Sodium Ferulate Inhibits Neointimal Hyperplasia in Rat Balloon Injury Model

Background/Aim Neointimal formation after vessel injury is a complex process involving multiple cellular and molecular processes. Inhibition of intimal hyperplasia plays an important role in preventing proliferative vascular diseases, such as restenosis. In this study, we intended to identify whether sodium ferulate could inhibit neointimal formation and further explore potential mechanisms involved. Methods Cultured vascular smooth muscle cells (VSMCs) isolated from rat thoracic aorta were pre-treated with 200 µmol/L sodium ferulate for 1 hour and then stimulated with 1 µmol/L angiotensin II (Ang II) for 1 hour or 10% serum for 48 hours. Male Sprague-Dawley rats subjected to balloon catheter insertion were administrated with 200 mg/kg sodium ferulate (or saline) for 7 days before sacrificed. Results In presence of sodium ferulate, VSMCs exhibited decreased proliferation and migration, suppressed intracellular reactive oxidative species production and NADPH oxidase activity, increased SOD activation and down-regulated p38 phosphorylation compared to Ang II-stimulated alone. Meanwhile, VSMCs treated with sodium ferulate showed significantly increased protein expression of smooth muscle α-actin and smooth muscle myosin heavy chain protein. The components of Notch pathway, including nuclear Notch-1 protein, Jagged-1, Hey-1 and Hey-2 mRNA, as well as total β-catenin protein and Cyclin D1 mRNA of Wnt signaling, were all significantly decreased by sodium ferulate in cells under serum stimulation. The levels of serum 8-iso-PGF2α and arterial collagen formation in vessel wall were decreased, while the expression of contractile markers was increased in sodium ferulate treated rats. A decline of neointimal area, as well as lower ratio of intimal to medial area was observed in sodium ferulate group. Conclusion Sodium ferulate attenuated neointimal hyperplasia through suppressing oxidative stress and phenotypic switching of VSMCs.


Introduction
Percutaneous coronary intervention (PCI) is the main therapeutic approach to coronary heart disease. Although the obvious therapeutic effects are encouraging, the clinical complications following PCI operation cannot be ignored. Neointimal formation after vascular injury is the main cause reducing cure rate and leading to rehospitalization or revascularization [1]. With the advent of drug-eluting stent, a great progress has been made in preventing restenosis [2]. Antiproliferative drugs are coated onto the stent and released slowly, interrupting cell proliferation cycle and significantly reducing in-stent restenosis [3]. However, a series of clinical consequences, including poor re-endothelialization, delayed vessel function restoration and higher cardiac risk, demonstrate that present drug-eluting stent introduction still needs careful consideration [4,5]. Therefore, to understand the complex mechanism involved in restenosis and seek a new optimal way to attenuate neointimal hyperplasia have important clinical significance.
Current evidence indicates intimal hyperplasia after artery injury is an intricate pathological process. It consists of thrombosis, inflammation, proliferation and extracellular matrix production [6]. Immediate endothelial denudation after angioplasty results in the exposure of circulating cells to subendothelial matrix [7]. The activated platelet, numerous cytokines and reactive oxygen species (ROS) are recruited, thus accelerating thrombosis, inflammation and oxidative stress in vessel lesions [8,9,10]. In response to these environmental changes induced by endothelial screen deprivation, the medial vascular smooth muscle cells (VSMCs) turned from contractile to synthetic subtype. The new dedifferentiated phenotype facilitates VSMCs migration to the intima and begins to proliferate and secrete extracellular matrix, which leads to the formation of the thick neointimal after acute phase [6]. Unlike the skeletal and cardiac muscle cells, VSMCs phenotype transition is reversible and regulatory [7]. Many studies have demonstrated that phenotype modulation from synthetic to contractile shows great suppression on neointimal hyperplasia [11,12,13,14].
Sodium ferulate (3-methoxy-4-hydroxy-cinnamate sodium, SF), the salt of ferulate acid, is widely distributed in Ligusticum, Chuanxiong, Propolis and other herbs [15,16]. As ferulate acid of active ingredient, sodium ferulate is an approved clinical drug with better solubility and lower toxicity than ferulate acid [17]. Previous studies reveal sodium ferulate or ferulate acid possesses a variety of beneficial pharmacological effects on human diseases. Due to its favorable anti-inflammatory and anti-oxidant properties, ferulate acid could attenuate ischemic injuries, atherosclerosis and diabetic related damages [18,19,20]. The latest research suggests ferulate acid is a novel inhibitor of presynaptic glutamate release, and sodium ferulate can protect cortical neurons against glutamate toxicity [21,22]. In addition, ferulate acid has a great capacity of anti-platelet aggregation, anti-hypertension and anti-tumor [23]. Curcumin, a derivative of ferulate acid, is an inhibitor of rabbit VSMCs proliferation [24]. However, the effect of ferulate acid or sodium ferulate on neointimal hyperplasia after vessel injury remains unclear. The aim of this study was to identify whether sodium ferulate has inhibitory effect on neointimal formation and the underlying processes.

Ethics statement
All rats used in present study were purchased from Animal Center of Renmin Hospital of Wuhan University. All experimental procedures were authorized by Animal Care and Use Committee of Wuhan University, and strictly complied with guidelines for the Care and Use of Laboratory Animals by the National Institutes of Health (NIH publication No. 85-23, revised 1996).

Cell proliferation and migration assay
VSMCs proliferation was performed with the cell counting kit-8 (CCK-8) according to the manufacturer's protocol (Dojindo Laboratories, Kumamoto, Japan). In brief, 8,000 cells were seeded in each well of 96-well plate. After synchronous growth in DMEM including 0.5% FBS for 24 hours, VSMCs were pre-incubated in a variety of concentrations of sodium ferulate (0, 50, 100 and 200 mmol/L) (Suzhou Changtong Chemical, Suzhou, China) for 1 hour and then stimulated with Ang II (1 mmol/L) (Peprotech, CT, USA) for another 48 hours. Following addition with CCK-8 reagent, OD values were measured at 450 nm using microplate spectrophotometer.
Transwell chamber was used for detecting VSMCs migration. 10 4 synchronized cells were seeded into the upper chamber containing 200 mmol/L sodium ferulate in 200 mL DMEM. The lower chamber was filled with 1 mmol/L angiotensin II in 600 mL DMEM. After incubation for 8 hours, non-migrated cells were removed from filter membrane. The lower membrane was fixed in methanol and stained with 0.1% crystal violet. Cells from five random fields were calculated under microscope at a magnification of 100.

Reactive oxygen species assay
A fluorescent probe (Beyotime, Shanghai, China) was used to detect intracellular ROS level. VSMCs were seeded onto 24-well plate at a density of 1.5610 4 (each well). Synchronized for 24 hours, cells were incubated with 200 mmol/L sodium ferulate for 1 hour and then exposed to 1 mmol/L Ang II for another 1 hour. The fluorescent probe DCFH-DA was added at a final concentration of 10 mmol/L and incubated at 37uC for 20 minutes. Pictures were taken under fluorescence microscope and the fluorescence mean density was measured from three random fields for each well.

NADPH oxidase and SOD activity assay
Cultured VSMCs were pre-treated with sodium ferulate and induced by Ang II as described above, and NADPH oxidase activity was detected following previous study [25]. Briefly, cells were incubated in lysis buffer on ice for 30 minutes and then centrifuged at 14,000 rpm for 5 minutes. The supernatant was added to assay buffer in presence of 100 mM lucigenin and 100 mM NADPH (Sigma, MO, USA) in a final volume of 1 mL. Sample chemiluminescence was measured immediately, and the NADPH oxidase activity was expressed as units/sec.mg protein.
For SOD activity, an assay kit (Cell Biolabs, CA, USA) was used according to the manufacturer's instructions.

Cell phenotypic transition
Following synchronization in medium with 0.5% serum for 24 hours, VSMCs were cultured with non-serum for 72 hours. Under this starved circumstance, VSMCs started to present contractile phenotype. Then 10% serum was added with or without sodium ferulate pre-treatment to promote cell phenotype to transfer into the synthetic pattern. After incubation at 37uC for another 48 hours, the evolution process of cell phenotypic transition was completed.

Western blot analysis
The total, cytoplasmic and nuclear proteins were isolated using protein extraction kit (Beyotime) according to the manufacturer's instructions. After quantitation (BCA kit, Beyotime), protein was separated by electrophoresis on 10% SDS-polyacrylamide gels and transferred onto PVDF membrane. Antibodies against b-actin, bcatenin, smooth muscle myosin heavy chain protein (SM-MHC), Notch1 (Santa Cruz, CA, USA), GAPDH, histone H3, phospho-p38, phospho-ERK1/2 (Cell Signaling Technology, MA, USA) and smooth muscle a-actin (SM a-actin) (Sigma) were used to probe with the interest blots. Finally, protein expression was detected with HRP-conjugated second antibody (Santa Cruz) and ECL (Pierce, IL, USA) luminescence method.

Quantitative real-time RT-PCR
The total RNA was extracted using the picopure RNA isolation kit (Applied Biosystems, CA, USA). The purified RNA was immediately reversely transcribed into cDNA using first-strand synthesis system (Invitrogen, CA, USA). Addition with primer sequences and Sybergreen supermixt (Bio-Rad, CA, USA) in the proper proportions, purified cDNA was get into amplification. Data were normalized to GAPDH and analyzed by 2 2DDCt method. The primer sequences used in this study were displayed in Table 1.

Balloon injury model in rat carotid artery
Thirty-six male Sprague-Dawley rats weighing 350-400 g were randomly divided into three groups: sham group, saline group and sodium ferulate group (n = 12 for each group). Animals were anesthetized via intraperitoneal injection with 2% sodium pentobarbital (Sigma) at a dose of 40 mg/kg. The common, internal and external carotid arteries were exposed sequentially. A balloon catheter (balloon diameter 1.25 mm, length 15 mm; Medtronic, MN, USA) was introduced into the common carotid in the case of systematic heparinization (100 U/kg, intravenous injection). The balloon was inflated and passed three times with rotation. Rats in sham group underwent the same procedure, except for artery injury. Animals in sodium ferulate group were administrated with 200 mg/kg sodium ferulate through intragastric way once a day, from the first day after surgery. Rats in saline group were taken in equal volume saline as control. All animals were fed with conventional diet until sacrificed.

Serum 8-iso-PGF2a detection
Blood samples from external jugular vein were collected 7 days after injury, when rats were sacrificed. After centrifugation at 3000 rpm for 10 min, the upper serum were extracted and stored at -80uC. An ELISA kit (Cell Biolabs) was used to quantify serum 8iso-PGF2a according to manufacturer's instructions.

Histomorphological analysis
At the 7th day, the common arteries on operated side were collected and fixed in 4% paraformaldehyde. After embedded in paraffin, arteries were cut into sections of 4 mm thickness and stained with hematoxylin and eosin. The lumen, neointimal and medial areas were measured and the ratio of intimal to medial area was also calculated by image analysis software (Image Pro Plus 6.0). Immunofluorescence assay and Masson trichrome staining were also used to evaluate vessel sections.

Statistical analysis
All statistical analysis was performed with Statistical Product and Service Solutions 13.0 software (SPSS 13.0). Data was represented as means 6 SD. Statistical analysis was performed by Student-Newman-Keuls post-test for comparison between two groups, and one-way ANOVA for multiple groups. P,0.05 was considered as statistically significant.

Sodium ferulate inhibits Ang II induced VSMCs proliferation and migration
In response to Ang II stimulation, VSMCs acquired strong capacity of proliferation when compared to the non-stimulated cells. Sodium ferulate reduced this increase at three concentrations, and 200 mmol/L sodium ferulate performed the best effect ( Figure 1A and Figure S1). Consisted with the results of CCK-8 test, Ang II-treated cells presented approximately two-fold migratory effect to the control. Pre-treatment with sodium ferulate decreased the total number of migratory cells significantly, less than one fifth of the stimulated group ( Figure 1B).

Sodium ferulate attenuates oxidative stress under Ang II stimulation
Exposed to Ang II for 1 hour, the intracellular ROS level was markedly increased compared to the negative control group. Sodium ferulate inhibited the activation induced by Ang II, and less than half of stimulated cells were identified with ROS expression (Figure 2A). NADPH oxidase, the main source of ROS production [26], was also inhibited by sodium ferulate at a small dose ( Figure 2B). Meanwhile, the SOD activity was elevated by sodium ferulate in response to Ang II stimulation ( Figure 2C). In present study, p38 MAPK was detected as the potential mechanism explaining for the underlying inhibitory effect of sodium ferulate on oxidative stress. The western blot results demonstrated the Ang II-induced p38 phosphorylation was blocked by sodium ferulate, but no significant difference of ERK1/2 phosphorylation was found ( Figure 2D).

Sodium ferulate upregulates SM a-actin and SM-MHC protein expression under 10% serum stimulation
Contractile proteins of SM a-actin and SM-MHC in VSMCs were induced during serum starvation, while such feature was alleviated by 10% serum. Pretreatment with Sodium ferulate prevented the decrease caused by serum and reversed both SM aactin and SM-MHC expression to the levels before serum stimulation ( Figure 3).

Sodium ferulate downregulates Notch and Wnt pathway in response to serum
In order to understand the possible mechanism accounting for the protective effects of sodium ferulate on VSMCs phenotype switching, the Notch and Wnt signaling transduction were observed under serum stimulation. As expected, the nuclear Notch-1 protein, Jagged-1, Hey-1, Hey-2 mRNA were increased significantly by 10% serum when compared to the unstimulated group. Sodium ferulate reversed these effects and decreased Notch-1 to 0.64-fold and Jagged-1, Hey-1 and Hey-2 mRNA production by more than 5-fold, 4-fold and 7-fold to the simulated group, respectively (Figure 4).
Similar with the Notch pathway, Wnt pathway components containing b-catenin and Cyclin D1 were also greatly blocked by sodium ferulate under serum induction. In sodium ferulate group, total b-catenin protein and Cyclin D1 mRNA were 0.53-fold and 0.31-fold to the serum group ( Figure 5).    Sodium ferulate inhibits oxidative stress in vivo 8-iso-PGF2a, an indicator of oxidative stress in vivo [27], was evaluated 7 days after artery injury. Compared to the sham group, the serum level of 8-iso-PGF2a was increased significantly after exposure to balloon catheter insertion. In contrast, sodium ferulate decreased 8-iso-PGF2a expression by more than 2-fold compared to the saline group (Table 2).

Sodium ferulate enhances contractile markers expression and attenuates collagen generation in injured arteries
As indicated above, VSMCs in synthetic state lose contractile ability as well as lowered expression of contractile makers [7]. Sodium ferulate suppressed this decrease, and increased the SM aactin and SM-MHC expression in arteries. Compared to the saline group, vessels in sodium ferulate group presented much stronger fluorescence throughout the wall ( Figure 6A). In addition, the greater synthesis induced by catheter insertion was also blocked by sodium ferulate. The results of Masson staining showed the collagen composition in sodium ferulate group was much less than saline group ( Figure 6B).

Sodium ferulate inhibits neointimal hyperplasia induced by balloon injury
As results of histomorphological detection shown, a layer of loose neointima close to the internal elastic membrane was generated and the perfusion lumen area was lost markedly in arteries of saline group when compared to the sham group. However, the deterioration on vessels with catheter intervention was alleviated by sodium ferulate. The neointimal formation in sodium ferulate administration group was thinner and intimal cells arranged more orderly than the saline group ( Figure 7A). In addition to the smoother intimal surface, the intimal area and ratio of intimal area to medial area were both decreased ( Figure 7B, 7C).

Discussion
In the present study, we demonstrated for the first time that sodium ferulate could inhibit neointimal hyperplasia after artery injury. This conclusion was established based on the observations below. First, sodium ferulate decreased Ang II-induced VSMCs proliferation and migration, which was attributed to significant reduction of excessive ROS generation. Second, sodium ferulate blocked VSMCs phenotypic switching and promoted the synthetic alteration to the contractile phenotype. In addition, the accompanied modulation on molecular mechanism, including inhibition of p38 MAPK, Notch and Wnt signaling were found.
Ang II, a most well-known active peptide of renin-angiotensin system, is also a multiple modulator involved in vasoconstriction, inflammation, thrombosis and vascular remodeling [28]. In response to arterial injury, Ang II is found highly expressed in lesions and plays as an important mediator to trigger oxidative stress and cell growth [29,30]. In this study, our data confirmed Ang II induced significant increase of ROS in cultured VSMCs, while sodium ferulate inhibited Ang II-induced ROS generation, and attenuated VSMCs proliferation and migration. However, the relationship between sodium ferulate and oxidative stress in VSMCs remains unknown. According to previous studies, imbalance between ROS and anti-oxidants results in oxidative damage [31]. Combination of Ang II to Ang II type 1 receptor stimulates non-phagocytic NADPH oxidase, including NOX1 and NOX4 in VSMCs [32,33,34]. The activated NOX accept electrons donated by NADPH and converts molecular oxygen (O 2 ) into unstable superoxide state (?O 22 ) and hydrogen peroxide (H 2 O 2 ) [35]. These emerging oxidants are regulated by antioxidants such as SODs, catalase, heme oxygenase-1, glutathione and other molecules [26]. Sodium ferulate not only decreased NADPH oxidase activity, but also up-regulated total SOD, the important anti-oxidant substance, contributing to sequential suppression of ROS production. In addition, this inhibitory effect was also confirmed in vivo. Rats administrated with sodium ferulate displayed a much lower serum level of 8-iso-PGF2a, which indicated oxidative stress triggered by intimal injury was interrupted by sodium ferulate.
Accumulating evidence suggests ROS participate in a variety of signaling transduction as second messengers [36]. The regulation of ROS on signal pathway is dependent on oxidative modification of target proteins. Generally, these proteins contain redox-reactive cysteine (Cys) residues, which can be oxidized into forms of sulfenic acid (2SOH) or other further oxidative products [37]. Mitogen activated protein kinases (MAPK), as the major targets of ROS, are regulated by Ang II in many cell types [38]. There are two ways for ROS to modify MAPK activity. MAPK phosphatases and protein tyrosine phosphatases share a conserved redoxsensitive cysteine residue, both of which facilitate the turn of MAPK into active state through oxidative modification by ROS [39,40]. Our data implied p38 and ERK1/2 phosphorylation increased along with aggravation of intracellular ROS under Ang II stimulation, while sodium ferulate blocked p38 activation with no effect on ERK1/2. However, the relationship between sodium ferulate and MAPK pathway lacks of adequate experimental evidence. In rat hippocampus, sodium ferulate suppressed p38 activation, but enhanced ERK1/2 phosphorylation under amyloid-beta stimulation [41]. In human vascular endothelial cells, ferulate acid inhibited JNK rather than p38 and ERK1/2 when exposed to radiation [42]. However, ferulate acid attenuated activation of ERK1/2 and JNK rather than p38 in VSMCs after stimulation with 1 mM Ang II for 15min [43]. Accordingly, the regulation of sodium ferulate on MAPK is dependent on cell type, drug dose, stimulus, stimulation time and other factors.
Previous findings revealed VSMCs could alter from contractile to synthetic state when exposed to various environmental factors, for instance, growth factor, cytokines, hypoxia, injury and mechanical force [44]. The synthetic type cells not only downregulate contractile protein, including SM a-actin, SM-MHC, calpolin, myocardin and smoothenin, but also acquire strong capacity of synthesis [7]. Collagen, the main component of ECM, is an important product of dedifferentiated VSMCs [45]. In the present study, serum, a known potent stimulator of phenotypic switching on VSMCs [46], induced much lower expression of SM a-actin and SM-MHC. Sodium ferulate blocked this deterioration and up-regulated them as well as in balloon injured arteries.    Moreover, the collagen content in neointimal was also decreased by sodium ferulate. Our findings above suggested sodium ferulate attenuated synthetic VSMCs activation induced by vessel injury, which played a pivotal role in inhibition of sequential neointimal hyperplasia. However, the effect of sodium ferulate on vascular relaxation is not clear. Although ferulate acid was considered as an inactive component to induce vascular relaxation in Radix Angelica [47], Chen et al. have demonstrated sodium ferulate is a non-selective relaxant to VSMCs [48]. Unfortunately, we have not made any relevant studies on this issue. Protective relaxation after endothelium deprivation plays a critical role in vascular health, and the relationship between sodium ferulate and vascular relaxation should be investigated in the future study. The transcriptional pathway involved in regulating VSMCs specific gene contains GATA-6, serum response factor (SRF) and myocardin [49]. Recently, Notch has been found to be an additional pathway critical for regulating VSMCs phenotype [49]. Generally, interaction of Notch ligands (Jagged 1 and 2; Delta-like 1, 3 and 4) and receptors (Notch 1 to 4) induces Notch intracellular domain (NICD) cleavage from membrane and translocation into nucleus, further leading to transcription activation of the effect genes (Hes, Hey) [50]. Sodium ferulate may decrease proteolytical release of NICD through down-regulation of Notch 1 and Jagged 1, which ultimately caused inhibition of Hey-1 and Hey-2 expression. Previous studies suggest Notch signaling participates in regulating VSMCs differentiation and identify SM a-actin and SM-MHC as targets of Notch [51,52]. Indeed, Jagged 1-Notch interactions can induce SM a-actin expression via NCID/CBF-1 complex binding to specific region of SM a-actin promoter [51]. Over-expression of Hey-1 and Hey-2 turns off this signal by interfering the binding process rather than interruption on NCID/ CBF-1 complex [53]. In this way, the decline of Hey-1 and Hey-2 caused by sodium ferulate, relieved inhibition on binding of NCID/CBF-1 complex and marker gene, contributing to increase of SM a-actin and SM-MHC expression.
Wnt pathway is another emerging spot on modulation of VSMCs phenotype alteration. Following with ligands association with receptors, the activation of Wnt signaling facilities target genes expression, which includes Cyclin D1, cMyc and insulin growth factor [54]. Since the decrease of Cyclin D1 dependent on b-catenin/TCF axis, sodium ferulate attenuated cell cycle activation and arrested VSMCs proliferation. In addition, inactive Wnt pathway mediated lower matrix matalloproteinase activity and less ECM composition [55], which contributes to suppression of neointimal generation. Considerable observations have shown that there are several intersections between Notch and Wnt pathway. Here, we focus on three of them, which would involve in sodium ferulate simultaneous inhibition on Notch and Wnt pathway in our study. First, Delta-like-1(Dll-1), a ligand of Notch pathway, is also a downstream gene of Wnt [56]. Sodium ferulate inhibited Wnt signaling activation would down-regulate the Dll-1, which resulted in depression of Notch components. Second, NICD can attach with Dishevelled (Dsh) and thus interrupt Wnt signals as well as Notch pathway itself [57]. Sodium ferulate may promote this combination, so as to enable inactivation of Notch and Wnt transduction. Third, glycogen synthase kinase-3b (GSK-3b) can bind and phosphorylate Notch, inducing NICD degradation [58]. Sodium ferulate could intensify GSK-3b, which caused b-catenin decline and Notch transcriptional arrest. Hence, we deducted that crosstalk between Wnt and Notch pathway would be an important part of complex regulatory networks sodium ferulate acts on VSMCs behavior.
In summary, sodium ferulate could attenuate neointimal hyperplasia after artery insult through blockade of oxidative stress and VSMCs phenotypic switching. The regulatory action involved with complex molecular mechanism, such as MAPK, Notch and Wnt pathway. Based on observations of the beneficial effect showed in this study, sodium ferulate might be a potent therapeutic approach to improve vascular function, and reduce the occurrence of restenosis after PCI. Figure S1 Effect of sodium ferulate on Ang II-induced VSMCs proliferation. VSMCs were pre-incubated with sodium ferulate at dose of 200 and 300 mmol/L for 1 hour and stimulated with 1 mmol/L Ang II for 48 hours. Following addition with CCK-8 reagent, OD values were measured at 450 nm (n = 6). * P,0.05 vs. the control group, * * P,0.05 vs. the Ang II group. (TIF)