Sulforaphane inhibits platelet-derived growth factor-induced vascular smooth muscle cell proliferation by targeting mTOR/p70S6kinase signaling independent of Nrf2 activation

Abstract

Activation of nuclear factor erythroid 2-related factor 2 (Nrf2, a transcription factor) and/or inhibition of mammalian target of rapamycin (mTOR) are implicated in the suppression of vascular smooth muscle cell (VSMC) proliferation. The present study has examined the likely regulatory effects of sulforaphane (SFN, an antioxidant) on Nrf2 activation and platelet-derived growth factor (PDGF)-induced mTOR signaling in VSMCs. Using human aortic VSMCs, nuclear extraction and siRNA-mediated downregulation studies were performed to determine the role of Nrf2 on SFN regulation of PDGF-induced proliferative signaling. Immunoprecipitation and/or immunoblot studies were carried out to determine how SFN regulates PDGF-induced mTOR/p70S6K/S6 versus ERK and Akt signaling. Immunohistochemical analysis was performed to determine SFN regulation of S6 phosphorylation in the injured mouse femoral artery. SFN (5 μM) inhibits PDGF-induced activation of mTOR without affecting mTOR association with raptor in VSMCs. While SFN inhibits PDGF-induced phosphorylation of p70S6K and 4E-BP1 (downstream targets of mTOR), it does not affect ERK or Akt phosphorylation. In addition, SFN diminishes exaggerated phosphorylation of S6 ribosomal protein (a downstream target of p70S6K) in VSMCs in vitro and in the neointimal layer of injured artery in vivo. Although SFN promotes Nrf2 accumulation to upregulate cytoprotective genes (e.g., heme oxygenase-1 and thioredoxin-1), downregulation of endogenous Nrf2 by target-specific siRNA reveals an Nrf2-independent effect for SFN-mediated inhibition of mTOR/p70S6K/S6 signaling and suppression of VSMC proliferation. Strategies that utilize local delivery of SFN at the lesion site may limit restenosis after angioplasty by targeting mTOR/p70S6K/S6 axis in VSMCs independent of Nrf2 activation.

Graphical abstract

Sulforaphane (SFN) inhibits downstream targets of mTORC1, including p70S6K/S6 and 4E-BP1 phosphorylation to attenuate PDGF-induced VSMC proliferation. Although SFN activates Nrf2 transcription factor to induce HO-1 expression, Nrf2 downregulation using target-specific siRNA reveals that Nrf2/HO-1 signaling does not contribute to SFN inhibition of VSMC proliferation.

1. Introduction

Neointimal hyperplasia is one of the key biological events that underlie restenosis after angioplasty [1, 2]. Upon endothelial injury during revascularization procedures, vascular smooth muscle cells (VSMCs) undergo phenotypic transition from the contractile to proliferative state [3], thereby contributing to neointima formation. Despite the use of drug-eluting stents (e.g., rapamycin/sirolimus) to limit neointimal hyperplasia, in-stent restenosis still remains a major clinical problem [4]. It is well accepted that antioxidants play a significant role in reducing the risk of cardiovascular disease [5, 6]. Studies by several investigators including our recent findings demonstrate that sulforaphane (SFN, an antioxidant derived from cruciferous vegetables) inhibits neointima formation in vivo and VSMC proliferation in vitro [7–11]. In particular, SFN inhibits: i) the expression/activation of pro-inflammatory transcription factors such as NF-κB (p65 subunit), GATA6, and AP-1 in the injured artery or VSMCs [7, 8]; ii) platelet-derived growth factor (PDGF)-induced VSMC proliferation through upregulation of p53 and p21 that promote cell cycle arrest at G1/S phase [8, 10]; and iii) leptin-induced VSMC proliferation through diminished p70S6 kinase (p70S6K)/S6 ribosomal protein signaling [9]. It is noteworthy that PDGF, a potent mitogen released at the site of arterial injury [12], induces VSMC proliferation through the intermediary activation of signaling components such as mitogen-activated protein/ERK kinase1 (MEK1)/ extracellular signal-regulated kinase (ERK), phosphoinositide 3-kinase (PI 3-kinase)/Akt, and mammalian target of rapamycin (mTOR)/p70S6K [13–18]. However, the likely regulatory effects of SFN on PDGF-induced key proliferative signaling events have not yet been fully examined in VSMCs.

Of importance, mTOR exists as rapamycin-sensitive mTOR complex 1 (mTORC1) and rapamycin-insensitive mTOR complex 2 (mTORC2). mTORC1 signaling is activated in VSMCs upon arterial injury or in response to PDGF [13, 14, 17–21]. The downstream targets of mTORCl include p70S6K/S6 and eukaryotic translation initiation factor 4E-binding protein 1 (4E-BP1) [13–18, 22]. While mTOR interacts with regulatory associated protein of mTOR (raptor) in mTORC1 to regulate the phosphorylation of p70S6K and 4E-BP1 [23–25], mTORC2 complex that contains rapamycin-insensitive companion of mTOR (rictor) has been shown to phosphorylate Akt on serine 473 residue [26, 27]. Thus, it is important to examine how SFN regulates PDGF-induced mTORC1 signaling in VSMCs.

As an isothiocyanate, SFN is known to interact with specific cysteine residues of kelch-like ECH associated protein 1 (Keap1), thereby promoting the activation of nuclear factor erythroid 2-related factor 2 (Nrf2) transcription factor [28]. Due to its essential role as an Nrf2 activator, SFN is a potent inducer of Nrf2-mediated, antioxidant response element (ARE)-driven cytoprotective genes, including heme oxygenase-1 (HO-1). Studies involving Nrf2 overexpression or downregulation have documented a critical role for Nrf2 in suppressing VSMC proliferation [29] and neointima formation [30, 31]. Nevertheless, it is unclear as to whether SFN-mediated Nrf2 activation contributes to inhibition of VSMC proliferation.

The objectives of the present study are to determine the extent to which SFN regulates PDGF-induced mTOR/p70S6K signaling versus Nrf2 expression in VSMCs. We hypothesize that SFN attenuates VSMC proliferation through inhibition of PDGF-induced mTOR/p70S6K signaling and/or upregulation of Nrf2 expression. In the present study, we examined the effects of SFN on PDGF-induced activation of p70S6K versus 4E-BP1 (downstream targets of mTOR), in conjunction with Akt (a downstream target of PI 3-kinase) and ERK1/2 (a downstream target of MEK1). In addition, this study has examined SFN regulation of S6 ribosomal protein, a downstream target and a functional read-out of p70S6K [32]. Since the activation of p70S6K/S6 is determined by the complex formation between mTOR and raptor (key components of mTOR complex 1; mTORC1) [25], select studies have examined how SFN regulates the association between mTOR and raptor in VSMCs. Furthermore, studies involving target-specific downregulation of Nrf2 have examined the likely intermediary role of Nrf2 toward SFN regulation of VSMC proliferative phenotype.

2. Materials and Methods

2.1. Materials

SFN (cat# 574215) and pure proteome protein A/G mix magnetic beads (cat# LSKMAGAG02) were purchased from EMD Millipore (Billerica, MA). All surgical tools were purchased from Roboz Surgical Instrument (Gaithersburg, MD). The primary antibody for SM α-actin was purchased from Abcam (cat# ab5694; Cambridge, MA). The primary antibody for Ki-67 (cat# RM-9106-S1), Mayer’s hematoxylin (cat# TA-125-MH), and NE-PER Nuclear and Cytoplasmic Extraction kit (cat# 78835) were purchased from Thermo Scientific (Hanover, IL). The primary antibody for thioredoxin-1 (Trx-1; cat# sc-13526) was purchased from Santa Cruz Biotechnology, Inc. (Dallas, TX). The primary antibodies for cyclin D1 (cat# 2978), phosphorylated retinoblastoma protein at serine 795 (pRb; cat# 9301), phosphorylated protein kinase B at serine 473 (pAkt^Ser473; cat# 4060) and threonine 308 (pAkt^Thr308; cat# 9275), Akt (cat# 4691), phosphorylated extracellular signal-regulated kinase1/2 at Threonine 202/204 (pERK1/2; cat# 4370), ERK1/2 (cat# 4695), phosphorylated ribosomal protein S6 at serine 235/236 (pS6; cat# 4857), S6 (cat# 2217), phosphorylated p70 S6 kinase at threonine 389 (pp70S6K; cat# 9234), p70S6K (cat# 2708), phosphorylated eukaryotic translation initiation factor 4E-binding protein 1 at serine 65 (p4E-BP1; cat# 9451), 4E-BP1 (cat# 9644), phosphorylated PDGFRβ at tyrosine 751 (pPDGFRβ; cat# 3161), PDGFRβ (cat# 3169), phosphorylated glycogen synthase kinase3β at serine 9 (pGSK3β; cat# 9336), GSK3β (cat# 9315), phosphorylated mTOR at serine 2448 (pmTOR; cat# 5536], mTOR (cat# 2972), regulatory associated protein of mTOR (Raptor; cat# 2280), Keap1 (cat# 4617S), Nrf2 (cat# 12721), heme oxygenase-1 (HO-1; cat# 5061), cAMP-response element-binding protein (CREB; cat# 9197), β-actin (cat# 4970), GAPDH (cat# 5174), anti-rabbit IgG, HRP-linked secondary antibody (cat# 7074), SignalStain antibody diluent (cat# 8112L), and SignalStain Boost IHC Detection Reagent (HRP, Rabbit) [cat# 8114S] were purchased from Cell Signaling Technology (Danvers, MA). Goat anti-mouse IgG (H+L)-HRP conjugate (cat# 1721011) was purchased from Bio Rad (Hercules, CA). Goat anti-rabbit secondary antibody (A-11037), prolong gold anti-fade mountant with DAPI (P36931), Silencer Select Negative Control #1 siRNA (cat# 4390843) and NFE2L2 (Nrf2) silencer select siRNA (cat# 4392420: s9493) were purchased from Life Technologies (Grand Island, NY). 10% neutral buffered formalin (cas# 50-00-0) was purchased from Globe Scientific Inc. (Paramus, NJ). ImmPACT DAB Peroxidase (HRP) Substrate (cat# SK-4105) was purchased from Vector Laboratories, Inc. (Burlingame, CA). All other chemicals were from Fisher Scientific (Fair Lawn, NJ) or Sigma-Aldrich (St. Louis, MO).

2.2. Methods

2.2.1. Cell culture and treatments

Human aortic VSMCs were purchased from ATCC (Manassas, VA) and maintained in vascular cell basal media with smooth muscle growth supplement (SMGS) and antibiotic-antimycotic solution in a humidified atmosphere of 95% air and 5% CO₂ (passage 3 to 4), as described [22]. After the attainment of confluence (~7–10 days), VSMCs were trypsinized and seeded onto petri dishes. Subconfluent VSMCs were subjected to SMGS (serum)-deprivation for 48 h to achieve quiescence. Serum-deprived VSMCs were then exposed to 30 ng.mL⁻¹ PDGF with or without SFN treatment, as described in the legends to respective figures. The chosen concentration of PDGF was based on our previous studies with human aortic VSMCs [18, 22]. Dimethyl sulfoxide (DMSO; 0.05%) was used as vehicle control for SFN (5 μM), as described [33]. The concentration of SFN (5 μM) used in the present study was within the biologically relevant SFN concentration (2 to 20 μM), previously reported in plasma samples of rodents or humans after ingestion of SFN or SFN-rich diet [34, 35].

To determine the acute stimulatory effects of PDGF on key proliferative signaling components, VSMCs were incubated with PDGF for 6 min, as described [18]. To determine the long-term effects of PDGF on signaling events including cyclin D1 and pRb, VSMCs were exposed to PDGF for 48 h [18, 22].

To determine whether SFN regulates mTOR signaling (e.g., pS6, a functional read-out of p70S6K) and Nrf2 expression as a function of time, VSMCs were exposed to SFN between 6 min and 48 h. These findings provided the basis for studying the effects of acute (3 h) and prolonged (48 h) SFN treatment on PDGF-induced proliferative signaling and Nrf2 nuclear translocation in VSMCs, as described in the respective figure legends.

2.2.2. Cell counts

VSMCs were equally seeded onto petri dishes (~4 × 10⁵ cells/60 mm dish) and maintained in culture for 48 h. After the attainment of uniform subconfluency in all dishes, VSMCs were deprived of serum for 48 h. Subsequently, VSMCs were treated with SFN (5 μM) or vehicle control (DMSO, 0.05%) for 30 min followed by exposure to PDGF (30 ng.mL⁻¹) for 96 h. At the 48 h time point, the respective cells were replenished with fresh media with and without SFN or PDGF. After 96 h, VSMCs were trypsinized and the changes in cell number were determined using Countess Counter (Life Technologies), as described [36].

2.2.3. Immunoblot analysis

VSMC lysates (15 μg protein each) were electrophoresed using precast 4–12% NuPage minigels (Life Technologies), and the resolved proteins were transferred to PVDF membranes (EMD Millipore). The membranes were blocked with 5% milk in tris-buffered saline with tween (TBST) and probed with the indicated primary antibodies. After extensive washes, the immunoreactivity was detected using horseradish peroxidase-conjugated secondary antibodies followed by enhanced chemiluminescence. The protein bands were quantified using image J software, as described [37].

2.2.4. Immunoprecipitation studies

Serum-deprived VSMCs were exposed to SFN (5 μM) or vehicle control (DMSO, 0.05%) for 3 h followed by stimulation with or without PDGF (30 ng.mL⁻¹) for 6 min. VSMC lysates were prepared using 1× RIPA buffer with protease inhibitor cocktail (P8340, Sigma Aldrich, St. Louis, MO) and phosphatase inhibitor cocktail (524629, EMD Millipore, Billerica, MA). The lysates were then vortexed and centrifuged at 5000 × g at 4°C for 10 min. Aliquots of supernatants (200 μg protein) were incubated overnight at 4°C with either anti-raptor (cat# 2280, Cell Signaling Technology, Danvers, MA) or anti-mTOR (cat# 2972, Cell Signaling Technology, Danvers, MA) primary antibody at a dilution of 1:50, with continuous mixing. The respective antigen-antibody complexes were subjected to immunoprecipitation by mixing with magnetic beads (50 μL per sample) for 30 min at room temperature (RT). The tubes containing immunocomplexes were placed in a magnetic stand and the supernatant was discarded. The beads were washed three times with phosphate-buffered saline containing 0.05% tween 20 (PBS-T). Elution step was performed by mixing the beads with 2× Laemmli buffer containing dithiothreitol and bromophenol blue and heating at 90°C for 10 min. The eluted samples were then used for immunoblotting.

2.2.5. Nuclear and cytoplasmic protein extractions

Nuclear and cytoplasmic protein fractions were obtained using nuclear and cytoplasmic extraction kit according to manufacturer’s instructions (Thermo Scientific). Briefly, serum-deprived VSMCs were exposed to SFN (5 μM) or vehicle control (DMSO, 0.05%) for 3 h. Cells were harvested with trypsin-EDTA and then centrifuged at 500 × g for 5 min. The cell pellet was washed using ice-cold PBS. Subsequently, ice-cold cytoplasmic extraction reagents (CER I and CER II) were mixed with protease/phosphatase inhibitor cocktail and added to the cell pellet to facilitate cell membrane disruption and the release of cytoplasmic contents. After recovering the nuclear pellet from the cytoplasmic extract by centrifugation (16,000 × g for 5min), the nuclear proteins were extracted using nuclear extraction reagent (NER) plus protease/ phosphatase inhibitor cocktail.

2.2.6. Nucleofection of VSMCs with Nrf2 siRNA

Subconfluent VSMCs were transfected with 500 pmoles of Nrf2 siRNA using Amaxa nucleofector-II device U-025 program (Lonza, Germany). Scrambled (Scr.) siRNA-transfected cells served as controls. Nrf2-and Scr. siRNA-transfected VSMCs were incubated in complete medium for 48 or 72 h. Subsequently, VSMCs were serum-deprived for 45 or 24 h, respectively, and then subjected to treatments as described in the legends to the respective figures.

2.2.7. In vivo studies

2.2.7.1. Animal welfare and ethical statement

All animal experiments were performed in accordance with the Charlie Norwood Veterans Affairs Medical Center Institutional Animal Care and Use Committee (IACUC) guidelines and were approved by the committee. Male C57BL/6J mice (12 weeks old, 24–25 g body weight, Jackson Laboratories, Bar Harbor, ME) were maintained in a room at a controlled temperature of 23°C with a 12:12-h dark-light cycle. Control diet (CON; D14022802) was obtained from Research Diet (New Brunswick, NJ) and fed to mice ad libitum.

2.2.7.2. Model selection

In the chosen mouse model of arterial injury, femoral artery was subjected to guide wire injury to induce neointimal hyperplasia, as described [38].

2.2.7.3. Grouping and randomization of mice

Mice were divided into 2 treatment groups (7 mice per group), where mice were assigned randomly to the groups. In one group, SFN solution (in 0.12% DMSO vehicle) was injected (s.c.) daily at a dose of 0.5 mg.kg⁻¹ [9]. In the other group, 0.12% DMSO was injected (s.c.) daily at a dose of 10 mL.kg⁻¹ (equivalent to injection volume in SFN-treated group).

2.2.7.4. Experimental protocol

Left femoral artery injury was performed in all mice one day after the start of treatments according to the previously described protocol [38] and according to the IACUC-approved ACORP number 14-09-077. Under anesthesia with 2% isoflurane, the skin was shaved at the incision site and cleansed with povidone-iodine pads 3 times. After exposing the left femoral artery, a small muscular branch was isolated, ligated distally and looped proximally. Transverse arteriotomy was performed in the muscular branch and the incision was extended slightly using microsurgery forceps to allow the insertion of a straight spring wire (0.38 mm in diameter, No. C-SF-15-15, Cook, Bloomington, IN) into the femoral artery for 1 min, allowing denudation and dilatation of the artery. The proximal portion of the muscular branch was then secured by suturing to restore the blood flow in the main femoral artery. Right femoral arteries were used as sham controls as described [9]. Mice were monitored for body weight and signs of distress for 3 days after surgery. Twenty one days after femoral artery injury (23^rd day of SFN treatment), mice were anesthetized with 2% isoflurane and then perfused with 0.9% NaCl solution via the left ventricle, followed by perfusion fixation with 4% paraformaldehyde (PFA) in PBS, pH 7.4. The femoral arteries were carefully excised, fixed in 4% PFA for 1–2 days before being processed for frozen sections.

2.2.7.5. Morphometric analysis of femoral artery

From every femoral artery, cross-sections (5 μm each) were stained with haematoxylin and eosin (H&E) or elastic van gieson (EVG). Images were taken by AxioCam high-resolution camera (HRc) attached to an Observer Z1 microscope (Carl Zeiss Microimaging, Inc., Thornwood, NY) using 20× magnification power. Microscopic images of the arterial sections were analyzed using image analysis software (Axiovision, release 4.8.2 SP3). The intimal area from each section was determined by subtraction of the lumen area from the internal elastic lamina (IEL) area (i.e., intimal area = IEL area − lumen area). The medial area was determined by subtraction of the IEL area from the external elastic lamina (EEL) area (i.e., medial area = EEL area − IEL area) as described [9]. Note: In the control group, the final n value was 5 since 2 out of 7 mice died during the experimental period. Cryosectioning, H&E staining and EVG staining of injured femoral arteries were performed in a blinded manner.

2.2.7.6. Immunofluorescence analysis of femoral artery

Immunofluorescence analysis was performed for smooth muscle α-actin (SM α-actin) to confirm that the neointimal layer contains smooth muscle cells. In addition, immunofluorescence analysis was carried out for Ki-67, a marker of cell proliferation. The cross sections of injured femoral arteries were placed in a humidified chamber and fixed in 4% PFA, and then blocked by incubation with 5% normal goat serum for 1 h. This was followed by exposure of the sections to the primary antibody specific for SM α-actin (1:200 dilution) or Ki-67 (1:30 dilution) for 1 h at RT or overnight at 4°C, respectively. The sections were then processed as described previously [9] and the images were captured using confocal microscope at 20× magnification power.

2.2.7.7. Immunohistochemical analysis of femoral artery

Immunohistochemical analysis was performed for pS6 to elucidate the effect of SFN on mTORC1 activation in the neointima. The cross-sections of injured femoral arteries were washed, placed in a humidified chamber, fixed in 10% neutral buffered formalin, and then blocked by incubation with 5% normal goat serum for 1 h. After decanting the blocking serum, the sections were exposed to the primary antibody specific for pS6 (1:75 dilution in signalstain antibody diluent) for 1 h at RT and washed 3 times. Hydrogen peroxide was added to the sections to block endogenous peroxidases and left for 10 min at RT. The sections were then washed 3 times, incubated with signalstain immunohistochemistry detection reagent (HRP-rabbit) for 30 min at RT, washed again 3 times and incubated with diaminobenzidine (DAB) working solution for 8 min. All previous washing steps were done using tris-buffered saline (TBS) with 0.025% triton X-100. The slides were then rinsed in running tap water for 5 min, counterstained with Mayer’s hematoxylin, and rinsed again in running tap water for 5 min. The slides were then dehydrated by immersing in 95% ethanol (2 times, 10 seconds each), 100% ethanol (2 times, 10 seconds each), cleared by immersing in xylene (2 times, 10 seconds each), mounted with mounting medium (SP15-500, Fisher Scientific), and stored at RT for subsequent microscopic visualization. The images were taken using AxioCam high resolution camera (HRc) attached to an Observer Z1 microscope (Carl Zeiss Microimaging, Inc., Thornwood, NY) at 20 × magnification power.

2.2.8. Statistical analysis

Protein bands for western blot (for each in vitro experiment) were normalized using appropriate controls as mentioned in the figure legends, and expressed as either fold mean control to limit unwanted sources of variation or as arbitrary units when protein bands were undetectable under control conditions for the most part. Data shown in the bar graphs reflect the means ± SEM. Morphometric analyses data are expressed as means ± SEM of 5–7 mice per group.

Statistical analysis was performed using unpaired student t-test for Rb phosphorylation, nuclear Nrf2, Nrf2 expression, and morphometric analyses of injured femoral arteries. For time-dependent changes in pS6, Keap1, Nrf2 and HO-1, the data were analyzed by repeated measures one-way ANOVA followed by Tukey’s multiple comparisons test. All other measured parameters were compared using regular two-way ANOVA followed by Tukey’s multiple comparisons test. For siRNA data, the effects of SFN treatment (SFN versus vehicle) and PDGF exposure (PDGF versus control) were compared using regular two-way ANOVA followed by Tukey’s multiple comparisons test for each siRNA (Nrf2 and Scr.) separately. Values of P<0.05 were considered statistically significant. Statistical analyses were carried out using GraphPad Prism software (GraphPad Software Inc. V6.0f, San Diego, CA, USA).

3. Results

3.1. SFN inhibits PDGF-induced VSMC proliferation, cyclin D1 expression, and Rb phosphorylation

Previous studies have shown that SFN at 0.5 to 10 μM concentrations inhibits PDGF– induced VSMC proliferation [8, 10] by preventing cell cycle progression from G1 to S phase [10]. In the present study, SFN treatment at 5 μM concentration led to a marked decrease in PDGF-induced VSMC proliferation by 93.5% (Figure 1A). In addition, SFN diminished PDGF-induced increase in cyclin D1 expression and Rb phosphorylation by 82.9% and 70.5%, respectively (Figure 1B).

Figure 1.

Effects of SFN on PDGF-induced key proliferative signaling in VSMCs. Serum-deprived VSMCs were pretreated with SFN (5 μM) or DMSO (vehicle) for 30 min followed by incubation with or without PDGF (30 ng.mL⁻¹) for 96 h or 48 h. (A) At 96 h time point, the changes in cell number were determined using automated counter (n = 6). (B and C) At 48 h time point, the cell lysates were subjected to immunoblot analysis using primary antibodies specific for cyclin D1, pRb, pERK, ERK, pAkt^Ser473, Akt, pS6 and S6, as described (n = 4–5). β-actin was used as an internal control. The data shown in the bar graphs are the means ± SEM. ^#, * P<0.05 compared with CON (− SFN) and PDGF (− SFN), respectively; two-way ANOVA followed by Tukey’s multiple comparisons test (for all except pRb; unpaired student t-test).

3.2. SFN suppresses basal and PDGF-induced phosphorylation of S6, but not ERK1/2 and Akt, upon prolonged exposure in VSMCs

PDGF-induced VSMC proliferation is mediated by Ras/Raf/MEK1/ERK, PI3-kinase/Akt, and mTOR/p70S6K signaling. Previously, we have reported that exposure of VSMCs to PDGF for a prolonged time interval (48 h) leads to a sustained increase in the phosphorylation of ERK1/2, Akt, and p70S6K [18]. Figure 1C shows that SFN treatment (5 μM) did not result in significant changes in basal and PDGF (48 h)-induced phosphorylation of ERK1/2 and Akt. However, it inhibited basal S6 phosphorylation by 90.2% and completely abolished PDGF-induced increase in S6 phosphorylation.

3.3. SFN inhibits the basal phosphorylation of S6 in a sustained manner at increasing time intervals

As shown in Figure 2, treatment of VSMCs with a fixed concentration of SFN (5 μM) led to a transient but insignificant increase in S6 phosphorylation at 6 and 20 min time intervals. This was followed by a decrease in S6 phosphorylation (by 58.2%) at 3 h time point that persisted up to 48 h time point (92.4%).

Figure 2.

Time-dependent effects of SFN on the phosphorylation of S6 ribosomal protein in VSMCs. Serum-deprived VSMCs were treated with SFN (5 μM) at increasing time intervals (6 min to 48 h). VSMC lysates were then subjected to immunoblot analysis using primary antibody specific for pS6 and S6, as described. β-actin was used as an internal control. The data shown in the bar graph are the means ± SEM from 5 separate experiments. ^#, * P<0.05 compared with SFN-treated for 6 min and 20 min, respectively; repeated measures one-way ANOVA followed by Tukey’s multiple comparisons test.

SFN suppresses basal and PDGF-induced phosphorylation of p70S6K and S6, but not ERK and Akt, upon acute exposure in VSMCs

To further confirm the inhibitory effects of SFN on basal and PDGF-induced S6 phosphorylation and the lack of effects on ERK1/2 and Akt phosphorylation, VSMCs were treated with SFN (5 μM) for 3 h based on the time course study. SFN-treated VSMCs were then subjected to acute exposure to PDGF (6 min). As shown in Figure 3A, SFN did not affect the phosphorylation of PDGF receptor-β and ERK1/2. In addition, SFN did not affect PDGF-induced phosphorylation of Akt^Ser473 and Akt^Thr308, suggesting lack of inhibitory effects on the full activation of Akt [39]. Furthermore, SFN did not affect PDGF-induced phosphorylation of GSK-3β, a downstream target of Akt [40]. However, SFN inhibited basal and PDGF-induced increase in phosphorylation of p70S6K by 96.4% and 54.4% and completely abolished both basal and PDGF-induced increase in S6 phosphorylation (Figure 3B). With regard to 4E-BP1 (an inhibitor of translation initiation), SFN treatment showed a modest inhibition of basal phosphorylation (18.5%), but led to a significant inhibition of PDGF-induced increase in phosphorylation by 71.6%.

Figure 3.

Effects of SFN on key proliferative signaling events induced by acute PDGF exposure in VSMCs. Serum-deprived VSMCs were pretreated with SFN (5 μM) or DMSO (vehicle) for 3 h followed by stimulation with or without PDGF (30 ng.mL⁻¹) for 6 min. (A and B) VSMC lysates were then subjected to immunoblot analysis using primary antibodies specific for pPDGFRβ, PDGFRβ, pERK1/2, ERK1/2, pAkt^Ser473, pAkt^Thr308, Akt, pGSK3β, GSK3β, pp70S6K, p70S6K, pS6, S6, p4E-BP1 and 4E-BP1, as described. GAPDH or β-actin was used as internal control. The data shown in the bar graphs are the means ± SEM from 3 separate experiments. ^#, * P<0.05 compared with CON (− SFN) and PDGF (− SFN), respectively; two-way ANOVA followed by Tukey’s multiple comparisons test.

3.4. SFN attenuates mTOR phosphorylation without affecting mTOR-raptor association

Since mTOR (an upstream kinase) regulates the phosphorylation state of p70S6K/S6 and 4E-BP1 [41, 42], the likely regulatory effect of SFN (5 μM, 3 h) on basal and PDGF (6 min)-induced mTOR activation was determined. Of note, mTOR has four phosphorylation sites (Thr2446 [43], Ser2448, Ser2481 [44] and Ser1261 [45]). In addition, a decrease in mTOR phosphorylation at Ser2448 residue correlates with a decrease in mTORC1 activity [32]. As shown in Figure 4A, SFN showed a modest inhibition of basal mTOR^Ser2448 phosphorylation (30.3%), but significantly suppressed PDGF-induced increase in mTOR^Ser2448 phosphorylation by 96.4%.

Figure 4.

Effects of SFN on mTOR phosphorylation and mTOR-raptor association in VSMCs. (A) Serum-deprived VSMCs were pretreated with SFN (5 μM) or DMSO (vehicle) for 3 h followed by stimulation with or without PDGF (30 ng.mL⁻¹) for 6 min. VSMC lysates were then subjected to immunoblot analysis using primary antibodies specific for pmTOR and mTOR. The data shown in the bar graph are the means ± SEM from 3 separate experiments. ^#, * P<0.05 compared with CON (− SFN) and PDGF (− SFN), respectively; two-way ANOVA followed by Tukey’s multiple comparisons test. (B and C) Serum-deprived VSMCs were treated with SFN (5 μM) or DMSO (vehicle) for 3 h followed by stimulation with or without PDGF (30 ng.mL⁻¹) for 6 min. The cell lysates were then subjected to immunoprecipitation (IP) using primary antibody specific for raptor (B) or mTOR (C). The respective immunocomplexes were used for immunoblot (IB) analysis using the indicated primary antibodies. Blots representative of 2 individual n values are shown for each. The data shown in the bar graph are the means ± SEM from 3 separate experiments.

Although it has been shown that mTOR interacts with raptor in mTORC1 to regulate the phosphorylation of p70S6K and 4E-BP1 [23], a decrease in mTORC1 activity may occur independent of a change in the association between mTOR and raptor [32]. Hence, select studies were performed to determine the interaction between mTOR and raptor in SFN-treated VSMCs. Immunoprecipitation studies using raptor (Figure 4B) and mTOR (Figure 4C) primary antibodies did not show significant differences in the association with mTOR and raptor, respectively.

3.5. SFN enhances the expression of Nrf2 and HO-1 proteins in VSMCs and promotes Nrf2 accumulation in the nuclear fraction

Nuclear translocation of Nrf2 is required for Nrf2-mediated induction of antioxidant and phase 2 genes [46]. SFN has been shown to upregulate Nrf2 [47], promote its nuclear translocation [48] and increase the expression of antioxidant genes including HO-1 in rat VSMCs [47–49]. In the present study using human aortic VSMCs, SFN (5 μM) treatment enhanced the expression levels of Nrf2 and HO-1 proteins in a time-dependent manner (Figure 5). In particular, a significant increase in Nrf2 expression was observed at 1 to 48 h time points. In addition, a significant increase in HO-1 occurred at 6 to 48 h time points. SFN treatment did not affect Keap1 protein expression up to 24 h. However, it led to significant (~2.9 fold) increase in Keap1 expression at 48 h time point.

Figure 5.

Time-dependent effects of SFN on the expression of Nrf2, HO-1, and Keap1 proteins in VSMCs. Serum-deprived VSMCs were treated with SFN (5 μM) at increasing time intervals (6 min to 48 h). VSMC lysates were then subjected to immunoblot analysis using primary antibodies specific for Nrf2, HO-1, and Keap1, as described. β-actin was used as an internal control. The data shown in the bar graphs are the means ± SEM from 3 (Keap1) or 4 (Nrf2 and HO-1) separate experiments. ^# P<0.05 compared with control (vehicle-treated cells); repeated measures one-way ANOVA followed by Tukey’s multiple comparisons test.

Under control conditions, Nrf2 expression was relatively higher in the nuclear fraction compared with cytoplasmic fraction (Figure 6A). SFN treatment (5 μM, 3 h) led to marked increases in Nrf2 expression in the nuclear and cytoplasmic fractions. In particular, SFN enhanced Nrf2 expression in the nuclear fraction by ~2.7 fold (Figure 6B).

Figure 6.

Effects of SFN on Nrf2 protein expression in the nuclear fraction of VSMCs. (A–B) Serum-deprived VSMCs were treated with SFN (5 μM) or DMSO (vehicle) for 3 h. Subsequently, cytoplasmic (Cyt) and nuclear (Nuc) proteins were extracted for immunoblot analysis using primary antibody specific for Nrf2. GAPDH and CREB were used as cytoplasmic and nuclear markers, respectively. Nuclear Nrf2 bands were normalized to their respective CREB bands. The data shown in the bar graph are the means ± SEM from 4 separate experiments. ^# P<0.05 compared with control (vehicle-treated cells); unpaired student t-test.

3.6. SFN attenuates mTOR signaling independent of Nrf2 activation

Nrf2 activators have been shown to exhibit anti-proliferative effects in vitro and in vivo, in part, through Nrf2-mediated upregulation of HO-1 [50–54]. To determine whether SFN-induced Nrf2/HO-1 and/or mTOR signaling contribute to inhibition of PDGF receptor signaling, VSMCs were subjected to target-specific downregulation of Nrf2 using specific siRNA. Figure 7 shows that transfection of VSMCs with Nrf2 siRNA resulted in the downregulation of Nrf2 protein (by ~89%), which prevented SFN-mediated upregulation of HO-1 and Trx1. However, Nrf2 downregulation did not affect SFN-mediated inhibition of PDGF receptor signaling including phosphorylation of p70S6K, 4E-BP1 and S6. In addition, SFN-mediated inhibition of PDGF-induced cyclin D1 expression remained essentially the same under Nrf2 downregulated conditions.

Figure 7.

Effects of Nrf2 downregulation on HO-1/Trx-1 expression and PDGF-induced key proliferative signaling in VSMCs. VSMCs were transfected with scrambled (Scr.) or Nrf2 siRNA followed by maintenance in culture for 72 h or 48 h, and subsequent serum deprivation for 45 h or 24 h, respectively. (A) Representative immunoblots showing Nrf2 downregulation and its effects on SFN (3 h and 48 h)-induced HO-1 and Trx-1 expression, respectively, using specific primary antibodies. The data shown in the bar graph are the means ± SEM (n = 3). ^# P<0.05 compared with Scr siRNA; unpaired student t-test. (B) After 45 h serum deprivation, VSMCs were treated with SFN (5 μM) or DMSO (vehicle) for 3 h followed by exposure to PDGF (30 ng.mL⁻¹) for 6 min. VSMC lysates were then subjected to immunoblot analysis using primary antibodies specific for pp70S6K, p70S6K, pS6, S6, p4E-BP1 and 4E-BP1 (n = 3). (C) After 24 h serum deprivation, VSMCs were pretreated with SFN (5 μM) or DMSO (vehicle) for 30 min followed by exposure to PDGF (30 ng.mL⁻¹) for 48 h. VSMC lysates were then subjected to immunoblot analysis using primary antibody specific for cyclin D1 (n = 3). β-actin was used as an internal control. The data shown in the bar graphs are the means ± SEM. ^#, * P<0.05 compared with CON (− SFN) and PDGF (− SFN) of the same siRNA group, respectively; two-way ANOVA followed by Tukey’s multiple comparisons test.

3.7. SFN attenuates injury-induced neointima formation and Ki-67 expression

Previously, SFN has been shown to attenuate neointima formation after arterial injury in rats with normal metabolic milieu [8, 10, 11] and in mice with diet-induced obesity and metabolic aberrations [9]. In accordance with these earlier studies, SFN treatment in mice (subcutaneous injection, 23 days) significantly attenuated neointima formation in the injured femoral artery, as evidenced by decreases in neointima/media ratio (by 31.4%) and neointimal area (by 36%) compared with control group (Figure 8A–B). SFN treatment did not result in significant changes in lumen or medial areas (data not shown). Immunofluorescence analysis of injured femoral arteries showed that the neointimal layer is mainly composed of smooth muscle cells (Figure 8C). In addition, immunofluorescence analysis of the injured femoral arteries also showed a marked decrease in Ki-67 immunoreactivity (marker of cellular proliferation) in the neointimal layer in SFN-treated group (Figure 8D).

Figure 8.

Effects of SFN on neointima formation and S6 phosphorylation in mouse femoral artery after guide wire injury. SFN was injected daily (0.5 mg.kg⁻¹, s.c.) starting 1 day before femoral artery injury and for the next 21 days until sacrifice. (A) H&E and EVG staining of injured femoral arteries from SFN or vehicle-treated mice (B) Morphometric analyses of injured femoral arteries that include neointima/media ratio and neointimal area. The data shown are the means ± SEM. Number of mice in each group is shown at the bases of the bars. ^# P<0.05 compared with control (vehicle-treated mice); unpaired student t-test. (C) Representative images from confocal immunofluorescence analysis of smooth muscle (SM) α-actin (red), nuclei (DAPI, blue), elastin autofluorescence (laminae, green) and merged images in the injured femoral arteries from SFN or vehicle-treated mice. (D) Representative images from confocal immunofluorescence analysis of Ki-67 (red) and the merged [Ki-67 (red) merged with elastin autofluorescence (green)] in the injured femoral arteries from SFN or vehicle-treated mice. Magnification 20×, and scale bars represent 100–200 μm. Arrows indicate internal and external elastic laminae. (E) Representative images from immunohistochemical analysis of pS6 in the injured femoral arteries from SFN or vehicle-treated mice. Positive cells were visualized using diaminobenzidine (DAB) staining. Magnification 20×; scale bars represent 200 μm (upper panel). Black arrows indicate internal and external elastic laminae, while yellow arrows indicate phosphorylated S6 (lower panel).

3.8. SFN attenuates S6 phosphorylation in the neointimal layer of the injured femoral arteries

As shown in Figure 8E, immunohistochemical analysis revealed a decrease in the phosphorylation state of S6 ribosomal protein in the neointimal layer of SFN-treated group compared with control. These data suggest that the diminution in S6 phosphorylation, an index of decreased mTORC1 activity, may contribute in part to SFN-mediated attenuation of neointima formation at the lesion site.

4. Discussion

Previous studies have shown that activation of Nrf2 transcription factor [29–31] or inhibition of mTOR/p70S6K signaling [13–22] plays a critical role toward attenuating VSMC proliferation/neointima formation after arterial injury. Despite the fact that SFN is an Nrf2 activator [28], the likely intermediary role of activated Nrf2 toward SFN inhibition of VSMC proliferation has not yet been investigated. In addition, there are no reports that document the potential inhibitory effects of SFN on mTOR/p70S6K signaling induced by PDGF, a potent mitogen released at the site of arterial injury. The present study demonstrates that SFN inhibits PDGF-induced activation of mTOR/p70S6K/S6 axis in a sustained manner thereby abrogating enhanced VSMC proliferation, independent of its role in Nrf2 activation. In particular, SFN at 5 μM concentration inhibits PDGF-induced activation of mTOR without affecting mTOR association with raptor in VSMCs. In addition, SFN inhibits PDGF-induced phosphorylation of p70S6K and 4E-BP1, downstream targets of mTOR. Furthermore, SFN abolishes exaggerated phosphorylation of S6 ribosomal protein (a downstream target of p70S6K), which is induced upon acute or prolonged PDGF exposure. Notably, SFN-induced decrease in neointima formation is accompanied by a decrease in S6 phosphorylation in the mouse model of femoral artery injury. While SFN inhibits mTOR signaling in VSMCs, there are no detectable changes in PDGF-induced activation of ERK and Akt. Of importance, downregulation of endogenous Nrf2 by target-specific siRNA reveals an Nrf2-independent effect for SFN inhibition of mTOR/p70S6K/S6 axis, which may contribute to diminished cyclin D1 expression and suppression of VSMC proliferation. Strategies that utilize local delivery of SFN at the lesion site may limit restenosis after angioplasty by targeting mTOR/p70S6K/S6 axis in VSMCs independent of Nrf2 activation.

From a mechanistic standpoint, SFN-mediated increase in Nrf2 accumulation and nuclear translocation observed in the present study with VSMCs may be attributed to its well-characterized role as an electrophile [55]. In this regard, SFN has been shown to modify distinct cysteine residues in Keap1 by interacting with sulfhydryl groups. The consequent disruption of cytoplasmic Keap1-Nrf2 complex would prevent Nrf2 from undergoing proteasomal degradation, thereby facilitating its accumulation and nuclear translocation [56–58]. Nrf2 has been shown to heterodimerize with small Maf protein to promote the transcription of ARE-driven cytoprotective genes, including HO-1, Trx-1, and rate-limiting enzymes of glutathione synthesis [30, 59, 60]. Accordingly, adenoviral overexpression of Nrf2 induces the expression of HO-1 and key enzymes of glutathione synthesis in VSMCs, as described [29]. The present findings reveal that activation of endogenous Nrf2 by SFN results in the upregulation of HO-1 and Trx-1 expression in VSMCs.

Of importance, Nrf2 downregulation using target-specific siRNA suppresses SFN-mediated upregulation of HO-1 and Trx-1 in the present study, suggesting a critical relationship between Nrf2 activation and cytoprotective gene expression in VSMCs. However, under these conditions, SFN-mediated inhibition of PDGF-induced cyclin D1 expression and key proliferative signaling remains essentially the same, suggesting that SFN-mediated Nrf2 activation does not play an intermediary role toward suppressing PDGF-induced VSMC proliferation. At this juncture, it is important to note that pharmacological agents such as curcumin, trans-resveratrol, sulfasalazine (anti-inflammatory/immune-modulatory agent), and dimethylfumarate (an anti-psoriasis drug) inhibit VSMC proliferation in vitro and in vivo presumably through Nrf2-mediated upregulation of HO-1, as reported in recent studies [50–54]. However, none of these studies has utilized Nrf2 gene-silencing strategy to support an obligatory role of Nrf2 activation toward inhibition of VSMC proliferation. Notably, Ashino et al. have shown that Nrf2 downregulation by target-specific siRNA does not affect PDGF-induced VSMC proliferation [30]. However, a study by Levonen et al. has demonstrated that adenoviral delivery of Nrf2 gene results in anti-proliferative effects that are counterbalanced by its anti-apoptotic effects in vivo, thereby contributing to no significant changes in neointimal hyperplasia [29]. Together, although the role of Nrf2 toward inhibition of VSMC proliferation/neointima formation remains controversial, the present findings support the notion that Nrf2 activation is not required at least for SFN-mediated inhibitory effects on PDGF-induced VSMC proliferation. Future studies should employ Nrf2-null mice to confirm the intermediary role of Nrf2 transcription factor toward SFN-mediated inhibition of neointima formation after arterial injury.

In addition to targeting Keap1, SFN has been shown to interact with cysteine residues in several other proteins, including NF-κB [61], AP-1 [62], tubulin [63], and TLR4 [64], thereby accounting for its anti-inflammatory and/or anti-proliferative effects. As a thiol-reactive agent [65], SFN may thus have the potential to directly regulate mTORC1 activity in VSMCs. Previously, a redox-sensitive mechanism has been reported for the control of mTORC1 activity, where phenylarsine oxide is shown to react with the sulfhydryl groups of two closely spaced cysteine residues in mTOR to enhance mTORC1 activity. This is evidenced by increased p70S6K phosphorylation and destabilization of mTOR-raptor complex in HEK293 cells [25]. In addition, conditions that create a reducing environment could inhibit mTORC1 activity possibly through reduction of disulfide bonds in the FATC domain of mTOR [66]. However, the present study with VSMCs reveals that SFN does not affect the association between mTOR and raptor. It diminishes the phosphorylation of mTORC1 downstream targets, suggesting decreased mTORC1 activity by SFN. Thus, the likely interaction of SFN with cysteine residues in mTOR does not provide a plausible explanation for its inhibitory effect on mTORC1 activation in VSMCs. As an alternative mechanism, SFN may dysregulate tuberous sclerosis complex (TSC) and/or Rheb, upstream regulators of mTORC1 [67–69]. Future studies are clearly warranted to determine whether SFN affects TSC and/or Rheb signaling to inhibit mTOR/p70S6K/S6 axis in VSMCs.

Of importance, SFN-mediated inhibition of mTORC1 activation is clearly different from classical mTORC1 inhibitors (e.g., rapamycin and ATP-competitive inhibitors) with regard to mechanistic and functional considerations. For instance, SFN inhibits PDGF-induced phosphorylation of mTOR and its downstream signaling components, p70S6K/S6 and 4E-BP1, thereby suppressing VSMC proliferation. Rapamycin (sirolimus) is known to interact with FK506-binding protein 12 (FKBP12) to form a complex, which inhibits mTORC1 activity in an allosteric manner by binding to FKBP12-Rapamycin-Binding (FRB) domain of mTOR [67, 70]. Its role in mTORC1 inhibition is reflected by diminished phosphorylation of p70S6K and S6, whereas 4E-BP1 phosphorylation remains refractory with prolonged rapamycin treatment [70]. This increase in 4E-BP1 phosphorylation by rapamycin may allow protein synthesis to occur sufficient enough to promote cell proliferation [70]. However, at least in VSMCs, SFN treatment is not associated with an increase in 4E-BP1 phosphorylation. Furthermore, ATP-competitive mTOR inhibitors that compete with ATP in the catalytic domain of mTOR are currently in clinical trials, and these inhibitors have been shown to inhibit kinase-dependent functions of mTORC1 and mTORC2 [70, 71]. While SFN inhibits mTORC1 signaling, it does not appear to diminish mTORC2 activation (as reflected by its lack of effect on the phosphorylation Akt^Ser473, a downstream target of mTORC2). Further studies are clearly warranted that should determine whether SFN regulates mTORC2 (rictor-mTOR) kinase activity in VSMCs [26, 27]. From a therapeutic standpoint, rapamycin is used clinically for the treatment of restenosis after angioplasty. While it is beneficial in inhibiting exaggerated VSMC proliferation, its use is also associated with diminished endothelial cell proliferation. This would prevent reendothelialization at the lesion site, thereby increasing the risk of thrombosis with rapamycin treatment [72]. Future studies should determine the likely beneficial effects of SFN toward promoting reendothelialization in the injured vessel wall.

In conjunction with the observed inhibitory effects of SFN on VSMC proliferation/neointima formation (present study), it is pertinent to note that SFN administration to wild-type rodents does not result in changes in body weight, food intake, lipid profile, blood pressure, and heart rate, as reported in previous studies [9, 73, 74]. However, in high fat diet- or western diet-fed obese mice, SFN attenuates weight gain with no significant effects on food intake [9, 74]. In addition, SFN treatment improves systemic lipid profile. For instance, it decreases the circulating levels of total cholesterol [74] and triglycerides [9] in diet-induced obese mice. Furthermore, SFN decreases systolic blood pressure in spontaneously hypertensive stroke-prone rats and obese mice [9, 73], with no effect on heart rate [9].

In conclusion, SFN administration would provide a realistic treatment option to attenuate the development and/or progression of restenosis after angioplasty through its ability to target PDGF-induced mTOR, p70S6K, and 4E-BP1 signaling in VSMCs independent of Nrf2 activation. Furthermore, in view of the potential link between HO-1 induction and plaque stabilization [75], Nrf2-mediated upregulation of HO-1 by SFN may prevent plaque destabilization/rupture in vulnerable vessels thereby offering protection from acute coronary events.

Figure 9.

Sulforaphane (SFN) inhibits downstream targets of mTORC1, including p70S6K/S6 and 4E-BP1 phosphorylation to attenuate PDGF-induced VSMC proliferation. Although SFN activates Nrf2 transcription factor to induce HO-1 expression, Nrf2 downregulation using target-specific siRNA reveals that Nrf2/HO-1 signaling does not contribute to SFN inhibition of VSMC proliferation.

Abbreviations

ARE

antioxidant response element

CREB

cAMP-response element-binding protein

DAB

diaminobenzidine

DMSO

dimethyl sulfoxide

4E-BP1

eukaryotic translation initiation factor 4E-binding protein 1

EEL

external elastic lamina

ERK1/2

extracellular signal-regulated kinase1/2

EVG

elastic van gieson

FKBP12

FK506-binding protein 12

FRB domain

FKBP12-Rapamycin-Binding domain

GSK3β

glycogen synthase kinase3β

H&E

haematoxylin and eosin

HO-1

heme oxygenase-1

IEL

internal elastic lamina

Keap1

kelch-like ECH associated protein 1

mTOR

mammalian target of rapamycin

mTORC1

mTOR complex 1

mTORC2

mTOR complex 2

MEK1

mitogen-activated protein/ERK kinase1

Nrf2

nuclear factor erythroid 2-related factor 2

PBS

phosphate-buffered saline

PDGF

platelet-derived growth factor

PFA

paraformaldehyde

PI 3-kinase

phosphoinositide 3-kinase

pRb

phosphorylated retinoblastoma protein

p70S6K

p70S6 kinase

raptor

regulatory associated protein of mTOR

Rheb

Ras homolog enriched in brain

rictor

rapamycin-insensitive companion of mTOR

RT

room temperature

SFN

sulforaphane

SM α-actin

smooth muscle α-actin

SMGS

smooth muscle growth supplement

TBST

tris-buffered saline with tween

Trx-1

thioredoxin-1

TSC

tuberous sclerosis complex

VSMCs

vascular smooth muscle cells