The potent vasoconstrictor endothelin-1 (ET-1), acting on the endothelin-A (ETA) receptor, promotes intimal lesion formation following vascular injury. The endothelin-B (ETB) receptor, which mediates nitric oxide release and ET-1 clearance in endothelial cells, may moderate lesion formation, but this is less clear. We used selective ET receptor antagonists and cell-specific deletion to address the hypothesis that ETB receptors in the endothelium inhibit lesion formation following arterial injury.
Neointimal proliferation was induced by wire or ligation injury to the femoral artery in mice treated with selective ETA (ABT-627) and/or ETB antagonists (A192621). Measurement of lesion formation by optical projection tomography and histology indicated that ETA blockade reduced lesion burden in both models. Although ETB blockade had little effect on ligation injury-induced lesion formation, after wire injury, blockade of the ETB receptor increased lesion burden (184% of vehicle; P < 0.05) and reversed the protective effects of an ETA antagonist. Selective deletion of ETB receptors from the endothelium, however, had no effect on neointimal lesion size.
These results are consistent with ETB receptor activation playing an important role in limiting neointimal lesion formation following acute vascular injury, but indicate that this protective effect is not mediated by those ETB receptors expressed by endothelial cells. These data support the proposal that selective ETA antagonists may be preferable to mixed ETA/ETB antagonists for targeting the arterial response to injury.
It is well established that the endothelin (ET) system contributes to the response to acute vascular injury, leading to neointimal lesion formation and luminal occlusion. ET-1 is a potent vasoconstrictor peptide released from endothelial and other cell types, which acts via endothelin-A (ETA) and endothelin-B (ETB) receptor subtypes. ETA receptors on vascular smooth muscle mediate vasoconstriction,1 as well as mitogenesis,2 reactive oxygen species production, and adhesion molecule expression,3,4 while those on leucocytes mediate cytokine release and chemotaxis.5 Consistent with these pro-mitogenic and pro-inflammatory activities of ETA activation, neointima formation in animal models is diminished by selective ETA receptor antagonists.
The role of the ETB receptor in vascular lesion development is more difficult to infer because ETB receptors on different cells mediate actions that might have opposing effects on the response to acute vascular injury. On the one hand, ETB receptors mediate many detrimental actions common to ETA, such as induction of smooth muscle proliferation6 and enhancement of inflammatory responses known to be crucial to neointimal lesion formation.7 On the other hand, ETB mediates actions that might be expected to limit lesion growth. For example, stimulation of ETB in the endothelium promotes production of anti-proliferative, anti-inflammatory nitric oxide (NO),8,9 regrowth of damaged endothelial cells,10 and clearance of circulating ET-1.11,12 Similarly, ETB receptors in vascular smooth muscle stimulate apoptosis13 and those in the kidney suppress blood pressure.14 This specific role of ETB in different cell types has been demonstrated by studies of cell-specific ETB receptor deletion. Selective deletion from EC, for example, results in impaired endothelium-dependent vasodilatation without altering blood pressure, despite elevated plasma ET-1 concentrations,15 whereas deletion from collecting duct epithelial cells produces salt-sensitive hypertension.14
Much of the literature apparently indicates that any role for ETB in neointima formation is modest since mixed ETA/ETB antagonists,16–18 as well as selective ETA19–21 antagonists reduce lesion formation in animal models of acute vascular injury to a similar extent, although this has rarely been directly compared. More direct evidence for a role for ETB comes from two recent studies. Murakoshi et al.22 and Kitada et al.23 both described increased neointimal lesion formation following acute vascular injury in ETB-deficient mice and rats, respectively. Murakoshi et al.22 report that this is associated with reduced tissue levels of NO oxidation products, which may suggest that augmented lesion formation is a response to loss of protective ETB-stimulated NO release, but this remains to be confirmed.
Here, we address the hypothesis that ETB receptor activation in vascular endothelial cells inhibits lesion formation and, therefore, that selective ETA antagonism might be preferable to mixed ETA/ETB antagonism in limiting the response to vascular injury. Using two complementary models of arterial injury in mice, we report that ETB blockade augments neointima formation following acute vascular injury in the presence and absence of concurrent ETA blockade, an effect which is not mediated by ETB receptors expressed on endothelial cells.
All experiments were performed on male 25–30 g mice, in accordance with the Animals (Scientific Procedures) Act (UK), 1986 and approved by the University of Edinburgh ethical review committee (PPL 60/3867). Wild-type C57Bl/6 mice were purchased from Harlan Laboratories, UK. EC-specific ETB receptor deficient mice, on a mixed Ola/129/BKW background, were previously produced in our laboratory by cross of homozygous floxed ETB mice (ETBf/f) with transgenic Tie2-Cre mice24 and cell-specific deletion validated.12,15 Selective deletion of ETB receptors in the endothelium has been confirmed in these mice using a combination of functional, autoradiographic, and immunohistochemical techniques.12,15 As for previous studies using the Cre-loxP recombination system,12,14,15 floxed, Cre-negative littermates (EC ETBf/f) were used as experimental controls for EC-specific ETB deficient mice (EC ETB−/−). To remove any confounding effect of genetic background, key vascular injury studies were repeated using EC ETBf/f and EC ETB−/− mice back-crossed for 10 generations onto a C57Bl/6 background.
The orally active ET receptor antagonists ABT-627 (∼2000-fold ETA-selective; 10 mg/kg/day), A192621 (∼1300-fold ETB-selective; 30 mg/kg/day), or their vehicle (0.2% methyl cellulose) were administered in chow from 1 week before femoral artery injury until completion of the study. ABT-627 and A192621 were kind gifts of Abbott Laboratories, USA.
Blood pressure measurement
Systolic blood pressure was measured by tail cuff plethysmography. Measurements were started 2 weeks before the period of drug administration, and were recorded weekly for the duration of the study.
Femoral artery injury
Intra-luminal wire injury was performed to the left femoral artery according to the method of Sata et al.21 Briefly, a 0.014′ diameter straight-sprung angioplasty guide wire was advanced ∼1.5 cm proximally into the isolated femoral artery through an arteriotomy in the popliteal branch. After its withdrawal, the popliteal branch was ligated and re-perfusion of the injured femoral artery confirmed. In the same animals, the contra-lateral femoral artery was injured using an adaptation of the model of carotid artery ligation described by Kumar and Lindner.25 Briefly, the right femoral artery was isolated, and permanently ligated with 5-0 silk at its junction with the popliteal artery. Surgical procedures were performed under isoflurane inhalation anaesthesia with buprenorphine post-operative analgesia. Depth of anaesthesia was monitored by pedal withdrawal reflex. Mice were allowed to recover for 14 or 28 days before they were sacrificed by trans-cardiac perfusion fixation under terminal pentobarbital anaesthesia. 5-bromo-2-deoxyuridine (BrdU; 100 mg/kg, i.p.) was administered 2 h prior to sacrifice to label proliferating cells.
Optical projection tomography (OPT)
For non-destructive three-dimensional evaluation of lesion formation in injured femoral arteries, vessels were imaged by optical projection tomography (OPT), according to the method we have recently described.26 Briefly, agarose-embedded vessels were optically cleared in benzyl alcohol/benzyl benzoate, and intrinsic fluorescent emission imaged (excitation filter: 425/40 nm; emission filter: 475 nm low pass) using a Bioptonics 3001 OPT tomograph. Data were re-constructed by filtered back-projection using the NRecon software (Skyscan, Belgium). Volumetric measurements were generated by semi-automated tracing of the position of the internal elastic lamina and the neointima delineated from the lumen using a grey level threshold.
Following OPT imaging, femoral arteries were processed to paraffin wax. Serial 3 μm sections were cut through the length of the vessel and, at 50 μm intervals, stained using the ‘United States trichrome’ method.27 Measurements of the maximal intimal area, medial area, and luminal area were recorded by image analysis (Photoshop CS3 Extended, Adobe Systems, USA). To evaluate lesion collagen content, adjacent sections were stained with picro-sirius red (Sigma, UK) and the area of staining within the neointimal layer calculated by colour deconvolution. Measurements were performed by a blinded observer.
Rehydrated paraffin sections were blocked with goat serum (20% v/v) then incubated with anti-α-smooth muscle actin (αSMA; 1:400; 30 min; Sigma, UK) or anti-Mac2 (1:6000; overnight; Cederlane, USA) primary antibodies. After washing, sections were treated with goat anti-mouse or goat anti-rat secondary antibodies, respectively (1:400, 30 min; Vector Labs, UK) before incubation with streptavidin-conjugated horseradish peroxidase (Extravidin; 1:400; 30 min; Sigma, UK). Stains were developed by application of 3,3-diaminobenzidine (Dako, UK) for 1 min. αSMA and Mac2 immunoreactivity were quantified by colour deconvolution (Photoshop CS3 Extended, Adobe Systems, USA) and expressed as a fraction of the total neointimal area. BrdU immunohistochemistry was performed as for αSMA, except that sections were first pre-treated (each 30 min; 37°C) with 2 N HCl (VWR, UK) and 0.1% trypsin (Sigma, UK) before blocking, and the anti-BrdU primary antibody (1:200; Sigma, UK) was incubated for 1 h. BrdU incorporation was recorded by blinded, manual count of immunoreactive nuclei.
Left and right femoral arteries were cleaned of peri-adventitial tissue and divided into ∼2 mm rings and mounted in a Danish Myo Technology 610M myography containing physiological salt solution (PSS; composition: 119.0 mM NaCl, 3.7 mM KCl, 2.5 mM CaCl2, 1.2 mM MgSO4, 25.0 mM NaHCO3, 1.2 mM KH2PO4, 27.0 µM EDTA, 5.5 mM D-glucose). Resting tension of 8 mN was applied, in accordance with previous studies,28 and isometric tension developed by each vascular ring was recorded. In some vessels, the endothelium was removed by rubbing the luminal surface with a human hair. Contractility was initially assessed to 125 mM KCl before recording cumulative concentration response curves (CRCs) to phenylephrine and, after pre-contraction of vessels with an EC80 concentration of phenylephrine, acetylcholine (ACh). After washing, vessels were incubated with the ETA antagonist BQ123 (1 μM), the ETB antagonist A192621 (100 nM), their combination, or vehicle for 30 min and CRCs to ET-1 recorded. Some femoral arteries were used to determine the effect of ETB receptor stimulation by adding a single concentration (100 nM) of the ETB-selective agonist sarafotoxin S6c to femoral arteries following NO synthase inhibition with L-nitro-arginine methyl ester (L-NAME; 10−4M, 30 min).
Statistics and data analysis
Statistical testing and data analysis were performed using the Prism 4.0 software (Graphpad software, USA). Data sets comprising two groups were compared using Student's unpaired t-test. Data sets comprising three or more groups were compared using one-way ANOVA with Dunnett's post hoc test. Functional responses to vasoconstrictor agents were normalized to the tension elicited by 125 mM KCl. Functional responses to acetylcholine were normalized to the tension existing immediately before its first addition. Summary data for concentration–response relationships were calculated by fitting data to a logistic equation. All data are presented as mean ± standard error.
Vasomotor responses to ET agonists
To determine the functional expression of ET receptors in the uninjured mouse femoral artery, vasomotor responses to ET-1 were measured by wire myography in wild-type C57Bl/6 mice. ET-1 produced robust, concentration-dependent contractions of femoral arteries (Figure 1). Removal of the endothelium potentiated the vasoconstriction produced by ET-1 (Figure 1A) and phenylephrine (data not shown). In endothelium-denuded rings, pre-incubation with the ETA antagonist BQ123 (1 μM) resulted in a parallel (none of the treatments significantly altered the Hill slope) rightward shift of the concentration response curve to ET-1 consistent with competitive blockade (Figure 1B). BQ-123 was used for these functional investigations as it has proven selectivity (∼1000-fold) for ETA receptors and has been used previously with isolated arteries.29 The ETB antagonist A192621 (100 nM) did not alter the sensitivity of the responses to ET-1 either in the absence or presence of BQ123 (Figure 1B), but did increase the maximum contraction (P < 0.05) induced by ET-1. In the presence of L-NAME, sarafotoxin S6c produced a small contraction (5.7 ± 1.6% KPSS; n = 6) in isolated murine femoral arteries.
In wild-type mice intended to undergo femoral artery injury, systolic blood pressure (SBP) was reduced by selective ETA blockade (vehicle: 102.4 ± 0.5 mmHg; ABT-627: 96.0 ± 1.6 mmHg; n = 6–8; P < 0.05) or ETA + ETB blockade (ABT-627 + A192621: 96.2 ± 1.0 mmHg; P < 0.05), and increased by selective ETB blockade (A192621: 110.2 ± 1.6 mmHg; P< 0.01). These changes in blood pressure were maintained after femoral artery injury and for the study duration.
Wire injury-induced lesion formation: effect of ET receptor blockade
Femoral artery wire injury resulted in neointimal lesion formation in all treatment groups when examined 28 days after injury and this was quantified volumetrically by OPT imaging. In animals treated with the ETB antagonist A192621, total lesion volume was significantly increased, when compared with that in vehicle-treated animals (Figure 2A). The distribution of lesion formation also appeared to be altered by ETB blockade, with lesions appearing both larger in cross-sectional area and persisting for a greater length of vessel than in vehicle-treated animals (Figure 2B). In animals treated with the ETA antagonist, ABT-627, lesion volume was reduced compared with vehicle-treated animals (Figure 2A). When A192621 was administered concurrently, the reduction in lesion volume produced by ABT-627 was abolished (Figure 2A).
These effects on lesion morphology were confirmed by subsequent histological analysis of the same vessels. A192621-treatment produced an increase in the maximum lesion cross-sectional area (Figure 2C) but did not alter luminal dimensions (Figure 2D). ABT-627 treatment did not affect the neointimal cross-sectional area at this time of maximum lesion formation (28 days after injury) (Figure 2C), but significantly increased the luminal cross-sectional area compared with vehicle-treated mice (Figure 2D). In animals treated with the combination of ABT-627 and A192621, no effects on lesion or luminal dimensions were observed (Figure 2C and D). Medial dimensions were not altered by any treatment (data not shown).
Wire injury-induced lesion composition: effect of ET receptor blockade
Wire injury-induced neointimal lesions exhibited abundant smooth muscle-like immunoreactivity (α-smooth muscle actin; αSMA) and collagen staining (picro-sirius red) across all treatment groups (Figure 3). Macrophage-like immunoreactivity (Mac2) was also present but exhibited a more punctate staining pattern (Figure 3). The fraction of lesion area exhibiting αSMA- or Mac2-immunoreactivity was not altered by any treatment (data not shown). The fraction of lesion area staining for collagen, however, was reduced by ABT-627 treatment (vehicle: 60.8 ± 4.9%; ABT-627: 38.4 ± 5.0%; P< 0.05), with or without concurrent A192621 treatment (ABT-627+A192621: 30.6 ± 8.7%; P < 0.05). A192621 treatment alone did not alter lesion collagen content (A192621: 52.2 ± 5.7%; P> 0.05). Proliferating cells, as indicated by BrdU incorporation, were comparatively rare in all treatment groups. The absolute number of these was not altered by treatment with ABT-627 or ABT-627+A192621, but was increased in lesions from animals treated with A192621 alone (vehicle: 1.3 ± 0.4 cells; A192621: 5.1 ± 2.2 cells; P < 0.05).
Ligation-induced neointima formation: effect of ET receptor blockade
Ligation of the femoral artery across the femoro-popliteal bifurcation also induced formation of neointimal lesions, although these were smaller than those induced by wire injury in both cross-sectional area and axial extent (Figure 4). OPT imaging demonstrated that total lesion volume was reduced by ABT-627 treatment to a similar extent regardless of concurrent A192621 treatment, and that A192621 had no effect when administered alone (Figure 4A). Analysis of lesion distribution suggested that this effect of ETA blockade was primarily to reduce the length of vessel afflicted by lesion formation rather than the maximal cross-sectional area (Figure 4B). In agreement, when the same vessels were examined by traditional histological means at the point of maximum lesion size, no difference was observed between any treatment groups in neointimal (Figure 4C) or luminal (Figure 4D) cross-sectional area.
Wire injury-induced lesion formation: effect of EC-specific ETB deficiency
To determine the contribution of ETB receptors expressed by endothelial cells to the ability of ETB receptors to suppress neointima formation after femoral artery wire injury, additional studies were undertaken in EC-specific ETB-deficient mice (EC ETB−/−) and littermate control animals (EC ETBf/f). As in pharmacological experiments, wire injury elicited robust neointima formation in all animals at both 14 and 28 days after injury. EC-specific ETB deficiency had no effect on the extent of luminal stenosis (Figure 5A) or lesion formation (Figure 5B) at either time point. When these lesions were stained for αSMA, the fraction of the neointimal area exhibiting immunoreactivity was increased by EC-specific ETB deficiency at 14 days (EC ETBf/f: 19.0 ± 7.9%; EC ETB−/−: 51.7 ± 6.9%; P < 0.05) but not 28 days (EC ETBf/f: 46.3 ± 5.2%; EC ETB−/−: 55.9 ± 6.6%; P > 0.05). Mac2 immunoreactivity was not different between genotypes at either time point (data not shown).
To confirm that the discrepancies between pharmacological ETB blockade and EC-specific ETB deletion on wire-induced lesion formation was not due to the differing genetic background of the mice used, additional experiments were carried on EC ETBf/f and EC ETB−/− mice back-crossed for 10 generations onto a C57Bl/6 background. Also, in these animals, EC-specific ETB deficiency had no effect on neointimal cross-sectional area, 28 days after wire injury-induced lesion formation (EC ETBf/f: 68180 ± 14120 μm2; EC ETB−/−: 70680 ± 12010 μm2; n = 5; P = 0.90).
No effect of EC-specific ETB deficiency on ligation-induced lesion morphology or composition was observed (data not shown).
Vasomotor responses: effect of EC-specific ETB deletion
To determine whether changes in ET receptor expression and endothelial function in EC-specific ETB-deficient mice might explain the absence of effect on wire injury-induced lesion formation, we investigated vascular function in healthy femoral arteries from these animals (Figure 6). ET-1 elicited vasoconstriction of femoral artery rings from mice of both genotypes and this response was not different between vessels taken from EC ETB−/− mice and EC ETBf/f mice (Figure 6A). As in those from C57Bl/6 mice, the vasoconstrictor response to ET-1 in femoral arteries from EC ETBf/f mice was sensitive to the ETA antagonist BQ123, but not the ETB antagonist A192621. Acetylcholine stimulated concentration-dependent relaxation of phenylephrine pre-contracted femoral artery rings (Figure 6B). This action was almost abolished by pre-incubation with the NO synthase inhibitor, L-NAME, but no parameter of this response was different between EC ETB−/− and EC ETBf/f vessels. Pre-constricted vessels from EC ETB−/− and EC ETBf/f mice also relaxed in response to increasing concentrations of the NO donor drug sodium nitroprusside, a response that was also not altered by the genotype (data not shown).
This investigation applied pharmacological and cell-specific gene deletion techniques to two distinct models of arterial injury to address the hypothesis that activation of ETB receptors in the vascular endothelium inhibits lesion formation. ETB receptor blockade did, indeed, increase lesion formation following arterial injury, but this was dependent on the model used, with augmented lesion formation apparent after intra-luminal wire injury but not peri-vascular ligation injury. While this demonstrates a protective role for ETB in this setting, selective deletion of ETB receptors from the endothelium had no effect, suggesting that expression of these receptors in other tissues, for instance, vascular smooth muscle, may mediate this response. In addition, we demonstrated that not only was selective ETB blockade detrimental when applied alone, but also that ETB blockade abolished the protective effect of concurrent ETA blockade, supporting the notion that selective ETA antagonists may be preferable to mixed ETA/ETB antagonists for targeting the arterial response to injury.
We first characterized the ET receptors present in the mouse femoral artery by studies of vascular function. These demonstrated that the vasoconstrictor response to ET-1 in this vessel was modulated by generation of endothelium-derived relaxing factors (possibly including a component mediated by ETB-dependent release of endothelium-derived relaxing factors) and was mediated primarily by the ETA receptor. The small response to sarafotoxin S6c indicated that ETB receptors are present in the murine femoral artery but make little contribution to ET-1-mediated contraction. In agreement, ETB blockade had little effect in endothelium-denuded femoral arteries, consistent with a minor role for contractile ETB receptors in this response.
We next considered how pharmacological blockade of ET receptors would alter intimal lesion formation in response to injury to the femoral artery. To do this, we employed two distinct models of acute vascular injury: intra-luminal wire injury and peri-vascular ligation injury. Many potential differences exist between these models. In the wire-injury model, lesion formation is triggered by medial trauma and endothelial denudation and so bears some resemblance to the injury caused by percutaneous coronary interventions in man. Wire injury has been shown to cause extensive medial atrophy, mediated in part by stretch-induced smooth muscle apoptosis and, as such, may depend on recruitment of bone marrow-derived progenitors for neointimal growth.30 The mechanisms responsible for ligation-induced lesion formation are less clear and may involve blood stasis-driven thrombosis or hypoxia of the vessel wall. In contrast, to the wire-injury model, lesions appear to be formed predominantly from local smooth muscle cells30 rather than recruitment of circulating cells. Further, this model does not feature endothelial denudation and may therefore be more sensitive to modulation by the endothelium.
The role of ETA in the neointimal lesion formation was investigated by administration of ABT-627, a highly selective ETA receptor antagonist, to mice undergoing vascular injury. The effectiveness of the dosing regimen was confirmed by a reduction in SBP in accordance with published values in mice.31 Following wire injury, ETA blockade reduced several measures of lesion size and increased luminal cross-sectional area compared with vehicle-treated animals, clearly demonstrating the importance of ETA receptor activation in neointima formation following arterial injury. Importantly, we examined lesions 28 day following injury, a time point that represents a maximum stable lesion size in this model.21 Therefore, our data indicate that ETA blockade produces a change in the absolute extent of lesion development rather than merely altering the rate at which they form. A similar protective effect of ETA blockade was seen in the femoral artery ligation model, in which administration of ABT-627 produced a reduction in lesion volume. This reduction was clearly evident with the three-dimensional analysis but was not detected using standard histological methods, highlighting the limitations of traditional two-dimensional methods of lesion analysis and the potential value provided by three-dimensional imaging techniques in studies of this type. The mechanism responsible for reduced lesion formation in response to ETA antagonism was not explicitly investigated, although ABT-627-treated animals did have reduced blood pressure, which may be protective against neointimal growth. It is important to note, however, that combined ETA/ETB blockade reduced blood pressure to a similar extent without any effect on lesion size, indicating that hypotension alone is insufficient to attenuate lesion growth. Local effects of vascular wall ETA receptors may also be important. Indeed, it has been previously suggested that ETA blockade prevents lesion growth by inhibition of ET-1-mediated smooth muscle cell mitogenesis.17,20
The selective ETB antagonist A192621 was administered to study the role of ETB in the same models of arterial injury. The effectiveness of this regimen was confirmed by an increase in SBP consistent with the previous literature.31 These data also suggest that the blockade of ETB receptors was selective since our data and that of others31 demonstrate that substantial concurrent ETA blockade reverses the hypertensive effect of ETB blockade. In these A192621-treated mice, we observed an increase in several measures of wire injury-induced lesion formation. This is consistent with previous reports22,23 and the concept that ETB receptors act to protect against excessive neointimal growth following vascular injury. Three-dimensional analysis of lesion structure using OPT indicated that the greatest effect of ETB blockade occurred at the edges of lesions, with only small increases apparent in parts of the artery exhibiting large occlusive neointimal growths. Moreover, lesions from A192621-treated mice contained an increased number of proliferating cells, as determined by BrdU incorporation, suggesting that even at this time point, which ordinarily reflects a stable maximum lesion size,21 lesions from mice subject to ETB blockade are continuing to actively develop.
In contrast to the positive results on wire injury-induced lesion formation, ETB blockade did not alter lesion size or volume in the ligation-injury model, either in the presence or absence of concurrent ETA blockade. Therefore, in the femoral artery ligation-injury model, ETB receptors appear to have little role in lesion development and do not modulate the response to ETA blockade. This apparent difference in the actions of ETB between wire- and ligation-induced injuries is surprising for two reasons. First, given the many potentially protective actions of ETB in the endothelium, one might expect the impact of ETB blockade to be more dramatic in an injury model that does not involve endothelial denudation. Secondly, Murakoshi et al.22 describe reduced lesion size following a similar ligation injury to the mouse carotid artery with ETB deletion or blockade. This difference may reflect differences in ETB expression or smooth muscle phenotype between carotid and femoral arteries. It is also important to consider the phase of lesion growth analysed; while we have studied lesions that have achieved their maximum size,21 Murakoshi et al.22 showed an increase in ligation-induced lesion formation at 14 days post-injury, a time point that in this model very much reflects rate of lesion formation rather than absolute size.32
It is not clear why ETB antagonism increased lesion volume following wire injury but not following arterial ligation, although possible explanations can be suggested. It has been shown that the mechanisms of remodelling are different following wire-induced and ligation-induced injury.30 Wire insertion, which causes a denuding injury with extensive stretch-induced damage in the medial wall, disruption of the elastic laminae, and some development of thrombus on exposed surfaces produce a lesion primarily comprised of cells attracted from the circulation. In contrast, ligation induces a non-denuding injury without mechanical stretch or medial damage, resulting in a smaller lesion comprised predominantly of mural cells. Therefore, it is possible that ETB inhibition has a more dramatic effect on lesions involving cells originating in the circulation (wire model) rather than on proliferation and migration of mural smooth muscle cells (ligation model).
In order to determine whether selective ETA antagonism might be superior to mixed ETA/ETB antagonism for reducing lesion development, we also studied of the effect of combined administration of ABT-627 and A192621. These experiments demonstrated that co-administration of A192621 prevented the protective effects of ABT-627 on lesion formation. Importantly, this occurred in the absence of any rise in blood pressure. These results are in contrast to recent work by Kitada et al.23 where comparable inhibitory effects of ABT-627, and the mixed ETA/ETB antagonist J-104143 were observed on lesions resulting from rat carotid artery balloon injury. However, whether this reflects the role of ETB or a different degree of ETA blockade achieved with different drugs is not clear. Our experiments also provide some insight into the mechanism by which ETB blockade might increase lesion size following femoral artery wire injury. We suggest that this is not simply a response to the small elevation (∼8 mmHg) of blood pressure induced by A192621, since in the presence of concurrent ETA blockade A192621 had no effect on blood pressure. These data may also suggest that increased lesion formation is not a result of the impaired plasma [ET-1] clearance that results from ETB blockade12 as the activity of any excess of ET-1 would be blunted in the face of strong ETA + ETB blockade.
ETB receptors on the vascular endothelium mediate a number of processes that would be predicted to prevent lesion formation, including stimulation of NO release,16,17 EC mitogenesis,10 and ET-1 clearance.11,12 However, it is not possible to distinguish the roles of ETB receptors expressed on different cell types using a pharmacological approach. Therefore, to address the hypothesis that the moderating action of ETB on wire-induced lesion formation we observed was conferred by receptors expressed in the vascular endothelium, we used a previously described and validated model of EC-specific ETB deletion.15 Our results demonstrated that selective deletion of ETB from the endothelium had no effect on any measure of lesion size or volume, either at the mature 28-day time point, or at an earlier 14-day time point which reflects a stage of active lesion growth.21 More subtle alterations to wire injury-induced lesion composition were noted, however, and possibly indicate that stimulation of EC ETB delays lesion maturation. Despite this, our data suggest that any role for EC ETB in the regulation of intimal lesion formation is small. This contrasts with the strong effect of pharmacological ETB blockade on wire-induced lesion formation, indicating that ETB receptors expressed by other cell types offer protection against intimal thickening in this model. Several plausible alternative mechanisms for this response can be suggested, such as ETB-mediated inhibition of apoptosis,13 or stimulation of eNOS activity in neointimal vascular smooth muscle cells,33 but these remain to be confirmed.
In agreement with this weak phenotype of EC ETB deficiency on the response to vascular injury, vascular function studies indicated that the loss of these receptors has little impact on femoral artery function. Endothelium-dependent, NO-mediated vasodilatation of femoral arteries in response to acetylcholine was unaltered by EC ETB deficiency. This contrasts with the demonstration that the response to acetylcholine was impaired in the aortas of these mice.15 ET-1-induced contraction of the femoral artery was mediated by ETA receptor activation (showing a rightward shift in the presence of an ETA antagonist) but was also unaffected by EC ETB deletion. Taken together, these results suggest that EC ETB receptors have a minimal role in regulating vascular function in the mouse femoral artery.
In summary, we have demonstrated that ETB receptor blockade exacerbates, and ETA receptor blockade attenuates, intimal lesion formation following acute vascular injury. In the femoral artery wire-injury model, ETB blockade was not only detrimental when applied alone, but also prevented the beneficial effects of concurrent ETA blockade. This effect of ETB was not lost when it was specifically ablated from ECs, suggesting a novel protective action of non-endothelial cell ETB receptors in this setting, perhaps those in the vascular smooth muscle. These data clearly indicate that in the prevention of intimal hyperplasia, selective ETA receptor antagonists may be preferred to agents that produce substantial concurrent ETB blockade.
Conflict of interest: D.J.W. has provided advice to Abbott, Encysive, Pfizer, and Roche in relation to the clinical development of ET receptor antagonists.
N.S.K. was the recipient of a University of Edinburgh Studentship. K.D. was funded by a BHF project grant (PG/08/068/25461). The authors were supported by the BHF-funded Centre of Research Excellence (CoRE) at The Queen's Medical Research Institute.