NAD(P)H oxidase/nitric oxide interactions in peroxisome proliferator activated receptor (PPAR)α- mediated cardiovascular effects

Abstract

Activation of peroxisome proliferator activated receptor (PPAR)α and its protective role in cardiovascular function has been reported but the exact mechanism(s) involved is not clear. As we have shown that PPARα ligands increased nitric oxide (NO) production and cardiovascular function is controlled by a balance between NO and free radicals, we hypothesize that PPARα activation tilts the balance between NO and free radicals and that this mechanism defines the protective effects of PPARα ligands on cardiovascular system. Systolic blood pressure (SBP) was greater in PPARα knockout (KO) mice compared with its wild type (WT) litter mates (130 10 mmHg versus 107 4 mmHg). L-NAME (100 mg/L p.o.), the inhibitor of NO production abolished the difference between PPARα KO and WT mice. In kidney homogenates, tissue lipid hydroperoxide generation was greater in KO mice (11.8 1.4 pM/mg versus 8.3 0.6 pM/mg protein). This was accompanied by a higher total NOS activity (46 6%, p < 0.05) and a 3 fold greater Ca2+-dependent NOS activity in kidney homogenates of untreated PPARα WT compared with the KO mice. Clofibrate, a PPARα ligand, increased NOS activity in WT but not KO mice. Bezafibrate (30 mg/kg) reduced SBP in conscious rats (19 4%, p < 0.05), increased urinary NO excretion (4.06 0.53–7.07 1.59 µM/24 h; p < 0.05) and reduced plasma 8-isoprostane level (45.8 15 µM versus 31.4 8 µM), and NADP(H) oxidase activity (16 5%). Implantation of DOCA pellet (20 mg s.c.) in uninephrectomized mice placed on 1% NaCl drinking water increased SBP by a margin that was markedly greater in KO mice (193 13 mmHg versus 130 12 mmHg). In the rat, DOCA increased SBP and NAD(P)H oxidase activity and both effects were diminished by clofibrate. In addition, clofibrate reduced ET-1 production in DOCA/salt hypertensive rats. Thus, apart from inhibition of ET-1 production, PPARα activation exerts protective actions in hypertension via a mechanism that involves NO production and/or inhibition of NAD(P)H oxidase activity.

Keywords: NAD(P)H oxidase; Cardiovascular; Nitric oxide; PPARα; Interaction

1. Introduction

Peroxisome proliferator activated receptors (PPARs) are ligand-dependent nuclear transcription factors that form a subfamily of the nuclear receptor superfamily. PPARs comprise three isoforms, viz., α, β/6 and γ that exhibit tissue-specific distribution and ligand-specific affects. Upon heterodimerisation with retinoic acid receptor (RXR), PPARs regulate target gene expression by binding to specific peroxisome proliferator response element (PPREs) in enhancer sites of regulated genes. PPARα is expressed in tissues with very active fatty acid metabolism, such as the heart, kidney, liver and the endothelium [1], and vascular smooth muscle cells (VSMC) [2] suggesting that PPARα may exert direct beneficial effects on the vascular wall [3]. Beneficial effects of fibrates, prototype PPARα ligands, in the improvement of cardiovascular function has long been reported [4,5]. These beneficial effects in cardiovascular diseases have been suggested to involve the NO/NOS system [6]. Indeed, clinical observations indicate that treat- ment with PPARα activators, e.g. fibrates, lowers the progression of atherosclerotic lesions [7,8], an effect that might be ascribed to increased NO which is known to be antiatherogenic and anti-inflammatory [9]. This observation is consistent with the recent observation that PPARα activators improve endothelial-mediated NO vasodilation [10] and amplify iNOS expression [1,11,12]. Similarly, in our recent study, PPARα ligand-clofibrate, WY14643 and fenofibrate were shown to increase NO production in the kidney [13].

However, as cardiovascular function is regulated by the balance between NO and free radicals, and an enhanced superoxide (O2•− generation causes a diminished NO bioavailability by an oxidative reaction that inactivates NO, exaggerated O2•− and a low NO availability may lead to endothelial dysfunction and hypertrophy of vascular cells [14]. Indeed, increased production of reactive oxygen species (ROS) espe- cially, O2•−, contributes significantly to the functional and structural alterations in hypertension [15,16]. Since the balance between NO/NOS system and ROS is cru-
cial for optimal cardiovascular function, it is possible that besides improving NO availability, PPARα activa- tors may also influence either the antioxidant ability of the system and/or the production of ROS. Thus, experimental studies have shown a PPARα activator-mediated reduction of oxidative stress [10], a reduced p22phox message expression in primary endothelial cells [17], and increased Cu+–Zn+ SOD expression in the liver [18].

NAD(P)H oxidase is the source of O2•− in vascu- lar cells and earlier studies showed that PPARα ligands decreased the expression of NAD(P)H oxidase directly or indirectly by affecting the synthesis of hormonal agents that are known stimulants for NAD(P)H oxi- dase activity and free radical generation [17,19]. Of the many models of hypertension, the mineralocorticoid model involves the administration of deoxycorticos- terone acetate (DOCA) leading to an increase in blood pressure that was ascribed to diminished salt excretion. DOCA/salt hypertension an endothelin-1 (ET-1) sensi- tive model of hypertension [20], was shown to be char- acterized by increased NAD(P)H oxidase activity [15] and reduced NOS activity and expression [21]. When combined with the fact that PPARα also regulates the activity and production of ET-1 and therefore, influ- ence its pro-oxidant ability [22], a potential role exists for a protective effect following PPARα activation in DOCA/salt hypertension. In this study, we therefore, tested the hypothesis that PPARα-mediated reduction of NAD(P)H oxidase activity and/or increased NO pro- duction accounts for the beneficial effects of PPARα ligands in DOCA/salt hypertension.

2. Materials and methods

Unless specified otherwise in the text, all chem- icals were obtained from Sigma–Aldrich (St. Louis, MO) and are of the highest analytical grade. 3[H] L- arginine was purchased from Amersham Pharmacia Biotech (Piscataway, NJ). This study was approved by the Animal Care Com- mittee of the Texas Southern University and conforms to the institutional guidelines on animal care and use. This study was carried out in both rats (250–275 g, male, Sprague–Dawley; Harlan Sprague–Dawley, Houston, TX) and mice (18–20 g, male; Jackson Labo- ratory, Bar Harbor, ME) using bezafibrate or clofibrate as PPARα ligands. In the rat study, animals were ran- domly divided into vehicle and treatment groups. The treatment group received bezafibrate (30 mg/kg i.p.), alone or in combination with L-NAME (100 mg/L in drinking water), a NOS inhibitor, for 7 days. The control group received same volume of vehicle (mineral oil; i.p.), the diluent for bezafibrate. Animals were placed in metabolic cages and 24 h excretion of urine was collected at the beginning and at the end of the treatment. Systolic blood pressure (SBP) was mea- sured using tail cuff plethysmography (SC1000; Hat- teras Instruments, Cary, NC). Animals were sacrificed under sodium pentobarbital anesthesia (50 mg/kg i.p.) and plasma level of 8-isoprostane, an indicator of ROS generation, and NAD(P)H oxidase activity in the aorta was determined.

In another set of experiments, uninephrectomized rats (UNx) received implants of DOCA pellets (25 mg s.c.) and 1% NaCl in their drinking water. Rats were divided into two treatment groups—clofibrate (250 mg/kg i.p., daily) or vehicle (mineral oil, same vol- ume as clofibrate, i.p., daily) for 3 weeks. Blood pres- sure was measured before (Day 1) and after treatment (Day 21) and 24 h urine was collected for biochemical analyses. Rats were sacrificed under sodium pento- barbital anesthesia (50 mg/kg; i.p.) and tissue samples were collected for subsequent measurement of ET-1 levels and NAD(P)H oxidase enzyme activity.

In the mice study, age- and weight-matched adult (18–22 g) male PPARα knockout ( / ; KO) mice (Jackson Laboratory, Bar Harbor, ME) or their wild type (+/+; WT) litter mates (Controls) were randomly divided into groups—controls (mineral oil; 1 mL/kg; i.p., daily) and treatment (clofibrate 250 mg/kg, i.p., daily). SBP was measured before and after treatment. Plasma and kidney tissue were collected after the treat- ment period for measurement of lipid hydroperoxide and for total and Ca2+-independent NOS activity. In another set of experiments, both WT and KO mice were uninephrectomized, and DOCA (20 mg pellet; s.c.) was implanted. Mice were given 1% NaCl in the drinking water for 3 weeks and SBP was determined weekly.

2.1. Biochemical analyses

Urinary excretion of NO (UNOXV) and plasma lipid hydroperoxides were determined by fluorometric kits from Cayman Chemicals (Ann Arbor, MI). Urinary protein was determined by a colorimetric kit from Sigma–Aldrich (St. Louis, MO). Plasma level of 8- isoprostane, an indicator of free radical generation, was determined by EIA using a kit from Cayman Chemi- cals following manufacturers recommended protocols.Plasma levels of ET-1 were measured by EIA using a kit from Cayman Chemicals (Ann Arbor, MI).

2.2. NOS enzyme activity assay

NOS enzyme activity was determined by a radioac- tive method (conversion of 3[H] L-arginine to 3[H] L-citruline) using a kit from Cayman Chemicals (Ann Arbor, MI). Briefly, kidney was dissected and minced in a petri dish and then homogenized in 10 volumes of homogenization buffer (from assay kit). One milliliter of homogenate was pipetted into microcentrifuge tubes and spun at full speed for 5 min at 4 ◦C. The supernatant was transferred to a fresh tube and used for NOS activity assay. The reaction mixture was prepared by mixing 250 µL 2 reaction buffer, 50 µL 10 mM NAD(P)H (freshly prepared in 10 mM Tris–HCl; pH 7.4), 10 µL [3H] L-arginine (1 µCi/µL; Amersham), 50 µL 6 mM CaCl2, 0.1 µM calmodulin and 40 µL of distilled water. Fourty microliter of the reaction mixture was added to 10 µL of tissue extract and the mixture was incu- bated for 60 min at 37 ◦C. The reaction was stopped
by adding 400 µL of stop buffer and 100 µL of equilibrated resin to the samples. The samples were then transferred into spin cups and centrifuged at full speed for 30 s in a microcentrifuge. The eluate was trans- ferred to scintillation vials, mixed thoroughly with a scintillation cocktail (Beckman Coulter, Fullerton, CA) and radioactivity was measured by a liquid scintillation counter model LS6500 (Beckman Coulter, Fullerton, CA). Calcium-independent NOS activity was deter- mined by eliminating calcium and calmodulin from the reaction mixture.

2.3. NAD(P)H oxidase enzyme activity assay

NAD(P)H oxidase activity was estimated by a mod- ified enzyme activity assay [23]. Briefly, aorta were homogenized and centrifuged in HEPES buffer (pH 7.4) and protein concentration was measured with Bradford reagent. Tetrazolium salt radical detector (Cayman Chemicals, Ann Arbor, MI), dissolved in HEPES buffer (200 µL) was added to a 96-microwell plate. Equal volumes (20 µL) of protein samples were then added to separate wells and the reaction was initiated by 100 µM NADH and 100 µM NADPH. Absorbance at 450 nm was measured with a plate reader EL808IU (Bio-Tek Instruments Inc., Winoski,VT) every 10 min for 1 h at room temperature. Par- allel experiments were carried out in the presence of apocynin (10−4 M), an inhibitor of NADPH oxidase. Changes in absorbance were than normalized to the protein concentration.

2.4. Immunoblotting of iNOS protein expression

Fourty microgram of protein from kidney homogenate was electrophoresed on 12% polyacry- lamide gels and transferred to PVDF membrane (Amersham Pharmacia, NJ). Blots were probed with rabbit polyclonal anti-iNOS antibody (Santa Cruz Biotechnology, Santa Cruz, CA) at 1:300 dilution, followed by addition of secondary antibody at 1:8000 dilution. Immunocomplexes were visualized using an enhanced chemiluminescence (ECL-Plus) detection system from Amersham Pharmacia. The intensity of the bands was scanned and quantified using personal densitometer SI scanner and ImageQuant analysis software (Molecular Dynamics, Sunnyvale, CA).

2.5. Statistical analysis

Values were presented as mean S.E.M. Means were compared between groups and between treat- ments for significant difference using ANOVA. In all cases p < 0.05 was considered as statistically signifi- cant.

3.2. Effect of PPARα activation on NOS activity

Total NOS activity in the kidneys of untreated PPARα (WT) (n = 5) and (KO) mice (n = 5) was similar (Fig. 2a). However, clofi- brate almost doubled total NOS activity in PPARα (WT) mice (n = 5), from 0.99 0.14 to 1.96 0.14 pmol citruline/(mg protein min) (p < 0.05) whereas clofibrate was without effect in the PPARα (KO) mice (n = 4). Fig. 2b shows that basal Ca2+- independent NOS activity was 3 fold greater in PPARα (KO) mice compared to the PPARα (WT) litter mates. Treatment with clofibrate did not affect Ca2+-independent NOS activity in the PPARα (KO) mice but increased the activity in PPARα (WT) mice by 151 ± 30% from a basal value of 0.37 ± 0.08 pmol/(mg protein min) (p < 0.05).

3.3. Effect of PPARα activation on NOS protein expression

Western blot analysis (Fig. 2c) revealed a higher basal expression of iNOS in the PPARα (KO) mice compared to the PPARα (WT) mice (p < 0.05). Treat- ment with clofibrate reduced iNOS expression in PPARα (KO) mice (81 23%; p < 0.05, n = 6) but increased it in the PPARα (WT) mice (86 18%; p < 0.05, n = 6). However, clofibrate treatment was without effect on eNOS and nNOS protein expres- sion in both PPARα (WT) and (KO) mice (data not shown).

Fig. 2. (a) Total nitric oxide synthase (NOS) activity, (b) Ca2+- independent NOS activity and (c) iNOS protein expression in PPARα WT and KO mice before and after treatment with clofibrate (250 mg/kg i.p. for 7 days). The upper tracing in (c) shows a typical immunoblot of iNOS (130 kDa) in the different groups. Values are mean S.E.M, p < 0.05 vs. untreated PPARα WT mice (N = 5 per group).

3.4. Effect of PPARα activation on lipid hydroperoxide generation

PPARα ligands exert effects on free radical genera- tion (Inoue et al., [17,18]) which can be modulated by NO, therefore, we evaluated the capacity for free rad- ical generation in the presence or absence of PPARα. Fig. 3 illustrates that in untreated PPARα (KO) mice (n = 5), basal plasma lipid hydroperoxide generation was 30 3% (p < 0.05) lower than that obtained in the untreated PPARα (WT) mice (n = 5). Clofibrate elicited a dual effect, reducing lipid hydroperoxide gen- eration in the PPARα (WT) mice (34 ± 4%; p < 0.05) but increasing it (33 2%; p < 0.05) in PPARα (KO) mice.

Fig. 3. Plasma lipid hydroperoxides in PPARα (WT) and (KO) mice treated with clofibrate (250 mg/kg i.p. for 7 days). Values are mean S.E.M, @p < 0.05 vs. untreated PPARα WT mice, p < 0.05 vs. untreated PPARα KO mice (N = 5 per group).

3.5. Effect of PPARα activation on blood pressure, NO production and free radical generation

Fig. 4a illustrates that compared to controls (n = 4), bezafibrate (30 mg/kg orally for 5 days, n = 5) reduced

4. Discussion

The effect of PPARα ligands on NO production is the subject of active research interest but has not been resolved due to conflicting data from many lab- oratories. For example, PPARα activators inhibited iNOS in macrophages [3] and cytokine-stimulated astroglial cells [24] but amplified iNOS expression in cytokine-stimulated mesangial cells [12]. Our recent study also support that PPARα activators increased renal NO production as measured by urinary excretion of nitrite/nitrate [13].

In this study, we sought to clarify the role of PPARα in NO production by evaluating the effects of PPARα agonists on the NO/NOS system in mice in which the PPARα gene is deleted. The selective increase by clofi- brate, the prototype PPARα ligand, in total and Ca2+- independent renal NOS activity as well as increased iNOS expression in PPARα (WT) mice, corroborates our previous study in the rats showing that activation of PPARα stimulates NO/NOS system [13]. The lack of effect by clofibrate on NOS activity in PPARα KO mice supports that PPARα is required for NO pro- duction/NOS expression and indicates that this effect occurs probably at the transcriptional and/or transla- tional levels. However, the observation that clofibrate inhibited iNOS expression in PPARα KO mice; effects opposite to that obtained in PPARα WT mice, suggests that this effect is independent of PPARα activation. This is consistent with the observations of Pahan et al. who reported a PPARα-independent reduction of iNOS expression by another fibrate, gemfibrozil [24]. It is noteworthy that free radical generation was reduced alongside the increase by clofibrate in NOS activ- ity in WT mice. This observation corroborates data demonstrating that PPARα activation resulted in the induction of antioxidant enzymes [25] and that fibrates increased the mRNA for Cu2+/Zn2+ superoxide dis- mutase (SOD), decreased the mRNA for NAD(P)H oxidase in human cultured endothelial cells [17,18] and reduced the expression of vascular NAD(P)H oxidase p22phox subunit in diabetes [19]. However, it is not clear the extent to which PPARα-mediated effect on free radicals contributes to regulation of blood pressure. In this study, bezafibrate reduced SBP in normal rats, i.e. under normal physiologic conditions. This hypotensive effect was accompanied by reductions in plasma 8-isoprostane and NAD(P)H oxidase activity, suggesting that free radicals contribute to PPARα-mediated regulation of blood pressure. This is consistent with the proposed role for free radicals in normal cardiovascular function as superoxide anion (O2•−) was demonstrated to be a major regulator of vascular and renal function [26,27]. Other corroborating evidence came from stud- ies in which SBP was different in mice lacking genes for NAD(P)H oxidase subunits, e.g. gp91phox subunit [27] or in Cu2+/Zn2+ SOD ( / ) mice. In this study, PPARα activation elicited an antihypertensive effect in DOCA/salt hypertensive rats, an effect that appears to involve reduction in generation of reactive oxygen species (ROS) inasmuch as NAD(P)H oxidase activ- ity and plasma levels of 8-isoprostane were reduced. Apart from reduction in NAD(P)H oxidase activity, PPARα activation also blunted the renal damage that is characteristic of DOCA/salt hypertension. It, thus, appears that PPARα activation-mediated reduction in NAD(P)H oxidase activity may define the antihyper- tensive and renoprotective effects of PPARα ligands in DOCA/salt hypertension. This observation contrasts with that reported in SHR and angiotensin-II-induced hypertension in which PPARα ligands did not affect blood pressure elevation but rather blunted the vas- cular remodelling that accompanies the hypertension [28,29]. The reduction in plasma ET-1 levels by clofi- brate in this study is not surprising as PPARα activation is known to inhibit the production and transcriptional regulation of ET-1 [22]. The reduction in NAD(P)H oxidase activity following PPARα activation may also derive from the reduction in ET-1 levels since ET-1 is a known stimulant for NAD(P)H oxidase activity and free radical generation [17–19].

The next question to address is the interactions of PPARα, NO and NAD(P)H oxidase in the regulation of cardiovascular function. A balance between O2•− and NO is known to define the overall control of cardiovascular function under physiological [26,27,30] and pathological conditions [31–33]. Thus, an increase in NO will be expected to antagonize the effect/generation of O2•−. It is also possible that excessive generation of NO will produce peroxynitrite following combina- tion with O2•−, especially, in pathologic conditions. This being the case, our data seem to indicate that a dual mechanism involving PPARα-mediated increase in NO and/or reduction of NAD(P)H oxidase contribute to the effects of fibrates. Evidence in support of this comes from the observation that higher basal NOS activity was accompanied by reduced generation of lipid hydroperoxides in PPARα WT mice and that clofi- brate selectively increased NOS activity but reduced generation of hydroperoxides in PPARα WT but not KO mice. Additional evidence was provided by data in which bezafibrate increased urinary NO excretion but reduced plasma 8-isoprostane and NAD(P)H oxidase activity in normal as well as in DOCA/salt hypertensive rats. However, it is unclear whether inhibition of free radical(s) or stimulation of NO production is the pri- mary mechanism for the physiologic effects of PPARα ligands. Thus, we reasoned that if PPARα ligands still modulate cardiovascular function in the absence of NO, a case can be made for the involvement of ROS, hence, a dual mechanism of action. Thus, in rats treated with L-NAME to inhibit NO production, SBP increased as did plasma 8-isoprostane and NAD(P)H oxidase activ- ity. The blunting of these effects by clofibrate suggests that though NO may be the primary target for the blood pressure-lowering effect of PPARα ligands in the normal rat, reduction in free radical generation may become a major mechanism in the absence of NO. This mechanism may be of profound relevance in diseases characterized by diminished NO produc- tion/availability, e.g. in DOCA/salt hypertension [21]. In conclusion, these data suggest that PPARα activa- tion contributes to normal regulation of blood pressure and exerts protective actions in hypertension via a dual mechanism that involves NO production and inhibition of generation of free radicals.