¹BHF Blood Pressure Group, Department of Medicine and Therapeutics, University of Glasgow, Glasgow, UK

²Departments of Internal Medicine and Program in Gene Therapy, University of Iowa College of Medicine and VA Medical Center, Iowa City, IA, USA

Correspondence to: A F Dominiczak, BHF Blood Pressure Group, Department of Medicine and Therapeutics, Western Infirmary, University of Glasgow, Glasgow G11 6NT, UK

Gene transfer may be appropriate for therapeutic protocols targeted at the vascular endothelium. Endothelial dysfunction is the principal phenotype associated with atherosclerosis and hypertension. Oxidative stress has been implicated in the development of endothelial dysfunction. We have explored the ability of overexpressing anti-oxidant genes (superoxide dismutases; SODs) in vitro and in vivo to assess their potential for reversing endothelial dysfunction in a rat model, the stroke-prone spontaneously hypertensive rat (SHRSP). Western blotting and immunofluorescence assays in vitro showed efficient overexpression of MnSOD and ECSOD with respect to localisation to the mitochondria and extracellular surface, respectively. Transgene functional activity was quantified with SOD activity assays. MnSOD and ECSOD overexpression in intact SHRSP vessels in vivo led to endothelial and adventitial overexpression. Pharmacological assessment of transduced vessels following in vivo delivery by basal NO availability quantification demonstrated that the 'null' adenovirus and MnSOD adenovirus did not significantly increase NO availability. However, AdECSOD-treated carotid arteries showed a significant increase in NO availability (1.91 ± 0.04 versus 0.75 ± 0.08 g/g, n = 6, P = 0.029). In summary, efficient overexpression of ECSOD, but not MnSOD in vivo, results in improved endothelial function in a rat model of hypertension and has important implications for the development of endothelial-based vascular gene therapy.

Gene Therapy 2001 9, 110-117. DOI: 10.1038/sj/gt/3301633

adenovirus; gene transfer; superoxide dismutase; nitric oxide; SHRSP

Introduction

The use of viral vectors to efficiently deliver genes to the vasculature using local delivery approaches has long been established. These approaches have broad applications in the development of efficacious therapies for diverse vascular pathologies, including restenosis following balloon angioplasty, vein graft intimal thickening and endothelial cell dysfunction associated with atherosclerosis and hypertension.¹

Reduced nitric oxide (NO) availability is a common finding in many forms of vascular disease and a reduction in either NO or NO-mediated vasodilatation has been found in endothelial dysfunction due to causes as diverse as hypertension,² hypercholesterolaemia,³ cigarette smoking⁴ and diabetes.⁵ Reduced NO availability is the result of either impaired production or enhanced oxidative degradation. The endothelium is the primary source of constitutive NO release in the vasculature and the loss of NO-dependent vasodilatation is common to all forms of endothelial dysfunction. Increasing NO availability, by stimulation of NO synthesis or prevention of NO degradation, holds promise as a means to improve the phenotype of endothelial dysfunction.

Reactive oxidative species (ROS) especially superoxide (O₂^-), are known to play a role in the aetiology of endothelial dysfunction.^6,7,8,9 NO is scavenged by its reaction with O₂^- which produces the deleterious molecule peroxynitrite (ONOO^-) and removes the vasodilatory, anti-atherogenic, anti-inflammatory, and anti-proliferatory effects of NO. In addition, ONOO^- itself has a range of detrimental effects on cellular proteins, lipids and DNA.

The stroke-prone spontaneously hypertensive rat (SHRSP) model displays a phenotype resembling human essential hypertension. SHRSP develop a progressive and severe increase in blood pressure, left ventricular hypertrophy and both cerebral infarcts and haemorrhages. The reduced NO availability in this model has been shown to be the result of NO scavenging by the raised level of vascular O₂^-.¹⁰ Our previous work has shown that local NO availability in the SHRSP can be restored to levels comparable to those seen in the normotensive Wistar-Kyoto rat using adenovirus-mediated endothelial nitric oxide synthase (NOS) gene transfer.¹¹

Increasing NO production directly by NOS gene transfer in tissues with raised O₂^- levels runs the theoretical risk of increasing ONOO^- production. An alternative to increasing NO production is to prevent its degradation by O₂^- using SOD gene transfer. There are three SOD isoforms in humans, intracellular copper zinc SOD (CuZnSOD), mitochondrial manganese SOD (MnSOD) and extracellular SOD (ECSOD), all of which reduce O₂^- to diffusible hydrogen peroxide (H₂O₂), which is then reduced to H₂O by either catalase or glutathione peroxidase.¹² Increasing SOD expression has been shown to protect against oxidative stress in various experimental models including myocardial stunning,¹³ liver transplantation¹⁴ and aging.¹⁵ Adenovirus-mediated gene transfer of intracellular CuZnSOD or MnSOD has been shown to reduce O₂^- and LDL oxidation in cultured aortic endothelial cells.¹⁶ Previous in vivo experiments using SOD protein infusion in the vasculature were ineffective due to the rapid degradation of the SOD protein and the inability of plasma SOD to penetrate the cell membrane.¹⁷ Polyethylene-glycolated SOD protein (PEGSOD) resulted in a marked improvement in endothelium-dependent vascular relaxation in cholesterol-fed rabbits.¹⁸ The intravenous infusion of a membrane-permeable SOD-mimetic normalised mean arterial pressure and renal vascular resistance in the spontaneously hypertensive rat.¹⁹ Adenovirus-mediated SOD gene transfer in various models of cardiovascular disease has given differing results.^20,21,22,23 In vivo adenovirus-mediated intracellular CuZnSOD gene transfer in the SHRSP did not result in a significant increase in NO availability.²⁴ We therefore investigated the potential of adenovirus-mediated transfer of the human MnSOD and ECSOD genes to improve NO availability and endothelial function in the SHRSP model of essential hypertension.

Results

Western blotting for human MnSOD and ECSOD

We first assessed the endogenous expression of MnSOD and ECSOD, as well as the levels evoked by overexpression by adenovirus-mediated gene transfer. MnSOD was not detected in endothelial cells infected with the -galactosidase expressing virus (Adgal) or in uninfected cells (Figure 1a and b). Analysis of cell lysates harvested from cells infected with the MnSOD adenoviral vector (AdMnSOD) revealed efficient production of the 22-kDa protein. A second, MnSOD precursor band was also visible at 24 kDa at higher multiplicities of infection (MOI) Figure 1a.

ECSOD was not detected in the cells or conditioned media of Adgal-infected or mock-infected cells. Endothelial cells infected with the ECSOD adenovirus (AdECSOD) produced a dose-dependent increase in protein expression in total cell lysates with no 32-kDa ECSOD band detected in concentrated conditioned media. As expected, the addition of heparin competitively released ectopic ECSOD into the conditioned media Figure 1c. This confirmed the expression of MnSOD and ECSOD and localised ECSOD to the cell cytoplasm and the cell surface.

Immunofluorescent labelling of MnSOD and ECSOD

Immunofluorescent staining was used to confirm MnSOD and ECSOD expression and to analyse colocalisation of MnSOD to the mitochondria. Both AdMnSOD- and AdECSOD-infected cells showed increased fluorescence intensity with increasing MOI (Figure 2c and f, and Figure 2b and eFigure 2c and f, and Figure 2b and e, respectively). Both the uninfected cells and those infected with Adgal had almost undetectable levels of either MnSOD or ECSOD Figure 2a and d. No fluorescence was detected using a non-immune sheep IgG for MnSOD and rabbit IgG for ECSOD in either infected or uninfected cells. Colocalisation experiments using FITC labelling of MnSOD, a mitochondrial-specific probe and confocal microscopy in AdMnSOD-infected endothelial cells showed that the MnSOD protein was localised to the mitochondria Figure 2g-i. These experiments confirmed the overexpression of the human MnSOD and ECSOD and localised the exogenous MnSOD to the mitochondria.

Evaluation of in vitro SOD activity

SOD activity was examined in both adenovirus-infected and -uninfected human umbilical vein endothelial cells (HUVECs) (Figure 3). Infection with Adgal did not result in any significant increase in SOD activity in comparison to uninfected cells (uninfected: 0.07 ± 0.02 U/mg, Adgal 500 0.04 ± 0.01 U/mg). Infections using AdMnSOD resulted in a dose-dependent, statistically significant increase in total SOD activity (MOI 100: 0.35 ± 0.03 U/mg, P = 0.01; MOI 500: 0.83 ± 0.18 U/mg, P = 0.001). AdECSOD infection also increased SOD activity, but was only significant at the higher MOI (MOI 100: 0.11 ± 0.03 U/mg, non-significant; MOI 500: 0.39 ± 0.12 U/mg, P = 0.05).

In vivo gene delivery

We next used a local delivery procedure¹¹ to overexpress MnSOD or ECSOD in carotid arteries of both hypertensive SHRSP animals and normotensive WKY controls. SOD overexpression in rat carotid arteries was evaluated on transverse histological sections of the arteries harvested 24 h after in vivo infection with either AdMnSOD or AdECSOD. No staining under these conditions was observed in the contralateral uninfused carotid artery in either case (Figure 4a and c) or in control virus-infected controls (not shown). Antibody staining showed the expression of both the human MnSOD and the ECSOD was localised to the endothelial and the adventitial layers of the vessel wall Figure 4b and d consistent with previous studies using this local delivery approach.^11,24

Vascular contractility studies

We used an established ex vivo pharmacological approach to evaluate and quantify any potential benefit to local NO bioavailability evoked by MnSOD or ECSOD overexpression. NO availability was measured by the increase in the pressor response to phenylephrine in the presence of NOS blockade by N^G-nitro-L-arginine methylester (L-NAME) and expressed as an area under the curve (AUC). As expected, control adenovirus (RAd66) had no effect on vascular contractility (Figure 5) compared with uninfected control carotid arteries (RAd66 AUC: 0.96 ± 0.17 g/g, untreated AUC: 1.18 ± 0.11 g/g, n = 6, P = 0.384, 95% CI: -0.831, 0.381). AdMnSOD carotid artery infusions did not significantly alter NO bioavailability relative to the contralateral control arteries (AdMnSOD AUC: 0.76 ± 0.11 g/g, untreated AUC: 0.75 ± 0.19 g/g, n = 6, P = 0.944, 95% CI: -0.564, 0.597). However, in AdECSOD-infused vessels, L-NAME had a greater effect on pressor responses to phenylephrine than in the contralateral control arteries as shown in Figure 5. The AUC was increased from 0.75 ± 0.08 g/g in the uninfused control arteries to 1.91 ± 0.04 g/g in the AdECSOD-infused arteries (n = 6, P = 0.029, 95% CI: 0.179, 2.128). When responses to carbachol (EC₅₀ and E_max) were compared, no significant differences between the infused vessel and its contralateral control were found in any group. E_max values were 53 ± 5% and 48 ± 4% for control and RAd66, 60 ± 11% and 57 ± 9% for control and AdMnSOD and 57 ± 6% and 61 ± 11% for control and AdECSOD. These experiments clearly demonstrate that ECSOD, but not MnSOD overexpression improved NO bioavailability in hypertensive rat carotid arteries.

Discussion

We used adenoviral vectors to successfully transfer functional human MnSOD and ECSOD genes both in vitro and in vivo. In vivo experiments showed that AdECSOD, but not AdMnSOD, successfully increased NO availability in the SHRSP model of hypertension. Western analysis of HUVECs infected with AdMnSOD or AdECSOD showed a dose-dependent increase in the corresponding protein with increasing MOI. Interestingly, at higher levels of infection, a second band was detected in each case. Western analysis for the 22-kDa MnSOD protein revealed a 24-kDa band, which we believe may be a MnSOD precursor resulting from saturation of a precursor cleavage mechanism at high levels of MnSOD production. This doublet has also been reported by Fang et al¹⁶ in studies on MnSOD gene transfer into cultured endothelial cells. Similarly, at higher MOIs, Western analysis for the 32-kDa ECSOD monomer reveals a second smaller (~29 kDa) human ECSOD band. This is thought to be the result of intracellular proteolysis of ECSOD in which the C-terminal heparin-binding domain is removed.^25,26,27 Western analysis of conditioned media showed detectable levels of ECSOD in the media of AdECSOD-infected cells. Heparin treatment of AdECSOD-infected cells resulted in a large release of ECSOD into the media, indicating that ECSOD was present on the cell surface heparan sulphate proteoglycans. Uninfected and Adgal-infected cells did not release any detectable levels of ECSOD in the media, even after heparin treatment. Further confirmation of increased protein production in infected cells was shown by immunofluorescent labelling of MnSOD and ECSOD, which localised the proteins within the mitochondria and the cytosol, respectively. Increased SOD function in AdMnSOD- and AdECSOD-infected cells was confirmed by SOD activity assay of cell lysates, which demonstrated an increase in activity in AdMnSOD and AdECSOD-infected cells relative to the basal levels of SOD activity seen in Adgal-infected and uninfected cells. Local NO availability was significantly increased in carotid arteries infected with AdECSOD, but not AdMnSOD. Interestingly, the increase in NO availability obtained by AdECSOD parallels the increase in NO availability found after AdeNOS gene transfer in the SHRSP model.²⁴

Although basal NO availability was increased after ECSOD gene transfer, no effect was observed on carbachol-stimulated NO levels. In vivo, the basal release of NO has a major role in determining basal blood flow²⁸ and in patients, the impaired basal production of NO is associated with increased systemic blood pressure.²⁹ Basal and stimulated NO release diverge under a number of circumstances,^30,31 and Mian and Martin³² have reported that basal NO has a greater sensitivity to destruction by superoxide than acetylcholine-stimulated NO activity.

In this study, SOD vectors were used because eNOS is a source of superoxide under conditions of reduced L-arginine³³ or tetrahydrobiopterin (BH4) availability³⁴ and it has been shown to be an important source of O₂^- in the SHRSP.¹⁰ Elevating NO availability in a disease model that has raised levels of O₂^-could also result in increased production of detrimental ONOO^-. Elevating NO availability by reducing O₂^- directly may provide a means to avoid this. Raised ONOO^- levels are associated with a range of harmful effects including DNA damage, lipid peroxidation, nitration of tyrosine residues in proteins, and inactivation of MnSOD.^35,36,37 Recently, ONOO^- has been implicated in the production of O₂^- via the oxidation of the vital eNOS cofactor BH4.^38,39 Conversely, SOD gene transfer may result in increased H₂O₂ production as more O₂^- is dismutated. However, Liochev and Fridovich⁴⁰ have proposed that increasing dismutation of O₂^- would prevent the formation of increased levels of H₂O₂ by other reactions. This is supported by the finding that cell lines overexpressing CuZnSOD had reduced levels of H₂O₂.⁴¹

We hypothesise that ECSOD successfully increased NO availability because like eNOS, it is bound to the cell membrane and is able to act beyond the infected cell. It is found both intracellularly and extracellularly, binding to heparan sulphate proteoglycans on the cell surface and in the extracellular matrix. Therefore it has a larger distance of effect, probably extending into the extracellular matrix of the smooth muscle layer of the vessel wall. Unlike NO, O₂^- is unable to diffuse between cellular compartments and its effects are much more localised. MnSOD is contained within the mitochondria and is therefore only able to affect local mitochondrial O₂^- concentration (Figure 6). The major sources of O₂^- production in the SHRSP and in many other forms of endothelial dysfunction are NAD(P)H oxidase and eNOS.^10,12 The cell surface location of ECSOD makes it well suited to react with the O₂^- produced by both these membrane-bound enzymes. Both eNOS and ECSOD have a sphere of influence that extends beyond infected cells, which is not the case for the intracellular CuZnSOD and MnSOD. The effectiveness of AdECSOD gene transfer in this work is in agreement with the recent findings of Li et al²³ who used AdECSOD gene transfer combined with heparin to reduce myocardial infarct size by 50% in a rabbit model of myocardial infarction and reperfusion.

Nakane et al²⁰ found that the AdECSOD vector did not improve NO-mediated relaxation in aortae from an angiotensin II-infused rabbit model of hypertension. Although there were many differences between their studies and ours (species, different sources of O₂^-, vessels used and ex vivo rather than in vivo infection), our results agree in that neither showed any improvement in stimulated NO release after AdECSOD infection. Basal NO production was not measured in their SOD-infused, angiotensin II rabbit model of hypertension. Our results showed an important increase in basal NO production after AdECSOD, but not AdMnSOD infection. Recently, AdMnSOD gene transfer has been shown to improve NO-mediated vasorelaxation in models of atherosclerosis and diabetes.^21,42 The apparent discrepancy in results between these findings and ours may be explained by the different disease models used, the larger viral dose used and the probability that a 1-h ex vivo infection is more effective than a 20-min in vivo infection.

Further evidence for the improved effectiveness of ECSOD relative to the other two SOD isoforms is the longer half-life of the ECSOD protein when injected into the circulation in vivo. The half-lives of CuZnSOD and MnSOD have been estimated to be 6 min and 6 h, respectively,⁴³ whereas the half-life of ECSOD has been calculated at 20 h.⁴⁴ This is almost certainly the result of the ability of ECSOD to bind to the endothelial cell surface heparan sulphate proteoglycans. If this correlates to the half-life of SODs in the vessel wall, the combination of the longer-lasting SOD isoform with continuous protein production from gene transfer will provide a definite improvement over earlier SOD protein infusion methods and forms a promising basis for future longer-term experiments.

Another potential reason why ECSOD, but not MnSOD, successfully increases NO availability in the SHRSP is that rat vascular cells have much lower levels of ECSOD than human cells.⁴⁵ Immunohistochemistry of normal human arteries showed ECSOD throughout the vessel wall.⁴⁶ In rats, the vast majority of ECSOD exists as the ECSOD-A and ECSOD-B fractions, which have little or no affinity for heparin and exist primarily in the plasma. In humans and many other mammals, the reverse is true, much of the ECSOD exists as the C fraction and is present in the vessel wall and forms an equilibrium between the plasma and the endothelium. Therefore the addition of a functional human ECSOD gene into the vessel wall of the rat which does not express its own ECSOD-C to any great extent, results in a significant increase in the dismutation of superoxide. This could result in the increase in NO availability observed. As in rats, reduced vascular ECSOD-C activity has been found in patients with coronary artery disease⁴⁷ and both these situations may be more responsive to vascular ECSOD gene transfer.

In conclusion, our work has shown that local adenovirus-mediated transfer of the human ECSOD gene significantly improved NO availability in the SHRSP model of hypertension and endothelial dysfunction, whereas MnSOD gene transfer had no effect. To our knowledge, this is the first example of in vivo ECSOD gene transfer resulting in an improvement in the phenotype of a hypertensive model. These findings give an indication of the feasibility and potential of ECSOD gene transfer in endothelial dysfunction.

Materials and methods

Adenoviral vectors

First-generation AdMnSOD and AdECSOD vectors were obtained from Dr Beverly Davidson, director of Gene Transfer Core, University of Iowa, IA, USA. The first-generation control Adgal virus and 'null' adenovirus (RAd66) were obtained from Dr G Wilkinson, University of Wales, Cardiff, UK, and have been described previously.^48,49 High-titre stocks were generated and titred using established protocols.⁵⁰ The presence of replication-competent adenovirus was excluded by end point dilution on HeLa cells (ECAC, UK).

Western blotting for human MnSOD and ECSOD

HUVECs were purchased from TCS Cellworks, Botolph-Claydon, UK. For all experiments, low-passage (<6) cells were used. For in vitro studies, HUVECs were seeded into 24-well plates and infected in triplicate for 16 h at a range of MOIs with either AdMnSOD, AdECSOD or Adgal at 70% confluence. The media was then replaced and the cells harvested 48 h later. Heparin (100 IU/ml) was added to half of the wells infected with AdECSOD and their corresponding controls 4 h before harvesting. Conditioned media was concentrated using Centricons (Millipore, Bedford, MA, USA) as per manufacturer's instructions. The cells were washed twice in PBS then lysed in 200 l of 1 ´ reducing sample lysis buffer (1% -mercaptoethanol, 2% SDS, 5% glycerol, 62.5 mmol/l Tris pH 6.8, 0.003% bromophenol blue) to each well. Samples were subjected to 10% polyacrylamide gel elecrophoresis and transferred to Hybond-P membranes as previously described.¹¹

Human MnSOD was detected using a 1/1000 solution of a rabbit anti-human primary antibody, which was a kind gift from Professor Taniguchi, Osaka University Medical School, Osaka, Japan. The human ECSOD was detected using a 1/2000 solution of a polyclonal rabbit anti-human primary antibody, which was a generous gift from Professor JD Crapo, National Jewish Medical Research Center, Denver, CO, USA. Both primary antibodies were detected using a 1/2500 solution of horseradish peroxidase conjugated anti-rabbit antibody (SAPU, Carluke, UK). Bands were visualised using the enhanced chemiluminescence plus (ECL+) system (Amersham Pharmacia Biotech, Little Chalfont, UK) and exposed to X-ray film.

Immunofluorescent labelling of MnSOD and ECSOD

HUVECs were seeded into 24-well plates containing sterile 13-mm coverslips. At 70% confluency, cells were mock-infected or infected for 16 h in triplicate with either AdMnSOD, AdECSOD or Adgal at a range of MOIs up to 300. The medium was changed and immunofluorescent labelling performed 48 h later. MnSOD was detected using a 1/100 dilution of a sheep anti-human MnSOD primary antibody and non-immune sheep IgG (Sigma, Poole, UK) was used as a negative control. Blocking was performed in 10% rabbit serum. A fluorescein isothiocyanate (FITC)-labelled rabbit anti-sheep antibody (Dako, Ely, UK) was used for visualisation. ECSOD was detected using a 1/200 dilution of a rabbit anti-human ECSOD primary antibody and a non-immune rabbit IgG (Sigma) was used as a negative control. Blocking was performed using 10% swine serum. FITC-labelled swine anti-rabbit antibody (Dako) was used for detection. Coverslips were mounted using Vectashield (Vector Laboratories, Burlingame, CA, USA) and visualised using an Olympus BX40 microscope with BX-FLA fluorescence attachment. FITC-labelled cells were excited using a wavelength of 490 nm and detected using a 515-nm filter. Pictures were taken and analysed using Image-Pro Plus software.

For MnSOD colocalisation studies, endothelial cells were grown on 22-mm coverslips in six-well plates. On reaching 70% confluency, they were mock-infected or infected for 16 h with either AdMnSOD or Adgal. Immunofluorescent labelling was done 48 h later. MitoTracker Red (Molecular Probes, Leiden, The Netherlands) mitochondrial fluorescent dye was added 45 min before immunolabelling to produce a final concentration of 300 nmol/l. MnSOD immunolabelling was performed as described above except coverslips were mounted in Immu-mount (Shandon Laboratories, Pittsburg, PA, USA). The cells were visualised using a laser scanning confocal microscope (Zeiss Axiovert 100). FITC was excited using a 488-nm krypton/argon laser and detected with a 515-540 nm band pass filter. The MitoTracker was excited at 543 nm and detected with a 570-nm filter. All images were visualised using Zeiss LSM or MetaMorph software to assess colocalisation.

Evaluation of in vitro SOD activity

SOD activity was assessed by infection of HUVECs at increasing MOIs using the Adgal, AdMnSOD and AdECSOD vectors. Cells were harvested 48 h later, lysed and 20-l aliquots assayed using a SOD assay kit (Calbiochem, San Diego, CA, USA) as described previously.²⁴ The assay is based on the oxidation of 5,6,6a,11b-tetrahydro-3,9,10-trihydroxybenzo[c]fluorene (THBF) and is measured by the rate of increase in optical density at 525 nm. Results were expressed as SOD-525 units/mg protein (U/mg).

In vivo carotid artery gene transfer

The study conforms to the Guide for the Care and Use of Laboratory Animals published by the National Institute of Health (NIH Publication No 85-23, revised 1996) and carried out under project license from the UK Home Office. Male SHRSP rats were studied at 12 weeks of age. They were anaesthetised using inhaled halothane. The left carotid arteries were infused as described¹¹ previously with 30 l of either RAd66, AdMnSOD or AdECSOD at a titre of 8 ´ 10⁹ pfu/ml. Briefly, the left common carotid artery was exposed and isolated between microvascular clamps. The virus solution was instilled into the lumen and incubated for 20 min. Normal circulation was then restored, the incision was sutured and the animal was allowed to recover. In each experiment, the uninfected right contralateral carotid artery from each animal was used as a matched control. Twenty-four h after infusion, rats were killed by halothane overdose and the carotid arteries were removed and used for isometric tension recording or immunohistochemistry. Three groups of SHRSPs were used for isometric tension recording (n = 18). Each virus (RAd66, AdMnSOD and AdECSOD) was infused into six animals.

Carotid artery immunohistochemistry

Paraffin-embedded sections were treated as previously described.¹¹ For both SOD isoforms, SOD was detected by the rabbit anti-SOD primary antibodies described above. The secondary antibody was a horseradish-peroxidase labelled sheep anti-rabbit IgG (SAPU). Binding was visualised using 3', 3'-diaminobenzadine (DAB Liquid Substrate System; Sigma) and 0.01% H₂O₂ as a chromogen. All sections were air-dried and counterstained with haematoxylin and examined by light microscopy.

Vascular contractility studies

Carotid arteries from the experimental animals were infused, harvested and suspended in organ baths as previously described.¹¹ Isometric tension studies were performed using a force transducer and recorded using a MacLab dedicated computer. Vessel rings were contracted twice with 0.1 mol/l KCl, washed out and were then allowed to relax after each contraction. Cumulative concentration-response curves to phenylephrine (PE) (0.01-10 mol/l) were constructed, first in the absence, and after washout, in the presence of 100 mol/l L-NAME to inhibit NO synthase. The increase in tension in the presence of L-NAME provides a measure of the effect of NO on basal tone.³⁰ The increase in tension in the presence of L-NAME was calculated for each ring over the full concentration-response curve and expressed as an area under the curve (AUC). After constructing the first concentration-response curve to PE the rings were relaxed to carbachol (0.01-10 mol/l) as a measure of agonist stimulated NO release. E_max (the maximal relaxation) and EC₅₀ (the concentration of carbachol required to cause 50% of maximal relaxation) were calculated for each ring. Carotid artery rings were maintained in Kreb's buffer, which contained indomethacin (0.02 mmol/l) to inhibit any prostanoid-mediated effects.

Statistical analysis

Data were expressed as mean ± s.e.m. For SOD activity assays, ANOVA with Dunnett's correction for multiple comparisons was used. In each vascular contractility experiment, n refers to the number of animals used. The increase in tension in the presence of L-NAME in the infused rings (AdMnSOD or AdECSOD or RAd66) and the E_max and EC₅₀ for carbachol were compared with that obtained in the contralateral untreated control rings using Student's paired t test. A level of P < 0.05 was accepted as statistically significant.

Acknowledgements

This work is supported by the British Heart Foundation Program Grant RG/97009 and Project Grant PG/200023 to AFD. JPF is funded by a National Heart Research Fund studentship. We are very grateful to Professor Gwyn Gould for his help with confocal microscopy. We would also like to thank E Beattie, E Jardine and D McSharry for technical assistance.

1 Yla-Herttuala S., Martin J.F. Cardiovascular gene therapy. Lancet 2000; 355: 213, 213. MEDLINE

2 Panza J.A., Quyyumi A.A., Brush J.E.J., Epstein S.E. Abnormal endothelium-dependent vascular relaxation in patients with essential hypertension. New Engl J Med 1990; 323: 22, 22. MEDLINE

3 Creager M.A.et al. . Impaired vasodilation of forearm resistance vessels in hypercholesterolemic humans. J Clin Invest 1990; 86: 228, 228. MEDLINE

4 Kiowski W.et al. . Diminished vascular response to inhibition of endothelium-derived nitric oxide and enhanced vasoconstriction to exogenously administered endothelin-1 in clinically healthy smokers. Circulation 1994; 90: 27, 27. MEDLINE

5 Calver A., Collier J., Vallance P. Inhibition and stimulation of nitric oxide synthesis in the human forearm arterial bed of patients with insulin-dependent diabetes. J Clin Invest 1992; 90: 2548, 2548. MEDLINE

6 Bouloumié A.et al. . Endothelial dysfunction coincides with an enhanced nitric oxide synthase expression and superoxide anion production. Hypertension 1997; 30: 934, 934. MEDLINE

7 Warnholtz A.et al. . Increased NADH-oxidase-mediated superoxide production in the early stages of atherosclerosis: evidence for involvement of the renin-angiotensin system. Circulation 1999; 99: 2027, 2027. MEDLINE

8 Miller F.J.J.et al. . Superoxide production in vascular smooth muscle contributes to oxidative stress and impaired relaxation in atherosclerosis. Circ Res 1998; 82: 1298, 1298. MEDLINE

9 Ohara Y., Peterson T.E., Harrison D.G. Hypercholesterolemia increases endothelial superoxide anion production. J Clin Invest 1993; 91: 2546, 2546. MEDLINE

10 Kerr S.et al. . Superoxide anion production is increased in a model of genetic hypertension: role of the endothelium. Hypertension 1999; 33: 1353, 1353. MEDLINE

11 Alexander M.Y.et al. . Gene transfer of endothelial nitric oxide synthase improves nitric oxide-dependent endothelial function in a hypertensive rat model. Cardiovasc Res 1999; 43: 798, 798. MEDLINE

12 McIntyre M., Bohr D.F., Dominiczak A.F. Endothelial function in hypertension: the role of superoxide anion. Hypertension 1999; 34: 539, 539. MEDLINE

13 Li Q.et al. . Gene therapy with extracellular superoxide dismutase attenuates myocardial stunning in conscious rabbits. Circulation 1998; 98: 1438, 1438. MEDLINE

14 Lehmann T.G.et al. . Delivery of Cu/Zn-superoxide dismutase genes with a viral vector minimizes liver injury and improves survival after liver transplantation in the rat. Transplantation 2000; 69: 1051, 1051. MEDLINE

15 Parkes T.L.et al. . Extension of Drosophila lifespan by overexpression of human SOD1 in motorneurons. Nat Genet 1998; 19: 171, 171. Article MEDLINE

16 Fang X.et al. . Overexpression of human superoxide dismutase inhibits oxidation of low-density lipoprotein by endothelial cells. Circ Res 1998; 82: 1289, 1289. MEDLINE

17 Garcia C.E.et al. . Effect of copper-zinc superoxide dismutase on endothelium-dependent vasodilation in patients with essential hypertension. Hypertension 1995; 26: 863, 863. MEDLINE

18 Mugge A.et al. . Chronic treatment with polyethylene-glycolated superoxide dismutase partially restores endothelium-dependent vascular relaxations in cholesterol-fed rabbits. Circ Res 1991; 69: 1293, 1293. MEDLINE

19 Schnackenberg C.G., Welch W.J., Wilcox C.S. Normalization of blood pressure and renal vascular resistance in SHR with a membrane-permeable superoxide dismutase mimetic: role of nitric oxide. Hypertension 1998; 32: 59, 59. MEDLINE

20 Nakane H.et al. . Gene transfer of endothelial nitric oxide synthase reduces angiotensin II-induced endothelial dysfunction. Hypertension 2000; 35: 595, 595. MEDLINE

21 Zanetti M., Sato J., Katusic Z.S., O'Brien T. Gene transfer of superoxide dismutase isoforms reverses endothelial dysfunction in diabetic rabbit aorta. Am J Physiol. Heart Circ Physiol 2001; 280: H2516, H2516.

22 Lund D.D., Faraci F.M., Miller F.J.J., Heistad D.D. Gene transfer of endothelial nitric oxide synthase improves relaxation of carotid arteries from diabetic rabbits. Circulation 2000; 101: 1027, 1027. MEDLINE

23 Li Q.et al. . Gene therapy with extracellular superoxide dismutase protects conscious rabbits against myocardial infarction. Circulation 2001; 103: 1893, 1893. MEDLINE

24 Alexander M.Y.et al. . Gene transfer of endothelial nitric oxide synthase, but not Cu/Zn superoxide dismutase restores nitric oxide availability in the SHRSP. Cardiovasc Res 2000; 47: 609, 609. MEDLINE

25 Enghild J.J.et al. . The heparin-binding domain of extracellular superoxide dismutase is proteolytically processed intracellularly during biosynthesis. J Biol Chem 1999; 274: 14818, 14818. MEDLINE

26 Fukai T.et al. . Modulation of extracellular superoxide dismutase expression by angiotensin II and hypertension. Circ Res 1999; 85: 23, 23. MEDLINE

27 Ohta H., Adachi T., Hirano K. The nature of heterogeneous components of extracellular-superoxide dismutase purified from human umbilical cords. Free Rad Biol Med 1993; 15: 151, 151. MEDLINE

28 Vallance P., Collier J., Moncada S. Effects of endothelium-derived nitric oxide on peripheral arteriolar tone in man. Lancet 1989; 2: 997, 997. MEDLINE

29 Panza J.A., Casino P.R., Kilcoyne C.M., Quyyumi A.A. Role of endothelium-derived nitric oxide in the abnormal endothelium-dependent vascular relaxation of patients with essential hypertension. Circulation 1993; 87: 1468, 1468. MEDLINE

30 McIntyre M.et al. . Sex differences in the abundance of endothelial nitric oxide in a model of genetic hypertension. Hypertension 1997; 30: 1517, 1517. MEDLINE

31 Dowell F.J., Martin W., Dominiczak A.F., Hamilton C.A. Decreased basal despite enhanced agonist-stimulated effects of nitric oxide in 12-week-old stroke-prone spontaneously hypertensive rat. Eur J Pharmacol 1999; 379: 175, 175. MEDLINE

32 Mian K.B., Martin W. Differential sensitivity of basal and acetylcholine-stimulated activity of nitric oxide to destruction by superoxide anion in rat aorta. Br J Pharmacol 1995; 115: 993, 993. MEDLINE

33 Vergnani L.et al. . Effect of native and oxidized low-density lipoprotein on endothelial nitric oxide and superoxide production: key role of L-arginine availability. Circulation 2000; 101: 1261, 1261. MEDLINE

34 Cosentino F.et al. . Tetrahydrobiopterin alters superoxide and nitric oxide release in prehypertensive rats. J Clin Invest 1998; 101: 1530, 1530. MEDLINE

35 Wink D.A., Mitchell J.B. Chemical biology of nitric oxide: insights into regulatory, cytotoxic, and cytoprotective mechanisms of nitric oxide. Free Rad Biol Med 1998; 25: 434, 434. MEDLINE

36 Squadrito G.L., Pryor W.A. Oxidative chemistry of nitric oxide: the roles of superoxide, peroxynitrite and carbon dioxide. Free Rad Biol Med 1998; 25: 392, 392. MEDLINE

37 MacMillan-Crow L.A., Thompson J.A. Tyrosine modifications and inactivation of active site manganese superoxide dismutase mutant (Y34F) by peroxynitrite. Arch Biochem Biophys 1999; 366: 82, 82. Article MEDLINE

38 Laursen J.B.et al. . Endothelial regulation of vasomotion in apoE-deficient mice: implications for interactions between peroxynitrite and tetrahydrobiopterin. Circulation 2001; 103: 1282, 1282. MEDLINE

39 Milstien S., Katusic Z. Oxidation of tetrahydrobiopterin by peroxynitrite: implications for vascular endothelial function. Biochem Biophys Res Commun 1999; 263: 681, 681. Article MEDLINE

40 Liochev S.I., Fridovich I. The role of O2- in the production of HO: in vitro and in vivo. Free Rad Biol Med 1994; 16: 29, 29. MEDLINE

41 Teixeira H.D., Schumacher R.I., Meneghini R. Lower intracellular hydrogen peroxide levels in cells overexpressing CuZn-superoxide dismutase. Proc Natl Acad Sci USA 1998; 95: 7872, 7872. MEDLINE

42 Zanetti M.et al. . Gene transfer of manganese superoxide dismutase reverses vascular dysfunction in the absence, but not in the presence of atherosclerotic plaque. Hum Gene Ther 2001; 12: 1407, 1407. Article MEDLINE

43 Gorecki M.et al. . Recombinant human superoxide dismutases: production and potential therapeutical uses. Free Rad Res Commun 1991; 12-13: 401, 401.

44 Karlsson K.et al. . Pharmacokinetics of extracellular-superoxide dismutase in the vascular system. Free Rad Biol Med 1993; 14: 185, 185. MEDLINE

45 Karlsson K., Marklund S.L. Extracellular superoxide dismutase in the vascular system of mammals. Biochem J 1988; 255: 223, 223. MEDLINE

46 Stralin P., Karlsson K., Johansson B.O., Marklund S.L. The interstitium of the human arterial wall contains very large amounts of extracellular superoxide dismutase. Arterioscler Thromb Vasc Biol 1995; 15: 2032, 2032. MEDLINE

47 Landmesser U.et al. . Vascular extracellular superoxide dismutase activity in patients with coronary artery disease: relation to endothelium-dependent vasodilation. Circulation 2000; 101: 2264, 2264. MEDLINE

48 Wilkinson G.W., Akrigg A. Constitutive and enhanced expression from the CMV major IE promoter in a defective adenovirus vector. Nucleic Acids Res 1992; 20: 2233, 2233. MEDLINE

49 Baker A.H., Zaltsman A.B., George S.J., Newby A.C. Divergent effects of tissue inhibitor of metalloproteinase-1, -2, or -3 overexpression on rat vascular smooth muscle cell invasion, proliferation, and death in vitro. TIMP-3 promotes apoptosis. J Clin Invest 1998; 101: 1478, 1478. MEDLINE

50 Nicklin S.A., Baker A.H. Simple methods for preparing recombinant adenoviruses for high-efficiency transduction of vascular cells. Baker AH (eds); Vascular Disease: Molecular Biology and Gene Therapy Protocols Humana Press Inc, 1999, pp 271-283.

Figure 1 Western blot analysis of SOD expression in infected HUVECs. HUVECs were infected at a range of multiplicities of infection (MOI) up to 300 with either AdMnSOD, AdECSOD or Adgal. The media was replaced and the cells harvested 48 h after infection. Heparin (100 IU/ml) was added to half of the wells infected with AdECSOD and their corresponding controls before cell harvesting. Protein extracts were subjected to polyacrylamide gel electrophoresis on a 10% polyacrylamide gel and transferred to Hybond-P membranes. (a) Increasing the multiplicity of AdMnSOD infection resulted in a corresponding increase in MnSOD protein expression. (b) Increasing the multiplicity of AdECSOD infection resulted in a similar increase in ECSOD protein both in heparin-treated and -untreated cells. (c) Heparin treatment of AdECSOD-infected cells resulted in a large increase in ECSOD protein release into the conditioned media. Uninfected and Adgal-infected cells did not show detectable levels of MnSOD or ECSOD.

Figure 2 Immunofluorescent labelling of MnSOD and ECSOD in infected cells. Immunofluorescent FITC-labelling did not detect either SOD protein in either uninfected or Adgal-infected cells (a and d). FITC-labelling of ECSOD demonstrated increased ECSOD expression in AdECSOD infected cells (b and e). Similarly, FITC labelling of MnSOD revealed a much greater presence of MnSOD in AdMnSOD-infected endothelial cells (c, f and h). A fluorescent mitochondrial-specific probe (MitoTracker) was used to label cell mitochondria (g). MnSOD was colocalised to the mitochondria by confocal microscopy (g-i).

Figure 3 SOD activity in infected HUVECs. SOD activity was assessed by infection of HUVECs at increasing MOIs using the Adgal, AdMnSOD and AdECSOD vectors. Cells were harvested 48 h later, lysed and assayed using a SOD assay kit (Calbiochem) based on the oxidation of 5,6,6a,11b-tetrahydro-3,9,10-trihydroxybenzo[c]fluorene (THBF). SOD activity was measured as the rate of increase in optical density at 525 nm. All samples were performed in triplicate. HUVECs infected with both AdMnSOD (diagonally lined bars) and AdECSOD (horizontally lined bars) resulted in a dose-dependent increase in SOD activity in infected cells. Adgal infection (white bar) did not significantly increase SOD activity beyond the basal level of uninfected cells (filled bar). All SOD activities are expressed as a mean ± s.e.m., measured in arbitrary units per mg of protein.

Figure 4 Immunohistochemical analysis of MnSOD and ECSOD expression in infused carotid arteries. Carotid arteries were locally infected in vivo with 30 l of 8 ´ 10⁹ p.f.u./ml of AdMnSOD or AdECSOD. After 24 h, the arteries were removed, fixed and immunocytochemical staining was performed with either MnSOD- or ECSOD-specific antibodies and counterstained with haematoxylin. Increased MnSOD expression (brown DAB staining) (b) and increased ECSOD expression (d) was detected in the endothelium and adventitia of the infused rat carotid arteries when compared with the contralateral uninfused carotid arteries (a and c).

Figure 5 NO availability in infected rat carotid arteries. Carotid arteries were harvested 24 h after infection and NO availability was assessed by the difference in cumulative dose-response curves to increasing doses of phenylephrine both in the presence and absence of L-NAME. The contralateral uninfused vessels () were used as controls. Results are expressed as a mean ± s.e.m. RAd66 () and AdMnSOD () infusion did not result in a significant improvement in NO availability in the SHRSP (n = 6). However, ECSOD gene transfer () significantly improved NO availability.

Figure 6 Hypothesis for actions of SOD isoforms in the vessel wall. AdMnSOD-infected vascular cells only express MnSOD within mitochondria. This limited area of MnSOD expression results only in local protection against superoxide and did not produce a detectable effect on NO availability in this model of endothelial dysfunction. However, AdECSOD-infected cells release ECSOD both within the cytoplasm and extracellularly to the heparan sulphate proteoglycans of the cell surface and extracellular matrix. The larger sphere of influence of ECSOD may explain the significant increase in NO availability observed.