Microenvironment-responsive coating for vascular stents to regulate coagulation-inflammation interaction and promote vascular recovery

Abstract

Early coagulation-inflammation interaction and late in-stent restenosis undermine the efficacy of vascular stents after implantation. Targeting the interplay between inflammation and coagulation, and smooth muscle cell (SMC) proliferation, we presented a microenvironment-responsive coating designed to regulate tissue responses and vascular regeneration throughout the remodeling process. Coagulation was inhibited by incorporating anticoagulant tirofiban into the coating. MMP9-responsive nanoparticles embedded in the coating released salvianolic acid A to modulate inflammatory cell behavior and inhibit SMC dysfunction. By effectively interfering with clotting and inflammation, the coating suppressed platelet-fibrin interaction and formation of platelet-monocyte aggregates, thereby mitigating adverse effects on reendothelialization. Its ability to influence SMC proliferation and migration resulted in reduced intimal hyperplasia. Coated stents were shown to significantly regulate tissue regeneration, improve the vascular environment and even reduced the lipid content in the narrowed atherosclerotic vessels in vivo. This direct approach enhanced the vascular tissue regeneration after stent implantation, and offered promising insights for optimizing vascular stent design.

Graphical abstract

Highlights

• •An innovative microenvironment-responsive coating for stents to regulate tissue responses and enhance vascular regeneration.
• •The coating releases salvianolic acid A in response to MMP9, modulating monocyte differentiation and SMC proliferation.
• •The coating sustainedly releases tirofiban to interfere the coagulation-inflammation interaction with salvianolic acid A.
• •The coating reduces intimal hyperplasia, and improves vascular healing in an atherosclerotic rabbit model.

1. Introduction

Coronary artery disease is associated with inflammation, endothelial dysfunction and lipid deposition, severely impacting human longevity. Cardiovascular stents serve a crucial role in its clinical management [1]. In recent years, various types of vascular stents have been manufactured and utilized. Implanted stents provide mechanical support for narrowed vessels and prevent acute occlusion of the vessel. However, suboptimal reendothelialization, troublesome thrombosis, persistent inflammation, and late in-stent restenosis still impede the optimal functionality of vascular stents over time [2,3].

As a typical blood-contacting material, vascular stents undergo a sequence of events including vascular endothelial injury, inflammation, coagulation and intimal hyperplasia [[4], [5], [6]]. All uncontrolled tissue reactions compromise the reliability and effectiveness of stents, ultimately leading to delayed reendothelialization and late in-stent restenosis [7,8]. Both of these undesirable consequences are closely linked to the proliferation of endothelial cells and smooth muscle cells (SMCs), underscoring the critical importance of regulating reendothelialization and SMC proliferation to enhance vascular stent performance [1,7,[9], [10], [11], [12]].

During the initial contact phase, inflammation and coagulation are the most active tissue responses, contributing to delayed reendothelialization and intimal hyperplasia [13,14]. The inflammatory response occurs on damaged vascular tissue, and monocytes are recruited around the stent. They undergo phenotypic transformation and release various cytokines [15,16]. Overexpression of reactive oxygen species and matrix metalloproteinase is a typical phenomenon [17,18]. Upon contact with vascular stents, intrinsic and extrinsic coagulation reactions are immediately activated. During thrombosis, platelets are critical players as activated platelets accelerate the coagulation cascade and form clots along with fibrin and leukocytes [19,20]. Inhibiting platelet adhesion and activation is a direct way to improve the hemocompatibility of blood-contacting materials.

Additionally, coagulation and inflammation are closely related, involving cells and cytokines. Thrombin, tissue factor (TF), and platelets in the coagulation response play important roles in the inflammatory response [16,21]. Similarly, various inflammatory pathways induce platelet activation. Platelets are the link between coagulation and inflammation. Platelet aggregation and the binding of platelets to fibrin are essential steps in the process of thrombosis [22,23]. In addition to directly affecting monocyte activation and recruitment, platelets also form aggregates with monocytes through receptors such as CD62 antigen-like family member P (CD62P, also called P-selectin). Platelet-monocyte aggregates further increase the secretion of various inflammatory cytokines and chemokines [20,24]. Coagulation and inflammation also exacerbate late in-stent restenosis [17]. Cytokines secreted by macrophages and platelets, such as tumor necrosis factor-α (TNF-α), lead to enhanced migration and proliferation of SMCs [25]. They also affect cellular behavior, such as endothelial cell adhesion and permeability [17,26]. Therefore, methods to regulate coagulation and inflammation in delicate ways are worth exploring.

In the later implantation phase, intimal hyperplasia is a major challenge for vascular stents. Due to the absence of endothelium and stimulation of cytokines, SMCs migrate to the intima and transform into a synthetic phenotype. Activated SMCs constantly proliferate and secrete extracellular matrix (ECM). Abnormal SMC proliferation and ECM remodeling ultimately culminate in intimal hyperplasia and subsequent lumen loss [2,27]. Therefore, encouraging healthy SMC proliferation is important for the regenerative repair of vascular tissue.

Based on this, a microenvironment-responsive coating for the whole-process regulation of tissue response was proposed. Since coagulation and inflammation are closely related with reendothelialization and neointimal regeneration, dual regulation of coagulation and inflammation is essential to effectively mitigate adverse tissue responses. Platelets are one of the important crossroads of coagulation and inflammation, and inhibition of their activation is indispensable for regulating these processes. Here, we chose the platelet integrin receptor (Gp IIb/IIIa) antagonist tirofiban as an anticoagulant substance. In addition, matrix metalloproteinase 9 (MMP9) overexpression in inflamed tissues induces the release of inflammatory factors and accelerates extracellular matrix degradation. As the primary cells involved in inflammation, the phenotypic changes of monocyte-macrophages directly influence tissue responses around vascular stents [28]. To precisely modulate the inflammatory response and vascular regeneration, MMP9-responsive nanoparticles were developed. Salvianolic acid A (Sal A), rich in phenolic hydroxyl groups, was embedded in nanoparticles and released from nanoparticles upon MMP9 activation. Sal A not only inhibits MMP9 function, but also regulates the behavior of inflammatory cells [29]. Furthermore, Sal A modulates SMC differentiation and migration, thereby preventing intimal dysplasia and ECM remodeling [12,30]. At the same time, Sal A does not restrain the cell viability and proliferation of endothelial cells (ECs) [31]. Those functional substances were incorporated into the coating through layer-by-layer self-assembly modification. This coating strategy was designed to regulate tissue responses throughout the remodeling process, facilitating reendothelialization and healthy intimal regeneration. The effectiveness of the coating was verified in atherosclerotic rabbit model, and it was expected to interfere with coagulation-inflammation and SMC behavior. The coating showed potential for ultimately ensuring the long-term functionality of vascular stents in atherosclerotic blood vessels (see Fig. 1).

Fig. 1.

Illustrated schematic of coating preparation and working mode. PDA polydopamine, PEI polyethyleneimine, GS GS nanoparticles, Sal A Salvianolic acid A, TIR tirofiban, OxHA oxidized hyaluronic acid.

2. Results and discussion

2.1. Design of a responsive coating

To prepare nanoparticles with gelatin and Sal A, 4-Carboxyphenylboronic Acid (CPBA)-GEL was synthesized by crosslinking CPBA and gelatin using the 1-(3-Dimethylaminopropyl)-3-ethylcarbodiimide (EDC)/N-Hydroxysuccinimide (NHS) method. The phenylboronic acid group of CPBA-GEL was confirmed via ¹H Nuclear Magnetic Resonance Spectroscopy (NMR) spectra (Fig. S1). Subsequently, the Sal A nanoparticle (named GS) was crosslinked between the phenylboronic acid group of CPBA-GEL and the catechol group of Sal A through boron esterification reaction (Fig. 2A). The particle size of GS is about 100 nm (Fig. 2B). Given that gelatin can be cleaved by the MMP9 peptide and serves as the foundation for the inflammation-responsive ability of nanoparticles, the release of Sal A from nanoparticles in response to MMP9 was investigated. As shown in Fig. 2C, the rate and total amount of Sal A released increased proportionally with the concentration of MMP9. Within 20 min, the nanoparticles rapidly reacted to MMP9, triggering the release of Sal A. After 180 min, approximately 80 % of Sal A was released when exposed to 10 nM MMP9 while only 20 % was released when no MMP9 was present, which is an acceptable background release level [32,33]. The backgroud release of Sal A was attributed to the slightly hydrolysis of phenylboronic ester bonds and the desorption of Sal A molecules attached to the surface of the GS during preparation [34,35]. These results indicated that GS nanoparticles could release Sal A in response to the stimulation of MMP9, which underlined the release of anti-inflammatory substances from the coating in response to inflammation.

Fig. 2.

Characterization of GS nanoparticles and coatings. (A) Schematic diagram of nanoparticle preparation. (B) Particle size of GS nanoparticles. (C) Cumulative Sal A released from MMP9-responsive GS nanoparticles in PBS containing 0, 1, 5 and 10 nM MMP9 at 37 °C. (mean ± SD, n = 3 independent samples). (D) UV–vis spectra of Sal A, PEI/OxHA, PEI/GS/OxHA, PEI/TIR/OxHA and PEI/TIR/GS/OxHA. (E) Cumulative tirofiban release from PEI/TIR/OxHA and PEI/TIR/GS/OxHA in PBS at 37 °C. (mean ± SD, n = 3 independent samples). (F-G) SEM and AFM images of samples with different coatings. Scale bar, 10 μm.

We next prepared coatings embedded in functional modules by layer-by-layer self-assembly modification. The presence of GS was confirmed by UV–vis spectra of different coatings (Fig. 2D). PEI/GS/OxHA and PEI/TIR/GS/OxHA showed characteristic UV absorption bands at around 300 nm, which was attributed to Sal A in GS [36,37]. The absorption bands were mainly derived from aromatic rings and catechol groups of Sal A, as well as conjugation of hydroxyl and carboxyl groups. Electron transitions in these functional groups, especially π→π∗ and n→π∗ transitions, led to this characteristic UV absorption. Given the Sal A was the key component of GS, these results confirmed the successful introduction of GS into these coatings. As a typical polyphenol with abundant phenolic hydroxyl groups, Sal A could produce hydrogen bonding and electrostatic interaction with substances in the coating. Some quinone groups in Sal A interacted with polyethyleneimine (PEI) through Michael addition or Schiff base reaction, and carboxyl groups and amino groups in gelatin interacted with PEI and oxidized hyaluronic acid (OxHA).

As the anticoagulant agent, the sustained release behavior of tirofiban is desirable. The release profile of tirofiban was shown in Fig. 2E. Due to the presence of GS, the release rate of PEI/TIR/GS/OxHA was lower than that of PEI/TIR/OxHA within 20 days (Fig. S2). GS, composed of polyphenol Sal A and gelatin, interacted with tirofiban through multiple mechanisms, including hydrogen bonding, π-π stacking, and electrostatic interactions. Additionally, the GS nanoparticles acted as a physical barrier, further hindering the diffusion of tirofiban from the coating. After 20 days, the release profiles of the two groups tended to converge, likely due to prolonged immersion leading to partial dissolution or degradation of the coating. This structural loosening reduced the effect of the GS nanoparticles, diminishing their influence on the release rate. With in vitro release of tirofiban lasting approximately 50 days, the PEI/TIR/GS/OxHA exhibited sustained release properties, which was the basis for the long-term anticoagulant ability.

The morphology of coatings was examined using scanning electron microscope (SEM) and atomic force microscope (AFM) (Fig. 2F and G). Compared to other groups, an abundance of nanoparticles was observed on the surface of PEI/TIR/GS/OxHA. However, there was no significant increase in the surface roughness of PEI/TIR/GS/OxHA was detected. The roughness of PEI/GS/OxHA was similar to that of PEI/OxHA, while the roughness of PEI/TIR/OxHA was increased. These results suggested that the GS nanoparticle contributed to the improvement of coating uniformity. Above all, GS nanoparticles not only provided functional diversity to the coating, but also enhanced stability and uniformity.

2.2. Coating suppresses platelet activation

Platelets are a key intersection between coagulation and inflammation, and inhibiting platelet activation is integral to regulating both processes. The anticoagulant tirofiban was introduced into the coating to control the coagulation process. To investigate the anti-thrombotic capacity, samples were incubated with platelet-rich plasma (PRP) for 1 h. As shown in Fig. 3A, a large number of activated platelets were present on the surface of the control group. Compared to the control group, the number of activated platelets on PEI/OxHA and PEI/GS/OxHA was reduced. Due to the effect of tirofiban, PEI/TIR/OxHA significantly inhibited the degree and amount of platelet activation. Among all groups, PEI/TIR/GS/OxHA showed the best anti-platelet activation performance.

Fig. 3.

Hemocompatibility and anti-inflammatory ability of coatings. (A) Platelet adhesion and activation on coatings with and without H₂O₂ stimulation. Scale bar, 10 μm. (B) Images of cell morphology and ROS generation of THP-1 on coatings after treatment with PMA and LPS for 48 h. Scale bar, 50 and 100 μm. (C) Quantitative statistics of ROS expression in THP-1 treated with PMA and LPS (mean ± SD, n = 5 independent samples). (D) IL 6 and IL 1β levels of THP-1 measured by ELISA kits (mean ± SD, n = 3 independent samples). (E) Expression of MMP9 in THP-1 measured by ELISA kits (mean ± SD, n = 3 independent samples). (F-G) Quantitative results and cell flow cytometric prolies of THP-1 treated with PMA and LPS. Cyan, red, and yellow dots represented CD68⁺ CD206^- CD86⁻, CD68⁺ CD206⁺ CD86⁻, and CD68⁺ CD206^- CD86⁺ cell populations, respectively. (mean ± SD, n = 3 independent samples). One-way analysis of variance (ANOVA, t-test) was used to determine the difference between different samples (p value < 0.05 were considered statistically significant).

Hydrogen peroxide (H₂O₂) not only affects the inflammatory response, but also significantly aggravates the coagulation response. GS nanoparticles composed of the antioxidant Sal A can eliminate free radicals and exhibit potential to support anticoagulant activity. Therefore, the coatings were incubated with PRP containing H₂O₂ to investigate the effect of GS nanoparticles on anticoagulant functions (Fig. 3A). H₂O₂ remarkedly aggravated the degree of platelet activation in the control and PEI/OxHA groups. The milder increase in platelet activation in the PEI/GS/OxHA group and the PEI/TIR/OxHA group may be related to the antioxidant capacity of Sal A and the anticoagulant capacity of tirofiban [38]. Notably, PEI/TIR/GS/OxHA still showed the best anti-platelet activation performance with no significant increase in platelet activation. The antioxidant GS nanoparticles boosted the anticoagulant effect of the coating under a combination of oxidative stress and coagulation reactions. Therefore, tirofiban is the major contributor to the anticoagulant capacity, with antioxidant ability providing important support.

2.3. Coating regulates monocyte behavior

The anti-inflammatory ability is critical for vascular stents, since they are exposed to monocyte recruitment after implantation. Recruited monocytes are stimulated to undergo phenotypic transformation, exacerbating the coagulation response and subsequent endothelialization disorder. THP-1 cells, stimulated by Phorbol 12-myristate 13-acetate (PMA) and Lipopolysaccharides (LPS), were co-cultured with different coatings [39,40]. After 48 h of incubation, the morphology of cells adhering to the samples was shown in Fig. 3B The number of THP-1 cells on the coatings was reduced compared to the control group. Due to the anti-inflammatory effect of GS, the number and activation of cells on PEI/GS/OxHA were reduced, and most of the cells were ovoid. The anti-inflammatory capacity of PEI/TIR/OxHA was weaker, and the number of adherent macrophages was reduced, but there were still some cells that were pseudopod-like and in a weakly activated state. Compared with other groups, PEI/TIR/GS/OxHA showed desirable inhibition of monocyte activation. Intracellular reactive oxygen species (ROS) production, another indicator of monocyte activation, was probed by DCFH-DA (Fig. 3B, C and Fig. S3). The results of ROS production and the trend of cell morphology were generally consistent. The PEI/TIR/GS/OxHA exhibited an excellent suppression of intracellular ROS generation. In addition, the pro-inflammatory cytokines IL-1β and IL-6 of PEI/TIR/GS/OxHA were significantly decreased compared to the control group, indicating that PEI/TIR/GS/OxHA regulated the release of monocyte cytokines (Fig. 3D). In the inflammatory microenvironment, monocytes/macrophages express MMP9. Due to GS nanoparticles, the expression of MMP9 in PEI/GS/OxHA and PEI/TIR/GS/OxHA groups were also significantly suppressed after PMA and LPS treatment (Fig. 3E). Furthermore, the phenotypic changes of THP-1 were studied using flow cytometry (Fig. 3F, G and Fig. S4). Differentiation of the THP-1 cells to the M2 phenotype is characterized by changes in the expression of surface markers, in particular the up-regulation of CD206 (mannose receptor) and the down-regulation of CD86 (co-stimulatory molecule) [41,42]. The increased expression of CD206 correlates with the anti-inflammatory properties of M2 macrophages. CD86 is commonly associated with the M1 phenotype, which is pro-inflammatory and involved in pathogen clearance. The presence of CD86 inhibits macrophage differentiation to the M2 phenotype [43,44]. The ratio of CD206 to CD86 expression was significantly increased in cells treated with PEI/TIR/GS/OxHA compared to the control group, suggesting that THP-1 cells on PEI/TIR/GS/OxHA tended to differentiate toward the M2 phenotype.

Collectively, it was substantiated that PEI/TIR/GS/OxHA exhibited superior anti-inflammatory capacity by regulating the morphology, differentiation, and expression of secreted cytokines and proteins in monocytes.

2.4. Effects on interaction between coagulation and inflammation

Thrombus formation primarily involves fibrin, platelets, and leucocytes, which are closely tied to the interaction between coagulation and inflammation. In the process of thrombosis, activated platelets bind to fibrinogen via the integrin receptor CD61 (Platelet membrane glycoprotein IIb-IIIa, Gp IIb/IIIa), which is the final critical step in platelet aggregation. The anticoagulant tirofiban, a Gp IIb/IIIa antagonist in the coating, suppresses platelet aggregation and fibrinogen binding. Therefore, fibrinogen and CD61 were observed by CLSM after immunofluorescence staining (Fig. 4A). As shown in Fig. 4B and D, the control group and PEI/OxHA exhibited an extensive fibrin network with dense platelets and numerous fibrin-platelet interactions. In contrast, the fibrin network on PEI/GS/OxHA was reduced, with inhibited fibrin-platelet interactions. Sal A has been reported to inhibit platelet spreading on immobilized fibrinogen [45], but tirofiban provided a superior anticoagulant property. As shown in Fig. 4B and E, compared to other groups, the fibrin-platelet interactions on PEI/TIR/OxHA and PEI/TIR/GS/OxHA were significantly suppressed. The average branch length of fibrin reflected the extension of fibrin and platelet activation (Fig. 4E), and the fibrin network structure was also investigated (Fig. 4C). The average branch length of fibrin was significantly decreased on PEI/TIR/OxHA and PEI/TIR/GS/OxHA compared to the control group. Overall, PEI/TIR/GS/OxHA effectively inhibited the fibrin network development and fibrin-platelet interactions due to the Gp IIb/IIIa antagonist tirofiban.

Fig. 4.

Interaction analysis of thrombosis and inflammation. (A) Schematic diagram of cell preparation. (B) Images of fibrin and CD61⁺ platelets adhered on coatings after treatment with LPS for 24 h. The location of the amplification was marked with white squares. Scale bar, 40 μm and 20 μm. (C) Images of CD11b⁺ monocytes and CD62p⁺ platelets on coatings after treatment with LPS for 24 h. Scale bar, 20 μm. (D) Area ratio of fibrin to CD61⁺ cells (mean ± SD, n = 5 independent samples). (E) Average branch length of fibrin on coatings (mean ± SD, n = 5 independent samples). (F) Area ratio of CD11b⁺ CD62p⁺ region to CD62p⁺ region (mean ± SD, n = 3 independent samples). One-way analysis of variance (ANOVA, t-test) was used to determine the difference between different samples (p value < 0.05 were considered statistically significant).

Monocytes, a type of leucocytes, play a significant role in thrombosis. Leukocytes are actively recruited to sites of thrombosis by interacting with platelet and endothelial cell adhesion molecules. In the process of coagulation, leucocytes can be induced to express pro-inflammatory and pro-coagulant molecules [[46], [47], [48]]. These mediators can affect various aspects of thrombosis, including platelet activation and adhesion, and activation of the coagulation pathways. Platelets form platelet-monocyte aggregates with monocytes via the integrin receptor CD62p, which contributes to both thrombosis and inflammation [17]. The effects of the coating on the interaction between monocytes and platelets were investigated, with the results shown in Fig. 4C–F. In the control group, CD11b⁺ monocytes displayed irregular shapes. CD62p⁺ platelets were observed around monocytes, some of which were co-localized with monocytes, suggesting the formation of platelet-monocyte aggregates. Compared to the control group, CD11b⁺ monocytes on the modified coatings had rounder nuclei. CD62p⁺ platelets on PEI/OxHA tended to aggregate around monocytes, and platelet-monocyte aggregates were observed. The number of CD11b⁺ monocytes was reduced on both PEI/GS/OxHA and PEI/TIR/OxHA, with monocytes on PEI/TIR/OxHA being rounder and more dispersed. Platelet aggregation and platelet-monocyte aggregates were reduced on PEI/TIR/OxHA, attributable to tirofiban. The number of CD11b⁺ monocytes and platelet-monocyte aggregates on PEI/TIR/GS/OxHA was the lowest, attributed to the combination of GS and tirofiban.

GS and tirofiban are key components of this coating, both of which contributed to the excellent anti-inflammatory and anticoagulant properties of PEI/TIR/GS/OxHA. The PEI/TIR/GS/OxHA inhibited the platelet aggregation and fibrin-platelet interactions by antagonizing Gp IIb/IIIa on platelets, thereby inhibiting thrombosis. Both GS and tirofiban suppressed the formation of platelet-monocyte aggregates, but tirofiban showed superior performance. The PEI/TIR/GS/OxHA coating inhibited the activation of monocytes, and interfered with platelet-monocyte aggregation. The coatings showed the potential to regulate thrombosis and inflammation by affecting the behavior of monocytes and platelets, in particular their interactions and the interaction between fibrin and platelets.

2.5. Regulation of SMCs on proliferation and migration

SMC dysfunction following implantation significantly impacts tissue repair. Activated SMCs contribute to lumen loss through excessive proliferation and synthetic transformation, emphasizing the need to modulate their behavior for functional coatings. Platelet-derived growth factor BB (PDGF-BB) is a representative mediator of SMC activation [49]. To mimic in vivo conditions, HUASMCs were co-cultured with coatings and stimulated by PDGF-BB.

HUASMC proliferation was assessed using the 5-Ethynyl-2′-deoxyuridine (EdU) assay. As shown in Fig. 5A, a large number of HUASMCs were on the control groups, many of which were in an active state of proliferation. The amount of HUASMCs on the functional coatings was reduced compared to the control group. The cell proliferation on PEI/GS/OxHA and PEI/TIR/GS/OxHA was significantly inhibited due to the Sal A in GS nanoparticles [30,38]. The statistics (Fig. 5B) also confirmed that PEI/GS/OxHA and PEI/TIR/GS/OxHA were effective in inhibiting HUASMC proliferation. Given the presence of both endothelial and smooth muscle cells around vascular stents, we also investigated cell proliferation on the coating surface using both the monoculture and co-culture model to investigate the effect of the coating in the competitive growth of ECs and SMCs. The results in Fig. S5 and S6A-F demonstrated that PEI/TIR/GS/OxHA promoted the growth of HUAECs while suppressing the growth of HUASMCs, which was related to the protective effect of Sal A in GS nanoparticles and tirofiban on the ECs [[50], [51], [52], [53]]. As shown in Fig. 5C and D, HUASMCs on the surface of PEI/TIR/GS/OxHA was disadvantaged in the competitive growth, with the ratio of HUASMC to HUAEC being the lowest among all groups. Furthermore, we examined the expression level of genes and proteins associated with HUASMC proliferation. The mRNA levels of PCNA and CCND1 [54,55], encoding PCNA and Cyclin D1 proteins respectively, were significantly lower in the PEI/GS/OxHA and PEI/TIR/GS/OxHA groups compared to the control group (Fig. 5E). PCNA and Cyclin D1 proteins expression was also suppressed in PEI/TIR/GS/OxHA group due to GS nanoparticles (Fig. 5F–H).

Fig. 5.

Regulation of smooth muscle cell proliferation and migration. (A) HUASMC proliferation on coating after 48 h of PDGF-BB treatment. Scale bar, 200 μm. (B) Quantitative analysis of EdU⁺ cells ratio (mean ± SD, n = 3 independent samples). (C-D) Immunofluorescence images quantitative statistics and of HUASMCs and HUAECs co-cultured on coating after 24 h. Scale bar, 100 μm. (mean ± SD, n = 3 independent samples). (E) mRNA levels of PCNA and CCND1 in HUASMCs after PDGF-BB treatment for 48 h (mean ± SD, n = 5 independent samples). (F-H) Western blot analysis of PCNA and Cyclin D1 in HUASMCs following PDGF-BB treatment (mean ± SD, n = 4 independent samples). (I) mRNA levels of MMP2 and MMP9 in HUASMCs after PDGF-BB treatment for 48 h (mean ± SD, n = 8 independent samples). (J) Expression of MMP2 and MMP9 in HUASMCs measured by ELISA kits (mean ± SD, n = 3 independent samples). (K-L) Quantitative statistics and images of Transwell-migrated HUASMCs stained with crystal violet at 12 and 24 h. Scale bar, 200 μm. (mean ± SD, n = 5 independent samples). One-way analysis of variance (ANOVA, t-test) was used to determine the difference between different samples (p value < 0.05 were considered statistically significant).

Inhibition of HUASMC migratory capacity is also crucial in regulating cell behavior. MMPs degrade ECM components, facilitating cell migration [56,57]. The mRNA levels of MMP2 and MMP9, detected by qPCR, were significantly inhibited in the PEI/TIR/GS/OxHA group compared to the control and PEI/OxHA groups (Fig. 5I). Enzyme linked immunosorbent assay (ELISA) results further confirmed this inhibition at the protein level (Fig. 5J). Transwell experiments showed that HUASMC migration was inhibited in all coating groups compared to the control group after 12 and 24 h. The PEI/TIR/GS/OxHA group showed the best inhibitory effect at 24 h, attributed to the regulation of HUASMC behavior by GS nanoparticles (Fig. 5K and L) [29].

Collectively, these results demonstrated that incorporating GS nanoparticles enhanced the function of the material in influencing smooth muscle behavior. The coating effectively modulated the expression of proteins and genes related to the proliferation and migration of smooth muscle cells, displaying potential in regulating intimal stenosis caused by smooth muscle cell disorders.

2.6. Coating improves tissue regeneration on vascular stents in vivo

Clinically, vascular stents are used to treat narrowed blood vessels, which undergo atherosclerosis and suffer from a complex environment. Here, we constructed an atherosclerotic rabbit model via balloon injury and high-fat feeding (Fig. 6A). PLA stents with various coatings were implanted into the abdominal aorta, and the neointima on the inner surface was examined.

Fig. 6.

Stent implantation in atherosclerotic rabbit. (A) Study protocol of stent implantation in rabbit. (B-E) Representative immunohistochemistry images and quantitative statistics of CD68 and SM22α expression in the abdominal aorta implanted with stents. Scale bar, 100 μm. (mean ± SD, n = 3 independent samples) (F) Histological analysis of the abdominal aorta implanted with stents. Scale bar, 400 and 100 μm. (G) Quantitative analysis of neointimal thickness of stents post implantation (mean ± SD, n = 3 independent samples). (H-I) Representative images and quantitative statistics of stents stained with ORO. Scale bar, 400 and 40 μm. (J) Representative SEM images of the abdominal aorta and inner surface of vascular stents in rabbits. The representative oriented cells were marked with red arrows. Scale bar, 50 μm. (K) CD31 and eNOS expression in vessels and neointimal tissue. The struts of vascular stents were marked with white dashed lines. Scale bar, 200 μm. (L-M) Quantitative statistics of CD31 and eNOS expression in stented abdominal aorta. (mean ± SD, n = 3 independent samples) One-way analysis of variance (ANOVA, t-test) was used to determine the difference between different samples (p value < 0.05 were considered statistically significant).

CD68 expressed in macrophages reflected the degree of tissue inflammatory response. The vessel wall and tissue around plaques in injured vessels exhibited an exaggerated inflammatory response compared to the normal vessels. CD68 was highly expressed in the PLA group, indicating an active inflammatory response at the neointima (Fig. 6B, C and Fig. S7, 8). As shown in the images and statistical data, CD68 expression was significantly suppressed in PEI/GS/OxHA, PEI/TIR/OxHA and PEI/TIR/GS/OxHA. The anti-inflammatory and anti-oxidant GS nanoparticles resulted in reduced CD68 expression in PEI/GS/OxHA [31]. The desirably low expression of CD68 in PEI/TIR/OxHA may be associated with its ability to modulate the platelet-macrophage aggerates (Fig. 4), as the aggerates could exacerbate inflammation. PEI/TIR/GS/OxHA coated stents performed remarkable anti-inflammatory ability, attributed to the regulation of macrophages and the interaction between inflammation and coagulation (Fig. 3, Fig. 4).

Dysfunction of SMCs was closely related to restenosis, and contractile SMCs with the proper elasticity and contraction were crucial for healthy vascular regeneration. The expression level of α smooth muscle actin (αSMA) and smooth muscle 22 alpha (SM22α), markers of contractile SMCs, was detected. Images and statistical data showed a significant increase in the expression of αSMA and SM22α in PEI/GS/OxHA and PEI/TIR/GS/OxHA compared to PLA (Fig. 6D, E and Fig. S9, 10). Due to GS nanoparticles, PEI/TIR/GS/OxHA effectively regulated SMC proliferation and migration (Fig. 5), and further showed the highest αSMA and SM22α levels among all groups. Neointimal thickness was calculated based on H.E. images to assess lumen occlusion (Fig. 6F and G). The neointima of the PLA group was the thickest among all groups. All coated stents showed thinner neointimal compared to PLA, and the PEI/TIR/GS/OxHA was the thinnest.

Lipid content of atherosclerotic plaques on the aorta were determined by Oil red O (ORO) (Fig. 6H and I). Plaque was observed in the injured vessels. After implantation of the bare and PEI/OxHA stent, the plaques remained and were merely physically encapsulated by neoplastic tissue. In the PEI/GS/OxHA and PEI/TIR/OxHA groups, the lipid content of the plaques was reduced, especially in PEI/GS/OxHA. This phenomenon may be related to the capacity to modulate the inflammatory response (Fig. 3, Fig. 4). Under the combined effect of GS nanoparticles and tirofiban, PEI/TIR/GS/OxHA, with the best anti-inflammatory ability, showed the potential to reduce lipids and stabilize the plaque environment.

Furthermore, the morphology of vessel inner surfaces and regenerative tissue on stents were observed. As shown in Fig. 6H, dense fibrous tissue was observed in the inner wall of normal blood vessels, whereas injured vessels were disorganized, suggesting successful injury modeling. The direction of blood flow was shown by the arrow in Fig. 6J, and the direction of stents was consistent with the blood flow direction. The neointima on bare stents was disordered and showed no obvious orientation. PEI/OxHA was better, but the cell layer on PEI/GS/OxHA, PEI/TIR/OxHA and PEI/TIR/GS/OxHA coated stents were more regular and oriented. Among all coated groups, PEI/TIR/GS/OxHA exhibited the densest and most organized neointimal layer.

In addition to morphology, the expression of CD31 and endothelial NO synthase (eNOS) was evaluated by immunofluorescence staining (Fig. 6K, L, M). CD31, found at endothelial intercellular tight junctions, is mainly used to confirm endothelial cells [7]. A regular pattern of CD31 distribution was observed in the intima of healthy vessels, whereas the intima of injured vessels had haphazard regions of CD31 staining. In the PLA and PEI/OxHA groups, the tissue covered the stent surface well according to SEM images, but CD31 expression was minimal. The area of CD31 expression increased in the PEI/GS/OxHA and PEI/TIR/OxHA, with almost the entire stent area in the PEI/TIR/OxHA group expressing CD31. These findings suggested that GS nanoparticles and tirofiban favored endothelial regeneration, with tirofiban performing better. CD31 was most strongly expressed on the surface of the PEI/TIR/GS/OxHA group, with the largest number of CD31-positive cells. The above results, along with the statistical data on the percentage of the CD31-positive area, demonstrated that PEI/TIR/GS/OxHA effectively enhanced reendothelialization after implantation. Additionally, the expression of eNOS was also investigated. eNOS is an important enzyme in ECs and is regarded as a key player in NO production [7]. The trend of eNOS expression in each group was largely consistent with that of CD31.

Overall, PEI/TIR/GS/OxHA showed superior performance in modulating inflammation, promoting vascular SMC function, reducing atherosclerotic plaque lipid content, and enhancing reendothelialization. These findings suggested that controlling coagulation and inflammatory responses benefited subsequent reendothelialization. When combined with SMC regulation, this approach can effectively address atherosclerosis-related vascular issues and significantly enhance vascular healing and stability.

2.7. Coating improves tissue regeneration on vascular stents in vivo

Six weeks post-implantation of abdominal aortic stents in rabbits, transcriptomic analysis was conducted on the involved vessels to assess the effect of modified coatings on re-endothelialization and neointimal formation. Gene expression profiles were visualized using gene heatmaps (Fig. 7A). Differentially expressed genes between the PEI/OxHA, PEI/GS/OxHA, PEI/TIR/OxHA, PEI/TIR/GS/OxHA groups, and the PLA group were compared (Figs. S11A–C). Inter-group relationships were depicted using Venn diagrams (Fig. 7B). In general, representative gene expression profiles related to endothelial cell function, SMC phenotypic transition, coagulation processes, and inflammatory responses were presented using stacked violin plots (Fig. 7C–F). Integrin α3 (ITGA3) promotes platelet aggregation during coagulation process [58], while P-selectin, mainly expressed in activated endothelial cells and platelets, influences the recruitment of leukocytes and leukocyte microparticles to thrombi [59]. The results revealed that the expression of ITGA3 and P selectin (SELP) decreased following tirofiban treatment, with further suppression observed upon the addition of GS nanoparticles. M2 macrophage markers like interleukin-10 (IL-10) and arginase-1 (ARG-1) were upregulated after GS nanoparticle treatment [60], while the expression of inflammatory factors interleukin-18 (IL-18) and interleukin-1β (IL-1β) decreased [61]. This indicated that GS nanoparticles promoted an anti-inflammatory macrophage phenotype, with tirofiban having minimal additional effect. Concurrent changes in endothelial cell and SMC functional markers validated that the combined action of GS nanoparticles and tirofiban promoted endothelialization and facilitated SMC phenotypic transition.

Fig. 7.

Transcriptomic analysis of modified stent coatings in vascular repair. (A) Gene expression heatmaps among five groups. (B) Venn diagram of differentially expressed genes between the PEI/OxHA, PEI/GS/OxHA, PEI/TIR/OxHA, PEI/TIR/GS/OxHA groups, and the PLA group. (C-F) Stacked violin plots representative gene expression profiles considering endothelial cell function, SMC phenotypic transition, coagulation processes, and inflammatory responses. (G-I) GO enrichment analysis of PLA with PEI/GS/OxHA, PEI/TIR/OxHA and PEI/TIR/GS/OxHA, respectively. (J-L) KEGG enrichment analysis of PLA with PEI/GS/OxHA, PEI/TIR/OxHA and PEI/TIR/GS/OxHA, respectively. (M-O) GSEA Enrichment analysis between the PEI/TIR/GS/OxHA group and PLA group. (P) Network map of top 10 Hub genes depicted by CytoHubba in Cytoscape. Top 10 nodes were ranked using MCC scores and presented by purple from dark to light.

Enrichment analysis was conducted separately on differentially expressed genes among these groups. The results showed that, compared to the PLA group, the PEI/GS/OxHA group had enriched pathways involved in regulating macrophage differentiation and Mitogen-Activated Protein Kinase (MAPK) cascade (Fig. 7G–J). Meanwhile, the PEI/TIR/OxHA group primarily exhibited enrichment in pathways related to complement and coagulation cascades, as well as inflammatory pathways involving tumor necrosis factor (TNF) and interleukin-17 (IL-17) (Fig. 7H–K). In comparison between the PEI/TIR/GS/OxHA and PLA groups, pathways involving complement and coagulation responses, macrophage activation, participation in ROS pathways, and signaling pathways related to endothelial cell function like vascular endothelial growth factor (VEGF) were enriched (Fig. 7I–L). Additionally, functions related to lipid transport and foam cell differentiation were enriched in this group, suggesting that the combination of tirofiban and GS nanoparticles played a role in modulating lipids, which explained the regulatory effects of PEI/TIR/GS/OxHA on plaques in Fig. 6. Previous studies have demonstrated that Sal A regulates lipid accumulation in metabolic diseases such as fatty liver disease and obesity-induced diabetes [62,63]. Meanwhile, gene set enrichment analysis (GSEA) analysis indicated the enrichment of pathways such as lipid and atherosclerosis, peroxisome proliferators-activated receptors (PPAR) signaling pathway, and cholesterol metabolism in the PEI/TIR/GS/OxHA groups, where lipid-associated pathways also elucidated the variability of plaques in the samples. (Fig. 7M − O, Figs. S11D–F). A protein-protein interaction (PPI) network was constructed using cn.string-db.org, analyzing differentially expressed genes between the PEI/TIR/GS/OxHA and PLA groups. The maximal clique centrality (MMC) scores of individual nodes were computed using CytoHubba in Cytoscape software. Among all the differentially expressed genes, the top 10 key genes were identified. Collagen type I α2 (COL1A2) encodes the α chain of type I collagen and is defined as a marker of coagulation [64,65]. Collagen type V alpha 2 (COL5A2) is one of the subunits of Type V collagen (COL5), which is a fibrillar collagen co-distributed with type I collagen [66]. And Serpin Family F Member 2 (SERPINF2) is an inhibitor of fibrinolysis [67]. COL1A2, COL5A2, and SERPINF2 mainly play roles in collagen fibril organization [[68], [69], [70]], while both SERPINF2 and Calponin 2 (CCN2) also affect cell differentiation processes [71,72]. These findings align with the previously mentioned anticoagulant effects and the regulation of smooth muscle cell phenotype transition. In summary, based on transcriptomic results, PEI/GS/OxHA primarily counteracted inflammatory responses in stented vessels. PEI/TIR/OxHA mainly exerted anticoagulant effects and influenced inflammatory responses. Combined with GS and tirofiban, PEI/TIR/GS/OxHA further reduced lipid and cholesterol deposition at the injury site while mitigating inflammatory and coagulation responses, thereby stabilizing the vascular microenvironment.

3. Conclusion

In this work, we introduced a self-responsive coating for vascular stents, which is capable of interfering tissue responses and regulating regeneration. Due to anticoagulant tirofiban, the coating effectively inhibited thrombus formation, and interfered the interaction between coagulation and inflammation with the assistance of GS nanoparticles. With the presence of Sal A, the nanoparticles biased the monocytes on coatings towards the M2 phenotype. Meanwhile, it modulated the proliferation and migration of SMCs in vitro. These functions make it possible for coating to suppress the drawbacks from coagulation and inflammation, and guarantee healthy endothelization in vivo. The ability to inhibit synthetic SMCs enabled the coating to reduce intimal hyperplasia. The coating addressed atherosclerosis-related vascular issues, significantly improving vascular healing and stability. By incorporating different essential functions, it even showed potential for plaque improvement. The whole-process regulation of tissue response post-implantation escorted the long-term functionality and ideal vascular regeneration. Through this work, functional coating could provide insights into overcoming limits of vascular stents, and optimizing the design of modification on stents even other blood-contacting materials.

4. Experimental section/methods

4.1. Preparation and characterization of nanoparticles

CPBA (0.3 g) was dissolved in 2 ml of DMSO. EDC (0.052 g) and NHS (0.025 g) were dissolved in 2 ml of DMSO, and mixed with the aforementioned solution under stirring. Following 24 h, 2 ml of DMSO containing 0.1 g of gelatin was added and the reaction proceeded for an additional 24 h. The CPBA-GEL was acquired through dialysis against water and freeze-drying.

To synthesize the nanoparticle, 5 mg of CPBA-GEL was dissolved in 1 ml of PBS, while 5 mg of Sal A was dissolved in 0.6 ml of DMSO. The CPBA-GEL solution was added dropwise to Sal A solution under stirring. After 12 h, the Sal A nanoparticles (named as GS) were obtained through dialysis against PBS.

The particle size of nanoparticles was determined by dynamic light scattering (DLS).

To investigate the drug release behavior of nanoparticles, GS solution was added into a dialysis bag and immersed in solutions of varying MMP9 concentrations (0, 1, 5, 10 nm) at 37 °C under shaking. At specific intervals, 1 ml of the solution was withdrawn and substituted with 1 ml of fresh MMP9 solution. The amount of released Sal A was quantified with UV spectroscopy.

4.2. Coating preparation

Sodium periodate (NaIO₄) solution (2.87 M) was dropped into a hyaluronic acid solution (25 mg/ml). The mixture was stirred in darkness for 24 h, followed by the addition of ethylene glycol (0.6 ml) to quench the reaction. Oxidized hyaluronic acid (Ox-HA) was obtained by dialysis in deionized water (DI water) and subsequent freeze-drying.

Before modification, substrates were immersed in 2 mg/ml dopamine solution (Tris buffer, 10 mM, pH 8.5) for 2 h. Substrates were sequentially dipped into PEI (2 mg/ml), GS (2 mg/ml), and tirofiban (1 mg/ml)-OxHA (2 mg/ml) solution. The cycle was repeated for 10 times, and the obtained coating was named PEI/TIR/GS/OxHA. The coating composed of PEI and OxHA was denoted as PEI/OxHA. The coating comprising PEI, tirofiban and OxHA was labeled as PEI/TIR/OxHA, and the coating consisting of PEI, GS and OxHA was designated as PEI/GS/OxHA.

The morphology and roughness of coatings were investigated by scanning electron microscopy (SEM) and atomic force microscope (AFM).

To determine the existence of Sal A, the UV spectra of coated quartz plates were measured using a UV/Vis spectrophotometer.

The PEI/TIR/OxHA@GS was immersed in 2 ml of PBS at 37 °C. At predetermined time points, 1 ml of PBS was withdrawn, and an equivalent volume of fresh PBS was added. The quantification of released tirofiban was performed using UV measurements.

4.3. In vitro hemocompatibility evaluation

All animal experiments procedures were approved by the Chinese Institutional Animal Care and Use Committee at Sir Run Run Shaw Hospital, School of Medicine, Zhejiang University. New Zealand white rabbits were supplied by Zhejiang Chinese Medical University (China). Samples were respectively incubated with 500 μl rabbit platelet-rich plasma (PRP) at 37 °C. To investigate the relationship between antioxidant capacity and anticoagulant capacity of coatings, H₂O₂ was added to PRP, and then the samples were incubated in the PRP with 100 μmol H₂O₂. After 1 h, all samples were taken out and rinse with PBS. Adherent platelet and clots were observed by SEM.

4.4. In vitro anti-inflammation evaluation

Samples were placed into 24-well plates and then 1 ml of THP-1 cell (Pricella Life Science&Technology, CL-0233) suspension containing 160 nM PMA and 100 ng/ml LPS was added to each well. After 48 h, the suspensions were collected, and the expression of IL-6 and IL-1β was assessed using ELISA kits. To observe the morphology of THP-1, cells were fixed for 20 min, stained with rhodamine-conjugated phalloidin (100 nM) and 4’,6-diamidino-2-phenylindole dyes (DAPI), and observed by confocal laser scanning microscopy (CLSM, Nikon A1 Ti Confocal Laser Scanning Microscope, Nikon, Japan). For ROS detection, cells were stained with 2,7-Dichloro di-hydrofluorescein diacetate (DCFH-DA) (10 μM). The phenotype of THP-1 was determined by cell flow cytometry (CytoFLEX LX Flow cytometer, Beckman, the USA) after labeling with CD68, CD86 and CD206.

4.5. Interaction analysis of macrophages and platelets

Mouse blood was collected from the orbital venous plexus under anesthesia, and diluted with PBS at a 2:1 ratio. After gentle mixing, the monocyte separation medium was added, and the mixture was centrifuged at 300×g for 30 min without brake. The upper platelet-rich layer and the middle monocyte layer were extracted. 75 nM PGI₂ was added to the platelet-rich supernatant. Platelets were counted and stored at room temperature. The monocyte layer was washed with 1 × PBS, and a lysis buffer was added to lyse red blood cells. Monocytes were resuspended in RPMI medium containing 10 % FBS and 1 % penicillin-streptomycin, then counted and stored at room temperature.

Monocytes (2 × 10⁵ cells/ml) were seeded on the surface of coatings, and platelets (2 × 10⁷ cells/ml) were added. 100 ng/mL LPS was added to stimulate cells. After 24 h of incubation, cells were fixed and permeabilized. Coatings with cells were incubated overnight with CD62p and Fibrinogen antibodies, or CD11b and CD61 antibodies. Subsequently, cells were incubated with corresponding fluorescent secondary antibodies and observed using CLSM.

4.6. Migration and proliferation of HUASMCs

4.6.1. Proliferation assay of HUASMCs using EDU staining

To investigate the proliferation, Human umbilical artery smooth muscle cells (HUASMCs) were seeded onto samples at a density of 2 × 10⁵ cells/ml, and phenotypically differentiated with 20 ng/ml PDGF-BB. After 48 h, cells were incubated with 20 μM EDU working solution at 37 °C for 2 h. Subsequently, the cells were fixed for 20 min, and permeabilized with 0.5 % Triton X-100 for 5 min. Click additive solution was added in the dark. 30 min later, cells were stained with Hoechst 33342 staining solution for 10 min in the dark. The percentage of EDU positive cells were analyzed using CLSM.

4.6.2. mRNA expression analysis via qPCR

Total RNA from HUASMCs was extracted using the RNA Quick Extraction Kit. RNA quality was assessed by Nanodrop and qualified when the absorbance ratio of 260 nm/280 nm was between 1.8 and 2.0. Subsequently, the quantified RNA was mixed with RT Premix for reverse transcription (LifeECO PCR, BIOER, China). The resulting cDNA was then mixed with corresponding primers and SYBR Green Master Mix, and real-time quantitative PCR (qPCR) was performed using the Applied Biosystems quantstudio 6 Flex system (QuantStudio 6 Quantitative Real-time PCR System, Applied Biosystems, the USA). Each sample was analyzed in duplicate, and the relative expression levels of genes were quantified using the ΔΔCT method. GAPDH was used as a reference gene, and specific primer sequences are provided in Table S1.

4.6.3. MMP9 expression in HUASMC

After processing the cells, discard the cell supernatant, digest the cells with trypsin, and centrifuge to collect the cells. Collect the cell lysate by ultrasonic cell disruption combined with freeze-thaw cycles. Following centrifugation to remove cell debris, the MMP9 expression in cells is measured using the MMP9 ELISA kit.

4.6.4. Protein expression analysis via Western blot

After digestion and centrifugation, RIPA lysate containing 1 mm PMSF was added to the cell precipitate and lysed on ice for 10 min, and the supernatant was taken by centrifugation at 12000 rpm for 30 min. The protein concentration was subsequently determined by BCA kit, quantified and denatured at high temperature by adding SDS-PAGE Loading Buffer. Subsequently, proteins were added to the gel for electrophoresis (80V, 2h) by Biorad Protein Electrophoresis and Blotting System, and then transferred from the gel to the PVDF membrane (300 mA, 1.5h). The membranes were closed with TBST buffer containing 5 % skimmed milk powder (0.5 % Tween-20 in TBS buffer) for 2 h at room temperature. The proteins were then sequentially incubated with primary antibodies overnight at 4 °C and secondary antibodies of the corresponding species for 2 h at room temperature, during which the membranes were washed three times with TBST for 5 min each time (Mini-PROTEAN Tetra cell, Mini Trans-Blot Module, and PowerPac Basic Power Supply, BioRad, the USA). After final incubation with ECL working solution, images were captured using a biorad chemidoc system (ChemiDoc MP, BioRad, the USA) and analyzed with Image J software.

4.6.5. Transwell assay for HUASMC migration analysis

HUASMCs (2 × 10⁵ cells/ml) were seeded on samples, and the cells were induced to phenotypically differentiate with 20 ng/ml PDGF-BB. After 48 h, the cells were digested and then seeded at a density of 5 × 10⁴ cells/ml into Transwell chambers (pore size of 8 μm) and ECM medium was added to the bottom chambers. Following incubation for 12 and 24 h, the cells were fixed for 20 min and incubated with crystal violent staining solution for 20 min. The upper layer of cells was removed. The number of cells migrated to the bottom chamber was observed using a microscope.

4.7. Competitive proliferation of HUAECs and HUASMCs

4.7.1. Proliferation assay of HUASMCs and HUAECs

To investigate the proliferation separately, HUASMCs and Human umbilical artery endothelial cells (HUAECs) were respectively seeded onto samples at a density of 2 × 10⁵ cells/ml. After culturing 24 h, cell proliferation was assessed through the EdU assay with all subsequent procedures adhering to the methodology outlined in section 4.6.1. The cell counts and percentages of EdU positive cells were analyzed using CLSM.

4.7.2. Co-culture of HUASMCs and HUAECs

HUASMCS were pre-labeled by incubating in RPMI 1640 medium containing 10 μM celltracker™ Green CMFDA probe for 20 min, while HUAECs were pre-labeled by incubating in RPMI 1640 medium containing 10 μM celltracker™ Orange CMTMR probe for 20 min. All cells were washed and digested, then resuspended in RPMI 1640 medium containing 10 % FBS and 1 % penicillin-streptomycin at a 1:1 ratio. They were co-cultured on samples for 1 day. The growth of both cells was observed using CLSM.

4.8. Induction of atherosclerosis and stent placement

New Zealand White rabbits (2.8–3.0 kg) were fed an atherogenic diet (1 % cholesterol and 6 % peanut oil) for 5 weeks. At the end of the first week, balloon injury of abdominal aorta was induced using a balloon catheter. After 4 weeks, a PLA vascular stents (2.75 mm) were implanted in the abdominal aorta and the diet was switched to reduced cholesterol (0.025 % cholesterol) until euthanasia. All animals orally received aspirin (25 mg/d) and clopidogrel (5 mg/d) for 3 days, and were injected with penicillin (800000 units/d) for 3 days. Six weeks after implantation, animals were anesthetized and the vessels with embedded stents were fixed in PFA. Vessels and stents with surrounding tissue were stained with oil red O (ORO) and hematoxylin-eosin (H.E.).

After staining with anti-CD31, anti-eNOS and DAPI, the neointima on the vessel surface was observed by SEM and CLSM. The expression of CD68 and α-SMA in the vessel cross-section was detected by immunohistochemical staining and quantified by ImageJ software.

Six weeks post-stent implantation, vascular tissue was collected from the implantation site for Bulk RNA-seq analysis. Total RNA was extracted, and the enriched mRNA was fragmented and amplified using PCR. Libraries were constructed based on the resulting cDNA products and sequenced using the Illumina Hiseq2000 platform from BGI-Shenzhen, China. Data screening and quality control were conducted using the SOAPnuke system (https://github.com/BGI-flexlab/SOAPnuke), with subsequent data processing performed on the Dr. TOM analysis platform.

4.9. Statistical analysis

All experiments were repeated at least three times, and the results were expressed as mean ± standard deviation. One-way analysis of variance (ANOVA, t-test) was used to determine statistical significances between different samples undergoing the same treatment, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. Histomorphometric, immunohistochemical and immunofluorescent staining results were analyzed by Image-Pro Plus software.

CRediT authorship contribution statement

Qiongjun Zhu: Writing – review & editing, Writing – original draft, Resources, Methodology, Investigation, Data curation, Conceptualization. Zhezhe Chen: Resources, Methodology, Data curation. Dan'an Wang: Methodology, Investigation. Xiaolu Jiao: Investigation, Conceptualization. Yi Luan: Resources, Methodology. Min Wang: Resources, Data curation. Rifang Luo: Writing – review & editing, Conceptualization. Yunbing Wang: Writing – review & editing, Supervision, Resources. Guosheng Fu: Supervision, Investigation, Conceptualization. Yanan Wang: Writing – review & editing, Writing – original draft, Supervision, Investigation, Funding acquisition, Data curation, Conceptualization. Wenbin Zhang: Writing – review & editing, Supervision, Resources, Investigation, Funding acquisition, Conceptualization.

Data availability statement

The main data supporting the results of this study are available within the paper and its Supplementary Information. Any additional requests for information can be directed to, and will be fulfilled by, the corresponding author.

Ethics approval and consent to participate

All animal experiments were approved by the Laboratory Animal Research Center of Sir Run Run Shaw Hospital, Zhejiang University School of Medicine. The animal protocols were assessed and approved by the Animal Laboratory Ethics Committee of Zhejiang University (SYXK2017-0006).

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 B. Supplementary data

The following is the Supplementary data to this article: