Improvement of Endothelial Dysfunction of Berberine in Atherosclerotic Mice and Mechanism Exploring through TMT-Based Proteomics

Atherosclerosis is a multifactorial vascular disease triggered by disordered lipid metabolism, characterized by chronic inflammatory injury, and initiated by endothelial dysfunction. Berberine is the main active alkaloid of the herbal medicine Coptidis Rhizoma (Huanglian). Notably, berberine has been shown to have beneficial effects against atherosclerosis. However, the mechanisms of berberine in preventing atherosclerosis are still unclear. This study is aimed at investigating the effects and mechanisms of berberine in protecting the aorta and ameliorating atherosclerosis in apolipoprotein E-deficient (ApoE−/−) mice. Here, we demonstrated that berberine reduced serum lipid levels, antagonized hepatic lipid accumulation, improved intima-media thickening, and alleviated atherosclerotic lesions in ApoE−/− mice fed a western-type diet for 12 weeks. Meanwhile, berberine reduced aortic reactive oxygen species (ROS) generation and reduced the serum levels of malondialdehyde (MDA), oxidized low-density lipoprotein (ox-LDL), and interleukin-6 (IL-6). In aortic ring assay, berberine restored aortic endothelium-dependent vasodilatation in vivo and in vitro. Furthermore, 4,956 proteins were identified by proteomic analysis, and 199 differentially expressed proteins regulated by berberine were found to be involved in many biological pathways, such as mitochondrial dysfunction, fatty acid β-oxidation I, and FXR/RXR activation. Summarily, these data suggested that berberine ameliorates endothelial dysfunction and protects against atherosclerosis, and thus may be a promising therapeutic candidate for atherosclerosis.


Introduction
Atherosclerosis, a progressive chronic inflammatory disease characterized by lipid deposition, smooth muscle cell proliferation, mononuclear cell infiltration, calcification, and plaque formation, usually occurs in large and mediumsized arteries and is the main cause and pathological basis of cardiovascular disease, which is the leading cause of human death, accounting for 31.5% of all deaths worldwide [1][2][3]. Many risk factors affect atherosclerosis, including gender, smoking, obesity, diabetes, and hypertension, and interaction of multiple risk factors leads to the occurrence and development of atherosclerosis [4]. Although the pathogenesis of atherosclerosis is still not completely clear, it is known to involve disorder of lipid metabolism, endothelial dysfunction, inflammation, and oxidative stress. Endothelial dysfunction is an early key step in the development of atherosclerosis, initiates the progression of atherosclerosis, and is involved in plaque formation. Moreover, the status of endothelial function can potentially reflect the propensity of atherosclerosis occurrence [5,6]. Oxidized low-density lipoprotein (ox-LDL) is a major contributor to endothelial dysfunction in atherosclerosis; its formatting is triggered by enhanced blood lipid levels and oxidative stress, and it is deposited in the vascular endothelium [7]. Inflammation is also involved in endothelial dysfunction, with the activation of inflammatory cells and molecules, including adhesion molecules and chemokines [8]. When the endothelium is injured, it stimulates the expression of inflammatory mediators, including vascular cell adhesion molecule-1 (VCAM-1), intercellular adhesion molecule-1 (ICAM-1), interleukin 6 (IL-6), and tumor necrosis factor α (TNF-α), which activate oxidative stress and the inflammatory response and induce vascular smooth muscle cell proliferation and foam cell formation, leading to plaque formation and eventually atherosclerotic lesions [9,10].
Berberine, an isoquinoline alkaloid, is the main active ingredient of the herbal medicine Coptidis Rhizoma (Huanglian), which has traditionally been used for treating diarrhoea in China and other Asian countries. Recently, increasing in vitro and in vivo studies have found that berberine has significant pharmacological effects on the cardiovascular system, including lowering serum lipid levels [11], lowering blood sugar [12], protecting the vascular endothelium [13], and anti-inflammatory [14] and antioxidant [15] effects. Notably, berberine has shown beneficial effects against atherosclerosis. Berberine was found to suppress atherogenesis through activation of AMPK-dependent UCP2 expression [16] and stabilize atherosclerosis plaques in hyperhomocysteinemia mice via activation of peroxisome proliferator-activated receptor gamma (PPARγ) [17]. Moreover, the atheroprotective activities of berberine could be linked to inhibition of cholesterol accumulation in macrophages [18]. Nevertheless, there is still a lack of understanding of the underlying mechanisms by which berberine protects the aorta and ameliorates atherosclerosis.
With the recent advances in mass spectrometry (MS) technology, MS-based proteomics can rapidly identify and accurately quantify the proteomes of various samples within a single experiment and has been widely used for target identification and mechanistic studies [19,20]. As one proteomics method, Tandem Mass Tag-(TMT-) based liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis can identify and quantify proteins in up to 10 biological samples simultaneously [21]. This method has been an essential tool in quantitative proteomics and can provide an overview of treatment mechanisms.

Animals and Treatment.
Male wild-type C57BL/6 mice (6 to 8 weeks old) and ApoE -/mice (6 to 8 weeks old) on a C57BL/6 genetic background were obtained from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China), and the permission numbers are SCXK (Jing) 2016-0006 and SCXK (Jing) 2016-0011. All animals were housed in standard cages (≤5 mice/cage) with clean sawdust bedding under specific-pathogen-free environmental conditions (20-25°C, 40-60% relative humidity, 12 h light/dark cycle) and had free access to water and food. All procedures were strictly in accordance with the Guidelines for the Care and Use of Laboratory Animals issued by the Ministry of Science and Technology of China and approved by the Laboratory Animal Ethics Committee of Tianjin University of Traditional Chinese Medicine (approval number: TCM-LAEC2016014).
After being acclimatized to the housing environment for one week, fifty two ApoE -/mice fed a western-type diet containing 21% fat and 0.15% cholesterol were randomly divided into four groups: the vehicle group (ApoE -/-, n = 15), the atorvastatin group (ATO, 2.6 mg·kg -1 , n = 15), the berberine 78 mg·kg -1 group (BBR-78 mg·kg -1 , n = 7), and the berberine 156 mg·kg -1 group (BBR-156 mg·kg -1 , n = 15). The wild-type C57BL/6 mice fed a normal chow diet were used as the control group (WT, n = 15). The vehicle and control mice were treated with physiological saline. All the mice were administered treatments by oral gavage for 12 weeks.
2.3. Ultrasonography Analysis. Ultrasonography was performed after 12 weeks of administration using a Vevo 2100 ultra-high-resolution small animal ultrasound imaging system (VisualSonics, Toronto, Canada) equipped with an MS550D transducer operating at 40 MHz. Mice were anaesthetized with 1.5-2.0% inhalant isoflurane in 100% oxygen. After the hair was carefully removed from the chest and neck, the mice were laid on a heated platform in the supine position. First, the parasternal long-axis view was employed to visualize the left ventricle and ascending aorta, and Mmode and PW Doppler mode images were acquired. Then, the scan head was moved to the neck, where the right common carotid artery was visualized using a long-axis view. On B-mode, ultrasound images of the right common carotid artery were obtained. All the cine loops and frames were recorded and stored digitally for image analysis using an off-line workstation Vevo LAB 2.1.0 (VisualSonics, Toronto, Canada).
At the end of the experiment, mice were sacrificed, followed by blood sample, liver, aorta, and heart collection.
2.4. Organ Chamber. The organ chamber experiments were carried out as described previously [23]. Mice were sacrificed by decapitation, and the thoracic aorta was isolated and immediately immersed in oxygenated ice-cold Krebs-Henseleit (K-H) solution (composition in mM: NaCl 118, KCl 4.75, MgSO 4 ·7H 2 O 1.2, KH 2 PO 4 1.2, CaCl 2 2.5, NaHCO 3 25, and glucose 11; pH 7.3-7.4). After careful removal of the adhering perivascular tissues, 2-3 mm rings were cautiously cut. The aortic rings were suspended in an organ bath containing 10 mL K-H solution maintained at 37°C and bubbled with 95% O 2 and 5% CO 2 between two parallel stainless hooks. The isometric tension during the experiments was measured and recorded by a computerassisted data acquisition system (PowerLab/4SP; ADInstruments Pty Ltd., Australia). Each ring was stretched progressively to a baseline tension of 0.8 g and allowed to equilibrate for 60 min prior to the experiment. After the equilibration, the rings were initially contracted with 60 mM KCl to elicit their contractile response, and then, they were rinsed with K-H solution three times to restore tension to the basal level. Subsequently, NE (10 -6 M) was used to precontract the rings. At the plateau of contraction, ACh (10 -9 -10 -4 M) or SNP (10 -9 -10 -5 M) was added cumulatively into the organ bath to induce endothelium-dependent or endothelium-independent relaxation.
2.6. Hepatic Histology. After being photographed and weighed, a piece of liver was fixed in 10% neutral formalin, and the remaining portion was frozen in liquid nitrogen and preserved at -80°C. Liver tissues fixed in formalin were dehydrated with gradient ethanol and embedded in paraffin.
The paraffin-embedded specimens were sliced into 4 μm thick serial sections, which were stained with haematoxylin-eosin (H&E, Wuhan Servicebio Biological Technology Co., Ltd., Wuhan, China) in accordance with the standard protocol. To observe lipid accumulation in the liver, the frozen tissues were embedded in optimum cutting temperature compound and cut into 6 μm thick cryosections. The sections were stained with Oil Red O (Sigma-Aldrich, St. Louis, MO, USA). Images of all sections were obtained with a light microscope (Eclipse CI, Nikon Corporation, Tokyo, Japan).
2.7. Assessment of Aortic Atherosclerosis. Mice were perfused with 0.9% ice-cold saline through the left ventricle after blood collection. The hearts containing the aortic root were removed and fixed in 10% neutral formalin for paraffin sectioning or frozen in liquid nitrogen for cryosectioning. H&E, Oil Red O, and Verhoeff-Van Gieson (VVG, Wuhan Servicebio Biological Technology Co., Ltd., Wuhan, China) staining were performed. All the images were captured with a light microscope. Quantification of positive Oil Red O staining was performed using Image-Pro Plus 6.0 software (Media Cybernetics Inc., USA).

Immunofluorescence and TUNEL Staining.
To evaluate vascular endothelial function, the expression of eNOS and ET-1 in the aortic root was determined by immunofluorescence staining. CD31 and terminal deoxynucleotidyl transferase-mediated dUTP nick-end labelling (TUNEL) double immunofluorescence staining was performed to measure endothelial apoptosis. The cryosections of aortic sinuses were permeabilized with 0.2-0.5% Triton X-100 for 10 min. The primary antibodies anti-eNOS, anti-ET-1, and anti-CD31 were diluted (1 : 200) and incubated overnight at 4°C with the sections. After washing in PBS (3 × 3 min), the sections were incubated with secondary antibodies (donkey anti-rat IgG (ab150154, Abcam) and donkey anti-rabbit IgG (ab150073, Abcam, Cambridge, UK), 1 : 200 dilution) for 1 h at room temperature. Thereafter, TUNEL costaining was performed using an In Situ Cell Death Detection Kit (11684795910, Roche, Germany) in accordance with the manufacturer's protocol for double fluorescence staining, and the cell nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) (H1200, Vector Laboratories, CA, USA) for all sections. Fluorescence staining was visualized under a fluorescence microscope (DFC500, Leica, Wetzlar, Germany).  Oxidative Medicine and Cellular Longevity samples were centrifuged, and the supernatant was collected. The protein concentration was quantified using the Bradford method [24]. After quantification, the protein extracted from each sample was digested with 0.5 μg·μL -1 trypsin solution (Thermo Fisher Scientific, NJ, USA). Then, TMT labelling was performed with a 10-plex TMT kit (Thermo Fisher Scientific, NJ, USA). Peptides in each group were labelled with different TMT labels: three biological repeats of the wild-type mice group were labelled TMT-127N, TMT-127C, and TMT-128N; three biological repeats of the ApoE -/group were labelled TMT-128C, TMT-129N, and TMT-129C; and three biological repeats of the berberine 156 mg·kg -1 group were labelled TMT-130N, TMT-130C, and TMT-131. All the labelled samples were mixed, vacuum dried, and stored at -20°C for further analysis. LC-MS/MS analysis was performed with an EASY-nLC 1000 System (Nano HPLC, Thermo Fisher Scientific, USA) coupled online to a Q Exactive Mass Spectrometer with an EASY-Spray Ion Source (Thermo Fisher Scientific, USA). The samples were loaded onto an Acclaim PepMap 100 precolumn (2 cm × 100 μm, 5 μm, C18, Thermo Fisher Scientific, USA). Peptide separation was conducted using an EASY-Spray column (12 cm × 75 μm, 3 μm, C18, Thermo Fisher Scientific, USA) with buffer A (100% ultrapure water and 0.1% formic acid) and buffer B (100% acetonitrile and 0.1% formic acid) at a flow rate of 350 nL·min -1 . The eluted peptides were analysed using the Q Exactive online system. MS data acquisition was performed using a data-dependent top 20 method.

Protein Identification and Quantitative Analysis.
The LC-MS/MS raw data were searched against the mouse FASTA database from UniProt (Uniprot_mouse_54106_ 20170101.fasta) using Proteome Discoverer version 2.1 (Thermo Fisher Scientific, USA). The false discovery rate (FDR) was set as <0.01. Detection of at least one unique peptide per protein was set as the requirement for protein identification. The protein quantitative analysis was based on reporter ion peak intensity. Differentially expressed proteins were identified through ratio-fold change as well as P value calculated with a t-test. The up-and downregulation thresholds were set at an average ratio − fold change > 1:2 with a P value < 0.05 and an average ratio − fold change < 0:83 with a P value < 0.05; the proteins that satisfied the set thresholds were collected as "Differentially expressed proteins" for further bioinformatics analysis.

Bioinformatics
Analysis. Functional annotations of differentially expressed proteins were performed through Gene Ontology (GO) enrichment analysis (http://www .geneontology.org/). In addition, Ingenuity Pathway Analysis (IPA) software (QIAGEN, CA, USA, http://www.ingenuity .com) was used to analyse canonical pathways, diseases and functions, and biological networks associated with the lists of differentially expressed proteins. The dataset containing differentially expressed proteins and corresponding expression values was uploaded into the IPA, and then, IPA "Core Analysis" was performed.
2.11. Statistical Analysis. All the experimental data are expressed as the mean ± SEM. Statistical analyses were performed with SPSS software (version 17.0, SPSS Inc., Chicago, IL, USA). One-way ANOVA with Tukey's post hoc test or a Kruskal-Wallis test followed by Dunn's multiple comparison was used for comparisons among multiple groups. Student's t-test was used for comparisons of two groups. Differences with P < 0:05 were considered statistically significant. GraphPad Prism 5.0 (GraphPad Software Inc., La Jolla, CA, US) was used for the graphical work.

Berberine Reduced Serum FFA and TG Levels and
Antagonized Hepatic Lipid Accumulation. ApoE -/mice fed a western-type diet for 12 weeks were used as the atherosclerosis model, and berberine was administered by oral gavage. Throughout the course of the experiments, mouse body weights were recorded weekly (Figure 1(a)), and the weights of all the mice increased compared with that before the experiment. After 12 weeks of administration, the weight of ApoE -/mice increased significantly, and 156 mg·kg -1 berberine dramatically inhibited the weight increase (Figure 1(b)). With weight increase, the hair of ApoE -/mice appeared dim and depilated, which could be improved by berberine (Figure 1(c)). To confirm the hypolipidaemic effect of berberine, serum lipid profiles were determined. 4 Oxidative Medicine and Cellular Longevity

Oxidative Medicine and Cellular Longevity
Compared with wild-type mice, in ApoE -/mice, the FFA, TG, TC, and LDL-C serum levels were significantly increased and the HDL-C level was significantly decreased by the western-type diet. Berberine (156 mg·kg -1 ) lowered the FFA (Figure 1(d)) and TG (Figure 1(e)) levels but had no significant effect on TC (Figure 1(f)), HDL-C (Figure 1(g)), and LDL-C (Figure 1(h)).
Compared with wild-type mice, the liver index of ApoE -/mice increased (Figure 1(i)) and the liver colour was reduced (Figure 1(j)), implying hepatic lipid accumulation. Berberine (78 and 156 mg·kg -1 ) blocked the liver colour change and reduced the liver index. In Figure 1(k), H&E staining showed slight inflammatory cell infiltration and vacuoles in hepatocytes of ApoE -/mice, indicating steatosis of the    (Figure 2(a)). The intima-media thickness (IMT) of the carotid artery in ApoE -/mice increased significantly, and 156 mg·kg -1 berberine treatment for 12 weeks improved the IMT thickening, but there was no difference in the carotid artery inner diameter in mice of each group (Figure 2(b)). To further determine the antiatherosclerotic effect of berberine, we performed Oil Red O, H&E, and VVG staining of the aortic root. As shown in Figure 2  Oxidative Medicine and Cellular Longevity 9 Oxidative Medicine and Cellular Longevity apparent lipid accumulation and plaque formation were present in the aortic root of ApoE -/mice fed the western-type diet, whereas no lipid plaque accumulation was found in wild-type mice. Administration of berberine or atorvastatin showed a tendency to reduce the plaque area. As shown in Figure 2(d) (H&E staining), compared with wild-type mice, aortic intimal hyperplasia, lipid deposition, and plaque formation were observed in the ApoE -/mice, accompanied by inflammatory cell infiltration. The sinus lesions were improved in berberine-treated ApoE -/mice, demonstrating the antiatherosclerotic effect of berberine. Figure 2(e) (VVG staining) shows arterial lesions in ApoE -/mice, displaying an apparent disruption, decrease, and even loss of elastin fibres. However, berberine prevented these pathologies and increased the elastin content in the arterial wall.

Berberine Improved Endothelial Dysfunction in ApoE -/-
Mice. Thoracic aortic rings were isolated to investigate whether berberine can protect against endothelial dysfunction in ApoE -/mice fed a high-fat diet. Endothelial function was determined by measuring endothelium-dependent vasodilatation in response to ACh. As shown in Figure 3(a), after 12 weeks of administration, endothelium-dependent vasodilatation induced by ACh was significantly impaired in ApoE -/mice compared with wild-type mice. Importantly, the impairment of ACh-induced vasodilatation was restored by administration of berberine. However, there was no difference in SNP-induced endothelial-independent vasodilatation between the groups (Figure 3(b)), demonstrating that the protective effect of berberine on the aorta involves maintaining endothelial function. Similarly, in vitro experiments showed that aortic endothelium-dependent relaxation was strikingly preserved by berberine in ApoE -/mice (Figure 3(c)), and endothelium-independent vasodilatation was not altered in any of the groups (Figure 3(d)), indicating a significant improvement in endothelial function by berberine.
To further evaluate vascular endothelial function, aortic root eNOS, ET-1 immunofluorescence, and TUNEL staining were performed. Representative images are shown in Figures 3(e) and 3(f). Downregulated expression of eNOS (green) and overexpression of ET-1 (red) in the aortic root were observed in ApoE -/mice, and berberine treatment reversed the expression of eNOS and ET-1 in part (Figure 3(e)). Figure 3(f) shows increased apoptotic cells (green) in the sinus lesion area in ApoE -/mice compared with the wild-type mice, but berberine administration reduced the number of TUNEL-positive cells. Together, these data indicate the protective effect of berberine on the aorta of ApoE -/mice.

Berberine Attenuated Aortic Oxidative Stress in ApoE -/-
Mice. DHE staining was performed to detect endothelial ROS generation to evaluate the effect of berberine on oxidative stress. Compared with wild-type mice, DHE fluorescence intensity increased in ApoE -/mice, indicating that the total aortic ROS formation increased, whereas berberine treatment reduced endothelial fluorescence (Figure 4(a)). Serum levels of the oxidative stress products MDA and ox-LDL were remarkably higher in ApoE -/mice than in wild-type mice. After treatment with berberine or

12
Oxidative Medicine and Cellular Longevity atorvastatin, the alterations in MDA and ox-LDL levels were significantly inhibited (Figures 4(b) and 4(c)). Furthermore, administration of berberine decreased the level of the inflammatory cytokine IL-6 in ApoE -/mice (Figure 4(d)).

TMT Protein Profiling.
To obtain a global overview of proteins regulated by berberine and better understand the mechanism by which berberine protects the aorta of ApoE -/mice, a proteomic profiling analysis based on the TMT technique was carried out. The workflow of this study is shown in Figure 5(a). Through LC-MS/MS analysis, a total of 59,751 spectra were obtained, corresponding to 31,414 peptides. In total, 4,956 proteins (unique peptides ≥ 1) were identified from 28,957 unique peptides ( Figure 5(b)), and 4,841 proteins were successfully quantified (Supplementary Table 1). The proteins that exhibited a P value < 0.05 and a ratio − fold change > 1:2 or <0.83 were defined as differentially expressed proteins. Compared with wild-type mice, 871 upregulated proteins and 385 downregulated proteins were identified in ApoE -/mice fed the westerntype diet, whereas berberine treatment upregulated 132 proteins and downregulated 222 proteins ( Figure 5(c)). In Figure 5(d), the Venn diagram shows the distribution and overlap of differentially expressed proteins among the two comparison groups. There were 199 common differentially expressed proteins among these comparison groups, including 67 upregulated proteins and 132 downregulated proteins. The expression of these common differentially expressed proteins was clustered through hierarchical clustering ( Figure 5(e)), and the top 20 are shown in Table 1.

Bioinformatics Analysis of Differentially Expressed
Proteins and Western Blotting Validation. To clarify the biological significance of the 199 differentially expressed proteins, we first performed GO enrichment analysis to annotate the functions of these proteins according to cellular component, molecular function and biological process. 628 GO terms with P < 0:05 were significantly enriched, 79 in cellular component, 160 in molecular function, and 389 in biological process. The data of GO analysis are shown in Supplementary Table 2 and the top 20 terms of each category are shown in Figure 6. From the cellular component enrichment terms, it can be noted that the differentially expressed proteins are likely primarily located in mitochondrion (Figure 6(a)). In molecular function enrichment, high-density lipoprotein particle binding was the top enriched term, while receptor binding contained the most differentially expressed proteins (Figure 6(b)). Regarding biological process, the differentially expressed proteins were remarkably enriched in fatty acid betaoxidation, which might play important roles in berberine protection of the aorta in atherosclerosis (Figure 6(c)). Next, IPA was performed for further biological function investigation. Figure 7 (Figure 7(b)), in which NDUFA4 (NDUFA4 mitochondrial complex associated) and Mitochondrial complex 1 are the core proteins, is primarily involved in metabolic disease, developmental disorder, and hereditary disorder. Network 2 (Figure 7(c)) mediates biological processes, such as energy production, lipid metabolism, and small molecule biochemistry, and ACOX1 (acyl-CoA oxidase 1), CD36, and EHHADH (enoyl-CoA hydratase and 3hydroxyacyl CoA dehydrogenase) are the core proteins. In Network 3 (Figure 7(d)), APOA1 may play a key role. This network is primarily related to lipid metabolism, small molecule biochemistry, and vitamin and mineral metabolism. IPA was also adopted for grouping of differentially expressed proteins into three categories: "Diseases and Disorders," "Molecular and Cell Functions," and "Physiological System Development and Function" (Supplementary Figure 2). In "Diseases and Disorders," metabolic disease was found to be the top ranked, and cardiovascular disease was one of the related diseases (Supplementary Figure 2a). The molecular profiles of the top 4 categories in metabolic disease and cardiovascular disease are shown in Supplementary Figure 3.
Three proteins, including PON1, APOA1, and GPD2, involved in atherosclerosis, cardiovascular disease, and lipid metabolism, were selected for validation via western blotting. The expression levels of PON1 and APOA1 were found to be remarkably downregulated in ApoE -/mice compared with wild-type mice, whereas GPD2 showed significant upregulation. Berberine treatment upregulated the expression of PON1 and APOA1 and downregulated the expression of GPD2 (Figure 7(e)). These results were consistent with the results observed in the TMT-based proteomics analysis.

Discussion
Hyperlipidaemia, characterized by dyslipidaemia, including elevated blood levels of TG, TC, and LDL-C and a decreased HDL-C level, is a major risk factor for atherosclerosis [25]. Regulation of disorder of lipid metabolism is one method used to control atherosclerosis development. Our results showed that berberine treatment significantly lowered the TG and FFA levels in western-type diet-induced ApoE -/mice, indicating the lipid-lowering effect of berberine. In addition, the liver is a crucial organ of lipid metabolism and is responsible for maintaining the homeostasis of cholesterol and LDL [26]. We observed hepatic lipid accumulation. As shown in Figures 1(i)-1(k)

16
Oxidative Medicine and Cellular Longevity steatosis and reduced liver lipid accumulation, which indicates that berberine may have a potential therapeutic effect on fatty liver and further investigation is needed. Taken together, these data suggest that the protective effect on the aorta and the antiatherosclerotic effect of berberine might be partially due to its regulation of lipid metabolism. Endothelial dysfunction is an early step in atherosclerosis [6] and an independent predictor of future cardiovascular events [27]. Thus, targeting endothelial dysfunction represents an attractive pharmacological approach [28]. Impaired endothelium-dependent vasodilatation is the hallmark of endothelial dysfunction [29]. Improving endotheliumdependent vasodilatation and restoring endothelial function in patients with atherosclerosis may prevent cardiovascular events [30]. In addition, preventing impairment of endothelium-dependent relaxation was found to attenuate atherosclerosis in ApoE -/mice [31]. Our results revealed impaired endothelium-dependent vasodilatation induced by ACh in ApoE -/mice fed a western-type diet, suggesting endothelial dysfunction in ApoE -/mice. However, berberine intervention significantly restored endothelium-dependent vasodilatation in vivo and in vitro, indicating the protective effect of berberine on endothelial function. Many factors, including dyslipidaemia, oxidative stress, and inflammation, are involved in the progression of endothelial dysfunction. Increased FFA concentration in plasma results in an increase in ROS generation [32], while elevated circulating LDL deposits in the subendothelial matrix and is oxidized to form ox-LDL through increased ROS production. Ox-LDL injures endothelial cells directly and contributes to endothelial dysfunction via overexpression of lectin-like oxidized low-density lipoprotein receptor-1 (LOX-1), which induces a further rise in intracellular ROS [33][34][35]. ROS oxidize tetrahydrobiopterin, leading to eNOS uncoupling [36], and react with nitric oxide (NO) to form peroxynitrite (ONOO − ), thus reducing NO generation and bioavailability and accelerating endothelial dysfunction [37]. On the other hand, ROS are able to trigger activation of the transcription factor nuclear factor kappa B (NF-κB) [38], which can activate a variety of proinflammatory cytokines, such as IL-6, IL-8, E-selectin, ICAM-1, and VCAM-1 [39,40], thereby activating inflammation. In turn, IL-6 induces oxidative stress and endothelial dysfunction through overexpression of angiotensin II type 1 receptor in the atherosclerosis process [41]. A vicious circle exists between oxidative stress and inflammation in atherosclerosis [42].
In the current study, decreased ROS generation in the aortic root and lower serum levels of MDA, ox-LDL, and IL-6 were detected in the berberine-treated ApoE -/mice, demonstrating the anti-inflammatory and antioxidative properties of berberine, which may be the mechanism by which berberine improves endothelial dysfunction. Carotid IMT is a surrogate marker for atherosclerosis [43]. An increase in carotid IMT is related to endothelial dysfunction, and both of them, respectively, assess the atherosclerotic process from different aspects of anatomy and function [44]. Ultrasonography is a noninvasive approach based on high-resolution scanning and a computerized image-analysis system and has been a powerful tool to assess vascular atherosclerosis in mice. In the present study, we used long-axis views obtained with an ultrasound imaging system to measure carotid IMT. The ultrasound results showed that berberine significantly reduced IMT in the carotid artery, indicating that berberine could improve arterial wall lesions. We further investigated the pathological changes in the aortic root in ApoE -/mice, which exhibited an obvious accumulation of lipid, plaque formation, and severe atherosclerotic lesions, suggesting successful establishment of the atherosclerosis model. Treatment with berberine alleviated the severity of the sinus atherosclerotic lesions in ApoE -/mice. These results demonstrate that berberine treatment inhibits atherosclerosis development.
To further reveal the mechanism of berberine in ameliorating atherosclerosis and protecting the aorta of ApoE -/-mice, a TMT-based quantitative proteomics analysis was performed to globally screen berberine-regulated proteins, followed by bioinformatics analysis and validation. Based on the data from TMT protein profiling and GO analysis, we found 199 proteins regulated by berberine in the treatment of atherosclerosis, which are likely primarily located in mitochondrion and participate in the biological process of fatty acid beta-oxidation (Figure 6), indicating that berberine might protect the aorta by regulating mitochondrial function and lipid metabolism. IPA canonical pathways analysis showed that mitochondrial dysfunction (−log ðP valueÞ = 8:66) was the most significant pathway related to the differentially expressed proteins (Figure 7(a)). Mitochondria contain a multitude of redox carriers and are the main sites of aerobic respiration, producing adenosine triphosphate (ATP) via oxidative phosphorylation; thus, mitochondria are known as the "power house" of cells. In addition to producing energy, mitochondria are involved in regulating many other cellular processes, such as calcium signalling [45], cell death [46], and inflammation [47]. Moreover, mitochondria are an important source of ROS, approximately 90% of which can be traced back to mitochondria [48]. Under the stimulation of risk factors, such as hypercholesterolemia, hyperlipidaemia, and hypoxia, the mitochondrial membrane potential can be altered, triggering excessive ROS production in the mitochondria, which results in mitochondrial damage and dysfunction [49]. In turn, mitochondrial dysfunction leads to more ROS production, and the enhanced ROS production induces endothelial dysfunction by uncoupling eNOS and triggering inflammation. Mitochondrial dysfunction is present in human and mouse atherosclerosis and promotes the development of atherosclerosis by affecting endothelial function, vascular smooth muscle cell proliferation or apoptosis, and macrophage polarization [50,51]. As a consequence, mitochondria-targeted intervention may provide a novel therapeutic strategy for atherosclerosis. It has been found that chlorogenic acid can mitigate ox-LDLinduced endothelial oxidative stress and mitochondrial dysfunction by modulating the SIRT1/AMPK/PGC-1 pathway, thus providing beneficial effects in atherosclerosis [52]. Our results showed that the differentially expressed proteins regulated by berberine were most related to mitochondrial dysfunction, suggesting that regulation of mitochondrial dysfunction may be one of the mechanisms by which berberine protects the atherosclerotic aorta.
The IPA network analysis showed that APOA1 accounted for a large proportion of Network 3 and was upregulated by berberine (Figure 7(d)). APOA1 is the major protein component of HDL, which determines the levels of HDL in the plasma [53]. Elevated APOA1 levels are strongly associated with a reduced risk of cardiovascular events in patients [54], and a damaging mutation in APOA1 confers an increased risk of atherosclerosis [55]. Animal studies have also shown that APOA1 inhibits the progression of atherosclerosis. Somatic gene transfer of human APOA1 reduced atherosclerosis progression in human APOA1-transgenic mice and ApoE -/mice [56]. Infusions of synthetic HDL containing trimeric human APOA1 stabilized atherosclerotic plaques in hypercholesterolemic rabbits [57]. Guo et al. 18 Oxidative Medicine and Cellular Longevity found that perhexiline reduced atherosclerosis by regulating APOA1 transcription and increasing the APOA1 level in mice [58]. Consequently, increasing APOA1 synthesis may become a new approach for treating atherosclerosis. Herein, APOA1 expression was downregulated in ApoE -/mice, and the expression downregulation was reversed by berberine. The "Diseases and Disorders" analysis in the IPA showed that APOA1 was associated with disorder of lipid metabolism, coronary disease, acute coronary syndrome, and occlusion of blood vessels (Supplementary Figure 3). These results indicate that berberine has a regulatory effect on APOA1, which might be a potential target of berberine.

Conclusions
In summary (Figure 8), our study demonstrated that berberine can ameliorate endothelial dysfunction to protect against atherosclerosis by restoring endothelium-dependent vasodilatation, attenuating oxidative stress and inflammation, regulating lipid metabolism, and alleviating atherosclerotic lesions. Regulation of mitochondrial dysfunction and targeting of APOA1 might be potential mechanisms underlying this protection. The exact mechanisms remain to be further studied and characterized, but together, the findings presented here provide novel insight into the molecular antiatherosclerosis mechanism of berberine and evidence to support the potential of berberine as a promising therapeutic candidate for atherosclerosis. In atherosclerotic mice, berberine regulated lipid metabolism, alleviated atherosclerotic lesions, attenuated oxidative stress and inflammation, and ameliorated endothelial dysfunction, thus restoring endothelium-dependent vasodilatation to protect against atherosclerosis. Regulation of mitochondrial dysfunction might be potential mechanisms underlying this protection. 19 Oxidative Medicine and Cellular Longevity PON1:

Data Availability
The data used to support the findings of this study are included within the article.

Conflicts of Interest
The authors declare no conflict of interest.