INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THE VASCULAR ENDOTHELIUM has an important role in the regulation of vessel reactivity. Furchgott and Zawadsky (4) first reported that acetylcholine-induced vascular relaxation of rabbit aorta was suppressed when the endothelium was mechanically removed and identified the generation of a labile vasodilating substance by the endothelium, later named endothelium-derived relaxing factor (EDRF) (3). Subsequent studies identified EDRF as the free radical nitric oxide (NO ·) (21). In the cardiovascular system, a basal release of NO · is responsible for the physiological vasodilatory tone that regulates blood pressure and vascular flow to tissues, including the brain, heart, and lungs (19). In addition, this radical reduces platelet adhesive function (23) and blocks leukocyte activation (8) contributing to the homeostatic control of blood vessels.

Endothelial injury with loss of endothelium-dependent vasodilation occurs in a number of disease processes, including hypertension, diabetes, dyslipidemia, and postischemic reperfusion injury (33). In isolated coronary rings (13, 29, 31) and in vivo postischemic heart models of endothelial dysfunction (16), marked reduction of acetylcholine-dependent vasodilation was observed. These studies showed that vascular relaxation was impaired when endothelium-dependent agents were utilized, including the receptor-independent calcium ionophore A23187, but was normal when the coronary vessels were stimulated with direct vascular smooth muscle relaxants such as the NO · donor sodium nitroprusside. Because NO · is the major mediator of endothelium-dependent vasodilation and restores normal dilation of impaired vessels, alterations in NO · generation were suggested to account for endothelial dysfunction. Despite the fact that changes in NO · generation are thought to occur, the mechanisms of impaired NO · generation are not well understood.

Endothelial nitric oxide synthase (NOS III) is the major source of NO · in the vascular endothelium (25), and changes in its enzymatic function have been suggested to be the basis for alterations in NO · formation and endothelial dysfunction (13, 26). Other investigators have questioned the nature of changes in NO · generation by the dysfunctional endothelium. It has been suggested that NO · formation may be unchanged or increased but scavenged by concurrent superoxide (O₂·) generation (14, 26, 30, 36).

Triton X-100 detergent treatment has been used as a physiological model of endothelial dysfunction. In this study, we characterize the mechanism of the endothelial dysfunction induced by dilute Triton treatment. We characterize the role of alterations in NOS III activity, expression, and intracellular localization in the loss of endothelium-dependent reactivity and correlate these observations with measurements of NO · generation.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Isolated heart perfusion. Female Sprague-Dawley rats (250-300 g) were heparinized and anesthetized with intraperitoneal pentobarbital. The hearts were excised, the aorta was cannulated, and retrograde perfusion was initiated. Hearts were perfused at constant flow using Krebs bicarbonate buffer (in mM: 17 glucose, 120 sodium chloride, 25 sodium bicarbonate, 2.5 calcium chloride, 0.5 EDTA, 5.9 potassium chloride, and 1.2 magnesium chloride) and bubbled with 95% O₂-5% CO₂ gas at 37°C as described previously (28). To measure contractile function, a latex balloon was inserted into the left ventricular cavity and connected to a pressure transducer via a hydraulic line. The balloon was initially inflated to achieve an end-diastolic pressure of 8-14 mmHg. Pressures were recorded with a Gould recorder.

Study protocol. Hearts were preequilibrated with perfusate at constant pressure of 80 mmHg for 15 min and then coronary flow was measured. Flow was then maintained at this baseline value throughout the experiment by initiating perfusion at constant flow with a roller pump (Masterflex-Cole Parmer Instrument, Chicago, IL). After another 15 min of heart perfusion at constant flow, the vasodilatory effects of the endothelium-dependent vasodilators histamine and calcium ionophore A23187 and the direct smooth muscle relaxant nitroprusside were evaluated under basal conditions. Any hearts that failed to exhibit baseline vasoreactivity or NO · formation following histamine were excluded. All three vasodilating agents were infused by pump injection (Harvard Apparatus) and mixed with the perfusate just above the aortic valve at a rate calculated to deliver to the coronary circulation final concentrations of 10⁵ M histamine, 1 µM calcium ionophore, and 3 × 10⁶ M nitroprusside. In a set of preliminary experiments in which hearts were pretreated with the NOS inhibitor N-nitro-L-arginine methyl ester (L-NAME), 1 mM preinfusion for 5 min, it was observed that both the histamine- and calcium ionophore-induced vasodilations were totally blocked, whereas that from nitroprusside was not blocked.

Hearts were then allocated randomly to either saline (control group) or Triton X-100 (Triton group) bolus injections. Pilot experiments using Triton X-100 at concentrations of 1:1,000, 1:500, 1:400, 1:200, and 1:100 were performed to determine the lowest detergent concentration capable of abolishing coronary vasodilatory response to histamine and calcium ionophore without significant impairment of myocardial contractile function. A 1:400 or 0.25% Triton concentration was found to be the optimum dilution. Triton was injected as a short bolus over a 1-s period. In control hearts a comparable volume of saline was administered as a bolus. The effects of histamine, calcium ionophore, and nitroprusside on coronary perfusion pressure were then reassessed in both control and Triton-treated groups to determine the ability of the coronary vasculature to vasodilate after treatment.

EPR spectroscopy and spin trapping. Spin-trapping measurements of oxygen radical generation were performed with 5,5'-dimethyl-1-pyrroline-N-oxide (DMPO), whereas the complex Fe²⁺-N-methyl-D-glucamine (Fe-MGD) was used to trap NO · formation from the heart. DMPO was purified by double distillation, and fresh stock solutions of Fe-MGD (1:5) were prepared by adding ferrous ammonium sulfate to aqueous solutions of MGD. A sidearm in the perfusion line allowed direct infusion of Fe-MGD or DMPO along with histamine proximal to the coronary circulation. All solutions were infused by a Harvard pump and mixed with the perfusate at a rate calculated to deliver a final concentration of 1 mM Fe-MGD or 40 mM DMPO and 10⁵ M histamine. Hearts were infused with histamine and with either spin trap, after Triton or saline treatment, with effluent collected during the 90 s of infusion. The effluent was immediately transferred to electron paramagnetic resonance (EPR) quartz flat cells. EPR spectra of NO · trapped by Fe-MGD were recorded at room temperature with a Bruker ER 300 spectrometer operating at X-band, 9.77 GHz, with 100-kHz modulation frequency and a TM₁₁₀ cavity, as described previously (37). EPR spectra of oxygen radicals were obtained at room temperature with a microwave frequency of 9.77 GHz using 20 mW of microwave power and 0.5 G modulation amplitude. The microwave frequency and magnetic field were precisely measured using an EIP 575 microwave frequency counter and Bruker ER 035M NMR gauss meter.

Heart tissue homogenate preparation. Heart tissue homogenate was prepared for NOS III activity measurements and Western blotting studies as follows. Immediately upon preparation hearts were frozen, finely ground under liquid nitrogen, and suspended in 3 ml of ice-cold homogenizing buffer. This buffer consisted of Tris · HCl and Tris base (50 mM, pH 7.4) containing 0.1 mM EDTA, 0.1 mM EGTA, 12 mM mercaptoethanol, and protease inhibitors (2 mM phenylmethylsulfonylfluoride and 4 µM leupeptin). The suspension was homogenized (Homogenizer 5000; Omni International, Waterbury, CT) and centrifuged at 100,000 g for 60 min (L-5 50B Ultracentrifuge; Beckman Instruments, Fullerton, CA) at 5°C. The particulate fraction containing NOS III was subsequently washed in 3 ml of ice-cold KCl buffer (homogenizing buffer containing 1 M KCl) for 5 min, and the homogenate was centrifuged at 100,000 g for 30 min at 5°C. Finally, the pellet was resuspended in the same homogenizing buffer containing calmodulin (330 nM) and tetrahydrobiopterin (BH₄, 10 µM).

NOS III activity measurements. NOS III activity was measured by determining the conversion rate of L-[¹⁴C]arginine to L-[¹⁴C]citrulline. The heart tissue particulate fraction was assayed in Tris buffer (pH 7.4) containing NADPH and CaCl₂. The final concentrations in the reaction mixture were 3.0 mM NADPH, 200 µM CaCl₂, 30 µM EDTA, 30 µM EGTA, 100 nM calmodulin, and 3 µM BH₄. Reaction was initiated by the addition of purified L-[¹⁴C]arginine (317 mCi/mmol) to produce a 10 µM concentration and conversion conducted for 30 min at room temperature. Termination of the reaction was achieved with the use of 3 ml of ice-cold stop buffer (20 mM HEPES and 2 mM EDTA, pH 5.5). Experiments were also performed in the presence of either L-NAME (250 µM) or EGTA (5.0 mM) to determine NOS specificity and Ca²⁺ dependence. L-[¹⁴C]citrulline content was determined by liquid scintillation counting (1209 Racket Liquid Scintillation Counter; LKB Wallac, Finland) after filtering the reaction mixture through a column of the resin Dowex AG 50WX-8 (500 µl of the Na⁺ form). NOS activity was determined by subtracting total counts from L-NAME-blocked counts and normalized for protein content (Bradford method) and conversion time. It was observed that all measured NOS activity was Ca²⁺ dependent and inhibited by 5.0 mM EGTA, indicating that it was derived from NOS III. The isotope L-[¹⁴C]arginine was purified from DuPont/NEN stock and stored at 20°C. Duplicate incubations were performed for each sample.

Western blots of NOS III in whole heart homogenates. Control and Triton-treated heart homogenates were aliquoted into 100-µg samples and lysed by the application of 62.5 mM Tris · HCl buffer (pH 6.8) containing 25% glycerol, 2% SDS, and 5% -mercaptoethanol with heat denaturation at 96°C for 5 min. The samples were loaded into the wells of a 7.5% SDS-PAGE gel, and electrophoresis was carried out for 30 min at 200 V. Proteins were transferred to a nitrocellulose membrane and blocked with 5% milk solution prepared in Tris-buffered saline (TBS) with 0.05% (vol/vol) Tween 20 (TTBS). Rabbit polyclonal primary antibodies (1:500 dilution) raised against the polypeptides 1030-1209 of the human NOS III were incubated with the membranes overnight at 4°C followed by biotinylated goat anti-rabbit secondary antibodies and streptavidin-biotinylated alkaline phosphatase complex. Membrane color was developed for ~10 min using a 5-bromo-4-chloro-3-indolyl phosphate/nitroblue tetrazolium color development solution.

Immunohistochemical analysis. Five-micrometer cryostat sections of heart tissue were fixed in alumina-filtered acetone for 10 min and mounted on Histostik-coated slides. The slides were rinsed twice with TBS, incubated for 20 min in a humidity chamber with 0.1 mg/ml avidin followed by 0.05 mg/ml biotin, and quenched with 0.05% H₂O₂ (vol/vol). Heart sections were blocked with 0.5% dried milk and 1% normal goat serum for 10 min. Immunostaining was accomplished by sequential application of a mouse monoclonal antibody directed against the peptides 1030-1209 of the human NOS III (1:100 dilution), biotinylated goat anti-mouse Fc IgG, and streptavidin-alkaline phosphatase complex (pH 8.2). The chromogen Fast Red substrate was used for labeling. The slides were counterstained with Mayers modified hematoxylin, dehydrated in increasing concentrations of ethanol, cleared with xylene, and coverslipped in Permount. As negative controls, heart sections were incubated with secondary and tertiary antibodies in the absence of the primary.

Transmission electron microscopy. Control and Triton X-100-treated hearts were perfusion fixed with 2% paraformaldehyde and 2% glutaraldehyde in 0.1 M Millonig buffer (pH 7.4) for 10 min, cut into 1-mm cubes, and immersed in the same fixative solution overnight. Samples were microwave postfixed in KFeCN-reduced 1% osmium tetroxide and stained in uranyl acetate for 15 min. The heart tissue was then quickly dehydrated through a graded series of ethanol, infiltrated with Eponate in propylene oxide, and cured overnight at 50°C. Ultrathin sections were sliced on a low-angle Diatome diamond knife and collected on 400-mesh formvar-coated nickel grids. Sections were stained with lead citrate.

Immunoelectron microscopy. For immunogold labeling, hearts were processed as described for standard electron microscopy. Grids were floated in 1% sodium periodate and incubated with 0.14% glycine in TBS for 15 min. Samples were blocked with 5% bovine serum albumin in TTBS and incubated with mouse anti-NOS III monoclonal antibody (1:100) overnight at room temperature. Matched control experiments were performed with identical processing but in the absence of the primary anti-NOS III antibody. Grids were reblocked and floated on TBS solution containing 12-nm gold-conjugated goat anti-mouse IgG (1:50) for 1 h. Staining was performed with uranyl acetate only. All samples were viewed and photographed in a Zeiss 10B transmission electron microscopy electron microscope operating at 60 kV.

Chemicals and reagents. Histamine, calcium ionophore A23187, sodium nitroprusside, L-NAME, and other chemicals and reagents were from Sigma Chemical (St. Louis, MO) unless noted otherwise. The Mn-SOD mimetic SC-55858, which reaches intracellular sites, was obtained from Monsanto. L-[¹⁴C]arginine was obtained from NEN/DuPont (Boston, MA). Protein assay reagents and secondary and tertiary antibodies were from Bio-Rad (Hercules, CA). N-methyl-D-glucamine and carbon disulfide required for Fe-MGD synthesis and DMPO were purchased from Aldrich (Milwaukee, WI). S-nitroso-N-acetylpenicillamine was from BIOMOL Research Laboratories (Plymouth Meeting, PA). Monoclonal and polyclonal NOS III primary antibodies were purchased from Transduction Laboratories (Lexington, KY) and electron microscopy Gold complexes from Jackson Immunoresearch Laboratories (West Grove, PA).

Statistical analysis. All values are expressed as means ± SE, and statistical significance of difference between the means was evaluated by Student's unpaired t-test with equal variance. A P value of 0.05 or less was considered to indicate statistical significance.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Vasomotor response to histamine, calcium ionophore A23187, and nitroprusside. Similar reduction of the coronary pressure was observed during infusion of histamine, calcium ionophore, or nitroprusside under basal conditions (Table 1). Triton infusion led to a significant increase in coronary perfusion pressure, which was not observed in the control saline-treated groups (control: 81.4 ± 0.8 mmHg; Triton: 97.9 ± 1.2 mmHg; P < 0.0001). In addition, Triton bolus injection almost completely abolished the vasorelaxant effects of the endothelium-dependent vasodilators histamine and calcium ionophore while preserving the response to nitroprusside (Fig. 1, Table 1). In the control group response to vasodilators after infusion of saline solution was comparable to basal levels.

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Effect of histamine, A23187, and nitroprusside on coronary perfusion pressure before and after Triton treatment. Data are expressed as percent decrease in coronary pressure induced by each vasodilator. Triton X-100 0.25% abolished histamine- and A23187-induced coronary vasodilation but did not modify vasorelaxant response to nitroprusside, suggesting that detergent treatment produced endothelial but not vascular smooth muscle dysfunction (*P < 0.001).

EPR measurement of NO · and oxygen radicals. No EPR spectrum is observed from Fe-MGD in the absence of NO · (Fig. 2A). However, after addition of the NO · donor compound, S-nitroso-N-acetylpenicillamine, a prominent EPR triplet signal is observed with a central g-value of 2.04 and hyperfine splitting of 13.2 gauss (Fig. 2B). In the coronary effluent of control hearts a distinct triplet NO · adduct signal was seen after histamine infusion (Fig. 3B), whereas no histamine-induced signal was detected in the coronary effluent of hearts pretreated with the NOS inhibitor L-NAME (1 mM) (Fig. 3C). In Triton-treated hearts, the histamine-stimulated NO · formation was abolished, indicating that decreased NO · formation contributes to the endothelial dysfunction seen (Fig. 3D). This loss of NO · was not reversed in the presence of Cu,Zn superoxide dismutase (SOD), 200 U/ml.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Electron paramagnetic resonance (EPR) spectra of Fe2+-N-methyl-D-glucamine (MGD) in presence and absence of nitric oxide (NO ·). A: spectrum of 1 mM Fe2+, 5 mM MGD in 50 mM HEPES buffer, pH 7.4. B: spectrum after incubation with a 2 mM concentration of NO · donor S-nitroso-N-acetylpenicillamine. Spectra were recorded at room temperature with microwave frequency of 100 kHz using 1 mW microwave power and a modulation amplitude of 4.0 G. Each spectrum is a 60-s spectral acquisition of 100 G sweep width with a time constant of 0.32 s. Note spectrum in A is shown at 10-fold higher gain than that in B.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   EPR spectra of NO · trapping in coronary effluent of saline- and Triton-perfused hearts labeled with Fe-MGD. A: coronary effluent of a saline-perfused heart under basal conditions. B: coronary effluent of a saline-perfused heart stimulated with 105 M histamine showing a typical triplet NO · adduct signal. C: perfusion of control hearts with NO synthase (NOS) III inhibitor N-nitro-L-arginine methyl ester (L-NAME 1 mM) for 10 min before histamine administration blocks NO · generation from endothelium. D: infusion of Triton X-100 0.25% completely suppresses NO · signal in heart effluent. Spectra were recorded with a microwave frequency of 9.77 GHz using 20 mW microwave power and a modulation amplitude of 4.0 G. Each spectrum was obtained from the sum of 20 60-s spectral acquisitions of 100-G sweep width with a time constant of 0.32 s.

Because failure to detect NO · could alternatively be due to scavenging by oxygen radicals such as superoxide, measurements of oxygen radical generation were performed using the spin-trap DMPO. EPR spectra of the effluent from control hearts under basal conditions showed no detectable oxygen radical generation. As previously reported (34, 35), a characteristic 1:2:2:1 signal indicative of DMPO-OH was observed during conditions that stimulate oxygen radical generation of reflow following 30 min of global heart ischemia, and this signal was quenched by SOD (200 U/ml) or by the SOD mimetic SC-55858 (20 µM). However, no EPR signal was seen following Triton treatment. Furthermore, treatment with SOD or SC-55858 did not prevent the Triton-mediated alterations in vasomotor response. This indicates that endothelial impairment was produced without any stimulation of O₂· or other oxygen radicals.

NOS III activity. The conversion rate of L-[¹⁴C]arginine to L-[¹⁴C]citrulline in homogenates of hearts pretreated with Triton was unchanged compared with homogenates of control hearts (control: 0.37 ± 0.01 pmol · min¹ · g protein¹; Triton: 0.36 ± 0.01 pmol · min¹ · mg protein¹; P = not significant), demonstrating that NOS III catalytic function is preserved after Triton treatment while reduced NO · synthesis and endothelial dysfunction are present.

Western blots and immunohistochemical analysis of NOS. Color development of the nitrocellulose membrane showed a protein of ~135 kDa consistent with the molecular weight of NOS III (Fig. 4). NOS III bands of similar size and density were observed in control and Triton-treated hearts, suggesting that treatment of the hearts with the low concentrations of the detergent used did not remove substantial amounts of NOS III from endothelial cells and myocytes.

View larger version (39K): [in this window] [in a new window]   Fig. 4.   Western blot analysis of NOS III protein in extracts of control and Triton-treated hearts. NOS III bands of similar size and density were seen in control (lane B) and Triton-treated hearts (lane C), suggesting that the enzyme was not removed from the heart tissue by Triton X-100 treatment despite partial solubilization of membranes induced by detergent treatment. Rat aortas were used as positive control for NOS III detection (lane A). Band at 135 kDa corresponds to that of endothelial NOS (eNOS), whereas band at 80 kDa corresponds to that of tissue alkaline phosphatase (AP).

Comparable results were obtained from immunohistology studies. The use of NOS III monoclonal antibodies in sections of control rat ventricles produced typical staining of the endothelium of arterioles, capillaries, and venules along with weak staining of myocytes. This staining pattern was maintained after subjecting hearts to Triton treatment (Fig. 5).

View larger version (94K): [in this window] [in a new window]   Fig. 5.   Immunohistochemical detection of NOS III in sections of ventricular myocardium of saline- and Triton-treated hearts. A: strongly positive staining of endothelium of arterioles and capillaries along with diffuse light staining of myocardial cells in presence of anti-NOS III antibody is seen in control hearts. B: same pattern is noted for hearts treated with Triton X-100. C: in absence of NOS III monoclonal antibody, only hematoxylin background staining is observed in heart tissue. Magnification of ×100 for A and C and ×120 for B.

Electron microscopy. Control heart sections displayed typical arrangement of the tissue structure. At higher magnification, endothelial plasma membrane and intracellular organelles appear normal (Fig. 6A). In contrast, brief exposure to Triton caused partial solubilization of the luminal surface of the plasma membrane of capillary endothelial cells (Fig. 6B). Subendothelial myocytes remained intact after detergent treatment. A similar pattern of partial solubilization of the luminal surface of the plasma membrane was also seen in arterioles, whereas underlying smooth muscle was intact. Thus the Triton treatment caused only mild disruption of the endothelial plasma membrane without significant myocyte injury.

View larger version (76K): [in this window] [in a new window]   Fig. 6.   Transmission electron microscopy of endothelial cells of control and Triton-treated hearts. A: sharp outline of plasma, nuclear, and intracellular membranes of capillary endothelial cell is evident in control hearts stained with lead citrate. B: after Triton treatment partial solubilization of plasma membrane and intracellular vesicles occurs with no damage to nuclear membrane. Magnification ×10,000.

Immunoelectron microscopy with the primary anti-NOS III antibody exhibited prominent deposition of the 12-nm gold particles, indicating the presence of NOS III in both endothelial cells and myocytes (Fig. 7A). In the absence of the primary antibody, no significant gold particle deposition was seen. Despite mild injury imposed by Triton to the cell membrane, NOS III labeling was largely preserved in endothelial cells. No change in NOS III distribution between plasma and intracellular membranes was seen. Myocardial cells also showed no change in labeling (Fig. 7B). Therefore, the mild membrane solubilization observed after detergent exposure did not remove membrane-bound NOS III or alter the enzyme distribution in the cells of the heart.

View larger version (85K): [in this window] [in a new window]   Fig. 7.   Immunoelectron microscopy of NOS III in rat heart ventricular myocardium of control and Triton-treated hearts. A: labeling of endothelial cell plasma and intracellular membranes along with diffuse staining of myocytes is present in control hearts. B: despite partial solubilization of membranes, staining is largely unchanged after Triton X-100 0.25% infusion. Magnification ×6,000.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Endothelial dysfunction occurs in association with hypertension (22), diabetes mellitus (7), atherosclerosis (11, 20), and postischemic injury. Studies in humans have shown not only reduced vasodilation in response to acetylcholine but also paradoxical vasoconstriction (11). This abnormal vasodilatory response is thought to be involved in the mechanisms that precipitate thrombus formation leading to ischemia and necrosis (17). In this way, endothelial dysfunction contributes to the development of myocardial infarction and stroke.

Despite the pathophysiological significance of endothelial dysfunction, the basic mechanisms responsible for this process remain unclear. Because NO · is the primary mediator of endothelium-dependent vasodilation, either an impairment of NO · generation or NO · activity at the level of the vascular smooth muscle cell must occur. It has been reported that the response to activators or inhibitors of NOS is decreased after reflow, whereas reactivity to exogenous NO · from NO · donors is maintained, suggesting that endothelial NO · synthesis is impaired (13, 26). Investigations in postischemic models also showed direct scavenging of NO · by O₂· formed upon reflow (6, 24) and a protective effect of SOD on endothelial function (15, 29). Thus questions remain regarding the basic mechanisms of impaired vasoreactivity following endothelial injury and the presence, significance, and cause of alterations in the NO · synthetic pathway.

In this study, vascular endothelium was subjected to mild injury by intracoronary administration of a brief bolus of Triton X-100 detergent, which has been used as an experimental model of endothelial injury (2, 9). The use of low Triton concentrations altered vascular reactivity with little if any impairment in cardiac contractile function. Suppressed receptor- and nonreceptor-mediated, endothelium-dependent responses as well as preserved endothelium-independent vasodilation was observed. Endothelial plasma membrane damage was seen by electron microscopy in detergent-treated hearts. Triton caused membrane damage by its direct effect of solubilizing lipids rather than by oxygen radical-based lipid peroxidation that occurs in postischemic myocardium.

We observed that histamine-induced NO · formation measured in the coronary effluent of Langendorff-perfused rat hearts was abolished after Triton treatment. The characteristic triplet EPR signal on direct trapping with Fe-MGD found in the control group was no longer present in Triton-treated hearts, indicating that endothelial dysfunction was due to a loss of NO · synthesis. In addition to abolishing stimulated vasodilatory response to endothelium-dependent dilators, Triton induced a sustained increase in basal coronary perfusion pressure. This pressure increase suggests a reduction in the basal release of NO · from endothelial cells, although it may also result from a dysfunction of one of the other endothelium-dependent vasodilatory mechanisms such as decreased release of prostaglandins or endothelium-dependent hyperpolarizing factor that contribute to the maintenance of the basal vasodilatory tone (12). Preservation of heart function after Triton injection and only minor perivascular edema without compression of the vessel lumen, along with maintained response to endothelium-independent agents, confirm that the vasoconstriction and loss of reactivity is endothelium dependent.

NOS III located in the coronary endothelium accounts for most of the NO · production in the heart, and the regulation of its enzymatic activity controls the amounts of this endogenous vasodilator (25). NOS III is a membrane-bound enzyme, and because of Triton's ability to solubilize the plasma membrane, the mechanism of decreased NO · production could be due to loss of this enzyme. However, measurements of NOS III in the heart indicated that there was no loss of enzyme after detergent treatment. Normal NOS III activity was detected demonstrating that the enzyme is functionally active in the presence of added substrates and cofactors. Western blots showing bands of comparable size and density, immunohistology showing similar staining patterns, and immunoelectron microscopy displaying equal antibody labeling also demonstrate that the amounts and localization of the enzyme in the heart were unchanged by Triton treatment. From these observations it is clear that reduced NO · synthesis did not require loss of NOS III but was due to depletion of requisite substrates or cofactors. Lack of response to the calcium ionophore A23187, which elicits NO · release by promoting calcium influx into the endothelial cells, excluded the possibility of an impairment in signal transduction at the membrane level.

NO · synthesis occurs through the oxidation of the amino acid L-arginine by NOS III to yield L-citrulline plus NO ·. This oxidation involves the transfer of electrons among NADPH, FAD, FMN, tetrahydrobiopterin, and heme. Ca²⁺ and the Ca²⁺-binding protein calmodulin are also required (10, 18). Depletion of one or more of these substrates and cofactors would impair NO · synthesis. Prior publications in reperfused hearts (32) and hypercholesterolemic patients (1) reported restoration of endothelium-dependent vasodilation after L-arginine infusion. It is probable that Triton-induced membrane damage leads to leakage of both substrates such as L-arginine and cofactors such as BH₄ through the plasma membrane as occurs with other intracellular compounds. The lack of response to the calcium ionophore suggests that calcium deficiency is not involved in the genesis of reduced NO · production in this model.

Whereas Triton-induced endothelial injury results in a model with endothelial dysfunction, the particular alterations observed do not necessarily reflect those that occur in any given disease process. The knowledge obtained in the present set of studies provides a characterization of the endothelial injury and dysfunction caused by membrane damage as induced by mild detergent treatment. Whereas knowledge from this study should be useful in understanding possible alterations that occur in disease pathophysiology where endothelial membrane injury occurs or endogenous membrane phospholipids are increased, they may not equate with those in any given disease.

In summary, impaired endothelial reactivity was induced by infusion of low concentrations of Triton X-100. This endothelial dysfunction occurred without stimulation of O₂· radical generation and resulted from decreased NO · synthesis. Reduced NO · synthesis and endothelial dysfunction occurred in the presence of fully active NOS III, indicating that substrate and/or cofactor depletion are responsible for endothelium-derived vascular impairment. In disease processes with endothelial injury accompanied by cell membrane damage, a similar process of altered endothelial function and decreased NO · formation may occur due to loss of NOS substrates or cofactors rather than the enzyme. Therefore, supplementation of these NOS substrates and cofactors might be efficacious in restoring endothelium-dependent vasoreactivity in these settings.