The effects of stenting on coronary endothelium from a molecular biological view: Time for improvement?

Abstract Coronary artery stenting following balloon angioplasty represents the gold standard in revascularization of coronary artery stenoses. However, stent deployment as well as percutaneous transluminal coronary angioplasty (PTCA) alone causes severe injury of vascular endothelium. The damaged endothelium is intrinsically repaired by locally derived endothelial cells and by circulating endothelial progenitor cells from the blood, leading to re‐population of the denuded regions within several weeks to months. However, the process of re‐endothelialization is often incomplete or dysfunctional, promoting in‐stent thrombosis and restenosis. The molecular and biomechanical mechanisms that influence the process of re‐endothelialization in stented segments are incompletely understood. Once the endothelium is restored, endothelial function might still be impaired. Several strategies have been followed to improve endothelial function after coronary stenting. In this review, the effects of stenting on coronary endothelium are outlined and current and future strategies to improve endothelial function after stent deployment are discussed.

underlying smooth muscle cells, resulting in increased blood flow and inhibition of the coagulation cascade. Nitric oxide also inhibits platelet adhesion and aggregation by inhibiting the thromboxane A2 receptor. 8 In addition to these vasorelaxant and antithrombotic effects, endothelium-derived NO contributes to the inhibitory effect of endothelial cells on VSMCs. An incubation of co-cultures of VSMCs and endothelial cells with L-arginine, the substrate of NOsynthesis, led to a decrease of VSMC growth while this effect was decreased when cells were incubated with the L-arginine analogue nitro-L-arginine, thus blocking the NO synthase. 9 Studies performed in rabbit models of atherosclerosis 10 and of balloon injury 11 evidenced an inhibition of neointima formation after long-term oral administration of L-arginine.
In concert with NO, the significant role of the endothelium in maintaining the balance between vasoconstriction and vasodilatation is carried out by the production and secretion of vasoconstrictive substances such as endothelin, reactive oxygen species, endothelium-derived cyclo-oxygenase-dependent vasoconstricting factor, prostaglandin H 2 , thromboxan A 2, and angiotensin II. 12,13 Thus, the functional vascular endothelium has a striking role in keeping the balance between blood coagulation and fibrinolysis and the vascular tone. 14 Beneath its non-thrombogenic properties that maintain the fluidity of the blood exerted by NO and other molecules such as thrombomodulin, heparin-like molecules, and prostacyclin, prothrombotic endothelial molecules including von Willebrand factor and plasminogen activator inhibitor-1 act as counterparts. 15,16 Endothelial cell-derived Prostaglandin I 2 inhibits platelet activation 17 and therefore acts synergistically to NO that inhibits platelet adhesion and aggregation as mentioned above. 18 Taken together, the initiation and progression of atherosclerosis is characterized by disturbances of normal endothelial functions. 19 The endothelial dysfunction leads to an impaired vascular homeostasis, notably an impaired endothelium-dependent vasodilation and endothelial activation, associated with a pro-inflammatory and procoagulant environment. 20 Particularly the bioavailability of NO is reduced, caused by enhanced degradation of NO and decreased eNOS expression. 21 Besides the resulting impairment of vasodilation the reduced bioavailability of NO leads to apoptosis of endothelial cells, additionally triggered by local inflammatory processes. 19 The endothelial cell turnover is accelerated, however the regenerated cells may be senescent, lack endothelial barrier integrity and may be unable to produce sufficient NO, which results in increased oxidation of LDL and further progression of atherosclerosis. 22 Intravascular processes such as balloon angioplasty and stent implantation inevitably cause severe vascular injury as well. 23 It has been demonstrated that balloon injury induces splitting of the atheromatous plaque and stretching of the vessel wall, notably of media and adventitia, and lysis of some of the cells, mainly VSMCs. These effects have been described both in animal models 1,24 and in human postmortem arteries 3,25 and are accompanied by important alterations in the mechanical environment of the vessel wall. 26 On a macroscopic point of view the angioplasty immediately induces a change in the external size of the artery at the injured site, but the geometry of the vessel segment returns approximately to the 3-D shape it had before the plaque intruded into the lumen. 27 Initially, the effect of elastic recoil occurring seconds to hours after PTCA could be observed in 5%-10% patients undergoing angioplasty. 28 Additionally, local arterial injury provokes an increased local release of vasoconstrictive agents such as serotonin and thromboxane, possibly resulting in vasospasms. 29,30 Angioplasty produces endothelial denudation, resulting in a disturbance of the integrity of structures inside the diseased arterial wall 2,24 and followed by rapid platelet deposition and attraction of leukocytes. 31,32 VSMC proliferation and deposition of extracellular matrix proteins contributes to an intimal thickening, neointimal hyperplasia, and finally the development of a restenosis following PCI. 33 The release of chemotactic factors and mitogens for VSMCs by the platelets, such as platelet-derived growth factor (PDGF), together with the mechanical forces to the VSMCs lead to their activation, resulting in proliferation, migration and shift from a contractile to a synthetic phenotype. 34 Only some 15 years ago, Indolfi et al demonstrated that the extent of balloon injury is directly proportional to VSMCs proliferation. 35 After inserting a Fogarty catheter into the lumen of a rat common carotid artery, different inflation pressures of 0, 0.5, 1.0, 1.5, and 2.0 ATM were applied. Remarkably it was evidenced for the first time that the proliferative response of VSMCs was proportional to the degree of vascular injury, confirmed by histopathologic findings as well as by the extraction of RNA at 30 minutes after balloon injury, demonstrating that an increase of pressure resulted in higher c-fos expression and drives neointimal formation and proliferation.
Another long-term process has been identified as one major determinant of restenosis after balloon angioplasty in humans, known as constrictive vascular remodelling. 36 Arterial remodelling in general represents an adaptive or compensatory response of blood vessels to hemodynamic stress, arterial injury, and cellular proliferation and can either be constrictive or dilative. 37 Constrictive vascular remodelling may be the consequence of vessel constriction due to a retractile scar. Compensatory dilation on the other hand delays the development of focal stenosis in native atherosclerotic arteries despite significant plaque accumulation as the outer vessel diameter increases. 37 In stented segments a compensatory dilation by increase of the outer vessel wall is limited in parts due to the stiffness of the device.
The potential role of the endothelium in vascular remodelling after balloon injury has been discussed. 38 Langille and O′Donnel demonstrated that a structural reduction in vessel size induced by a long-term decrease in blood flow is dependent on an intact endothelium. 39 On the other hand endothelium-derived relaxing factor NO (EDRF-NO) is involved in the adaptive enlargement of the vessel in response to increased blood flow. 40 Measurements of EDRF-NO levels following balloon injury in porcine coronary arteries demonstrated a decreased production of EDRF-NO. 41 As EDRF-NO is a potent inhibitor of VSMC growth, the PCI-induced damage of the endothelium is suggested to influence neointimal hyperplasia as well as the development of restenosis.

ENDOTHELIALIZATION AFTER PTCA/ STENT DEPLOYMENT
Arterial healing after denudation involves regrowth of the endothelium from remaining endothelial cells within the treated segment, from proximal and distal to the lesion as well as from side-branch ostia. 42 Circulating endothelial progenitor cells (EPCs) might also contribute to re-endothelialization. 43 The process begins within the first 24 hours after arterial denudation. 44 A breakpoint of reendothelialization was observed at 6-10 weeks in several animal models. 44 In humans however there is limited information on the time-course of re-endothelialization following PCI. 23 Delayed endothelial recovery has been identified as one of the major contributing factors of late stent thrombosis at autopsy. 45,46 The risk of thrombosis is substantially increased in stents with >30% uncovered struts compared to stents with complete coverage. 46 Even beyond 1 year after implantation uncovered stent struts were identified in first-generation sirolimus-and paclitaxel-eluting stents, especially under high-risk implantation conditions like acute myocardial infarction, bifurcation and ostial lesions, lesions in bypass grafts, lesions of the left main artery, chronic total occlusions (CTO), long lesions (>30 mm), and in-stent restenosis. 47,48 Delayed arterial healing has also been observed in stents penetrating into the necrotic cores of atherosclerotic plaques and overlapping stents. 49,50 The biological factors controlling the re-endothelialization process have not been completely elucidated. Both vascular endothelial growth factor (VEGF) and fibroblast growth factor (FGF) represent growth factors for endothelial cells whereas FGF also has trophic effects on VSMCs. Balloon injury induces a release of FGF and an increased expression of FGF mRNA in endothelial cells and VSMCs. 51 Similarly, an increased expression of VEGF mRNA in rats could be observed. 52 Studies performed by Lam et al in humans undergoing PTCA showed increased levels of circulating FGF, VEGF, and tumor growth factor b1 (TGF-b1), suggesting an operative role of these factors in re-endothelialization in humans. 53

| THE IMPACT OF STENT DESIGN ON ENDOTHELIAL REGROWTH
Today, a broad variety of stents is available. There have been significant developments concerning the design of stent platforms as well as the stent coatings including novel polymers, polymer-free stents and bioresorbable stents. The endothelial recovery after stent implantation is influenced significantly by the stent design. The protrusion of stent struts leads to perturbations in the local flow patterns notably to the development of small regions with disturbed shear stress between the stent struts. 54 Alterations in shear stress and blood flow dynamics are known to impact endothelial growth. 55 In an experimental setting with flow and shear conditions similar to human arteries, the endothelial cell coverage area and migration was found to depend on object thickness and significantly decreased in objects with 75 μm thickness or greater. 56 Relating to coronary stents, improved re-endothelialization was demonstrated in newer generation stents with lower strut thicknesses. 57 In line with that, re-endothelialization was delayed in novel but comparably thick-strut bioabsorbable stents as compared with thin-strut everolimus-eluting stents in a study of Koppara et al who performed stent implantation into iliofemoral arteries in a healthy rabbit model with contemporary stents used in clinical practice. 58 Clinical significance was proven in the randomized multicentre Strut Thickness Effect on Restenosis Outcome" (ISAR-STEREO)-Trial demonstrating a significant reduction of angiographic and clinical restenosis after coronary artery stenting with thinner-strut devices. 59,60 In addition to the strut thickness also the shape of the struts influences re-endothelialization process by changes in local vascular flow conditions. Non-streamlined stent struts promote flow separation in the regions proximal and distal to the stent strut. High shear at the edges of the struts can activate platelets through the release of thromboxane A 2 , whereas areas of low shear rates adjacent to non-streamlined struts are associated with inhibition of re-endothelialization, potentially enabling procoagulant and pro-inflammatory elements to accumulate which contribute to thrombus formation. 54 In contrast, streamlined strut geometry reduces flow separation and high shear peak, resulting in rapid re-endothelialization and inhibition of platelet activation. Furthermore the areas affected by low shear stress are smaller adjacent to streamlined struts, contributing to a rapid re-endothelialization. 54

| TH E IMPACT OF STENTS ON ENDOTHE LIAL FUNCTION
Angioplasty and stent implantation cause denudation of the endothelium. The regeneration and regrowth of the denuded endothelium originates from the remaining endothelial cells and uninjured segments at the stent edges and from side-branch ostia.
However, vessels that need to undergo an interventional revascularization process usually are affected by atherosclerosis. The endothelium of atherosclerotic vessels is dysfunctional a priori, characterized by the impairment of vasomotion, disruption of the hemostatic balance, and a pro-inflammatory milieu. 20 Thus, the regenerated endothelium after stent placement is found to be inadequate in terms of both barrier integrity and functionality with impaired endothelium-dependent vasodilation and increased permeability. 61 Concerning the vasomotor functions the rigid stent prevents relaxation and constriction of the artery. The production and release of bioactive substances in stented vessel segments are affected, resulting in limited ability to respond sufficiently to changes in the physical, chemical, and humoral environment.
Notably both the barrier integrity and the extent of dysfunctionality depend on the presence or absence of antiproliferative stent coverage. The majority of the currently available drug-eluting stents release the mammalian target of rapamycin (mTOR) inhibitor sirolimus or one of its analogues (everolimus, zotarolimus, or biolimus).

ENDOTHELIALIZATION AFTER STE NT IMPLANTATION
Incomplete re-endothelialization has been demonstrated particularly after the implantation of drug-eluting stents due to the unselective antiproliferative effect of the drugs on both VSMCs and endothelial cells. It has been suggested that DES should ideally have a selective antiproliferative effect on VSMCs but be inert towards EC or better still promote their proliferation. The latter might be achieved by targeting factors influencing VSMC and EC growth. In this regard the impact of the administration of growth factors as well as the capture of EPCs have been investigated. In an effort to promote and accelerate the process of re-endothelialization after stent implantation, novel stent designs have been developed, summarized as pro-healing stents.
Van Belle et al demonstrated a promotion of re-endothelialization after stent implantation by local administration of VEGF in a rabbit iliac model. 66 Seven days after stent implantation, a near complete re-endothelialization of the stented vessel was detected whereas the endothelial coverage of the stented arterial segment without VEGF was only 30%. 67 Based on these findings, growth-factor delivering stents have been developed, however they failed to promote re-endothelialization in vivo and increased neointimal proliferation instead. 68 Another strategy to improve re-endothelialization is followed by capturing EPCs from the blood flow. EPC capture has been attempted by coating stents with antibodies that target EPC markers such as anti-CD133, anti-VE-cadherin (anti-CD 144), and anti-CD34. 69 Coating of stents with anti-CD133 antibodies did not influence re-endothelialization or neointimal thickening in a porcine model. 70  Among EPC subsets, outgrowth endothelial cells (OECs) preferentially express VE-cadherin, and exhibit greater vasculogenic activity with more rapid proliferation and more active migration. 72 Therefore capturing OECs more selectively with anti-VE-cadherin antibody could be responsible for a more efficient re-endothelialization and anti-thrombogenicity of anti-VE-cadherin antibody-coated stents in animal models. 73 The Genous ® (OrbusNeich) stent, coated with anti-CD34 antibodies to capture EPCs, was the first device of its kind to be evaluated in humans. In preclinical porcine studies the early reendothelialization process was enhanced without affecting neointimal proliferation. 69 Likewise, in humans neointimal thickness was not reduced as confirmed by angiographic and intravascular ultrasound follow-up. 74 Compared to first-generation paclitaxel-eluting stents, the TRIAS trial showed no significant difference in mortality, myocardial infarction, and target vessel revascularization at 2 years. 75 As inferiority to everolimus-eluting Xience V stent. 77 However, in the prevention of in-stent restenosis in complex lesions the dual endothelial capturing stent technology is less effective compared with drug-eluting stents 78 ; further data is to be expected.
A combination of the administration of growth factors and the concept of EPC capture is followed by Song et al who performed in vitro studies with stainless metallic steel coated with VEGF and anti-CD34 antibody. 79 With co-coating of VEGF and anti-CD34 the differentiation of EPCs in vitro was enhanced compared to single VEGF coating and bare metal, so simultaneously coating stents with VEGF and anti-CD34 antibody might be a novel research direction for facilitating re-endothelialization in order to reduce restenosis after stent implantation.
Cyclic-RGD (Arg-Gly-Asp)-peptide has been identified as a recognition sequence for integrins and thus represents a binding motif specifically attracting endothelial progenitor cells. 80 The use of RGDcontaining peptides on surfaces is known to enhance the adhesion, growth, and spreading of endothelial cells. 81 A novel stent coated with cRGD was investigated earlier by our group. 82 The cRGD coating clearly supported the outgrowth, recruitment, and migration of EPCs in vitro. Scanning electron microscopy indicated enhanced endothelial coverage on cRGD-loaded stents at 4 weeks in a porcine model that might contribute to a reduction of restenosis seen at 12 weeks. To achieve a specific attraction of OECs, a combinational stent coating with certain signalling molecules has been investigated. 83 84 The recruitment and differentiation of EPCs towards an endothelial lineage was improved, however the device has not been tested in preclinical models. ACE inhibitors have been described as beneficial on intimal hyperplasia following balloon angioplasty. 85 Besides the inhibition of conversion of angiotensin I to the vasoconstrictive agent angiotensin II, ACE inhibitors are known to inhibit kinin hydrolysis with stimulation of NO release. 86 The administration of ACE inhibitors has demonstrated an inhibition of neointimal hyperplasia in different animal models of balloon denudation. 85  BVS-Implantation might be based on more physiological flow conditions. 26 Since its commercial launch the BVS has been studied in registries 93,94 and randomized trials, but long-term follow-up is required. 95 Within 1 year there was no significant difference between the BVS and a thin strut second generation DES (Xience) in rates of cardiac death, target-vessel revascularization or target-vessel myocardial infarction. 95 However, stent thrombosis occurred more frequently following implantation of BVS within 1 year, but it has to be pointed out that the studies were underpowered to this event. Nevertheless the finding of an increased rate of stent thrombosis might be related to an incomplete re-endothelialization process, as previous animal studies suggested. 58,95 Improvements in stent design notably in terms of strut thickness might be an area for future research.