Exosome-eluting stents for vascular healing after ischaemic injury

Abstract

Drug-eluting stents implanted after ischaemic injury reduce the proliferation of endothelial cells and vascular smooth muscle cells and thus neointimal hyperplasia. However, the eluted drug also slows down the re-endothelialization process, delays arterial healing, and can increase the risk of late restenosis. Here, we show that stents releasing exosomes derived from mesenchymal stem cells in the presence of reactive oxygen species enhance vascular healing in rats with renal ischaemia-reperfusion injury, promoting endothelial-cell tube formation and proliferation, and impairing the migration of smooth muscle cells. Compared with drug-eluting stends and bare-metal stents, the exosome-coated stents accelerated re-endothelialization and decreased in-stent restenosis 28 days after implantation. We also show that exosome-eluting stents implanted in the abdominal aorta of rats with unilateral hindlimb ischaemia regulated macrophage polarization, reduced local vascular and systemic inflammation, and promoted muscle-tissue repair.

The most common cause of peripheral artery disease is atherosclerosis¹. Markers of inflammation and thrombosis, elevated lipoprotein and homocysteine levels, and chronic kidney disease are also associated with peripheral artery disease². Angioplasty and stent placement are major approaches to open blocked arteries³. Bare metal stents (BMS) induce proliferative and inflammatory reactions due to foreign body reactions. This causes neointimal hyperplasia, followed by in-stent restenosis (ISR)⁴. As a mitigation strategy, anti-proliferative drug-eluting stents (DES) have been widely applied to reduce ISR. However, as a reparative process resembling wound healing, ISR remains a major issue limiting the long-term efficacy of DES⁵. Recent studies have been focusing on the development of polymer-free stent to avoid the inflammatory reactions against the polymers^{6, 7}, on functional coating of DES or BMS to capture circulating endothelial progenitor cells for fast endothelization⁸, and on novel metal materials to prevent restenosis and thrombosis⁹. We sought to develop a stent with a coating of biologics that could not only regulate vascular remodeling and inflammation but also promote the regeneration of the injured tissue.

As the small size fraction of extracellular vesicles, exosomes derived from mesenchymal stem cells (MSCs) are known to ameliorate inflammation and promote endothelial proliferation and migration¹⁰. Injection of exosomes secreted from MSCs (MSC-XOs) has shown promise in treating ischemic injuries such as hindlimb ischemia¹¹, myocardial infarction¹², and renal ischemia injury¹³. MSC-XOs have been proposed to treat a variety of inflammatory and degenerative conditions in numerous clinical trials for immunomodulation and tissue regeneration^{14, 15}, including a phase II/III clinical trial to ameliorate chronic kidney disease¹⁶ and a phase I/II clinical trial to treat ischemic stroke¹⁷. In addition, pro-angiogenesis effects of MSC-XOs in ischemic diseases have been studied¹⁸. In the present study, we tested the hypothesis that vascular stents can be ideal carriers to deliver therapeutic exosomes to the ischemic tissue. Synergistically, the exosome coating on stents can improve biocompatibility, inhibit ISR, and promote vascular healing.

To those ends, we designed and tested a bioresponsive exosome-eluting stent (EES), taking advantage of the elevated level of reactive oxygen species (ROS) from the mechanical injury from stent deployment^{19, 20}. ROS is also reportedly a biomarker in vascular diseases, including atherosclerosis²¹. Studies have shown that an increase in oxidative stress and inflammatory response led to neointimal formation, vascular smooth muscle cell (SMC) proliferation²², and adverse extracellular matrix deposition²³.

Results

Fabrication and characterization of exosome-eluting stents

The fabrication strategy of exosome-eluting stents (EES) is shown in Fig. 1a and Fig. S1. Briefly, 1,2-distearoyl-sn-glycero-3-phosphoethanolamine (DSPE)-conjugated stents were prepared first and then integrated with MSC-XOs with an ROS linker. To coat MSC-XOs onto the stent, we first modified the stents’ surface by 3-aminopropyltriethoxysilane (APTES) to generate amines groups for further modification. Silanization with APTES was performed after surface hydroxylation. Then, 4-Carboxyphenylboronic acid was reacted with amino groups on the stent. DSPE-PEG₅₀₀₀-NHS was reacted with 3-Amino-1,2-propanediol to provide the dihydroxyl groups. The dihydroxyl modified DSPE was then incubated with the stent overnight to generate DSPE-coated stents²⁴. After that, the DSPE-modified stent was incubated with 10¹² exosomes in 4 °C overnight to generate the final product EES. MSC-XOs were collected and characterized as previously described¹². The size at the peak concentration was 127.1 nm (Fig. 1b). TEM image confirmed the morphology of MSC-XOs (Fig. 1c). Western blot further proved the expressions of common exosomal surface markers such as Alix, TSG101, and CD81 (Fig. S2a). MicroRNA sequencing revealed specific cargo miRNAs in MSC-XOs (Fig. S2b), including those that have been reported to mediate angiogenesis and tissue repair²⁵, such as miR-23²⁶ and the let-7 family²⁷. Furthermore, proteomics studies of MSC-XOs identified a variety of proteins, including fibronectin, collagen alpha-1, and thrombospondin-2 (Fig. S2c). EES was characterized by X-ray photoelectron spectroscopy (XPS, Fig. S3) and time-of-flight secondary ion mass spectrometry (ToF-SIMS, Figs. 1d&e). XPS demonstrated the binding energy change of Fe 2p 3/2 during surface hydroxylation and silanization. The area changes of N 1s, C 1s, and O 1s indicated the conjugation of DSPE and MSC-XOs. To confirm the efficiency of coating, ToF-SIMS spectrometry was employed to examine the molecular compositions of EES by direct chemical detection of exosome membranes²⁸. The lack of Cr signal (m/z 52) from BMS indicated an exosome coverage (Fig. 1d). Membrane phospholipids of exosomes showed signals at C₃H₆NO₂⁺ (m/z 88) from the amino acid serine, which was present in phosphatidylserine, and C₅H₁₂N⁺ (m/z 86) from phosphocholine, which was present in phosphatidylcholine²⁹ (Fig. 1e). Those signals confirmed the effective coating of MSC-XOs on the stent. SEM imaging showed a smooth surface of BMS but a nanoscale roughness surface of EES due to exosome coating (Fig. 1f). The exosome-based nanoscale coating could accelerate endothelial cell adhesion and proliferation^28,30 and inhibit the adhesion and activation of platelets³¹.

Fig. 1 |. Fabrication and characterization of EES.

a, Schematic illustration of the EES fabrication process. In brief, from bare metal stent (BMS) to exosome eluting stent (EES), the preparation process includes hydroxylation, silanization, reactive oxygen species (ROS) responsive linker, DSPE layer and exosome layer. PEG₅₀₀₀ serves as a spacer to reduce steric hindrance and improves flexibility of DSPE. DSPE is a diacyl lipid and it has been widely used for the exosome membrane insertion. b, Size distribution of MSC derived exosomes (MSC-XOs) examined by NanoSight. n=5. c, Representative TEM image of MSC-XOs. Scale bar, 500 nm. d-e) ToF-SIMS spectrometry showing the surface signal of Cr from BMS (d) and C₅H₁₂N⁺ and C₃H₆NO₂⁺ signal from EES (e). Three technical replicates. f, Representative SEM images of EES and BMS at different magnifications. The red squares are enlarged images.

In vitro exosome release from EES and blood compatibility

To release MSC-XOs into the bloodstream and peripheral ischemic tissue timely, benzeneboronic acid pinacol ester group was used here to react with local ROS stimuli³² (Fig. 2a). Under normal conditions, MSC-XOs were slowly released from the EES. Around 20% of exosomes were released from EES after 48 hours. With the addition of H₂O₂, around 40%~60% exosomes were released (Fig. 2b) in 48 hours. To confirm the bioactivity of exosomes after eluting, released exosomes and EES were further characterized. The size of released exosomes didn’t change (Fig. 2c). Fig. 2d showed the morphology of EES after 2-days of release studies in PBS and 0.1 mM H₂O₂, respectively. To verify the biocompatibility of EES, a small piece of EES (with red fluorescent DiD-labeled MSC-XOs) was co-cultured with human umbilical vein endothelial cells (HUVECs) for 12 hours on Matrigel (Fig. 2e). HUVEC tube formation and uptake of the released MSC-XOs is evident (Fig. 2e). In addition, blood compatibility of EES was evaluated in vitro by incubating BMS or EES with platelet-rich plasma (PRP) and inflammatory cells. Compared to BMS, EES significantly reduced the adhesion of both activated platelets and TNF-α activated U937 monocytes (Figs. 2f–h).

Fig. 2 |. In vitro ROS-trigged exosome release and biocompatibility of EES.

a, Schematic illustration of ROS-responsive EES. Upon exposure to H₂O₂, the aryl boronic ester group is oxidized and subsequently hydrolyzed to unmask a phenol group. b, Accumulative release of MSC-XOs from EES in PBS with or without H₂O₂ (100 μM or 1 mM). n=3. c, Size distribution of MSC-XOs released from EES determined by NanoSight. n=5. d, Representative SEM images of EES after release studies. Top image: EES in PBS; bottom image: EES in 100 μM H₂O₂. Scale bar, 20 μm. e, MSC-XOs (DiD-labeled) release and uptake by HUVECs (at 4 h and 12 h). Scale bar, 100 μm. n=5. f-g, Representative fluorescent images and SEM images of BMS and EES after incubation with platelet-rich plasma (f) and activated monocytes (g). Adhered platelets and U937 monocytes were pointed by red arrows. h, Quantification of adherent platelets and monocytes. n=5. P values are shown on the graph. Student’s two-tailed independent t-test was used to determine differences between two groups.

EES promotes endothelial cell proliferation and inhibits smooth muscle cell migration

Released exosomes are likely to interact with endothelial cells (ECs), SMCs, and the adjacent injured tissue (Fig. 3a). 4 hours of co-culture with EES promoted endothelial tube formation (Fig. 3b), with an increase in the numbers of endothelial network nodes (Fig. 3c) as well as tube lengths. HUVEC adhesion to EES was evident (Fig. S4). MSC-XOs reduced the expression of von Willebrand Factor (vWF) in HUVEC (Figs.3d & e). It has been reported that downregulation of vWF promotes the proliferation and migration of endothelial cells³³. CCK-8 assay further verified the enhanced proliferation of HUVECs with EES (Fig. 3f). To examine endothelialization in vitro, stents were cultured with confluent HUVEC for additional 4-days and imaged by scanning electron micrograph (SEM). SEM revealed the remarkable HUVEC coverage on EES but not on BMS (Figs. 3g & h). We also observed similar effects of EES on human primary coronary artery endothelial cells (HCAECs, Fig. S5). Collectively, MSC-XOs eluted from EES encouraged the proliferation and tube formation of both HUVECs and HCAECs. In contrast, EES inhibited trans-well migration capacity of SMCs (Fig. S6; Figs. 3h–j). We also observed an up-regulation of α-SMA expression in SMCs co-cultured with EES (Figs. 3k&l), as compared to BMS. This suggests phenotypic modulation effects from the released exosomes³⁴. There was no significant change in SMC proliferation (Fig. 3m). Taken together, those in vitro experiments indicated that EES promotes endothelial tube formation and proliferation but inhibits the migration of SMCs.

Fig. 3 |. EES promotes the proliferation and migration of endothelial cells but inhibits the migration of smooth muscle cells.

a, Schematic illustration of the effects of EES on endothelial cells, smooth muscle cells, and injured tissue. b, HUVEC tube formation with BMS or EES. Scale bars, 50 μm. c, Quantification of HUVEC network nodes. n=5. Three technical replicates. d, von Willebrand Factor (vWF) expression (green) of HUVECs with BMS or EES (exosomes were labeled with DiD (red)) were imaged via confocal microscopy, scale bars, 50 μm, and e, Quantification of relative vWF expression. f, CCK-8 cell proliferation assay on HUVECs co-cultured with BMS or EES for 48-h. g, Representative images of in vitro endothelialization of HUVECs on BMS and EES. Scale bars, 20 μm. h, Quantification of HUVEC coverage on stent struts. i, With BMS or EES in the lower chamber, trans-well migrated smooth muscle cells (SMCs) were stained by crystal violet, scale bars, 50 μm. j, Quantification of migrated SMCs. k, Representative microscopic images showing α-SMA (green) expression in SMCs co-cultured with BMS or EES. Scale bars, 100 μm. l, Corresponding quantification. m, CCK-8 cell proliferation assay of SMC co-cultured with BMS or EES for 48-h. n=5. Three technical replicates. P values are shown on the graphs. Student’s two-tailed independent t-test was used to determine differences between two groups.

Abdominal aorta stenting in rats with renal ischemia-reperfusion

Two rat models, namely bilateral renal ischemia-reperfusion injury and unilateral hindlimb ischemia, were employed to examine the biocompatibility and therapeutic benefits of EES (Fig. 4a). Briefly, the abdominal aorta³⁵ was separated from surrounding tissue and placed with either a commercially-available BMS or an EES (made from the same BMS). Three days after stenting, a sub-group of rats were euthanized. Major organs, including the heart, liver, spleen, lungs, and kidneys were excised for ex vivo imaging using the IVIS system (Fig. S7). The biodistributions of MSC-XOs in the spleen and the kidneys were significantly higher in rats with renal ischemia-reperfusion injury compared to sham-operated rats. Seven days after stenting, aortas with BMS or EES placement were harvested for SEM imaging analysis (Fig. 4b). The BMS was full of irregular adhesions, while EES exhibited a relatively smooth surface. Furthermore, thrombosis-related PCR array in the local blood vessel tissues 7-days after stenting (Fig. 4c) revealed that EES stenting reduced the expressions of Angpt2 (angiopoietin-2), Col1a1 (collagen, type I, alpha 1), Hif1a (Hypoxia Inducible Factor 1 Subunit Alpha), IL-6 (Interleukin 6), Mmp2 (Matrix Metallopeptidase 2), Pdgfra and Pdgfrb (platelet-derived growth factor receptor), while enhancing the expressions of Angpt1 (Angiopoietin 1) and Nfe2l2 (Nuclear Factor, Erythroid 2 Like 2), as compared to BMS stenting. Angpt1 positively regulates vascular remodeling³⁶. The reduced expressions of IL-6 and Mmp2 suggested that EES treatment may have ameliorated inflammation and fibrosis in the local environment. Nfe2l2 drives the expression of cytoprotective genes in response to oxygen stress³⁷.

Fig. 4 |. Stenting in the abdominal aorta of rats.

a, Schematic illustrating stent placement in the rat abdominal aorta. To demonstrate the therapeutic effects of EES in ischemia, hindlimb ischemia or renal ischemia-reperfusion injury was induced right before placing stents. b, SEM images of aorta with BMS or EES deployed on day 7. n=3. c, PCR array revealing thrombosis-related gene expression in the stented vessels from sham, BMS-, or EES-stented animals. Values in EES and BMS groups were normalized to the values from the sham group. n=3 animals. P values <0.05 are considered significant and shown in the graph. d, Representative elastin trichrome (ET) and hematoxylin and eosin (H&E) staining of stented aortas 28 days after stent deployment. The samples were evaluated via morphometric analysis and semi-quantitative histopathologic evaluation. Lumen area (inner area), IEL (internal elastic lamina), and EEL (external elastic lamina) were outlined by yellow lines on ET images. Areas of vessel wall injury in media and adventitia outlined by green lines characterized by loss of black elastic fiber staining and increased connective tissue (blue staining) within media and adventitia. Quadrant mural inflammation was analyzed and scored by the infiltration of inflammatory cells in neointima, media and strut-centered area respectively from H&E staining. The asterisks indicate struts that were fully covered. Scale bar, 1 mm. Quantification of (e) neointimal area, (f) average neointima thickness, (g) percent area stenosis, (h) vessel wall injury score, (i) inflammation score and (j) strut coverage in BMS, DES and EES groups. n=5. P values are shown on the graphs. Comparison of more than two groups were performed using one-way analysis of variance (ANOVA) followed by Tukey’s multiple comparisons. Grouped data were analyzed by using two-way ANOVA followed by Tukey’s multiple comparisons.

For clinical relevance, we added a DES control group. The aortas were collected and sectioned on day 28, and then stained with hematoxylin & eosin (H&E) or elastin trichrome (ET). The histopathologic measurements are summarized in Table 1. There were decreased lumen area and increased neointimal area in the BMS-treated animals, suggesting in-stent stenosis. The BMS group had a significantly thicker neointima and severe lumen restenosis on day 28 as compared to the other two groups. The EES and DES groups had similar neointimal areas and lumen areas. However, EES outperformed DES in strut coverage. DES produced a dramatic reduction in stenosis. However, due to the negative effects of the antiproliferation drugs on DES³⁸, strut coverage in DES (red asterisks, Fig. 4d and Fig. S8) was remarkably low. It has been reported that incomplete strut coverage was considered as a driver of late in-stent restenosis in DES-implanted patients^{39, 40}. Neointimal area, average neointima thickness, percent area stenosis, vessel wall injury score, inflammation score, and strut coverage of BMS, DES, and EES groups were quantified and displayed in Figs. 4e–j.

Table 1 |.

Histopathologic measurements on Day 28.

Group	BMS		DES		EES
	Mean	St. Dev.	Mean	St. Dev.	Mean	St.Dev.
EEL (mm2)	5.14	0.0801	5.19	0.0763	5.24	0.0549
IEL (mm2)	4.84	0.170	4.84	0.0892	4.91	0.125
Lumen area (mm2)	2.61	1.14	4.61	0.244	4.53	0.196
Neointimal area (mm2)	2.22	1.18	0.232	0.171	0.377	0.209
Average Neointimal thickness (μm)	351	208	30.5	22.7	49.1	27.4
Stenosis (%)	29.3	12.1	4.50	3.30	6.96	3.58
Injury score	0.952	0.449	1.07	0.708	0.200	0.291
Inflammation score	0.950	0.291	0.450	0.187	0.500	0.158

EEL, external elastic lamina; IEL, internal elastic lamina.

The reparative effects of EES stenting were evaluated in rats with renal ischemia-reperfusion injury (Fig. S9). H&E staining of the outer medulla of each kidney revealed that kidney tubule necrosis and inflammatory cells were significantly reduced by EES treatment (Fig. S9a). Severe tubular necrosis, including cell necrosis and tubular dilation were observed in the BMS-treated group, while EES significantly ameliorated tubular necrosis (Fig. S9b). Trichrome staining indicated severe fibrosis after ischemia-reperfusion injury. EES treatment significantly ameliorates the fibrosis process (Fig. S9c). Furthermore, EES promoted endogenous repair as indicated by a higher percentage of proliferative (Ki67-positive) cells (Fig. S9d). Moreover, TUNEL (Terminal deoxynucleotidyl transferase dUTP nick end labeling) staining revealed that EES treatment significantly decreased kidney tubules apoptosis (Fig. S9e). In addition, kidney functions of EES-treated animals were improved remarkably as compared to the ones treated with BMS (Figs. S9f&g).

Abdominal aorta stenting in ApoE^−/− atherosclerotic rats

ApoE knockout (ApoE^−/−) rodents have been widely used as an atherosclerosis model and have provided valuable insights into the mechanisms underlying this disease⁴¹. Consistent with published literature⁴², ApoE^−/− rats exhibited increased lipid cores (Fig. S10) and lesions on the aortic vessel walls (Fig. S11). To better understand the vascular healing process, we studied the effects of different types of stents on the proliferation/migration of SMCs and ECs during neointimal formation in ApoE^−/− rats. As shown in Fig. 5a&b, the expression of α-SMA was uniform in both the media and the intima of the EES group. However, the expression of α-SMA in the BMS group and the DES group were irregular, with increased thickness of the vessel wall. SMCs with a reduced expression of α-SMA hold a higher potential of proliferation and migration⁴³. Glucose transporter-1 (GluT1) is a marker of proliferating immature endothelial cells⁴⁴. We used GluT1 to evaluate intimal neovascularization. As shown in Fig. 5c–e, intimal neovascularization was reduced in the EES group and the distribution of CluT1 positive (GluT1⁺) cells was closer to the edge of the intima. The incidence of intimal neovascularization (small vessels in the intimal area and around the strut area), highly related to late stent complications⁴⁵, was reduced in the EES-treated group. We used CD31 to stain functional endothelial cells (Fig. 5f&g). There was no obvious difference among groups on day 7 after stent deployment. Neointimal microvessels can lead to fragile premature vessels, which are unstable and rupture-prone⁴⁶. Our results indicate a higher expression of GluT1 and relatively low expression of CD31 in DES-treated vessels as compared to EES-treated ones. It is possible that antiproliferation drugs eluted from DES inhibited the growth of both smooth muscle cells and endothelial cells but could not prevent the formation of intimal microvessels. We speculate that EES released MSC-XOs to promote the maturation and function of endothelial cells, favoring the formation of a healthy neointimal layer.

Fig. 5 |. Neointimal formation with different stents.

a, Representative confocal images showing α-SMA expression around struts. n=6. Ten technical replicates per animal, the average value per animal was used. b, Quantification of α-SMA expression in the intima compared to the media. c, GluT1 expressions in the intimal area. n=6. Ten technical replicates. d, Quantification of the density of intimal neovascularization based on GluT1 expression. e, Quantification of the relative intimal GluT1 expression. f, CD31 staining showing endothelial cells. g, Quantification of the relative intimal CD31 expression. Scale bar, 100 μm. n=5. Ten technical replicates. P values are shown on the graphs. Comparison of more than two groups were performed using one-way analysis of variance (ANOVA) followed by Tukey’s multiple comparisons.

MSC-XOs play a vital role in macrophage polarization to blunt inflammation and favor wound healing⁴⁷ in atherosclerosis^{48, 49}. On day 7, stented abdominal aortas were harvested. Both dihydroethidium (DHE) staining (Fig. 6a) and ROS assay (Fig. 6b) showed an overall enhanced ROS level of aortas after stent deployment. EES reduced the ROS level as compared to the BMS and DES groups. To reveal the inflammation regulatory effects of EES, we performed macrophage polarization-related PCR array (Figs. 6c&d). The upregulation of FABP4 gene⁵⁰ in the BMS group favors atherosclerosis. Compared to DES, EES treatment further reduced the expressions of inflammatory mediator chemokines such as (C-C motif) ligand 2 (CCL2), IL-1β, and IL1R1, which are reportedly associated with inflammation and adverse vascular remodeling⁵¹. Compared to both BMS and DES groups, EES treatment led to a higher expression in M2 macrophage markers, MRC-1 and CD163, and a higher expression of anti-inflammatory cytokine, IL10. EES also favors the M2 polarization by showing a higher expression of CD163 and MRC1. Immunohistochemistry (Figs. 6e&f) confirmed a decrease in total macrophage numbers and an increased ratio of M2 macrophages (CD206⁺/CD68⁺ ratio) in the EES-treated group.

Fig. 6 |. Local inflammation- and immuno-modulation effects of stent implantation.

a, Dihydroethidium (DHE) staining of aortas and b, Quantification of ROS levels in aortas of ApoE^−/− rats without stenting, with BMS, DES or EES were measured using the ROS/RNS assay kit. Scale bar, 100 μm. n=5. Ten technical replicates per animal. c, Heat map of PCR array of stented aortas. d, Corresponding quantification of the expression of genes in the PCR array. n=4. e, Representative CD68 (green) and CD206 (red) double staining of aortas from different groups. Scale bar, 100 μm. f, Corresponding quantification of the numbers of CD68 and CD206 positive cells (left Y-axis), and the ratio of them (right Y-axis). n=5 animals. Ten technical replicates per animal. P values are shown on the graphs. Comparison of more than two groups were performed using one-way analysis of variance (ANOVA) followed by Tukey’s multiple comparisons.

Treatment effects of EES in ApoE^−/− rats with hindlimb ischemia

To demonstrate the therapeutic benefits of EES on peripheral arterial disease, BMS, DES, and EES were deployed in ApoE^−/− rats with unilateral hindlimb ischemia injury. Laser Doppler perfusion imaging demonstrated the therapeutic efficacy of EES in restoring blood flow (Figs. 7a&b). H&E staining showed that muscle bundles were severely impaired in both BMS and DES groups, while EES treatment led to a preservation/repair of healthy muscle morphology (Fig. 7c). We performed MHC II and Dystrophin staining. MHC-II is mainly expressed by CD4⁺ T cells⁵². As shown in Figs. 7d–f, EES favored a quick immune response and myofiber repair process. The numbers of Dystrophin-positive fibers in the EES group were significantly higher than those from the BMS and DES groups. Measurement of fiber cross-sectional area indicated a decrease in myofiber sizes on the ischemic sides compared to nonischemic limbs in all groups. EES treatment led to intact and regular morphology as compared to the other two stent groups, indicating either a protective or reparative effect from EES treatment. The density of CD31 positive capillaries in the EES group was significantly higher than those from the BMS and DES groups on day 7, suggesting a pro-angiogenesis role of EES (Figs. 7g–j). In addition, EES treatment increased the number of Ki67 positive endothelial cells ((yellow, Fig. 7h). As shown in Fig. S12, the number of CD68 positive cells remained at a high level in both BMS and DES groups. In contrast, infiltration of CD68 positive cells in the EES group dropped to a normal level on day 28.

Fig. 7 |. Restoration of blood flow and muscle repair in the ischemic limbs of ApoE^−/− rats after EES treatment.

a, Representative laser Doppler perfusion images taken on various time points after ischemia and stenting procedure. (left: nonischemic leg; right: ischemic leg). Three technical replicates. b, Quantitative analysis of hindlimb blood perfusion as indicated by ischemic/nonischemic ratio. n=6. c, Representative H&E staining of nonischemic leg and ischemic leg of each group. Scale bar, 100 μm. n=6. Ten technical replicates per animal. d, Representative Dystrophin (green) and MHC-II (red) double staining images of legs from different groups showing the reconstruction and inflammation of each leg sample. Scale bar, 50 μm. n=6. Ten technical replicates per animal. e-f, Quantitative analysis of MHC-II+ cells and mean cross-sectional fiber area according to the morphology of Dystrophin. g, Immunohistochemistry of CD31 expression. Scale bar, 100 μm. n=6. Ten technical replicates per animal. h, Representative immunofluorescent images showing CD31 (green) and Ki67 positive cells (red). They overlay (yellow) of CD31 and Ki67 means proliferating endothelial cells. Scale bar, 100 μm. i-j, Quantitative analysis of CD31⁺ and Ki67⁺ cells. n=6. Ten technical replicates per animal. P values are shown on the graphs. Comparison of more than two groups were performed using one-way analysis of variance (ANOVA) followed by Tukey’s multiple comparisons.

Discussion

Despite its wide use in vascular medicine, BMS can cause mechanical injury to the blood vessel⁵, followed by local inflammatory response that further stimulates the migration of vascular SMCs and impedes endothelialization⁵³. As an improvement to BMS, DES were later on developed to release anti-proliferative drugs and reduce incidents of restenosis. However, some early generations of DES were associated with higher rates of stent thrombosis than BMS, particularly beyond the first few months after implantation. Furthermore, neither BMS nor DES carries biologically active substances to promote tissue regeneration.

Naturally derived exosomes are promising therapeutic agents due to their safety, biocompatibility, and stability^{54, 55}. Here, we chose MSC-XOs as our regenerative cargo to coat the stents due to the reported role in tissue repair and excellent safety profile. The biodegradable linker we used to link the therapeutic exosomes to the stent, benzeneboronic acid pinacol ester group, is highly sensitive to ROS³². EES with arylboronic ester derivatives were designed to be sensitive to elevated reactive oxygen species (ROS, 50 μM~100 μM), which is induced by both local inflammation⁵⁶ after stenting and atherosclerosis²¹. We estimated the numbers of exosomes coated onto stents. Assuming the diameter of exosomes is 100 nm, the cross-sectional area of exosomes would be:

$${\text{A}}_{\text{exo}}=\pi r^2=\pi ·{(50\times {10}^{-6}mm)}^2=7.85\times {10}^{-9}{mm}^2$$

The surface area of the stents varies depending on the length of the stents. Taking Boston Scientific Rebel PtCr OTW coronary stent 2.25 mm× 12 mm for example, the folded average stent profile is 1.07 mm in diameter, the surface (internal and external) would be:

$$A_{\mathit{\text{stent}}}=2\times \pi \text{d}\times l=2\pi ·1.07mm·12mm\approx 80{mm}^2$$

The number of monolayer exosomes can be coated:

$$\text{n}=\raisebox{1ex}{$A_{\mathit{\text{stent}}}$}\!\left/ \!\raisebox{-1ex}{${\text{A}}_{\text{exo}}$}\right.=\raisebox{1ex}{$80$}\!\left/ \!\raisebox{-1ex}{$7.85\times {10}^{-9}$}\right.\approx {10}^{10}$$

The stents’ struts are close to a cylinder, so the whole contact area of the deployed stent with exosomes would be larger. There also can be multiple layers of exosomes linked to the stent. Therefore, the estimated loading number of exosomes is around 10¹⁰~10¹¹ per stent. We also quantified MSC-XOs per stent by counting the number of MSC-XOs before and after EES fabrication, and we found that 1.0~1.5×10¹⁰ MSC-XOs were coated onto the stent. In one of the previous studies¹², the dose of intravenous MSC-XOs was 1×10¹²/kg mice. However, most exosomes were rapidly taken up by macrophages in the reticuloendothelial system, which greatly limits the application of systemic administration of exosomes. According to Gallet et al.⁵⁷, the dose of exosomes by intracoronary or open-chest intramyocardial delivery was 4.1×10¹⁰ and 2.1×10¹⁰/kg, respectively in Yucatan pigs. Our EES system could deliver 10¹⁰~10¹¹ exosomes per stent through a minimally invasive approach, which is enough to demonstrate therapeutic effects via local delivery.

We studied the release profile of MSC-XOs from EES. Within 48 hours, around 40%~60% of exosomes could be released upon ROS stimulation in vitro as compared to a 20% release under normal physical environment. This demonstrates that EES can quickly respond to elevated ROS levels to release exosomes. In vivo, the release of exosomes into the circulation could be shorter as the struts of the stent are gradually covered by blood cells, vascular cells, and extracellular matrix deposited by the cells.

Coating with a biocompatible material is reportedly to improve the safety of medical devices. Human origin MSC-XOs were widely used in small animal studies due to their hypo-immunogenicity⁵⁸. Unlike stem cells, exosomes could be sterile filtered, and they would not cause abnormal tissue growth due to their nonviable property⁵⁹. In addition, the phospholipid bilayer of exosomes provided a superior alternative to synthetic polymer coating due to its similar composition as compared to cell membranes⁶⁰. We confirmed reduced adhesion of platelets and monocytes on EES as compared to BMS (Figs. 2f–h).

EES promoted the proliferation of ECs and their migration to the stents. Reportedly those are also observed as a benefit of MSC-XO treatment⁶¹. The mechanisms underlying such effects remain elusive, but they are likely due to the microRNA cargos in MSC-XOs (Fig. S2). For example, miR-23a-3p⁶² and let-7b-5p⁶³ in MSC-XOs target genes related to angiogenesis and regulate vascular repair. Besides miRNAs, abundant extracellular-associated proteins, like fibronectin and collagen α1 were identified in MSC-XOs (Fig. S2). Those exosomal proteins also play critical roles in angiogenesis⁶⁴. In addition, we investigated the effects of EES on the proliferation and migration of SMCs. EES favors the modulation of SMCs in a contractile phenotype with a higher expression of α-SMA as compared to BMS (Fig. 3k). This was confirmed in vivo in the ApoE^−/− rat model of atherosclerosis (Fig. 5a). SMCs are the primary cell type in the pre-atherosclerotic intima. The phenotypic modulation of intimal SMCs occurs in response to environmental change⁶⁵. The proliferation and migration of SMCs in the intima require the transition of SMCs from a contractile to a synthetic phenotype²². EES didn’t slow down the proliferation of SMCs as this is an important step for early-stage stent coverage (Fig. 3m).

Anti-inflammation and pro-angiogenesis mechanisms underlie MSC-XO-mediated tissue repair. The inflammation and the vascular remodeling after stenting are highly related to local monocyte’s behaviors. MSC-XOs were reportedly to have the ability to modulate macrophage polarization and inhibit inflammation in the tissue remodeling process. It has been well established that M1 macrophages mediated the secretion of pro-inflammatory factors and vascular smooth muscle cell migration, while M2 macrophages encourage wound healing and re-endothelization⁶⁶. Mounting lines of evidence correlate enhanced level of anti-inflammatory M2 macrophages leads with less in-stent restenosis⁶⁷. In addition, M2 macrophage-derived exosomes⁶⁸ were studied for their effects on SMC dedifferentiation and vascular repair process. Immunostaining and PCR array (Fig. 6) showed reduced inflammation and M2 macrophage polarization with EES treatment, suggesting a positive role of EES in vascular healing and remodeling^{23, 49}. Further, the effects of MSC-XOs on major organs were investigated. As shown in Fig. S13, lipid deposition in the liver was evident in all groups. Histology of the heart and liver was found to be normal. ApoE^−/− rats displayed mast cell infiltration in the lung, mesangial proliferative glomerulonephritis, and splenomegaly. We found a decrease in inflammation in the spleen and the glomeruli of the EES group. The spleen is an important organ for atherosclerosis-associated immunity⁶⁹, and the change of inflammatory cells in the spleen with MSC-XOs (released from EES) deserves further investigation.

As summarized in Supplementary Table 1, mounting lines of evidence have confirmed the advantages of DES over BMS⁷⁰. However, late in-stent restenosis rate of current DES remains high (~3%−20%) after stent implantation⁷¹. Long-term follow up studies of the performance of DES are likely needed to understand the safety profile thoroughly. In this study, we avoided the use of synthetic polymers and anti-proliferative drugs in DES, but used MSC-XOs as a biocompatible and bioactive coating. Given MSC-XOs have shown promise in attenuating atherosclerosis^{47, 48} and promoting re-endothelialization⁷², they are ideal coating materials for vascular stents. In addition, MSC-XOs has been tested in a wide range of ischemia diseases, such as myocardial infarction^{12, 73} and peripheral and renal ischemic injury^{11, 13, 64}.

In this study, the EES group showed a smaller neointimal area compared to the BMS group, and a higher strut coverage rate compared to the DES group 28 days after deployment. The histological analysis proved the functional advantage of EES. Additionally, the regenerative effects of EES on ischemic tissue were demonstrated in rats with renal ischemic or hindlimb ischemic injury. Our study offers translational values. Given the excellent stability of exosomes, we foresee EES can be an off-the-shelf product to be applied in acute settings. As a “reopen and regenerate” strategy, this biological stent holds the potential to not only mechanically keep the vessel patent but also to repair the injured tissue, which is something not accomplished by current stent products. Our study also has several limitations. First, our animal studies are relatively short. Long-term follow up studies of EES and DES are likely needed to fully understand the safety and efficacy of EES. Secondly, small animal models cannot precisely reflect the real clinical application scenarios for coronary stents. Large animal studies in porcine models are needed for future translation. Thirdly, storage stability studies would be helpful to evaluate under what storage conditions the exosomes on EES remain intact and functional.

In summary, we have designed, fabricated, and tested a biocompatible and bioactive stent that could release therapeutic exosomes under ROS. EES is free of polymer coating and anti-proliferative drugs (used in DES) and decreases the risks of thrombosis and inflammation⁷⁴. EES accelerates the vascular healing process via promoting endothelial proliferation and early-stage cell coverage, while reducing inflammation and SMC migration. In addition, EES treatment promotes tissue repair in renal ischemia and hind limb ischemia models via pro-angiogenesis mechanisms.

Methods

Generation of MSC exosomes

Human MSCs were obtained from ATCC (PCS-500–012). Authentication and validation were performed to confirm the identity of the cells by flow cytometry characterization of common MSC surface markers⁷⁵. Exosomes were collected as previously described¹². In brief, MSCs were cultured until 80% confluence and washed by serum-free medium three times. Then, the cells were incubated with serum-free medium for another three days to allow exosome secretion. The conditioned media were then collected and filtered through 0.22 μm to remove residual cells and debris. We then performed ultracentrifugation at 100,000 g for 2 hours to collect MSC-XOs (Beckman Coulter XL90 ultracentrifuge). In several experiments, fluorescently labeled exosomes were used. Purified MSC-XOs were mixed with 1 μM DiD or DiR (Invitrogen, Life Technologies) and incubated for 30 min at 4°C, then free dye was removed through centrifugal filters (10 KDa). MSC-XOs were washed three times with PBS.

Characterization of exosomes

The concentration of exosomes was examined with a NanoSight LM10 (Malvern Instruments Ltd., UK). The morphology of exosomes was visualized using a transmission electron microscope (TEM, JEOL JEM-2000FX). RNA sequencing and proteomics of MSC-XOs were performed as previously described⁷⁶. Briefly, exosomal RNA was isolated using a total exosome RNA isolation kit (Qiagen’s exoRNeasy Serum Plasma Kit). Libraries were quantified by a Quant-iT^™ dsDNA High Sensitivity Assay Kit (ThermoFisher) and sequenced on an Illumina NextSeq500 using a mid-output V2 kit. LC/MS/MS analysis of the exosome samples was performed on an Easy Nano ultra-high-pressure liquid chromatography coupled to a Q Exactive HF-X Hybrid Quadrupole-Orbitrap mass spectrometer (ThermoFisher Scientific).

Fabrication of EES

All materials were purchased from Sigma Aldrich. All reagents were used as received. Stents were purchased from eSutures. Medtronic Integrity RX Coronary Stent System 2.5 mm ×12 mm and Boston Scientific Rebel PtCr Monorail Coronary Stent was used in the modification study. EES made from that were used in animal studies for a head-to-head comparison. Medtronic Resolute Integrity RX Zotarolimus-Eluting Coronary Stent System (2.5 mm × 12 mm) and Boston Scientific Ion Monorail Paclitaxel-Eluting Platinum Chromium Coronary Stent System was used as the DES control in our studies. To fabricate EES, stents were first expanded and removed at a pressure of 4 atm by an angioplasty balloon. Then, stents were ultrasonically cleaned with acetone, ethanol, and deionized water. After drying, the stents were placed in mixed acid solution (1:1 (v/v) nitric acid- hydrofluoric acid) for 30 min at room temperature (RT). Following acid treatment, the stents were washed with deionized water three times and placed in 10 N NaOH at 80°C for 30 min. Stents were rinsed with deionized water another three times. The hydroxylated stents were then immersed in a 5% (v/v) solution of (3 aminopropyl) triethoxysilane (APTES) in ethanol overnight and kept thermostatic at 60°C under nitrogen flow. Following rinsing with deionized water and blow-dried under a stream of nitrogen, the modified stents were stored in individual containers for further usage. 4-Carboxyphenylboronic acid was activated using 1-Ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) and N-hydroxysuccinimide (NHS) and reacted with amino groups on the stent. 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[amino(polyethylene glycol)-5000]-N- hydroxysuccinimide (DSPE-PEG₅₀₀₀-NHS) was reacted with 3-Amino-1,2-propanediol with a mole ratio of 1:3 to provide dihydroxyl groups. The obtained products were dialyzed against DI water (MWCO: 3 KDa) and lyophilized. Finally, the dihydroxyl-modified DSPE was added to the stents overnight to generate DSPE-conjugated stents. DSPE-modified stent was incubated with 10¹² exosomes in 4 °C overnight to fabricate the final product EES.

Physiochemical Characterization of EES

Time-of-flight secondary ion mass spectrometry (ToF-SIMS) analyses were conducted using a TOF SIMS V (ION TOF, Inc. Chestnut Ridge, NY) instrument equipped with a Bi_n^m+ (n=1–5, m=1, 2) liquid metal ion gun, Cs⁺ sputtering gun, and electron flood gun for charge compensation. X-ray Photoelectron Spectroscopy (XPS) analysis was conducted by SPECS XPS/UVS System with PHOIBOS 150 Analyzer (SPECS Surface Nano Analysis GmbH, Berlin, Germany). XPS data analysis was performed with the curve fitting program (CasaXPS, Version 2.3.17PR1.1). Morphology of stents was visualized via scanning electron microscopy (SEM) imaging. Samples were fixed with 2% glutaraldehyde, and then dehydrated in gradient ethanol successively for 10 min each, at last, dried in hexamethyldisilazane (Sigma-Aldrich) for imaging (JEOL 6010LA SEM, JEOL ltd, Japan).

Cell lines

Human primary umbilical vein endothelial cells (HUVEC, PCS-100–013), human primary aortic smooth muscle cells (SMC, PCS-100–012), primary coronary artery endothelial cells (HCAEC, PCS-100–020), and human bone marrow-derived mesenchymal stem cells (BMMSC, PCS-500–012) were obtained from ATCC. U937 human monocytes were purchased from Millipore Sigma.

Biological characterization of EES

For exosome release studies, one EES stent was incubated with 1 mL PBS in 37 °C, and the solution was taken out at different time points for NanoSight quantification of exosome concentrations. For tube formation assay, HUVECs were seeded onto Matrigel (BD Biosciences)-coated plate wells and a piece of BMS or EES was placed gently on the cells. Images were taken by an epi-fluorescent microscope (Olympus, Tokyo, Japan). Tube nodes were quantified by ImageJ (NIH). For in vitro endothelialization assay, HUVECs were seeded on glass slides and allowed to be confluent. A piece of BMS or EES was placed gently on top of the HUVEC monolayer. Four days later, stents were gently removed, washed with PBS, and fixed with glutaraldehyde for SEM imaging.

Cell proliferation assay

Cells were seeded in 96 well plates overnight and incubated with a piece of BMS or EES for additional two days. Then, BMS or EES were removed gently, and Cell Counting Kit-8 (CCK-8, Sigma) reagent was added. Cell viability was determined by measuring absorbance wavelength of 450 nm and reference wavelength of 630 nm with a VersaMax^™ Microplate Reader.

Smooth muscle cell migration assay

HUVECs were seeded in 24-well plates. A piece of BMS or EES was gently placed on HUVECs. Trans-well chambers with an 8-μm pore size were placed. SMCs were seeded on the trans-well chambers. The plates were incubated for 6 h then the chambers were removed and fixed with 4% formaldehyde. The non-migrated SMCs were removed with cotton swap gently. The migrated SMCs on the backside of the filter were stained with 0.05% crystal violet and viewed via a microscope (Olympus, Tokyo, Japan).

In vitro thrombosis and adhesion assay

Fresh platelet-rich plasma (PRP) was collected from whole blood of healthy Sprague Dawley rats (male, 8 weeks old). BMS or EES were co-cultured with PRP for 30 min at 37°C. Then, BMS or EES were washed with PBS and fixed in 2.5% Glutaraldehyde for SEM imaging. FITC-CD42b (eBioscience) staining was used to confirm the platelets on the stent’s surface. BMS or EES were incubated with U937 monocytes (1×10⁶ cells/mL) and TNF-α (10 ng/mL) under rotation for 4 hours to stimulate in vivo conditions. After that the stents were washed with PBS and fixed in 2.5% Glutaraldehyde for further SEM imaging. FITC-CD11c (eBioscience) staining was performed to confirm the monocytes on stent surface.

Antibodies

Primary antibodies used in this study: Anti-Von Willebrand Factor antibody (ab6994) (vWF, ABCAM), Anti-alpha smooth muscle Actin antibody (ab7817) (α-SMA, ABCAM), Anti-Ki67 antibody (ab15580) (ABCAM), Rhodamine labeled Lens Culinaris Agglutinin (LCA, VECTOR LABORATORIES, INC.), CD81 antibody (sc-166029, Santa Cruz Biotechnology), Anti-ALIX antibody (ab186429, ABCAM), TSG101 antibody (NB200–112, NOVUS Biologicals), Anti-CD68 antibody (ab955) (ABCAM), Anti-Dystrophin (ab15277, ABCAM), Anti-Glucose Transporter GLUT1 (ab40084, ABCAM), Anti-Mannose Receptor (ab64693, ABCAM), Anti-MHC Class II (ab23990, ABCAM), Anti-CD31 (ab119339, ABCAM) Anti-VE Cadherin (ab7047, ABCAM). Secondary antibodies used in this study: Goat Anti-Rabbit IgG H&L (Alexa Fluor® 488) (ab150077), Goat Anti-Mouse IgG H&L (Alexa Fluor® 488) (ab150113), Goat Anti-Mouse IgG H&L (Alexa Fluor® 594) (ab150116).

Immunocytochemistry

Cells were rinsed once with PBS and fixed in 4% paraformaldehyde (PFA) (Electron Microscopy Sciences; 15710) for 20 min, followed by permeabilization and blocking with Dako Protein blocking solution (DAKO; X0909) containing 0.1% saponin (Sigma-Aldrich; 47036) for 1 h at room temperature to prevent non-specific binding. Cells were incubated in primary antibodies overnight at 4 °C and then in secondary antibodies for 1.5 h at room temperature. ProLong^™ Gold Antifade Mountant with DAPI (ThermoFisher) was used to counter-stain nuclei and slow the fade of the fluorophore. All antibodies and kits were used according to instructions from the vendors. Cells were examined with a confocal microscope (Zeiss LSM 710).

Sprague Dawley rat studies

All experimental protocols were approved by the Institutional Animal Care and Use Committee at North Carolina State University. Sprague Dawley (SD) rats were purchased from Charles River Laboratories. The use of the rat abdominal artery to test ISR due to stenting has been supported by the literature for decades³⁵. SD rats (300–500 g, male) were anesthetized with 2% isoflurane in oxygen. After median laparotomy, the abdominal aorta was isolated to give enough length for stents deployment (3 cm). Then, two vascular clips were placed to create an aorta segment. A small incision was created at one end. This segment was then flushed with heparin carefully. A balloon catheter was inserted through the small incision and inflated to deploy a stent into the aortic segment. After that, the balloon was removed, and the small incision was closed with a 9–0 suture. Finally, the vascular clips were removed. Intravenous heparin (100 U/kg) was administered after stent deployment. The abdomen was closed with 4–0 sutures, and the animals were allowed for recovery. Healthy SD rats were randomized into three groups (n=6): Sham, BMS, EES. Stents were collected on day 7 and used for PCR arrays (Figure 4b, n=3), SEM (Figure 4c, n=3), or immunofluorescent staining (Figure 5, n=6).

Renal ischemia-reperfusion injury model

Bi-lateral renal ischemia-reperfusion (RIR) injury model was performed as previously described with slight modifications⁷⁷. Briefly, bilateral renal pedicle occlusions were induced by vascular clamping of 45 min. Ischemia was confirmed by the color change of the kidney. With the removal of the clamp, a color change in the kidney indicated proper reperfusion. Stenting was performed during the ischemia surgery. The animals were randomized into four groups (n=5): Sham, RIR, RIR + BMS stenting, and RIR + EES. On days 1, 7, and 28, blood samples (1 mL) were collected to test kidney functional biomarkers BUN (blood urea nitrogen, Urea Nitrogen Colorimetric Detection Kit, EIABUN, ThermoFisher) and Cre (Creatinine Assay Kit (ab65340), AMCAM). After euthanasia, kidneys were harvested and bisected longitudinally. Kidney tubular necrosis, inflammatory cells, and vacuolar degeneration were analyzed with H&E staining. Fibrosis was assessed by Trichrome staining. In Situ Cell Death Detection Kit, Fluorescein (TUNEL, Sigma) was used to detect apoptotic cells. BMS and EES were collected from euthanized animals on day 28 after treatment for histological analysis (HE and ET staining). A group of animals (RIR injury model) were added for DES stenting and histology (n=5, figure 4d–j).

ApoE^−/− rat studies and hind limb ischemia model

ApoE^−/− rats (8-week, male) were purchased from Horizon Discovery and fed with a high-fat diet during the entire study. Hindlimb ischemia was induced on the right limb before stent deployment. Briefly, rats were anesthetized with 2% isoflurane in oxygen and the surgical area was shaved. The targeted vessels were visualized by creating an incision overlying the proximal, medial portion of the right hindlimb. The femoral artery and vein were dissected away from the femoral nerve and ligated to target the proximal removal of a 1 cm segment. The entry incision through the skin was then closed with a suture (5–0 silk). Blood flow in both hindlimbs was assessed using a Laser Doppler perfusion imager (PeriCam PSI NR). The rats were randomly divided into three groups (n=12): BMS, DES, and EES to receive stent deployment in their abdominal aortas as previously described. A subgroup of rats (n=6) was sacrificed 7 days later. The aortas were cut into two parts: one for PCR arrays (Figure 6c&d, n=4) and the other for immunofluorescent staining (Figure 6a–b, 6e–f, n=5). All other animals were followed until day 28.

Histopathological analysis

The harvested stented aortas were washed with PBS three times and fixed with 4% PFA. After that, the aortas were embedded in methylmethacrylate and sent for sectioning. Sections with 5-μm thickness were cut from the proximal and middle parts of the stented aorta and mounted on 3-Aminopropyltriethoxysilane (APES) coated slides. Afterwards, the slides were de-plastified and stained with Hematoxylin & Eosin (H&E) and Elastin Trichrome (ET). Immunostaining was performed on tissues (organs, limbs, or stented aortas). Tissue sections were fixed with 4% PFA for 20 min and then blocked with Dako Protein Blocking solution containing 0.1% saponin for 1 h. Then, slides were incubated with primary antibodies overnight at 4 °C and then with secondary antibodies for 1.5 h at room temperature. ProLong^™ Gold Antifade Mountant with DAPI was used. The slides were imaged with a confocal microscope (Zeiss LSM 710). Elastin trichrome staining was used for histomorphometry analysis using Image-Pro Plus 7 software^{78, 79}. All cross-sections were measured, and the following morphometric data were collected: external elastic lamina (EEL) area (mm²), internal elastic lamina (IEL) area (mm²), and lumen area (mm²). From those direct measurements, all other histomorphometric parameters were calculated, such as neointimal area (mm²): IEL area minus lumen area; percent area stenosis based on IEL area (%): 100×neointimal area/IEL area; Average neointima thickness (μm): 1000× (√(IEL area/π)-√(lumen area/π)). Samples were divided into four sections, and quadrant injury was scored. Device biocompatibility was assessed based on H&E staining and quadrant inflammation was scored. Semi-quantitative scoring was as follows: 0, not present; 1, present, but minimal feature; 2, notable feature, mild; 3, prominent feature that does not disrupt tissue architecture, moderate; 4, overwhelming feature, severe.

Dihydroethidium staining

The aortas were harvested, and the stents were carefully removed. Cryosections were prepared for dihydroethidium (DHE) staining⁸⁰. Briefly, slides were rinsed once in pure water to wash out OCT compound, then placed in DHE staining solution (5 μM) immediately. After 20 min incubation at room temperature, the slides were immersed in pure water for washing (1 min, three times). The slides were imaged with a confocal microscope (Zeiss LSM 710).

ROS detection

OxiSelect^™ In Vitro ROS/RNS Assay Kit (Green Fluorescence) was used for ROS detection. 2 mg aorta in 100 μL PBS was homogenized on ice. Then, spin at 10,000 g for 5 min to remove any tissue debris. The homogenate was assayed. The plate was read with a fluorescence plate reader at 480 nm excitation / 530 nm emission (Fluorescence Spectroscopy, Perkin Elmer Model LS-3B).

PCR array

Total RNA was isolated from the stented arteries using the RNeasy mini kit (QIAGEN, Hilden, Germany). Complementary DNA (cDNA) was synthesized using iScript^™ cDNA Synthesis Kit (Bio-Rad). Rat thrombosis-related qPCR array was purchased from Bio-Rad, and the data was analyzed using the Bio-Rad qPCR analysis software. GeneQuery rat macrophage polarization markers qPCR array kit was purchased from ScienCell. SsoAdvanced^™ Universal SYBR® Green Supermix (Bio-Rad) was used to perform the qPCR studies. The plates were read on Roche LightCycler® 480. Data has been submitted to NCBI (GSE155793).

Statistics

All quantitative experiments were done in triplicate unless otherwise indicated. Data were shown as mean± S.D. Biological replicates (n) and technical replicates were indicated in figure captions. GraphPad Prism was used for statistical analysis. Student’s two-tailed independent t-test was used to determine differences between two groups. Comparison of more than two groups were performed using one-way analysis of variance (ANOVA) followed by Tukey’s multiple comparisons. Grouped data were analyzed by using two-way ANOVA followed by Tukey’s multiple comparisons. P<0.05 was considered statistically significant.

Reporting summary.

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability

The main data supporting the results in this study are available within the paper and its Supplementary Information. The raw and analysed datasets generated during the study are too large to be publicly shared, yet they are available for research purposes from the corresponding authors on reasonable request. GeneQuery rat-macrophage-polarization-markers qPCR-array data are available on the NCBI database with the identifier GSE155793. ToF-SIMS, XPS and histopathological data are available on reasonable request.