Synergistic fusion of CD47, VE-cadherin and mussel adhesion protein promotes endothelialization and suppresses inflammation in vascular stents

Abstract

Endothelial cell (EC)-specific coatings for vascular stents are crucial for enhancing their biocompatibility and preventing complications such as restenosis and thrombosis. This study developed an innovative CD47-VE-cadherin-Mfp5 (CD47-VE-M) fusion protein coating for cardiovascular stents that integrates three distinct functional domains: endothelial adhesion enhancement (VE-cadherin EC1-2), macrophage inhibitory signaling (CD47), and substrate adhesion reinforcement (Mfp5). In vitro, CD47-VE-M coatings significantly promoted EC adhesion (3.4-fold increase vs. bare-metal stent (BMS) (p < 0.001)), directional migration (accelerated 62 % compared to BMS at 24 h) and proliferation (2.3-fold increase vs. BMS (p < 0.01)), with increased VE-cadherin expression and improved tight junction formation (1.5-fold higher than BMS (p < 0.001)). Additionally, the CD47-VE-M coating reduced macrophage phagocytosis by 59 % (p < 0.01). Compared with BMS, synergistic CD47-VE-M fusion protein-coated stents showed accelerated endothelialization and reduced neointimal hyperplasia and restenosis by 64.4 % (p < 0.001) in vivo. Besides, the coating also decreased the presence of M1 pro-inflammatory macrophages (64.74 % decrease vs. BMS (p < 0.01)), which mitigated the inflammatory response. This novel coating strategy overcomes the limitations of current drug-eluting stent (DES) by simultaneously enhancing endothelial regeneration and suppressing pathological inflammation.

Graphical abstract

Synergistic integration of CD47, VE-Cadherin, and mussel adhesion protein enhances endothelialization and attenuates inflammatory responses in vascular stents.

Highlights

• •CD47-VE-Cadherin-Mfp-5 (CD47-VE-M) fusion protein was synthesized for the first time.
• •CD47-VE-M coatings promote endothelial repair, suppress macrophage phagocytosis in vitro.
• •Dual-functional stent accelerates endothelialization and reduces restenosis in vivo.

1. Introduction

Vascular stents are widely used in clinical practice to treat cardiovascular diseases. However, their long-term success remains limited by complications like restenosis, thrombosis, and endothelial dysfunction [1,2]. Although DES effectively reduce the rate of restenosis, there are associated with delayed re-endothelialization and late stent thrombosis larged due to persistent inflammation and neovascularization [3,4]. Although the antiproliferative drugs of DES, such as sirolimus, inhibit smooth muscle cell (SMC) hyperplasia, they also hinder EC repair, which results in long-term exposure of the stent surface to the blood stream and increases the risk of thrombosis [5,6]. Furthermore, the durable polymer coating may trigger chronic inflammation and further delay vascular healing [7,8]. A critical challenge in stent therapy is delayed or insufficient endothelialization, which is essential for restoring endothelial barrier function and preventing pathological responses such as SMC hyperplasia and platelet aggregation [9,10]. It is crucial to develop bioactive coatings that promote EC adhesion, migration, and proliferation to improve the biocompatibility and performance of these implants [11]. Innovations in biliary stents focus on biodegradable materials and functional coatings to overcome restenosis and replacement challenges [12]. Among the emerging strategies, fusion proteins that integrate endothelial-specific adhesion molecules with bioactive factors represent a promising approach to enhance vascular implant biocompatibility and accelerate endothelial repair.

To address these challenges, particular attention has been directed toward endothelial integrity, which is crucial for vascular homeostasis and the long-term success of stent implantation. Endothelial integrity is crucial for vascular homeostasis and the long-term success of stent implantation [13,14]. Vascular endothelial cadherin (VE-cadherin, VE-Cad) is a key adhesion molecule responsible for maintaining endothelial barrier function and intercellular junction stability that plays a central role in the regulation of endothelial permeability, mechanotransduction, and vascular remodeling [15]. VE-cadherin also influences EC proliferation, migration, and angiogenesis, which are critical processes for vascular repair [16,17]. Despite the established role of VE-cadherin in endothelial function, its potential in enhancing stent biocompatibility has not been fully explored. Loss of VE-Cad expression following stent implantation leads to increased endothelial permeability [18], which facilitates leukocyte infiltration and platelet adhesion, further exacerbating inflammatory responses and thrombotic complications [19]. By incorporating VE-Cad into our synergistic fusion protein, the aim of this study was to promote EC‒cell adhesion, accelerate endothelial monolayer formation, and enhance vascular barrier function. Improved endothelial integrity not only prevents thrombosis but also reduces the risk of excessive SMC proliferation and neointimal formation, both of which are major contributors to in-stent restenosis (ISR). Building on the critical role of VE-cadherin in endothelial repair, we designed a synergistic fusion protein that integrates the extracellular domain of VE-cadherin with CD47, a glycoprotein with well-established anti-inflammatory properties.

While promoting endothelial integrity is essential, equally important is the control of inflammation, which strongly influences the success of stent implantation. Inflammation plays a critical role in the early response to stent implantation, influencing EC adhesion, proliferation, and functional recovery [20]. The insertion of extraneous matter into the vasculature triggers an immediate immune response, leading to the recruitment of monocytes and macrophages, which secrete pro-inflammatory cytokines and ROS [21]. Persistent inflammation can delay endothelial regeneration and contribute to neointimal hyperplasia, ultimately leading to ISR [22]. CD47, a transmembrane protein ubiquitously expressed on vascular cells, acts as a “don't eat me” signal by interacting with signal regulatory protein alpha (SIRPα) on macrophages [23]. This interaction inhibits excessive phagocytic activity, thereby reducing inflammatory responses and protecting ECs from immune-mediated damage. By incorporating CD47 into a synergistic fusion protein coating, this study aimed to suppress inflammatory activation and create a favorable microenvironment for endothelial regeneration.

In addition to enhancing endothelial stability and suppressing inflammation, stent coatings must achieve robust adhesion under physiological conditions. Mussel foot protein 5 (Mfp5), derived from the Mytilus species, possesses exceptional adhesive properties in wet environments due to its catechol functional groups [24]. which form strong interactions with metal and polymer surfaces, thereby ensuring coating stability and durability [25,26]. Studies have shown that mussel-inspired adhesion results in the formation of a robust coating, enabling nitric oxide (NO) catalysis and heparin grafting for enduring antithrombotic protection in blood-contacting devices [27]. Corporating Mfp5 into the synergistic fusion protein enables sustained retention of CD47 and VE-Cad on the stent surface, preserving their bioactivity in vivo. Such adhesion is particularly important in addressing neointimal hyperplasia, the leading cause of ISR, which arises from excessive smooth muscle cell proliferation and extracellular matrix deposition. While traditional DES coated with antiproliferative agents such as sirolimus or paclitaxel effectively reduce neointimal growth but also impair endothelial recovery, which leads to late stent thrombosis [28,29]. Unlike conventional drug-eluting stents that inhibit SMC growth but delay endothelial repair, the combined effects of CD47-mediated immunomodulation, VE-Cad–driven junction stabilization, and Mfp5-assisted adhesion are expected to promote rapid endothelialization while limiting pathological SMC proliferation, supporting long-term vascular patency.

We constructed a prokaryotic expression vector for the CD47-VE-M fusion protein and confirmed its successful expression and purification in E. coli. The purified protein was coated onto stainless steel (SS) substrates to evaluate its ability to promote endothelial cell adhesion. Drawing inspiration from previous fusion protein coatings with antithrombotic and pro-endothelial effects, the CD47-VE-M construct was designed to combine VE-cadherin–mediated endothelial attachment and migration with CD47-driven immunomodulation and Mfp5-assisted adhesion. The performance of the CD47-VE-M–coated stent was then assessed in vitro and in vivo, focusing on endothelialization, macrophage-mediated inflammation, and smooth muscle cell regulation. This strategy aims to accelerate endothelial repair, suppress inflammation, and prevent restenosis, thereby improving long-term vascular outcomes.

2. Methods and materials

2.1. Construction of the vector and expression of CD47-VE-M

The cDNA sequences encoding the VE-cadherin EC1-2 domain and CD47 extracellular domain were amplified from Human umbilical vein endothelial cells (HUVECs) via PCR (primer sequences were see in Table 1). The Mfp5 coding sequence was synthesized on the basis of the previously reported sequence from Mussel edulis. The Mfp5 sequence was fused to the C-terminus of CD47-VE-cadherin via a ClonExpress Ultra One Step Cloning Kit. The resulting CD47-VE-M coding sequence was inserted into the pCold TF vector via the EcoRI and BamHI sites.

Table 1.

PCR primer sequences for the target gene fragments.

Gene	Primer sequences
VE-cad	Forward primer, 5′- CCATGGATGCACATTGATGAAGAGAA -3′ Reverse primer, 5′- TGAACCACCACCACCGTCTTGCAGAGTGAC -3′
CD47	Forward primer, 5′-TGGCTGATATCGGATCCATGTGGCCCCTGGT AGCG -3′ Reverse primer, 5′- TTCGCCGCCGCCGCTGCCGCCGCCGCCATT TTCATTTGGAGAAAACCATGAAACAA -3′
Mfp5	Forward primer, 5′- CTCGAGCTAACTGCTACCACC -3′ Reverse primer, 5′- GGTGGTGGTGGTTCAAGTTCTG -3′

The recombinant plasmid pCold TF-CD47-VE-M was transformed into E. coli (DE3), and protein expression was induced with 0.05 mM IPTG. After 12 h, the cells were harvested via centrifugation and lysed ultrasonically. The soluble fraction was collected and purified via immobilized metal affinity chromatography (Ni‒NTA beads). The eluted CD47-VE-M protein was dialyzed against 1 % acetic acid buffer at 4 °C overnight. Purity was confirmed by SDS‒PAGE and western blotting.

2.2. Coating preparation

The purified CD47-VE-M fusion protein was diluted to 100 μg/mL in PBS (pH 7.4) and then sterilized by filtration through a 0.2 μm membrane. For in vitro studies, the 316 L SS discs (Φ 20 mm, 1 mm) were sequentially polished, ultrasonically cleaned in acetone, ethanol, and ultrapure water, and dried at 65 °C. Sterile protein solution was add to the discs at a surface density of 1 μg/cm² and incubated at 37 °C for 12 h. The discs were then rinsed with ultrapure water and dried under nitrogen. Control coatings, including VE and VE-M fusion protein, were prepared using the same procedure. For in vivo application, sterilized 316 L SS stents (Φ 2.5 mm, 16 mm) were immersed in sterile CD47-VE-M protein solution (0.5 μg/mL) at 37 °C for 12 h, rinsed, and subsequently crimped onto angioplasty balloons (Φ 2.5 mm, 18 mm; Branden®, Shandong, China).

2.3. Stent balloon dilation test

The flexibility and ductility of the synergistic CD47-VE-M fusion protein coating were evaluated via a stent balloon dilation test. The stent transport system was used to expand the bare 316 L SS stent and the 316 L SS stent with the synergistic CD47-VE-M fusion protein coating at 8 atmospheric pressures for 30 s, and the process was repeated three times. The integrity of the coating before and after stent expansion was observed via SEM.

2.4. In vivo stent implantation

All the implantation experiments were the same as reported previously [30]. Male Sprague‒Dawley rats aged 12 weeks were used for all the experiments (n = 5 for per group at 1, 3 and 6 months). Under aseptic conditions, general anesthesia was induced with 3 % pentobarbital sodium (40 mg/kg), and systemic anticoagulation was administered prior to the procedure. The abdominal aorta was then surgically exposed. Using a percutaneous transluminal angioplasty balloon catheter, one of three stent types, namely, BMS (316 L bare-metal stents, Φ 2.5 mm × 16 mm, Branden®, Shandong, China), DES (Sirolimus drug-eluting stents, Φ 2.5 mm × 16 mm, Branden®, Shandong, China), or CD47-VE-M fusion protein-coated stents (prepared by surface modification of the BMS), was deployed into the abdominal aorta. The balloon was inflated at a nominal pressure of 8 atm for 30 s to ensure proper stent deployment. To minimize the risk of acute thrombosis induced by the implantation procedure, dual antiplatelet therapy was initiated 3 days before the procedure and continued for 7 days afterward. Aspirin (10 mg/kg/day) and clopidogrel (7.5 mg/kg/day) were administered during this period. Follow-up evaluations were conducted at 1 month, 3 months, and 6 months post-implantation. At each time point, patency and vessel conditions were assessed by Doppler ultrasound and optical coherence tomography (OCT). The rats were immediately subjected to imaging after being sacrificed. The abdominal aorta containing the implanted stent was carefully explanted, and the vessels were flushed with heparinized saline. Under continuous aseptic conditions, the tissues were fixed in 4 % paraformaldehyde (after a 45 min of fixation), with the heart perfused with 0.9 % saline containing heparin. The samples were then stored at 4 °C for subsequent analysis. All the animal studies were approved by the Committee of Ethics on Experimentation of Chongqing University (Approval number: CQU-IACUC-RE-202305-007). The experiments were conducted following the National Institutes of Health guidelines for the care and use of animals.

2.5. Cell culture

HUVECs, human aortic smooth muscle cells (HASMCs) and mouse monocyte macrophage leukemia cells (RAW 264.7) were cultured in DMEM supplemented with 10 % FBS and 1 % penicillin‒streptomycin. All the cells were maintained in a cell culture incubator at 37 °C with 5 % CO₂. The culture medium was changed every other day. When the cells reached approximately 90 % confluence in the culture flask, they were sub-cultured.

2.6. Atomic force microscopy observation of coating surface morphology

Atomic force microscopy (AFM) can obtain information about the surface morphology and surface roughness of materials with nanometer-level resolution by detecting extremely weak atomic interactions between the surface of the sample under test and a microforce-sensitive sensor within the instrument. In this study, AFM was used to observe and measure the surface morphology and surface roughness of bare 316 L SS discs, as well as 316 L SS discs coated with VE protein, VE-M fusion protein, and CD47-VE-M fusion protein.

2.7. Scanning electron microscopy observation of the coating surface morphology

Scanning electron microscopy (SEM) was used to characterize the surfaces of the bare SS discs and those coated with the VE protein, VE-M fusion protein, or CD47-VE-M fusion protein. The surface morphology of the different protein coatings was observed.

2.8. Static water contact angle measurement

Measuring the water contact angle (WCA) of a material's surface is important for evaluating the hydrophilicity or hydrophobicity of the material. In this experiment, a laboratory-made WCA instrument was used to measure the WCA of deionized water on the surfaces of bare discs and different protein-coated surfaces to evaluate the hydrophilicity or hydrophobicity of the coatings. The WCA was measured at six randomly selected areas on each coating surface via a WCA instrument.

2.9. Hemolysis rate test

Following ASTM F756-17, fresh rabbit blood was collected, anticoagulated with EDTA, and diluted with saline (4:5). The samples were immersed in 10 mL of saline at 37 °C for 30 min. After 0.2 mL of diluted blood was added to each sample, the mixture was incubated at 37 °C for 1 h. The samples were subsequently centrifuged at 1500 rpm for 10 min, after which the absorbance of the supernatant at 540 nm was measured. Distilled water and saline were used as positive and negative controls, respectively. The hemolysis rate was calculated via the following formula:

$$\text{Hemolysis}\text{Rate}(\%)=\frac{\mathrm{A}{}_{\text{sample}-}{\mathrm{A}}_{\text{negative}\text{control}}}{\mathrm{A}{}_{\text{positive}\text{control}-}{\mathrm{A}}_{\text{negative}\text{control}}}\times 100$$

The A _sample, A _{positive control} and A _{negative control} represent the absorbance values of the sample, negative control, and positive control, respectively. A hemolysis rate less than 5 % is considered acceptable for biocompatible materials.

2.10. Platelet adhesion test

Fresh rabbit blood was anticoagulated with sodium citrate and centrifuged to prepare platelet-rich plasma (PRP). The sterilized test samples were placed in a 24-well plate, and 200 μL of PRP was added to each well. After a 1-h incubation at 37 °C, non-adherent platelets were removed with PBS. Adherent platelets were fixed with 2.5 % glutaraldehyde, washed with PBS, dehydrated with ethanol, dried under vacuum, and sputter-coated with gold. The samples were observed via SEM, and platelet adhesion was quantified by counting the platelets in random fields of view.

2.11. Endothelial cell proliferation and infiltration assays

316 L SS substrates and protein coatings were prepared as previously described. HUVECs (1 × 10⁵ cells/mL) were seeded onto coated discs in a 24-well plate and incubated at 37 °C in 5 % CO₂ for 2 h. Non-adherent cells were removed by washing, and fresh medium was added for an additional 24 h of incubation. The cells were then fixed with 4 % paraformaldehyde, blocked, and incubated overnight with a Ki67 primary antibody, followed by incubation with a secondary antibody. The samples were subsequently stained with FITC-phalloidin and DAPI and analyzed via laser confocal microscopy.

For infiltration, after the coatings were dried, a layer of FITC-labeled gelatin was applied. ECs were added to the 24-well plates and incubated at 37 °C with 5 % CO₂ for 24 h. The cells were fixed with 4 % PFA, stained with DAPI, and stored at 4 °C. The samples were observed and photographed via a laser confocal microscope.

2.12. Endothelial cell migration assay

316 L SS and protein coatings were prepared as described. ECs were trypsinized and resuspended at 1 × 10⁵ cells/mL. The suspension (2 mL) was added to a 6-well plate with SS discs and incubated at 37 °C with 5 % CO₂ until it reached confluence. A scratch was made in the monolayer with a 10 μL pipette tip. The detached cells were washed with PBS, and the cells were cultured in RPMI 1640 without FBS. Scratch healing was observed and photographed at 0, 6, 12, and 24 h. The migration rate was calculated via ImageJ with the following formula:

$$\text{MR}=\frac{{\mathrm{S}}_0-{\mathrm{S}}_{\mathrm{t}}}{{\mathrm{S}}_0}$$

where MR = migration rate; S0 = scratch area at 0 h; and St = scratch area at 6, 12, or 24 h.

2.13. ELISA for endothelial cell eNOS content

The steps for EC digestion, seeding, and adhesion are as follows. Following the instructions of the ELISA kit for measuring endothelial nitric oxide synthase (eNOS), the OD values of the supernatants from ECs on different coatings were determined. The eNOS expression levels in ECs subjected to different treatments were calculated via a regression equation from the standard curve.

2.14. Macrophage phagocytosis in vitro

Poly (lactic-co-glycolic acid) (PLGA) (40 mg) was dissolved in 1 mL of N,N-Dimethylformamide (DMF), stirred for 24 h, and mixed with DiD ((MCE, HY-D1028, 10 mg in 1 mL of DMF) to form DiD nano particles (DiDNPs) by dropwise addition to 3 mL of ultrapure water. The mixture was dialyzed for 24 h, and the DiDNPs were stored at 4 °C. RAW 264.7 cells (5 × 10⁵ cells/mL) were seeded onto prepared 316 L SS discs in a 24-well plate and incubated overnight. DiDNPs (100 μL) were added, incubated for various durations, washed, fixed with PFA, stained with DAPI, and analyzed via confocal microscopy. For neutralization, an anti-CD47 antibody was added to the CD47-VE-M-coated discs, which were subsequently incubated for 1 h and processed as described above.

2.15. SEM analysis of neointimal recovery in stented vessels

Stented vessels were harvested after 1, 3, and 6 months, rinsed with PBS, and fixed in 2.5 % glutaraldehyde at 4 °C for 24 h. After fixation, the samples were dehydrated through a graded ethanol series and subjected to critical point drying. The stented segments were dissected to expose the luminal surface, mounted on stubs with conductive carbon adhesive, and sputter-coated with gold-palladium.

SEM analysis was performed via a JSM-7800F instrument (JEOL, Japan) at 3 kV to assess neointimal coverage, with a focus on the EC distribution and the presence of thrombotic material. Images were captured at various magnifications to document the cellular morphology and surface integrity, and recovery across the different time points and stent types was compared.

2.16. Optical coherence tomography analysis

OCT imaging was performed after stent implantation in the rat abdominal aorta using the following parameters: Wavelength: 1300 nm; scanning speed: 50000 A-scans/s; frame rate: 100 frames per second. Following specimen collection, the stented vessels were flushed with saline to remove residual blood and connected to the OCT system (Abbott, USA). The stented segments were scanned in real time to evaluate stent apposition, neointimal formation, and restenosis, and representative images were captured for further analysis.

2.17. Histology and immunofluorescence staining

Tissues from stented rats (stented artery, heart, liver, spleen, lung, and kidney) were collected. The samples were fixed in 4 % PFA, followed by electrolytic removal of metallic struts under constant voltage (7.5 V) in an electrolyte solution containing 5 % citric acid and 5 % sodium chloride. During electrolysis, 316 L SS was dissolved as ions (Fe²⁺/Fe³⁺, Cr³⁺, Ni²⁺ and MoO₄²⁻). The samples were then processed for paraffin embedding and sectioning (five-micron-thick), with any residual struts being manually removed. The sections were deparaffinized, hydrated through a series of graded alcohols, and stained with hematoxylin for 5–10 min. The samples were rinsed, differentiated in acid alcohol, and counterstained with eosin for 2–5 min. The samples were dehydrated, cleared, and mounted. For Verhoeff-Van Gieson (VVG) staining on adjacent sections from the same blocks, slides were treated with Verhoeff's working solution (hematoxylin–ferric chloride–iodine), differentiated in 2 % ferric chloride to accentuate elastic laminae, cleared in sodium thiosulfate, counterstained with Van Gieson (acid fuchsin–picric acid), dehydrated, cleared, and mounted. Nuclei/elastic fibers appear blue-black, collagen red, and cytoplasm/yellowish. Images were acquired on a bright-field microscope and analyzed in ImageJ with identical exposure and threshold settings across groups. For IF staining, rehydrated stented arteries were blocked with goat serum. The sections were incubated with primary antibodies against pro-inflammatory and anti-inflammatory markers of macrophages overnight at 4 °C. The samples were rinsed and incubated with secondary antibodies conjugated to fluorophores, and mounted with anti-fade medium. A Leica SP8 confocal microscope was used to image and analyze the images with ImageJ.

2.18. Antibodies

The primary antibodies used in this study were as follows: anti-CD47 antibody (ab300124; Abcam), anti-VE cadherin antibody (ab33168; Abcam), Ki-67 antibody (sc-23900; Santa Cruz Biotechnology), anti-SM22α antibody (ab14106; Abcam), anti-CD86 antibody (26903-1-AP; Proteintech), anti-CD68 antibody (ab955; Abcam) and anti-CD31 antibody (ab119339; Abcam). The secondary antibodies used in this study were as follows: goat anti-rabbit IgG H&L (Alexa Fluor® 488) (ab150077; Abcam), goat anti-mouse IgG H&L (Alexa Fluor® 488) (ab150113; Abcam), goat anti-rabbit IgG H&L (Alexa Fluor® 555) (ab150078; Abcam) and goat anti-mouse IgG H&L (Alexa Fluor® 594) (ab150116; Abcam).

2.19. Statistical analysis

Statistical analyses were performed via the Statistical Package for Social Sciences (SPSS). The data are expressed as the means ± standard deviations (SDs), with each experiment conducted in triplicate to ensure reliability. Significant differences were assessed via Tukey's multiple comparison test, the Mann–Whitney U test, and Student's t-test, as appropriate. Statistical significance is denoted as not significant (ns), ∗P < 0.05, ∗∗P < 0.01, and ∗∗∗P < 0.0001.

3. Results

3.1. Construction of the synergistic CD47-VE-M fusion protein prokaryotic vector

We constructed a prokaryotic expression vector for the synergistic CD47-VE-M fusion protein via molecular cloning techniques, as shown in Fig. 1A. The pCold TF vector was digested with EcoRⅠ and BamHⅠ, and the CD47-VE-M gene was inserted via homologous recombination, resulting in the vector pCold TF-CD47-VE-M. We also synthesized the control vectors pET32a-VE and pET32a-VE-M for the proteins VE and VE-M.

Fig. 1.

Bioinformatics analysis of CD47-VE-M. (A) Schematic diagram of the prokaryotic expression vector of the synergistic CD47-VE-M fusion protein. (B) Gel electrophoresis images of the recombinant vectors pCold TF-CD47-VE-M, pET32a-VE-M and pET32a-VE. (C) SDS‒PAGE and (D) Western blot of the purified fusion protein. VE, VE-cadherin; VE-M fusion protein; CD47-VE-M fusion protein. (E) Hydrophobicity analysis of CD47-VE-M; (F). Secondary structure analysis of CD47-VE-M. Blue line, alpha helix; red line, extension chain; green line, beta corner; purple line, random curl. (G) Predictive simulation of the tertiary structure of CD47-VE-M.

Following the linearization of the vectors, complete digestion was confirmed (Fig. S1), and the linearized vectors were gel-purified for homologous recombination. Primers with 15 bp homologous regions allowed specific amplification of target fragments (CD47, VE-cad, Mfp5). Homologous recombination successfully connected these fragments to form pCold TF-CD47-VE-M, pET32a-VE, and pET32a-VE-M, as confirmed by sequencing. To facilitate fusion protein construction, we designed five types of linkers (Table S1). SDS‒PAGE analysis revealed that CD47-VE-M proteins with different linkers presented similar molecular weights, with VE at approximately 40 kDa, VE-M at 62 kDa, and CD47-VE-M at 137 kDa. The purified VE, VE-M, and CD47-VE-M were expressed in a soluble form in E. coli, making them suitable for direct use in coating applications.

The VE and VE-M fusion proteins were identified with an N-terminal VE-cadherin antibody, while a CD47 antibody was used for CD47-VE-M. Western blot results (Fig. 1D) confirmed the presence of the VE-cadherin EC1-2 domain in the purified proteins and revealed that CD47-VE-M specifically bound to the CD47 antibody, indicating retention of the extracellular domain of CD47. Owing to a lack of specific antibodies for Mfp5, we could not assess this component by immunoblotting.

These five purified synergistic CD47-VE-M fusion proteins were coated onto 316 L SS to evaluate EC adhesion. Coatings with linker-1 and linker-2 resulted in 2.5-fold and 2.2-fold increases in the number of adhered ECs compared with bare SS (Fig. S2A), confirming their functional activity. Conversely, compared with the control, linkers 3, 4, and 5 did not enhance adhesion (Fig. S2B), indicating that they negatively affected the synergistic fusion protein activity, likely by altering the VE-cadherin structure.

Given that linker-1 produced the strongest adhesion, this construct was selected for further analysis. The basic physicochemical properties of the synergistic CD47-VE-M fusion protein with linker-1 were assessed via an online prediction server, which revealed that it was hydrophilic (Fig. 1E). The secondary structure comprised α-helices (18.06 %), extended strands (30.32 %), β-turns (14.12 %), and random coils (37.50 %) (Fig. 1F). The predicted tertiary structure, modeled via SWISS-MODEL, is shown in Fig. 1G.

3.2. Characterization of the synergistic CD47-VE-M fusion protein coating

The results from the AFM and SEM observations of the uncoated and protein-coated substrates are shown in Fig. 2. SEM images (Fig. 2A) revealed the morphological characteristics of 316 L SS, VE, VE-M, and CD47-VE-M coatings at 1000 × and 5000 × magnifications. The hydrophilicity, shown in Fig. 2B, varies with protein coating, where the VE coating has a WCA of 72.5 ± 4.5°, whereas the hydrophilicity of the 316 L SS is 38.4 ± 4.1°, indicating strong hydrophilicity. The VE-M and CD47-VE-M coatings had contact angles of 56.3 ± 2.7° and 43.9 ± 3.3°, respectively, suggesting that CD47-VE-M may offer better biocompatibility for vascular tissue engineering. The AFM results (Fig. 2C) indicate surface roughness values of 9.35 ± 2.1 nm, 13.05 ± 1.6 nm, 16.42 ± 1.8 nm, and 23.12 ± 0.9 nm for the respective coatings. Although the roughness increased with increasing protein coating, all the roughness values remained below the protein adsorption threshold (<50 nm), meeting the standards for hemocompatibility and preventing platelet aggregation. Immunofluorescence confirmed successful CD47-VE-M coating on stents, with VE-cadherin and CD47 absent on BMS (Fig. S3A). Under pulsatile flow, the coating remained stable, retaining both markers after 12, 24, and 48 h (Fig. S3B). In mechanical tests using an in vitro stent model, SEM images (Fig. S4) revealed that the CD47-VE-M coating remained uniform and intact after expansion at 8 atm for 30 s, with only minor localized detachment.

Fig. 2.

(A) The specific morphologies of different coating surfaces were observed via SEM. Scale bars = 10 μm and 1 μm. (B) The specific morphologies of different coating surfaces were observed via SEM. (C) The surface morphology and roughness of different coatings were observed via AFM. Ra, roughness. (D) Hemolysis rates and (E) statistics of SS and various protein coatings. (F) Morphology and (G) quantity of platelets adhered to 316 L SS and different protein coatings. Scale bars = 10 μm and 5 μm. 316 L SS, 316 L medical SS; VE, VE-cadherin; VE-M, fusion protein; CD47-VE-M, fusion protein.

Hemolysis rates were measured (Fig. 2D and E), with values of 1.147 ± 0.09 % for 316 L SS, 1.953 ± 0.08 % for VE, 1.280 ± 0.10 % for VE-M, and 1.227 ± 0.03 % for CD47-VE-M, all below the safety standard of 5 %. Platelet adhesion (Fig. 2F and G) was significantly lower on the VE-M and CD47-VE-M coatings than on the 316 L SS coating, with CD47-VE-M demonstrating the strongest inhibition of platelet activation and adhesion, indicating excellent blood compatibility.

3.3. Evaluation of the synergistic CD47-VE-M fusion protein coating in vitro

Normal vascular ECs inhibit smooth muscle contraction, platelet aggregation, SMC proliferation, leukocyte adhesion, and thrombosis, serving as a barrier to vascular integrity. To assess whether the synergistic fusion protein containing the extracellular domain of VE-cadherin promotes the adhesion and proliferation of ECs, we seeded HUVECs on various coatings (Fig. 3A). There were greater numbers of ECs on the VE-M and CD47-VE-M coatings (Fig. 3B), but no significant difference in the number of Ki67⁺ cells was observed in the collagen- and VE-coated groups (Fig. 3C). Immunofluorescence staining of F-actin revealed that ECs on 316 L SS had cytoskeletons concentrated in the cytoplasm, whereas those on the VE-M and CD47-VE-M coatings had greater cytoskeletal distributions at the cell membrane, which enhanced intercellular connections and prevented the leakage of cytokines (Fig. 3D).

Fig. 3.

CD47-VE-M promotes the proliferation and migration of vascular ECs. (A) Fluorescence images of the adherent number and proliferation ratio of ECs on 316 L SS and different protein coatings. Scale bar = 50 and 25 μm. (B–D) Statistical chart of the number of adherent and proliferating ECs on 316 L SS and different protein coatings. (E, F) Migration velocity and statistical data of ECs cultured on 316 L SS and different protein coatings. 316 L SS, 316 L SS; VE, VE-cadherin; VE-M fusion protein; CD47-VE-M fusion protein. Statistical significance was determined using one-way ANOVA followed by Tukey's post hoc test. Symbols indicate: ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; ns, not significant.

We evaluated EC infiltration by analyzing the hydrolysis of FITC-labeled gelatin. Compared with those on bare 316 L SS, ECs on the collagen, VE, VE-M, and CD47-VE-M coatings presented significantly increased hydrolysis areas (Fig. S5), indicating enhanced infiltration and migration capacities. After VE-cadherin binding sites were blocked with antibodies, the adhesion of ECs to the VE, VE-M, and CD47-VE-M coatings decreased significantly, approaching the levels observed on 316 L SS (Fig. S6). These finding confirm that the specific adhesion to CD47-VE-M is due to active structural domain of VE-cadherin.

In a scratch wound healing assay, ECs on VE-M demonstrated the fastest migration (Fig. 3E and F). Compared with the 316 L SS and collagen coatings, the CD47-VE-M coating also enhanced migration. These results indicate that VE, VE-M, and CD47-VE-M coatings improve EC migration, and promote healing at injury sites.

During the process of vascular remodeling and repair, the competitive coverage of vascular SMCs and ECs at the interface of implanted materials is crucial for determining the rate of healing. We performed co-culture experiments with ECs (green) and SMCs (red) at a 1:1 ratio. The results, shown in Fig. S7A, demonstrate that both VE-M and synergistic CD47-VE-M fusion protein coatings exhibit remarkable adhesion properties for ECs, with significantly greater numbers of adhered ECs than 316 L SS, collagen, and VE coatings. As illustrated by the statistical analysis of the adhesion ratios of ECs to SMCs on the coatings (Fig. S7C), the ratios of ECs adhered to SMCs on the VE-M and CD47-VE-M coatings were more than two-fold greater than those on the controls. To further validate cell identity and confirm the adhesion proportion of SMCs on the surface, we performed immunofluorescence staining for the SMC marker SM22α (Fig. S7B) and quantified the percentage of SM22α⁺ cells among total nuclei (Fig. S7D). This enhanced specificity for EC adhesion, with minimal effects on SMCs, may be attributed to the presence of VE-cadherin in the coatings, which interacts with endogenous VE-cadherin in ECs, thereby promoting their adhesion.

3.4. Synergistic CD47-VE-M fusion protein enhances cell tight junction and eNOS levels

VE-cadherin is a specific adhesion molecule found at EC junctions that is crucial for regulating vascular permeability and leukocyte extravasation. In addition to its adhesive role, VE-cadherin influences cell proliferation, apoptosis, and the functionality of vascular endothelial growth factor receptors, which makes it essential for endothelial repair and angiogenesis.

To examine the effect of a fusion protein on endogenous VE-cadherin in HUVECs, we conducted immunofluorescence staining after the ECs were seeded on different protein-coated surfaces. The results (Fig. 4A) revealed that the ECs on the VE-M and CD47-VE-M coatings presented uniform junctions with a high density of continuous VE-cadherin connections. Statistical analysis (Fig. 4B) revealed significantly greater fluorescent areas of VE-cadherin at EC junctions on the VE-M and CD47-VE-M coatings than on the 316 L SS and collagen coatings, suggesting tighter intercellular connections.

Fig. 4.

Synergistic CD47-VE-M fusion protein enhances tight connections between HUVECs. (A) Immunofluorescence staining of VE-cadherin between HUVECs on 316 L SS and different protein coatings. Scale bars = 50 μm and 20 μm (enlarged view). (B) Statistics of the fluorescence intensity and (C) transcription level of VE-cadherin in HUVECs cultured on 316 L SS and different protein coatings. (D–F) Transcription levels of ZO-1, Occludin and Claudin-5 in HUVECs cultured on 316 L SS and different protein coatings. (G) The expression of eNOS in the supernatant of ECs adhered to 316 L SS and different protein coatings. 316 L SS, 316 L medical SS; VE, VE-cadherin; VE-M fusion protein; CD47-VE-M fusion protein.

Since adhesion requires VE-cadherin, its increased expression in ECs is essential. We measured endogenous VE-cadherin transcription levels in adhered ECs across different surfaces. The results of quantitative PCR (qPCR) (Fig. 4C) confirmed that VE-cadherin levels were significantly greater in ECs on VE-M and CD47-VE-M coatings than in those on 316 L SS, with no significant differences among the other groups. These findings suggest that these coatings enhance endogenous VE-cadherin expression, improving EC barrier function and connectivity. In addition to VE-cadherin, tight junctions (TJs), which are mediated by proteins such as ZO-1, Occludin, and Claudin-5, also play roles in intercellular connections. We assessed the transcription levels of these TJs in adhered ECs via qPCR. The results (Fig. 4D–F) indicated that ZO-1 levels were significantly increased on the VE-M and CD47-VE-M coatings, whereas Claudin-5 and Occludin levels were not significantly different among the groups.

Endothelial dysfunction is critical for delaying stent re-endothelialization and late restenosis.NO produced by ECs helps dilate blood vessels and inhibits platelet aggregation and SMC proliferation. Endothelium-derived NO is a multifunctional signaling molecule that serves as an effective vasodilator and reduces ROS production.

To evaluate the effect of the synergistic CD47-VE-M fusion protein on endothelial function, we measured eNOS levels in ECs on various coatings using an ELISA kit. The results (Fig. 4G) revealed that compared with 316 L SS, VE, VE-M and CD47-VE-M coatings significantly enhanced eNOS expression. In parallel, NO levels in the cultured medium were assessed by Griess assay, and the trends in NO production were consistent with intracellular eNOS expression (Fig. S8). These results indicate that the synergistic CD47-VE-M fusion protein coating promotes eNOS-mediated NO production, highlighting its potential as an endothelial-friendly surface.

3.5. Evaluation of the synergistic CD47-VE-M fusion protein coating in vivo

We examined the extent of endothelial regrowth via SEM after the stents were implanted for 1 month, 3 months or 6 months. As shown in Fig. 5 (arranged by blood flow direction), regenerated ECs completely covered the BMS, DES and CD47-VE-M coated stents at 3 months and 6 months. In the one-month group, only the CD47-VE-M coated stents group presented complete restoration of the neointima covering the entire vascular segment of the stent filament surface. There was even a small amount of activated platelet adhesion and blood cells in the DES group (no platelets or fibrin were observed on the CD47-VE-M coated stents). These results suggest that CD47-VE-M coated stents are vital for improving the barrier function of the regenerated endothelium.

Fig. 5.

SEM images of neointimal ECs on the BMS, DES and CD47-VE-M fusion protein coated stents. SEM images showing in-stent structures after 1 month of implantation (A), 3 months of implantation (B) and 6 months of implantation (C), Scale bars = 1 mm, 500 μm, 100 μm and 50 μm. Key: white arrows indicate blood cells.

OCT analysis on multiple cross-sectional areas, which revealed that in-stent restenosis predominantly occurred in the distal part of the stented vessels (Fig. S9). Furthermore, longitudinal sections were obtained and subjected to VVG and SMC staining, which corroborated the OCT findings, showing abundant ECM and SMC accumulation in the distal part of BMS-treated vessels at 6 months (Fig. S10). OCT morphology and histomorphometric analyses were employed to evaluate intimal hyperplasia and in-stent stenosis, with OCT imaging providing a comprehensive in vivo assessment of vascular parameters. Notably, at 6 months post-implantation, the BMS group exhibited significant neointimal hyperplasia and considerable lumen area loss (Fig. 6A). Statistically, the CD47-VE-M coating effectively inhibited excessive neointimal proliferation and the occurrence of ISR (Fig. 6B–D). Compared with the OCT morphology, the VVG-stained images of the stented artery are depicted in Fig. 6E. Endothelial coverage was observed at 1 month, further confirming the rapid endothelialization of the stent. The neointimal area subsequently gradually increased over time (Fig. 6F–H). Importantly, the restenosis rate obtained from VVG staining trended to be consistent with the previous H&E results (Fig. S11). Immunofluorescence staining on vascular sections at 1, 3, and 6 months post-implantation (Fig. 6I). Except for the DES group at 1 month, all groups showed continuous endothelial coverage (CD31⁺) on the neointimal surface, while the neointimal cells were predominantly proliferative SMCs (SM22α⁺). Over time, as the ECs fully covered the stent struts, the arrangement and intercellular connections of the ECs in the CD47-VE-M-coated group increasingly resembled those of normal vascular morphology.

Fig. 6.

Morphometric analysis of the BMS, DES and CD47-VE-M fusion protein coated stent groups at 1, 3 and 6 months after stent placement. (A) Representative OCT images showing arterial stents at 1, 3 and 6 months after implantation of the BMS, DES and CD47-VE-M fusion protein coated stents. Scale bar = 500 μm. Green ∗ indicates uncovered neointima stent filaments, and red # indicates thrombus formation. At 1, 3 and 6 months after stent implantation, measurements of intrastent restenosis, including (B) the size of the neointima, (C) the size of the vascular lumen, and (D) the percentage of intrastent restenosis, were analyzed. ∗∗p < 0.01, ∗∗∗p < 0.001 compared towith the BMS-stented groups. (E) Representative VVG stained sections from the BMS, DES and CD47-VE-M fusion protein coated stent are shown at 1, 3 and 6 months post-stenting. Scale bars = 500 μm and 50 μm. (F) Neointimal area, (G) luminal area, and (H) percentage of stenosis were counted. N = 5 for each group. (I) Immunofluorescence staining of the distal part of stent-implanted vascular sections showing CD31 (green), SM22α (red), and nuclei (DAPI, blue). Scale bars = 500 μm and 100 μm ∗∗P < 0.01, ∗∗∗P < 0.001 versus the BMS-stented group (one-way, repeated-measures ANOVA).

3.6. Regulation of inflammatory cells by the CD47-VE-M fusion protein coating

Considering the early acute reaction after stent implantation, we performed histomorphological examination and routine blood examination of the heart, liver, spleen, lung and kidney of the rats with stent implantation. In terms of histomorphology, the CD47-VE-M coated stents implanted in the abdominal aorta of the rats did not cause changes in the heart, liver, spleen, lung or kidney (Fig. S12). Compared with the BMS group, the CD47-VE-M coated stent group presented a decrease in the number of lymphocytes and platelets but no significant changes in other blood-related components (Fig. S13 & Table S2).

To investigate the anti-phagocytic effects of the synergistic CD47-VE-M fusion protein, we evaluated its activity in RAW264.7 cells. The phagocytosis of macrophages on the VE and VE-M coatings was similar to 316 L SS, whereas those on the CD47-VE-M coatings exhibited a significant reduction in DiDNP internalization (Fig. 7A & C). Disruption of the CD47 with antibodies or recombinant proteins activated macrophage phagocytosis. Blocking the CD47 domain on CD47-VE-M with CD47 antibodies restored phagocytosis to levels similar to those on the 316 L, VE, and VE-M coatings (Fig. 7B & D). These results align with the findings of Weissman on the enhancement of macrophage phagocytosis via anti-CD47 antibodies, confirming that the CD47-VE-M fusion protein effectively inhibits macrophage phagocytosis [23], which reduces inflammatory cell adhesion and inflammation and improves biocompatibility.

Fig. 7.

The synergistic CD47-VE-M fusion protein reduced macrophage recruitment and phagocytosis in the stent implantation area. (A, C) Effects and statistics of 316 L SS and various protein coatings on macrophage phagocytosis in vitro. (B, D) Macrophage phagocytosis in vitro after blocking CD47 on 316 L SS and protein-coated surfaces. Scale bar = 10 μm. (E) Macrophage detection via CD68 and CD86 immunofluorescence at 1 and 3 months post-stent implantation. Scale bar = 100 μm. (F, G) Statistical analysis of CD68-positive cells and CD68⁺CD86 double-positive cells in stented vascular sections. BMS, bar metal stent; DES, drug-eluting stent; CD47-VE-M, synergistic CD47-VE-M fusion protein-coated stent.

The presence of inflammation is directly linked to neointimal healing and the development of stent-in restenosis. The inflammatory microenvironment of the vascular segment implanted with synergistic CD47-VE-M fusion protein-coated stents was assessed, as illustrated in Fig. 7E. Synergistic CD47-VE-M fusion protein coating effectively reduced the recruitment and infiltration of macrophages (CD68⁺ cells) and M1 pro-inflammatory macrophages (CD68⁺ & CD86⁺ cells) within the neointima. Compared with the BMS group, the synergistic CD47-VE-M fusion protein coating resulted in 59 % and 65 % reductions in macrophages and M1 macrophages, respectively, at one month post-implantation (Fig. 7F & G). At 3 months after stent implantation, the numbers of macrophages and M1 pro-inflammatory macrophages remained low among the groups. These results indicate that the synergistic CD47-VE-M fusion protein coating inhibits the polarization of recruited macrophages in the vascular segment following stent implantation.

4. Discussion

The development of biocompatible coatings for vascular stents is essential to improve the long-term success of these implants, as they must promote EC adhesion and migration while preventing complications such as restenosis and thrombosis [31,32]. This study demonstrated that a synergistic CD47-VE-M fusion protein coating, which combines the extracellular domain of VE-cadherin with CD47, significantly enhances endothelialization and reduces inflammation in both in vitro and in vivo models. Specifically, in vitro experiments revealed that compared with BMS, the CD47-VE-M coating enhanced EC adhesion by 3.4-fold, accelerated cell migration by 62 %, and increased cell proliferation by 2.3-fold. These functional enhancements were accompanied by a 1.5-fold increase in VE-cadherin expression and elevated eNOS levels, which together suggest improved endothelial barrier function and cell–cell junction integrity. By promoting EC adhesion, migration, and proliferation while also mitigating the inflammatory response, this coating offers a promising solution for improving vascular stent biocompatibility and performance.

Endothelial barrier integrity is a critical factor in vascular homeostasis and stent performance. VE-Cad, a major component of endothelial adherens junctions, plays a pivotal role in maintaining endothelial cohesion and preventing permeability-related complications [33,34]. Compared with the other coatings, the synergistic CD47-VE-M fusion protein resulted in superior EC adhesion in vitro, including VE and VE-M alone. The ability of this protein to enhance cell adhesion and migration is attributed to the presence of the VE-cadherin extracellular domain, which mediates EC attachment and promotes tighter intercellular connections [35]. Our study revealed that, compared with those on control stents, ECs on fusion protein-coated stents exhibited greater junctional stability and fewer intercellular gaps. Vinculin phosphorylation impairs VE-cadherin/catenin complex disruption to increase endothelial permeability and atherogenesis [36]. Moreover, functional assays revealed increased expression of tight junction proteins and reduced endothelial permeability, indicating improved vascular barrier function. This enhanced junctional stability is essential for preventing thrombus formation and maintaining vascular integrity post-stent implantation [21,37]. This improved structural integrity is particularly important, as it is associated with decreased thrombus formation and a reduced risk of restenosis, as evidenced by the in vivo observation that the fusion protein coating led to a 64.4 % reduction in neointimal hyperplasia and restenosis as confirmed by OCT and histological analysis.

Another crucial aspect of this study is the anti-inflammatory properties of the synergistic CD47-VE-M fusion protein. Inflammation is a significant contributor to the development of restenosis and thrombosis following stent implantation [38,39]. Recent studies have also demonstrated that peptide-coated stents can improve both the mechanical and physiological properties necessary for re-endothelialization, supporting the concept of peptide-functionalized coatings as viable alternatives [40]. Our in vitro data indicated that the coating reduced macrophage phagocytosis by 59 %, which aligns with the findings of previous studies demonstrating the anti-inflammatory effects of CD47 [41]. In vivo, CD47-VE-M-coated stents exhibited a 64.7 % reduction in M1 pro-inflammatory macrophage infiltration at 1 month post-stent implantation, which was associated with reduced neointimal hyperplasia and restenosis [42,43]. Notably, bioresorbable stent degradation accelerates endothelial senescence and reduces shear stress, further promoting ISR in aged vasculature [44]. Recent advances have also highlighted the importance of spatiotemporal regulation of therapeutic signals to prevent thrombosis and restenosis [45], suggesting that temporal control of anti-inflammatory and pro-healing cues may further enhance vascular repair. Similarly, CO-loaded hemoglobin-EGCG nanoparticle coatings effectively modulated inflammation in vascular implants [46], highlighting the potential of integrating anti-inflammatory and pro-regenerative functions within a single coating platform. These findings are significant because they indicate that the CD47-VE-M coating not only accelerates endothelialization but also limits the inflammatory response, which is a key factor in preventing restenosis and thrombosis. The reduction in inflammatory cells, particularly M1 macrophages, further underscores the anti-inflammatory potential of this coating, which could be instrumental in improving the biocompatibility and longevity of vascular stents.

Alternative protein coatings, such as CD31 mimetic peptide coatings, have been explored. CD31 facilitates endothelial healing by inhibiting platelet and leukocyte activation [47]. However, its mechanism is relatively singular and does not offer the immunomodulatory benefits inherent to CD47. In contrast, the synergistic CD47-VE-M fusion coating integrates anti-inflammatory properties, promotes endothelialization, and enhances mechanical stability. Similarly, the CCN5 recombinant protein coating reduces restenosis by inhibiting SMC proliferation, although it may inadvertently delay endothelialization [38]. Notably, the synergistic CD47-VE-M fusion coating accelerated endothelial coverage through VE-Cadherin mediation while concurrently mitigating inflammation via CD47, potentially achieving more balanced vascular repair. Furthermore, although the SIRPα-Fc fusion protein can block the CD47-SIRPα pathway, it lacks direct endothelial barrier protection [48,49]. In addition, multifunctional surface modification strategies have been developed to provide simultaneous antibacterial and anticoagulant properties, or to mimic mussel-inspired adhesive peptides to enhance endothelialization [50,51], as well as versatile peptide-mimic strategies based on mussel-inspired and bioclickable chemistry [52], further supporting the concept that multi-targeted biofunctional coatings may outperform single-mechanism approaches. Similarly, endothelium-mimicking multifunctional coatings developed via a stepwise metal–catechol–(amine) surface engineering strategy [53] further exemplify how rational surface engineering can provide simultaneous endothelialization promotion, anticoagulation, and anti-inflammatory benefits in vascular stents. While bio-inspired lotus-fiber and mussel-based hydrogels show super-stretchability, self-healing, strong adhesion, and antibacterial properties [54], suggesting that integrating such cues could enhance vascular coating performance. A tailored recombinant humanized type III collagen coating enhances anticoagulation, endothelial healing, and smooth muscle phenotype conversion, reducing restenosis [5]. In addition to protein coating, other coating processing or application methods can replace some of the functions of drug coating. Ultrathin Zn and additively manufactured Mo stents show superior strength, biocompatibility, and controlled degradation for next-generation bioresorbable applications [55,56]. In addition, Cu²⁺ chitosan coatings enhance endothelialization, curb degradation, and suppress intimal hyperplasia in magnesium stents [31,57]. Moreover, nitric oxide–releasing coatings and spatiotemporal delivery systems that combine therapeutic gases with growth factors have demonstrated significant potential in simultaneously preventing thrombosis and promoting endothelial repair [58,59], Hierarchical capillary coatings to biofunctionalize drug-eluting stents for improving endothelium regeneration [60] also highlight the promise of combining structural design with biological cues to accelerate re-endothelialization, further underscoring the translational promise of drug-free functional coatings. Collectively, the protein functional coating or biomimetic functional coating drug-free system represents a promising strategy for next-generation of stents.

The long-term success of vascular stents depends on their ability to balance endothelialization with the controlled inhibition of restenosis [5]. In addition to the fusion protein coating, a peptide vaccine targeting COMP and TSP-1 degradation inhibited vascular SMC migration while promoting re-endothelialization, thereby ameliorating atherosclerosis and post-injury neointimal hyperplasia [61]. Compared with conventional stents, our coated stents presented significantly lower restenosis rates and better vascular remodeling outcomes. The integration of biofunctional molecules in the coating not only enhanced endothelial function but also mitigated adverse immune responses [62], highlighting the potential of fusion protein-based strategies for vascular interventions. Recent analyses have highlighted a shift in drug-eluting stent research from drug loading to surface functionalization [63], supporting the concept of bioactive protein coatings. Similarly, zinc-based biomaterials show promise for biodegradable stents, while insights from biliary stents emphasize how active materials and surface modifications can enhance biocompatibility—principles directly applicable to vascular stents [55,64]. Future studies will focus on optimizing the long-term stability of the coating and evaluating its performance in diabetic and high-risk atherosclerotic models [12]. Overall, the results suggest that the synergistic CD47-VE-M fusion protein coating represents a promising approach for next-generation vascular stents, offering a synergistic strategy to enhance endothelialization while reducing inflammation and restenosis.

5. Conclusion

Synergistic CD47-VE-M fusion protein coating represents a novel strategy for improving the performance of vascular stents. By combining the endothelial-specific adhesion properties of VE-cadherin with the anti-inflammatory effects of CD47, this synergistic fusion protein enhances EC adhesion, migration, and function while simultaneously reducing inflammation. These properties are crucial for the successful integration of stents into the vascular system, as they promote rapid endothelialization and reduce the risk of restenosis and thrombosis. Our findings suggest that the CD47-VE-M coating could significantly improve the biocompatibility and long-term success of vascular implants. Further clinical studies and trials are warranted to fully evaluate the efficacy and safety of this coating in human patients, but the results from this study provide a strong foundation for future development in the field of vascular tissue engineering.

CRediT authorship contribution statement

Wenhua Yan: Writing – review & editing, Writing – original draft, Investigation, Funding acquisition, Formal analysis. Shuyu Li: Methodology, Investigation, Formal analysis. Tian Zhang: Investigation. Junli Huang: Writing – review & editing, Investigation. Chengchen Deng: Investigation.Kunshan Yuan: Investigation. Nan Huang: Writing – review & editing, Supervision. Haijun Zhang: Writing – review & editing, Validation. Guixue Wang: Writing – review & editing, Writing – original draft, Supervision, Resources, Project administration, Methodology, Investigation, Funding acquisition, Conceptualization.

Ethics approval and consent to participate

All the animal studies were approved by the Committee of Ethics on Experimentation of Chongqing University (Approval number: CQU-IACUC-RE-202305-007). The experiments were conducted following the National Institutes of Health guidelines for the care and use of animals.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Data availability

The data will be made available upon request.