Vascular NAD(P)H oxidase mediates endothelial dysfunction in basilar arteries from Otsuka Long-Evans Tokushima Fatty (OLETF) rats

Atherosclerosis 192 (2007) 15–24

Vascular NAD(P)H oxidase mediates endothelial dysfunction in basilar arteries from Otsuka Long-Evans Tokushima Fatty (OLETF) rats Takayuki Matsumoto a , Tsuneo Kobayashi a , Hiroshi Wachi b , Yoshiyuki Seyama b , Katsuo Kamata a,∗ b

a Department of Physiology and Morphology, Institute of Medicinal Chemistry, Hoshi University, Shinagawa-ku, Tokyo 142-8501, Japan Department of Clinical Chemistry, School of Pharmacy and Pharmaceutical Science, Hoshi University, Shinagawa-ku, Tokyo 142-8501, Japan

Received 29 August 2005; received in revised form 13 April 2006; accepted 2 June 2006 Available online 10 July 2006

Abstract We examined the responses of basilar arteries taken from Otsuka Long-Evans Tokushima Fatty (OLETF) rats, a type 2 diabetes model. Both the nitric oxide (NO)-mediated relaxation and the cyclic 3 ,5 -guanosine monophosphate (cGMP) production elicited by acetylcholine (ACh) were much weaker in OLETF rats than in age-matched control Long Evans Tokushima Otsuka (LETO) rats. The contraction induced by an NO synthase (NOS) inhibitor [NG -nitro-l-arginine (l-NNA)] was weaker in the OLETF group. In that group, application of apocynin, an NAD(P)H oxidase inhibitor, normalized (i) ACh-induced relaxation, (ii) l-NNA-induced contraction, and (iii) ACh-induced cGMP production to the LETO levels. Superoxide anion production was greater in basilar arteries from OLETF rats than in those from LETO rats. The protein expression of gp91phox , an NAD(P)H oxidase subunit, was upregulated in the OLETF arteries (versus LETO ones). These results suggest that the existence of endothelial dysfunction in basilar arteries in type 2 diabetes is related to increased oxidative stress mediated via NAD(P)H oxidase. Possibly, an impairment of NO-dependent relaxation responses and a basal impairment of NO signaling may be responsible for the increased risk of adverse cerebrovascular events in type 2 diabetes. © 2006 Published by Elsevier Ireland Ltd. Keywords: gp91phox ; Nitric oxide; Relaxation; Type 2 diabetes

1. Introduction Large arteries such as the basilar artery make important contributions to total vascular resistance in the cerebral circulation, and they are major determinants of local microvascular pressure in that circulation [1]. While the basic principles of blood flow regulation apply to all vascular beds, there are some important differences between cerebral blood vessels and vessels in other organs in their response to humoral, neural, and metabolic stimuli, in their response to hypercapnia/hypoxia, and in their autoregulation [2]. Diabetes mellitus increases the risk for all manifestations of atherosclerotic vascular disease, coronary heart disease, cerebrovascular disease, and peripheral vascular disease [3]. ∗

Corresponding author. Tel.: +81 3 5498 5856; fax: +81 3 5498 5856. E-mail address: [email protected] (K. Kamata).

0021-9150/$ – see front matter © 2006 Published by Elsevier Ireland Ltd. doi:10.1016/j.atherosclerosis.2006.06.005

Diabetes mellitus is a risk factor in the pathogenesis of many cerebrovascular events, including cerebral ischemia and stroke [3]. An accumulating body of evidence obtained in animal models and patients indicates that endotheliumdependent relaxation is impaired in peripheral blood vessels in type 2 diabetes [4,5], as is also the case in type 1 diabetes [4,6–9]. For example, Otsuka Long-Evans Tokushima Fatty (OLETF) rats [10] develop hyperglycemia, obesity, and insulin resistance, and as a result, serve as a model for human type 2 diabetes. It is now realized that this animal model exhibits endothelial dysfunction in peripheral blood vessels [11]. In contrast, very little is known about the endotheliumdependent responses of cerebral vessels in type 2 diabetes [12–14]. Moreover, the changes that occur in the mechanisms regulating the cerebral circulation – changes presumably leading to the scourge of cerebral vascular dysfunction – remain virtually unexplored.

16

T. Matsumoto et al. / Atherosclerosis 192 (2007) 15–24

Nitric oxide (NO) is a potent vasodilator that produces relaxation of cerebral blood vessels both in vitro and in vivo [2]. After its release by the endothelium, via NO synthase (NOS) activation, NO stimulates soluble guanylyl cyclase in vascular smooth muscle, resulting in an increase in the intracellular concentration of cyclic 3 ,5 -guanosine monophosphate (cGMP) and hence to relaxation [2,9]. Moreover, an accumulating body of evidence indicates that the constitutive levels of NOS expression in the endothelium are sufficient to influence tone in cerebral blood vessels under basal conditions [2]. Basal levels of cGMP are much greater in cerebral arteries when the endothelium is present than when it is absent. In vitro, inhibitors of NOS decrease the basal level of cGMP and produce a constriction of cerebral arteries that is endothelium-dependent [2]. In vivo, inhibitors of NOS constrict cerebral blood vessels and decrease cerebral blood flow under basal conditions in several species [2,15]. However, there is little information as to whether the NO/cGMPmediated response in cerebral blood vessels is altered in type 2 diabetic models. Although the mechanisms mediating endothelial dysfunction in diabetic states are as yet poorly understood, a considerable body of evidence implicates oxidative stress as an important pathogenic element in such dysfunction [4,9,16]. NO reacts extremely efficiently with superoxide, resulting in a loss of NO bioavailability [16]. Oxidative stress, defined as an increase in the steady-state levels of reactive oxygen species (ROS), may occur as a result of increased free-radical generation – an event largely attributable to NAD(P)H oxidase activation within the vascular system [17–19] – and/or to a decline in anti-oxidant defense mechanisms, such as superoxide dismutase (SOD) [4,9]. To date, very little is known regarding the importance of the levels of NAD(P)H oxidase expression and whether it is an important source of ROS in cerebral blood vessels during diseases such as hypertension and subarachnoid hemorrhage [16]. We recently reported that the impairment of endothelium-dependent relaxation seen in basilar arteries from streptozotocin (STZ)-induced diabetic rats may be attributed to increased oxidative stress, and that this increment may be due to an overexpression of the mRNA for the gp91phox subunit, which is an active catalytic subunit of NAD(P)H oxidase [15]. However, no study has yet provided an insight into the relationship between endothelial dysfunction and NAD(P)H oxidase in type 2 diabetic basilar arteries. Since the incidence of type 2 diabetes is increasing and since the cerebral vascular responses in type 2 diabetes have not yet been subjected to detailed scrutiny, we designed the present study to test whether the relaxations of basilar arteries induced by endothelium-dependent and -independent agonists are altered in a type 2 diabetic model. A study of endothelial function seemed particularly appropriate because endothelial dysfunction has emerged as an independent predictor of adverse clinical events. In addition, we attempted to clarify whether NAD(P)H oxidase contributes to any endothelial dysfunction we detected in such basilar arter-

ies. We also asked whether basilar arteries from control and diabetic rats might differ in their expressions of NAD(P)H oxidase subunits.

2. Materials and methods 2.1. Reagents NG -Nitro-l-arginine (l-NNA), sodium nitroprusside (SNP), 9,11-dideoxy-11␣,9␣-epoxymethanoprostaglandin F2␣ (U46619), phenylmethylsulfonyl fluoride (PMSF), nitrite (NO2 − ), nitrate (NO3 − ), indomethacin, ␤nicotinamide adenine dinucleotide (NADH), polyethyleneglycolated superoxide dismutase (PEG-SOD), nitro blue tetrazolium (NBT), and monoclonal ␤-actin antibody were all purchased from Sigma Chemical Co. (St. Louis, MO, USA), while acetylcholine chloride (ACh) was from Daiichi Pharmaceuticals (Tokyo, Japan). The NAD(P)H oxidase inhibitor apocynin and the SOD mimetic 4-hydroxy2,2,6,6-tetramethylpiperidine-1-oxyl (Tempol) were from Calbiochem-Novabiochem Corporation (La Jolla, CA, USA). All drugs were dissolved in saline, except where otherwise noted. All concentrations are expressed as the final molar concentration of the base in the organ bath. Horseradish peroxidase (HRP)-linked secondary anti-mouse antibody was purchased from Promega (Madison, WI, USA), while an antibody against gp91phox was obtained from BD Biosciences (San Jose, CA, USA). Finally, p22phox and HRP-linked secondary anti-goat antibody were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). 2.2. Animals and experimental design Five-week-old male rats [OLETF rats and Long-Evans Tokushima Otsuka (LETO) rats, a genetic control for OLETF] were supplied by the Tokushima Research Institute (Otsuka Pharmaceutical). Food and water were given ad libitum in a controlled environment (room temperature 21–22 ◦ C, room humidity 50 ± 5%), until the rats were 60–63 weeks old. This study was conducted in accordance with the Guide for the Care and Use of Laboratory Animals adopted by the Committee on the Care and Use of Laboratory Animals of Hoshi University (which is accredited by the Ministry of Education, Culture, Sports, Science and Technology, Japan). 2.3. Measurement of plasma glucose, cholesterol, triglyceride, insulin, and blood pressure Plasma parameters and blood pressure were measured as described previously [5,15,20]. 2.4. Measurement of isometric force Vascular isometric force was recorded as in our previous papers [15]. Rats were anesthetized with diethyl ether and

T. Matsumoto et al. / Atherosclerosis 192 (2007) 15–24

euthanized by decapitation at 60–63 weeks of age. A section of the basilar artery was then removed and placed in ice-cold, oxygenated, modified Krebs–Henseleit solution (KHS). This solution consisted of (in mM) 118.0 NaCl, 4.7 KCl, 25.0 NaHCO3 , 1.8 CaCl2 , 1.2 NaH2 PO4 , 1.2 MgSO4 , and 11.0 dextrose. Each basilar artery was separated from surrounding connective tissue and cut into rings (2 mm long). The ring segments were suspended by a pair of stainless-steel pins in a well-oxygenated (95% O2 –5% CO2 ) bath of 10mL KHS at 37 ◦ C. The rings were stretched until an optimal resting tension of 0.3 g was loaded, and then allowed to equilibrate for at least 60 min. Force generation was monitored by means of an isometric transducer (model TB-611T; Nihon Kohden, Tokyo, Japan). Tension was readjusted when necessary, and the bath fluid was changed every 15 min. After this period of equilibration, the reactivity of the rings was checked by depolarization with 64 mM KCl. There were no significant differences in the response to KCl between the OLETF (n = 30) and LETO (n = 30) groups (280 ± 17 and 259 ± 18 mg, respectively). Once the contraction induced by U46619, a thromboxane analogue (1 ␮M), was established, a concentration–response curve was constructed for the relaxation induced by ACh or SNP. After the addition of sufficient aliquots of the agonist to produce the chosen concentration, a plateau response was allowed to develop before the addition of the next dose of the same agonist. In addition, to investigate the influence of oxidative stress on ACh- or NO2 − -induced relaxation in the diabetic basilar artery, an inhibitor of NAD(P)H oxidase (apocynin, 100 ␮M), a SOD mimetic (Tempol, 1 mM), a cell-permeant superoxide anion scavenger (PEG-SOD, 41 U/mL), or a cyclooxygenase (COX) inhibitor (indomethacin, 10 ␮M) was added to the organ bath 30 min before precontraction with U46619, and this was followed by application of 10 ␮M ACh or 1 mM NO2 − . In a preliminary experiment, when basilar arteries were treated with apocynin for 30 min neither the 5-HT- or U46619-induced contraction nor the SNP-induced relaxation showed any alteration (data not shown). In separate experiments, we examined the changes in vascular tone seen after application of l-NNA (100 ␮M), a NOS inhibitor. In some of these experiments, 100 ␮M apocynin was applied 30 min before the l-NNA application and was present thereafter. These contractile responses were each expressed as a percentage of the contraction previously induced by 64 mM KCl. 2.5. Enzyme immunoassay for cGMP The cGMP level in the ACh-stimulated basilar artery was quantified as in our previous paper [5]. Basilar artery rings from OLETF or LETO rats were incubated for 30 min at 37 ◦ C in well-oxygenated KHS in the presence or absence of 1 mM Tempol or 100 ␮M apocynin. The rings were frozen in liquid N2 after a 1-min addition of ACh (10 ␮M), then stored at −80 ◦ C. cGMP was subsequently extracted in 6%

17

trichloroacetic acid, followed by neutralization with watersaturated diethyl ether and enzyme immunoassay (Amersham Biosciences, UK). 2.6. Quantification of superoxide anion by measurement of the amount of NBT reduced Basilar rings were incubated with NBT to allow the superoxide generated by the tissue to reduce the NBT to blue formazan. The details of the assay have been published previously [21]. Briefly, basilar arteries from LETO and OLETF groups were placed for 30 min in 1 mL KHS (at 37 ◦ C) containing 10−5 M NADH, and this was followed by application of NBT (10−4 M) for 180 min. The NBT reduction was stopped by addition of 0.5N HCl (1 mL). After this incubation, the rings were minced and homogenized in a mixture of 0.1N NaOH and 0.1% SDS in water containing 40 mg L−1 diethylentriaminepentaacetic acid. The mixture was centrifuged at 16,000 × g for 30 min, and the resultant pellet resuspended in 250 ␮L of pyridine at 80 ◦ C for 60 min to extract formazan. The mixture was then subjected to a second centrifugation at 10,000 × g for 10 min. The absorbance of the formazan was determined spectrophotometrically at 540 nm. The amount of NBT reduced (=quantity of formazan), was calculated as follows: amount of NBT reduced = A × V/(T × Wt × ε × l), where A is the absorbance, V the volume of pyridine, T the time for which the rings were incubated with NBT, Wt the blotted wet weight of the aortic rings, ε the extinction coefficient (0.7 L mmol−1 /mm) and l is the length of the light path. The results are reported in pmol min−1 /mg wt. 2.7. Western blotting The protein levels of NAD(P)H oxidase subunits were quantified using immunoblotting procedures, essentially as described before [5]. Basilar arterial tissues (for each group, five pooled vessels per experiment) were homogenized in ice-cold lysis buffer containing 50 mM Tris–HCl (pH 7.2), 150 mM NaCl, 1% Nonidet P-40, 1% sodium deoxycholate, and 0.1% SDS containing 1 mM PMSF. The lysate was cleared by centrifugation at 16,000 × g for 10 min at 4 ◦ C. The supernatant was collected, and the proteins were solubilized in Laemmli’s buffer containing mercaptoethanol. The protein concentration was determined by means of a bicinchoninic acid (BCA) protein assay reagent kit (Pierce, IL, USA). Samples (20 ␮g/lane) were resolved by electrophoresis on 12% SDS-PAGE gels, then transferred onto PVDF membranes. Briefly, after blocking the residual protein sites on the membrane with Block ace (Dainippon-pharm., Osaka, Japan), the membrane was incubated with anti-gp91phox (58 kDa; 1:1000) or anti-p22phox (22 kDa; 1:200) in blocking solution. HRP-conjugated, anti-mouse, or -goat antibody was used at a 1:10,000 dilution in Tween PBS, followed by detection using SuperSignal (Pierce). To normalize the data, we used ␤-actin as a housekeeping protein, the ␤-actin

18

T. Matsumoto et al. / Atherosclerosis 192 (2007) 15–24

protein levels being determined after stripping the membrane and probing with ␤-actin monoclonal primary antibody (42 kDa; 1: 5000), with HRP-conjugated anti-mouse IgG as the secondary antibody. The optical densities of the bands on the film were quantified using densitometry, with correction for the optical density of the corresponding ␤-actin band. 2.8. Statistical analysis Data are expressed as the mean ± S.E. When appropriate, statistical differences were assessed by Dunnett’s test for multiple comparisons after a one-way analysis of variance, a probability level of P < 0.05 being regarded as significant. Statistical comparisons between concentration–response curves were made using a two-way ANOVA, with Bonferroni’s correction for multiple comparisons being performed post hoc (P < 0.05 again being considered significant).

3. Results 3.1. General parameters In accord with a previous report [22], at the time of the experiment all OLETF rats (non-fasted) exhibited hyperglycemia, their blood glucose concentrations being significantly higher than those of the age-matched nondiabetic control LETO rats (also non-fasted) [357.6 ± 45.3 and 161.0 ± 3.6 mg/dl in OLETF (n = 20) and LETO (n = 20), respectively (P < 0.001)]. The body weight of the OLETF rats was greater than that of LETO rats [648.1 ± 34.1 and 563.5 ± 9.1 g in OLETF (n = 20) and LETO (n = 20), respectively (P < 0.05)]. The plasma total cholesterol [204.4 ± 17.7 and 98.8 ± 1.5 mg/dl in OLETF (n = 20) and LETO (n = 20), respectively (P < 0.001)] and triglyceride levels [404.9 ± 50.4 and 96.1 ± 5.6 mg/dl in OLETF (n = 20) and LETO (n = 20), respectively (P < 0.001)] were significantly higher in OLETF rats than in LETO rats, while the plasma insulin concentrations [1263.7 ± 160.5 and 1398.1 ± 121.4 pg/ml in OLETF (n = 20) and LETO (n = 20), respectively] were similar between the two groups. The systolic blood pressure of OLETF rats was higher than that of LETO rats [137.4 ± 3.0 and 113.3 ± 1.5 mmHg in OLETF (n = 18) and LETO (n = 18), respectively (P < 0.001)]. 3.2. Relaxation responses to ACh and SNP The tension developed in response to 1 ␮M U46619 did not differ significantly between the two groups [143 ± 14 mg in LETO (n = 14) and 181 ± 15 mg in OLETF (n = 14)]. In rings from age-matched LETO rats, ACh (10−9 to 10−5 M) induced a concentration-dependent relaxation, with the maximum response at 10−5 M (Fig. 1A). This relaxation was significantly weaker in rings from OLETF rats (P < 0.001 versus LETO group). The EC50 values for the ACh-induced relaxations were 0.27 ± 0.05 ␮M (n = 7) and

Fig. 1. Concentration–response curves for ACh-induced (A) and SNPinduced (B) relaxations of isolated rings of basilar arteries obtained from LETO and OLETF rats and SNP-induced (C) effects of apocynin (100 ␮M) and Tempol (1 mM) on SNP-induced relaxation in basilar arteries from OLETF rats. Details are given under Section 2. Each data-point represents the mean ± S.E. from 3 to 7 experiments; the S.E. is included only when it exceeds the dimension of the symbol used. *** P < 0.001, OLETF vs. LETO.

0.67 ± 0.12 ␮M (n = 7) in LETO and OLETF rats, respectively (P < 0.01). SNP primarily releases NO intracellularly within smooth muscle cells, and it thus produces relaxation in a non-endothelium-dependent manner. SNP-induced a dosedependent relaxation that was similar between LETO and OLETF rats (Fig. 1B). The SNP-induced relaxation responses in basilar arteries from OLETF rats (Fig. 1C) and LETO rats (data not shown) were not significantly altered by preincubation with either the SOD mimetic Tempol (1 mM) or the NAD(P)H oxidase inhibitor apocynin (100 ␮M).

T. Matsumoto et al. / Atherosclerosis 192 (2007) 15–24

19

3.3. Effect of various drugs on ACh-induced relaxation Since O2 − reacts rapidly with NO, thereby reducing its bioactivity, we examined the ACh (10 ␮M)-induced relaxation in basilar arteries from OLETF rats to establish the effects of (i) apocynin, an inhibitor of NAD(P)H oxidase, a major enzymatic source of O2 − within the vascular system (see Section 1) and (ii) two free-radical scavengers (PEGSOD and Tempol) (Fig. 2). When apocynin (100 ␮M) was applied to basilar artery rings isolated from the OLETF group, this compound did not itself cause tension development nor did it alter the U46619-induced contraction (data not shown). In LETO rats, the vasodilator response to ACh was similar between the absence and presence of apocynin. In contrast, the impaired basilar artery dilatation seen in response to ACh in OLETF rats was restored to normal levels by apocynin. In line with the results obtained with apocynin, the vasodilator response to ACh in LETO rats was similar between the absence and presence of either PEG-SOD (41 U/mL) or Tempol (1 mM), neither of which by itself induced tension development or altered the U46619-induced contraction (data not shown). Moreover, the impaired AChinduced dilatation of basilar arteries seen in OLETF rats was significantly improved by both PEG-SOD and Tempol. In addition, to assess the possible effects of PGI2 and other inhibitory products of COX metabolism, basilar arterial rings were preincubated with 10 ␮M indomethacin for 30 min (Fig. 2). This treatment had no significant effect on the subsequently induced endothelium-dependent relaxation (versus that in untreated rings) in either group of rats. It should be noted that the ACh-induced relaxation observed in the presence of indomethacin was significantly weaker in rings from OLETF rats than in those from LETO rats (P < 0.001). The above result indicates that products of COX metabolism do not make a prominent contribution to ACh-induced relaxation in this preparation.

Fig. 3. Effects of Tempol (Temp; 1 mM) and PEG-SOD (41 U/mL) on NO2 − (1 mM)-induced relaxation in basilar arteries from LETO and OLETF rats. Details are given under Section 2. Each column represents the mean ± S.E. from 3 to 6 experiments. * P < 0.05 vs. vehicle (Veh)-treated LETO. # P < 0.05, ## P < 0.01 vs. Veh-treated OLETF.

3.4. Relaxation response to NO2 − It has been reported that when O2 − reacts with NO, it produces less-potent vasodilators such as nitrite (NO2 − ) and nitrate (NO3 − ) [23]. In both LETO and OLETF groups, NO2 − (Fig. 3) but not NO3 − (data not shown) produced vasodilatation in basilar arteries. In basilar arteries from OLETF rats, this NO2 − -induced relaxation was significantly impaired, but it was significantly improved by both Tempol and PEG-SOD (Fig. 3). In LETO rats, the vasodilator response to NO2 − was similar between the absence and presence of Tempol or PEG-SOD (Fig. 3). 3.5. Effect of a nitric oxide synthase inhibitor on vascular tone To examine the involvement of NO in setting vascular tone in the basilar artery, l-NNA (100 ␮M), a representative NOS inhibitor, was applied to basilar artery rings. This agent induced a contractile response that was significantly weaker in the OLETF group than in the LETO group (Fig. 4). Interestingly, this diminished contractile response in the OLETF basilar artery showed significant recovery in the presence of 100 ␮M apocynin, although apocynin had no effect on l-NNA-induced contraction in LETO rats (Fig. 4). The lNNA-induced contraction was completely relaxed by SNP in both OLETF and LETO rats (data not shown). 3.6. Measurement of cGMP production

Fig. 2. Effects of apocynin (Apo; 100 ␮M), Tempol (Temp; 1 mM), PEGSOD (41 U/mL), and indomethacin (indo; 10 ␮M) on ACh (10 ␮M)-induced relaxation in basilar arteries from LETO and OLETF rats. Details are given under Section 2. Each column represents the mean ± S.E. from 3 to 8 experiments. *** P < 0.001 vs. vehicle (Veh)-treated LETO. ## P < 0.01, ### P < 0.001 vs. Veh-treated OLETF.

cGMP levels were measured in rat basilar artery rings treated or not treated with 1 mM Tempol or 100 ␮M apocynin (Fig. 5). Under our conditions, basal cGMP level, as in the vehicle-treated group, was significantly lower in basilar arteries from OLETF rats than in those from LETO rats. An ACh (10 ␮M)-induced cGMP accumulation was seen in both groups. However, this cGMP accumulation much smaller in

20

T. Matsumoto et al. / Atherosclerosis 192 (2007) 15–24

Fig. 6. Quantification of basilar arterial superoxide-anion production by measurement of amount of reduced nitro blue tetrazolium (NBT). Tissues were obtained from OLETF rats and age-matched LETO rats. Details are given under Section 2. Each column represents the mean ± S.E. from 5 experiments. * P < 0.05 vs. LETO group. Fig. 4. Time course of changes in contractile response to the nitric oxide synthase inhibitor, NG -nitro-l-arginine (l-NNA; 100 ␮M) in basilar arteries from LETO and OLETF rats. Details are given under Section 2. Each datapoint represents the mean ± S.E. from 5 to 6 experiments; the S.E. is included only when it exceeds the dimension of the symbol used. *** P < 0.001, Vehtreated LETO vs. Veh-treated OLETF. ### P < 0.001, Veh-treated OLETF vs. apocynin-treated OLETF.

rings isolated from OLETF rats than in those from LETO rats. In LETO rings, the ACh-stimulated cGMP level was not altered by pretreatment with either Tempol or apocynin, whereas in OLETF rings it was greatly increased by each of these pretreatments. 3.7. Quantification of superoxide anion by measurement of the amount of NBT reduced To extend our investigation of the effect of oxidative stress via NAD(P)H oxidase in basilar arteries, rings were prein-

cubated with NADH (10−5 M) for 30 min. It is likely that vascular NAD(P)H oxidase utilizes NADH as an electron donor for the generation of superoxide, and consequently NADH has been commonly employed to increase NAD(P)H oxidase activity in intact vascular segments, intact vascular cells, and vascular cell homogenates [24]. To judge from our measurements of the amount of NBT reduced by superoxide, the superoxide level in the presence of NADH (10−5 M) was greater in basilar arteries from OLETF rats than in those from LETO rats (P < 0.05) (Fig. 6). 3.8. Expression of NAD(P)H oxidase subunits To investigate the possible mechanisms underlying (a) the abnormal vascular responsiveness seen in basilar arteries from OLETF rats and (b) the normalization of these alterations seen upon apocynin treatment, we examined whether the protein expression of NAD(P)H oxidase subunits might be altered in the diabetic basilar artery. Interestingly, the level of gp91phox protein, a major subunit responsible for NAD(P)H oxidase activity, was significantly higher in OLETF rats than in LETO rats (Fig. 7). An other membrane NAD(P)H oxidase subunit, p22phox , tended to be at a higher level in basilar arteries from OLETF rats, although statistical significance was not reached (Fig. 7).

4. Discussion

Fig. 5. cGMP levels in basilar arteries from LETO to OLETF rats. Tempol (1 mM), apocynin (100 ␮M), or vehicle (water) was applied 30 min before ACh (10 ␮M) application, and was present thereafter. Details are given under Section 2. Each column represents the mean ± S.E. from 3 to 5 experiments. * P < 0.05 vs. vehicle-treated LETO. † P < 0.05 vs. vehicle-treated OLETF. ‡‡ P < 0.01 vs. ACh-treated LETO. ## P < 0.01, ### P < 0.001 vs. ACh-treated OLETF.

In the present investigation, we demonstrated that in OLETF rats, a model of type 2 diabetes, the endotheliumdependent relaxation in basilar arteries is significantly weaker than that seen in LETO rats, and that the mechanism underlying this endothelial dysfunction may be related to a decrease in NO bioactivity mediated via NAD(P)H oxidase activation. OLETF rats are characterized by an early increase in serum insulin, and by late-onset hyperglycemia, mild obesity, and mild type 2 diabetes [10]. To clarify the mechanism underlying cerebral vascular dysfunction in the chronic stage of the disease, we used basilar arteries isolated from aged

T. Matsumoto et al. / Atherosclerosis 192 (2007) 15–24

21

Fig. 7. Analysis of protein expression of membrane components of NAD(P)H oxidase expression in basilar arteries from LETO to OLETF rats. (A) Representative Western blots of gp91phox , p22phox , and ␤-actin. Proteins (20 ␮g) were subjected to 12% PAGE and then transferred to polyvinylidene difluoride membranes. These were then incubated with a primary antibody specific for gp91phox (58 kDa) or p22phox (22 kDa), and also with secondary anti-mouse or anti-goat antibody. In addition, the membranes were stripped and probed with ␤-actin (42 kDa) primary and secondary anti-mouse antibodies. (B) Bands were quantified as described in Section 2. Ratios were calculated for the optical density of gp91phox or p22phox over that of ␤-actin. Values are the mean ± S.E. of 4 (gp91phox /␤-actin) or 5 (p22phox /␤-actin) determinations. * P < 0.05 vs. LETO.

OLETF rats. First, we examined the endothelium-dependent and -independent relaxations induced by ACh and SNP (Fig. 1). The ACh-induced relaxation of this artery appears to be mediated by activation of the l-arginine/NO biosynthetic pathway. We [15] and others [2] have shown that ACh relaxes the rat basilar artery, and that enzymatic inhibitors of NOS attenuate this relaxation. In the present study, the ACh-induced relaxation, but not that induced by SNP, was significantly impaired in the OLETF group (Fig. 1), as was ACh-induced cGMP accumulation (Fig. 5). These results suggest that in OLETF rats, agonist-induced NOS-dependent dilation of the basilar artery is impaired. Moreover, several pieces of evidence suggest that the constitutive level of NOS expression in the endothelium of cerebral blood vessels is sufficient to influence tone under basal conditions [2]. Indeed, inhibitors of NOS induce contraction of cerebral blood vessels and decrease cerebral blood flow under basal conditions in several species [2,15]. In the present study, both the lNNA-induced contractile response (Fig. 4) and the basal cGMP production (Fig. 5) were found to be significantly impaired in the OLETF rat basilar artery. These results suggest that NO bioactivity is impaired in OLETF rat basilar arteries. Although very little is known about the endotheliumdependent responses of cerebral vessels in type 2 diabetes, our findings are consistent with the recent demonstration that endothelium-dependent responses are impaired in cerebral blood vessels both in Zucker rats, a model of type 2 diabetes with hypertension [13,14], and in db/db mice, a genetic model of type 2 diabetes characterized by hyperinsulinemia, insulin resistance, hyperglycemia, and obesity [12].

NO reacts extremely efficiently with O2 − , resulting in a loss of NO bioavailability [4,9]. It has been reported that when O2 − reacts with NO, it produces the less-potent vasodilators, peroxynitrite, NO2 − , and NO3 − [23]. The chemical liability of NO in cells and tissues has been attributed to its rapid oxidation to both NO2 − and NO3 − . Indeed, it has been found that the primary inactivation product of NO in aerobic aqueous solution is NO2 − , and that further oxidation to NO3 − requires the presence of oxidizing species such as oxyhemoglobin and O2 − [25]. We recently published evidence that NO is metabolized by O2 − to NO3 − , not just to NO2 − , and that the consequent rapid inactivation of NO may be responsible for the impairment of endotheliumdependent relaxation seen in aortic strips from diabetic rats [7]. Furthermore, we found that although NO2 − may contribute to smooth muscle relaxation (presumably through NO formation), NO3 − was not able to relax aortic strips [7], and also that NO2 − -induced relaxation was impaired in the diabetic aorta [7]. Moreover, Modin et al. [26] suggested that (i) non-enzymatic reduction of inorganic nitrite to NO took place predominantly during acidic/reducing conditions, (ii) nitrite-evoked relaxation was effectively prevented by coadministration of an inhibitor of soluble guanylyl cyclase, (iii) this nitrite-induced relaxation was further potentiated by ascorbic acid (which effectively scavenges superoxide [27]). In the present study, NO2 − -induced relaxation was impaired in the diabetic (OLETF) basilar artery (Fig. 3), and this impaired relaxation was improved by both PEGSOD treatment and SOD-mimetic treatment [as also was the ACh-induced relaxation (Fig. 2)]. Moreover, superoxide pro-

22

T. Matsumoto et al. / Atherosclerosis 192 (2007) 15–24

duction of was increased in the OLETF basilar artery (Fig. 6). Taking all this together leads us to suggest that NO, which is generated by the endothelium and by nitrite, is rapidly inactivated by superoxide before reaching the target molecule (e.g., guanylyl cyclase) in smooth muscle. Thus, we suggest that the impairment of endothelium-dependent relaxation seen in basilar arteries from OLETF rats may be due to the presence of an abnormal oxidative metabolism of NO. Recent biochemical studies have shown that an NAD(P)H oxidase is the major source of O2 − in the vascular wall [17,19]. This vascular NAD(P)H oxidase shares several characteristics with the multicomponent enzyme complex reported in neutrophils [17,19]. The endothelial, the adventitial, and the neutrophil NAD(P)H oxidases consist of the flavocytochrome b558 subunits gp91phox (also known as Nox2) and p22phox as well as the cytosolic factors p47phox and p67phox and the small molecular weight G protein rac-1 [17,19]. Szocs et al. [28] recently demonstrated that increases in p22phox and Nox mRNA levels within vascular cells are associated with increased NAD(P)H oxidase activity. As far as extracranial vessels in diabetic states are concerned, NAD(P)H oxidase activity has been shown to be increased (in parallel with the levels of its subunit proteins) in the aorta in an animal model of diabetes [18]. We previously reported that the mRNA for the NAD(P)H oxidase p22phox subunit was significantly increased in the aorta of the STZ-induced diabetic rat [6]. In addition, gp91phox mRNA has been found to be upregulated in the steady state in the STZ-induced diabetic aorta [18]. These results are consistent with NAD(P)H oxidase playing a crucial role in the development of diabetic vascular complications. As far as cerebral blood vessels are concerned, O2 − production by canine basilar artery homogenates has been shown to be increased by the addition of exogenous NADH or NADPH, and this increased O2 − generation can be abrogated by diphenyl iodonium chloride, a flavoprotein inhibitor [29]. Furthermore, Didion and Faraci [24] suggested that NADH- and NADPH-induced changes in cerebral vascular tone are mediated by the O2 − produced by a flavin-containing enzyme (most likely NAD(P)H oxidase, but not xanthine oxidase or NOS). Recently, Ago et al. [30] suggested that the endothelial cells of rat basilar arteries have a unique set of Nox homologues, Nox1 being highly expressed along with Nox2 and Nox4, and these may be functionally active within the endothelial cells. Although both the expression and activity levels of NAD(P)H oxidase in the diabetic basilar artery are uncertain, we recently demonstrated that the expression of gp91phox mRNA is increased in the STZ-induced diabetic basilar artery [15]. It is known that apocynin (4 -hydroxy-3 -methoxy-acetophenone or acetovanilone), a non-toxic compound isolated from the medicinal plant Picrohiza kuroa, is a potent inhibitor of NAD(P)H oxidase. Apocynin has been reported to inhibit NAD(P)H oxidase activity by causing p47phox to dissociate from gp91phox [17,31]. In the present study, treatment of basilar artery rings from OLETF rats with apocynin improved (i) endothelium-dependent relaxation (Fig. 2), (ii) l-NNA-

induced contraction (Fig. 4), and (iii) ACh-induced cGMP production (Fig. 5). Moreover, the superoxide level in the presence of NADH was higher in basilar arteries from OLETF rats than in those from LETO rats (Fig. 6). Furthermore, the expression of gp91phox protein was found to be significantly increased in the OLETF basilar artery (versus control arteries) (Fig. 7), and the expression of p22phox protein tended to be elevated in the OLETF basilar artery (Fig. 7). Taken together, the above evidence strongly suggests that the defect in NO bioavailability seen in the type 2 diabetic basilar artery is attributable to increased oxidative stress, such as is associated with increased NAD(P)H oxidase activity. Although the mechanisms underlying such differences in NAD(P)H oxidase activities between OLETF and LETO rats remain unclear, metabolic and or hormonal alterations might be involved. It has been reported (a) that high glucose levels stimulate superoxide generation via PKC-dependent activation of vascular NAD(P)H oxidase in cultured aortic endothelial cells and smooth muscle cells [32] and (b) that in the rabbit basilar artery in high glucose solution in vitro, rapidly increasing ROS production, largely derived from NAD(P)H oxidase, reduces the relaxation to ACh [31]. Moreover, in an in vivo study Sonta et al. [33] noted a hyperglycemia-induced and insulin resistance-induced PKC-dependent activation of NAD(P)H oxidase that they felt might be attributable to the increased oxidative stress present in both diabetes and obesity. Taking the above evidence together, we speculate that in long-term hyperglycemic and insulin-resistant states in type 2 diabetes, these altered cellular mechanisms might activate NAD(P)H oxidase in the basilar artery. However, further investigation will be required on this point. Type 2 diabetes and insulin resistance increase the prevalence of cerebrovascular events, and if such patients suffer stroke they are subject to a more severe progression, a slower recovery, and a higher mortality. In addition, obesity and insulin resistance in older people are risk factors for dementia, particularly Alzheimer’s disease, which might be related to cerebrovascular dysfunction and chronic hypoperfusion of the brain. A recent study by Erdos et al. [13] demonstrated that cerebrovascular dysfunction is mediated by oxidative stress and PKC in Zucker obese rats, while Phillips et al. [34] revealed that in Zucker obese rats, the vasodilator reactivity of the middle cerebral artery to endothelium-dependent dilator stimuli is impaired, and they felt that this may represent a reduced bioavailability of signaling molecules due to oxidant scavenging. Furthermore, Didion et al. [12] provided evidence for superoxide-mediated impairment of endothelialdependent responses in cerebral vessels in type 2 diabetes, and they suggested that reductions in NO bioavailability are associated with increases in Rho-kinase activity. In addition, we have demonstrated here that the endothelial dysfunction in OLETF rat basilar arteries may be mediated by increased oxidative stress, and further provided evidence for this increased oxidative stress being mediated by increased NAD(P)H oxidase activity. Since the basilar artery makes an important contribution to cerebral vascular resistance, and

T. Matsumoto et al. / Atherosclerosis 192 (2007) 15–24

along with other cerebral vessels is a major determinant of local microvascular pressure within the cerebral circulation [1], our data may have important implications for the mechanisms by which type 2 diabetes contributes to the pathogenesis of cerebrovascular abnormalities. However, several limitations of the present study should be mentioned. Although we used OLETF rats, which manifest stable clinical and pathological features resembling human type 2 diabetes [10], this model also exhibits hypercholesterolemia, high levels of triglyceride, and hypertension. All of these risk factors can cause endothelial dysfunction. Hence, the contribution made by risk factors other than diabetes is difficult to determine in this model’s basilar artery. Further investigation is therefore required, for example of the endothelial dysfunction in the basilar artery present in models with these other risk factors. In conclusion, the present study suggests that NAD(P)H oxidase may play a key role in the impairment of endothelial function seen in cerebral blood vessels in OLETF rats. We believe that our findings should stimulate further interest in NAD(P)H oxidase as a therapeutic target in the continuing efforts to reduce diabetes-associated cerebrovascular complications.

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16] [17]

[18]

Acknowledgments [19]

We thank K. Wakabayashi, S. Yoshiyama, K. Miyamori, and E. Noguchi for technical help. We also thank Otsuka Pharmaceutical for providing LETO and OLETF rats. This study was supported in part by the Ministry of Education, Culture, Sports, Science and Technology, Japan, by the Promotion and Mutual Aid Cooperation for Private Schools of Japan, and by the Suzuken Memorial Foundation, Japan.

[20]

[21]

[22]

References [23] [1] Faraci FM, Heistad DD. Regulation of large cerebral arteries and cerebral microvascular pressure. Circ Res 1990;66:8–17. [2] Faraci FM, Heistad DD. Regulation of the cerebral circulation: role of endothelium and potassium channels. Physiol Rev 1998;78:53– 97. [3] Laakso M, Lehto S. Epidemiology of risk factors for cardiovascular disease in diabetes and impaired glucose tolerance. Atherosclerosis 1998;137:S65–73. [4] De Vriese AS, Verbeuren TJ, Van de Voorde J, et al. Endothelial dysfunction in diabetes. Br J Pharmacol 2000;130:963–74. [5] Kobayashi T, Taguchi K, Yasuhiro T, et al. Impairment of PI3-K/Akt pathway underlies attenuated endothelial function in aorta of type 2 diabetic mouse model. Hypertension 2004;44:956–62. [6] Kanie N, Matsumoto T, Kobayashi T, et al. Relationship between peroxisome proliferator-activated receptors (PPAR alpha and PPAR gamma) and endothelium-dependent relaxation in streptozotocin-induced diabetic rats. Br J Pharmacol 2003;140:23–32. [7] Kobayashi T, Kamata K. Effect of chronic insulin treatment on NO production and endothelium-dependent relaxation in aortae from established STZ-induced diabetic rats. Atherosclerosis 2001;155:313–20. [8] Matsumoto T, Kobayashi T, Kamata K. Alterations in EDHF-type relaxation and phosphodiesterase activity in mesenteric arteries from

[24]

[25]

[26]

[27]

[28]

[29]

[30]

23

diabetic rats. Am J Physiol Heart Circ Physiol 2003;285:H283– 91. Pieper GM. Review of alterations in endothelial nitric oxide production in diabetes: protective role of arginine on endothelial dysfunction. Hypertension 1998;31:1047–60. Kawano K, Hirashima T, Mori S, et al. Spontaneous long-term hyperglycemic rat with diabetic complications. Otsuka Long-Evans Tokushima Fatty (OLETF) strain. Diabetes 1992;41:1422–8. Minami A, Ishimura N, Harada N, et al. Exercise training improves acetylcholine-induced endothelium-dependent hyperpolarization in type 2 diabetic rats, Otsuka Long-Evans Tokushima Fatty rats. Atherosclerosis 2002;162:85–92. Didion SP, Lynch CM, Baumbach GL, et al. Impaired endotheliumdependent responses and enhanced influence of Rho-kinase in cerebral arterioles in type II diabetes. Stroke 2005;36:342–7. Erdos B, Snipes JA, Miller AW, et al. Cerebrovascular dysfunction in Zucker obese rats is mediated by oxidative stress and protein kinase C. Diabetes 2004;53:1352–9. Schwaninger RM, Sun H, Mayhan WG. Impaired nitric oxide synthasedependent dilatation of cerebral arterioles in type II diabetic rats. Life Sci 2003;73:3415–25. Matsumoto T, Yoshiyama S, Wakabayashi K, et al. Effects of chronic insulin on endothelial dysfunction of basilar arteries from established streptozotocin-diabetic rats. Eur J Pharmacol 2004;504:119–27. Faraci FM. Oxidative stress: the curse that underlies cerebral vascular dysfunction? Stroke 2005;36:186–8. Cai H, Griendling KK, Harrison DG. The vascular NAD(P)H oxidases as therapeutic targets in cardiovascular diseases. Trends Pharmacol Sci 2003;24:471–8. Hink U, Li H, Mollnau H, et al. Mechanisms underlying endothelial dysfunction in diabetes mellitus. Circ Res 2001;88:e14–22. Lassegue B, Clempus RE. Vascular NAD(P)H oxidases: specific features, expression, and regulation. Am J Physiol Regul Integr Comp Physiol 2003;285:R277–97. Matsumoto T, Sato A, Suenaga H, et al. Modulations of shear stressinduced contractile responses and agonist-induced vasodilation in hypercholesterolemic rats. Atherosclerosis 2004;175:31–8. Wang HD, Pagano PJ, Du Y, et al. Superoxide anion from the adventitia of the rat thoracic aorta inactivates nitric oxide. Circ Res 1998;82:810–8. Sugawara T, Fujii S, Zaman AK, et al. Coronary capillary network remodeling and hypofibrinolysis in aged obese diabetic rats: implications for increased myocardial vulnerability to ischemia. Mol Cell Biochem 2003;248:165–70. Beckman JS, Chen J, Ischiropoulos H, et al. Oxidative chemistry of peroxynitrite. Methods Enzymol 1994;233:229–40. Didion SP, Faraci FM. Effects of NADH and NADPH on superoxide levels and cerebral vascular tone. Am J Physiol Heart Circ Physiol 2002;282:H688–95. Ignarro LJ, Fukuto JM, Griscavage JM, et al. Oxidation of nitric oxide in aqueous solution to nitrite but not nitrate: comparison with enzymatically formed nitric oxide from l-arginine. Proc Natl Acad Sci USA 1993;90:8103–7. Modin A, Bjorne H, Herulf M, et al. Nitrite-derived nitric oxide: a possible mediator of ‘acidic-metabolic’ vasodilation. Acta Physiol Scand 2001;171:9–16. Jackson TS, Xu A, Vita JA, Keaney Jr JF. Ascorbate prevents the interaction of superoxide and nitric oxide only at very high physiological concentrations. Circ Res 1998;83:916–22. Szocs K, Lassegue B, Sorescu D, et al. Upregulation of Nox-based NAD(P)H oxidases in restenosis after carotid injury. Arterioscler Thoromb Vasc Biol 2002;22:21–7. Wambi-Kiesse CO, Katusic ZS. Inhibition of copper/zinc superoxide dismutase impairs NO-mediated endothelium-dependent relaxations. Am J Physiol Heart Circ Physiol 1999;276:H1043–8. Ago T, Kitazono T, Kuroda J, et al. NAD(P)H oxidases in rat basilar arterial endothelial cells. Stroke 2005;36:1040–6.

24

T. Matsumoto et al. / Atherosclerosis 192 (2007) 15–24

[31] Sercombe R, Vicaut E, Oudart N, et al. Acetylcholine-induced relaxation of rabbit basilar artery in vitro is rapidly reduced by reactive oxygen species in acute hyperglycemia: role of NADPH oxidase. J Cardiovasc Pharmacol 2004;44:507–16. [32] Inoguchi T, Li P, Umeda F, et al. High glucose level and free fatty acid stimulate reactive oxygen species production through protein kinase C-dependent activation of NAD(P)H oxidase in cultured vascular cells. Diabetes 2000;49:1939–45.

[33] Sonta T, Inoguchi T, Tsubouchi H, et al. Evidence for contribution of vascular NAD(P)H oxidase to increased oxidative stress in animal models of diabetes and obesity. Free Radic Biol Med 2004;37:115– 23. [34] Phillips SA, Sylvester FA, Frisbee JC. Oxidative stress and constrictor reactivity impair cerebral artery dilation in obese Zucker rats. Am J Physiol Regul Integr Comp Physiol 2005;288:R522– 30.