Dihydromyricetin increases endothelial nitric oxide production and inhibits atherosclerosis through microRNA‐21 in apolipoprotein E‐deficient mice

Abstract Natural products were extracted from traditional Chinese herbal emerging as potential therapeutic drugs for treating cardiovascular diseases. This study examines the role and underlying mechanism of dihydromyricetin (DMY), a natural compound extracted from Ampelopsis grossedentata, in atherosclerosis. DMY treatment significantly inhibits atherosclerotic lesion formation, proinflammatory gene expression and the influx of lesional macrophages and CD4‐positive T cells in the vessel wall and hepatic inflammation, whereas increases nitric oxide (NO) production and improves lipid metabolism in apolipoprotein E‐deficient (Apoe− / −) mice. Yet, those protective effects are abrogated by using NOS inhibitor L‐NAME in Apoe− / − mice received DMY. Mechanistically, DMY decreases microRNA‐21 (miR‐21) and increases its target gene dimethylarginine dimethylaminohydrolase‐1 (DDAH1) expression, an effect that reduces asymmetric aimethlarginine (ADMA) levels, and increases endothelial NO synthase (eNOS) phosphorylation and NO production in cultured HUVECs, vascular endothelium of atherosclerotic lesions and liver. In contrast, systemic delivery of miR‐21 in Apoe− / − mice or miR‐21 overexpression in cultured HUVECs abrogates those DMY‐mediated protective effects. These data demonstrate that endothelial miR‐21‐inhibited DDAH1‐ADMA‐eNOS‐NO pathway promotes the pathogenesis of atherosclerosis which can be rescued by DMY. Thus, DMY may represent a potential therapeutic adjuvant in atherosclerosis management.


| INTRODUC TI ON
Endothelial cell (EC) activation and dysfunction is the initial step that plays a critical role in the pathogenesis of atherosclerosis. [1][2][3] For example, activated ECs express multiple adhesion molecules, such as vascular cell adhesion molecule-1 (VCAM-1) and intercellular adhesion molecule-1 (ICAM-1), to facilitate the circulating leukocytes into the vascular wall, where they differentiate into macrophages and eventually become foam cells. 1,4,5 Accumulating studies demonstrate that reduced nitric oxide (NO) production, caused by impaired endothelial NO synthase (eNOS) function, is a hallmark of EC dysfunction that contributes importantly to atherosclerosis. 3,6 Asymmetrical dimethylarginine (ADMA), mainly metabolized by dimethylarginine dimethylaminohydrolase-1 (DDAH1), is the foremost endogenous NOS inhibitor through directly competing with the physiological precursor L-arginine from the enzyme. 7,8 Recently, the DDAH1-ADMA-eNOS pathway has emerged as an important regulator in mediating NO production and atherosclerosis. The elevated levels of ADMA are positive correlated with all established cardiovascular risk factors and acted as a biomarker and/or inducer of endothelial dysfunction. 9 Overexpression of DDAH1 increases NO production and improves endothelial function through an ADMA-dependent manner. 10 Moreover, genetic overexpression of DDAH1 reduces atherosclerotic plaque formation in apolipoprotein E-deficient (Apoe −/− ) mice by lowering ADMA levels and improving endothelial function. 11 Thus, targeting DDAH1-ADMA-mediated eNOS-NO activation and associated endothelial function holds great potential for developing novel therapeutic approaches for atherosclerosis.
Accumulating evidence demonstrates that traditional Chinese herbal medicines are emerging as promising therapeutic drugs for treating cardiovascular diseases. For instance, Ampelopsis grossedentata, also known as vine tea and widely distributed in southern China, has been consumed as health tea and herbal medicine for hundreds of years. 12 Dihydromyricetin (DMY) is the most abundant and bioactive flavanonol compound that can be robust extracted from the stems and leaves of A grossedentata using response surface methodology. 13 Recent studies demonstrated that DMY attenuates both pressure overload-and angiotensin II-induced cardiac hypertrophy through ameliorating oxidative stress reaction. 14,15 In addition, in a mouse model of acute myocardial infarction, DMY reducesischemia/ reperfusion-induced cardiomyocytes apoptosis, resulting in less infarct area and the improvement of cardiac function. 16 Moreover, DMY increases glucose uptake in skeletal muscle, thereby improving insulin resistance, a major risk factor in the development of cardiovascular diseases. 17 Though it has been reported that DMY ameliorates atherosclerosis, 18,19 the signals and molecular mechanisms of how DMY attenuates endothelial function, vascular inflammation and atherosclerosis are largely unknown.
MicroRNAs (miRNAs), such as miR-21, have emerged as important regulators of endothelial activation and dysfunction that contribute importantly to the development of atherosclerosis. [20][21][22] Our previous studies identified that miR-21 was increased in ECs in response to tumour necrosis factor alpha (TNF-α) and 4-hydroxynonenal (4-HNE) stimulation and miR-21-mediated DDAH1-ADMA-eNOS activation plays a critical role in mediating DMY's protective effects on TNF-α-induced endothelial dysfunction. 20,23 In this study, we found that DMY decreases miR-21 expression, improves EC function and thereby inhibits vascular inflammation, lipid metabolism and atherosclerosis in Apoe −/− mice. We identified an important role of endothelial miR-21-DDAH1-ADMA-eNOS-NO signalling in DMYameliorated atherosclerotic lesion formation, indicating that DMY supplementation may serve as a potential therapeutic adjuvant for treating atherosclerosis.

| Animal studies
All animal procedures were approved by the Institutional Animal Care and Use Committee at Second Xiangya Hospital, Central South University. Male Apoe −/− mice and C57BL/6J mice from 8-to 10-week-old were purchased from the Beijing Vital River Laboratory Animal Technology Co. in China. All mice were maintained on a gene dimethylarginine dimethylaminohydrolase-1 (DDAH1) expression, an effect that reduces asymmetric aimethlarginine (ADMA) levels, and increases endothelial NO synthase (eNOS) phosphorylation and NO production in cultured HUVECs, vascular endothelium of atherosclerotic lesions and liver. In contrast, systemic delivery of miR-21 in Apoe −/− mice or miR-21 overexpression in cultured HUVECs abrogates those DMY-mediated protective effects. These data demonstrate that endothelial miR-21-inhibited DDAH1-ADMA-eNOS-NO pathway promotes the pathogenesis of atherosclerosis which can be rescued by DMY. Thus, DMY may represent a potential therapeutic adjuvant in atherosclerosis management.

K E Y W O R D S
atherosclerosis, dihydromyricetin, endothelial cell, microRNA, nitric oxide 12-hour light/dark cycle in a pathogen-free animal facility. Mice were kept on a standard chow diet or on a 1.25% high cholesterol diet (HCD; D12108C, Research Diets) for 12 weeks. Mice had free access to food and water. For DMY intervention study, Apoe −/− mice were administered daily an intragastric gavage with DMY (500 mg/kg; D101549, Aladdin; n = 8), DMY puls L-NAME (50 mg/ kg; N5751, Sigma; n = 7) or same dosage of solution control (n = 10).
For in vivo systemic overexpression of miR-21 efficient assay experiment, C57BL/6N mice were treated with miR-21 mimics (21-m; miR10000076-1-5, RiboBio) or miRNA non-specific control (NSm; miR1N0000001-1-5, RiboBio) for two consecutive days (once a day, 20 nmol/injection, iv) and harvested after 7 days (n = 3 for each group). For in vivo miR-21 accumulation assay, 8-week-old male Apoe −/− mice were kept on a HCD for 4 weeks followed by tail vein injection of FITC-labelled or unlabelled miR-21 mimic (20 nmol/injection, iv) and harvested 4 hours after injection. For miR-21 intervention study, Apoe −/− mice were kept on a HCD and daily intragastric gavage with DMY (500 mg/kg) for 12 weeks. Eight weeks after HCD, mice were tail vein injected with 21-m or NS-m for two consecutive days and then followed by once a week for 3 weeks (20 nmol/ injection; n = 7 for each group). Systemic delivery of miRNA was performed according to the established protocol described in Ref. 24. Briefly, 20 nmol 21-m or NS-m was dissolved in 100 μL dPBS (solution 1). Lipofectamine 2000 (30 μL; 11668019, Invitrogen) was mixed with 70 μL dPBS by pipetting up and down (solution 2), and placed at room temperature for 15 minutes. Then, solution 1 and solution 2 were mixed by pipetting up and down. After incubating at room temperature for 30 minutes, the mixture (200 μL) was injected into mice by tail vein injection. All mice in the current study were randomly assigned to groups. After 12 weeks, mice were humanely killed, followed by cardiac puncture blood collection, and aortic root and liver were harvested. Aortic roots were embedded in optimum cutting temperature (OCT) compound and frozen at −80°C, while part of liver was fixed in 4% paraformaldehyde (PFA) and the rest were frozen at −80°C for further experiments.

| Atherosclerotic lesions characterization and immunohistological analysis
Serial cryostat sections (6 μm) from OCT-embedded aortic roots were prepared using a Lab-Tek tissue processor Leica CM1950.
Paraffin sections (6 μm) were prepared from 4% PFA fixed liver tissues. To determine atherosclerotic lesion size, aortic root sections and the descending thoracoabdominal aorta were stained with oil red O.
For immunohistological analysis, serial cryostat sections from aortic root were fixed and permeabilized with cold-acetone for 5 minutes and blocked in PBS containing 10% normal goat serum for 1 hour at room temperature. Paraffin sections from liver were deparaffinized, and antigen retrieval was performed using citrate buffer purchased from Abcam (ab93678). Then, sections were stained for macrophages (anti-Mac-2, 1:100, CL8942AP, Cedarlane), T cells (anti-CD4, 1:100, 553043, BD Pharmingen) antibodies for 1.5 hours, followed by appropriated biotin-conjugated secondary antibodies

| Intima RNA isolation from aorta tissue
Intima RNA isolation from aorta was performed according to established protocol described in a previous study. 24 Briefly, aortas were gently flushed with ice-cold PBS, followed by intima peeling using TRIzol reagent (15596018, Invitrogen). TRIzol was carefully flushed for 10, and 10 seconds pause, followed by another 10 seconds flushed and collected in an RNA-free Eppendorf tube and snapfrozen in liquid nitrogen.

| Plasma RNA isolation
Total RNA isolated from plasma was isolated using total RNA purification kit (37500, Norgen Biotek) according to the manufacturer's instruction. During the RNA isolation procedure, miR-39-3p (miRB0000010-3-1) was added as an endogenic control.

| Lipid profile analysis
The plasma levels of total cholesterol (A111-1-1; Nanjing Jiancheng were measured using commercial colorimetric enzymatic assay kits according to the manufacturers' recommendation.

| Plasma and liver ADMA and NO levels measurement
Same volume of plasma samples were used for measuring AMDA (E-EL-0042c; Elabscience) and NO (S0023; Beyotime) levels from mice using the relevant commercial ELISA kits according to manufacturers' instruction. To detect AMDA and NO concentrations in liver, equal amount of protein from each mouse were measured by using the relevant AMDA and NO kit. Ref. [20]. Then transfected cells were treated with ox-LDL, or ox-LDL plus DMY as described above.

| Cell adhesion assay
Cell adhesion assay was performed as previously described in Ref. [25]. Briefly, HUVECs were plated on a 96-well fluorescence plate and treated as described in the cell culture methods part. THP-1 cells were washed with serum-free RPMI 1640 medium and sus-

| Measurement of intracellular NO production
The production of NO in HUVECs was examined using commercial DAF-FM DA kit (S0019; Beyotime) according to the manufacturers' instruction. Briefly, HUVECs after treatment were washed three times in PBS and then incubated with DAF-FM DA (5 µmol/L) in PBS for 20 minutes at 37°C. After incubation, cells were washed three times in PBS and images were captured by a microscopy (Carl Zeiss).
The mean fluorescence intensity per view was used to show the NO production in ECs and quantified from randomly acquired images.

| Protein extraction and immunoblot
Tissues were homogenized using TissueLyser II (QIAGEN) according to the manufacture's instruction. Homogenized liver tissue and cultured HUVECs were isolated RIPA buffer supplemented with protease inhibitor (CW2200, CWBIO) and phosphorylase inhibitor (CW2383, CWBIO). Lysates were separated by 10% SDS-PAGE gels, transferred to PVDF membranes (IPVH00010; Millipore) and blocked in 5% non-fat milk in TBST for 1 hour at room temperature.  normally distributed continuous variables. For multiple groups comparison, one-way analysis of variance (ANOVA) followed by Tukey multiple comparison analysis were used. For data without normal distribution, non-parametric Mann-Whitney U test or Kruskal-Wallis test was used. All data were expressed as mean ± SEM P < .05 was considered significant for all tests.

| DMY inhibits atherosclerosis by increasing NO production and improving endothelial function in Apoe −/− mice
Nitric oxide-dependent endothelial dysfunction contributes importantly to the pathogenesis of atherosclerosis. 3,6 We previously found that DMY-ameliorated TNF-α-induced dysfunction of HUVECs by increasing NO generation. 20 To investigate whether DMY attenuates atherosclerosis through NO production, Apoe −/− mice were fed a HCD for 12 weeks and administered daily intragastric gavage with vehicle, DMY or DMY plus NOS inhibitor L-NAME, respectively. As shown in Figure 1A  Taken together, these data indicate that DMY treatment improves endothelial function, hepatic inflammation and lipid metabolism by increasing NO production.

| DMY increases endothelial NO generation through increasing DDAH1-ADMA-eNOS pathway activation
In ECs, NO is synthesized directly by eNOS through metabolizing substrate L-arginine. During this process, DDAH1 plays an important role in controlling eNOS activation by catalyzing the endogenous eNOS inhibitor ADMA. 29 Thus, we next investigated whether  Figure 3D,). Finally, in HUVECs exposed to ox-LDL, DMY F I G U R E 2 Dihydromyricetin (DMY) attenuates plasma lipid levels and hepatic endothelial activation and inflammation depending on nitric oxide (NO) production in Apoe −/− mice. A, ELISA analysis of NO levels in livers from vehicle control, DMY or DMY combined with L-NAME-treated Apoe −/− mice fed with HFD for 12 wk. B, Circulating lipid levels (total cholesterol, triglycerides, LDL-C, HDL) in HFD-fed Apoe −/− mice treated with vehicle control, DMY or DMY combined with L-NAME after 12 wk. C, Representative images and quantification show lipid accumulation in livers from vehicle control, DMY or DMY combined with L-NAME-treated Apoe −/− mice fed with HFD for 12 wk. Scale: 100 μm. D, Representative images and quantification show Mac-2-positive macrophages in livers. Scale: 100 μm. E, Real-time qPCR analysis of indicated macrophage M1 markers in liver. The expression of genes was normalized to mouse β-actin. F, Western blot analysis of VCAM-1 expression in liver. G, Real-time qPCR analysis of indicated endothelial cells activated markers in liver. The expression of genes was normalized to mouse β-actin. Data shown are mean ± SEM (n = 7-10 mice per group). *P < .05 pretreatment increased DDAH1 protein expression and phosphorylation of eNOS (ser1177) and decreased the intracellular ADMA levels, respectively, compared with control cells (Figure 3F,G).
Collectively, these data suggest that DMY increases endothelial NO generation and improves endothelial function by activating DDAH1-ADMA-eNOS pathway.

| Systemic delivery of miR-21 decreases NO generation and blocks DMY's protective effects on atherosclerotic plaque formation, lipid metabolism and hepatic inflammation in HCD-fed Apoe −/− mice
MiR-21 is a conserved small non-coding RNA in human and mice, and plays a critical role in the pathogenesis of atherosclerosis. We have shown that DMY improves TNF-α-induced endothelial dysfunction by decreasing miR-21 expression. 20 Moreover, compared with vehicle-treated Apoe −/− mice, the miR-21 expression was decreased by 47% in plasma and 41% in liver, respectively, in those DMY-treated Apoe −/− mice ( Figure S2A,B). Furthermore, in HUVECs in response to ox-LDL, miR-21 expression was significantly increased compared with control, as measured by qPCR, which was attenuated by DMY ( Figure S2C). To examine whether DMY inhibits atherosclerosis by down-regulating miR-21 expression, we first tested whether systemic delivery of exogenous miR-21 mimics (21-m) by tail vein injection could increase miR-21 expression in EC-enriched aortic intima and liver. As shown in Figure S3A Figure 4A). The efficiency of miR-21 overexpression was verified by qPCR using plasma and liver samples ( Figure S4A T cells ( Figure 4E) and a 5.5-fold increase in SMC accumulation ( Figure 4F). Moreover, systemic delivery of miR-21 reduced circulating NO levels ( Figure 4G) and increased VCAM-1 expression in ECs at aortic sinus ( Figure 4H). Furthermore, overexpression of miR-21 in HUVECs inhibited NO production and secretion ( Figure 5A-C), increased VCAM-1 protein and mRNA expression ( Figure 5D,) and the numbers of attached monocytes to ECs ( Figure 5F), and blocked the protective effects of DMY on ECs ( Figure 5A-F). The miR-21 transfection efficiency in HUVECs was verified by qPCR ( Figure 5G).
Next, we investigated whether overexpression of miR-21 abolishes the protective effects of DMY on NO production, lipid metabolism and hepatic inflammation in liver. Compared with Apoe −/− mice received DMY and NS-m, systemic delivery of miR-21 mimics significantly increases miR-21 expression ( Figure S4B) and decreases NO levels in liver from mice treated with DMY ( Figure 6A). Overexpression of miR-21 increased plasma cholesterol, triglyceride and LDL levels, and decreased HDL levels ( Figure 6B). In addition, systemic delivery of miR-21 mimics significantly increased lipid accumulation in liver by Oil Red O staining ( Figure 6C). Moreover, compared with NS-m

| Overexpression of miR-21 abrogates DMYimproved endothelial eNOS activation and NO production through targeting DDAH1
Our previous studies demonstrated that DDAH1 is a direct target of miR-21 in HUVECs. 20,23,30 Thus, we hypothesized that miR-21 abrogates the protective effects of DMY on atherosclerosis by decreasing DDAH1 expression. Indeed, DMY-increased DDAH1 protein expression in ECs in aortic sinus, liver and HUVECs ( Figure S1, Figures 2   and 3). In contrast, overexpression of miR-21 significantly decreased DMY-induced DDAH1 protein expression in HUVECs ( Figure 7A).

| D ISCUSS I ON
Here, we show that DMY increases NO production and inhibits endothelial activation and leukocyte accumulation in liver and aortic vessel wall, thereby attenuates lipid metabolism and atherosclerotic plaque formation in Apoe −/− mice. Moreover, we demonstrate that DMY exerts these beneficial effects, at least in part, by reducing miR-21 expression in ECs, an effect leading to the activation of DDAH1-ADMA-eNOS-NO pathway. Taken together, these data demonstrate that de-repression of DDAH1-ADMA-eNOS-NO pathway due to the reduction of miR-21 plays a critical role in the protective effects of DMY on endothelial function and atherosclerosis.
Decreased NO production in ECs caused by reduced eNOS activity contributes greatly to endothelial dysfunction, impaired vascular homeostasis and atherosclerosis. 31,32 Therefore, pharmacological approaches to maintain or restore eNOS activity in ECs are of therapeutic interest. DMY is a natural flavanonol compound that exhibited multiple pharmacological effects including anti-oxidation. 33 We and others recent found that DMY treatment ameliorated endothelial function in cultured HUVECs in response to TNF-α stimulation or angiotensin II-induced cardiomyocyte hypertrophy by increasing eNOS (ser1177) phosphorylation and NO production. 15,20 Moreover, reduced eNOS activity can also produce excessed ROS. 6 It is possible that DMY-mediated protective effects largely depend on eNOS-NO pathway activation. Consistent with this hypothesis, we observed that elevated NO production and eNOS function in cultured HUVECs, endothelium and liver by DMY administration were associated with improvement of lipid metabolism and inhibition of hepatic inflammation and atherosclerosis. In support of this, NOS inhibitor L-NAME completely abrogates those DMY-mediated beneficial effects. While eNOS dysfunction can be caused, alone or in combination, by abnormal coupling of EC membrane receptors, insufficient supply of substrate (l-arginine) or cofactors (tetrahydrobiopterin), or endogenous inhibitors (ADMA), 29 this study focuses on a role of DMY in modulating DDAH1-ADMA pathway activation.
Different from prior studies, 18,19 the current study demonstrated that DMY supplementation potently increased DDAH1 expression and reduced ADMA levels, resulting in increased eNOS activity and NO production in ECs. We also detected elevated NO levels in plasma and liver, because ECs is the major source of circulating and localization of NO. 32 However, our study did not test a direct role of DDAH1 in DMY-mediated reduction of ADMA and eNOS activation, which will be examined in future studies. Indeed, clinically observed studies identified that miR-21 expression is much higher in plasma and atherosclerotic lesions from patients compared with health control subjects and the circulating levels of miR-21 is a sensitive predictive biomarker of coronary artery disease and future myocardial infarction. [38][39][40][41] Moreover, miR-21 is significantly increased in response to multiple well-established proatherogenic stimulus including ox-LDL, TNF-α and oscillatory shear stress in ECs. 20,36 Overexpression of miR-21 potently promotes endothelial activation, dysfunction and senescence. 36,37 More importantly, miR-21-deficient Apoe −/− mice developed less atherosclerotic plaque than miR-21-sufficient Apoe −/− mice. 42 We previous identified that DDAH1 is a direct target of miR-21. 20,23,30 Thus, we hypothesized that DMY modulates DDAH1 expression by decreasing miR-21, which can enhance the activation of eNOS-NO pathway and improve endothelial function. Our results showed that overexpression of miR-21 inhibited eNOS-NO pathway in both DMY-treated Apoe −/− mice and cultured HUVECs by directly targeting DDAH1, the reduction of which increased the level of ADMA-an inhibitor of eNOS.
These data revealed the molecular basis by which DMY attenuates EC dysfunction and atherosclerosis.
Limitations of the study are (a) eNOS is the dominant isoform expressed in blood vessel wall and the major source of endothelium-derived NO, 43 while L-NAME can inhibit all three NOS function, not only eNOS and (b) gain-of-function of miR-21 is not EC-specific. We cannot rule out other NOS isoforms and miR-21 in other cell types also involved in the beneficial effects of DMY on endothelial function and atherosclerosis.

| CON CLUS IONS
In summary, our findings demonstrate that DMY reduces lipid burden and inhibits atherosclerosis, at least in part, by decreasing miR-21 expression and in turn activating DDAH1-medaited ADMA-eNOS-NO pathway in ECs ( Figure S5). Recent clinical trials found that in patients with non-alcoholic fatty liver disease supplemented with 600 mg DMY per day or patients with type 2 diabetes mellitus with A grossedentata at dose 10 g/d significantly improves glucose and lipid metabolism. 44,45 Given that the intense clinical interest in developing novel preventive and therapeutic strategies that can improve lipid metabolism and lower inflammation, 26,46 our results provide new insights into how DMY supplementation attenuates atherosclerosis, indicating that DMY may be a potential therapeutic adjuvant for treating cardiovascular diseases.

CO N FLI C T O F I NTE R E S T
The authors declare that they do not have any competing interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings described in this study are available in the article.