Human umbilical vein endothelial cells (HUVECs) were obtained from American Type Culture Collection (ATCC). Cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS) (Invitrogen, USA), 100 units/ml penicillin and 100 mg/ml streptomycin. Cells were plated in 6-well plates or 100-mm tissue culture dishes 1 day before experiments. Near-confluent cultures were starved overnight in medium containing 0.5% FBS before NaHS treatment or stimulation with LPS.

All treatments were performed in serum-free culture medium with penicillin and streptomycin. Cells were washed twice with serum-free culture medium and pre-incubated with saline or NaHS (50, 100, and 200 μM) for 24 h. Some cells were then stimulated with LPS (100 ng/ml) for 24 h in the continuous presence of NaHS. In the time course experiment, cells were pre-incubated with saline or NaHS (100 μM) for 0, 6, 24, or 48 h. For experiments using inhibitors, cells were pre-treated with DX600 (1 μM, an ACE2 inhibitor, Phoenix Pharmaceuticals; Pedersen et al., 2011) for 2 h before pre-incubation with NaHS (100 μM) for 24 h and subsequent stimulation with LPS (100 ng/ml) for 24 h.

C57BL/6J male apoE^−/− mice were purchased from the Animal Center of the Beijing University, Beijing, China. The study was carried out in accordance with the Guidelines for the Care and Use of Laboratory Animals of the Shanghai JiaoTong University School of Medicine. The protocol was approved by the Committee on the Ethics of Animal Experiments of the Shanghai JiaoTong University School of Medicine (Permit Number: [2015]-117). At age of 8 weeks, mice were randomly assigned to partial ligation of left carotid artery (LCA) or sham operation, followed by being fed a high-fat diet for 4 weeks that contained 10% fat from lard and was supplemented with 2% (w/w) cholesterol. LCA was partially ligated as previously described (Sullivan, 2002; Nam et al., 2009) with slight modifications. In brief, mice were intrapeitoneally anesthetized by the mixture of xylazine (10 mg/kg) and ketamine (80 mg/kg). The neck was epilated and then disinfected with iodophor. LCA was exposed by a ventral midline incision (4–5 mm) in the neck. Except that the superior thyroid artery remained intact, left external carotid, internal carotid and occipital artery were ligated with 6–0 silk (Figure 1A). The skin was sutured and mice were then kept in a warm chamber until recovery. After LCA partial ligation or sham operation, mice were randomly given saline, NaHS (1 mg/kg/day, i.p.) or DL-propargylglycine (PAG, 10 mg/kg/day, i.p.). Some NaHS-treated mice were intraperitoneally given MLN-4760, a selective ACE2 inhibitor (0.5 mg/kg, daily, Millennium Pharmaceuticals) 2 weeks after initiation of NaHS treatment (Ye et al., 2012). Four weeks postligation, all the mice were sacrificed. Samples of carotid arteries and blood were collected and stored at −80°C. At the beginning and the end of the study, systolic blood pressure (SBP) was monitored by a tail cuff system (Blood Pressure Analysis System BP- 98AW monitor, Japan).

Anesthetized mice underwent left ventricle perfusion with buffered saline and 10% neutralized formalin at 100 mmHg. LCA was collected en bloc with the trachea and esophagus. LCA were embedded in paraffin. Fifteen serial sections (5 μm) were taken 500–1,000 μm proximal to the location of ligation. Five sections from each mouse were stained with hematoxylin and eosin. The staining were then examined by light microscopy (n = 6 mice in each group, objective lens magnification of × 20; eyepiece magnification of × 10). The lesions were quantified by Image J software (National Institutes of Health, USA) as described in the literatures (Lessner et al., 2002; Sullivan, 2002; Nam et al., 2009; Zhang et al., 2012, 2015). Measurement of intima and media of carotid arteries was acquired by tracing the border of the lumen and the internal and external elastic laminae (Figure S1). The mean area of the intima and media was measured by tracing multiple sections, and the ratio of the area of intima to media was calculated. Sections were also stained with Masson's trichrome to display collagen components in a given plaque. To calculate proportions of collagen content to neointima, we measured intimal lesion areas and Masson's trichrome positive blue areas in at least three sequential sections (Tasaki et al., 2013). Every morphological parameter was quantified by one investigator blinded for the treatment.

LCA were deparaffinized in xylene and rehydrated in aqueous solutions with decreasing alcohol content, followed by a wash in PBST (1 × PBS with 0.5% Tween 20, pH 7.4). Antigen retrieval was achieved by heating the slides in 10 mM sodium citrate (PH 6.0) at 95°C for 20 min. Slides were gradually cooled down at room temperature and washed in H₂O and PBST. To inactivate endogenous peroxidase, sections were treated with 3% H₂O₂ for 15 min, followed by incubation with 10% normal goat serum for 20 min to block nonspecific staining. After incubation overnight at 4°C with primary antibody (rabbit anti-mouse ACE2 in a dilution of 1:75, Abcam, USA), slides were incubated with secondary antibody (HRP-conjugated, goat anti-rabbit IgG in a dilution of 1:200, Santa Cruz Biotechnology, USA) for 2 h at room temperature. Specific staining for ACE2 was developed by the reaction with 3, 3′-diaminobenzidine and counterstaining was applied with hematoxylin. The immunohistochemical staining was viewed by light microscopy (objective lens magnification of × 40; eyepiece magnification of × 10).

Total RNA from cells was extracted using Trizol ® reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol. The concentration of isolated RNA was assayed by measuring the absorbance at 260 nm and the integrity of RNA was visualized by ethidium bromide staining of 18S and 28S on a denaturing agarose gel. One microgram RNA was reversely transcribed using iScript™ cDNA Synthesis Kit (Biorad, USA) at 25°C for 5 min, 42°C for 30 min, followed by 85°C for 5 min. The resulting cDNA was then used as a template for real time PCR amplification. The forward and reverse primers of ACE2, ACE, Mas, and β-actin gene were shown in Table 1. Real-time PCR was performed using Lightcycler 2.0 system (Roche Applied Science, USA). Relative expression of gene mRNA was analyzed using a comparative method described in the user bulletin. The data were calculated with 2^−ΔΔCT method and normalized to β-actin expression.

Cells (3 × 10⁶) were collected and lysated at 4°C using radioimmunoprecipitation assay lysis buffer. Tissues of LCA were homogenized on ice in radioimmunoprecipitation assay lysis buffer. Cell lysates or tissue homogenates were centrifuged at 14,000 g for 10 min at 4°C. Protein concentration in the soluble fraction was determined by Bradford method. Proteins (20 μg) were size-fractionated by 8% SDS-PAGE and transferred onto nitrocellulose membranes. The membranes were blocked for 2–3 h with 5% nonfat milk and then probed overnight at 4°C with rabbit polyclonal anti-ACE2 (1:500, molecular weight 97 kDa, Abcam, USA), β-actin antibodies (1:1,000, Santa Cruz Biotechnology, USA) and mouse monoclonal anti-CSE antibody (1:1,000, molecular weight 45 kDa, Abnova, Taiwan) respectively, followed by secondary antibody for 2 hs with a 1:2,000 dilution of HRP-conjugated, goat anti-rabbit IgG or goat anti-mouse IgG (Santa Cruz Biotechnology, USA). The blots were visualized using a standard enhanced chemiluminescence system.

MCP-1, TNF- α and IL-6 (Quantikine, R&D systems), as well as Ang-(1-7) and Ang II (Cloud-Clone Corp., USA) were assayed using ELISA kits according to the manufacturers' instructions. Results for the levels of Ang-(1-7) and Ang II in carotid arteries were expressed as pg/mg protein after correction for the protein concentration in tissue homogenates (determined using the Bradford assay).

Ten percent Trichloroacetic acid (120 μl), 1% zinc acetate (60 μl), 20 μM N, N-dimethyl-p-phenylenediamine sulfate (40 μl) in 7.2 M hydrochloride acid and 30 μM FeCl₃ (40 μl) in 1.2 M hydrochloride acid were mixed with plasma (120 μl) and distilled water (100 μl). After 10 min, the absorbance of the mixture was assayed by spectrophotometry at 670 nm (Tecan Systems Inc.). Plasma level of H₂S was calculated using a standard curve of NaHS with a concentration from 3.125 to 100 μM.

H₂S synthesizing activity in LCA was measured as described previously (Zhang et al., 2006). In brief, 4.5% w/v tissue homogenate (430 μl) in 100 mM potassium phosphate buffer (pH 7.4) was mixed with 20 mM L-cysteine (20 μl), 2 mM pyridoxyal 5′-phosphate (20 μl) and saline (30 μl). The reaction started in tightly sealed tubes after tubes were transferred from ice to water bath at 37°C. After incubation for half an hour, 1% zinc acetate (250 μl) was added and trapped the evolved H₂S, followed by 10% trichloroacetic acid (250 μl) to cease the reaction. Afterward, 20 μM N, N-dimethyl-p-phenylenediamine sulfate (133 μl) in 7.2 M hydrochloride acid was added, immediately followed by 30 μM FeCl3 (133 μl) in 1.2 M hydrochloride acid. The absorbance of the resulting mixture was assayed by spectrophotometry at 670 nm (Tecan Systems Inc). H₂S concentration was calculated using a standard curve of NaHS with a range from 3.125 to 100 μM. Results were expressed as nmoles H₂S produced per mg protein in tissue homogenates (determined using the Bradford assay).

The data were expressed as mean ± SEM. The significance of differences among groups was evaluated by analysis of variance (ANOVA) with post-hoc Tukey's test when comparing three or more groups. The significance of differences between two groups was evaluated by T-test. A P < 0.05 was regarded as statistically significant.

Atherosclerosis is known to be closely associated with disturbed flow characterized by low and oscillatory shear stress (Sullivan, 2002; Pedersen et al., 2011). However, studies directly linking H₂S to disturbed flow condition in atherogenesis are lacking. Here, we investigated the alterations of H₂S metabolism in a murine model of acutely disturbed flow-induced atherosclerosis by partial carotid ligation (Figure 1A). ApoE^−/− mice underwent either LCA partial ligation or sham operation and then were fed a high-fat diet for 1 or 4 weeks. By the first week, LCA showed slight evidences of atherosclerotic lesions as determined by H&E staining (Figure 1B). By 4 weeks, LCA developed accelerated atherosclerosis (Figure 1B). However, only minor or no lesions were observed in carotid arteries isolated from mice fed a high-fat diet for 1 or 4 weeks after sham operation (Figure 1B).

Then, we examined the time-dependent alterations of carotid H₂S bio-synthesis in disturbed flow-induced atherosclerosis. As shown in Figures 1C,D, H₂S synthesizing activity and CSE expression in LCA decreased in a time-dependent fashion, with a significant reduction 4 weeks after ligation. Furthermore, plasma H₂S concentration gradually declined in a time-dependent manner (Figure 1E). There was a statistically significant reduction in plasma H₂S level by 4 weeks after ligation. These data suggest both local and systemic H₂S insufficiency in LCA partial ligation induced atherosclerosis. In another word, acute flow disturbance induced by partial ligation inhibited carotid CSE expression and CSE activity, thus resulting in an evident decline in production of endogenous H₂S.

Next, we explored the alterations of carotid ACE2 over time in disturbed flow-induced atherosclerosis (Figure 2A). Immunostaining showed that ACE2 was finely expressed in vascular endothelial cells from normal mice. By 1 week postligation, expression of ACE2 in LCA endothelim was significantly upregulated. However, by 4 weeks, the endothelial expression of ACE2 dramatically decreased with the progression of atherosclerosis (Figures 2A,B). Four weeks after ligation, LCA endothelium exhibited less intense staining with ACE2 than 1 week after ligation. As a result, the level of Ang-(1-7) in LCA was high at the beginning of ligation and then gradually reduced in a time dependent manner (Figure 2C). The level of Ang II in LCA changed oppositely (Figure 2D). These data indicate that disturbed flow in carotid arteries initially induced the endothelial expression of ACE2 in an attempt to inhibit the initiation of atherosclerosis. However, with the development and progression of atherosclerosis, endothelial ACE2 expression was inhibited, therefore shifting a balance from anti-atherosclerotic Ang-(1-7) to pro-atherosclerotic Ang II.

In the time course study, we found that the biosynthesis of H₂S significantly decreased 4 weeks after PLA whereas the carotid ACE2 expression started rising 1 week after PLA (Figures 1, 2). This finding suggested that H₂S seems to play a part in regulating ACE2 expression in the advanced stage of atherosclerosis (4 weeks after PLA) but not in the early phase of atherosclerosis (1 week after PLA). In addition to H₂S, there are some other factors contributing to ACE2 expression in different stages of atherosclerosis. Here, we investigated the effect of H₂S on carotid ACE2 expression 4 weeks after PLA.

NaHS (1 mg/kg/day, i.p.), DL-propargylglycine (PAG, 10 mg/kg/day, i.p.) or saline was randomly given to mice after partial ligation. After 4 weeks, the extent of atherosclerotic lesions was assessed by histological analysis (Figure 3A). Histological analysis showed that NaHS treatment significantly impeded the plaque development, as characterized by alleviated neointimal hyperplasia and less atherosclerotic lesions in LCA (Figures 3B,C). On the other hand, inhibition of H₂S formation by PAG induced advanced atherosclerosis, as evidenced by more severe neointimal hyperplasia and more obvious narrowness in LCA (Figures 3B,C). Masson's staining revealed that NaHS reduced the amount of collagen deposits whereas PAG enhanced hyperplasia of collagen fibers in neointimal lesions (Figures 3D,E). In addition, NaHS or PAG had negligible effect on body weight, plasma lipid profiles, and blood pressure (Table 2).

Immunohistochemistry revealed that supplement with NaHS enhanced the expression of ACE2 in endothelial cells whereas PAG inhibited it in atherosclerotic lesions (Figures 4A,B). Then, we used Western blot to quantify the ACE2 expression level in LCA. Consistent with the findings obtained by immunostaining, NaHS significantly upregulated carotid ACE2 expression while PAG downregulated it (Figure 4C). As a result, NaHS notably promoted the local production of Ang (1-7) in LCA but PAG inhibited it (Figure 4D). The carotid level of Ang II was blunted by NaHS while it was raised by PAG (Figure 4E).

In order to reinforce the influence of H₂S on ACE2-Ang-(1-7), MLN-4760, a selective and potent inhibitor against mouse ACE2 was applied (Ye et al., 2012). As shown in Figures 4D,E, blockage of ACE2 activity by MLN-4760 significantly reversed the elevation of carotid Ang-(1-7) level and the decline of carotid Ang II level induced by NaHS. The anti-atherosclerotic property of NaHS was also significantly abolished by treatment with MLN-4760 (Figures 3A,D).

Since ACE2 is highly expressed in carotid endothelium, we sought to examine the effect of H₂S on ACE2-Ang-(1-7) in HUVECs. We found that NaHS time-dependently (Figures 5A–C) and dose-dependently (Figures 5D–F) increased the mRNA and protein expression of ACE2 in HUVECs. Subsequently, NaHS enhanced the production of Ang-(1-7) (Figures 5G,I) but decreased the level of Ang II (Figures 5H,J) in a time-dependent and dose-dependent manner. Moreover, similar effect of NaHS was observed in LPS-stimulated HUVECs (Figure 6). LPS at a concentration of 100 ng/ml significantly downregulated the expression of ACE2 and Ang-(1-7) but increased the level of Ang II in HUVECs. The inhibition of endothelial ACE2-Ang-(1-7) expression induced by LPS was reversed by NaHS in a dose dependent manner (Figure 6). NaHS also dose-dependently reduced the level of Ang II in LPS-stimulated HUVECs (Figure 6E). As Mas receptor is a functional receptor for Ang-(1-7) (Santos et al., 2003), we evaluated the effect of NaHS on the mRNA expression of Mas in HUVECs. NaHS did not alter the mRNA expression of Mas in unstimulated or LPS-stimulated HUVECs (Figure S2A). In addition, we examined the effect of H₂S on ACE, the classical component of RAS in HUVECs. After treatment with NaHS, the mRNA expression level of ACE remained unchangeable in unstimulated or LPS-stimulated HUVECs (Figure S2B).

As shown in Figure 7, the production of TNF-α, IL-6, and MCP-1 was dose-dependently repressed by NaHS in either unstimulated or LPS-stimulated HUVECs. The inhibitory effect of NaHS was significantly blunted by DX600, a selective ACE2 inhibitor (Nam et al., 2009), suggesting that H₂S may inhibit endothelial activation through an ACE2-dependent mechanism (Figure 8).

In addition to endothelial cells, positive immunostaining for ACE2 in atherosclerotic plaques was also present in foam cells and macrophages (Figure 4A). H₂S has been shown to be a potent regulator of monocyte/ macrophage activation (Zhi et al., 2007; Zhang et al., 2012), which is central to the pathogenesis of atherosclerosis. Therefore, we explored whether H₂S would affect the activation of macrophages via ACE2-Ang-(1-7). As shown in Figure S3, NaHS significantly inhibited the production of TNF-α and MCP-1 in unstimulated or LPS-stimulated RAW264.7 cells. The anti-inflammatory property of NaHS was not reversed by pretreatment with DX 600, suggesting that H₂S likely modulates macrophage functions through other pathways but not ACE2.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.