Matrix-Binding, N-Cadherin–Targeting Chimeric Peptide Inhibits Intimal Thickening but Not Endothelial Repair in Balloon-Injured Carotid Arteries

Abstract

BACKGROUND:

Treatment of occluded vessels can involve angioplasty, stenting, and bypass grafting, which can be limited by restenosis and thrombosis. Drug-eluting stents attenuate restenosis, but the current drugs used are cytotoxic, causing smooth muscle cell (SMC) and endothelial cell (EC) death that may lead to late thrombosis. N-cadherin is a junctional protein expressed by SMCs, which promotes directional SMC migration contributing to restenosis. We propose that engaging N-cadherin with mimetic peptides can act as a cell type–selective therapeutic strategy to inhibit polarization and directional migration of SMCs without negatively impacting ECs.

METHODS:

We designed a novel N-cadherin–targeting chimeric peptide with a histidine-alanine-valine cadherin-binding motif, combined with a fibronectin-binding motif from Staphylococcus aureus. This peptide was tested in SMC and EC culture assays of migration, viability, and apoptosis. Rat carotid arteries were balloon injured and treated with the N-cadherin peptide.

RESULTS:

Treating scratch-wounded SMCs with the N-cadherin–targeting peptide inhibited migration and reduced polarization of wound-edge cells. The peptide colocalized with fibronectin. Importantly, EC junction, permeability, or migration was not impacted by peptide treatment in vitro. We also demonstrated that the chimeric peptide persisted for 24 hours after transient delivery in the balloon-injured rat carotid artery. Treatment with the N-cadherin–targeting chimeric peptide reduced intimal thickening in balloon-injured rat carotid arteries at 1 and 2 weeks after injury. Reendothelialization of injured vessels after 2 weeks was unimpaired by peptide treatment.

CONCLUSIONS:

These studies show that an N-cadherin–binding and fibronectin-binding chimeric peptide is effective in inhibiting SMC migration in vitro and in vivo and limiting neointimal hyperplasia after balloon angioplasty without affecting EC repair. These results establish the potential of an advantageous SMC-selective strategy for antirestenosis therapy.

Highlights.

• We have designed a novel N-cadherin–targeting chimeric peptide with a matrix-binding domain.

• This peptide inhibited the migration of vascular smooth muscle cells but did not affect endothelial cell migration or function in vitro.

• This peptide delivered in vivo to rat carotid arteries was bound in the intima and inhibited intimal thickening after balloon injury without hindering endothelial repair.

Treatment of occluded blood vessels often involves angioplasty, stenting, or bypass grafting, but these approaches can be limited by restenosis, thrombosis, and graft failure.^1–3 Restenosis is caused by the directional migration of smooth muscle cells (SMCs) from the media to the intima, where they proliferate to become the dominant cell type in the neointima and impinge into the vascular lumen.⁴ The use of drug-eluting stents can effectively attenuate restenosis, but the cytotoxic drugs commonly used (ie, paclitaxel or sirolimus) contribute to endothelial cell (EC) dysfunction and death, leading to in-stent thrombosis and necessitating prolonged antiplatelet therapy.^5,6 Although advances in pharmacological and stent technology have enabled a reduction in stent failure, restenosis and stent thrombosis can still occur, and given the large number of percutaneous interventions performed annually, addressing these complications remains relevant.⁷ Further, whereas drug-eluting stents are commonly used in coronary arteries, balloon angioplasty alone is often used in peripheral arteries, with a consequent elevation in the incidence of restenosis.^7a Therefore, a cell type–selective therapy targeted at inhibiting SMC migration may offer a considerable advantage by reducing invasion of SMCs into the neointima without impairing EC regeneration to reestablish an antithrombotic covering.

See cover image

Balloon angioplasty with or without stenting causes mechanical injury to both ECs and SMCs. SMC migration and proliferation are activated in response to release of growth factors and cytokines from platelets and inflammatory cells depositing in the vessel wall.^8,9 SMCs also receive promigratory signals from adhesion molecules. N-cadherin is an adherens junction protein, which is a marker of the epithelial-to-mesenchymal transition in tumor growth and plays important roles in migration of cancer cells.¹⁰ N-cadherin is expressed in SMCs, and previous studies have described important roles for this protein regulating SMC migration, proliferation, and cell survival.^11–13 SMC-specific deletion of N-cadherin in mice reduced cell proliferation in response to vascular injury and stiffening.¹⁴ We have previously shown that N-cadherin is upregulated following experimental injury in vivo and in vitro.¹⁵ It localizes to posterior-lateral borders in wound-edge SMCs, an asymmetrical distribution essential for polarization of the microtubule organizing center (MTOC) to the leading edge of the migrating cell.¹⁶ MTOC polarization is a hallmark of directional cell migration and governs microtubule assembly and protein delivery to the leading edge.¹⁷ N-cadherin is also delivered to lamellipodial protrusions of migrating SMCs, where it functions to fuse protrusions and seal macropinocytotic vesicles.¹⁸ All of these activities are thought to facilitate cell migration.

Several studies have used N-cadherin mimetic peptides administered in soluble form in the cell culture media to attenuate cell migration and proliferation.^13,19–22 These peptides contain a histidine-alanine-valine (HAV) amino acid motif that forms a binding partner for a tryptophan residue on endogenous N-cadherin. We have designed a unique N-cadherin–targeting chimeric peptide that incorporates the HAV motif, as well as a matrix-binding domain, which includes a segment of FnBPA (fibronectin-binding protein A) from Staphylococcus aureus bacteria.²³ Our intention is to tether the peptide to fibronectin abundant in the vessel wall matrix and simultaneously engage N-cadherin on adjacent SMCs as they extend protrusions to migrate into the intima. The targeting peptide will thus restore receptor symmetry around the SMC to inhibit directional migration. With selective targeting of N-cadherin, the aim is to leave EC migration undisturbed, since the latter does not depend on N-cadherin but utilizes VE (vascular endothelial cell)-cadherin.¹⁶ In the present study, we report on the results from experiments using these peptides, in vitro and in vivo, in models of SMC migration after injury.

MATERIALS AND METHODS

In accordance with the Transparency and Openness Promotion Guidelines, the authors will make the data available upon reasonable request.

N-Cadherin–Targeting Peptides

Several putative N-cadherin–targeting peptides were designed and 3 versions chosen for in vitro testing (Figure 1). The chimeric peptide included a fibronectin-binding domain based on Fnbp1 from the S aureus bacteria²³ and an N-cadherin–binding domain based on the HAV amino acid motif responsible for homotypic cadherin binding. Short (linear) and cyclic versions were also designed lacking the fibronectin domain but maintaining the N-cadherin mimetic element and HAV-binding motif. The short and cyclic peptide sequences were derived from examples previously tested in the literature,^19,20 for the purposes of comparison to our novel matrix-anchoring configuration. Criteria for selection of chimeric N-cadherin mimetic peptides were inclusion of HAV domain with flanking residues for N-cadherin specificity, feasibility of synthesis (not excessive hydrophobicity, length), and inclusion of a matrix-anchoring domain from the literature with sufficient evidence of utility and affinity for a relevant matrix protein (fibronectin) found in abundance in injured arteries. Each peptide was paired with a scrambled control version in which the N-cadherin–binding domain contained the identical amino acids in a randomized sequence. Additionally, 5-carboxytetramethylrhodamine fluorescent tags were added to peptides in some experiments to allow for visualization by fluorescence microscopy. For experiments not requiring this, untagged versions of these peptides were used. Peptides were synthesized de novo by Lifetein, Inc (Somerset, NJ) to a purity >95% and shipped as lyophilized powder. Peptides were dissolved in pure dimethylsulfoxide to a concentration of 100 mg/mL before being diluted in cell culture media for in vitro administration or in PBS for use in vivo.

Figure 1.

N-cadherin (NCad)–targeting peptide design and sequences. Peptides were designed and synthesized de novo by Lifetein (Somerset, NJ). Chimeric peptides include a fibronectin-binding domain derived from Staphylococcus aureus FnbpA (fibronectin-binding protein A).

Cell Culture

Immortalized mouse vascular arterial SMCs (MOVASs; ATCC [American Type Culture Collection]; CRL-2797) were cultured in DMEM (11885084; Life Technologies, Carlsbad, CA) with 10% fetal bovine serum (12483020; Thermo Fisher Scientific, Waltham, MA) and 2% penicillin/streptomycin (15140122; Thermo Fisher Scientific). Rat neointimal SMCs (RNISMCs) were isolated from the intima of balloon-injured male Sprague-Dawley rat carotid arteries as we have described previously²⁴ and cultured in DMEM with 10% fetal bovine serum and 2% penicillin/streptomycin. Human umbilical vein ECs (HUVECs) and human arterial SMCs (HASMCs) were purchased from Lonza (Basel, Switzerland) and cultured in EGM-2 (endothelial growth medium-2; Lonza) or SmGm-2 (smooth muscle growth medium-2; Lonza) media supplemented with the included bullet kit according to the manufacturer’s instructions. RNISMC, HUVEC, and HASMC were used at passages 2 to 10, and MOVAS cells were used at passages 20 to 50. For scratch wound assays, cells were plated at a seeding density of 150 000 cells/mL in 24-well dishes, which had been coated with 20 μg/mL fibronectin (10838039001; Roche, Basel, Switzerland) for 1 hour at 37 °C and washed once with PBS (14190144; Thermo Fisher Scientific). Cells were incubated for 24 hours, scratched with a 200-μL pipette tip guided along a straightedge, washed with PBS, and media replaced. Short, cyclic, or N-cadherin–targeting chimeric peptides were diluted from 100 mg/mL in dimethylsulfoxide, by addition to DMEM with 0.1% fetal bovine serum and 2% penicillin/streptomycin. Peptides were used at final concentrations of 125 to 500 μg/mL based on previous similar studies.^11–13,16 Cells were treated for 24 hours (MOVAS) or 48 hours (RNISMC, HASMC, and HUVEC) and imaged to visualize the scratch wound area. The wound gap was manually traced, and the area measured using Nikon NIS Elements software. The wound gap area was subtracted from, and then divided by, the initial area of the wound to give a final measurement for percentage of wound area that had closed. For viability assays, cells were plated at 10 000 cells per well in fibronectin-coated (20 μg/mL) 96-well plates and cultured for 24 hours. Cells were treated with peptides for 24 hours before adding water-soluble tetrazolium salts (05015944001; Roche) to assess metabolic activity by changing color in response to reduced NAD(P)H (nicotinamide adenine dinucleotide phosphate) oxidation, which was recorded by measuring absorbance at 450 and 610 nm in a plate reader (PerkinElmer EnSpire Multimode Plate Reader). Comparing the cultures treated with vehicle, or the different scrambled control peptides that mimicked the active peptides, there were no statistically significant differences in wound closure, cell viability, or polarization (Figure S1A through S1D). Therefore, the data from these control groups were pooled for ease of visualization.

Immunostaining

Cells were fixed in 4% paraformaldehyde (or 100% ice-cold methanol for γ-tubulin) for 10 minutes and permeabilized in 0.2% triton X-100 for 10 minutes at room temperature. Cells were blocked in 0.2% BSA for 1 hour and stained overnight with primary antibodies: anti-N-cadherin (rabbit-anti-mouse, 1:800, 18203; Abcam), anti-γ-tubulin (rabbit-anti-mouse, 1:800, ab11317; Abcam), anti-fibronectin (rabbit-anti-mouse, 1:1000, ab23750; Abcam), and anti-cleaved-caspase 3 (CC3; rabbit-anti-mouse, 1:1000, 9661; Cell Signalling Technologies). All antibodies used were polyclonal and raised in rabbit against mouse epitopes. Cells were washed thrice with PBS and stained with 4′,6-diamidino-2-phenylindole (10236276001; Roche), phalloidin 488 or 568 (A12379, A12380; Invitrogen), or secondary antibodies: goat-anti-rabbit Alexa488 (A-11008; Invitrogen), goat-anti-rabbit Alexa568 (A-11011; Invitrogen) at 1:1000. Coverslips with cells were washed and mounted on slides in ProLong Gold antifade mounting media (P10144; Invitrogen). Images were obtained using a laser-scanning confocal microscope (Olympus FV3000).

Cell Polarity Measurement

For MTOC polarization experiments, cells were scratch-wounded and treated with 500 μg/mL short, cyclic, or chimeric N-cadherin–targeting peptides or their respective scrambled controls; then 6 hours post-injury, cells were fixed in ice-cold methanol for 10 minutes. Cells were stained for γ-tubulin to visualize the MTOC. Cells at the wound edge were imaged by confocal microscopy using a ×60 oil immersion objective lens to generate fields of view containing 4 to 10 wound-edge cells. Cells were grouped into 2 categories based on the position of the MTOC relative to the wound gap and the nucleus. A virtual line was drawn through the center of each nucleus parallel to the wound gap. Cells with an MTOC in front of this line were considered polarized, and those with an MTOC behind this line were considered nonpolarized. At least 20 fields of view were analyzed from at least 3 separate wells for each condition. In a field of nonmigrating, randomly polarized cells, 50% of cells would have the MTOC located to one side of the nucleus. Therefore, 50% is considered the baseline measured in nonpolarized cells. The 50% threshold representing nonpolarized cells is indicated by a dotted line on the graph of polarization (Figure 3C).

Figure 3.

N-cadherin–targeting chimeric peptide inhibited the polarization of wound-edge smooth muscle cells (SMCs). Mouse aortic SMCs were scratched and treated with 500 μg/mL short, cyclic, or N-cadherin–targeting chimeric peptide for 6 hours. A, Cells were fixed and stained for g-tubulin to label the microtubule polarizing center (MTOC), indicated by the white arrow. B, The position of the MTOC was determined relative to a centerline through the nucleus parallel to the wound gap for each cell per field of view. C, The percentage of polarized cells per field of view (FOV) is shown with means±SEM in n≥6 biological replicates. Scrambled control peptide data were pooled. Statistical significance was determined by 1-way ANOVA with Sidak multiple comparisons. DAPI indicates 4′,6-diamidino-2-phenylindole.

Animal Surgery

Animal experiments were performed in accordance with the guidelines of the Canada Council on Animal Care, with the approval of the University of Toronto Temerty Faculty of Medicine Animal Care Committee. Male Sprague-Dawley rats weighing 350 to 400 g were purchased from Charles River. Only male animals were used in experiments to eliminate potential confounding factors from the estrogen cycle on vascular remodeling. Carotid balloon injury surgery was performed as we have described previously, by inflating and withdrawing a 2F embolectomy catheter (Edwards) 3× through the carotid artery.^24,25 Following balloon injury, 500 μg/mL of N-cadherin–targeting chimeric or scrambled control peptide dissolved in 100 μL of PBS, also containing heparin at 50 IU/mL, was infused through a soft silicone sheath introduced over a 20G needle through the arteriotomy near the site of injury, and the vessel segment was held closed by metal clamps for 5 minutes to allow incubation of the chimeric peptide in the carotid segment. The animal surgeon was blinded regarding the identity of the active peptide or control treatment; so animals were selected at random for which treatment they received. The clamps were released and blood flow restored. Animals were allowed to recover for 24 hours (for peptide residency experiments), 1 week, or 2 weeks (for neointima formation experiments) and then sacrificed by CO₂ asphyxiation and bilateral thoracotomy. Animals were immediately perfusion fixed with 4% paraformaldehyde for 5 minutes at 120 mm Hg, through a needle inserted directly into the heart, to preserve vessel physiological shape and structure for histological or immunofluorescence microscopy. Carotid segments were excised and processed into cross sections 4 µm thick for mounting on slides, histological staining with hematoxylin and eosin, and imaging by conventional bright-field microscopy (Nikon Eclipse Ci upright microscope). Luminal, intimal, and medial areas were measured (Nikon NIS Elements software) by manually selecting each area landmarked by cells and elastic lamellae. Luminal area was determined by measuring the circumference and applying the formula for area of a circle (A=πr² and C=2πr) to correct for vessel deformation during processing. These measurements were used to calculate specific areas for each arterial layer and the intima/media ratio. Separate groups of animals and carotid segments were harvested 24 hours after injury, and the carotids opened longitudinally and imaged en face, for assessment of fluorescent peptide binding with confocal microscopy (Olympus FV3000). To measure reendothelialization, rats were injected with 0.5 mL of 5% Evans blue dye dissolved in sterile PBS through the tail vein. At 30 minutes following dye administration, animals were euthanized and perfusion fixed as described. Vessel segments were prepared en face and imaged using a dissection microscope (Nikon). Using the ImageJ software, unstained area was divided by total vessel area and multiplied by 100% to determine the percentage of the intimal surface with an intact EC barrier. Cross sections of carotid arteries were also stained by immunohistochemistry for Ki67 to assess cell proliferation. Perfusion-fixed rat carotid arteries were paraffin embedded and cross-sectioned on to microscope slides. Deparaffinization with stepwise ethanol baths was followed by permeabilization in 0.2% triton, blocking in 0.2% BSA, and incubation with primary anti-Ki67 antibody (1:1000, NB110-89717; Novus Biologicals). The total number of cells, and the total positive for Ki67, were counted in the intimal and medial layers, and the percentage of Ki67-positive cells was calculated. Rat carotid arteries were excluded from analysis in the event of one of the following criteria being met: premature death of the animal or artifactual dissociation of the intima. Two animals died prematurely and 1 had defective processing resulting in artifactual tissue sections for 2-week scrambled control, 1 animal had defective processing for 2-week N-cadherin peptide treatment group, and 1 animal died prematurely for reendothelialization in 2-week N-cadherin peptide–treated group.

Statistical Analysis

All statistical analyses were performed with the Prism 9 software (GraphPad). A value of P<0.05 was considered statistically significant. Where n was <6, a nonparametric analysis with Mann-Whitney U test was performed. Where n was ≥6, a parametric analysis was performed after analyzing for normality using the Kolmogorov-Smirnov test, and data were analyzed by Student t tests for pairwise comparisons. Two-way ANOVA with a Tukey correction for multiple analyses was used for comparisons between multiple groups where appropriate. Where indicated, multiple fields of view from each well were considered technical replicates and averaged together. Values from separate wells were then considered biological replicates for analysis. For the intima/media ratio calculation, measurements were made from cross sections taken from three locations along each injured carotid artery specimen corresponding to ¼, ½, and ¾ along the length of the injured segment. These measurements were averaged for each animal, and then the overall mean was calculated. Sample sizes were 6 animals per group for 1-week intimal thickening; 7 scrambled control and 9 N-cadherin–targeting chimeric peptide-treated animals for 2-week intimal thickening; 6 scrambled control and 5 N-cadherin–targeting peptide-treated animals to measure reendothelialization, 3 animals per condition for Ki67 staining at 1 week; 3 and 5 animals for scrambled control and N-cadherin–targeting peptide-treated animals for Ki67 at 2 weeks; and 3 animals per condition for CC3 at 1 week.

RESULTS

N-Cadherin–Targeting Chimeric Peptide Inhibited Migration of Rodent SMCs

Three versions of N-cadherin peptides, including short, cyclic, and the chimeric N-cadherin–targeting peptide with a matrix-binding domain (Figure 1), were designed and synthesized by solid-phase peptide synthesis (Lifetein, Inc). Scratch-wounding assays were performed to assess cell migration in the presence of N-cadherin–targeting peptide compared with controls. In MOVAS cells, the N-cadherin–targeting chimeric peptide significantly reduced wound closure at 250 or 500 μg/mL compared with the scrambled control-treated cells (Figure 2A). In RNISMC, the N-cadherin–targeting chimeric peptide significantly reduced wound closure at 250 or 500 μg/mL (Figure 2B). Neither the short nor the cyclic peptide significantly affected migration of MOVAS or RNISMC. There were no statistically significant differences in wound closure in cultures treated with vehicle or the different scrambled control peptides, which mimicked the active peptides (Figure S1A through S1D). The data from these scrambled control peptide groups are pooled for ease of comparison in Figure 2.

Figure 2.

N-cadherin–targeting chimeric peptide inhibited migration of rodent smooth muscle cells (SMCs) without reducing viability or inducing apoptosis. Scratch-wounded SMCs were treated with N-cadherin–targeting peptides or scrambled control peptides and imaged after 24 hours for mouse aortic SMC (MOVAS; A) or 48 hours for rat neointimal SMC (RNISMC; B). The area of the wound gap was subtracted from and divided by the initial gap area to give a percentage of the wound that had closed. For viability measurements, MOVAS (C) or RNISMC (D) was treated with N-cadherin mimetic peptides or scrambled control peptides and incubated for 24 hours. Metabolic activity was assessed by colorimetric change of water-soluble tetrazolium salt, measured as absorbance at 450 nm, with background at 610 nm subtracted. Apoptosis was assessed by staining for cleaved caspase 3 (CC3) in MOVAS treated with N-cadherin–targeting peptides or scrambled controls (E) and RNISMCs (F). Scrambled control peptide data were pooled. Data shown are means±SEM; n≥3 experimental replicates. Statistical significance was calculated by 2-way ANOVA (A–D), Kruskal-Wallis (E), or Mann-Whitney U (F) tests.

N-Cadherin–Targeting Peptides Did Not Impact Viability or Apoptosis of Rodent SMCs

MOVASs or RNISMCs were cultured in 96-well dishes and treated with N-cadherin peptides, scrambled control peptides, or vehicle for 24 hours. A water-soluble tetrazolium salt assay was performed to assess viability, measuring cell metabolism as a proxy, for the amount of viable cells present in each well. There were no statistically significant differences in viability in MOVAS or RNISMC treated with N-cadherin peptides compared with controls (Figure 2C and 2D). Apoptosis was measured by staining MOVAS cells for CC3 after 24 hours of peptide treatment at 500 μg/mL. There was no significant difference in the percentage of CC3+ cells in the peptide-treated groups compared with the controls (Figure 2E). Apoptosis was also assessed in RNISMCs, which were treated with 500 μg/mL of scrambled or N-cadherin–targeting peptide for 24 hours and stained for CC3 (Figure 2F), with no significant difference between groups. Each scrambled control peptide was tested individually, and differences between the peptides were not statistically significant (Figure S1B and S1E); therefore, the data from these groups were pooled for ease of comparison in Figure 2.

N-Cadherin–Targeting Chimeric Peptide Inhibited the Polarization of Wound-Edge SMCs

To determine whether cell polarity of wound-edge migrating VSMCs was inhibited by N-cadherin peptides, MOVAS cells were cultured to confluency in 24-well dishes and scratch wounded. Cells were treated with N-cadherin targeting peptides or scrambled controls, then fixed and stained for γ-tubulin, a marker for the MTOC (Figure 3A). The position of the MTOC was analyzed relative to the centerline of the nucleus, indicated by a line drawn through the center of the nucleus parallel to the wound gap (Figure 3B), and cells in which the MTOC was forward of this centerline relative to the direction of migration were counted as polarized. The N-cadherin–targeting chimeric peptide significantly reduced the number of polarized wound-edge cells by 21.7% (Figure 3C). The scrambled peptides did not significantly affect cell polarization (Figure S1C); so data from the scrambled control groups are pooled for ease of comparison in Figure 3. Our binary classification of front polarization would give a 50% result for a field of nonmigrating cells, which is indicated by a dotted line on the graph.

N-Cadherin–Targeting Chimeric Peptide Inhibited Migration but Did Not Affect the Viability of Human SMCs

HASMCs were subject to scratch wounding, then treated with N-cadherin–targeting chimeric peptide or scrambled control peptide for 48 hours. In HASMCs, the N-cadherin–targeting chimeric peptide significantly reduced wound closure at a dose of 500 μg/mL (Figure 4A). There were no statistically significant effects of N-cadherin–targeting chimeric peptides on viability of HASMCs (Figure 4B).

Figure 4.

N-cadherin–targeting chimeric peptide inhibited migration of human smooth muscle cells (SMCs), but not endothelial cells (ECs), and did not impact viability. Scratch-wounded human aortic SMCs (HASMCs; A) or human umbilical vein endothelial cells (HUVECs; C) were treated with N-cadherin–targeting chimeric peptide or scrambled control peptide and imaged after 48 hours. The area of the wound gap was subtracted from and divided by the initial gap area to give a percentage of the wound that had closed. For viability measurements, HASMCs (B) or HUVECs (D) were treated with N-cadherin–targeting chimeric peptide or scrambled control peptide and allowed to proliferate for 24 hours. Metabolic activity was assessed by colorimetric change of water-soluble tetrazolium salt, measured as absorbance at 450 nm, with background at 610 nm subtracted. Apoptosis was measured in HUVECs by staining for cleaved caspase 3 (CC3) and counting positively stained cells (E). Permeability was measured in HUVECs grown on transwell membranes with 0.4-μm pores where the top chamber media replaced with 5 μM fluorescein isothiocyanate (FITC)–dextran and N-cadherin peptide or scrambled control. Media in the abluminal (bottom) chamber was sampled into a 96-well dish and luminescence at 530 nm read by plate reader. FITC-dextran concentration calculated against a standard curve (F). Data shown are means±SEM; n≥3 biological replicates. Statistical significance was determined by 2-way ANOVA with Tukey multiple comparisons (A–D) or Mann-Whitney U tests (E and F).

N-Cadherin–Targeting Chimeric Peptide Did Not Impact Human EC Migration, Viability, Apoptosis, Permeability, Junctions or Activation

In HUVECs that were scratch wounded and treated with peptides, there was no significant impact of N-cadherin–targeting chimeric peptide treatment on cell migration (Figure 4C). Viability assessed using a water-soluble tetrazolium salt assay was not significantly altered by N-cadherin–targeting peptide (Figure 4D). Apoptosis was measured by staining HUVECs for CC3 and was not significantly altered by the N-cadherin–targeting peptide compared with scrambled control (Figure 4E). Permeability of fluorescein isothiocyanate–dextran was assessed in transwells containing confluent monolayers of HUVECs and was not significantly altered by N-cadherin–targeting peptide compared with the scrambled control (Figure 4F). Immunofluorescence staining of HUVECs for VE-cadherin was performed to assess cell junctions, and junctional staining was not altered by N-cadherin–targeting peptide treatment (Figure S2). Expression of adhesion molecules VCAM1 (vascular cell adhesion molecule 1), ICAM1 (intercellular adhesion molecule 1), and the antithrombotic protein thrombomodulin were measured by RT-qPCR (real-time quantitative polymerase chain reaction) after treatment of HUVECs for 24 hours with N-cadherin–targeting or scrambled control peptides and were not significantly altered (Figure S3).

N-Cadherin–Targeting Chimeric Peptide Colocalized In Vitro With Fibronectin and Was Retained In Vivo in Balloon-Injured Carotid Arteries

To determine whether the N-cadherin–targeting chimeric peptide could bind with fibronectin, RNISMCs were plated on fibronectin-coated coverslips, treated with 100 μg/mL tetramethylrhodamine-labeled N-cadherin–targeting or scrambled peptide and stained to visualize fibronectin by immunofluorescence microscopy. Significant colocalization of fluorescent peptide and fibronectin was observed, regardless of whether the N-cadherin–binding domain was active (Figure 5A) or scrambled (Figure 5B). To assess binding of peptide to damaged vessel wall in vivo, the carotid arteries in male Sprague-Dawley rats were injured using a balloon catheter to denude the endothelium. Immediately after balloon injury, the common carotid was clamped at both ends, and 100 μL of PBS with 500 μg/mL tetramethylrhodamine-labeled chimeric peptide was instilled locally for 5 minutes. After this, the clamps were removed, and blood flow was restored. Twenty-four hours later, binding of the tetramethylrhodamine-labeled peptide was confirmed by confocal microscopy of en face vessel segments (Figure 5C), whereas PBS-only control-treated vessels lacked any red fluorescence (Figure 5D). Autofluorescence of the elastic lamellae was evident as green signal, and Z stacks to visualize tissue in 3 dimensions revealed that the N-cadherin–targeting peptide was localized on the intimal surface of the vessel, as well as within the top layers of the media.

Figure 5.

N-cadherin–targeting chimeric peptide was colocalized in vitro with fibronectin and was found in vivo in balloon-injured rat carotid arteries. Rat neointimal smooth muscle cells were treated for 24 hours with tetramethylrhodamine (TAMRA)-labeled N-cadherin–targeting chimeric peptide (A) or scrambled control peptide (B). Cells were fixed and stained with 4′,6-diamidino-2-phenylindole (DAPI; nuclei), phalloidin (actin), and anti-fibronectin and imaged by confocal microscopy. Rat carotid arteries were balloon injured and infused with TAMRA-labeled N-cadherin–targeting chimeric peptide (C) or PBS only (D) for 5 minutes and allowed to recover for 24 hours. Vessels were perfusion fixed with paraformaldehyde and prepared en face for confocal microscopy.

N-Cadherin–Targeting Chimeric Peptide Inhibited Intimal Thickening in Injured Rat Carotid Arteries, Without Impairing Reendothelialization

Carotid arteries were injured using a balloon catheter to denude the endothelium. This disrupted the cell adhesions between ECs and SMCs and triggered migration of SMCs to the intima. Immediately following balloon injury, the common carotid artery segment was clamped and infused with N-cadherin–targeting chimeric or scrambled control peptide at a concentration of 500 µg/mL and a volume of 100 μL for 5 minutes, followed by removal of the clamps, restoration of blood flow, and closure of the surgical incisions, and the rats were allowed to recover. There was a marked reduction in intimal thickening in the vessels treated with the N-cadherin–targeting chimeric peptide compared with those treated with scrambled control peptide, which was apparent at both 1 and 2 weeks after balloon injury (Figure 6A). Measurement of the ratio of intima/media area revealed significant reductions in N-cadherin–targeting chimeric peptide-treated vessels compared with scrambled control peptide-treated vessels (63.8% reduction at 1 week and 32.4% reduction at 2 weeks; Figure 6B). Cell proliferation was assessed by immunostaining for Ki67, which is expressed in proliferating cells (Figure S4A). The percentage of medial cells proliferating was low (≈1%) at 1 and 2 weeks after balloon injury. Intimal cell proliferation was the highest at 1 week (≈30%), then declined to ≈7% at 2 weeks post-injury. There were no significant differences in the percentage of proliferating cells in the intima or media comparing N-cadherin–targeting and scrambled control peptide–treated rats at 1 or 2 weeks after balloon injury (Figure S4A and S4B), nor were there changes in the levels of apoptotic cells at 1 week as measured by CC3 staining (Figure S4C and S4D).

Figure 6.

N-cadherin–targeting chimeric peptide inhibited intimal thickening in rat carotid arteries without impairing reendothelialization. Carotid arteries were subject to balloon catheter denudation of the endothelium; then the carotids were locally infused with N-cadherin–targeting chimeric or scrambled control peptides for 5 minutes. Animals were allowed to recover for 1 or 2 weeks; then the carotids were perfusion-fixed and artery sections stained with H&E for histological analysis (A), and intima/media ratio was calculated for 1 week post-injury (B) and 2 weeks post-injury (C). For scrambled control (black) and N-cadherin peptide (red), n=6 and n=6 at 1 week, n=7 and n=9 at 2 weeks, respectively. Two weeks following injury and peptide treatment, animals were injected with Evans blue dye to measure reendothelialization, and photomicrographs of stained vessels (D) were used to measure blue stained and unstained areas. E, The percentage of unstained area—a measure of endothelial cell coverage—was compared between groups. n=6 for scrambled control and n=5 for N-cadherin–targeting peptide-treated groups. Data shown are means±SEM. Statistical significance was determined by the Student t test and normality by the Kolmogorov-Smirnov test.

Reendothelialization was measured by staining with Evans blue dye, which binds albumin in the circulation and stains artery segments lacking an endothelium but not those with an intact endothelial barrier. An uninjured carotid (white) and a fully denuded ballooned carotid artery (blue) were dissected and exposed en face and are shown in Figure S5. Balloon-injured carotid arteries from rats treated with scrambled control peptide and the N-cadherin–targeting chimeric peptide are shown in Figure 6C. There was no significant difference in the percentage of endothelialized intimal surface comparing vessels treated with N-cadherin–targeting chimeric peptide and the scrambled control peptide (Figure 6D). This suggests that the N-cadherin–targeting peptide did not impair the endothelial healing response.

DISCUSSION

Current interventional treatments for atherosclerosis such as angioplasty and stenting are undermined by damage to the endothelium and can fail due to restenosis and thrombosis.^1–3 These limitations highlight the need for a cell type–selective therapy that spares EC repair while inhibiting SMC-driven intimal thickening. In the current study, we have conceived and characterized a uniquely configured chimeric peptide with N-cadherin–binding and fibronectin-binding domains. We have demonstrated in vitro that targeting N-cadherin with this peptide inhibits SMC migration without impacting EC migration, junction formation, or permeability. We have used SMCs and ECs from different species including mouse, rat, and humans. Furthermore, the N-cadherin–targeting peptide bound to fibronectin and reduced the polarization of wound-edge SMCs, an important hallmark of directional cell migration that drives intimal thickening and restenosis. We have also demonstrated that the chimeric N-cadherin–targeting peptide was deposited in the walls of arteries in vivo and inhibited intimal thickening after vascular injury without impairing reendothelialization.

N-cadherin adhesions make connections between adjacent SMCs and also between SMCs and the overlying EC layer.²⁶ When these adhesions are interrupted, for example, by scratch wounding in vitro or by EC denudation in vivo, the SMCs are polarized with an asymmetrical distribution of N-cadherin at the lateral and trailing edges of the cells. We propose that adding the chimeric peptide to engage N-cadherin at the cell leading edge restores receptor symmetry and stops polarized migration. It may be particularly effective to focus on targeting early SMC migration, as approximately half of all neointimal SMCs are thought to derive from the migration of cells from the media and contributions from circulating and adventitial-resident progenitor cells, with the other half arising as a result of proliferation from the migrated cells.²⁷ Thus, arresting early SMC migration may also circumvent later proliferation, contributing to the reduction in intimal thickening. In the current study, we have utilized the rat carotid artery balloon injury model with complete EC denudation, which does not represent the condition of atherosclerosis, where plaques can develop under the EC layer. However, in atherosclerosis, the accumulation of lipids and inflammatory cells in the intima will also disrupt contacts between ECs and SMCs, thus polarizing SMCs at the edge of the media and leading to directional migration to the intima. Future studies should test the peptide in models with preexisting atherosclerosis.

Our data are largely consistent with previous studies that used antagonist peptides, neutralizing antibodies, or dominant negative N-cadherin to inhibit SMC migration and proliferation and attenuate intimal thickening in an organ culture model using the human saphenous vein.^11,28 However, whereas others found that inhibition or disruption of VSMC N-cadherin may induce apoptosis,^11–13 our N-cadherin–targeting chimeric peptide did not cause apoptosis. We have shown previously that N-cadherin is delivered continually to the membrane at the leading edge of migrating SMCs, and if another cadherin-expressing cell is encountered, N-cadherin binds and is stabilized at the cell membrane, but if another cadherin-expressing cell is not encountered, the receptor is internalized.¹⁸ We suggest that contact with the targeting peptide mimics the engagement of the SMC leading edge with an adjacent cell, engaging and stabilizing the adhesion site. The extracellular domain of N-cadherin engages neighboring cells through homotypic transinteractions,²⁹ while the cytoplasmic tail of the N-cadherin molecule is reported to bind catenins and actin filaments, forming a link between adherens junctions and the cytoskeleton and stabilizing N-cadherin adhesions.^30,31 We suggest that engagement of N-cadherin stabilizes the receptor at the cell surface and inhibits migration without causing apoptosis as previously reported peptides have. N-cadherin signals through Rho family GTPases, β-, α-, and p120 catenins, and growth factor tyrosine kinase receptors (ie, FGF [fibroblast growth factor]) to impact gene expression and regulation of migration and proliferation.³²

The fibronectin-binding domain was included in the chimeric peptide to take advantage of the high abundance of fibronectin in the vessel matrix and atherosclerotic lesions.^33,34 By binding the chimeric peptide to fibronectin, we aimed to target SMCs where they are most vulnerable to intimal migration. Moreover, the functions of cadherins are influenced by microenvironmental cues from matrix molecules such as fibronectin, which binds to integrins abundant in the protrusions of migrating cells and is intimately involved in cell adhesion and migration.^35–37 We reasoned that when the chimeric peptide binds to fibronectin, it will present the N-cadherin–binding domain at high concentration near cell protrusions where it is especially effective to engage cellular N-cadherin receptors. Subcellular structures including invadopodia, filopodia, and podosomes enable cell migration and adhesion and are important sites for integrin and cadherin cross talk.^38,39 Indeed, our novel N-cadherin–targeting chimeric peptide showed greater inhibition of migration and polarity compared with the short linear and cyclic variants. Another possible reason for this is that anchoring the peptide to the fibronectin matrix introduces mechanical resistance. N-cadherin is involved in cellular mechanosensing,^40–42 and increased resistance may enhance signaling.⁴³ The mechanisms of action of the N-cadherin–targeting chimeric peptide at cell protrusions in concert with matrix-sensing pathways and the interactions with endogenous cadherin-based mechanosensory pathways will be explored in future studies.

In the present study, we have demonstrated in vivo delivery of the N-cadherin–targeting chimeric peptide immediately after balloon injury, and retention in the matrix of the rat carotid artery for at least 24 hours. This was achieved using a transient intraluminal application of the peptide in solution to the injured carotid. Balloon injury in the carotid artery is followed by thickening of the neointimal layer caused by migration of SMCs from the media and proliferation in the intima. Early and transient treatment with the N-cadherin–targeting peptide resulted in a significant inhibition of neointimal thickening at 1 and 2 weeks after injury. SMC migration from the media to the intima is the predominant response early after balloon injury, and inhibition at this stage will leave fewer intimal cells to undergo subsequent proliferation. Our results are, therefore, consistent with the notion that inhibition of early SMC migration is sufficient to attenuate neointimal growth. Changes to proliferation of SMCs in the arterial media or intima were not detected, but others have previously observed reduced proliferation following administration of N-cadherin antagonists to isolated SMCs in vitro and in saphenous vein organ culture ex vivo,¹³ where there is likely much greater exposure of cells to antagonists. We are currently investigating more robust delivery methods, delivering peptides complexed with nanoparticles, which might result in attenuation of cell proliferation.

Following balloon injury, ECs regrow from the ends of the artery to cover the denuded surface.⁴⁴ Importantly, EC repair was not inhibited by N-cadherin–targeting peptide treatment; 80% reendothelialization of the carotid artery surface was achieved at 2 weeks post-injury. This is in agreement with our in vitro studies that showed that HUVEC migration was not affected by N-cadherin–targeting peptides and with previous work demonstrating that N-cadherin inhibition did not alter EC polarity during migration.¹⁶ Additional in vitro experiments confirmed that endothelial function was preserved, as junction stability and permeability were not impacted by treatment with the N-cadherin–targeting peptide. Expression of inflammatory or antithrombotic markers was also unchanged. Whereas other N-cadherin agonists may impact endothelial tube formation or adhesion,¹⁹ our data are consistent with the notion that N-cadherin–targeting peptides act in an SMC-selective manner and do not disturb endothelium integrity, a potentially advantageous combination for reducing restenosis without late thrombosis.

There are some limitations to the experimental methods used in the present study. First, the animal model of healthy Sprague-Dawley rats does not represent the reality of clinical angioplasty because the animals do not have preexisting atherosclerosis. Future studies should incorporate animal models of atherosclerosis with N-cadherin–targeting peptide treatment. Furthermore, the administration method of the therapeutic peptide by local delivery in solution and incubation for 5 minutes would not be practical for clinical use. Studies are ongoing to explore coating balloons with peptide and biocompatible nanoparticles that would enable tailored release kinetics and local delivery for efficient and translatable delivery of therapeutics to arteries in the clinic.

The present study establishes a basis for future translational work involving N-cadherin–targeting peptide therapies that inhibit SMC migration while allowing for EC repair. Our work also demonstrates that the fibronectin-binding matrix tether we derived from S aureus FnbpA may be useful for local drug retention. We envision that such therapy would be useful for the treatment of coronary and peripheral artery disease, alleviating the need for cytotoxic agents and for long-term antiplatelet treatments.

ARTICLE INFORMATION

Sources of Funding

The work was supported by grants-in-aid from the Heart and Stroke Foundation of Canada to M.P. Bendeck (G-18-0022067 and G-21-10031541) and a New Frontiers in Research Foundation Exploration Grant (2020-00180) to M.P. Bendeck and J.P. Santerre. J. Burke-Kleinman was supported by an Ontario Graduate Scholarship. J. Rubianto was supported by a scholarship from the Natural Sciences and Engineering Research Council of Canada.

Disclosures

None.

Supplemental Material

Supplemental Materials and Methods

Figures S1–S5

Major Resources Table