mechanisms of the lipid-regulating effects and the effect on modulating the gut microbiota of hawthorn leaf flavonoids (HLF) in the high-fat diet-induced hyperlipidemic rats. Methods The hypolipidemic effect of HLF was investigated in the high-fat diet-induced hyperlipidemic rats. The action targets of HLF in the treatment of hyperlipidemia were predicted by network pharmacology and KEGG enrichment bubble diagram, which were verified by the test of western blotting. Meanwhile, we used 16S rRNA sequencing to evaluate the effects of HLF on the microbes. Results The results of animal experiments showed that HLF could reduce the body weight and regulate the levels of serum lipid in high-fat diet (HFD) rats. Meanwhile, for the related targets of cholesterol metabolism, HLF could significantly upregulate the expression of LDLR, NR1H3, and ABCG5/ABCG8; reduce the expression of PCSK9; and increase the level of CYP7A1 in the intestinal tissue, whereas cholesterol biosynthetic protein expressions including HMGCR and SCAP were lowered by HLF. In addition, HLF increased the activities of plasma SOD, CAT, and GSH-Px and decreased the levels of Casp 1, NLRP3, IL-1β, IL-18, and TNF-α, improving the degree of hepatocyte steatosis and inflammatory infiltration of rats. Notably, HLF significantly regulated the relative abundance of major bacteria such as g_Lactobacillus, g_Anaerostipes, g_[Eubacterium]_hallii_group, g_Fusicatenibacter, g_Akkermansia, and g_Collinsella. Synchronously, we found that HLF could regulate the disorder of plasma HEPC and TFR levels caused by HFD. Conclusion This study demonstrates that HLF can regulate metabolic hyperlipidemia syndromes and modulate the relative abundance of major bacteria, which illustrated that it might be associated with the modulation of gut microbiota composition and metabolites.


Introduction
Hyperlipidemia, also known as dyslipidemia, refers to the increase of total cholesterol (TC), triglyceride (TG), and lowdensity lipoprotein cholesterol (LDL-C) and the decrease of high-density lipoprotein cholesterol (HDL-C) [1]. Te pathological process of hyperlipidemia is closely related to the physiological and pathological processes of many tissues and cells, such as metabolism [2], infammation [3], immunity, stress, and so on [4]. It is also an important risk factor for cardiovascular and metabolic diseases, such as atherosclerosis, fatty liver disease, obesity, hypertension, diabetes mellitus, coronary heart disease, and stroke [5,6]. At present, the drug therapy for hyperlipidemia in clinic are mainly statins, fbrates, and niacins, among which statins are the frst choice and are considered the cornerstone of preventing atherosclerotic and cardiovascular disease (ASCVD). In spite of the statins therapy-mediated positive efects on cardiovascular diseases, patient compliance is often poor due to their adverse efects [7]. It is worth mentioning that traditional Chinese medicine (TCM) has unique advantages that are low cost, efectiveness, and fewer side efects in the treatment of hyperlipidemia [8,9].
Hawthorn (Crataegi folium) leaves are the dried leaves of Crataegus pinnatifda Bge. of the Rosaceae plant, which have the efects of lipid lowering [10], antiatherosclerosis [11], antiliver damage [12], anti-infammation, and antioxidative stress [13]. A variety of hawthorn leaf preparations, such as Yixintong and Shanmei capsule, is clinically used to treat cardiovascular diseases such as hyperlipidemia, coronary heart disease, angina pectoris, and arrhythmia [14,15]. At present, a great many types of chemical constituents have been extracted from hawthorn leaves, including favonoids, favane and its polymers, pentacyclic triterpenes, monoterpenes, sesquiterpenes, lignans, organic acids, volatile oil, and so on [10,16,17], among which favonoids are considered being the main active ingredients of hawthorn leaves and important ingredients for the herb to exert its drug activity [18,19]. Our previous study [20] found that hawthorn leaf favonoids (HLF) can reduce the levels of blood lipid and improve the liver function in hyperlipidemic mice. Meanwhile, the protein expression profles of HMGCR in the liver were downregulated by HLF. However, the specifc mechanism by which HLF regulate lipid metabolism remains unclear.
Network pharmacology integrates multidisciplinary and multiomics databases, and systematically and integrally connects drugs and diseases through network construction tools, which can provide scientifc, technological, and theoretical support for the study of the action mechanism of TCM [21][22][23]. For instance, baicalin could regulate the gene expression of SLC2A1, TNF, NFKB1, SREBF1, and CASP3 to ameliorate obesity and hyperlipidemia through a network pharmacology approach [24]. Furthermore, the cholesterol metabolism, fat digestion and absorption, and PPAR signaling pathways were identifed as the potential mechanism of sea buckthorn favonoids extract, among which isorhamnetin could activate the PPAR-c/NR1H3/CYP7A1 pathway against HLP by network pharmacology analysis and experimental validation [25]. In this study, network pharmacology analysis was used to screen the active components of HLF to further predict the action targets of the active compounds. Te mechanism of HLF in the treatment of hyperlipidemia was further analyzed by constructing a network diagram and KEGG enrichment bubble diagram.
Gut microbiota is a key environmental factor that regulates body metabolism [26]. A balanced gut microbiota can maintain lipid homeostasis through pathways such as regulating hepatic cholesterol metabolism, promoting muscle lipid oxidation and adipose tissue energy storage, and maintaining the integrity of the gut barrier [27,28].
Imbalances in the gut microbiota can lead to proliferation of potential pathogenic bacteria, afect immune homeostasis, and induce the production of infammatory cytokines and adipokines. It is closely associated with the development and progression of chronic diseases such as hyperlipidemia, obesity, diabetes, and atherosclerosis [29][30][31]. Meanwhile, gut mucosal barrier damage induced by a high-fat diet exacerbates this condition [32,33]. It is demonstrated that the relative abundances of the benefcial genuses such as Lactobacillus and Oscillibacter were inhibited by the high-fat diet. Furthermore, Auricularia auricula and its polysaccharides could improve the intestinal microbial environment by enriching SCFA-producing bacteria to relieve liver damage and treat hyperlipidemia [34]. However, the microbiota-modulating efects of hawthorn leaf favonoids on diet-induced hyperlipidemia rats have not been revealed yet.
Terefore, the hypolipidemic efect of HLF was investigated in high-fat diet-induced hyperlipidemic rats. Te action targets of HLF in the treatment of hyperlipidemia were predicted by network pharmacology and KEGG enrichment bubble diagram, which were verifed by molecular docking and the test of western blotting. Meanwhile, the composition and richness of gut microbiota were tested by 16S rRNA sequencing, and the correlations with HLF intervention were analyzed accordingly.

Biochemical
Analysis. Blood samples were taken from the orbit vein and subsequently centrifuged. Te serum or plasma obtained was stored at −80°C until biochemical analysis. Te serum lipid levels (TC, TG, LDL-C, and HDL-C) were determined by a Beckman Coulter AU480 Automatic Biochemical Analyzer (USA).  1.30.2), and the diference between the groups is analyzed. Use the R language tool to make the rank abundance curve explain the diversity, draw the pan/core species curve to judge whether the sample size is enough, evaluate the total species richness and the number of core species in the feces, and make the Venn chart and community bar chart to show the species' composition and similarity. Based on two distance algorithms (unweight-ed_unifrac and weighted_unifrac), principal coordinate analysis (PCoA), statistical analysis, and nonmetric multidimensional scaling analysis (NMDS) were carried out by using Qiime (version v. 1.9.1) as well as R language tools to Evidence-Based Complementary and Alternative Medicine calculate the distance between the samples and obtain the distance matrix. LEfSe software is used to analyze the difference of species' relative abundance between the groups. In this study, all-against-all (more-strict) comparison strategies and bacteria with linear discriminant analysis (LDA) score >3 and p < 0.05 were selected as diferential bacteria.

ELISA Analysis.
In terms of alpha diversity, Student's test is used to test the diference of index values among the groups. In terms of beta diversity, ANOSIM is used to detect the diference of community composition among diferent groups. One-way ANOVA test and post hoc Tukey-Kramer test are used to compare the abundance of gut microbiota in each group. In LEfSe analysis, the nonparametric Kruskal-Wallis (KW) sum rank test is mainly used to detect the species' abundance diferences among diferent groups and signifcantly different species are obtained. Ten, Wilcoxon rank sum test is used to test the diference consistency of diferent species in diferent subgroups; fnally, LDA (linear discriminant analysis) is used to estimate the impact of these diferent species on the diference between the groups.

Compound Collection and Target Prediction for HLF.
Te active ingredients of HLF are extracted from the Traditional Chinese Medicine Database and Analysis Platform (TCMSP) [38], the chemical database, and supplemented with references to the related literature. Te acquired active substances are imported into the PubChem [39] for InChi and canonical smiles. Te acquired InChi and canonical smiles are input into the Comparative Toxicogenomics Database (CTD) [40], STITCH [41] according to the method of compound similarity search, the species are defned as Homo sapiens, the potential targets of the active ingredients are predicted, and all target information is normalized using UniProt [42].

Hyperlipidemia-Related Targets Collection and
Screening. "Hyperlipidemia" has been retrieved from GeneCards [43], and the top 200 as potential targets for hyperlipidemia have been tested by top-down score. After all target names are corrected, merged, and removed duplicates and TCM targets are imported into the Jvenn [44] to draw a Venn diagram, and its intersection is taken to obtain 83 HLF antihyperlipidemia possible targets.

PPI Network Construction.
Proteins rarely function as a single substance, but as members in a dynamic network. Te accumulation of evidence suggests that protein-protein interactions (PPI) are critical to many biological processes in living cells [45]. For clarity of the interaction relationship between HLF-related targets and hyperlipidemia targets, we submitted the 83 intersection targets to STRING [46] with the species set as "Homo sapiens".

Analysis of GO Function and KEGG Pathway
Enrichment. Intersection targets are entered into the Metascape [47], setting the species to "H. sapiens," min overlap � 3, p value cutof � 0.01, min. enrichment � 1.5, and p < 0.01. Mainly analyze GO molecular functions (MF), GO components (CC), GO biological processes (BP), and KEGG pathway; save the data results; and import the Origin 2021 software to draw the bubble chart. 2.7.5. Construction of the "Active Ingredient-Target-Biological Processes" Network Diagram. Te −log p top 25 biological processes and related components and targets are imported into the model using Cytoscape 3.8.2. Nodes are used to represent active ingredients and predicted targets, and nodes are connected to the edges to represent subordination. Network analysis is used to analyze network topological properties.

Western Blotting Assay.
Total proteins were obtained from the rat hepatic tissue and intestinal tissue homogenates with RIPA bufer supplemented with phenylmethylsul-fonyl fuoride and protease inhibitor cocktail. Protein samples were separated on 10% separation gel and then transferred to polyvinylidene fuoride membranes. After blocking with 5% fetal bovine serum for 1 h, we then incubated separately with primary rabbit polyclonal antibodies against SCAP (1 : (1 : 1500), ABCG8 (1 : 1500), and mouse polyclonal antibodies against β-actin (1 : 10000) overnight at 4°C. After washing, the membranes were incubated at room temperature for 45 minutes with appropriate secondary antibodies. Finally, the membranes were treated according to the protocol of the enhanced chemiluminescence detection kit and protein bands were observed by Tanon 4200. Te intensities of protein bands were quantifed with the Image J software and the values normalized to β-actin.
2.9. Statistical Analysis. All results were presented as mean ± SD. Te statistical analysis was performed using SPSS (version 26.0). Diferences between the groups were statistically analyzed using one-way analysis of variance (ANOVA). A value of p < 0.05 was considered statistically signifcant. Diagrams are performed by GraphPad Prism version 9.1.

Changes in Body Weight.
To evaluate the efect on regulating the levels of blood lipid by HLF, we analyzed the body weight gain of rats for 6 weeks ( Figure 1). When compared with the NC group, the body weight of rats in the HFD group had considerably increased by 11.87% after 6 weeks (p < 0.05). However, after receiving HLF treatment for four weeks, compared with the HFD group, the body weight of rats in the HLF-M and HLF-H treatment groups was signifcantly reduced by 11.71% (p < 0.05) and 10.04% (p < 0.05), respectively. Evidence-Based Complementary and Alternative Medicine

Serum Lipid Levels in Rats.
We used an automatic biochemical analyzer to determine the serum lipid levels in rats. As shown in Figures 2(a)-2(d), the TC, TG, and LDL-C concentrations of the HFD group were signifcantly increased, and HDL-C levels were signifcantly decreased compared with the NC group (TC, LDL-C, and HDL-C: p < 0.001, TG, p < 0.01), indicating the successful establishment of the hyperlipidemia rat model. Compared with the HFD group, the TC and LDL-C levels in the HLF-L, HLF-M, and HLF-H treatment groups were signifcantly reduced, with TC levels falling by 37.52% (p < 0.001), 37.20% (p < 0.001), and 49.10% (p < 0.001), respectively, LDL-C levels falling by 41.11% (p < 0.001), 43.10% (p < 0.001), and 53.14% (p < 0.001), respectively. Meanwhile, the TG in the HLF-M and HLF-H groups were dramatically decreased by 42.52% (p < 0.001) and 25.66% (p < 0.05), respectively, and the HDL-C in the HLF-L and HLF-H groups were signifcantly increased by 44.19% (p < 0.01) and 54.26% (p < 0.05), respectively, when compared with the HFD group. Tese fndings suggested that HLF can signifcantly improve the levels of serum lipid in hyperlipidemic rats. Atherosclerosis is the chronic accumulation of cholesterol-rich plaques within the arteries, which is associated with a range of cardiovascular diseases including peripheral vascular disease, aortic aneurysm, myocardial infarction, and stroke [48]. Atherogenic index (AI, AI � (TC − HDL-C)/HDL-C) is considered as a strong marker to predict the risk of atherosclerosis and coronary heart disease [49]. As shown in Figure 2(e), the AI levels in the HFD group were dramatically raised (p < 0.001), compared with the NC group. However, the AI levels in the HLF-L, HLF-M, and HLF-H treatment groups were signifcantly lowered by 58.68% (p < 0.001), 42.00% (p < 0.001), and 69.18% (p < 0.001), respectively, compared with the HFD group. It is suggested that HLF has potential to inhibit the progression of atherosclerosis in hyperlipidemia rats, which needs further study.

Antioxidant Profles in Plasma.
Tere are multiple mechanisms which can be completed through key antioxidants such as SOD, CAT, and GSH-PX in the human body to prevent oxidative stress caused by free radicals [50]. As shown in Figures 3(a)-3(c), compared with the NC group, the activities of plasma SOD, CAT, and GSH-PX in the HFD group were signifcantly decreased (p < 0.05 or p < 0.01), illustrating that the oxidative stress response of rats fed the high-fat diet were aggravated. While the activities of SOD, CAT, and GSH-PX in the HLF-L, HLF-M, and HLF-H treatment groups were signifcantly enhanced, with SOD being increased by 7.18% (p < 0.05), 9.30% (p < 0.01), and 7.80% (p < 0.05), respectively; CAT being raised by 8.92% (p < 0.05), 9.24% (p < 0.05), and 9.07% (p < 0.05), respectively; and GSH-PX being increased by 14.83% (p < 0.05), 31.96% (p < 0.01), and 18.75% (p < 0.05), respectively, compared with the HFD group, which manifested that HLF could signifcantly inhibit the oxidative stress response in hyperlipidemic rats.

Anti-Infammatory in Plasma and Intestinal Tissue.
We measured the levels of proinfammatory cytokines, such as plasma IL-1β, IL-6, IL-18, TNF-α, Casp1, and NLRP3 in the intestinal tissue of rats to better understand the antiinfammatory efects of HLF. As shown in Figures 4(a)-4(f ), the levels of NLRP3, Casp1, IL-1β, IL-6, IL-18, and TNF-α in the HFD group were signifcantly increased (p < 0.05 or p < 0.01 or p < 0.001), compared with the NC group, which demonstrated that the infammatory response of rats fed the high-fat diet were exacerbated.

Hepatic Morphology (HE Staining).
We used HE staining to analyze the pathological changes of the liver tissue. As shown in Figure 5, the NC group appeared with : Te changes in the body weight of rats before and after HFD feeding (0W: before feeding HFD, 6 W: after 6 weeks of feeding HFD). Te data are presented as the mean ± SD (n � 8).
Note: compared with the HFD group, * p < 0.05. Compared with the NC group, # p < 0.05.
Evidence-Based Complementary and Alternative Medicine the normal hepatic lobular structure, normal hepatocytes, no fatty vacuoles in the cytoplasm, and no steatosis or necrosis. While the HFD group showed various degrees of steatosis and a great quantity of lipid vacuoles, the hepatocyte degeneration was mostly round, enlarged in size, and partially was infltrated by infammatory cells, which indicated that the high-fat diet induced hepatic steatosis in rats. Te three dosages of HLF signifcantly decreased lipid droplets and lessened the infltration of infammatory cells in the liver to diferent degrees, especially in the HLF-M group, which manifested that HLF could ameliorate the accumulation of lipid droplets and inhibit infammation in hepatic of rats fed high-fat diet.
3.6. Intestinal sIgA. As shown in Figure 6, the levels of intestinal sIgA in the HFD group were signifcantly declined (p < 0.05), compared with the NC group, suggesting that the intestinal immune functions of rats fed high-fat diet were seriously impaired. Compared with the HFD group, the Tese data are presented as the mean ± SD (n � 8). Note: compared with the NC group, # p < 0.05; ## p < 0.01; ### p < 0.001. Compared with the HFD group, * p < 0.05; * * p < 0.01; * * * p < 0.001. 6 Evidence-Based Complementary and Alternative Medicine levels of intestinal sIgA in the HLF-H treatment group were signifcantly increased by 58.38% (p < 0.01), which indicated that HLF could improve the immune function of the gastrointestinal tract by regulating the levels of sIgA.

Plasma TRF and HEPC.
Our study showed that the levels of plasma TRF and HEPC in the HFD group of rats were signifcantly reduced (p < 0.05 or p < 0.001) compared with the NC group (Figure 7), suggesting that the high-fat diet could lower the levels of plasma iron and increase iron accumulation and deposition in the liver, leading to dysregulation of iron metabolism. Notably, the levels of plasma TRF and HEPC in the HLF-M and HLF-H treatment group were signifcantly raised, with TRF being increased by 27.17% (p < 0.01) and 52.44% (p < 0.001), respectively, and HEPC being grown by 23.99% (p < 0.05) and 26.95% (p < 0.01), respectively, which demonstrated that HLF could modulate the disorder of body iron metabolism.

Species Annotation and Assessment.
Annotations and species assessments include primarily OTU (operational classifcation unit) analysis, alpha diversity analysis, and rarefaction curve analysis. In this study, we conducted 16S rRNA sequences of 48 fecal microbiota samples for 2,425,668 high-quality sequences following quality control.
Te mean sequence length was 410. Te sequence length was mainly distributed in 420∼440 bp, followed by 400∼420 bp. Te NC group showed signifcant diferences (p < 0.05) in the number of OTUs compared to the rest of HFD. While compared with the number of OTUs in the HFD, only the HLF-M showed a signifcant diference (p < 0.05) (Figure 8(a)). In addition, pan/core analysis showed the total number of species in each increased gradually with increasing sample size, and the number of core species in each tended to remain stable with increasing sample size, indicating an adequate sample size for this experiment (Figures 8(b) and 8(c)). Based on the OTU levels, the results of α-diversity analysis showed that the indices of Sobs (Figure 8 and stable. Tis indicated that HFD causes a disturbance in the composition of the gut microbiota and HLF could reverse this situation.

Sample Comparison Analysis.
To further investigate the similarities or diferences between gut microbial compositions, we assessed β-diversity at the OTU level using nonmetric multidimensional scaling (NMDS) and principal coordinate analysis (PCoA), with ANOSIM-tests for differences between(Figures 9(a)-9(d), Table 1). Te results of PCoA based on unweighted and weighted, as well as NMDS, indicated the microbial composition of NC difered from that of HFD (p < 0.01), suggesting the formation of hyperlipidemia altered the composition of the whole gut microbial composition. According to the unweighted analysis, results showed the gut microbial composition of the HLF-H was completely diferent from that of the HFD (p < 0.01), a small portion of the intestinal microbiota of the other treatment groups overlapped with the HFD and only the microbiota of HLF-M showed obvious convergence towards  Tese data are presented as the mean ± SD (n � 8). Note: compared with the NC group, # p < 0.05; ## p < 0.01 ; ###p < 0.001. Compared with the HFD group, * p < 0.05; * * p < 0.01; * * * p < 0.001. 8 Evidence-Based Complementary and Alternative Medicine NC. However, the weighted analysis showed that there was increased overlap in the gut microbiota between each treatment and the HFD, and the HLF-H presented a more conspicuous convergence towards NC. Moreover, further analysis using the ANOSIM test revealed that although HLF and AVT showed signifcant diferences in the microbiota compared with HFD (p < 0.05), HLF-H had the strongest explanation for the diference from the HFD (R = 0.8114).

Species Diference Analysis.
In this study, we normalized each sample to equal sequencing depth and clustering according to the minimum sample sequence number. Data analysis obtained 906 OTUs with 97% similarity and 149 OTUs in common, detecting 15 phyla, 26 classes, 41 orders, 76 families, 197 genera, and 365 species (Figure 10(a)). Subsequently, we calculated the species richness of each sample at diferent taxonomic levels, classifed the species with an abundance ratio below 0.01 among all samples as others, and averaged the values to calculate within group samples. Five phyla, 15 families, and thirty genera were identifed, representing over 0.01% of all samples (Figures 10(b)-10(d)). By visually displaying the species abundance of each group at diferent taxonomic levels, we could intuitively show which dominant species each sample contains at the taxonomic level and the relative abundance of the dominant species. At the phylum level, p_Firmicutes prevailed in all subjects' gut microbiota, and smaller populations include  Note: compared with the NC group, # p < 0.05; ## p < 0.01. Compared with the HFD group, * p < 0.05; * * p < 0.01. p_Bacteroidetes, p_Proteobacteria, p_Actinobacteria, and p_Verrucomicrobia. Compared with NC, HFD showed a relatively increased abundance of p_Firmicutes, p_Proteobacteria, and p_Actinobacteria and a relatively decreased abundance of p_Bacteroidetes and p_Verrucomicrobia, indicating the imbalance of gut microbiota dysbiosis in hyperlipidemia. To a certain extent, HLF intervention attenuated the dysbiosis of the gut microbiota. Compared with HFD, the three doses of HLF could increase the relative abundance of p_Verrucobacteria and p_Bacteroidetes, and reduce the relative abundance of p_Actinobacteria. While HLF-L and HLF-H can decrease the relative abundance of p_Firmicutes, and HLF-M could decrease the relative abundance of p_Proteobacteria. At the family level, f_Lachnospiraceae prevailed in all subjects' gut microbiota, and the composition of gut microbiota in each group was partially diferent. Compared with the NC group, f_Lachnospiraceae, f_Erysipelotrichaceae, and f_Enterobacteriaceae in the HFD group increased signifcantly and f_Lactobacillus, f_Verrucomicrobiaceae, and f_Ruminococcaceae decreased signifcantly, while HLF could signifcantly improve the relative abundance of these species. At the genus level (Figures 10(e)-10(h)), one could clearly see that the structure and relative abundance of the principal microbiota of each group have clearly been altered. We used the signifcant diference test between the groups to analyze the species with a relative abundance ratio ≥0.01 and evaluate the signifcance level of the diference in species abundance. Compared with the NC group, the number of g_Blautia, g_Anaerostipes, and g_Allobaculum in the HFD group signifcantly increased, while the number of g_Lactobacillus, g_Akkermansia, and g_Alloprevotella signifcantly decreased. It is worth noting that HLF treatment could improve the dysbiosis, among which HLF-H could signifcantly reduce the relative abundance of g_Anaerostipes, g_Collinsella, g_Fusicatenibacter, and g_ [Eubacterium]_hallii_group (p < 0.05).
Besides, to further explore diferences in the specifc gut microbiota among all subjects, we used the linear discriminant analysis efect size (LEfSe) method to recognize the specifc altered bacterial phenotypes at each phylogenetic level (from phylum to genus), with linear discriminant analysis (LDA) > 3, ,p < 0.05 and multigroup comparison strategy of all-against-all (Figures 10(i)-10(n)). By comparing the signifcantly diferent species of NC and HFD, it could be seen that HFD caused a serious imbalance in the composition and relative abundance of intestinal microorganisms. After HLF treatment, the main species of gut microbiota (the proportion of species ≥0.01), such as p_Verrucomicrobia, f_Lactobacillaceae, g_Akkermansia, and g_Lactobacillus, changed signifcantly. Although the results of the LEfSe test were quite diferent from those of the Tukey-Kramer test, it still showed that HLF could signifcantly regulate the relative abundance of these species, indicating that these bacteria were associated with the HLF treatment of hyperlipidemia.

Active Ingredients and Targets for HLF.
A total of 75 favonoids compounds in hawthorn leaves were obtained by searching TCMSP, chemistry database, and related literature, as shown in Table 2. By predicting the potential targets of active compounds based on the STITCH and CTD platforms, the 83 targets were fnally screened for their possible association with the prevention and treatment of hyperlipidemia in HLF.

Construction and Analysis of PPI Network.
We introduced the abovementioned 83 intersection targets into the string platform, resulting in a column of proteinprotein interaction data and a PPI network map. After the minimum interaction threshold was set to "highest confdence" (0.900) and disconnected nodes were hidden, 68 closely linked targets were fnally obtained. Te PPI network diagram of this study included 68 nodes, 143 edges, and the average node degree was 3.45. Te local clustering coefcient was 0.46. Te PPI analysis considered that this network to be far more interactive than expected, meaning more protein-protein interactions than would be expected from a set of proteins randomly  Figure 7: (a) Plasma TRF; (b) plasma HEPC. Tese data are presented as the mean ± SD (n � 8). Note: compared with the NC group, # p < 0.05; ## p < 0.01; ### p < 0.001. Compared with the HFD group, * p < 0.05; * * p < 0.01; * * * p < 0.001. 10 Evidence-Based Complementary and Alternative Medicine drawn from the genome with the same size and degree distribution, as shown in Figure 11.

Network of HLF Active Ingredients-Antihyperlipidemic
Targets-Biological Processes. In this study, the compound, target, and biological processes' information obtained above were imported to Cytoscape 3.8.2 software to construct the "HLF active ingredient antihyperlipidemia target pathway            Evidence-Based Complementary and Alternative Medicine network" (Figure 13). Te network contained 131 nodes, 843 edges, and the network concentration was 0.396. It was predicted that quercetin was the main component of antihyperlipidemia in HLF, followed by (+)-catechin, epicatechin, and so on. Taken together the KEGG analysis and the degree values, we supposed that the targets of HLF against hyperlipidemia might be related to the biosynthesis, transport, and homeostasis regulation system of several lipids, including cholesterol, steroids, and fatty acids, and the targets were mainly enriched in APOE, LDLR, PPARG, and so on.  (Figures 14(a) and 14(b)), the level of CYP7A1 in HFD rats was decreased signifcantly (p < 0.01), and the HMGCR level was increased signifcantly (p < 0.01). Compared with the HFD group, the levels of CYP7A1 in the HLF-L, HLF-M, and HLF-H treatment group were signifcantly increased (p < 0.05), while the HMGCR levels in the three dosages of HLF were signifcantly reduced (p < 0.01). As shown in Figures 14(c) and 14(d), the protein expression profles of SCAP, PCSK9, HMGCR, SREBF2, and NLRP3 (p < 0.05 or p < 0.01) in the liver or intestine of HFD rats were signifcantly upregulated, while the expressions of LDLR, NR1H3, ABCG5, and ABCG8 were signifcantly downregulated (p < 0.05 or p < 0.01) compared with the NC group. Compared with the HFD group, the protein expression profles of SCAP, HMGCR, and NLRP3 were   18 Evidence-Based Complementary and Alternative Medicine downregulated by HLF (p < 0.05 or p < 0.01). Meanwhile, the expression profles of PCSK9 in the three dosages of the HLF group were signifcantly decreased (p < 0.05). However, the protein expression profles of ABCG5, ABCG8, LDLR, and NR1H3 were upregulated by HLF (p < 0.05 or p < 0.01).

Te Potential
Although HLF did not signifcantly inhibit the expression of SREBF2 compared with the HFD group, a very clear decreasing trend was still seen. Tese results indicated that the mechanism of HLF treatment for hyperlipidemia may be related to the regulation of cholesterol biosynthesis, metabolism, and transport.

Correlation Analysis of Intestinal Microbes.
Based on the bacteria at the genus level, we used Spearman's test to analyze the relationship between bacteria and each indicator in plasma and serum ( Figure 15). In blood lipid levels, the abundance of g_Lactobacillus and g_Akkermansia were signifcantly negatively correlated with the levels of LDL-C and TC, but signifcantly positively correlated with the levels of HDL-C. Te abundances of g_Anaerostipes, g_[Eubacterium]_hallii_group, g_Collinsella, and g_Fusicatenibacter were signifcantly positively correlated with the levels of LDL-C and TC, and negatively correlated with the levels of   Figure 11: PPI network of HLF. Nodes in the fgure represent proteins, and each edge represents a protein-protein interaction relationship, and the more lines represent a greater association. Te sky-blue line in the fgure represents protein-protein interactions obtained from the created database and the purple line represents experimentally determined protein-protein interactions.
HDL-C. In terms of cholesterol metabolism, CYP7A1 was a signifcant target with a signifcant correlation to bacterial production and a signifcant positive correlation to the abundance of g_Lactobacillus and a signifcant negative correlation to the abundance of g_[Eubacterium]_hal-lii_group, g_Collinsella, and g_Fusicatenibacter. In terms of antiinfammation, the abundance of g_Anaerostipes was signifcantly positively correlated with the level of NLRP3, but the abundance of g_Lactobacillus was signifcantly negatively correlated with the level of NLRP3. In terms of immunity, Casp1 and sIgA were important indicators that are signifcantly related to bacteria. Te abundances of g_ [Eubacterium]_hallii_group, g_Anaerostipes, g_Collinsella, and g_Fusicatenibacter were signifcantly positively correlated with the level of Casp1 and negatively correlated with the level of sIgA. However, the correlation of g_Lactobacillus was just the opposite of these bacteria. In terms of antioxidants, the abundance of g_Collinsella, g_ [Eubacterium] _hallii_group, and g_Anaerostipes was signifcantly negatively correlated with the level of CAT, while the abundance of g_Lactobacillus was signifcantly positively correlated with the level of SOD. In terms of iron metabolism, g_Anaerostipes and g_Fusicatenibacter were negatively correlated with the level of HEPC, while g_Fusicatenibacter and g_ [Eubacterium]_hallii_group were negatively correlated with the level of TFR and g_Akkermansia was positively correlated with the level of TFR.

Discussion
Hyperlipidemia, whose pathogenesis is very complex, is accompanied by the ascent of serum TC, TG, and LDL-C and the decrease of HDL-C levels, which is inseparable from the physiopathological processes, and cholesterol accumulation caused by any factor can lead to hyperlipidemia and aggravate the occurrence as well as development of cardiovascular disease [51,52]. Increased intestinal cholesterol absorption or increased liver cholesterol biosynthesis can easily cause the accumulation of cholesterol in the body, promoting the activation of NLRP3 infammasome and triggering the expression of infammatory factors such as interleukin-1 beta (IL-1β), interleukin-18 (IL-18), and tumor necrosis factor-α (TNF-α) [53,54]. Te high concentrations of cholesterol or cholesterol crystals can promote the activation of NLRP3 infammasome to start infammation associated with hyperlipidemia or atherosclerosis [55,56]. In this study, HLF reduced the serum TC, TG, and LDL-C levels and increased HDL-C levels in hyperlipidemia model rats, improving lipid profles. It is also shown that HLF could attenuate the liver tissue swelling and improve infammatory cell infltration or fatty lesions. Moreover, HLF could signifcantly reduce the levels of NLRP3, caspase-1, IL-1β, IL-6, IL-18, and TNF-α and increase the activities of plasma SOD, CAT, and GSH-PX, which suggested that HLF could efectively relieve the  Evidence-Based Complementary and Alternative Medicine infammatory response and oxidative stress induced by hyperlipidemia. Cholesterol synthesis and metabolism require a large number of enzymes to catalyze, among which 3-hydroxy-3methylglutaryl-coenzyme A reductase (HMGCR), a key enzyme in cholesterol biosynthesis, is regulated by SCAP/ SREBF2 (sterol regulatory element-binding protein cleavage-activating protein/sterol regulatory elementbinding protein 2) regulation [57,58]. When cholesterol is low in the cell, SCAP activates the cleavage of S1P protease and the release of the active fragment, which will activate the expression of the downstream target gene HMGCR, and ultimately contributes to increase cellular cholesterol uptake instead of endogenous synthesis [59][60][61]. When cells are high in cholesterol, SCAP inhibits this cleavage reaction, which leads to decrease the expression of the downstream target genes HMGCR [62,63]. Finally, compared with the endogenous synthesis, it leads to a decrease in cellular cholesterol uptake to maintain cholesterol homeostasis. In this study, HLF could lower the level of HMGCR in plasma and decrease the protein expression profles of HMGCR and SCAP in HFD rats, which demonstrated that HLF could inhibit cholesterol biosynthesis and improve the lipidlowering activity.
Low-density lipoprotein receptor (LDLR) is the liver surface receptor of LDL, responsible for removing LDL-C from human blood. After binding to LDL particles, LDLR is internalized into clathrin-coated pits and then transports LDL from the cytoplasm to the lysosome for degradation [64]. Proprotein convertase subtilisin/kexin type 9 (PCSK9) is a soluble protein and a ligand for LDLR. Extracellular PCSK9 binds to LDLR through protein-protein interactions and directly enters the lysosome as a PCSK9-LDLR complex for destruction, inhibiting LDLR recirculation and lowering plasma LDL-C levels [65,66]. Te conversion of cholesterol into bile acid (BA) in the liver is its main metabolic pathway. Cholesterol 7α-hydroxylase (CYP7A1) is the rate limiting enzyme in the conversion of cholesterol to Bas [67]. Te expression of CYP7A1 is regulated by the liver X receptor alpha (NR1H3), and its activity determines the rate of BA synthesis [68]. In addition, ATP-binding cassette subfamily G member 5 and 8 (ABCG5/G8) are a heterodimeric complex, mainly on the tubule membrane of hepatocytes and the apical membrane of intestinal cells, which can     [69,70]. Te expression of ABCG5/G8 can prevent liver fat accumulation by reducing cholesterol concentration and fatty acid intake [71]. In the liver, the main regulator of ABCG5 and ABCG8 mRNA expression is NR1H3, and NR1H3 promotes cholesterol excretion by regulating ABCG5/G8 transporters [72,73]. In this study, HLF could signifcantly downregulate the expression of PCSK9; increase the level of CYP7A1 in the intestinal tissue; and upregulate the expression of LDLR, NR1H3, ABCG5, and ABCG8 in HFD rats to decrease the intestinal absorption of cholesterol, promoting cholesterol excretion. HEPC is the most master regulator of the body iron metabolism, and increased uptake of iron by the liver leads to increased production and secretion of hepcidin, which regulates iron metabolism by inhibiting ferroportin located in intestinal enterocytes and macrophages. Contrary to the expected results, plasma HEPC and TRF levels were both decreased in the studies in which we used high-fat fed animals, a situation consistent with the experimental results of Ye [74]. Tis may be caused by dysregulation of iron metabolism due to hepatic impairment, as a manifestation of decreased hepcidin is also seen in patients with chronic liver disease [75]. In addition, in previous studies, patients with dyslipidemia and atherosclerosis also presented signifcant reductions in TRF levels [76,77]. In this study, HLF was able to signifcantly increase plasma TRF and HEPC levels, which suggested that HLF could regulate the disorder of the body iron metabolism.
Structural variations in the gut microbiome are associated with the health of the host, and studies on the composition of the gut microbes are helpful for the diagnosis and treatment of hyperlipidemia [78,79]. According to the results of the diference and LEfSe test for each group, the species in the HLF group with signifcant diferences were mostly concentrated in p_Firmicute, 、p_Actinobacteria, and p_Verrucomicrobia. Te p_Firmicutes were mainly enriched in f_Lachnospiraceae. F_Lachnospiraceae produce high amounts of short-chain fatty acids, which is the largest butyrate producing group of Firmicutes [80,81]. It has turned out to be important to maintain the metabolic health of the gut microbiota and the stability of the internal environment [82,83]. Tis microbial imbalance may be related to the changes in fatty acid levels in the gut [84,85]. Its abundance is closely related to glycolipid metabolism [86,87], host immune activation [88], and infammatory response [89], which in turn afects the level of bile acid [90]. Specifcally, HLF had signifcant efects on g_Lactobacillus, g_Anaerostipes, g_[Eubacterium]_hallii_group, and g_Fusicatenibacter belonging to f_Lachnospiraceae in the gut microbes of hyperlipidemia rats, indicating that HLF may afect the levels of short-chain fatty acids to realize the treatment of hyperlipidemia by regulating the relative abundance of these gut microbe.
g_Akkermansia is an important part of p_Verrucomicrobia, which can use mucin as the sole carbon and nitrogen source and release free forms of sulfate from mucin fermentation, resulting in improved host metabolism [91,92]. In the metabolic syndrome, obesity and hypertriglyceridemia were most strongly associated with  Figure 16: Te efects of hawthorn leaf favonoids on the physiopathological processes of hyperlipidemia through modulating lipid metabolism and gut microbiota. 24 Evidence-Based Complementary and Alternative Medicine g_Akkermansia; followed by reduced HDL cholesterol, hypertension, and hyperglycemia; and increasing g_Akkermansia abundance can reverse the efects of a high-fat, highcholesterol diet [93,94]. Studies have shown that the mechanism of adjusting the relative abundance of g_Akkermansia in the treatment of metabolic diseases may be related to its ability to stimulate GLP-1 secretion [95,96], promote 5-HT biosynthesis, and intestinal stem cellmediated epithelial development [97,98]. g_Collinsella belongs to f_Coriobacteriaceae and p_Actinobacteria, which produces lactate, formate, and butyrate [99] can modify host bile acids and infuence metabolism by altering intestinal cholesterol absorption, reducing hepatic glycogen production and increasing triglyceride synthesis [100][101][102]. In terms of infammation, g_Collinsella also increases intestinal permeability, decreases the expression of tight junction protein in epithelial cells, and induces the expression of IL-17 [103][104][105]. At present, the molecular mechanism by which g_Collinsella afects host metabolism is not yet clear [106]. It is certain that g_Collinsella is involved in the progression of ulcerative colitis [107], hyperlipidemia [108], and diabetes [109]. A recent study found that a lower dietary fber intake may lead to an increased abundance of g_Collinsella. A structured weight loss program could signifcantly reduce the abundance of g_Collinsella in patients [110]. Anaerostipes is closely related to eating habits and infammation in obese people, but the precise mechanism is unclear [103,111]. It is reported that by promoting propionate formation via inositol or phytate, anaerostipes may lower the risk of metabolic disorders [112]. Tese species, which can use a variety of substrates as well as lactate and acetate to create butyrate, are among the most efective lactate consumers in the human colon [113,114]. Tis is essential for maintaining healthy intestinal barrier function. Tis is essential for maintaining healthy intestinal barrier function [115].
Spearman's correlation analysis further proved that these genera were closely connected with the regulatory efects of HLF on lipid metabolism, cholesterol transport, metabolism, immunity, infammation, and oxidative stress. Te present study suggested that the mechanism of HLF in treating hyperlipidemia may be related to regulate signifcantly the relative abundance of some bacteria, such as g_Lactobacillus, g_Anaerostipes, g_ [Eubacterium]_hallii_group, g_Fusicatenibacter, g_Akkermansia, g_Collinsella, and other bacteria. Tese dominant bacterial genera altered by HLF showed strong correlations with the hyperlipidemia-related metabolic parameters in HFD-fed rats.

Conclusion
In conclusion, HLF could play a role in maintaining the normal level of blood lipids in a great deal of ways ( Figure 16). We have shown that HLF could improve disorders of lipid metabolism by inhibiting the absorption of intestinal cholesterol and promoting cholesterol excretion. Meanwhile, HLF played an important role in controlling the levels of cholesterol synthesis. In addition, HLF could efectively alleviate oxidative stress and infammatory response induced by hyperlipidemia. Furthermore, HLF could also regulate the relative abundance of gut microbiota, such as g_Lactobacillus, g_Anaerostipes, g_ [Eubacterium]_hallii_group, g_Fusicatenibacter, g_Akkermansia, g_Collinsella, and other bacteria, which may be an efective way to modulate lipid metabolism.

Data Availability
Te data are available from the corresponding author on reasonable request.

Supplementary Materials
Supplementary materials of this article contain the molecular docking diagram of HLF. Autodock tools-1.5.6 software is used to dock the active ingredients and target proteins of HLF, and the conformation with the lowest docking binding energy is selected for docking mode analysis, and PyMOL software is used for molecular docking mapping and display. Pharmacophore analysis was performed using the Ligplot + program to evaluate hydrogen bonds and van der Waals interaction residues. (Supplementary Materials)