Early modulation of macrophage ROS-PPARγ-NF-κB signalling by sonodynamic therapy attenuates neointimal hyperplasia in rabbits

Disruption of re-endothelialization and haemodynamic balance remains a critical side effect of drug-eluting stents (DES) for preventing intimal hyperplasia. Previously, we found that 5-aminolevulinic acid-mediated sonodynamic therapy (ALA-SDT) suppressed macrophage-mediated inflammation in atherosclerotic plaques. However, the effects on intimal hyperplasia and re-endothelialization remain unknown. In this study, 56 rabbits were randomly assigned to control, ultrasound, ALA and ALA-SDT groups, and each group was divided into two subgroups (n = 7) on day 3 after right femoral artery balloon denudation combined with a hypercholesterolemic diet. Histopathological analysis revealed that ALA-SDT enhanced macrophage apoptosis and ameliorated inflammation from day 1. ALA-SDT inhibited neointima formation without affecting re-endothelialization, increased blood perfusion, decreased the content of macrophages, proliferating smooth muscle cells (SMCs) and collagen but increased elastin by day 28. In vitro, ALA-SDT induced macrophage apoptosis and reduced TNF-α, IL-6 and IL-1β via the ROS-PPARγ-NF-κB signalling pathway, which indirectly inhibited human umbilical artery smooth muscle cell (HUASMC) proliferation, migration and IL-6 production. ALA-SDT effectively inhibits intimal hyperplasia without affecting re-endothelialization. Hence, its clinical application combined with bare-metal stent (BMS) implantation presents a potential strategy to decrease bleeding risk caused by prolonged dual-antiplatelet regimen after DES deployment.

Macrophage infiltration in the media of injured arteries. RAM-11-positive macrophages were determined by immunofluorescence in uninjured arteries (control group) and injured arteries harvested at 12,
ALA-SDT inhibits neointima formation. At day 28 after ALA-SDT treatment, an angiogram showed that the lumen diameter of the arteries in the SDT group was larger than the lumen diameters in the other groups ( Supplementary Fig. S1). Meanwhile, compared with the control group, the mean lumen area and the percentage of area stenosis measured by optical coherence tomography (OCT) in the SDT group were significantly increased by 83% (1.42 ± 0.06 vs. 2.61 ± 0.15 mm 2 ) and decreased by 60% (20.56 ± 3.12% vs. 8.24 ± 1.97%), respectively (Fig. 3A). Fractional flow reserve (FFR), which measures the pressure gradient across stenotic arteries, was increased in arteries with larger cross-sectional lumen areas 16,17 . The FFR value in the SDT group (Fig. 3B) increased significantly by 0.39-fold compared with the control group (0.60 ± 0.04 vs. 0.83 ± 0.03). Morphomet- Figure 2. ALA-SDT induces macrophage apoptosis and reduces TNF-α, IL-6 and IL-1β expression in the media of balloon-denuded rabbit femoral arteries at day 1 after treatment. (A) Rabbit tissue Western blot analysis of apoptosis-related proteins, including pro-caspase 3, cleaved caspase-3 and Bcl-xl. GAPDH was used as a loading control. (B) Immunofluorescence staining of TUNEL and RAM-11 to assess the macrophage apoptosis rate in the indicated groups (n = 6). Scale bar = 100 μm. (C) Representative sections of TNF-α, IL-6 and IL-1β expression stained by immunohistochemistry and quantitative analysis in the indicated groups (n = 7). Scale bar = 200 μm. Each data point represents one rabbit. *p < 0.05, **p < 0.01 and ***p < 0.001 versus control. ALA-SDT inhibits NF-κB signalling and subsequently suppresses the expression and secretion of TNF-α, IL-6 and IL-1β in THP-1-derived macrophages. To explore the underlying mechanisms by which ALA-SDT inhibited inflammation in the injured arteries, we performed Western blot analysis and found that the expression level of NF-κB p65 in THP-1-derived macrophages was significantly decreased in the nucleus and increased in the cytoplasm after ALA-SDT (Fig. 5A,B). IκBα phosphorylation was significantly decreased upon ALA-SDT treatment, suggesting an inhibitory effect on IκBα degradation that aids in sequestering NF-κB in the cytoplasm (Fig. 5C). The inhibitory effect of ALA-SDT on p65 nuclear translocation was further confirmed by confocal laser scanning (Fig. 5D). Furthermore, ALA-SDT significantly reduced the phosphorylation level of NF-κB p65 (p-p65) (Fig. 5E), which is considered a marker of NF-κBactivation. As shown in Fig. 5F-H, ALA-SDT prominently inhibited the expression and secretion of TNF-α, IL-6 and IL-1β in THP-1-derived macrophages.  Fig. S4A). When we treated TNF-α-stimulated HUASMCs with the same SDT parameters to induce macrophage apoptosis, we did not find any remarkable change in TNF-α, IL-6 or IL-1β secretion ( Supplementary Fig. S4B). Intriguingly, when ALA-SDT-treated THP-1-derived macrophages were co-cultured with TNF-α-stimulated HUASMCs, remarkable inhibition of HUASMC proliferation, migration and IL-6 production was observed (Supplementary Fig. S4C-G).
ALA-SDT inhibits NF-κB transcriptional activity while increasing PPARγ transcriptional activity in THP-1-derived macrophages. To clarify the mechanism by which ALA-SDT inhibited NF-κB signalling, we performed a transcription factor profiling array. Among the 96 transcription factors, PPAR was significantly up-regulated, and HIF, NF-κB and SP-1 were significantly down-regulated by ALA-SDT (Supplementary Fig. S5). Compared with the control group, ALA-SDT reduced NF-κB binding to DNA response elements by 66% (3.99 ± 0.86 vs. 1.36 ± 0.23) while increasing PPAR binding to DNA response elements by 182% (6.41 ± 1.52 vs. 18.10 ± 3.60). We further confirmed that ALA-SDT increased the protein level of PPARγ but not PPARα or PPARβ/δ in THP-1-derived macrophages ( Supplementary Fig. S6). www.nature.com/scientificreports/ ALA-SDT inhibits NF-κB signalling and the production of TNF-α, IL-6 and IL-1β by activating PPARγ. PPARγ has been generally identified as an anti-inflammatory transcription factor. A part of this anti-inflammatory regulation is mediated through negative interference with NF-κB 18 . In this study, ALA-SDT significantly increased the nuclear protein level of PPARγ in THP-1-derived macrophages (Fig. 6A). To determine whether ALA-SDT inhibits NF-κB signalling and if the production of TNF-α, IL-6 and IL-1β occurs through PPARγ activation, THP-1-derived macrophages were pre-treated with the well-known PPARγ inhibitor GW9662 or PPARγ siRNA. The inhibition of NF-κB signalling by ALA-SDT was reversed by the presence of GW9662 (  www.nature.com/scientificreports/

ALA-SDT induces THP-1-derived macrophage apoptosis via the PPARγ-NF-κB pathway.
As an upstream regulatorof NF-κB, PPARγ is also involved in the regulation of apoptosis 19 . To illustrate whether PPARγ-NF-κB signalling is involved in the ALA-SDT-induced apoptosis process, we detected the effects of ALA-SDT on apoptosis-related protein expression, including pro-caspase-9, cleaved caspase-9, pro-caspase 3, cleaved caspase-3, Bcl-2 and Bax, in the presence and absence of GW9662 in THP-1-derived macrophages. Here, we demonstrated that pretreatment with GW9662 reversed the effect of ALA-SDT on apoptosis-related protein expression (Fig. 7A). Furthermore, flow cytometry analysis indicated that GW9662 pretreatment diminished the increased percentage of apoptotic cells upon ALA-SDT (Fig. 7B).

Discussion
In this study, we performed early intervention of ALA-SDT at day 3 after balloon denudation in hypercholesterolemic rabbits since the maximum macrophage infiltration was reached at that time point. Previously, we demonstrated that macrophages represent the main target cell type for ALA-SDT in atherosclerotic plaques 12,22 .
Considering the pivotal role of macrophage-mediated inflammation in neointima formation 8 , ALA-SDT may www.nature.com/scientificreports/ exert protective effects by reducing macrophages from the early stage. Interestingly, fluorescence spectra analysis showed that ALA-PpIX peaks at 1 h in the injured femoral artery, which differs from our previous result that ALA-PpIX peaks at 2 h in established plaque 23 . Indeed, we found that most macrophages infiltrate into freshly injured arterial subintima and reside inside the media of pre-existing plaques. Thus, it would take less time for macrophages in freshly injured vessels to phagocytise ALA. As early as day 1 after ALA-SDT treatment, the rate of apoptotic macrophages had already increased, while TNF-α, IL-6 and IL-1β in the media of the femoral artery decreased significantly. At day 28 after ALA-SDT treatment, OCT, FFR and histopathological analysis showed significantly inhibited intimal hyperplasia with fewer macrophages and proliferative SMCs and reduced collagen content but increased elastin content without affecting endothelium regeneration. It is well known that infiltrated macrophages in the subintimal space produce pro-inflammatory cytokines, enzymes and chemokines to promote the proliferation and migration of SMCs. After dedifferentiation from the contractile to synthetic phenotype, SMCs migrate into the subendothelial space to excessively proliferate, and a large amount of collagen is secreted, accompanied by a decrease in elastin expression, leading to intima hyperplasia and stenosis of the lumen 7,24,25 . Thus, ALA-SDT induces depletion of activated macrophages and reduction in pro-inflammatory cytokines, which may promote phenotypic switching of SMCs, thereby inhibiting intimal hyperplasia.
In the present study, ALA-SDT parameters that induce THP-1-derived macrophage apoptosis had no direct effect on cytokine secretion in HUASMCs. However, after co-culturing ALA-SDT-treated macrophages with HUASMCs, the proliferation, migration and IL-6 production of HUASMCs were significantly inhibited. We previously found that after incubating 1 mM ALA with THP-1-derived macrophages and SMCs separately for 3 h, the fluorescence intensity of ALA-PpIX increased by 12-fold in THP-1-derived macrophages and by twofold in SMCs 20,26 . Additionally, after incubation of 1 mM ALA with SMCs for 24 h, ALA-SDT inhibited the proliferation and migration of SMCs and enhanced the expression of differentiated phenotypic markersin SMCs 26 . These results suggest that ALA-SDT had no direct effect on HUASMCs after incubation with ALA for 3 h, as HUASMCs uptake sono-sensitizers much slower than macrophages. Since inflammatory cytokines are widely recognized to activate SMCs 27-29 , we inferred that ALA-SDT-induced TNF-α, IL-6 and IL-1β reduction in macrophages indirectly inhibited HUASMC proliferation, migration, and IL-6 expression ( Supplementary Fig. S9).
In the current study, we observed the effects of ALA-SDT on macrophages via NF-κB signalling inhibition, which was further confirmed by the transcription factor profiling array. ALA-SDT inhibited the transcriptional activity of NF-κB but increased the transcriptional activity of PPARγ. Hence, inhibition of PPARγ reversed the effects of ALA-SDT effect on NF-κB activity. We previously proved that 1.16 W/cm 2 ultrasound intensitymediated SDT produces a large amount of ROS to induce macrophage necrosis, while low ultrasound intensitymediated SDT (0.48-0.84 W/cm 2 ) produces a small amount of ROS to induce macrophage apoptosis 30 . Unlike other studies showing that ROS directly activate NF-κB signalling, which in turn aggravates the inflammatory www.nature.com/scientificreports/ response 31 , our results demonstrated that small amounts of ROS produced by low intensity ALA-SDT inhibit NF-κB signalling by activating PPARγ. There are three limitations in this study. Although we confirmed that ALA-SDT inhibits intimal hyperplasia without affecting re-endothelialization in balloon-injured hypercholesterolemic rabbits, whether ALA-SDT can effectively exert the same effect upon BMS implantation requires further investigation. Moreover, the underlying cellular interaction between ALA-SDT-treated macrophages and SMCs needs further clarification. Finally, how ALA-SDT-induced ROS production activated PPARγ and subsequently inhibited NF-κB remains to be elucidated.
In summary, early intervention with ALA-SDT efficiently inhibits intimal hyperplasia without affecting reendothelialization; hence, its clinical application combined with BMS implantation represents a potential strategy for decreasing bleeding risk caused by prolonged dual-antiplatelet regimen after DES deployment.

Methods
Reagents. 5 Ltd., Shanghai, China) for one week before surgery. Then, the rabbits were intravenously anaesthetized with pentobarbital (30 mg/kg) and ketamine (10 mg/kg). A 6F sheath was inserted into the left common carotid artery after intravenous administration of heparin (100 IU/kg). A 2.75 × 14 mm balloon dilation catheter (JW Medical Systems, Shandong, China) was guided through the descending aorta into the right superficial femoral artery with guidewire assistance (Cordis, MiamiLakes, FL, USA), inflated to 6-8 atmospheres and withdrawn to the femoral bifurcation with a rotating motion three times to induce effective endothelial denudation. After the injury, the rabbits continued to receive the high cholesterol diet until being killed.
To examine macrophage infiltration in the balloon-denuded arteries, five healthy rabbits were used as uninjured controls. The injured rabbits were euthanized at 12, 24, 48 and 72 h (n = 5) after balloon denudation. To investigate the effects of ALA-SDT on neointima formation, 56 balloon-denuded rabbits were randomly assigned to the control, ultrasound, ALA and ALA-SDT groups (n = 14). Each group was divided into 2 subgroups (n = 7), and the rabbits were killed at day 1 and day 28 after ALA-SDT treatment.
Cell culture. Cells from the human THP-1 monocyte cell line were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA) and cultured as previously described 20 . The cells were seeded into 35-mm Petri dishes (1.0 × 10 5 cells/ml) and treated with 100 ng/ml PMA for 72 h for differentiation into macrophages. Human umbilical artery SMCs (HUASMCs) (BeNa Culture Collection, Beijing, China), which have been previously used as a cell model for intimal hyperplasia 32 , were cultured in DMEM (Gibco, Life Technologies) with 10% foetal bovine serum and activated with TNF-α (10 ng/ml) for 24 h before experiments.
SDT procedure. For the rabbit experiments, ALA (60 mg/kg) was bolus injected via the ear vein 1 h before ultrasound exposure. The ultrasonic procedure was consistent with a previous study 12 . For THP-1-derived macrophages and TNF-α-stimulated HUASMCs, the parameters of ALA-SDT (1 mM ALA, 0.5 W/cm 2 , 5 min ultrasound exposure time, 10% duty cycle) were the same as previously described 20 . Plasma cholesterol measurement. Rabbit blood was collected at day 0 (baseline), day 7 (balloon denudation time), day 11 (first harvest time) and day 38 (second harvest time). Plasma cholesterol was then measured by an automated analyser (AUC5800, Beckman Coulter, Inc., USA).

ALA-PpIX fluorescence intensity detection.
To delineate the time course of ALA-PpIX maximal accumulation in balloon-denuded arteries, the rabbits were sacrificed at 0, 30, 60, 90, 120 and 150 min after ALA injection (n = 3). The injured part of the femoral artery was removed, cut longitudinally to expose the intimal surface and washed thoroughly with phosphate buffered saline (PBS). Fluorescence emission spectra of PpIX were detected as described previously 23 . Emission spectra from 350 to 900 nm were collected every 3 mm across the entire area of the exposed intimal surface. Data were analysed by OriginPro7.0 software (OriginLab, Massachusetts, USA).
TNF-α-stimulated HUASMCs (2 × 10 5 cells/mL) seeded in 96-well plates were incubated with 1 mM ALA for 0, 2, 4, 8, 12 and 24 h. Then, the cells were washed three times with PBS, and ALA-PpIX fluorescence intensity was measured at 405 nm excitation and 630 nm emission wavelengths using a multimode microplate reader (Tecan Infinite M200 Pro, Mannedorf, Switzerland). www.nature.com/scientificreports/ Tissue preparation. The rabbits were euthanized by pentobarbital overdose (100 mg/kg). The arterial tree was perfused at 100 mmHg and stored in 4% paraformaldehyde for 24 h, embedded in paraffin and dissected for morphology studies. Each block of the injured artery was divided into 3 parts: both ends and a middle part. Serial transverse sections (5 μm) were cut at each part for histopathology and immunohistochemistry analysis. For immunofluorescence staining and ALA-PpIX fluorescence detection, the arteries were embedded in optimum cutting temperature compound (Sakura Fine Technical, Tokyo, Japan) and snap-frozen in liquid nitrogen.
Immunofluorescence staining. Frozen sections (7 μm) of rabbit arteries or HUASMCs were fixed, permeabilized and blocked, followed by incubation with primary antibody at 4 °C overnight. Rinsed sections or cells were then incubated with FITC-conjugated secondary antibody for 1 h at room temperature in the dark and with DAPI for another 2 min. The sections or cells were washed twice with PBS and analysed with a fluorescence microscope.
Colocalization of ALA-PpIX fluorescence and macrophages. One hour after ALA injection, the balloon-denuded arteries were snap-frozen and cryosectioned immediately for the fluorescence detection of PpIX. To detect the cellular localization of PpIX, the same cryosection was processed for immunofluorescence staining of macrophages, and colocalization was examined with a fluorescence microscope. For sections stained by immunohistochemistry, images were recorded with an Olympus IX70 microscope (Olympus, Tokyo, Japan) and analysed by Image Pro-plus 6.0 (Media Cybernetics, Rockville, MD, USA). The results of the stained area were expressed as percent positive area over the vascular media area or neointima. The number of macrophages and SMCs and the percentage ratio of macrophages and SMCs to the total cell number in neointima were semi-quantified. Re-endothelialization was scored for each cross-section by the extent of the circumference of the arterial lumen covered by endothelium: 1 = 25%; 2 = 25%-75%; and 3 = 75%-100% coverage, and data were expressed as averages of three independent scores.

Histopathology and immunohistochemistry.
Angiography. The angiography procedure was performed as previously described 33 . Optical coherence tomography (OCT) imaging. Seven normal healthy rabbits were euthanized as a baseline control. The OCT imaging procedure was performed as previously described 12 . Fractional flow reserve (FFR). The FFR measurement procedure was performed as previously described 12 . Western blots. The balloon-denuded arteries harvested at day 1 after ALA-SDT were cut into pieces (n = 3) and lysed in RIPA buffer containing PMSF on ice. THP-1-derived macrophages harvested at 6 h after ALA-SDT and HUASMCs harvested at 24 h after co-culturing with ALA-SDT-treated THP-1-derived macrophages were lysed in RIPA buffer containing PMSF on ice. The nuclear and cytosolic fractions were obtained by using a nuclear/cytosol fractionation kit (Beyotime). After quantification and denaturation, the proteins were electrophoresed in SDS-polyacrylamide gels and transferred onto PVDF membranes (Millipore, Schwalbach, Germany). The membranes were incubated with primary antibodies. Antibody labelling was detected using a ChemiDoc™ MP Imaging System (Bio-Rad, CA, USA) and quantified using Quantity One software analysis (Bio-Rad, USA). Data were collected from at least three independent experiments. Quantitative real-time PCR. Total RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, California, USA) and reverse-transcribed using ReverTra Ace qPCR RT Master Mix with gDNA Remover (TOY-OBO, Osaka, Japan). Then, 1 μl of cDNA was amplified in SYBR Green Real-time PCR Master Mix (TOYOBO) on an ABI 7500 Fast Sequence Detection System (Applied Biosystems, CA, USA). Gene expression values were normalized to β-actin. mRNA was quantified by the 2 −△△Ct method. Primers are listed in Supplementary Table 2.
Macrophage and HUASMC co-culture. THP-1-derived macrophages were divided into control and SDT groups. Six hours after ALA-SDT, THP-1-derived macrophages were harvested, transferred to the upper chamber of co-culture plates (Corning, NY, USA) at a concentration of 6 × 10 5 cells/ml and co-cultured with TNF-α-stimulated HUASMCs (3 × 10 5 cells/ml) in the lower chamber for 24 or 48 h using DMEM.
HUASMC proliferation assay. Ki67 immunofluorescence staining was used to detect proliferating HUASMCs after co-culture for 48 h with THP-1-derived macrophages treated with ALA-SDT. The percentage of Ki67-positive cells was calculated from total cells. At least 1,000 total cells from five random microscopic fields for each well at 20× magnification were counted. www.nature.com/scientificreports/ with ALA-SDT using serum-free DMEM, and their migration distance was recorded at 0, 12 and 24 h from the beginning of co-culture. Cell migration was photographed at identical locations for each time point. Phasecontrast images were recorded using a microscope (Olympus, Tokyo, Japan). The width of the wound in three random fields at 4× magnification in each well was analysed by Image Pro-plus 6.0 at each time point to determine the healed distance.

Enzyme linked immunosorbent assay (ELISA). The supernatants from THP-1-derived macrophages
and TNF-α-stimulated HUASMCs were collected at 4 h after ALA-SDT, and the supernatants from the coculture system were collected at 24 h after co-culture. TNF-α, IL-6 and IL-1β were measured using ELISA kits (Elabscience Biotechnology Co., Ltd., Wuhan, China) according to the manufacturer's instructions.
Flow cytometry analysis. Flow cytometry was performed as previously described 11 . Briefly, 6 h following ALA-SDT in the absence or presence of GW9662, THP-1-derived macrophages were collected and analysed by a fluorescence-activated cell sorting system within 1 h. Annexin V + /PI − and Annexin V + /PI + represent the apoptotic cells in the early and late phases, respectively. The data were analysed using BD FACSDiva Software 7.0 (Becton-Dickinson, USA).
Reactive oxygen species determination. Intracellular ROS generation was determined using DCFH-DA (Beyotime Biotechnology, Beijing, China) as described previously 20 . Briefly, seeded THP-1-derived macrophages (2 × 10 5 cells/ml) were pre-treated with NAC for 1 h and loaded with DCFH-DA for 30 min, followed by ALA-SDT treatment. Immediately after ALA-SDT treatment, the cells were analysed using a multimode microplate reader at 488 nm excitation and 525 nm emission wavelengths. Then, the cells were photographed under a fluorescence microscope (Olympus Corporation, Tokyo, Japan).
Confocal laser scanning. Confocal imaging was performed as previously described 11 . Briefly, six hours after ALA-SDT, THP-1-derived macrophages seeded in glass bottom cell culture dishes were stained with p65 and examined with a confocal laser scanning microscope (Zeiss, Germany).
Transcription factor profiling array. The activity of 96 different transcription factors in THP-1-derived macrophages was analysed by a Transcription Factor Activation Profiling Plate Array II (Signosis, Sunnyvale, CA, USA). Nuclear components were isolated using a Nuclear Extraction Kit (Signosis). In brief, both the control group and the SDT group contained ten samples from 35-mm Petri dishes of THP-1-derived macrophages.
In each group, the nuclear components were extracted 4 h after treatment from pooled 10 dish samples. Then, biotin-labelled probes were mixed with 15 μg of nuclear protein extract to form transcription factor/probe complexes. After separation, elution and hybridization, the captured DNA probe was detected with streptavidinhorseradish peroxidase (HRP), and signal intensity was measured with a microplateluminometer (Luminoskan Ascent, Thermo Scientific). Data were collected from three independent experiments.
Statistical analysis. All quantitative data are expressed as the mean ± SEM. The statistical analysis was performed by GraphPad 6.0 (La Jolla, CA, USA). The normality test (Shapiro-Wilk) was performed to determine whether the data were normally distributed. Student's unpaired test was used to determine the significant differences between two groups. One-way analysis of variance (ANOVA) with repeated measures and ANOVA followed by Tukey's test were used to determine the significance of differences among the groups in plasma cholesterol values and the other data, respectively. Two-way ANOVA with repeated measures followed by Dunnett's test was used to determine the significance of differences among the groups in HUASMC migration. A p value < 0.05 was considered to be statistically significant.

Data availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.