INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

FOLLOWING EARLY WORKS that identified xanthine oxidase as the main source of superoxide in endothelial cells submitted to hypoxia-reoxygenation (32), it was proposed that an NAD(P)H oxidase accounts for most of the reactive oxygen species (ROS) produced in the normoxic vessel wall (8). Although a number of critical features are still in question, most reports agree that this vascular oxidase is a flavoenzyme (8) associated with membrane or microsomes (16, 21). This enzyme(s) has been reported to share some similarity with the phagocytic NADPH oxidase regarding its structure (11, 19, 30), without, however, being identical (26). The recent cloning of homologues of the phagocytic NADPH oxidase gp91^phox subunit in nonphagocytic tissues raises the interesting possibility of the existence of a family of superoxide-generating enzymes with different structure and functional features but sharing some homology to the phagocytic enzyme (13).

The chemical identity of this enzyme or enzymes, however, has not yet been established. Nevertheless, much attention has been focused on its function under physiological and pathological conditions. The phagocytic NADPH oxidase reduces molecular oxygen to superoxide, using NADPH as an electron donor (2), and it has an estabilished function in immune defense. Questions remain regarding the physiological role for the vascular enzyme. There is also considerable controversy regarding some of its fundamental functional properties (8). The preferential substrate for the vascular NAD(P)H oxidase (NADH or NADPH), the ROS generated (superoxide or hydrogen peroxide), and the rate at which they are produced has not yet been clearly established.

It has been proposed that the vascular oxidase generates primarily superoxide, which then participates in a redox signaling system, transducing signals intracellularly, for example, from angiotensin II or growth factor receptors (7, 30).

The few studies that tried to quantitate the vascular oxidase output, however, reported that superoxide generation by this enzyme is very high (3, 5, 18, 21, 31), which could be toxic to cells, instead of allowing the oxidase to participate in signaling processes. Most of these studies, however, used methods that are not appropriate for accurate quantitative measurement of superoxide generation.

Therefore, the main objectives of this study were to obtain a more reliable estimate of the ROS generated by the vascular NAD(P)H oxidase, as well as to define its substrate utilization and preference. A novel electron paramagnetic resonance (EPR) spectroscopy spin-trapping method was developed to provide a more specific and quantitative measurement of superoxide generation. The stoichiometry of the substrates consumed in the process of ROS generation by this enzyme(s) was also determined.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Chemicals. The cell membrane-permeant superoxide dismutase (SOD) mimetic M40403 was obtained from MetaPhore Pharmaceuticals (St. Louis, MO) (26). The spin-trap 5,5'-dimethyl-pyrroline-N-oxide (DMPO) was obtained from Dojindo Laboratories (Kumamoto, Japan). SOD from bovine erythrocytes, diphenyleneiodonium (DPI), rotenone, NADH and NADPH, xanthine and xanthine oxidase were purchased from Sigma Chemicals (St. Louis, MO) and stored and diluted according to the manufacturer's directions.

EPR experiments with intact aortas. Male Sprague-Dawley rats (200-250 g) were euthanized with intraperitoneal injection of pentobarbital (100 mg/kg). The thoracic aorta was removed, freed from loose connective tissue, and cut in six fragments, ~5 mm long. Each one of these six aortic rings was placed in a distinct centrifuge tube containing DMPO (50 mM), rotenone (60 µM), diethylenetriamine pentaacetic acid (DTPA, 0.1 mM), and NADPH or NADH (0.1 mM); phosphate-buffered saline (PBS) was added to a final volume of 0.5 ml. In some experiments the SOD mimetic M40403 (40 µM) was added. The rings were maintained in this solution at 37°C. Ten minutes after addition of NADPH or NADH, one tube was taken out of the incubator, the ring was removed, and the tube with the spin trap was frozen in liquid nitrogen. The same process was repeated after 20, 30, 40, 50, and 60 min after addition of NADPH or NADH. On the same day, each sample was rapidly thawed and transferred to a quartz flat cell, and EPR spectra were recorded at room temperature (23°C) with a Bruker ER 300 spectrometer operating at X-band with a TM₁₁₀ cavity with the use of a modulation frequency of 100 kHz, modulation amplitude of 0.5 G, microwave power of 20 mW, and microwave frequency of 9.78 GHz as described (23, 25, 32). Ten serial 60-s acquisitions were accumulated to obtain the final spectrum.

Vessel homogenate and particulate fraction preparation. The thoracic aortas from 6 to 12 rats were excised, freed from periadventitial and loose connective tissue, and placed in cold Tris · HCl buffer (50 mM, pH = 7.40) containing mercaptoethanol (0.1%), phenylmethylsulfonyl fluoride(1.0 mM), and protease inhibitors (Complete Mini, Boehringer-Mannheim). The vessels were minced under liquid nitrogen, collected with 1 ml of the same buffer, sonicated 10 times (3 s each) on ice at medium power, and centrifuged at 1,000 g for 10 min at 4°C. The supernatant was collected and transferred to a 4.0-ml ultracentrifuge tube. It was then centrifuged at 20,000 g for 15 min at 4°C. The supernatant was then centrifuged at 100,000 g for 45 min at 4°C. A small pellet was harvested from the bottom of the centrifuge tube and resuspended in buffer. Protein was quantified by the Bradford method. This preparation yields the particulate fraction containing mainly membranes and microsomes (21). No mitochondria were detected in this particulate fraction with the use of the protein Bcl-2 as a marker for this organelle (data not shown). Also, rotenone did not inhibit oxygen consumption by the particulate fraction, either under basal conditions or after exposure to NADPH or NADH (data not shown).

Quantitation of DMPO-hydroxyl adduct. Quantitation of the observed DMPO-hydroxyl (DMPO-OH) spin adduct was performed by simulation fitting, as previously described (12). This procedure provides accurate quantitation of intensity and minimizes errors due to noise or baseline drifts. First, any baseline drift was removed by linear baseline correction. The DMPO-OH adduct spectrum was then simulated, using the observed nitrogen and hydrogen hyperfine splitting values of a_N = a_H = 14.9 G, to match the corresponding measured spectrum. The simulated spectrum was calibrated with respect to an appropriate aqueous solution of the nitroxide 2,2,6,6-tetramethylpiperidine-1-oxyl, 1 µM, measured under identical conditions (12).

Quantitation of superoxide formation. To calculate the concentration of superoxide formed in a given reaction system from the concentration of an observed spin-trap adduct, it is necessary to correct for the rate of adduct decay and then the efficiency of the trap (23, 24). The concentration of the observed spin adduct (A), determined by simulation fitting, can be defined as a function (f₁) of time (t) (see Fig. 1A), and the rate of the spin adduct concentration change can be expressed by

(1)

(as shown in Fig. 1B). The spin adduct DMPO-OH, however, decays to a nonradical (and therefore EPR undetectable) species. Thus the actual rate of the spin adduct formation is given by the rate of change of concentration plus the rate of spin adduct decay.

View larger version (25K): [in this window] [in a new window]   Fig. 1.   Graphs illustrating the process of superoxide quantitation based on measurement of the rates of spin-trap adduct formation and decay. A: experimental values of 5,5'-dimethyl-pyrroline-N-oxide hydroxyl (DMPO-OH) adduct formation measured over time (; a) and its fitted function (dashed line). Experimental values for decay of DMPO-OH adduct measured over time (; b) and its fitting (solid line) are also shown. Data are based on a typical example obtained with aortic particulate fraction in the presence of 0.1 mM NADPH. B: first derivative over time of functions a and b were calculated to determine the rate of DMPO-OH formation vs. time (function line c, dashed line) or rate of DMPO-OH decay vs. time (function line d, solid line). C: for each adduct concentration in A, the correspondent rate of adduct formation or decay was determined from B. In function line e, the data from function lines a and c are transformed to graph the rate of DMPO-OH adduct formation vs. concentration, and in line f, the data from function lines b and d were transformed to graph the rate of DMPO-OH adduct decay vs. concentration. In function line g, the corrected rate of DMPO-OH formation for a given concentration was calculated by subtracting the rate of DMPO-OH adduct decay (function line f) from the rate of formation (function line e). Data from function line g were used to calculate the corrected rate of DMPO-OH formation rate vs. time (open circles), and a function was fitted to these values (function line h, dashed line). The integral of this function shows the total amount of DMPO-OH formed over time (function line i, solid line). With adjustment for the efficiency of superoxide-mediated adduct formation, as described in the text, a quantitative estimate of the amounts and rates of superoxide formation are obtained.

(2)

(3)

(4)

1

1

1

1

Oxygen consumption. Electrochemical measurements of oxygen concentration were performed using an electrochemical oxygen probe, which consists of a working electrode and a reference electrode, connected to a CHI 832 electrochemical detector (CH Instruments). The probe was sealed in a water-jacketed chamber (Gilson) containing 1.5 ml of buffer solution. No head space was allowed above the buffer solution except a small hole for injecting samples. The potential on the working oxygen electrode was held at 0.7 V versus the reference electrode. After baseline stablization, samples were added to the buffer solution, and the current was measured to determine the change in O₂ concentration. Calibration of the probe was performed by adding air-equilibrated solution to the deaerated buffer solution.

NADH and NADPH oxidation. NADPH and NADH oxidation were followed spectrophotometrically at 340 nm in a Cary 300 Bio (Varian Instruments). The reactions observed contained the particulate fraction of the rat aortas (protein concentration = 0.02 mg/ml), diluted to a final volume of 1.0 ml in phosphate buffer, pH = 7.40. NADPH or NADH (0.05 mM) was added, just before the measurements were started, and followed for at least 60 min. The experiments were run at room temperature. The rate of NADPH and NADH oxidation was calculated using a molar extinction coefficient of 6.22 mM¹ · cm¹.

Statistical analysis. All data are expressed as means ± SE. Comparisons among groups were performed by Student's t-test or one-way ANOVA, with Student-Newman-Keuls tests. The significance level was 0.05. The Primer of Biostatistics computer program was used (version 3.01, 1992, McGraw-Hill).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Characterization of NAD(P)H oxidase from particulate fraction. Free radical generation by the particulate fraction of rat aortas, which contains mainly membranes and microsomes, was studied regarding the species produced and the substrates consumed. This fraction was chosen because it has been previously reported that the main site of superoxide generation in the vessel wall is a membrane-bound oxidase, which catalyzes the reduction of oxygen to superoxide using NADPH or NADH as electron donors (8).

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Electron paramagnetic resonance (EPR) spin-trapping measurements of rat aortic particulate fraction. Under basal conditions, no signal was detected (A). Addition of NADPH (B, left) or NADH (B, right) triggered the formation of a characteristic DMPO-OH adduct derived from superoxide as demonstrated because it is abolished by superoxide dismutase (SOD) (C) but not affected by catalase (D). Spectra were obtained 60 min after addition of NADPH or NADH and are representative of at least 3 experiments in each group.

1

1

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Superoxide generation, oxygen consumption, and NADPH oxidation by particulate fraction. Under basal conditions no superoxide generation was detected from particulate fraction. After addition of NADPH, there was an increase in oxygen consumption, NADPH was oxidized, and superoxide was generated. Values shown are means of at least 5 experiments for each type of measurement.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Superoxide generation, oxygen consumption, and NADH oxidation by particulate fraction. Under basal conditions no superoxide generation was detected from particulate fraction. After addition of NADH, there was an increase in oxygen consumption, NADH was oxidized, and superoxide was generated. Values shown are means of at least 5 experiments for each type of measurement.

1

1

1

1

Characterization of NAD(P)H oxidase in aortic rings. Radical generation by arterial rings or cultured cells (5, 27) when exposed to extracellularly added NADPH or NADH has been described previously. We observed that rat aortic rings incubated with rotenone (60 µM) showed negligible basal oxygen consumption and no detectable EPR signal when incubated with the spin-trap DMPO (50 mM). Exposure to NADH or NADPH increased the oxygen consumption by the aortic rings (Figs. 5 and 6), and a quartet EPR signal could be detected, characteristic of the DMPO-OH adduct. This signal was completely abolished by a membrane-permeable SOD mimetic, suggesting that these signals were derived from superoxide. As previously reported, the vascular NAD(P)H oxidase is the main source of the DMPO-OH adduct observed in aortic rings and inhibitors of other pathways, including mitochondrial respiration, xanthine oxidase, cyclooxygenase, and lipoxygenase, have no effect on this radical generation, whereas the commonly used NADPH oxidase inhibitor DPI totally inhibits this signal (26).

View larger version (12K): [in this window] [in a new window]   Fig. 5.   Oxygen consumption by aortic rings. Vessels incubated with rotenone (60 µM) consume negligible amounts of oxygen. After 10 min, NADH or NADPH (0.1 mM each) was added to the preparation. There was an increase in oxygen consumption by vessels, which is higher for NADPH, but not statistically significant. The average of at least 4 experiments is represented.

View larger version (12K): [in this window] [in a new window]   Fig. 6.   Superoxide generation by aortic rings. Exposure to NADPH or NADH (0.1 mM each) stimulates rat aortic rings to produce superoxide, which is trapped by DMPO and detected by EPR. The amount of superoxide produced by vessels was obtained after correction of the observed spectral data for the rate of spin-trap decay and efficiency of trapping. Data are the average of at least 4 experiments ± SE.

1

1

1

1

1

1

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Formerly, ROS were considered only effectors of tissue damage in pathological situations, such as ischemia-reperfusion, inflammation, or ionizing injury (6). Recently, however, it has been shown that exogenous administration of ROS stimulates many of the signaling pathways activated by stimulation of cell surface receptors with growth factors or agonists (8, 9). Therefore, it has been proposed that ligand-induced receptor activation increases intracellular ROS generation, which then plays a critical role in intracellular signaling.

Some studies suggest that the amount of ROS may be crucial to determine the fate of the cells where they are generated. Exposure of vascular smooth muscle cells to exogenously added superoxide (22) or H₂O₂ (28) was reported to stimulate cell growth. In higher concentrations, however, H₂O₂ can lead to cell apoptosis (4). Therefore, an attractive hypothesis is that the distinction between the role played by ROS, whether damage effector or second messenger, can be related to the amount of radicals produced by the tissue in response to a given stimulus. Although this issue is not addressed directly in this work, a reliable measurement of the ROS produced by a biological system is imperative to help clarify this question. In vascular tissues, this feature is even more important, because some of the most important vascular diseases, like atherosclerosis (15) and restenosis following angioplasty (1), are reported to have oxidant injury as one of their pathogenic mechanisms.

In vascular tissues the main site of ROS generation has been reported to be an NAD(P)H oxidase, which reduces molecular oxygen to superoxide, using NADH or NADPH as electron donors (8). The structural and functional features of this enzyme, or group of enzymes, are still not completely elucidated; however, it has been described to participate in the transduction of angiotensin II-related effects (8), as well as in the activation of nuclear transcription factors (27). Given its importance in vascular function, several attempts have been made to quantitate the output of the vascular NAD(P)H oxidase. Unfortunately, most of these measurements have used lucigenin-amplified chemiluminescence as the method to measure superoxide. This method, under the conditions employed, has been reported to increase superoxide production by itself (14). Addition of NADH to the system as a method to detect NADH-dependent oxidase activity has also been reported to introduce another source of error to the measurements, because NADH can increase the redox-cycling of lucigenin (8, 10). Therefore, lucigenin-amplifed chemiluminescence does not provide a good quantitative approach to measure superoxide generation.

This fact is reflected by the great discrepancy observed in the previous reports that tried to quantify the vascular NAD(P)H oxidase output using this method (3, 5, 18, 21, 31). Reports of superoxide generation from particulate fractions stimulated with NADH or NADPH varied 1,000-fold from less than 1 nmol · mg protein¹ · min¹ (3) to more than 1 µmol · mg protein¹ · min¹ (21). Also, reports of superoxide generation measurements made in arterial rings varied more than 20-fold (3, 17).

Some prior measurements are difficult to compare, because different preparations were used. Some authors measured the vascular NAD(P)H oxidase activity in particulate fractions obtained from cultured cells, whereas others used particulate fractions from whole artery homogenates (3, 18, 21, 31). Also, distinct animal species and different methods of particulate fraction preparation were used. In common, all authors used lucigenin-amplified chemiluminescence to quantitate the activity of the vascular NAD(P)H oxidase, which has potential problems of altering the process of radical generation and thus cannot provide true quantitation of intrinsic radical generation.

EPR spin trapping is a specific method to detect free radicals (12, 14, 25, 32), and it has been previously used to quantitatively assess superoxide generation by enzymatic systems (24) and leukocytes (23). Here, this technique was extended to measure the amount of superoxide produced by the vascular NAD(P)H oxidase(s). DMPO was chosen as the spin trap, instead of the more recently developed 5-(diethoxyphosphoryl)-5-methyl-1-pyrroline-N-oxide (DEPMPO), because it provides a specific adduct spectrum, easily identifiable, which is easily simulated and quantitated through simulation-fitting methods. Also, this nitrone spin trap is relatively simple to synthesize and purify and therefore is much cheaper and more readily available than the more recently developed spin traps, such as DEPMPO, or 5-ethoxycarbonyl-5-methyl-1-pyrroline N-oxide (EMPO) where problems with purification have limited their commercialization. Nitrone spin traps, such as DMPO, have an advantage over reduced spin labels, like 1-hydroxy-3-carboxy-pyrrolidine (CP-H), because they provide a specific adduct, allowing more definitive identification of the radical trapped. Hydroxylamines such as CP-H are reduced forms of nitroxide spin labels and do not trap the radical formed, and therefore, they work as nonspecific detectors of redox state. Unfortunately, the use of nitrone spin traps such as DMPO in in vivo applications is complicated by reduction of the spin adducts by plasma reducing equivalents and complex processes of bioreduction, thus the approach reported here while useful in isolated vessel or vascular homogenates may not be readily extendable to in vivo applications.

In this study, it was initially confirmed that superoxide was produced by a particulate fraction-associated enzyme(s). After procedures to quantitate the experimental spectra and correction for the spin trap efficiency and spin adduct decay were used, we were able to determine the amount of superoxide released. Superoxide was generated by NADPH- or NADH-dependent oxidase(s) with a rate of 0.41 ± 0.03 and 0.36 ± 0.04 nmol superoxide · mg protein¹ · min¹, respectively, which is lower than those rates previously reported in equivalent preparations (3, 5, 17, 18, 21, 30, 31). In agreement with previous reports (8), the NAD(P)H oxidase(s) activity was characterized, in our experiments, by a sustained release of low concentrations of superoxide, in contrast to the rapid release of relatively large amounts of ROS from the respiratory burst oxidase of the phagocytes (2, 23).

To better characterize the oxidase function, the oxygen consumption and the NADPH or NADH oxidation were measured under the same experimental conditions. It was shown that the consumption of these substrates had a time course similar to that of superoxide generation; however, the values for oxygen consumption and NAD(P)H oxidation are approximately three times higher than superoxide production. This fact is not surprising, because the oxidase could directly reduce molecular oxygen-generating hydrogen peroxide, in addition to superoxide. In fact, for a variety of other superoxide-generating enzymes such as xanthine oxidase, only 30-35% of the oxygen consumed is typically reduced to superoxide, whereas the balance is reduced to H₂O₂ (20, 24). In addition, this particulate fraction may contain other enzymes, which could also consume oxygen, NADPH, or NADH, but still not generate superoxide. Hence, complete stoichiometric conversion of oxygen to superoxide would not be expected.

It is interesting to note that the superoxide generation was only slightly higher after addition of NADPH compared with NADH addition. Both dinucleotides (each containing an adenine and a nicotinamide ring) are used as substates or cofactors in many cellular oxidation-reduction reactions. The only structural difference between NADH and NADPH is the presence of an extra phosphate group at the 2'-position of the adenine nucleotide moiety of NADPH. This extra phosphate group is far from the active redox region (of the nicotinamide ring), and it carries no importance in the electron transfer reaction. It appears, however, to confer specificity for the enzymes to which NADPH can bind as a coenzyme (29). It has been reported previously that the vascular oxidase uses both NADPH (18) and NADH (16) as electron donors. We observed that both NADH and NADPH were oxidized by membrane fraction of rat aortas and that oxygen consumption induced by addition of either substrate is very similar. These data suggest that the vascular oxidase should have similar affinity to NADPH and NADH. Another explanation could be the presence of more than one enzyme in this preparation, each one oxidizing NADPH or NADH and producing similar amounts of superoxide in the reaction. The data presented in this paper cannot discriminate between these two hypotheses.

Also intact aortic rings, when exposed to NADPH or NADH, showed an increase in superoxide generation and oxygen consumption. Increase in mitochondrial respiration was not responsible for this phenomenon, because vessels were previously incubated with rotenone, which abolished the basal oxygen consumption, proving its effectiveness. Under these conditions, the rates of superoxide generation are around 6.5 or 5 times lower than oxygen consumption, after addition of NADPH or NADH, respectively. The higher complexity of the system, with oxygen being also consumed by other enzymes and dismutated by Cu,Zn-SOD, can be responsible for these differences. In the particulate fraction experiments there was no Cu,Zn-SOD present, because it is a cytosolic enzyme. Alternatively, Mn-SOD is mitochondrially localized and also should be absent. Previous measurements of superoxide production by aortic rings using lucigenin-amplified luminescence showed higher values, even when the vessels were not exposed to NADH or NADPH (3, 21). Although some variation may be expected, because distinct animal species were used, differences in the method used to quantify superoxide may account for much of this discrepancy.

The fact that extracellular addition of substrates to intact tissues would cause an increase in superoxide generation and oxygen consumption raises the question that this enzyme may have a different orientation from the phagocytic oxidase. In phagocytes, intracellular NADPH binds to cytochrome b₅₅₈, which, then, transfers the electrons to extracellular oxygen (2). Addition of exogenous NADPH to phagocytes does not cause an increase in their superoxide production (unpublished observations). The mechanism of the vascular oxidase activity, thus, seems to be more complex. It is still not known whether the vascular oxidase(s) possesses an extracellular domain where NADH or NADPH can bind or whether they need to cross the membrane to reach intracellular binding sites.

In summary, the vascular NAD(P)H oxidase present in aortic segments and particulate fraction of aortic homogenates was studied, regarding the amount of superoxide that it produces. These measurements were correlated with rates of substrate oxidation and oxygen consumption. A lower rate of superoxide generation was observed compared with the values previously reported for similar preparations using less specific methods. The lower rates of superoxide generation measured in our experiments are more compatible with the proposed signaling role of this enzyme(s).