Superoxide dismutase mimetic M40403 improves endothelial function in apolipoprotein(E)-deficient mice

Abstract

1. Overproduction of superoxide anions in the vascular wall contributes to endothelial dysfunction in vascular disease. A superoxide-generating reduced β-nicotinamide adenine dinucleotide phosphate (NADPH) oxidase has recently been identified as a major source of oxidative radicals in vascular tissues. We studied the effects of a synthetic manganese-containing superoxide dismutase (SOD) mimetic, M40403, on NADPH oxidase-dependent superoxide generation and on endothelial dysfunction.

2. In rat aortic smooth muscle cells, NADPH (100 μM) markedly stimulated superoxide production as detected by lucigenin (5 μM)-enhanced chemiluminescence. M40403 reduced NADPH oxidase-dependent superoxide production in a concentration-dependent manner, with IC₅₀ being 31.6 μM. In contrast, native Cu/Zn SOD (up to 300 U ml⁻¹) had no effect. Angiotensin II (100 nM) increased the NADPH oxidase activity by 70%, and treatment with M40403 (10 μM) reduced this increased superoxide to the control level.

3. In aortae from apolipoprotein(E)-deficient mice (apoE⁰) with hyperlipidemia and atherosclerosis, superoxide production is largely derived from NADPH oxidase. The attenuation of endothelial nitric oxide vasodilator function parallels the increase in vascular superoxide production at different stages of the disease. Acute incubation of such aortic rings with M40403 significantly suppressed superoxide production and improved endothelium-dependent vasorelaxation to a level comparable to that in wildtype control mice.

4. In summary, the cell-permeable SOD mimetic M40403 was found to reverse endothelial dysfunction in apoE⁰ aorta ex vivo by decreasing NADPH oxidase-dependent superoxide levels. The advantages of synthetic SOD mimetics over the native Cu/Zn SOD enzyme, such as greater cell permeability and stability, confer significant therapeutic potential in vascular disease.

Introduction

Increased production of superoxide anions (O₂^−•) in vascular tissues has been implicated in many cardiovascular diseases such as hypertension, hyperlipidemia and atherosclerosis (Mugge et al., 1994; Miller et al., 1998; Kerr et al., 1999). Although reactive oxygen species may have important roles in various intracellular signal transduction pathways (Kunsch & Medford, 1999), excess superoxide molecules can cause multiple pathophysiological consequences. Superoxide rapidly inactivates endothelium-derived nitric oxide (NO), which is a predominant endogenous vasodilator, leading to decreased NO bioavailability and elevated blood pressure (Zicha et al., 2001). Superoxide and the product of its reaction with NO, peroxynitrite, promote lipid peroxidation that contributes to atherosclerotic lesion formation (White et al., 1994). Moreover, impairment of NO function by superoxide may result in vascular smooth muscle cell proliferation and migration and expression of a variety of proinflammatory molecules in vascular cells, all of which are operative mechanisms of atherosclerosis and restenosis (Dusting et al., 1998).

Recently, a superoxide-producing reduced β-nicotinamide adenine dinucleotide phosphate (NAD(P)H) oxidase has been identified in vascular tissues (Griendling et al., 2000). This enzyme is composed of two cytosolic subunits p47phox and p67phox, a cell membrane-bound cytochrome b558 which consists of gp91phox (or its homologues) and p22phox, and a small G protein rac. Upon assembly of these subunits in the membrane, this enzyme generates a superoxide by one-electron reduction of oxygen via its gp91phox subunit using NADPH or reduced β-nicotinamide adenine dinucleotide (NADH) as the electron donor (Griendling et al., 2000). The activity of this enzyme can be increased by angiotensin II (AngII) (Griendling et al., 1994), growth factors (Marumo et al., 1997) or cytokines (De Keulenaer et al., 1998). Emerging evidence has suggested that this NAD(P)H oxidase is the major source of superoxide in vascular tissues (Guzik et al., 2000; Paravicini et al., 2002) and generation of pathological amounts of superoxide by this enzyme has been implicated in cardiovascular diseases such as atherosclerosis (Yokoyama et al., 2000) or hypertension (Morawietz et al., 2001).

Metabolism of superoxide largely depends on its dismutation to hydrogen peroxide by endogenous superoxide dismutase (SOD). It has been shown that in vivo treatment with modified SOD enzyme improved endothelium-dependent vasorelaxations in atherosclerotic rabbits (Mugge et al., 1991). The therapeutic potential of native SOD enzyme is, however, restricted because it is a large peptide with limited stability and permeability, and several cell-permeable small molecule SOD mimetics have been synthesized. M40403 is a low-molecular weight manganese-containing SOD mimetic, which has marked protective effects in inflammation and septic shock by suppressing superoxide generation (Salvemini et al., 1999;2001; Macarthur et al., 2000). However, the actions of synthetic SOD mimetics in the context of cardiovascular diseases such as athero-sclerosis and hypertension, and in particular their interactions with the vascular NADPH oxidase, have not been well defined. Using cultured rat aortic smooth muscle cells (RASMC) and apolipoprotein(E)-deficient (apoE⁰) mice, which developed hyperlipidemia, hypertension, atherosclerosis and endothelial dysfunction (Plump et al., 1992; Yang et al., 1999; Jiang et al., 2001), we investigated (1) whether synthetic SOD mimetic may functionally antagonize the NADPH oxidase activity, and (2) the role of superoxide production in the endothelial NO dysfunction in apoE⁰ mice and the efficacy of M40403 ex vivo.

Methods

Animals and aortic tissue preparation

All animal studies were carried out in accordance with the guidelines of Howard Florey Institute Animal Ethics Committee and National Health and Medical Research Council of Australia. Male apoE⁰ mice with >99% C57BL/6J background were purchased from Animal Resource Centre (Western Australia) at 4 weeks of age and maintained with either a normal diet or a high-fat diet containing 0.15% cholesterol (Harlan Teklad, Madison, WI, U.S.A.). These animals developed hyperlipidemia and atherosclerosis (Jiang et al., 2001). Age-matched C57BL/6J mice were used as wildtype controls (WT). Animals were killed at 16 or 28 weeks of age by decapitation under general anesthesia. Abdominal aortae were isolated immediately for vascular reactivity and lucigenin-enhanced chemiluminescence assay.

Aortic smooth muscle cell (SMC) culture

SMCs were isolated from thoracic aorta of male Sprague–Dawley rats by enzymatic digestion with collagenase I (1 mg ml⁻¹), elastase (0.5 mg ml⁻¹) and trypsin (1.25 mg ml⁻¹). Cells were maintained in Dulbecco's modification of Eagle's medium (DMEM) supplemented with 10% fetal calf serum in a humidified incubator with 5% CO₂ at 37°C. Cells between passages 4 and 10 was used. The purity of the cells was determined as >95% by antismooth muscle α-actin immunocytochemistry.

Lucigenin-enhanced chemiluminescence assay

Lucigenin-enhanced chemiluminescence assay was based on the method described by Ohara et al. (1993) with some modifications. Aortic segments of about 0.5 cm long were equilibrated in Krebs–HEPES buffer (KHB) (pH 7.4) comprising (mM): NaCl 98.0, KCl 4.7, NaHCO₃ 25.0, MgSO₄ 1.2, KH₂PO₄ 1.2,CaCl₂ 2.5, D-glucose 11.1 and HEPES-Na 20.0 for 60 min. The chemiluminescence was detected with a microplate scintillation counter (Topcount model 9912, Packard) running in single-photon-count mode. Polystyrene 96-well plate (OptiPlate-96 Packard) was loaded with KHB (300 μl in each well) containing 5 μM lucigenin, with or without NADPH (100 μM). The specificity of this assay was confirmed by the concentration-dependent generation of superoxide by xanthine plus xanthine oxidase (X/XO) and its inhibition by Cu/Zn SOD.

To measure superoxide in SMCs, cells were suspended with collagenase I (1 mg ml⁻¹) dissolved in KHB, pelleted by centrifugation (1000 r.p.m. × 4 min) and then resuspended in 0.5 ml KHB containing lucigenin (5 μM) with or without NADPH. Each well was loaded with 300 μl of cell suspension. For various drug treatments, drugs were added to the culture plate at least 40 min before cell collection. To eliminate the influence of endogenous Cu/Zn SOD activity, all aortic segments and SMCs were treated with diethyldithiocarbamic acid (DETCA) (3 mM) for 60 min prior to chemiluminescence assay, unless otherwise mentioned. All drugs, except for DETCA, were also present in the cell suspension throughout the counting. To measure superoxide generation in cell homogenates, confluent SMCs grown in three six-well plates were scraped into 9 ml cold KHB containing 10 μg ml⁻¹ leupeptin. Cells were homogenized by sonication on ice and centrifuged at 1000 r.p.m. for 3 min. Chemiluminescence was measured in the supernatant as described in the cell suspension. To measure superoxide generated by X/XO, the assay was carried out in a reaction mixture of 300 μl containing xanthine (100 μM), xanthine oxidase (0.03 U ml⁻¹) and lucigenin (5 μM) in PBS with or without scavenging drugs. The reaction was initiated by the addition of xanthine oxidase into the mixture. As the superoxide signal produced by X/XO was transient, all counting commenced 1 min after xanthine oxidase addition.

Vascular reactivity study

Vascular reactivity was studied in aortic rings (∼1.5 mm long) using a four-channel myograph as described previously (Jiang et al., 2001). Briefly, rings were mounted onto two parallel stainless-steel pins through the lumen and placed in a chamber containing 6 ml physiological salt solution, which was maintained at 36±1°C and gassed with 5% CO₂ and 95% O₂. After a 20 min period of equilibration, the resting tension was raised stepwise to 5 mN; then a further equilibration period of 20 min was given. The tissues were maximally contracted with the thromboxane analogue 9,11-dideoxy-9α,11α-methanoepoxyprostaglandin F_2α (U46619) (1 μM). Relaxations were induced after the tone had been raised with U46619 to 40–50% of the maximum. In some experiments, M40403 was added 30 min prior to the relaxation experiment and present throughout.

Drugs used

The drugs used were β-Nicotinamide adenine dinucleotide phosphate (NADPH, ICN), 17-octadecynoic acid (17-ODYA), 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ), U46619 (Calbiochem, U.S.A.), acetylcholine perchlorate (ACh, BDH, England), allopurinol, AngII (Auspep, Australia), diethyldithiocarbamic acid (DETCA) (ICN), 4,5-dihydroxy-1,3-benzenedisulfonic acid (disodium salt, tiron), diphenyleneio-donium chloride (DPI), 4-hydroxy-2,2,6,6-tetramethylpiperidinyloxy (tempo, Aldrich), indomethacin, losartan (Merck), leupeptin (ICN), L-phenylephrine hydrochloride, M40403 (MetaPhore Pharmaceuticals, St Louis, MO, U.S.A.), N^G-nitro-L-arginine methyl ester hydrochloride (L-NAME), Cu/Zn SOD, X, XO. Drugs with unspecified source are from Sigma.

Data and statistics

The superoxide production level was expressed as count per second (CPS). Results were normalized to per mg wet tissue (aorta) or to a concentration of 10⁶ cell ml⁻¹ (cell suspension). Data were presented as mean±standard error of the mean (s.e.m.). One-way analysis of variance (one-way ANOVA) followed by Tukey's test (for multiple comparisons) or Student's t-test (for two groups) was used as appropriate for statistical analysis. A value of P<0.05 was regarded as significant. The regression analysis of dose–response curves and calculation of IC₅₀ (the molar concentration required to produce 50% inhibition of the enzyme activity) were performed using Prism software (GraphPad).

Results

Superoxide-scavenging efficacy of M40403

The in vitro superoxide-scavenging efficacy of M40403 was assessed in a biochemical system comprising X (100 μM) plus XO (0.03 U ml⁻¹) (X/XO). The chemiluminescence signal generated by X/XO was abolished by Cu/Zn SOD (Figure 1a). M40403 suppressed the superoxide generated by X/XO in a concentration-dependent manner, the IC₅₀ being 6.3 μM. This is in contrast to the efficacies of the nonspecific superoxide radical scavengers tempo and tiron under the same experimental conditions, the IC₅₀ values being 398.1 and 794.3 μM, respectively. The regression analysis of the dose–response curves of M40403, tempo and tiron are shown in Figure 1b.

Figure 1.

(a) Effects of Cu/Zn SOD on superoxide generated with xanthine (100 μM) plus xanthine oxidase (0.03 U ml⁻¹) (X/XO) detected by lucigenin-enhanced chemiluminescence. Results were expressed in CPS as mean±s.e.m., ^*P<0.05, n=4. (b) Regression analysis of the inhibitory effects of M40403, tempo and tiron on superoxide production in the X/XO system. (c) Regression analysis of the effect of M40403 on NADPH (100 μM)-stimulated superoxide production in RASMCs pretreated with DETCA (3 mM). In (b) and (c) each point represents the average of four measurements.

Effects of M40403 on NADPH-induced superoxide in RASMCs

In cultured RASMCs, superoxide could not be detected either under resting conditions or after inactivation of endogenous SOD by treatment with DETCA (3 mM). In DETCA-treated cells, NADPH (100 μM) stimulated superoxide production; the absolute magnitude (CPS) of the chemiluminescence signal is similar to that generated with X/XO. The superoxide production was inhibited by the NADPH oxidase inhibitor DPI, but not by the nitric oxide synthase (NOS) inhibitor L-NAME, the XO inhibitor allopurinol, the cyclooxygenase inhibitor indomethacin or the cytochrome P450 inhibitor 17-ODYA (Figure 2a). NADPH-stimulated superoxide production was suppressed by M40403 in a concentration-dependent manner, with IC₅₀ being 31.6 μM (Figure 1c). Pretreatment with AngII for 4 h significantly increased NADPH oxidase activity, an effect that was blocked by the AngII type-1 (AT₁) receptor antagonist losartan (Figure 2b). Incubation with DPI and M40403 also suppressed NADPH-stimulated superoxide in AngII-treated cells (Figure 2b).

Figure 2.

(a) The effects of DPI (D, 10 μm), L-NAME (N, 200 μM), allopurinol (A, 100 μM), indomethacin (I, 30 μM) and 17-ODYA (O, 10 μM) on NADPH-stimulated superoxide production in RASMCs pretreated with DETCA; (b) the effects of AngII (100 nM) alone or in the presence of losartan (L, 10 μM), M40403 (M, 10 μM) and DPI (D, 10 μM) on NADPH-stimulated superoxide production in RASMCs. Results are converted to percentage of control (C). Data are mean±s.e.m. ^*P<0.05 vs control; ^†P<0.05 vs AngII, n=3–10.

To elucidate whether the native Cu/Zn SOD has similar effects as M40403 on NADPH-stimulated superoxide in RASMCs, we measured the superoxide production in cell suspensions that were not pretreated with DETCA. Treatment with Cu/Zn SOD at 100 or 300 U ml⁻¹ had no effect on the superoxide production in intact cells (Figure 3a). In SMC homogenates, the superoxide production in the presence of NADPH was not affected by Cu/Zn SOD treatment either, whereas M40403 reduced the superoxide level in a concentration-dependent manner (Figure 3b).

Figure 3.

(a) Effects of Cu/Zn SOD on NADPH-stimulated superoxide in intact RASMCs (without DETCA pretreatment), n=3. (b) Effects of SOD and M40403 on NADPH-stimulated superoxide in RASMC homogenates (without DETCA pretreatment). Results are expressed as percentage of control (C) and data are mean±s.e.m. ^*P<0.05 vs control, n=5–6.

Superoxide generation in mouse aorta and effects of M40403

In aortic rings from 28-week-old WT and apoE⁰ mice (fed a normal diet) pretreated with DETCA (3 mM), superoxide could be detected under basal conditions. Inclusion of NADPH (100 μM) further stimulated this superoxide production, but there was no difference in the basal or NADPH-stimulated superoxide between WT and apoE⁰ (Table 1). To clarify the enzymatic source of the superoxide anion in the aorta, we examined the effects of various enzyme inhibitors in aortic rings from apoE⁰ mice. Both the basal and NADPH-stimulated superoxide productions were blocked by DPI (5 μM), but not significantly affected by L-NAME (200 μM), allopurinol (100 μM), indomethacin (30 μM) or 17-ODYA (10 μM) (Figure 4a–d). Similar as in RASMCs, Cu/Zn SOD treatment had no significant effect on the superoxide level in either the absence or presence of NADPH in these aortic rings (not pretreated with DETCA) (Figure 4e,f).

Table 1.

Effects of M40403 (10 μM) on basal and NADPH-stimulated superoxide production in mouse abdominal aorta

	Superoxide (CPS mg−1)	n
−NADPH
WT control	41±8	10
apoE0/normal diet	53±10	13
apoE0/high-fat diet	102±12ab	8
apoE0/normal diet+M40403 (10 μM)	16±3b	9
apoE0/high-fat diet+M40403	15±2c	7
+NADPH (100 μM)
WT control	311±125	10
apoE0/normal diet	180±43	13
apoE0/high-fat diet	1198±245ab	7
apoE0/normal diet+M40403	43±11b	9
apoE0/high-fat diet+M40403	112±12c	7

All tissues were pretreated with DETCA.

^aP<0.05 vs WT;

^bP<0.05 vs apoE⁰/normal diet;

^cP<0.05 vs apoE⁰/high-fat diet.

Figure 4.

Effects of L-NAME (N, 200 μM), allopurinol (A, 100 μM), indomethacin (I, 30 μM) and 17-ODYA (O, 10 μM) (a,b), DPI (D, 5 μM) (c,d) and Cu/Zn SOD (e,f) on the superoxide production in aortic rings from 28-week-old apoE⁰ (normal diet fed) mice, in the absence (a,c,e) or presence (b,d,f) of NADPH (100 μM) stimulation. Tissues were pretreated with DETCA (3 mM) except for (e) and (f). Results are expressed as CPS per mg wet tissue. Data are mean±s.e.m. C, control; ^*P<0.05 vs control, n=4.

In aortic rings from 28-week-old apoE⁰ mice (fed a high fat diet) pretreated with DETCA, both the basal and NADPH-stimulated superoxide productions were significantly elevated as compared to WT control and apoE⁰ on a normal diet (Table 1). Acute ex vivo incubation with M40403 (10 μM) significantly reduced the superoxide levels in these tissues (Table 1).

Vascular superoxide generation and endothelial NO function

We have previously shown that in the aorta of both WT and apoE⁰ mice, endothelium-dependent relaxations induced by ACh could be blocked by the NOS inhibitor L-NAME (Jiang et al., 2001), suggesting that this response is mediated entirely by endothelial NO. In pilot studies, we found that ACh-induced relaxations in the aorta from 28-week-old apoE⁰ (normal diet) were not different from those in WT control (data not shown). In contrast, ACh-induced vasorelaxations were significantly reduced in tissues of high-fat diet-fed apoE⁰ mice as compared to WT (see Figure 6), while the relaxations induced by the NO donor S-nitroso-N-acetylpenicillamine were not significantly different (Jiang et al., 2001). These data suggest that the increased vascular superoxide level in apoE⁰ mice (high fat) compromises endothelial NO function. We further studied endothelial function and vascular superoxide generation in apoE⁰ mice (high fat) at different ages (16 and 28 weeks). There was an inverse association between the ACh-induced vasorelaxation and the vascular superoxide level (without NADPH) or the NADPH oxidase activity (with NADPH) (Figure 5). A correlation analysis within tissue was not possible because the superoxide level and vascular reactivity cannot be measured in a single tissue.

Figure 6.

Acute effects of M40403 (10 μM) on ACh-induced vasorelaxations in the abdominal aorta from 28-week-old WT (open bars) and apoE⁰ (high fat, closed bars) mice. Relaxations are expressed as the percentage reduction of U46619-induced tone. Data are mean±s.e.m. ^*P<0.05 vs WT; ^†P<0.05 vs without M40403, n=7–13.

Figure 5.

Dynamic changes in the endothelial NO function (assessed as ACh-induced vasorelaxation) and the vascular superoxide production in the absence (a) or presence (b) of exogenous NADPH (100 μM) in isolated abdominal aorta from apoE⁰ (high fat diet treated) mice at 16 or 28 weeks of age (n=4–8). There is an obvious trend that the increase in vascular superoxide level or NADPH oxidase activity is accompanied by the deterioration of ACh-induced vasorelaxation. At 28 weeks, both of the endothelium-dependent relaxation and the superoxide production/NADPH oxidase activity are significantly different from those at 16 weeks (P<0.05).

Effects of M40403 on endothelial function

In aortic rings from 28-week-old apoE⁰ (high fat) mice, the endothelium-dependent relaxations induced by ACh (30–300 nM) were significantly reduced as compared to age-matched WT controls (Figure 6). Acute incubation of the apoE⁰ aorta with M40403 (10 μM) improved ACh-induced relaxations to a level comparable to those in WT (Figure 6). M40403 at the same concentration had no significant effect in aortic rings from WT mice (Figure 6). To confirm that superoxide affects the basal bioavailability of endothelial NO, the vasoconstrictor response induced by phenylephrine (1 μM) was studied in the presence and absence of superoxide generated exogenously. X/XO significantly increased phenylephrine-induced contraction, and this was abolished by M40403. Blockade of the NO-cyclic GMP pathway with L-NAME plus the soluble guanylate cyclase inhibitor ODQ markedly increased contractions, and X/XO had no further effect (Figure 7).

Figure 7.

Effect of basally released endothelial NO, as assessed by contractile responses to phenylephrine (1 μM), in normal aortic rings from WT mice. Exogenous superoxide was generated by xanthine (10 μM) plus xanthine oxidase (0.01 U ml⁻¹) combination (X/XO). All tissues were preincubated with DETCA (3 mM). Some tissues were pretreated with M40403 or L-NAME plus ODQ. Contractions are expressed as percentage of the responses induced by U46619 (1 μM). Numbers in the bar indicate n of experiments. Data are mean±s.e.m. ^*P<0.05.

Discussion

NADPH oxidase is expressed in vascular smooth muscle. The expression of the NADPH oxidase subunits p22phox, p47phox and the gp91phox homologues nox1 and nox4 has been identified in human and rat smooth muscle cells (Ushio-Fukai et al., 1996; Patterson et al., 1999; Lassegue et al., 2001). We have now showed that NADPH stimulated superoxide generation in RASMC and this was totally blocked by the flavin-binding NADPH oxidase inhibitor DPI, but was not affected by inhibitors of other potential superoxide-generating enzymes, including NOS, XO, cyclooxygenase or cytochrome P450 (Munzel et al., 1999). In the absence of substrate (NADPH), superoxide production was undetectable even after the inactivation of endogenous Cu/Zn SOD. Together with the observation that supplementation with NADPH dramatically enhanced the superoxide level in mouse aortic rings, it is suggested that physiologically, the availability of substrate appears to be a key factor that limits superoxide production by this enzyme.

M40403, whose superoxide-scavenging efficacy is much higher than the nonspecific superoxide scavengers tiron and tempo as determined in the cell-free system, effectively suppressed superoxide production in RASMC stimulated by NADPH, whereas exogenous native Cu/Zn SOD had little effect at concentrations that are high enough to abolish X/XO-produced superoxide of a similar level. Similar findings have also been obtained in isolated rings of apoE⁰ mouse aorta. These observations are in agreement with several previous studies showing that native SOD had limited efficacy in reversing redox-sensitive cellular events (Arai et al., 1998; Pueyo et al., 2000), while polyethylene-glycolated SOD (Mugge et al., 1991), which has a longer half-life and higher cellular permeability, and adenovirus-mediated SOD gene-transfer (Arai et al., 1998; Zanetti et al., 2001), which induced intracellular SOD protein expression, are effective. The contrast between the effects of SOD and M40403 indicates that the smooth muscle NADPH oxidase may generate superoxide intracellularly, and the poor cell permeability of SOD limits its direct therapeutic efficacy in cardiovascular disease. However, the lack of effect with Cu/Zn SOD in contrast to M40403 on NADPH-stimulated superoxide was also found in smooth muscle cell homogenates, where SOD has free access to the cytosolic components. The mechanism of this ineffectiveness of SOD is unknown. In neutrophils, it has been proposed that the membrane-bound subunits (gp91phox and p22phox) reside in the membrane of certain intracellular compartments, and the intracellularly produced superoxide is released to the extracellular environment through the orifice formed by fusion of the compartments with the plasma membrane (Kobayashi et al., 2001). Although the exact process of superoxide generation by the NADPH oxidase in smooth muscle cells is unclear, it is conceivable that subcellular compartmentalization of the enzyme may hinder the access of Cu/Zn SOD to the generated superoxide anions.

Several studies have demonstrated that proatherogenic stimuli such as AngII and proinflammatory cytokines may increase the activity of NADPH oxidase in vascular cells. For instance, in vascular SMC, AngII (Lassegue et al., 2001) and tumor necrosis factor-α (De Keulenaer et al., 1998) stimulate superoxide generation, and these effects appear to be associated with upregulated expression of the subunits of NADPH oxidase. In this study, we found that preincubation with AngII produced about two-fold increase in NADPH oxidase activity, and this effect was mediated by the AT₁ receptor. To further test the efficacies of M40403 in pathophysiological situations, we examined the effects of M40403 in AngII-treated cells. Coincubation with M40403 greatly reduced the increase in superoxide production by AngII, comparable with the effect of losartan. Taken together, our results indicate that M40403 can functionally antagonize the NADPH oxidase activity in smooth muscle cells under both normal and pathological conditions.

Endothelial NO function is compromised in apoE⁰ mice, and this is consistent with observations in other species that endothelial dysfunction is a prominent manifestation of early vascular changes in hyperlipidemia and atherosclerosis (Dusting et al., 1998). The role of superoxide in the pathogenesis of endothelial dysfunction in apoE⁰ mice has been investigated by several groups. Laursen et al. (2001) showed that there is an increase in superoxide production in apoE⁰ aorta. Moreover, d'Uscio et al. (2001) demonstrated that the endothelial dysfunction is partly reversed by another SOD mimetic Mn(III)tetrakis(4-benzoic acid)porphyrin (MnTBAP) (d'Uscio et al., 2001), suggesting that superoxide is involved in the development of endothelial dysfunction. We have now analyzed the relation between the time courses of changes in vascular superoxide production and the endothelial function in this model. Clearly, the decrease in endothelial function is closely associated with a concomitant elevation in vascular superoxide production, and decreased endothelial function was accompanied by an increase in the enzymatic activity of NADPH oxidase. All of these data indicate that NADPH oxidase-derived superoxide anions in the vascular wall have an important role in the impaired endothelial NO function in apoE⁰ mice. It should be noted that the cellular origin of superoxide in intact blood vessel rings might be diverse. In apoE⁰ aorta, a potential source of superoxide is phagocytic NADPH oxidase in macrophages (foam cells), which comprise the major cellular component of atherosclerotic lesions. Although the present study did not distinguish the origins of superoxide in apoE⁰ vessels, recent findings that the superoxide production in vascular tissues is impaired in p47phox knockout mice but not altered in gp91phox (a subunit present in phagocytes and endothelial cells but not smooth muscle cells) knockouts suggest that cells other than phagocytes may also be important in inducing oxidative stress in the vessel wall (Barry-Lane et al., 2001; Souza et al., 2001).

Acute treatment with M40403 improved the endothelium-dependent relaxation in apoE⁰ aorta to a level comparable to that in wildtype controls. The specificity of the effect of M40403 on superoxide is confirmed by the observation that exogenously generated superoxide diminishes the basal release of NO in normal vessels (indicated by increased contractions to phenylephrine and its sensitivity to the blockade of NO pathway), and this action is also prevented by M40403 pretreatment. Such observations are consistent with those reported by Laursen et al. (2001) showing that acute treatment with liposome-entrapped SOD normalized the endothelium-dependent relaxation in apoE⁰ aorta. The mechanisms of the impairment in endothelial function in the apoE⁰ model are controversial and complex. Apart from the inactivation of NO by superoxide, it has been suggested that the product of the reaction between these two species, peroxynitrite, may facilitate the degradation of the NOS cofactor tetrahydrobiopterin, leading to reduced NO production (Laursen et al., 2001). Whatever the mechanism, suppression of vascular superoxide generation would confer a protective effect on endothelial NO function. On the other hand, others have reported that endothelial NOS (eNOS) activity is reduced in high-fat-fed apoE⁰ mice aorta as compared to WT mice, whereas the protein level of eNOS is not changed. This alteration was not observed in our study or that of d'Uscio (d'Uscio et al., 2001), where the impairment in endothelium-dependent vasorelaxation was largely prevented by acute superoxide dismutation. The reason for this discrepancy is unclear.

In summary, we have demonstrated that the manganese-based, cell-permeable SOD mimetic M40403 improved the endothelial NO function in apoE⁰ aorta via a functional antagonism of the vascular NADPH oxidase activity. Clearly, synthetic SOD mimetics have therapeutic potential for improving endothelial function in cardiovascular disorders associated with elevated vascular oxidative stress.

Abbreviations

ACh

acetylcholine perchlorate

AngII

angiotensin II

apoE⁰

apolipoprotein(E)-deficient mouse

CPS

count per second

DETCA

diethyldithiocarbamic acid

tiron

4,5-dihydroxy-1,3-benzenedisulfonic acid

DMEM

Dulbecco's modification of Eagle's medium

DPI

diphenyleneiodonium chloride

KHB

Krebs–HEPES buffer

L-NAME

N^G-nitro-L-arginine methyl ester hydrochloride

NADPH

reduced β-nicotinamide adenine dinucleotide phosphate

NO

nitric oxide

NOS

nitric oxide synthase

ODQ

1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one

17-ODYA

17-octadecynoic acid

one-way ANOVA

one-way analysis of variance

RASMC

rat aortic smooth muscle cells

s.e.m.

standard error of the mean

SOD

superoxide dismutase

tempo

4-hydroxy-2,2,6,6-tetramethylpiperidinyloxy

U46619

9,11-dideoxy-9α,11α-methanoepoxyprostaglandin F_2α

X/XO

xanthine plus xanthine oxidase